linux_dsm_epyc7002/arch/powerpc/kvm/book3s_64_mmu_hv.c

2156 lines
53 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
/*
*
* Copyright 2010 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
*/
#include <linux/types.h>
#include <linux/string.h>
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/highmem.h>
#include <linux/gfp.h>
#include <linux/slab.h>
#include <linux/hugetlb.h>
#include <linux/vmalloc.h>
#include <linux/srcu.h>
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
#include <linux/anon_inodes.h>
#include <linux/file.h>
#include <linux/debugfs.h>
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
#include <asm/kvm_ppc.h>
#include <asm/kvm_book3s.h>
#include <asm/book3s/64/mmu-hash.h>
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
#include <asm/hvcall.h>
#include <asm/synch.h>
#include <asm/ppc-opcode.h>
#include <asm/cputable.h>
#include <asm/pte-walk.h>
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
#include "trace_hv.h"
//#define DEBUG_RESIZE_HPT 1
#ifdef DEBUG_RESIZE_HPT
#define resize_hpt_debug(resize, ...) \
do { \
printk(KERN_DEBUG "RESIZE HPT %p: ", resize); \
printk(__VA_ARGS__); \
} while (0)
#else
#define resize_hpt_debug(resize, ...) \
do { } while (0)
#endif
static long kvmppc_virtmode_do_h_enter(struct kvm *kvm, unsigned long flags,
long pte_index, unsigned long pteh,
unsigned long ptel, unsigned long *pte_idx_ret);
struct kvm_resize_hpt {
/* These fields read-only after init */
struct kvm *kvm;
struct work_struct work;
u32 order;
/* These fields protected by kvm->arch.mmu_setup_lock */
/* Possible values and their usage:
* <0 an error occurred during allocation,
* -EBUSY allocation is in the progress,
* 0 allocation made successfuly.
*/
int error;
/* Private to the work thread, until error != -EBUSY,
* then protected by kvm->arch.mmu_setup_lock.
*/
struct kvm_hpt_info hpt;
};
int kvmppc_allocate_hpt(struct kvm_hpt_info *info, u32 order)
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
{
unsigned long hpt = 0;
int cma = 0;
struct page *page = NULL;
struct revmap_entry *rev;
unsigned long npte;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
if ((order < PPC_MIN_HPT_ORDER) || (order > PPC_MAX_HPT_ORDER))
return -EINVAL;
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
page = kvm_alloc_hpt_cma(1ul << (order - PAGE_SHIFT));
if (page) {
hpt = (unsigned long)pfn_to_kaddr(page_to_pfn(page));
memset((void *)hpt, 0, (1ul << order));
cma = 1;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
}
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
if (!hpt)
mm, tree wide: replace __GFP_REPEAT by __GFP_RETRY_MAYFAIL with more useful semantic __GFP_REPEAT was designed to allow retry-but-eventually-fail semantic to the page allocator. This has been true but only for allocations requests larger than PAGE_ALLOC_COSTLY_ORDER. It has been always ignored for smaller sizes. This is a bit unfortunate because there is no way to express the same semantic for those requests and they are considered too important to fail so they might end up looping in the page allocator for ever, similarly to GFP_NOFAIL requests. Now that the whole tree has been cleaned up and accidental or misled usage of __GFP_REPEAT flag has been removed for !costly requests we can give the original flag a better name and more importantly a more useful semantic. Let's rename it to __GFP_RETRY_MAYFAIL which tells the user that the allocator would try really hard but there is no promise of a success. This will work independent of the order and overrides the default allocator behavior. Page allocator users have several levels of guarantee vs. cost options (take GFP_KERNEL as an example) - GFP_KERNEL & ~__GFP_RECLAIM - optimistic allocation without _any_ attempt to free memory at all. The most light weight mode which even doesn't kick the background reclaim. Should be used carefully because it might deplete the memory and the next user might hit the more aggressive reclaim - GFP_KERNEL & ~__GFP_DIRECT_RECLAIM (or GFP_NOWAIT)- optimistic allocation without any attempt to free memory from the current context but can wake kswapd to reclaim memory if the zone is below the low watermark. Can be used from either atomic contexts or when the request is a performance optimization and there is another fallback for a slow path. - (GFP_KERNEL|__GFP_HIGH) & ~__GFP_DIRECT_RECLAIM (aka GFP_ATOMIC) - non sleeping allocation with an expensive fallback so it can access some portion of memory reserves. Usually used from interrupt/bh context with an expensive slow path fallback. - GFP_KERNEL - both background and direct reclaim are allowed and the _default_ page allocator behavior is used. That means that !costly allocation requests are basically nofail but there is no guarantee of that behavior so failures have to be checked properly by callers (e.g. OOM killer victim is allowed to fail currently). - GFP_KERNEL | __GFP_NORETRY - overrides the default allocator behavior and all allocation requests fail early rather than cause disruptive reclaim (one round of reclaim in this implementation). The OOM killer is not invoked. - GFP_KERNEL | __GFP_RETRY_MAYFAIL - overrides the default allocator behavior and all allocation requests try really hard. The request will fail if the reclaim cannot make any progress. The OOM killer won't be triggered. - GFP_KERNEL | __GFP_NOFAIL - overrides the default allocator behavior and all allocation requests will loop endlessly until they succeed. This might be really dangerous especially for larger orders. Existing users of __GFP_REPEAT are changed to __GFP_RETRY_MAYFAIL because they already had their semantic. No new users are added. __alloc_pages_slowpath is changed to bail out for __GFP_RETRY_MAYFAIL if there is no progress and we have already passed the OOM point. This means that all the reclaim opportunities have been exhausted except the most disruptive one (the OOM killer) and a user defined fallback behavior is more sensible than keep retrying in the page allocator. [akpm@linux-foundation.org: fix arch/sparc/kernel/mdesc.c] [mhocko@suse.com: semantic fix] Link: http://lkml.kernel.org/r/20170626123847.GM11534@dhcp22.suse.cz [mhocko@kernel.org: address other thing spotted by Vlastimil] Link: http://lkml.kernel.org/r/20170626124233.GN11534@dhcp22.suse.cz Link: http://lkml.kernel.org/r/20170623085345.11304-3-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: Alex Belits <alex.belits@cavium.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Christoph Hellwig <hch@infradead.org> Cc: Darrick J. Wong <darrick.wong@oracle.com> Cc: David Daney <david.daney@cavium.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Mel Gorman <mgorman@suse.de> Cc: NeilBrown <neilb@suse.com> Cc: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-13 04:36:45 +07:00
hpt = __get_free_pages(GFP_KERNEL|__GFP_ZERO|__GFP_RETRY_MAYFAIL
|__GFP_NOWARN, order - PAGE_SHIFT);
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
if (!hpt)
return -ENOMEM;
/* HPTEs are 2**4 bytes long */
npte = 1ul << (order - 4);
/* Allocate reverse map array */
treewide: Use array_size() in vmalloc() The vmalloc() function has no 2-factor argument form, so multiplication factors need to be wrapped in array_size(). This patch replaces cases of: vmalloc(a * b) with: vmalloc(array_size(a, b)) as well as handling cases of: vmalloc(a * b * c) with: vmalloc(array3_size(a, b, c)) This does, however, attempt to ignore constant size factors like: vmalloc(4 * 1024) though any constants defined via macros get caught up in the conversion. Any factors with a sizeof() of "unsigned char", "char", and "u8" were dropped, since they're redundant. The Coccinelle script used for this was: // Fix redundant parens around sizeof(). @@ type TYPE; expression THING, E; @@ ( vmalloc( - (sizeof(TYPE)) * E + sizeof(TYPE) * E , ...) | vmalloc( - (sizeof(THING)) * E + sizeof(THING) * E , ...) ) // Drop single-byte sizes and redundant parens. @@ expression COUNT; typedef u8; typedef __u8; @@ ( vmalloc( - sizeof(u8) * (COUNT) + COUNT , ...) | vmalloc( - sizeof(__u8) * (COUNT) + COUNT , ...) | vmalloc( - sizeof(char) * (COUNT) + COUNT , ...) | vmalloc( - sizeof(unsigned char) * (COUNT) + COUNT , ...) | vmalloc( - sizeof(u8) * COUNT + COUNT , ...) | vmalloc( - sizeof(__u8) * COUNT + COUNT , ...) | vmalloc( - sizeof(char) * COUNT + COUNT , ...) | vmalloc( - sizeof(unsigned char) * COUNT + COUNT , ...) ) // 2-factor product with sizeof(type/expression) and identifier or constant. @@ type TYPE; expression THING; identifier COUNT_ID; constant COUNT_CONST; @@ ( vmalloc( - sizeof(TYPE) * (COUNT_ID) + array_size(COUNT_ID, sizeof(TYPE)) , ...) | vmalloc( - sizeof(TYPE) * COUNT_ID + array_size(COUNT_ID, sizeof(TYPE)) , ...) | vmalloc( - sizeof(TYPE) * (COUNT_CONST) + array_size(COUNT_CONST, sizeof(TYPE)) , ...) | vmalloc( - sizeof(TYPE) * COUNT_CONST + array_size(COUNT_CONST, sizeof(TYPE)) , ...) | vmalloc( - sizeof(THING) * (COUNT_ID) + array_size(COUNT_ID, sizeof(THING)) , ...) | vmalloc( - sizeof(THING) * COUNT_ID + array_size(COUNT_ID, sizeof(THING)) , ...) | vmalloc( - sizeof(THING) * (COUNT_CONST) + array_size(COUNT_CONST, sizeof(THING)) , ...) | vmalloc( - sizeof(THING) * COUNT_CONST + array_size(COUNT_CONST, sizeof(THING)) , ...) ) // 2-factor product, only identifiers. @@ identifier SIZE, COUNT; @@ vmalloc( - SIZE * COUNT + array_size(COUNT, SIZE) , ...) // 3-factor product with 1 sizeof(type) or sizeof(expression), with // redundant parens removed. @@ expression THING; identifier STRIDE, COUNT; type TYPE; @@ ( vmalloc( - sizeof(TYPE) * (COUNT) * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | vmalloc( - sizeof(TYPE) * (COUNT) * STRIDE + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | vmalloc( - sizeof(TYPE) * COUNT * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | vmalloc( - sizeof(TYPE) * COUNT * STRIDE + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | vmalloc( - sizeof(THING) * (COUNT) * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | vmalloc( - sizeof(THING) * (COUNT) * STRIDE + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | vmalloc( - sizeof(THING) * COUNT * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | vmalloc( - sizeof(THING) * COUNT * STRIDE + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) ) // 3-factor product with 2 sizeof(variable), with redundant parens removed. @@ expression THING1, THING2; identifier COUNT; type TYPE1, TYPE2; @@ ( vmalloc( - sizeof(TYPE1) * sizeof(TYPE2) * COUNT + array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2)) , ...) | vmalloc( - sizeof(TYPE1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2)) , ...) | vmalloc( - sizeof(THING1) * sizeof(THING2) * COUNT + array3_size(COUNT, sizeof(THING1), sizeof(THING2)) , ...) | vmalloc( - sizeof(THING1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(THING1), sizeof(THING2)) , ...) | vmalloc( - sizeof(TYPE1) * sizeof(THING2) * COUNT + array3_size(COUNT, sizeof(TYPE1), sizeof(THING2)) , ...) | vmalloc( - sizeof(TYPE1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(TYPE1), sizeof(THING2)) , ...) ) // 3-factor product, only identifiers, with redundant parens removed. @@ identifier STRIDE, SIZE, COUNT; @@ ( vmalloc( - (COUNT) * STRIDE * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - COUNT * (STRIDE) * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - COUNT * STRIDE * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - (COUNT) * (STRIDE) * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - COUNT * (STRIDE) * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - (COUNT) * STRIDE * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - (COUNT) * (STRIDE) * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | vmalloc( - COUNT * STRIDE * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) ) // Any remaining multi-factor products, first at least 3-factor products // when they're not all constants... @@ expression E1, E2, E3; constant C1, C2, C3; @@ ( vmalloc(C1 * C2 * C3, ...) | vmalloc( - E1 * E2 * E3 + array3_size(E1, E2, E3) , ...) ) // And then all remaining 2 factors products when they're not all constants. @@ expression E1, E2; constant C1, C2; @@ ( vmalloc(C1 * C2, ...) | vmalloc( - E1 * E2 + array_size(E1, E2) , ...) ) Signed-off-by: Kees Cook <keescook@chromium.org>
2018-06-13 04:27:11 +07:00
rev = vmalloc(array_size(npte, sizeof(struct revmap_entry)));
if (!rev) {
if (cma)
kvm_free_hpt_cma(page, 1 << (order - PAGE_SHIFT));
else
free_pages(hpt, order - PAGE_SHIFT);
return -ENOMEM;
}
info->order = order;
info->virt = hpt;
info->cma = cma;
info->rev = rev;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
return 0;
}
void kvmppc_set_hpt(struct kvm *kvm, struct kvm_hpt_info *info)
{
atomic64_set(&kvm->arch.mmio_update, 0);
kvm->arch.hpt = *info;
kvm->arch.sdr1 = __pa(info->virt) | (info->order - 18);
pr_debug("KVM guest htab at %lx (order %ld), LPID %x\n",
info->virt, (long)info->order, kvm->arch.lpid);
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
}
long kvmppc_alloc_reset_hpt(struct kvm *kvm, int order)
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
{
long err = -EBUSY;
struct kvm_hpt_info info;
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
mutex_lock(&kvm->arch.mmu_setup_lock);
if (kvm->arch.mmu_ready) {
kvm->arch.mmu_ready = 0;
/* order mmu_ready vs. vcpus_running */
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
smp_mb();
if (atomic_read(&kvm->arch.vcpus_running)) {
kvm->arch.mmu_ready = 1;
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
goto out;
}
}
if (kvm_is_radix(kvm)) {
err = kvmppc_switch_mmu_to_hpt(kvm);
if (err)
goto out;
}
if (kvm->arch.hpt.order == order) {
/* We already have a suitable HPT */
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
/* Set the entire HPT to 0, i.e. invalid HPTEs */
memset((void *)kvm->arch.hpt.virt, 0, 1ul << order);
/*
* Reset all the reverse-mapping chains for all memslots
*/
kvmppc_rmap_reset(kvm);
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
err = 0;
goto out;
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
}
KVM: PPC: Book3S HV: Fix host crash on changing HPT size Commit f98a8bf9ee20 ("KVM: PPC: Book3S HV: Allow KVM_PPC_ALLOCATE_HTAB ioctl() to change HPT size", 2016-12-20) changed the behaviour of the KVM_PPC_ALLOCATE_HTAB ioctl so that it now allocates a new HPT and new revmap array if there was a previously-allocated HPT of a different size from the size being requested. In this case, we need to reset the rmap arrays of the memslots, because the rmap arrays will contain references to HPTEs which are no longer valid. Worse, these references are also references to slots in the new revmap array (which parallels the HPT), and the new revmap array contains random contents, since it doesn't get zeroed on allocation. The effect of having these stale references to slots in the revmap array that contain random contents is that subsequent calls to functions such as kvmppc_add_revmap_chain will crash because they will interpret the non-zero contents of the revmap array as HPTE indexes and thus index outside of the revmap array. This leads to host crashes such as the following. [ 7072.862122] Unable to handle kernel paging request for data at address 0xd000000c250c00f8 [ 7072.862218] Faulting instruction address: 0xc0000000000e1c78 [ 7072.862233] Oops: Kernel access of bad area, sig: 11 [#1] [ 7072.862286] SMP NR_CPUS=1024 [ 7072.862286] NUMA [ 7072.862325] PowerNV [ 7072.862378] Modules linked in: kvm_hv vhost_net vhost tap xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables rpcrdma ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ib_srp scsi_transport_srp ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm iw_cxgb3 mlx5_ib ib_core ses enclosure scsi_transport_sas ipmi_powernv ipmi_devintf ipmi_msghandler powernv_op_panel i2c_opal nfsd auth_rpcgss oid_registry [ 7072.863085] nfs_acl lockd grace sunrpc kvm_pr kvm xfs libcrc32c scsi_dh_alua dm_service_time radeon lpfc nvme_fc nvme_fabrics nvme_core scsi_transport_fc i2c_algo_bit tg3 drm_kms_helper ptp pps_core syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm dm_multipath i2c_core cxgb3 mlx5_core mdio [last unloaded: kvm_hv] [ 7072.863381] CPU: 72 PID: 56929 Comm: qemu-system-ppc Not tainted 4.12.0-kvm+ #59 [ 7072.863457] task: c000000fe29e7600 task.stack: c000001e3ffec000 [ 7072.863520] NIP: c0000000000e1c78 LR: c0000000000e2e3c CTR: c0000000000e25f0 [ 7072.863596] REGS: c000001e3ffef560 TRAP: 0300 Not tainted (4.12.0-kvm+) [ 7072.863658] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE,TM[E]> [ 7072.863667] CR: 44082882 XER: 20000000 [ 7072.863767] CFAR: c0000000000e2e38 DAR: d000000c250c00f8 DSISR: 42000000 SOFTE: 1 GPR00: c0000000000e2e3c c000001e3ffef7e0 c000000001407d00 d000000c250c00f0 GPR04: d00000006509fb70 d00000000b3d2048 0000000003ffdfb7 0000000000000000 GPR08: 00000001007fdfb7 00000000c000000f d0000000250c0000 000000000070f7bf GPR12: 0000000000000008 c00000000fdad000 0000000010879478 00000000105a0d78 GPR16: 00007ffaf4080000 0000000000001190 0000000000000000 0000000000010000 GPR20: 4001ffffff000415 d00000006509fb70 0000000004091190 0000000ee1881190 GPR24: 0000000003ffdfb7 0000000003ffdfb7 00000000007fdfb7 c000000f5c958000 GPR28: d00000002d09fb70 0000000003ffdfb7 d00000006509fb70 d00000000b3d2048 [ 7072.864439] NIP [c0000000000e1c78] kvmppc_add_revmap_chain+0x88/0x130 [ 7072.864503] LR [c0000000000e2e3c] kvmppc_do_h_enter+0x84c/0x9e0 [ 7072.864566] Call Trace: [ 7072.864594] [c000001e3ffef7e0] [c000001e3ffef830] 0xc000001e3ffef830 (unreliable) [ 7072.864671] [c000001e3ffef830] [c0000000000e2e3c] kvmppc_do_h_enter+0x84c/0x9e0 [ 7072.864751] [c000001e3ffef920] [d00000000b38d878] kvmppc_map_vrma+0x168/0x200 [kvm_hv] [ 7072.864831] [c000001e3ffef9e0] [d00000000b38a684] kvmppc_vcpu_run_hv+0x1284/0x1300 [kvm_hv] [ 7072.864914] [c000001e3ffefb30] [d00000000f465664] kvmppc_vcpu_run+0x44/0x60 [kvm] [ 7072.865008] [c000001e3ffefb60] [d00000000f461864] kvm_arch_vcpu_ioctl_run+0x114/0x290 [kvm] [ 7072.865152] [c000001e3ffefbe0] [d00000000f453c98] kvm_vcpu_ioctl+0x598/0x7a0 [kvm] [ 7072.865292] [c000001e3ffefd40] [c000000000389328] do_vfs_ioctl+0xd8/0x8c0 [ 7072.865410] [c000001e3ffefde0] [c000000000389be4] SyS_ioctl+0xd4/0x130 [ 7072.865526] [c000001e3ffefe30] [c00000000000b760] system_call+0x58/0x6c [ 7072.865644] Instruction dump: [ 7072.865715] e95b2110 793a0020 7b4926e4 7f8a4a14 409e0098 807c000c 786326e4 7c6a1a14 [ 7072.865857] 935e0008 7bbd0020 813c000c 913e000c <93a30008> 93bc000c 48000038 60000000 [ 7072.866001] ---[ end trace 627b6e4bf8080edc ]--- Note that to trigger this, it is necessary to use a recent upstream QEMU (or other userspace that resizes the HPT at CAS time), specify a maximum memory size substantially larger than the current memory size, and boot a guest kernel that does not support HPT resizing. This fixes the problem by resetting the rmap arrays when the old HPT is freed. Fixes: f98a8bf9ee20 ("KVM: PPC: Book3S HV: Allow KVM_PPC_ALLOCATE_HTAB ioctl() to change HPT size") Cc: stable@vger.kernel.org # v4.11+ Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-07-21 12:41:49 +07:00
if (kvm->arch.hpt.virt) {
kvmppc_free_hpt(&kvm->arch.hpt);
KVM: PPC: Book3S HV: Fix host crash on changing HPT size Commit f98a8bf9ee20 ("KVM: PPC: Book3S HV: Allow KVM_PPC_ALLOCATE_HTAB ioctl() to change HPT size", 2016-12-20) changed the behaviour of the KVM_PPC_ALLOCATE_HTAB ioctl so that it now allocates a new HPT and new revmap array if there was a previously-allocated HPT of a different size from the size being requested. In this case, we need to reset the rmap arrays of the memslots, because the rmap arrays will contain references to HPTEs which are no longer valid. Worse, these references are also references to slots in the new revmap array (which parallels the HPT), and the new revmap array contains random contents, since it doesn't get zeroed on allocation. The effect of having these stale references to slots in the revmap array that contain random contents is that subsequent calls to functions such as kvmppc_add_revmap_chain will crash because they will interpret the non-zero contents of the revmap array as HPTE indexes and thus index outside of the revmap array. This leads to host crashes such as the following. [ 7072.862122] Unable to handle kernel paging request for data at address 0xd000000c250c00f8 [ 7072.862218] Faulting instruction address: 0xc0000000000e1c78 [ 7072.862233] Oops: Kernel access of bad area, sig: 11 [#1] [ 7072.862286] SMP NR_CPUS=1024 [ 7072.862286] NUMA [ 7072.862325] PowerNV [ 7072.862378] Modules linked in: kvm_hv vhost_net vhost tap xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables rpcrdma ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ib_srp scsi_transport_srp ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm iw_cxgb3 mlx5_ib ib_core ses enclosure scsi_transport_sas ipmi_powernv ipmi_devintf ipmi_msghandler powernv_op_panel i2c_opal nfsd auth_rpcgss oid_registry [ 7072.863085] nfs_acl lockd grace sunrpc kvm_pr kvm xfs libcrc32c scsi_dh_alua dm_service_time radeon lpfc nvme_fc nvme_fabrics nvme_core scsi_transport_fc i2c_algo_bit tg3 drm_kms_helper ptp pps_core syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm dm_multipath i2c_core cxgb3 mlx5_core mdio [last unloaded: kvm_hv] [ 7072.863381] CPU: 72 PID: 56929 Comm: qemu-system-ppc Not tainted 4.12.0-kvm+ #59 [ 7072.863457] task: c000000fe29e7600 task.stack: c000001e3ffec000 [ 7072.863520] NIP: c0000000000e1c78 LR: c0000000000e2e3c CTR: c0000000000e25f0 [ 7072.863596] REGS: c000001e3ffef560 TRAP: 0300 Not tainted (4.12.0-kvm+) [ 7072.863658] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE,TM[E]> [ 7072.863667] CR: 44082882 XER: 20000000 [ 7072.863767] CFAR: c0000000000e2e38 DAR: d000000c250c00f8 DSISR: 42000000 SOFTE: 1 GPR00: c0000000000e2e3c c000001e3ffef7e0 c000000001407d00 d000000c250c00f0 GPR04: d00000006509fb70 d00000000b3d2048 0000000003ffdfb7 0000000000000000 GPR08: 00000001007fdfb7 00000000c000000f d0000000250c0000 000000000070f7bf GPR12: 0000000000000008 c00000000fdad000 0000000010879478 00000000105a0d78 GPR16: 00007ffaf4080000 0000000000001190 0000000000000000 0000000000010000 GPR20: 4001ffffff000415 d00000006509fb70 0000000004091190 0000000ee1881190 GPR24: 0000000003ffdfb7 0000000003ffdfb7 00000000007fdfb7 c000000f5c958000 GPR28: d00000002d09fb70 0000000003ffdfb7 d00000006509fb70 d00000000b3d2048 [ 7072.864439] NIP [c0000000000e1c78] kvmppc_add_revmap_chain+0x88/0x130 [ 7072.864503] LR [c0000000000e2e3c] kvmppc_do_h_enter+0x84c/0x9e0 [ 7072.864566] Call Trace: [ 7072.864594] [c000001e3ffef7e0] [c000001e3ffef830] 0xc000001e3ffef830 (unreliable) [ 7072.864671] [c000001e3ffef830] [c0000000000e2e3c] kvmppc_do_h_enter+0x84c/0x9e0 [ 7072.864751] [c000001e3ffef920] [d00000000b38d878] kvmppc_map_vrma+0x168/0x200 [kvm_hv] [ 7072.864831] [c000001e3ffef9e0] [d00000000b38a684] kvmppc_vcpu_run_hv+0x1284/0x1300 [kvm_hv] [ 7072.864914] [c000001e3ffefb30] [d00000000f465664] kvmppc_vcpu_run+0x44/0x60 [kvm] [ 7072.865008] [c000001e3ffefb60] [d00000000f461864] kvm_arch_vcpu_ioctl_run+0x114/0x290 [kvm] [ 7072.865152] [c000001e3ffefbe0] [d00000000f453c98] kvm_vcpu_ioctl+0x598/0x7a0 [kvm] [ 7072.865292] [c000001e3ffefd40] [c000000000389328] do_vfs_ioctl+0xd8/0x8c0 [ 7072.865410] [c000001e3ffefde0] [c000000000389be4] SyS_ioctl+0xd4/0x130 [ 7072.865526] [c000001e3ffefe30] [c00000000000b760] system_call+0x58/0x6c [ 7072.865644] Instruction dump: [ 7072.865715] e95b2110 793a0020 7b4926e4 7f8a4a14 409e0098 807c000c 786326e4 7c6a1a14 [ 7072.865857] 935e0008 7bbd0020 813c000c 913e000c <93a30008> 93bc000c 48000038 60000000 [ 7072.866001] ---[ end trace 627b6e4bf8080edc ]--- Note that to trigger this, it is necessary to use a recent upstream QEMU (or other userspace that resizes the HPT at CAS time), specify a maximum memory size substantially larger than the current memory size, and boot a guest kernel that does not support HPT resizing. This fixes the problem by resetting the rmap arrays when the old HPT is freed. Fixes: f98a8bf9ee20 ("KVM: PPC: Book3S HV: Allow KVM_PPC_ALLOCATE_HTAB ioctl() to change HPT size") Cc: stable@vger.kernel.org # v4.11+ Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-07-21 12:41:49 +07:00
kvmppc_rmap_reset(kvm);
}
err = kvmppc_allocate_hpt(&info, order);
if (err < 0)
goto out;
kvmppc_set_hpt(kvm, &info);
out:
if (err == 0)
/* Ensure that each vcpu will flush its TLB on next entry. */
cpumask_setall(&kvm->arch.need_tlb_flush);
mutex_unlock(&kvm->arch.mmu_setup_lock);
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
return err;
}
void kvmppc_free_hpt(struct kvm_hpt_info *info)
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
{
vfree(info->rev);
info->rev = NULL;
if (info->cma)
kvm_free_hpt_cma(virt_to_page(info->virt),
1 << (info->order - PAGE_SHIFT));
else if (info->virt)
free_pages(info->virt, info->order - PAGE_SHIFT);
info->virt = 0;
info->order = 0;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
}
/* Bits in first HPTE dword for pagesize 4k, 64k or 16M */
static inline unsigned long hpte0_pgsize_encoding(unsigned long pgsize)
{
return (pgsize > 0x1000) ? HPTE_V_LARGE : 0;
}
/* Bits in second HPTE dword for pagesize 4k, 64k or 16M */
static inline unsigned long hpte1_pgsize_encoding(unsigned long pgsize)
{
return (pgsize == 0x10000) ? 0x1000 : 0;
}
void kvmppc_map_vrma(struct kvm_vcpu *vcpu, struct kvm_memory_slot *memslot,
unsigned long porder)
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
{
unsigned long i;
unsigned long npages;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 19:31:00 +07:00
unsigned long hp_v, hp_r;
unsigned long addr, hash;
unsigned long psize;
unsigned long hp0, hp1;
unsigned long idx_ret;
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 19:31:00 +07:00
long ret;
KVM: PPC: Book3S HV: Make the guest hash table size configurable This adds a new ioctl to enable userspace to control the size of the guest hashed page table (HPT) and to clear it out when resetting the guest. The KVM_PPC_ALLOCATE_HTAB ioctl is a VM ioctl and takes as its parameter a pointer to a u32 containing the desired order of the HPT (log base 2 of the size in bytes), which is updated on successful return to the actual order of the HPT which was allocated. There must be no vcpus running at the time of this ioctl. To enforce this, we now keep a count of the number of vcpus running in kvm->arch.vcpus_running. If the ioctl is called when a HPT has already been allocated, we don't reallocate the HPT but just clear it out. We first clear the kvm->arch.rma_setup_done flag, which has two effects: (a) since we hold the kvm->lock mutex, it will prevent any vcpus from starting to run until we're done, and (b) it means that the first vcpu to run after we're done will re-establish the VRMA if necessary. If userspace doesn't call this ioctl before running the first vcpu, the kernel will allocate a default-sized HPT at that point. We do it then rather than when creating the VM, as the code did previously, so that userspace has a chance to do the ioctl if it wants. When allocating the HPT, we can allocate either from the kernel page allocator, or from the preallocated pool. If userspace is asking for a different size from the preallocated HPTs, we first try to allocate using the kernel page allocator. Then we try to allocate from the preallocated pool, and then if that fails, we try allocating decreasing sizes from the kernel page allocator, down to the minimum size allowed (256kB). Note that the kernel page allocator limits allocations to 1 << CONFIG_FORCE_MAX_ZONEORDER pages, which by default corresponds to 16MB (on 64-bit powerpc, at least). Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix module compilation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-05-04 09:32:53 +07:00
struct kvm *kvm = vcpu->kvm;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
psize = 1ul << porder;
npages = memslot->npages >> (porder - PAGE_SHIFT);
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
/* VRMA can't be > 1TB */
if (npages > 1ul << (40 - porder))
npages = 1ul << (40 - porder);
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
/* Can't use more than 1 HPTE per HPTEG */
if (npages > kvmppc_hpt_mask(&kvm->arch.hpt) + 1)
npages = kvmppc_hpt_mask(&kvm->arch.hpt) + 1;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
hp0 = HPTE_V_1TB_SEG | (VRMA_VSID << (40 - 16)) |
HPTE_V_BOLTED | hpte0_pgsize_encoding(psize);
hp1 = hpte1_pgsize_encoding(psize) |
HPTE_R_R | HPTE_R_C | HPTE_R_M | PP_RWXX;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
for (i = 0; i < npages; ++i) {
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 19:31:00 +07:00
addr = i << porder;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
/* can't use hpt_hash since va > 64 bits */
hash = (i ^ (VRMA_VSID ^ (VRMA_VSID << 25)))
& kvmppc_hpt_mask(&kvm->arch.hpt);
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
/*
* We assume that the hash table is empty and no
* vcpus are using it at this stage. Since we create
* at most one HPTE per HPTEG, we just assume entry 7
* is available and use it.
*/
hash = (hash << 3) + 7;
hp_v = hp0 | ((addr >> 16) & ~0x7fUL);
hp_r = hp1 | addr;
ret = kvmppc_virtmode_do_h_enter(kvm, H_EXACT, hash, hp_v, hp_r,
&idx_ret);
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 19:31:00 +07:00
if (ret != H_SUCCESS) {
pr_err("KVM: map_vrma at %lx failed, ret=%ld\n",
addr, ret);
break;
}
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
}
}
int kvmppc_mmu_hv_init(void)
{
KVM: PPC: book3s_hv: Add support for PPC970-family processors This adds support for running KVM guests in supervisor mode on those PPC970 processors that have a usable hypervisor mode. Unfortunately, Apple G5 machines have supervisor mode disabled (MSR[HV] is forced to 1), but the YDL PowerStation does have a usable hypervisor mode. There are several differences between the PPC970 and POWER7 in how guests are managed. These differences are accommodated using the CPU_FTR_ARCH_201 (PPC970) and CPU_FTR_ARCH_206 (POWER7) CPU feature bits. Notably, on PPC970: * The LPCR, LPID or RMOR registers don't exist, and the functions of those registers are provided by bits in HID4 and one bit in HID0. * External interrupts can be directed to the hypervisor, but unlike POWER7 they are masked by MSR[EE] in non-hypervisor modes and use SRR0/1 not HSRR0/1. * There is no virtual RMA (VRMA) mode; the guest must use an RMO (real mode offset) area. * The TLB entries are not tagged with the LPID, so it is necessary to flush the whole TLB on partition switch. Furthermore, when switching partitions we have to ensure that no other CPU is executing the tlbie or tlbsync instructions in either the old or the new partition, otherwise undefined behaviour can occur. * The PMU has 8 counters (PMC registers) rather than 6. * The DSCR, PURR, SPURR, AMR, AMOR, UAMOR registers don't exist. * The SLB has 64 entries rather than 32. * There is no mediated external interrupt facility, so if we switch to a guest that has a virtual external interrupt pending but the guest has MSR[EE] = 0, we have to arrange to have an interrupt pending for it so that we can get control back once it re-enables interrupts. We do that by sending ourselves an IPI with smp_send_reschedule after hard-disabling interrupts. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:40:08 +07:00
unsigned long host_lpid, rsvd_lpid;
if (!mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE))
return -EINVAL;
/* POWER7 has 10-bit LPIDs (12-bit in POWER8) */
host_lpid = 0;
if (cpu_has_feature(CPU_FTR_HVMODE))
host_lpid = mfspr(SPRN_LPID);
rsvd_lpid = LPID_RSVD;
KVM: PPC: book3s_hv: Add support for PPC970-family processors This adds support for running KVM guests in supervisor mode on those PPC970 processors that have a usable hypervisor mode. Unfortunately, Apple G5 machines have supervisor mode disabled (MSR[HV] is forced to 1), but the YDL PowerStation does have a usable hypervisor mode. There are several differences between the PPC970 and POWER7 in how guests are managed. These differences are accommodated using the CPU_FTR_ARCH_201 (PPC970) and CPU_FTR_ARCH_206 (POWER7) CPU feature bits. Notably, on PPC970: * The LPCR, LPID or RMOR registers don't exist, and the functions of those registers are provided by bits in HID4 and one bit in HID0. * External interrupts can be directed to the hypervisor, but unlike POWER7 they are masked by MSR[EE] in non-hypervisor modes and use SRR0/1 not HSRR0/1. * There is no virtual RMA (VRMA) mode; the guest must use an RMO (real mode offset) area. * The TLB entries are not tagged with the LPID, so it is necessary to flush the whole TLB on partition switch. Furthermore, when switching partitions we have to ensure that no other CPU is executing the tlbie or tlbsync instructions in either the old or the new partition, otherwise undefined behaviour can occur. * The PMU has 8 counters (PMC registers) rather than 6. * The DSCR, PURR, SPURR, AMR, AMOR, UAMOR registers don't exist. * The SLB has 64 entries rather than 32. * There is no mediated external interrupt facility, so if we switch to a guest that has a virtual external interrupt pending but the guest has MSR[EE] = 0, we have to arrange to have an interrupt pending for it so that we can get control back once it re-enables interrupts. We do that by sending ourselves an IPI with smp_send_reschedule after hard-disabling interrupts. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:40:08 +07:00
kvmppc_init_lpid(rsvd_lpid + 1);
kvmppc_claim_lpid(host_lpid);
KVM: PPC: book3s_hv: Add support for PPC970-family processors This adds support for running KVM guests in supervisor mode on those PPC970 processors that have a usable hypervisor mode. Unfortunately, Apple G5 machines have supervisor mode disabled (MSR[HV] is forced to 1), but the YDL PowerStation does have a usable hypervisor mode. There are several differences between the PPC970 and POWER7 in how guests are managed. These differences are accommodated using the CPU_FTR_ARCH_201 (PPC970) and CPU_FTR_ARCH_206 (POWER7) CPU feature bits. Notably, on PPC970: * The LPCR, LPID or RMOR registers don't exist, and the functions of those registers are provided by bits in HID4 and one bit in HID0. * External interrupts can be directed to the hypervisor, but unlike POWER7 they are masked by MSR[EE] in non-hypervisor modes and use SRR0/1 not HSRR0/1. * There is no virtual RMA (VRMA) mode; the guest must use an RMO (real mode offset) area. * The TLB entries are not tagged with the LPID, so it is necessary to flush the whole TLB on partition switch. Furthermore, when switching partitions we have to ensure that no other CPU is executing the tlbie or tlbsync instructions in either the old or the new partition, otherwise undefined behaviour can occur. * The PMU has 8 counters (PMC registers) rather than 6. * The DSCR, PURR, SPURR, AMR, AMOR, UAMOR registers don't exist. * The SLB has 64 entries rather than 32. * There is no mediated external interrupt facility, so if we switch to a guest that has a virtual external interrupt pending but the guest has MSR[EE] = 0, we have to arrange to have an interrupt pending for it so that we can get control back once it re-enables interrupts. We do that by sending ourselves an IPI with smp_send_reschedule after hard-disabling interrupts. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:40:08 +07:00
/* rsvd_lpid is reserved for use in partition switching */
kvmppc_claim_lpid(rsvd_lpid);
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
return 0;
}
static long kvmppc_virtmode_do_h_enter(struct kvm *kvm, unsigned long flags,
long pte_index, unsigned long pteh,
unsigned long ptel, unsigned long *pte_idx_ret)
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 19:31:00 +07:00
{
long ret;
/* Protect linux PTE lookup from page table destruction */
rcu_read_lock_sched(); /* this disables preemption too */
ret = kvmppc_do_h_enter(kvm, flags, pte_index, pteh, ptel,
kvm->mm->pgd, false, pte_idx_ret);
rcu_read_unlock_sched();
KVM: PPC: Only get pages when actually needed, not in prepare_memory_region() This removes the code from kvmppc_core_prepare_memory_region() that looked up the VMA for the region being added and called hva_to_page to get the pfns for the memory. We have no guarantee that there will be anything mapped there at the time of the KVM_SET_USER_MEMORY_REGION ioctl call; userspace can do that ioctl and then map memory into the region later. Instead we defer looking up the pfn for each memory page until it is needed, which generally means when the guest does an H_ENTER hcall on the page. Since we can't call get_user_pages in real mode, if we don't already have the pfn for the page, kvmppc_h_enter() will return H_TOO_HARD and we then call kvmppc_virtmode_h_enter() once we get back to kernel context. That calls kvmppc_get_guest_page() to get the pfn for the page, and then calls back to kvmppc_h_enter() to redo the HPTE insertion. When the first vcpu starts executing, we need to have the RMO or VRMA region mapped so that the guest's real mode accesses will work. Thus we now have a check in kvmppc_vcpu_run() to see if the RMO/VRMA is set up and if not, call kvmppc_hv_setup_rma(). It checks if the memslot starting at guest physical 0 now has RMO memory mapped there; if so it sets it up for the guest, otherwise on POWER7 it sets up the VRMA. The function that does that, kvmppc_map_vrma, is now a bit simpler, as it calls kvmppc_virtmode_h_enter instead of creating the HPTE itself. Since we are now potentially updating entries in the slot_phys[] arrays from multiple vcpu threads, we now have a spinlock protecting those updates to ensure that we don't lose track of any references to pages. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-12 19:31:00 +07:00
if (ret == H_TOO_HARD) {
/* this can't happen */
pr_err("KVM: Oops, kvmppc_h_enter returned too hard!\n");
ret = H_RESOURCE; /* or something */
}
return ret;
}
static struct kvmppc_slb *kvmppc_mmu_book3s_hv_find_slbe(struct kvm_vcpu *vcpu,
gva_t eaddr)
{
u64 mask;
int i;
for (i = 0; i < vcpu->arch.slb_nr; i++) {
if (!(vcpu->arch.slb[i].orige & SLB_ESID_V))
continue;
if (vcpu->arch.slb[i].origv & SLB_VSID_B_1T)
mask = ESID_MASK_1T;
else
mask = ESID_MASK;
if (((vcpu->arch.slb[i].orige ^ eaddr) & mask) == 0)
return &vcpu->arch.slb[i];
}
return NULL;
}
static unsigned long kvmppc_mmu_get_real_addr(unsigned long v, unsigned long r,
unsigned long ea)
{
unsigned long ra_mask;
ra_mask = kvmppc_actual_pgsz(v, r) - 1;
return (r & HPTE_R_RPN & ~ra_mask) | (ea & ra_mask);
}
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
static int kvmppc_mmu_book3s_64_hv_xlate(struct kvm_vcpu *vcpu, gva_t eaddr,
KVM: PPC: Book3S PR: Better handling of host-side read-only pages Currently we request write access to all pages that get mapped into the guest, even if the guest is only loading from the page. This reduces the effectiveness of KSM because it means that we unshare every page we access. Also, we always set the changed (C) bit in the guest HPTE if it allows writing, even for a guest load. This fixes both these problems. We pass an 'iswrite' flag to the mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether the access is a load or a store. The mmu.xlate() functions now only set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot() instead of gfn_to_pfn() so that it can indicate whether we need write access to the page, and get back a 'writable' flag to indicate whether the page is writable or not. If that 'writable' flag is clear, we then make the host HPTE read-only even if the guest HPTE allowed writing. This means that we can get a protection fault when the guest writes to a page that it has mapped read-write but which is read-only on the host side (perhaps due to KSM having merged the page). Thus we now call kvmppc_handle_pagefault() for protection faults as well as HPTE not found faults. In kvmppc_handle_pagefault(), if the access was allowed by the guest HPTE and we thus need to install a new host HPTE, we then need to remove the old host HPTE if there is one. This is done with a new function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to find and remove the old host HPTE. Since the memslot-related functions require the KVM SRCU read lock to be held, this adds srcu_read_lock/unlock pairs around the calls to kvmppc_handle_pagefault(). Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore guest HPTEs that don't permit access, and to return -EPERM for accesses that are not permitted by the page protections. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-09-20 11:52:51 +07:00
struct kvmppc_pte *gpte, bool data, bool iswrite)
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
{
struct kvm *kvm = vcpu->kvm;
struct kvmppc_slb *slbe;
unsigned long slb_v;
unsigned long pp, key;
unsigned long v, orig_v, gr;
__be64 *hptep;
long int index;
int virtmode = vcpu->arch.shregs.msr & (data ? MSR_DR : MSR_IR);
if (kvm_is_radix(vcpu->kvm))
return kvmppc_mmu_radix_xlate(vcpu, eaddr, gpte, data, iswrite);
/* Get SLB entry */
if (virtmode) {
slbe = kvmppc_mmu_book3s_hv_find_slbe(vcpu, eaddr);
if (!slbe)
return -EINVAL;
slb_v = slbe->origv;
} else {
/* real mode access */
slb_v = vcpu->kvm->arch.vrma_slb_v;
}
preempt_disable();
/* Find the HPTE in the hash table */
index = kvmppc_hv_find_lock_hpte(kvm, eaddr, slb_v,
HPTE_V_VALID | HPTE_V_ABSENT);
if (index < 0) {
preempt_enable();
return -ENOENT;
}
hptep = (__be64 *)(kvm->arch.hpt.virt + (index << 4));
v = orig_v = be64_to_cpu(hptep[0]) & ~HPTE_V_HVLOCK;
if (cpu_has_feature(CPU_FTR_ARCH_300))
v = hpte_new_to_old_v(v, be64_to_cpu(hptep[1]));
gr = kvm->arch.hpt.rev[index].guest_rpte;
unlock_hpte(hptep, orig_v);
preempt_enable();
gpte->eaddr = eaddr;
gpte->vpage = ((v & HPTE_V_AVPN) << 4) | ((eaddr >> 12) & 0xfff);
/* Get PP bits and key for permission check */
pp = gr & (HPTE_R_PP0 | HPTE_R_PP);
key = (vcpu->arch.shregs.msr & MSR_PR) ? SLB_VSID_KP : SLB_VSID_KS;
key &= slb_v;
/* Calculate permissions */
gpte->may_read = hpte_read_permission(pp, key);
gpte->may_write = hpte_write_permission(pp, key);
gpte->may_execute = gpte->may_read && !(gr & (HPTE_R_N | HPTE_R_G));
/* Storage key permission check for POWER7 */
if (data && virtmode) {
int amrfield = hpte_get_skey_perm(gr, vcpu->arch.amr);
if (amrfield & 1)
gpte->may_read = 0;
if (amrfield & 2)
gpte->may_write = 0;
}
/* Get the guest physical address */
gpte->raddr = kvmppc_mmu_get_real_addr(v, gr, eaddr);
return 0;
}
/*
* Quick test for whether an instruction is a load or a store.
* If the instruction is a load or a store, then this will indicate
* which it is, at least on server processors. (Embedded processors
* have some external PID instructions that don't follow the rule
* embodied here.) If the instruction isn't a load or store, then
* this doesn't return anything useful.
*/
static int instruction_is_store(unsigned int instr)
{
unsigned int mask;
mask = 0x10000000;
if ((instr & 0xfc000000) == 0x7c000000)
mask = 0x100; /* major opcode 31 */
return (instr & mask) != 0;
}
int kvmppc_hv_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu,
unsigned long gpa, gva_t ea, int is_store)
{
u32 last_inst;
/*
* Fast path - check if the guest physical address corresponds to a
* device on the FAST_MMIO_BUS, if so we can avoid loading the
* instruction all together, then we can just handle it and return.
*/
if (is_store) {
int idx, ret;
idx = srcu_read_lock(&vcpu->kvm->srcu);
ret = kvm_io_bus_write(vcpu, KVM_FAST_MMIO_BUS, (gpa_t) gpa, 0,
NULL);
srcu_read_unlock(&vcpu->kvm->srcu, idx);
if (!ret) {
kvmppc_set_pc(vcpu, kvmppc_get_pc(vcpu) + 4);
return RESUME_GUEST;
}
}
/*
* If we fail, we just return to the guest and try executing it again.
*/
if (kvmppc_get_last_inst(vcpu, INST_GENERIC, &last_inst) !=
EMULATE_DONE)
return RESUME_GUEST;
/*
* WARNING: We do not know for sure whether the instruction we just
* read from memory is the same that caused the fault in the first
* place. If the instruction we read is neither an load or a store,
* then it can't access memory, so we don't need to worry about
* enforcing access permissions. So, assuming it is a load or
* store, we just check that its direction (load or store) is
* consistent with the original fault, since that's what we
* checked the access permissions against. If there is a mismatch
* we just return and retry the instruction.
*/
if (instruction_is_store(last_inst) != !!is_store)
return RESUME_GUEST;
/*
* Emulated accesses are emulated by looking at the hash for
* translation once, then performing the access later. The
* translation could be invalidated in the meantime in which
* point performing the subsequent memory access on the old
* physical address could possibly be a security hole for the
* guest (but not the host).
*
* This is less of an issue for MMIO stores since they aren't
* globally visible. It could be an issue for MMIO loads to
* a certain extent but we'll ignore it for now.
*/
vcpu->arch.paddr_accessed = gpa;
vcpu->arch.vaddr_accessed = ea;
return kvmppc_emulate_mmio(run, vcpu);
}
int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu,
unsigned long ea, unsigned long dsisr)
{
struct kvm *kvm = vcpu->kvm;
unsigned long hpte[3], r;
unsigned long hnow_v, hnow_r;
__be64 *hptep;
unsigned long mmu_seq, psize, pte_size;
unsigned long gpa_base, gfn_base;
unsigned long gpa, gfn, hva, pfn;
struct kvm_memory_slot *memslot;
unsigned long *rmap;
struct revmap_entry *rev;
struct page *page, *pages[1];
long index, ret, npages;
bool is_ci;
unsigned int writing, write_ok;
struct vm_area_struct *vma;
unsigned long rcbits;
long mmio_update;
struct mm_struct *mm;
if (kvm_is_radix(kvm))
return kvmppc_book3s_radix_page_fault(run, vcpu, ea, dsisr);
/*
* Real-mode code has already searched the HPT and found the
* entry we're interested in. Lock the entry and check that
* it hasn't changed. If it has, just return and re-execute the
* instruction.
*/
if (ea != vcpu->arch.pgfault_addr)
return RESUME_GUEST;
if (vcpu->arch.pgfault_cache) {
mmio_update = atomic64_read(&kvm->arch.mmio_update);
if (mmio_update == vcpu->arch.pgfault_cache->mmio_update) {
r = vcpu->arch.pgfault_cache->rpte;
psize = kvmppc_actual_pgsz(vcpu->arch.pgfault_hpte[0],
r);
gpa_base = r & HPTE_R_RPN & ~(psize - 1);
gfn_base = gpa_base >> PAGE_SHIFT;
gpa = gpa_base | (ea & (psize - 1));
return kvmppc_hv_emulate_mmio(run, vcpu, gpa, ea,
dsisr & DSISR_ISSTORE);
}
}
index = vcpu->arch.pgfault_index;
hptep = (__be64 *)(kvm->arch.hpt.virt + (index << 4));
rev = &kvm->arch.hpt.rev[index];
preempt_disable();
while (!try_lock_hpte(hptep, HPTE_V_HVLOCK))
cpu_relax();
hpte[0] = be64_to_cpu(hptep[0]) & ~HPTE_V_HVLOCK;
hpte[1] = be64_to_cpu(hptep[1]);
hpte[2] = r = rev->guest_rpte;
unlock_hpte(hptep, hpte[0]);
preempt_enable();
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
hpte[0] = hpte_new_to_old_v(hpte[0], hpte[1]);
hpte[1] = hpte_new_to_old_r(hpte[1]);
}
if (hpte[0] != vcpu->arch.pgfault_hpte[0] ||
hpte[1] != vcpu->arch.pgfault_hpte[1])
return RESUME_GUEST;
/* Translate the logical address and get the page */
psize = kvmppc_actual_pgsz(hpte[0], r);
gpa_base = r & HPTE_R_RPN & ~(psize - 1);
gfn_base = gpa_base >> PAGE_SHIFT;
gpa = gpa_base | (ea & (psize - 1));
gfn = gpa >> PAGE_SHIFT;
memslot = gfn_to_memslot(kvm, gfn);
trace_kvm_page_fault_enter(vcpu, hpte, memslot, ea, dsisr);
/* No memslot means it's an emulated MMIO region */
if (!memslot || (memslot->flags & KVM_MEMSLOT_INVALID))
return kvmppc_hv_emulate_mmio(run, vcpu, gpa, ea,
dsisr & DSISR_ISSTORE);
/*
* This should never happen, because of the slot_is_aligned()
* check in kvmppc_do_h_enter().
*/
if (gfn_base < memslot->base_gfn)
return -EFAULT;
/* used to check for invalidations in progress */
mmu_seq = kvm->mmu_notifier_seq;
smp_rmb();
ret = -EFAULT;
is_ci = false;
pfn = 0;
page = NULL;
mm = kvm->mm;
pte_size = PAGE_SIZE;
writing = (dsisr & DSISR_ISSTORE) != 0;
/* If writing != 0, then the HPTE must allow writing, if we get here */
write_ok = writing;
hva = gfn_to_hva_memslot(memslot, gfn);
mm/gup: change GUP fast to use flags rather than a write 'bool' To facilitate additional options to get_user_pages_fast() change the singular write parameter to be gup_flags. This patch does not change any functionality. New functionality will follow in subsequent patches. Some of the get_user_pages_fast() call sites were unchanged because they already passed FOLL_WRITE or 0 for the write parameter. NOTE: It was suggested to change the ordering of the get_user_pages_fast() arguments to ensure that callers were converted. This breaks the current GUP call site convention of having the returned pages be the final parameter. So the suggestion was rejected. Link: http://lkml.kernel.org/r/20190328084422.29911-4-ira.weiny@intel.com Link: http://lkml.kernel.org/r/20190317183438.2057-4-ira.weiny@intel.com Signed-off-by: Ira Weiny <ira.weiny@intel.com> Reviewed-by: Mike Marshall <hubcap@omnibond.com> Cc: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Borislav Petkov <bp@alien8.de> Cc: Dan Williams <dan.j.williams@intel.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: James Hogan <jhogan@kernel.org> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: John Hubbard <jhubbard@nvidia.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Rich Felker <dalias@libc.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 07:17:11 +07:00
npages = get_user_pages_fast(hva, 1, writing ? FOLL_WRITE : 0, pages);
if (npages < 1) {
/* Check if it's an I/O mapping */
down_read(&mm->mmap_sem);
vma = find_vma(mm, hva);
if (vma && vma->vm_start <= hva && hva + psize <= vma->vm_end &&
(vma->vm_flags & VM_PFNMAP)) {
pfn = vma->vm_pgoff +
((hva - vma->vm_start) >> PAGE_SHIFT);
pte_size = psize;
is_ci = pte_ci(__pte((pgprot_val(vma->vm_page_prot))));
write_ok = vma->vm_flags & VM_WRITE;
}
up_read(&mm->mmap_sem);
if (!pfn)
goto out_put;
} else {
page = pages[0];
KVM: PPC: Book3S HV: Fix physical address calculations This fixes a bug in kvmppc_do_h_enter() where the physical address for a page can be calculated incorrectly if transparent huge pages (THP) are active. Until THP came along, it was true that if we encountered a large (16M) page in kvmppc_do_h_enter(), then the associated memslot must be 16M aligned for both its guest physical address and the userspace address, and the physical address calculations in kvmppc_do_h_enter() assumed that. With THP, that is no longer true. In the case where we are using MMU notifiers and the page size that we get from the Linux page tables is larger than the page being mapped by the guest, we need to fill in some low-order bits of the physical address. Without THP, these bits would be the same in the guest physical address (gpa) and the host virtual address (hva). With THP, they can be different, and we need to use the bits from hva rather than gpa. In the case where we are not using MMU notifiers, the host physical address we get from the memslot->arch.slot_phys[] array already includes the low-order bits down to the PAGE_SIZE level, even if we are using large pages. Thus we can simplify the calculation in this case to just add in the remaining bits in the case where PAGE_SIZE is 64k and the guest is mapping a 4k page. The same bug exists in kvmppc_book3s_hv_page_fault(). The basic fix is to use psize (the page size from the HPTE) rather than pte_size (the page size from the Linux PTE) when updating the HPTE low word in r. That means that pfn needs to be computed to PAGE_SIZE granularity even if the Linux PTE is a huge page PTE. That can be arranged simply by doing the page_to_pfn() before setting page to the head of the compound page. If psize is less than PAGE_SIZE, then we need to make sure we only update the bits from PAGE_SIZE upwards, in order not to lose any sub-page offset bits in r. On the other hand, if psize is greater than PAGE_SIZE, we need to make sure we don't bring in non-zero low order bits in pfn, hence we mask (pfn << PAGE_SHIFT) with ~(psize - 1). Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-16 13:46:02 +07:00
pfn = page_to_pfn(page);
if (PageHuge(page)) {
page = compound_head(page);
pte_size <<= compound_order(page);
}
/* if the guest wants write access, see if that is OK */
if (!writing && hpte_is_writable(r)) {
pte_t *ptep, pte;
unsigned long flags;
/*
* We need to protect against page table destruction
* hugepage split and collapse.
*/
local_irq_save(flags);
ptep = find_current_mm_pte(mm->pgd, hva, NULL, NULL);
if (ptep) {
pte = kvmppc_read_update_linux_pte(ptep, 1);
if (__pte_write(pte))
write_ok = 1;
}
local_irq_restore(flags);
}
}
if (psize > pte_size)
goto out_put;
/* Check WIMG vs. the actual page we're accessing */
if (!hpte_cache_flags_ok(r, is_ci)) {
if (is_ci)
goto out_put;
/*
* Allow guest to map emulated device memory as
* uncacheable, but actually make it cacheable.
*/
r = (r & ~(HPTE_R_W|HPTE_R_I|HPTE_R_G)) | HPTE_R_M;
}
KVM: PPC: Book3S HV: Fix physical address calculations This fixes a bug in kvmppc_do_h_enter() where the physical address for a page can be calculated incorrectly if transparent huge pages (THP) are active. Until THP came along, it was true that if we encountered a large (16M) page in kvmppc_do_h_enter(), then the associated memslot must be 16M aligned for both its guest physical address and the userspace address, and the physical address calculations in kvmppc_do_h_enter() assumed that. With THP, that is no longer true. In the case where we are using MMU notifiers and the page size that we get from the Linux page tables is larger than the page being mapped by the guest, we need to fill in some low-order bits of the physical address. Without THP, these bits would be the same in the guest physical address (gpa) and the host virtual address (hva). With THP, they can be different, and we need to use the bits from hva rather than gpa. In the case where we are not using MMU notifiers, the host physical address we get from the memslot->arch.slot_phys[] array already includes the low-order bits down to the PAGE_SIZE level, even if we are using large pages. Thus we can simplify the calculation in this case to just add in the remaining bits in the case where PAGE_SIZE is 64k and the guest is mapping a 4k page. The same bug exists in kvmppc_book3s_hv_page_fault(). The basic fix is to use psize (the page size from the HPTE) rather than pte_size (the page size from the Linux PTE) when updating the HPTE low word in r. That means that pfn needs to be computed to PAGE_SIZE granularity even if the Linux PTE is a huge page PTE. That can be arranged simply by doing the page_to_pfn() before setting page to the head of the compound page. If psize is less than PAGE_SIZE, then we need to make sure we only update the bits from PAGE_SIZE upwards, in order not to lose any sub-page offset bits in r. On the other hand, if psize is greater than PAGE_SIZE, we need to make sure we don't bring in non-zero low order bits in pfn, hence we mask (pfn << PAGE_SHIFT) with ~(psize - 1). Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-16 13:46:02 +07:00
/*
* Set the HPTE to point to pfn.
* Since the pfn is at PAGE_SIZE granularity, make sure we
* don't mask out lower-order bits if psize < PAGE_SIZE.
*/
if (psize < PAGE_SIZE)
psize = PAGE_SIZE;
r = (r & HPTE_R_KEY_HI) | (r & ~(HPTE_R_PP0 - psize)) |
((pfn << PAGE_SHIFT) & ~(psize - 1));
if (hpte_is_writable(r) && !write_ok)
r = hpte_make_readonly(r);
ret = RESUME_GUEST;
preempt_disable();
while (!try_lock_hpte(hptep, HPTE_V_HVLOCK))
cpu_relax();
hnow_v = be64_to_cpu(hptep[0]);
hnow_r = be64_to_cpu(hptep[1]);
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
hnow_v = hpte_new_to_old_v(hnow_v, hnow_r);
hnow_r = hpte_new_to_old_r(hnow_r);
}
KVM: PPC: Book3S HV: Fix exclusion between HPT resizing and other HPT updates Commit 5e9859699aba ("KVM: PPC: Book3S HV: Outline of KVM-HV HPT resizing implementation", 2016-12-20) added code that tries to exclude any use or update of the hashed page table (HPT) while the HPT resizing code is iterating through all the entries in the HPT. It does this by taking the kvm->lock mutex, clearing the kvm->arch.hpte_setup_done flag and then sending an IPI to all CPUs in the host. The idea is that any VCPU task that tries to enter the guest will see that the hpte_setup_done flag is clear and therefore call kvmppc_hv_setup_htab_rma, which also takes the kvm->lock mutex and will therefore block until we release kvm->lock. However, any VCPU that is already in the guest, or is handling a hypervisor page fault or hypercall, can re-enter the guest without rechecking the hpte_setup_done flag. The IPI will cause a guest exit of any VCPUs that are currently in the guest, but does not prevent those VCPU tasks from immediately re-entering the guest. The result is that after resize_hpt_rehash_hpte() has made a HPTE absent, a hypervisor page fault can occur and make that HPTE present again. This includes updating the rmap array for the guest real page, meaning that we now have a pointer in the rmap array which connects with pointers in the old rev array but not the new rev array. In fact, if the HPT is being reduced in size, the pointer in the rmap array could point outside the bounds of the new rev array. If that happens, we can get a host crash later on such as this one: [91652.628516] Unable to handle kernel paging request for data at address 0xd0000000157fb10c [91652.628668] Faulting instruction address: 0xc0000000000e2640 [91652.628736] Oops: Kernel access of bad area, sig: 11 [#1] [91652.628789] LE SMP NR_CPUS=1024 NUMA PowerNV [91652.628847] Modules linked in: binfmt_misc vhost_net vhost tap xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack libcrc32c iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables ses enclosure scsi_transport_sas i2c_opal ipmi_powernv ipmi_devintf i2c_core ipmi_msghandler powernv_op_panel nfsd auth_rpcgss oid_registry nfs_acl lockd grace sunrpc kvm_hv kvm_pr kvm scsi_dh_alua dm_service_time dm_multipath tg3 ptp pps_core [last unloaded: stap_552b612747aec2da355051e464fa72a1_14259] [91652.629566] CPU: 136 PID: 41315 Comm: CPU 21/KVM Tainted: G O 4.14.0-1.rc4.dev.gitb27fc5c.el7.centos.ppc64le #1 [91652.629684] task: c0000007a419e400 task.stack: c0000000028d8000 [91652.629750] NIP: c0000000000e2640 LR: d00000000c36e498 CTR: c0000000000e25f0 [91652.629829] REGS: c0000000028db5d0 TRAP: 0300 Tainted: G O (4.14.0-1.rc4.dev.gitb27fc5c.el7.centos.ppc64le) [91652.629932] MSR: 900000010280b033 <SF,HV,VEC,VSX,EE,FP,ME,IR,DR,RI,LE,TM[E]> CR: 44022422 XER: 00000000 [91652.630034] CFAR: d00000000c373f84 DAR: d0000000157fb10c DSISR: 40000000 SOFTE: 1 [91652.630034] GPR00: d00000000c36e498 c0000000028db850 c000000001403900 c0000007b7960000 [91652.630034] GPR04: d0000000117fb100 d000000007ab00d8 000000000033bb10 0000000000000000 [91652.630034] GPR08: fffffffffffffe7f 801001810073bb10 d00000000e440000 d00000000c373f70 [91652.630034] GPR12: c0000000000e25f0 c00000000fdb9400 f000000003b24680 0000000000000000 [91652.630034] GPR16: 00000000000004fb 00007ff7081a0000 00000000000ec91a 000000000033bb10 [91652.630034] GPR20: 0000000000010000 00000000001b1190 0000000000000001 0000000000010000 [91652.630034] GPR24: c0000007b7ab8038 d0000000117fb100 0000000ec91a1190 c000001e6a000000 [91652.630034] GPR28: 00000000033bb100 000000000073bb10 c0000007b7960000 d0000000157fb100 [91652.630735] NIP [c0000000000e2640] kvmppc_add_revmap_chain+0x50/0x120 [91652.630806] LR [d00000000c36e498] kvmppc_book3s_hv_page_fault+0xbb8/0xc40 [kvm_hv] [91652.630884] Call Trace: [91652.630913] [c0000000028db850] [c0000000028db8b0] 0xc0000000028db8b0 (unreliable) [91652.630996] [c0000000028db8b0] [d00000000c36e498] kvmppc_book3s_hv_page_fault+0xbb8/0xc40 [kvm_hv] [91652.631091] [c0000000028db9e0] [d00000000c36a078] kvmppc_vcpu_run_hv+0xdf8/0x1300 [kvm_hv] [91652.631179] [c0000000028dbb30] [d00000000c2248c4] kvmppc_vcpu_run+0x34/0x50 [kvm] [91652.631266] [c0000000028dbb50] [d00000000c220d54] kvm_arch_vcpu_ioctl_run+0x114/0x2a0 [kvm] [91652.631351] [c0000000028dbbd0] [d00000000c2139d8] kvm_vcpu_ioctl+0x598/0x7a0 [kvm] [91652.631433] [c0000000028dbd40] [c0000000003832e0] do_vfs_ioctl+0xd0/0x8c0 [91652.631501] [c0000000028dbde0] [c000000000383ba4] SyS_ioctl+0xd4/0x130 [91652.631569] [c0000000028dbe30] [c00000000000b8e0] system_call+0x58/0x6c [91652.631635] Instruction dump: [91652.631676] fba1ffe8 fbc1fff0 fbe1fff8 f8010010 f821ffa1 2fa70000 793d0020 e9432110 [91652.631814] 7bbf26e4 7c7e1b78 7feafa14 409e0094 <807f000c> 786326e4 7c6a1a14 93a40008 [91652.631959] ---[ end trace ac85ba6db72e5b2e ]--- To fix this, we tighten up the way that the hpte_setup_done flag is checked to ensure that it does provide the guarantee that the resizing code needs. In kvmppc_run_core(), we check the hpte_setup_done flag after disabling interrupts and refuse to enter the guest if it is clear (for a HPT guest). The code that checks hpte_setup_done and calls kvmppc_hv_setup_htab_rma() is moved from kvmppc_vcpu_run_hv() to a point inside the main loop in kvmppc_run_vcpu(), ensuring that we don't just spin endlessly calling kvmppc_run_core() while hpte_setup_done is clear, but instead have a chance to block on the kvm->lock mutex. Finally we also check hpte_setup_done inside the region in kvmppc_book3s_hv_page_fault() where the HPTE is locked and we are about to update the HPTE, and bail out if it is clear. If another CPU is inside kvm_vm_ioctl_resize_hpt_commit) and has cleared hpte_setup_done, then we know that either we are looking at a HPTE that resize_hpt_rehash_hpte() has not yet processed, which is OK, or else we will see hpte_setup_done clear and refuse to update it, because of the full barrier formed by the unlock of the HPTE in resize_hpt_rehash_hpte() combined with the locking of the HPTE in kvmppc_book3s_hv_page_fault(). Fixes: 5e9859699aba ("KVM: PPC: Book3S HV: Outline of KVM-HV HPT resizing implementation") Cc: stable@vger.kernel.org # v4.10+ Reported-by: Satheesh Rajendran <satheera@in.ibm.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-08 10:44:04 +07:00
/*
* If the HPT is being resized, don't update the HPTE,
* instead let the guest retry after the resize operation is complete.
* The synchronization for mmu_ready test vs. set is provided
KVM: PPC: Book3S HV: Fix exclusion between HPT resizing and other HPT updates Commit 5e9859699aba ("KVM: PPC: Book3S HV: Outline of KVM-HV HPT resizing implementation", 2016-12-20) added code that tries to exclude any use or update of the hashed page table (HPT) while the HPT resizing code is iterating through all the entries in the HPT. It does this by taking the kvm->lock mutex, clearing the kvm->arch.hpte_setup_done flag and then sending an IPI to all CPUs in the host. The idea is that any VCPU task that tries to enter the guest will see that the hpte_setup_done flag is clear and therefore call kvmppc_hv_setup_htab_rma, which also takes the kvm->lock mutex and will therefore block until we release kvm->lock. However, any VCPU that is already in the guest, or is handling a hypervisor page fault or hypercall, can re-enter the guest without rechecking the hpte_setup_done flag. The IPI will cause a guest exit of any VCPUs that are currently in the guest, but does not prevent those VCPU tasks from immediately re-entering the guest. The result is that after resize_hpt_rehash_hpte() has made a HPTE absent, a hypervisor page fault can occur and make that HPTE present again. This includes updating the rmap array for the guest real page, meaning that we now have a pointer in the rmap array which connects with pointers in the old rev array but not the new rev array. In fact, if the HPT is being reduced in size, the pointer in the rmap array could point outside the bounds of the new rev array. If that happens, we can get a host crash later on such as this one: [91652.628516] Unable to handle kernel paging request for data at address 0xd0000000157fb10c [91652.628668] Faulting instruction address: 0xc0000000000e2640 [91652.628736] Oops: Kernel access of bad area, sig: 11 [#1] [91652.628789] LE SMP NR_CPUS=1024 NUMA PowerNV [91652.628847] Modules linked in: binfmt_misc vhost_net vhost tap xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack libcrc32c iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables ses enclosure scsi_transport_sas i2c_opal ipmi_powernv ipmi_devintf i2c_core ipmi_msghandler powernv_op_panel nfsd auth_rpcgss oid_registry nfs_acl lockd grace sunrpc kvm_hv kvm_pr kvm scsi_dh_alua dm_service_time dm_multipath tg3 ptp pps_core [last unloaded: stap_552b612747aec2da355051e464fa72a1_14259] [91652.629566] CPU: 136 PID: 41315 Comm: CPU 21/KVM Tainted: G O 4.14.0-1.rc4.dev.gitb27fc5c.el7.centos.ppc64le #1 [91652.629684] task: c0000007a419e400 task.stack: c0000000028d8000 [91652.629750] NIP: c0000000000e2640 LR: d00000000c36e498 CTR: c0000000000e25f0 [91652.629829] REGS: c0000000028db5d0 TRAP: 0300 Tainted: G O (4.14.0-1.rc4.dev.gitb27fc5c.el7.centos.ppc64le) [91652.629932] MSR: 900000010280b033 <SF,HV,VEC,VSX,EE,FP,ME,IR,DR,RI,LE,TM[E]> CR: 44022422 XER: 00000000 [91652.630034] CFAR: d00000000c373f84 DAR: d0000000157fb10c DSISR: 40000000 SOFTE: 1 [91652.630034] GPR00: d00000000c36e498 c0000000028db850 c000000001403900 c0000007b7960000 [91652.630034] GPR04: d0000000117fb100 d000000007ab00d8 000000000033bb10 0000000000000000 [91652.630034] GPR08: fffffffffffffe7f 801001810073bb10 d00000000e440000 d00000000c373f70 [91652.630034] GPR12: c0000000000e25f0 c00000000fdb9400 f000000003b24680 0000000000000000 [91652.630034] GPR16: 00000000000004fb 00007ff7081a0000 00000000000ec91a 000000000033bb10 [91652.630034] GPR20: 0000000000010000 00000000001b1190 0000000000000001 0000000000010000 [91652.630034] GPR24: c0000007b7ab8038 d0000000117fb100 0000000ec91a1190 c000001e6a000000 [91652.630034] GPR28: 00000000033bb100 000000000073bb10 c0000007b7960000 d0000000157fb100 [91652.630735] NIP [c0000000000e2640] kvmppc_add_revmap_chain+0x50/0x120 [91652.630806] LR [d00000000c36e498] kvmppc_book3s_hv_page_fault+0xbb8/0xc40 [kvm_hv] [91652.630884] Call Trace: [91652.630913] [c0000000028db850] [c0000000028db8b0] 0xc0000000028db8b0 (unreliable) [91652.630996] [c0000000028db8b0] [d00000000c36e498] kvmppc_book3s_hv_page_fault+0xbb8/0xc40 [kvm_hv] [91652.631091] [c0000000028db9e0] [d00000000c36a078] kvmppc_vcpu_run_hv+0xdf8/0x1300 [kvm_hv] [91652.631179] [c0000000028dbb30] [d00000000c2248c4] kvmppc_vcpu_run+0x34/0x50 [kvm] [91652.631266] [c0000000028dbb50] [d00000000c220d54] kvm_arch_vcpu_ioctl_run+0x114/0x2a0 [kvm] [91652.631351] [c0000000028dbbd0] [d00000000c2139d8] kvm_vcpu_ioctl+0x598/0x7a0 [kvm] [91652.631433] [c0000000028dbd40] [c0000000003832e0] do_vfs_ioctl+0xd0/0x8c0 [91652.631501] [c0000000028dbde0] [c000000000383ba4] SyS_ioctl+0xd4/0x130 [91652.631569] [c0000000028dbe30] [c00000000000b8e0] system_call+0x58/0x6c [91652.631635] Instruction dump: [91652.631676] fba1ffe8 fbc1fff0 fbe1fff8 f8010010 f821ffa1 2fa70000 793d0020 e9432110 [91652.631814] 7bbf26e4 7c7e1b78 7feafa14 409e0094 <807f000c> 786326e4 7c6a1a14 93a40008 [91652.631959] ---[ end trace ac85ba6db72e5b2e ]--- To fix this, we tighten up the way that the hpte_setup_done flag is checked to ensure that it does provide the guarantee that the resizing code needs. In kvmppc_run_core(), we check the hpte_setup_done flag after disabling interrupts and refuse to enter the guest if it is clear (for a HPT guest). The code that checks hpte_setup_done and calls kvmppc_hv_setup_htab_rma() is moved from kvmppc_vcpu_run_hv() to a point inside the main loop in kvmppc_run_vcpu(), ensuring that we don't just spin endlessly calling kvmppc_run_core() while hpte_setup_done is clear, but instead have a chance to block on the kvm->lock mutex. Finally we also check hpte_setup_done inside the region in kvmppc_book3s_hv_page_fault() where the HPTE is locked and we are about to update the HPTE, and bail out if it is clear. If another CPU is inside kvm_vm_ioctl_resize_hpt_commit) and has cleared hpte_setup_done, then we know that either we are looking at a HPTE that resize_hpt_rehash_hpte() has not yet processed, which is OK, or else we will see hpte_setup_done clear and refuse to update it, because of the full barrier formed by the unlock of the HPTE in resize_hpt_rehash_hpte() combined with the locking of the HPTE in kvmppc_book3s_hv_page_fault(). Fixes: 5e9859699aba ("KVM: PPC: Book3S HV: Outline of KVM-HV HPT resizing implementation") Cc: stable@vger.kernel.org # v4.10+ Reported-by: Satheesh Rajendran <satheera@in.ibm.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-08 10:44:04 +07:00
* by the HPTE lock.
*/
if (!kvm->arch.mmu_ready)
KVM: PPC: Book3S HV: Fix exclusion between HPT resizing and other HPT updates Commit 5e9859699aba ("KVM: PPC: Book3S HV: Outline of KVM-HV HPT resizing implementation", 2016-12-20) added code that tries to exclude any use or update of the hashed page table (HPT) while the HPT resizing code is iterating through all the entries in the HPT. It does this by taking the kvm->lock mutex, clearing the kvm->arch.hpte_setup_done flag and then sending an IPI to all CPUs in the host. The idea is that any VCPU task that tries to enter the guest will see that the hpte_setup_done flag is clear and therefore call kvmppc_hv_setup_htab_rma, which also takes the kvm->lock mutex and will therefore block until we release kvm->lock. However, any VCPU that is already in the guest, or is handling a hypervisor page fault or hypercall, can re-enter the guest without rechecking the hpte_setup_done flag. The IPI will cause a guest exit of any VCPUs that are currently in the guest, but does not prevent those VCPU tasks from immediately re-entering the guest. The result is that after resize_hpt_rehash_hpte() has made a HPTE absent, a hypervisor page fault can occur and make that HPTE present again. This includes updating the rmap array for the guest real page, meaning that we now have a pointer in the rmap array which connects with pointers in the old rev array but not the new rev array. In fact, if the HPT is being reduced in size, the pointer in the rmap array could point outside the bounds of the new rev array. If that happens, we can get a host crash later on such as this one: [91652.628516] Unable to handle kernel paging request for data at address 0xd0000000157fb10c [91652.628668] Faulting instruction address: 0xc0000000000e2640 [91652.628736] Oops: Kernel access of bad area, sig: 11 [#1] [91652.628789] LE SMP NR_CPUS=1024 NUMA PowerNV [91652.628847] Modules linked in: binfmt_misc vhost_net vhost tap xt_CHECKSUM ipt_MASQUERADE nf_nat_masquerade_ipv4 ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 xt_conntrack ip_set nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_mangle ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack libcrc32c iptable_mangle iptable_security iptable_raw ebtable_filter ebtables ip6table_filter ip6_tables ses enclosure scsi_transport_sas i2c_opal ipmi_powernv ipmi_devintf i2c_core ipmi_msghandler powernv_op_panel nfsd auth_rpcgss oid_registry nfs_acl lockd grace sunrpc kvm_hv kvm_pr kvm scsi_dh_alua dm_service_time dm_multipath tg3 ptp pps_core [last unloaded: stap_552b612747aec2da355051e464fa72a1_14259] [91652.629566] CPU: 136 PID: 41315 Comm: CPU 21/KVM Tainted: G O 4.14.0-1.rc4.dev.gitb27fc5c.el7.centos.ppc64le #1 [91652.629684] task: c0000007a419e400 task.stack: c0000000028d8000 [91652.629750] NIP: c0000000000e2640 LR: d00000000c36e498 CTR: c0000000000e25f0 [91652.629829] REGS: c0000000028db5d0 TRAP: 0300 Tainted: G O (4.14.0-1.rc4.dev.gitb27fc5c.el7.centos.ppc64le) [91652.629932] MSR: 900000010280b033 <SF,HV,VEC,VSX,EE,FP,ME,IR,DR,RI,LE,TM[E]> CR: 44022422 XER: 00000000 [91652.630034] CFAR: d00000000c373f84 DAR: d0000000157fb10c DSISR: 40000000 SOFTE: 1 [91652.630034] GPR00: d00000000c36e498 c0000000028db850 c000000001403900 c0000007b7960000 [91652.630034] GPR04: d0000000117fb100 d000000007ab00d8 000000000033bb10 0000000000000000 [91652.630034] GPR08: fffffffffffffe7f 801001810073bb10 d00000000e440000 d00000000c373f70 [91652.630034] GPR12: c0000000000e25f0 c00000000fdb9400 f000000003b24680 0000000000000000 [91652.630034] GPR16: 00000000000004fb 00007ff7081a0000 00000000000ec91a 000000000033bb10 [91652.630034] GPR20: 0000000000010000 00000000001b1190 0000000000000001 0000000000010000 [91652.630034] GPR24: c0000007b7ab8038 d0000000117fb100 0000000ec91a1190 c000001e6a000000 [91652.630034] GPR28: 00000000033bb100 000000000073bb10 c0000007b7960000 d0000000157fb100 [91652.630735] NIP [c0000000000e2640] kvmppc_add_revmap_chain+0x50/0x120 [91652.630806] LR [d00000000c36e498] kvmppc_book3s_hv_page_fault+0xbb8/0xc40 [kvm_hv] [91652.630884] Call Trace: [91652.630913] [c0000000028db850] [c0000000028db8b0] 0xc0000000028db8b0 (unreliable) [91652.630996] [c0000000028db8b0] [d00000000c36e498] kvmppc_book3s_hv_page_fault+0xbb8/0xc40 [kvm_hv] [91652.631091] [c0000000028db9e0] [d00000000c36a078] kvmppc_vcpu_run_hv+0xdf8/0x1300 [kvm_hv] [91652.631179] [c0000000028dbb30] [d00000000c2248c4] kvmppc_vcpu_run+0x34/0x50 [kvm] [91652.631266] [c0000000028dbb50] [d00000000c220d54] kvm_arch_vcpu_ioctl_run+0x114/0x2a0 [kvm] [91652.631351] [c0000000028dbbd0] [d00000000c2139d8] kvm_vcpu_ioctl+0x598/0x7a0 [kvm] [91652.631433] [c0000000028dbd40] [c0000000003832e0] do_vfs_ioctl+0xd0/0x8c0 [91652.631501] [c0000000028dbde0] [c000000000383ba4] SyS_ioctl+0xd4/0x130 [91652.631569] [c0000000028dbe30] [c00000000000b8e0] system_call+0x58/0x6c [91652.631635] Instruction dump: [91652.631676] fba1ffe8 fbc1fff0 fbe1fff8 f8010010 f821ffa1 2fa70000 793d0020 e9432110 [91652.631814] 7bbf26e4 7c7e1b78 7feafa14 409e0094 <807f000c> 786326e4 7c6a1a14 93a40008 [91652.631959] ---[ end trace ac85ba6db72e5b2e ]--- To fix this, we tighten up the way that the hpte_setup_done flag is checked to ensure that it does provide the guarantee that the resizing code needs. In kvmppc_run_core(), we check the hpte_setup_done flag after disabling interrupts and refuse to enter the guest if it is clear (for a HPT guest). The code that checks hpte_setup_done and calls kvmppc_hv_setup_htab_rma() is moved from kvmppc_vcpu_run_hv() to a point inside the main loop in kvmppc_run_vcpu(), ensuring that we don't just spin endlessly calling kvmppc_run_core() while hpte_setup_done is clear, but instead have a chance to block on the kvm->lock mutex. Finally we also check hpte_setup_done inside the region in kvmppc_book3s_hv_page_fault() where the HPTE is locked and we are about to update the HPTE, and bail out if it is clear. If another CPU is inside kvm_vm_ioctl_resize_hpt_commit) and has cleared hpte_setup_done, then we know that either we are looking at a HPTE that resize_hpt_rehash_hpte() has not yet processed, which is OK, or else we will see hpte_setup_done clear and refuse to update it, because of the full barrier formed by the unlock of the HPTE in resize_hpt_rehash_hpte() combined with the locking of the HPTE in kvmppc_book3s_hv_page_fault(). Fixes: 5e9859699aba ("KVM: PPC: Book3S HV: Outline of KVM-HV HPT resizing implementation") Cc: stable@vger.kernel.org # v4.10+ Reported-by: Satheesh Rajendran <satheera@in.ibm.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-08 10:44:04 +07:00
goto out_unlock;
if ((hnow_v & ~HPTE_V_HVLOCK) != hpte[0] || hnow_r != hpte[1] ||
rev->guest_rpte != hpte[2])
/* HPTE has been changed under us; let the guest retry */
goto out_unlock;
hpte[0] = (hpte[0] & ~HPTE_V_ABSENT) | HPTE_V_VALID;
/* Always put the HPTE in the rmap chain for the page base address */
rmap = &memslot->arch.rmap[gfn_base - memslot->base_gfn];
lock_rmap(rmap);
/* Check if we might have been invalidated; let the guest retry if so */
ret = RESUME_GUEST;
if (mmu_notifier_retry(vcpu->kvm, mmu_seq)) {
unlock_rmap(rmap);
goto out_unlock;
}
/* Only set R/C in real HPTE if set in both *rmap and guest_rpte */
rcbits = *rmap >> KVMPPC_RMAP_RC_SHIFT;
r &= rcbits | ~(HPTE_R_R | HPTE_R_C);
if (be64_to_cpu(hptep[0]) & HPTE_V_VALID) {
/* HPTE was previously valid, so we need to invalidate it */
unlock_rmap(rmap);
hptep[0] |= cpu_to_be64(HPTE_V_ABSENT);
kvmppc_invalidate_hpte(kvm, hptep, index);
/* don't lose previous R and C bits */
r |= be64_to_cpu(hptep[1]) & (HPTE_R_R | HPTE_R_C);
} else {
kvmppc_add_revmap_chain(kvm, rev, rmap, index, 0);
}
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
r = hpte_old_to_new_r(hpte[0], r);
hpte[0] = hpte_old_to_new_v(hpte[0]);
}
hptep[1] = cpu_to_be64(r);
eieio();
__unlock_hpte(hptep, hpte[0]);
asm volatile("ptesync" : : : "memory");
preempt_enable();
if (page && hpte_is_writable(r))
SetPageDirty(page);
out_put:
trace_kvm_page_fault_exit(vcpu, hpte, ret);
if (page) {
/*
* We drop pages[0] here, not page because page might
* have been set to the head page of a compound, but
* we have to drop the reference on the correct tail
* page to match the get inside gup()
*/
put_page(pages[0]);
}
return ret;
out_unlock:
__unlock_hpte(hptep, be64_to_cpu(hptep[0]));
preempt_enable();
goto out_put;
}
void kvmppc_rmap_reset(struct kvm *kvm)
{
struct kvm_memslots *slots;
struct kvm_memory_slot *memslot;
int srcu_idx;
srcu_idx = srcu_read_lock(&kvm->srcu);
slots = kvm_memslots(kvm);
kvm_for_each_memslot(memslot, slots) {
/* Mutual exclusion with kvm_unmap_hva_range etc. */
spin_lock(&kvm->mmu_lock);
/*
* This assumes it is acceptable to lose reference and
* change bits across a reset.
*/
memset(memslot->arch.rmap, 0,
memslot->npages * sizeof(*memslot->arch.rmap));
spin_unlock(&kvm->mmu_lock);
}
srcu_read_unlock(&kvm->srcu, srcu_idx);
}
typedef int (*hva_handler_fn)(struct kvm *kvm, struct kvm_memory_slot *memslot,
unsigned long gfn);
static int kvm_handle_hva_range(struct kvm *kvm,
unsigned long start,
unsigned long end,
hva_handler_fn handler)
{
int ret;
int retval = 0;
struct kvm_memslots *slots;
struct kvm_memory_slot *memslot;
slots = kvm_memslots(kvm);
kvm_for_each_memslot(memslot, slots) {
unsigned long hva_start, hva_end;
gfn_t gfn, gfn_end;
hva_start = max(start, memslot->userspace_addr);
hva_end = min(end, memslot->userspace_addr +
(memslot->npages << PAGE_SHIFT));
if (hva_start >= hva_end)
continue;
/*
* {gfn(page) | page intersects with [hva_start, hva_end)} =
* {gfn, gfn+1, ..., gfn_end-1}.
*/
gfn = hva_to_gfn_memslot(hva_start, memslot);
gfn_end = hva_to_gfn_memslot(hva_end + PAGE_SIZE - 1, memslot);
for (; gfn < gfn_end; ++gfn) {
ret = handler(kvm, memslot, gfn);
retval |= ret;
}
}
return retval;
}
static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
hva_handler_fn handler)
{
return kvm_handle_hva_range(kvm, hva, hva + 1, handler);
}
/* Must be called with both HPTE and rmap locked */
static void kvmppc_unmap_hpte(struct kvm *kvm, unsigned long i,
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
struct kvm_memory_slot *memslot,
unsigned long *rmapp, unsigned long gfn)
{
__be64 *hptep = (__be64 *) (kvm->arch.hpt.virt + (i << 4));
struct revmap_entry *rev = kvm->arch.hpt.rev;
unsigned long j, h;
unsigned long ptel, psize, rcbits;
j = rev[i].forw;
if (j == i) {
/* chain is now empty */
*rmapp &= ~(KVMPPC_RMAP_PRESENT | KVMPPC_RMAP_INDEX);
} else {
/* remove i from chain */
h = rev[i].back;
rev[h].forw = j;
rev[j].back = h;
rev[i].forw = rev[i].back = i;
*rmapp = (*rmapp & ~KVMPPC_RMAP_INDEX) | j;
}
/* Now check and modify the HPTE */
ptel = rev[i].guest_rpte;
psize = kvmppc_actual_pgsz(be64_to_cpu(hptep[0]), ptel);
if ((be64_to_cpu(hptep[0]) & HPTE_V_VALID) &&
hpte_rpn(ptel, psize) == gfn) {
hptep[0] |= cpu_to_be64(HPTE_V_ABSENT);
kvmppc_invalidate_hpte(kvm, hptep, i);
hptep[1] &= ~cpu_to_be64(HPTE_R_KEY_HI | HPTE_R_KEY_LO);
/* Harvest R and C */
rcbits = be64_to_cpu(hptep[1]) & (HPTE_R_R | HPTE_R_C);
*rmapp |= rcbits << KVMPPC_RMAP_RC_SHIFT;
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
if ((rcbits & HPTE_R_C) && memslot->dirty_bitmap)
kvmppc_update_dirty_map(memslot, gfn, psize);
if (rcbits & ~rev[i].guest_rpte) {
rev[i].guest_rpte = ptel | rcbits;
note_hpte_modification(kvm, &rev[i]);
}
}
}
static int kvm_unmap_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
unsigned long gfn)
{
unsigned long i;
__be64 *hptep;
unsigned long *rmapp;
rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
for (;;) {
lock_rmap(rmapp);
if (!(*rmapp & KVMPPC_RMAP_PRESENT)) {
unlock_rmap(rmapp);
break;
}
/*
* To avoid an ABBA deadlock with the HPTE lock bit,
* we can't spin on the HPTE lock while holding the
* rmap chain lock.
*/
i = *rmapp & KVMPPC_RMAP_INDEX;
hptep = (__be64 *) (kvm->arch.hpt.virt + (i << 4));
if (!try_lock_hpte(hptep, HPTE_V_HVLOCK)) {
/* unlock rmap before spinning on the HPTE lock */
unlock_rmap(rmapp);
while (be64_to_cpu(hptep[0]) & HPTE_V_HVLOCK)
cpu_relax();
continue;
}
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
kvmppc_unmap_hpte(kvm, i, memslot, rmapp, gfn);
unlock_rmap(rmapp);
__unlock_hpte(hptep, be64_to_cpu(hptep[0]));
}
return 0;
}
int kvm_unmap_hva_range_hv(struct kvm *kvm, unsigned long start, unsigned long end)
{
hva_handler_fn handler;
handler = kvm_is_radix(kvm) ? kvm_unmap_radix : kvm_unmap_rmapp;
kvm_handle_hva_range(kvm, start, end, handler);
return 0;
}
void kvmppc_core_flush_memslot_hv(struct kvm *kvm,
struct kvm_memory_slot *memslot)
{
unsigned long gfn;
unsigned long n;
unsigned long *rmapp;
gfn = memslot->base_gfn;
rmapp = memslot->arch.rmap;
if (kvm_is_radix(kvm)) {
kvmppc_radix_flush_memslot(kvm, memslot);
return;
}
for (n = memslot->npages; n; --n, ++gfn) {
/*
* Testing the present bit without locking is OK because
* the memslot has been marked invalid already, and hence
* no new HPTEs referencing this page can be created,
* thus the present bit can't go from 0 to 1.
*/
if (*rmapp & KVMPPC_RMAP_PRESENT)
kvm_unmap_rmapp(kvm, memslot, gfn);
++rmapp;
}
}
static int kvm_age_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
unsigned long gfn)
{
struct revmap_entry *rev = kvm->arch.hpt.rev;
unsigned long head, i, j;
__be64 *hptep;
int ret = 0;
unsigned long *rmapp;
rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
retry:
lock_rmap(rmapp);
if (*rmapp & KVMPPC_RMAP_REFERENCED) {
*rmapp &= ~KVMPPC_RMAP_REFERENCED;
ret = 1;
}
if (!(*rmapp & KVMPPC_RMAP_PRESENT)) {
unlock_rmap(rmapp);
return ret;
}
i = head = *rmapp & KVMPPC_RMAP_INDEX;
do {
hptep = (__be64 *) (kvm->arch.hpt.virt + (i << 4));
j = rev[i].forw;
/* If this HPTE isn't referenced, ignore it */
if (!(be64_to_cpu(hptep[1]) & HPTE_R_R))
continue;
if (!try_lock_hpte(hptep, HPTE_V_HVLOCK)) {
/* unlock rmap before spinning on the HPTE lock */
unlock_rmap(rmapp);
while (be64_to_cpu(hptep[0]) & HPTE_V_HVLOCK)
cpu_relax();
goto retry;
}
/* Now check and modify the HPTE */
if ((be64_to_cpu(hptep[0]) & HPTE_V_VALID) &&
(be64_to_cpu(hptep[1]) & HPTE_R_R)) {
kvmppc_clear_ref_hpte(kvm, hptep, i);
if (!(rev[i].guest_rpte & HPTE_R_R)) {
rev[i].guest_rpte |= HPTE_R_R;
note_hpte_modification(kvm, &rev[i]);
}
ret = 1;
}
__unlock_hpte(hptep, be64_to_cpu(hptep[0]));
} while ((i = j) != head);
unlock_rmap(rmapp);
return ret;
}
int kvm_age_hva_hv(struct kvm *kvm, unsigned long start, unsigned long end)
{
hva_handler_fn handler;
handler = kvm_is_radix(kvm) ? kvm_age_radix : kvm_age_rmapp;
return kvm_handle_hva_range(kvm, start, end, handler);
}
static int kvm_test_age_rmapp(struct kvm *kvm, struct kvm_memory_slot *memslot,
unsigned long gfn)
{
struct revmap_entry *rev = kvm->arch.hpt.rev;
unsigned long head, i, j;
unsigned long *hp;
int ret = 1;
unsigned long *rmapp;
rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
if (*rmapp & KVMPPC_RMAP_REFERENCED)
return 1;
lock_rmap(rmapp);
if (*rmapp & KVMPPC_RMAP_REFERENCED)
goto out;
if (*rmapp & KVMPPC_RMAP_PRESENT) {
i = head = *rmapp & KVMPPC_RMAP_INDEX;
do {
hp = (unsigned long *)(kvm->arch.hpt.virt + (i << 4));
j = rev[i].forw;
if (be64_to_cpu(hp[1]) & HPTE_R_R)
goto out;
} while ((i = j) != head);
}
ret = 0;
out:
unlock_rmap(rmapp);
return ret;
}
int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva)
{
hva_handler_fn handler;
handler = kvm_is_radix(kvm) ? kvm_test_age_radix : kvm_test_age_rmapp;
return kvm_handle_hva(kvm, hva, handler);
}
void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte)
{
hva_handler_fn handler;
handler = kvm_is_radix(kvm) ? kvm_unmap_radix : kvm_unmap_rmapp;
kvm_handle_hva(kvm, hva, handler);
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
}
static int vcpus_running(struct kvm *kvm)
{
return atomic_read(&kvm->arch.vcpus_running) != 0;
}
/*
* Returns the number of system pages that are dirty.
* This can be more than 1 if we find a huge-page HPTE.
*/
static int kvm_test_clear_dirty_npages(struct kvm *kvm, unsigned long *rmapp)
{
struct revmap_entry *rev = kvm->arch.hpt.rev;
unsigned long head, i, j;
unsigned long n;
unsigned long v, r;
__be64 *hptep;
int npages_dirty = 0;
retry:
lock_rmap(rmapp);
if (!(*rmapp & KVMPPC_RMAP_PRESENT)) {
unlock_rmap(rmapp);
return npages_dirty;
}
i = head = *rmapp & KVMPPC_RMAP_INDEX;
do {
unsigned long hptep1;
hptep = (__be64 *) (kvm->arch.hpt.virt + (i << 4));
j = rev[i].forw;
/*
* Checking the C (changed) bit here is racy since there
* is no guarantee about when the hardware writes it back.
* If the HPTE is not writable then it is stable since the
* page can't be written to, and we would have done a tlbie
* (which forces the hardware to complete any writeback)
* when making the HPTE read-only.
* If vcpus are running then this call is racy anyway
* since the page could get dirtied subsequently, so we
* expect there to be a further call which would pick up
* any delayed C bit writeback.
* Otherwise we need to do the tlbie even if C==0 in
* order to pick up any delayed writeback of C.
*/
hptep1 = be64_to_cpu(hptep[1]);
if (!(hptep1 & HPTE_R_C) &&
(!hpte_is_writable(hptep1) || vcpus_running(kvm)))
continue;
if (!try_lock_hpte(hptep, HPTE_V_HVLOCK)) {
/* unlock rmap before spinning on the HPTE lock */
unlock_rmap(rmapp);
while (hptep[0] & cpu_to_be64(HPTE_V_HVLOCK))
cpu_relax();
goto retry;
}
/* Now check and modify the HPTE */
if (!(hptep[0] & cpu_to_be64(HPTE_V_VALID))) {
__unlock_hpte(hptep, be64_to_cpu(hptep[0]));
continue;
}
/* need to make it temporarily absent so C is stable */
hptep[0] |= cpu_to_be64(HPTE_V_ABSENT);
kvmppc_invalidate_hpte(kvm, hptep, i);
v = be64_to_cpu(hptep[0]);
r = be64_to_cpu(hptep[1]);
if (r & HPTE_R_C) {
hptep[1] = cpu_to_be64(r & ~HPTE_R_C);
if (!(rev[i].guest_rpte & HPTE_R_C)) {
rev[i].guest_rpte |= HPTE_R_C;
note_hpte_modification(kvm, &rev[i]);
}
n = kvmppc_actual_pgsz(v, r);
n = (n + PAGE_SIZE - 1) >> PAGE_SHIFT;
if (n > npages_dirty)
npages_dirty = n;
eieio();
}
v &= ~HPTE_V_ABSENT;
v |= HPTE_V_VALID;
__unlock_hpte(hptep, v);
} while ((i = j) != head);
unlock_rmap(rmapp);
return npages_dirty;
}
void kvmppc_harvest_vpa_dirty(struct kvmppc_vpa *vpa,
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
struct kvm_memory_slot *memslot,
unsigned long *map)
{
unsigned long gfn;
if (!vpa->dirty || !vpa->pinned_addr)
return;
gfn = vpa->gpa >> PAGE_SHIFT;
if (gfn < memslot->base_gfn ||
gfn >= memslot->base_gfn + memslot->npages)
return;
vpa->dirty = false;
if (map)
__set_bit_le(gfn - memslot->base_gfn, map);
}
long kvmppc_hv_get_dirty_log_hpt(struct kvm *kvm,
struct kvm_memory_slot *memslot, unsigned long *map)
{
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
unsigned long i;
unsigned long *rmapp;
preempt_disable();
rmapp = memslot->arch.rmap;
for (i = 0; i < memslot->npages; ++i) {
int npages = kvm_test_clear_dirty_npages(kvm, rmapp);
/*
* Note that if npages > 0 then i must be a multiple of npages,
* since we always put huge-page HPTEs in the rmap chain
* corresponding to their page base address.
*/
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
if (npages)
set_dirty_bits(map, i, npages);
++rmapp;
}
preempt_enable();
return 0;
}
void *kvmppc_pin_guest_page(struct kvm *kvm, unsigned long gpa,
unsigned long *nb_ret)
{
struct kvm_memory_slot *memslot;
unsigned long gfn = gpa >> PAGE_SHIFT;
struct page *page, *pages[1];
int npages;
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
unsigned long hva, offset;
int srcu_idx;
srcu_idx = srcu_read_lock(&kvm->srcu);
memslot = gfn_to_memslot(kvm, gfn);
if (!memslot || (memslot->flags & KVM_MEMSLOT_INVALID))
goto err;
hva = gfn_to_hva_memslot(memslot, gfn);
mm/gup: change GUP fast to use flags rather than a write 'bool' To facilitate additional options to get_user_pages_fast() change the singular write parameter to be gup_flags. This patch does not change any functionality. New functionality will follow in subsequent patches. Some of the get_user_pages_fast() call sites were unchanged because they already passed FOLL_WRITE or 0 for the write parameter. NOTE: It was suggested to change the ordering of the get_user_pages_fast() arguments to ensure that callers were converted. This breaks the current GUP call site convention of having the returned pages be the final parameter. So the suggestion was rejected. Link: http://lkml.kernel.org/r/20190328084422.29911-4-ira.weiny@intel.com Link: http://lkml.kernel.org/r/20190317183438.2057-4-ira.weiny@intel.com Signed-off-by: Ira Weiny <ira.weiny@intel.com> Reviewed-by: Mike Marshall <hubcap@omnibond.com> Cc: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Borislav Petkov <bp@alien8.de> Cc: Dan Williams <dan.j.williams@intel.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: James Hogan <jhogan@kernel.org> Cc: Jason Gunthorpe <jgg@ziepe.ca> Cc: John Hubbard <jhubbard@nvidia.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Rich Felker <dalias@libc.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 07:17:11 +07:00
npages = get_user_pages_fast(hva, 1, FOLL_WRITE, pages);
if (npages < 1)
goto err;
page = pages[0];
srcu_read_unlock(&kvm->srcu, srcu_idx);
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
offset = gpa & (PAGE_SIZE - 1);
if (nb_ret)
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
*nb_ret = PAGE_SIZE - offset;
return page_address(page) + offset;
err:
srcu_read_unlock(&kvm->srcu, srcu_idx);
return NULL;
}
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
void kvmppc_unpin_guest_page(struct kvm *kvm, void *va, unsigned long gpa,
bool dirty)
{
struct page *page = virt_to_page(va);
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
struct kvm_memory_slot *memslot;
unsigned long gfn;
int srcu_idx;
put_page(page);
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
if (!dirty)
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
return;
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
/* We need to mark this page dirty in the memslot dirty_bitmap, if any */
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
gfn = gpa >> PAGE_SHIFT;
srcu_idx = srcu_read_lock(&kvm->srcu);
memslot = gfn_to_memslot(kvm, gfn);
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
if (memslot && memslot->dirty_bitmap)
set_bit_le(gfn - memslot->base_gfn, memslot->dirty_bitmap);
KVM: PPC: Book3S HV: Report VPA and DTL modifications in dirty map At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications done by the host to the virtual processor areas (VPAs) and dispatch trace logs (DTLs) registered by the guest. This is because those modifications are done either in real mode or in the host kernel context, and in neither case does the access go through the guest's HPT, and thus no change (C) bit gets set in the guest's HPT. However, the changes done by the host do need to be tracked so that the modified pages get transferred when doing live migration. In order to track these modifications, this adds a dirty flag to the struct representing the VPA/DTL areas, and arranges to set the flag when the VPA/DTL gets modified by the host. Then, when we are collecting the dirty log, we also check the dirty flags for the VPA and DTL for each vcpu and set the relevant bit in the dirty log if necessary. Doing this also means we now need to keep track of the guest physical address of the VPA/DTL areas. So as not to lose track of modifications to a VPA/DTL area when it gets unregistered, or when a new area gets registered in its place, we need to transfer the dirty state to the rmap chain. This adds code to kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify that code, we now require that all VPA, DTL and SLB shadow buffer areas fit within a single host page. Guests already comply with this requirement because pHyp requires that these areas not cross a 4k boundary. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-04-19 02:51:04 +07:00
srcu_read_unlock(&kvm->srcu, srcu_idx);
}
/*
* HPT resizing
*/
static int resize_hpt_allocate(struct kvm_resize_hpt *resize)
{
int rc;
rc = kvmppc_allocate_hpt(&resize->hpt, resize->order);
if (rc < 0)
return rc;
resize_hpt_debug(resize, "resize_hpt_allocate(): HPT @ 0x%lx\n",
resize->hpt.virt);
return 0;
}
static unsigned long resize_hpt_rehash_hpte(struct kvm_resize_hpt *resize,
unsigned long idx)
{
struct kvm *kvm = resize->kvm;
struct kvm_hpt_info *old = &kvm->arch.hpt;
struct kvm_hpt_info *new = &resize->hpt;
unsigned long old_hash_mask = (1ULL << (old->order - 7)) - 1;
unsigned long new_hash_mask = (1ULL << (new->order - 7)) - 1;
__be64 *hptep, *new_hptep;
unsigned long vpte, rpte, guest_rpte;
int ret;
struct revmap_entry *rev;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
unsigned long apsize, avpn, pteg, hash;
unsigned long new_idx, new_pteg, replace_vpte;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
int pshift;
hptep = (__be64 *)(old->virt + (idx << 4));
/* Guest is stopped, so new HPTEs can't be added or faulted
* in, only unmapped or altered by host actions. So, it's
* safe to check this before we take the HPTE lock */
vpte = be64_to_cpu(hptep[0]);
if (!(vpte & HPTE_V_VALID) && !(vpte & HPTE_V_ABSENT))
return 0; /* nothing to do */
while (!try_lock_hpte(hptep, HPTE_V_HVLOCK))
cpu_relax();
vpte = be64_to_cpu(hptep[0]);
ret = 0;
if (!(vpte & HPTE_V_VALID) && !(vpte & HPTE_V_ABSENT))
/* Nothing to do */
goto out;
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
rpte = be64_to_cpu(hptep[1]);
vpte = hpte_new_to_old_v(vpte, rpte);
}
/* Unmap */
rev = &old->rev[idx];
guest_rpte = rev->guest_rpte;
ret = -EIO;
apsize = kvmppc_actual_pgsz(vpte, guest_rpte);
if (!apsize)
goto out;
if (vpte & HPTE_V_VALID) {
unsigned long gfn = hpte_rpn(guest_rpte, apsize);
int srcu_idx = srcu_read_lock(&kvm->srcu);
struct kvm_memory_slot *memslot =
__gfn_to_memslot(kvm_memslots(kvm), gfn);
if (memslot) {
unsigned long *rmapp;
rmapp = &memslot->arch.rmap[gfn - memslot->base_gfn];
lock_rmap(rmapp);
KVM: PPC: Book3S HV: Unify dirty page map between HPT and radix Currently, the HPT code in HV KVM maintains a dirty bit per guest page in the rmap array, whether or not dirty page tracking has been enabled for the memory slot. In contrast, the radix code maintains a dirty bit per guest page in memslot->dirty_bitmap, and only does so when dirty page tracking has been enabled. This changes the HPT code to maintain the dirty bits in the memslot dirty_bitmap like radix does. This results in slightly less code overall, and will mean that we do not lose the dirty bits when transitioning between HPT and radix mode in future. There is one minor change to behaviour as a result. With HPT, when dirty tracking was enabled for a memslot, we would previously clear all the dirty bits at that point (both in the HPT entries and in the rmap arrays), meaning that a KVM_GET_DIRTY_LOG ioctl immediately following would show no pages as dirty (assuming no vcpus have run in the meantime). With this change, the dirty bits on HPT entries are not cleared at the point where dirty tracking is enabled, so KVM_GET_DIRTY_LOG would show as dirty any guest pages that are resident in the HPT and dirty. This is consistent with what happens on radix. This also fixes a bug in the mark_pages_dirty() function for radix (in the sense that the function no longer exists). In the case where a large page of 64 normal pages or more is marked dirty, the addressing of the dirty bitmap was incorrect and could write past the end of the bitmap. Fortunately this case was never hit in practice because a 2MB large page is only 32 x 64kB pages, and we don't support backing the guest with 1GB huge pages at this point. Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-10-26 12:39:19 +07:00
kvmppc_unmap_hpte(kvm, idx, memslot, rmapp, gfn);
unlock_rmap(rmapp);
}
srcu_read_unlock(&kvm->srcu, srcu_idx);
}
/* Reload PTE after unmap */
vpte = be64_to_cpu(hptep[0]);
BUG_ON(vpte & HPTE_V_VALID);
BUG_ON(!(vpte & HPTE_V_ABSENT));
ret = 0;
if (!(vpte & HPTE_V_BOLTED))
goto out;
rpte = be64_to_cpu(hptep[1]);
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
vpte = hpte_new_to_old_v(vpte, rpte);
rpte = hpte_new_to_old_r(rpte);
}
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
pshift = kvmppc_hpte_base_page_shift(vpte, rpte);
avpn = HPTE_V_AVPN_VAL(vpte) & ~(((1ul << pshift) - 1) >> 23);
pteg = idx / HPTES_PER_GROUP;
if (vpte & HPTE_V_SECONDARY)
pteg = ~pteg;
if (!(vpte & HPTE_V_1TB_SEG)) {
unsigned long offset, vsid;
/* We only have 28 - 23 bits of offset in avpn */
offset = (avpn & 0x1f) << 23;
vsid = avpn >> 5;
/* We can find more bits from the pteg value */
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
if (pshift < 23)
offset |= ((vsid ^ pteg) & old_hash_mask) << pshift;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
hash = vsid ^ (offset >> pshift);
} else {
unsigned long offset, vsid;
/* We only have 40 - 23 bits of seg_off in avpn */
offset = (avpn & 0x1ffff) << 23;
vsid = avpn >> 17;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
if (pshift < 23)
offset |= ((vsid ^ (vsid << 25) ^ pteg) & old_hash_mask) << pshift;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
hash = vsid ^ (vsid << 25) ^ (offset >> pshift);
}
new_pteg = hash & new_hash_mask;
if (vpte & HPTE_V_SECONDARY)
new_pteg = ~hash & new_hash_mask;
new_idx = new_pteg * HPTES_PER_GROUP + (idx % HPTES_PER_GROUP);
new_hptep = (__be64 *)(new->virt + (new_idx << 4));
replace_vpte = be64_to_cpu(new_hptep[0]);
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
unsigned long replace_rpte = be64_to_cpu(new_hptep[1]);
replace_vpte = hpte_new_to_old_v(replace_vpte, replace_rpte);
}
if (replace_vpte & (HPTE_V_VALID | HPTE_V_ABSENT)) {
BUG_ON(new->order >= old->order);
if (replace_vpte & HPTE_V_BOLTED) {
if (vpte & HPTE_V_BOLTED)
/* Bolted collision, nothing we can do */
ret = -ENOSPC;
/* Discard the new HPTE */
goto out;
}
/* Discard the previous HPTE */
}
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
rpte = hpte_old_to_new_r(vpte, rpte);
vpte = hpte_old_to_new_v(vpte);
}
new_hptep[1] = cpu_to_be64(rpte);
new->rev[new_idx].guest_rpte = guest_rpte;
/* No need for a barrier, since new HPT isn't active */
new_hptep[0] = cpu_to_be64(vpte);
unlock_hpte(new_hptep, vpte);
out:
unlock_hpte(hptep, vpte);
return ret;
}
static int resize_hpt_rehash(struct kvm_resize_hpt *resize)
{
struct kvm *kvm = resize->kvm;
unsigned long i;
int rc;
for (i = 0; i < kvmppc_hpt_npte(&kvm->arch.hpt); i++) {
rc = resize_hpt_rehash_hpte(resize, i);
if (rc != 0)
return rc;
}
return 0;
}
static void resize_hpt_pivot(struct kvm_resize_hpt *resize)
{
struct kvm *kvm = resize->kvm;
struct kvm_hpt_info hpt_tmp;
/* Exchange the pending tables in the resize structure with
* the active tables */
resize_hpt_debug(resize, "resize_hpt_pivot()\n");
spin_lock(&kvm->mmu_lock);
asm volatile("ptesync" : : : "memory");
hpt_tmp = kvm->arch.hpt;
kvmppc_set_hpt(kvm, &resize->hpt);
resize->hpt = hpt_tmp;
spin_unlock(&kvm->mmu_lock);
synchronize_srcu_expedited(&kvm->srcu);
if (cpu_has_feature(CPU_FTR_ARCH_300))
kvmppc_setup_partition_table(kvm);
resize_hpt_debug(resize, "resize_hpt_pivot() done\n");
}
static void resize_hpt_release(struct kvm *kvm, struct kvm_resize_hpt *resize)
{
if (WARN_ON(!mutex_is_locked(&kvm->arch.mmu_setup_lock)))
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
return;
if (!resize)
return;
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
if (resize->error != -EBUSY) {
if (resize->hpt.virt)
kvmppc_free_hpt(&resize->hpt);
kfree(resize);
}
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
if (kvm->arch.resize_hpt == resize)
kvm->arch.resize_hpt = NULL;
}
static void resize_hpt_prepare_work(struct work_struct *work)
{
struct kvm_resize_hpt *resize = container_of(work,
struct kvm_resize_hpt,
work);
struct kvm *kvm = resize->kvm;
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
int err = 0;
if (WARN_ON(resize->error != -EBUSY))
return;
mutex_lock(&kvm->arch.mmu_setup_lock);
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
/* Request is still current? */
if (kvm->arch.resize_hpt == resize) {
/* We may request large allocations here:
* do not sleep with kvm->arch.mmu_setup_lock held for a while.
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
*/
mutex_unlock(&kvm->arch.mmu_setup_lock);
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
resize_hpt_debug(resize, "resize_hpt_prepare_work(): order = %d\n",
resize->order);
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
err = resize_hpt_allocate(resize);
/* We have strict assumption about -EBUSY
* when preparing for HPT resize.
*/
if (WARN_ON(err == -EBUSY))
err = -EINPROGRESS;
mutex_lock(&kvm->arch.mmu_setup_lock);
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
/* It is possible that kvm->arch.resize_hpt != resize
* after we grab kvm->arch.mmu_setup_lock again.
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
*/
}
resize->error = err;
KVM: PPC: Book3S HV: Fix use after free in case of multiple resize requests When serving multiple resize requests following could happen: CPU0 CPU1 ---- ---- kvm_vm_ioctl_resize_hpt_prepare(1); -> schedule_work() /* system_rq might be busy: delay */ kvm_vm_ioctl_resize_hpt_prepare(2); mutex_lock(); if (resize) { ... release_hpt_resize(); } ... resize_hpt_prepare_work() -> schedule_work() { mutex_unlock() /* resize->kvm could be wrong */ struct kvm *kvm = resize->kvm; mutex_lock(&kvm->lock); <<<< UAF ... } i.e. a second resize request with different order could be started by kvm_vm_ioctl_resize_hpt_prepare(), causing the previous request to be free()d when there's still an active worker thread which will try to access it. This leads to a use after free in point marked with UAF on the diagram above. To prevent this from happening, instead of unconditionally releasing a pre-existing resize structure from the prepare ioctl(), we check if the existing structure has an in-progress worker. We do that by checking if the resize->error == -EBUSY, which is safe because the resize->error field is protected by the kvm->lock. If there is an active worker, instead of releasing, we mark the structure as stale by unlinking it from kvm_struct. In the worker thread we check for a stale structure (with kvm->lock held), and in that case abort, releasing the stale structure ourself. We make the check both before and the actual allocation. Strictly, only the check afterwards is needed, the check before is an optimization: if the structure happens to become stale before the worker thread is dispatched, rather than during the allocation, it means we can avoid allocating then immediately freeing a potentially substantial amount of memory. This fixes following or similar host kernel crash message: [ 635.277361] Unable to handle kernel paging request for data at address 0x00000000 [ 635.277438] Faulting instruction address: 0xc00000000052f568 [ 635.277446] Oops: Kernel access of bad area, sig: 11 [#1] [ 635.277451] SMP NR_CPUS=2048 NUMA PowerNV [ 635.277470] Modules linked in: xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter nfsv3 nfs_acl nfs lockd grace fscache kvm_hv kvm rpcrdma sunrpc ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ext4 ib_srp scsi_transport_srp ib_ipoib mbcache jbd2 rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm ocrdma(T) ib_core ses enclosure scsi_transport_sas sg shpchp leds_powernv ibmpowernv i2c_opal i2c_core powernv_rng ipmi_powernv ipmi_devintf ipmi_msghandler ip_tables xfs libcrc32c sr_mod sd_mod cdrom lpfc nvme_fc(T) nvme_fabrics nvme_core ipr nvmet_fc(T) tg3 nvmet libata be2net crc_t10dif crct10dif_generic scsi_transport_fc ptp scsi_tgt pps_core crct10dif_common dm_mirror dm_region_hash dm_log dm_mod [ 635.278687] CPU: 40 PID: 749 Comm: kworker/40:1 Tainted: G ------------ T 3.10.0.bz1510771+ #1 [ 635.278782] Workqueue: events resize_hpt_prepare_work [kvm_hv] [ 635.278851] task: c0000007e6840000 ti: c0000007e9180000 task.ti: c0000007e9180000 [ 635.278919] NIP: c00000000052f568 LR: c0000000009ea310 CTR: c0000000009ea4f0 [ 635.278988] REGS: c0000007e91837f0 TRAP: 0300 Tainted: G ------------ T (3.10.0.bz1510771+) [ 635.279077] MSR: 9000000100009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 24002022 XER: 00000000 [ 635.279248] CFAR: c000000000009368 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c0000000009ea310 c0000007e9183a70 c000000001250b00 c0000007e9183b10 GPR04: 0000000000000000 0000000000000000 c0000007e9183650 0000000000000000 GPR08: c0000007ffff7b80 00000000ffffffff 0000000080000028 d00000000d2529a0 GPR12: 0000000000002200 c000000007b56800 c000000000120028 c0000007f135bb40 GPR16: 0000000000000000 c000000005c1e018 c000000005c1e018 0000000000000000 GPR20: 0000000000000001 c0000000011bf778 0000000000000001 fffffffffffffef7 GPR24: 0000000000000000 c000000f1e262e50 0000000000000002 c0000007e9180000 GPR28: c000000f1e262e4c c000000f1e262e50 0000000000000000 c0000007e9183b10 [ 635.280149] NIP [c00000000052f568] __list_add+0x38/0x110 [ 635.280197] LR [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280253] Call Trace: [ 635.280277] [c0000007e9183af0] [c0000000009ea310] __mutex_lock_slowpath+0xe0/0x2c0 [ 635.280356] [c0000007e9183b70] [c0000000009ea554] mutex_lock+0x64/0x70 [ 635.280426] [c0000007e9183ba0] [d00000000d24da04] resize_hpt_prepare_work+0xe4/0x1c0 [kvm_hv] [ 635.280507] [c0000007e9183c40] [c000000000113c0c] process_one_work+0x1dc/0x680 [ 635.280587] [c0000007e9183ce0] [c000000000114250] worker_thread+0x1a0/0x520 [ 635.280655] [c0000007e9183d80] [c00000000012010c] kthread+0xec/0x100 [ 635.280724] [c0000007e9183e30] [c00000000000a4b8] ret_from_kernel_thread+0x5c/0xa4 [ 635.280814] Instruction dump: [ 635.280880] 7c0802a6 fba1ffe8 fbc1fff0 7cbd2b78 fbe1fff8 7c9e2378 7c7f1b78 f8010010 [ 635.281099] f821ff81 e8a50008 7fa52040 40de00b8 <e8be0000> 7fbd2840 40de008c 7fbff040 [ 635.281324] ---[ end trace b628b73449719b9d ]--- Cc: stable@vger.kernel.org # v4.10+ Fixes: b5baa6877315 ("KVM: PPC: Book3S HV: KVM-HV HPT resizing implementation") Signed-off-by: Serhii Popovych <spopovyc@redhat.com> [dwg: Replaced BUG_ON()s with WARN_ONs() and reworded commit message for clarity] Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-12-04 21:36:42 +07:00
if (kvm->arch.resize_hpt != resize)
resize_hpt_release(kvm, resize);
mutex_unlock(&kvm->arch.mmu_setup_lock);
}
long kvm_vm_ioctl_resize_hpt_prepare(struct kvm *kvm,
struct kvm_ppc_resize_hpt *rhpt)
{
unsigned long flags = rhpt->flags;
unsigned long shift = rhpt->shift;
struct kvm_resize_hpt *resize;
int ret;
if (flags != 0 || kvm_is_radix(kvm))
return -EINVAL;
if (shift && ((shift < 18) || (shift > 46)))
return -EINVAL;
mutex_lock(&kvm->arch.mmu_setup_lock);
resize = kvm->arch.resize_hpt;
if (resize) {
if (resize->order == shift) {
/* Suitable resize in progress? */
ret = resize->error;
if (ret == -EBUSY)
ret = 100; /* estimated time in ms */
else if (ret)
resize_hpt_release(kvm, resize);
goto out;
}
/* not suitable, cancel it */
resize_hpt_release(kvm, resize);
}
ret = 0;
if (!shift)
goto out; /* nothing to do */
/* start new resize */
resize = kzalloc(sizeof(*resize), GFP_KERNEL);
if (!resize) {
ret = -ENOMEM;
goto out;
}
resize->error = -EBUSY;
resize->order = shift;
resize->kvm = kvm;
INIT_WORK(&resize->work, resize_hpt_prepare_work);
kvm->arch.resize_hpt = resize;
schedule_work(&resize->work);
ret = 100; /* estimated time in ms */
out:
mutex_unlock(&kvm->arch.mmu_setup_lock);
return ret;
}
static void resize_hpt_boot_vcpu(void *opaque)
{
/* Nothing to do, just force a KVM exit */
}
long kvm_vm_ioctl_resize_hpt_commit(struct kvm *kvm,
struct kvm_ppc_resize_hpt *rhpt)
{
unsigned long flags = rhpt->flags;
unsigned long shift = rhpt->shift;
struct kvm_resize_hpt *resize;
long ret;
if (flags != 0 || kvm_is_radix(kvm))
return -EINVAL;
if (shift && ((shift < 18) || (shift > 46)))
return -EINVAL;
mutex_lock(&kvm->arch.mmu_setup_lock);
resize = kvm->arch.resize_hpt;
/* This shouldn't be possible */
ret = -EIO;
if (WARN_ON(!kvm->arch.mmu_ready))
goto out_no_hpt;
/* Stop VCPUs from running while we mess with the HPT */
kvm->arch.mmu_ready = 0;
smp_mb();
/* Boot all CPUs out of the guest so they re-read
* mmu_ready */
on_each_cpu(resize_hpt_boot_vcpu, NULL, 1);
ret = -ENXIO;
if (!resize || (resize->order != shift))
goto out;
ret = resize->error;
if (ret)
goto out;
ret = resize_hpt_rehash(resize);
if (ret)
goto out;
resize_hpt_pivot(resize);
out:
/* Let VCPUs run again */
kvm->arch.mmu_ready = 1;
smp_mb();
out_no_hpt:
resize_hpt_release(kvm, resize);
mutex_unlock(&kvm->arch.mmu_setup_lock);
return ret;
}
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
/*
* Functions for reading and writing the hash table via reads and
* writes on a file descriptor.
*
* Reads return the guest view of the hash table, which has to be
* pieced together from the real hash table and the guest_rpte
* values in the revmap array.
*
* On writes, each HPTE written is considered in turn, and if it
* is valid, it is written to the HPT as if an H_ENTER with the
* exact flag set was done. When the invalid count is non-zero
* in the header written to the stream, the kernel will make
* sure that that many HPTEs are invalid, and invalidate them
* if not.
*/
struct kvm_htab_ctx {
unsigned long index;
unsigned long flags;
struct kvm *kvm;
int first_pass;
};
#define HPTE_SIZE (2 * sizeof(unsigned long))
/*
* Returns 1 if this HPT entry has been modified or has pending
* R/C bit changes.
*/
static int hpte_dirty(struct revmap_entry *revp, __be64 *hptp)
{
unsigned long rcbits_unset;
if (revp->guest_rpte & HPTE_GR_MODIFIED)
return 1;
/* Also need to consider changes in reference and changed bits */
rcbits_unset = ~revp->guest_rpte & (HPTE_R_R | HPTE_R_C);
if ((be64_to_cpu(hptp[0]) & HPTE_V_VALID) &&
(be64_to_cpu(hptp[1]) & rcbits_unset))
return 1;
return 0;
}
static long record_hpte(unsigned long flags, __be64 *hptp,
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
unsigned long *hpte, struct revmap_entry *revp,
int want_valid, int first_pass)
{
unsigned long v, r, hr;
unsigned long rcbits_unset;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
int ok = 1;
int valid, dirty;
/* Unmodified entries are uninteresting except on the first pass */
dirty = hpte_dirty(revp, hptp);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
if (!first_pass && !dirty)
return 0;
valid = 0;
if (be64_to_cpu(hptp[0]) & (HPTE_V_VALID | HPTE_V_ABSENT)) {
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
valid = 1;
if ((flags & KVM_GET_HTAB_BOLTED_ONLY) &&
!(be64_to_cpu(hptp[0]) & HPTE_V_BOLTED))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
valid = 0;
}
if (valid != want_valid)
return 0;
v = r = 0;
if (valid || dirty) {
/* lock the HPTE so it's stable and read it */
preempt_disable();
while (!try_lock_hpte(hptp, HPTE_V_HVLOCK))
cpu_relax();
v = be64_to_cpu(hptp[0]);
hr = be64_to_cpu(hptp[1]);
if (cpu_has_feature(CPU_FTR_ARCH_300)) {
v = hpte_new_to_old_v(v, hr);
hr = hpte_new_to_old_r(hr);
}
/* re-evaluate valid and dirty from synchronized HPTE value */
valid = !!(v & HPTE_V_VALID);
dirty = !!(revp->guest_rpte & HPTE_GR_MODIFIED);
/* Harvest R and C into guest view if necessary */
rcbits_unset = ~revp->guest_rpte & (HPTE_R_R | HPTE_R_C);
if (valid && (rcbits_unset & hr)) {
revp->guest_rpte |= (hr &
(HPTE_R_R | HPTE_R_C)) | HPTE_GR_MODIFIED;
dirty = 1;
}
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
if (v & HPTE_V_ABSENT) {
v &= ~HPTE_V_ABSENT;
v |= HPTE_V_VALID;
valid = 1;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
}
if ((flags & KVM_GET_HTAB_BOLTED_ONLY) && !(v & HPTE_V_BOLTED))
valid = 0;
r = revp->guest_rpte;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
/* only clear modified if this is the right sort of entry */
if (valid == want_valid && dirty) {
r &= ~HPTE_GR_MODIFIED;
revp->guest_rpte = r;
}
unlock_hpte(hptp, be64_to_cpu(hptp[0]));
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
preempt_enable();
if (!(valid == want_valid && (first_pass || dirty)))
ok = 0;
}
hpte[0] = cpu_to_be64(v);
hpte[1] = cpu_to_be64(r);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
return ok;
}
static ssize_t kvm_htab_read(struct file *file, char __user *buf,
size_t count, loff_t *ppos)
{
struct kvm_htab_ctx *ctx = file->private_data;
struct kvm *kvm = ctx->kvm;
struct kvm_get_htab_header hdr;
__be64 *hptp;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
struct revmap_entry *revp;
unsigned long i, nb, nw;
unsigned long __user *lbuf;
struct kvm_get_htab_header __user *hptr;
unsigned long flags;
int first_pass;
unsigned long hpte[2];
Remove 'type' argument from access_ok() function Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument of the user address range verification function since we got rid of the old racy i386-only code to walk page tables by hand. It existed because the original 80386 would not honor the write protect bit when in kernel mode, so you had to do COW by hand before doing any user access. But we haven't supported that in a long time, and these days the 'type' argument is a purely historical artifact. A discussion about extending 'user_access_begin()' to do the range checking resulted this patch, because there is no way we're going to move the old VERIFY_xyz interface to that model. And it's best done at the end of the merge window when I've done most of my merges, so let's just get this done once and for all. This patch was mostly done with a sed-script, with manual fix-ups for the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form. There were a couple of notable cases: - csky still had the old "verify_area()" name as an alias. - the iter_iov code had magical hardcoded knowledge of the actual values of VERIFY_{READ,WRITE} (not that they mattered, since nothing really used it) - microblaze used the type argument for a debug printout but other than those oddities this should be a total no-op patch. I tried to fix up all architectures, did fairly extensive grepping for access_ok() uses, and the changes are trivial, but I may have missed something. Any missed conversion should be trivially fixable, though. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 09:57:57 +07:00
if (!access_ok(buf, count))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
return -EFAULT;
if (kvm_is_radix(kvm))
return 0;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
first_pass = ctx->first_pass;
flags = ctx->flags;
i = ctx->index;
hptp = (__be64 *)(kvm->arch.hpt.virt + (i * HPTE_SIZE));
revp = kvm->arch.hpt.rev + i;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
lbuf = (unsigned long __user *)buf;
nb = 0;
while (nb + sizeof(hdr) + HPTE_SIZE < count) {
/* Initialize header */
hptr = (struct kvm_get_htab_header __user *)buf;
hdr.n_valid = 0;
hdr.n_invalid = 0;
nw = nb;
nb += sizeof(hdr);
lbuf = (unsigned long __user *)(buf + sizeof(hdr));
/* Skip uninteresting entries, i.e. clean on not-first pass */
if (!first_pass) {
while (i < kvmppc_hpt_npte(&kvm->arch.hpt) &&
!hpte_dirty(revp, hptp)) {
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
++i;
hptp += 2;
++revp;
}
}
hdr.index = i;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
/* Grab a series of valid entries */
while (i < kvmppc_hpt_npte(&kvm->arch.hpt) &&
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
hdr.n_valid < 0xffff &&
nb + HPTE_SIZE < count &&
record_hpte(flags, hptp, hpte, revp, 1, first_pass)) {
/* valid entry, write it out */
++hdr.n_valid;
if (__put_user(hpte[0], lbuf) ||
__put_user(hpte[1], lbuf + 1))
return -EFAULT;
nb += HPTE_SIZE;
lbuf += 2;
++i;
hptp += 2;
++revp;
}
/* Now skip invalid entries while we can */
while (i < kvmppc_hpt_npte(&kvm->arch.hpt) &&
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
hdr.n_invalid < 0xffff &&
record_hpte(flags, hptp, hpte, revp, 0, first_pass)) {
/* found an invalid entry */
++hdr.n_invalid;
++i;
hptp += 2;
++revp;
}
if (hdr.n_valid || hdr.n_invalid) {
/* write back the header */
if (__copy_to_user(hptr, &hdr, sizeof(hdr)))
return -EFAULT;
nw = nb;
buf = (char __user *)lbuf;
} else {
nb = nw;
}
/* Check if we've wrapped around the hash table */
if (i >= kvmppc_hpt_npte(&kvm->arch.hpt)) {
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
i = 0;
ctx->first_pass = 0;
break;
}
}
ctx->index = i;
return nb;
}
static ssize_t kvm_htab_write(struct file *file, const char __user *buf,
size_t count, loff_t *ppos)
{
struct kvm_htab_ctx *ctx = file->private_data;
struct kvm *kvm = ctx->kvm;
struct kvm_get_htab_header hdr;
unsigned long i, j;
unsigned long v, r;
unsigned long __user *lbuf;
__be64 *hptp;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
unsigned long tmp[2];
ssize_t nb;
long int err, ret;
int mmu_ready;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
int pshift;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
Remove 'type' argument from access_ok() function Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument of the user address range verification function since we got rid of the old racy i386-only code to walk page tables by hand. It existed because the original 80386 would not honor the write protect bit when in kernel mode, so you had to do COW by hand before doing any user access. But we haven't supported that in a long time, and these days the 'type' argument is a purely historical artifact. A discussion about extending 'user_access_begin()' to do the range checking resulted this patch, because there is no way we're going to move the old VERIFY_xyz interface to that model. And it's best done at the end of the merge window when I've done most of my merges, so let's just get this done once and for all. This patch was mostly done with a sed-script, with manual fix-ups for the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form. There were a couple of notable cases: - csky still had the old "verify_area()" name as an alias. - the iter_iov code had magical hardcoded knowledge of the actual values of VERIFY_{READ,WRITE} (not that they mattered, since nothing really used it) - microblaze used the type argument for a debug printout but other than those oddities this should be a total no-op patch. I tried to fix up all architectures, did fairly extensive grepping for access_ok() uses, and the changes are trivial, but I may have missed something. Any missed conversion should be trivially fixable, though. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 09:57:57 +07:00
if (!access_ok(buf, count))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
return -EFAULT;
if (kvm_is_radix(kvm))
return -EINVAL;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
/* lock out vcpus from running while we're doing this */
mutex_lock(&kvm->arch.mmu_setup_lock);
mmu_ready = kvm->arch.mmu_ready;
if (mmu_ready) {
kvm->arch.mmu_ready = 0; /* temporarily */
/* order mmu_ready vs. vcpus_running */
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
smp_mb();
if (atomic_read(&kvm->arch.vcpus_running)) {
kvm->arch.mmu_ready = 1;
mutex_unlock(&kvm->arch.mmu_setup_lock);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
return -EBUSY;
}
}
err = 0;
for (nb = 0; nb + sizeof(hdr) <= count; ) {
err = -EFAULT;
if (__copy_from_user(&hdr, buf, sizeof(hdr)))
break;
err = 0;
if (nb + hdr.n_valid * HPTE_SIZE > count)
break;
nb += sizeof(hdr);
buf += sizeof(hdr);
err = -EINVAL;
i = hdr.index;
if (i >= kvmppc_hpt_npte(&kvm->arch.hpt) ||
i + hdr.n_valid + hdr.n_invalid > kvmppc_hpt_npte(&kvm->arch.hpt))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
break;
hptp = (__be64 *)(kvm->arch.hpt.virt + (i * HPTE_SIZE));
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
lbuf = (unsigned long __user *)buf;
for (j = 0; j < hdr.n_valid; ++j) {
__be64 hpte_v;
__be64 hpte_r;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
err = -EFAULT;
if (__get_user(hpte_v, lbuf) ||
__get_user(hpte_r, lbuf + 1))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
goto out;
v = be64_to_cpu(hpte_v);
r = be64_to_cpu(hpte_r);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
err = -EINVAL;
if (!(v & HPTE_V_VALID))
goto out;
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
pshift = kvmppc_hpte_base_page_shift(v, r);
if (pshift <= 0)
goto out;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
lbuf += 2;
nb += HPTE_SIZE;
if (be64_to_cpu(hptp[0]) & (HPTE_V_VALID | HPTE_V_ABSENT))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
kvmppc_do_h_remove(kvm, 0, i, 0, tmp);
err = -EIO;
ret = kvmppc_virtmode_do_h_enter(kvm, H_EXACT, i, v, r,
tmp);
if (ret != H_SUCCESS) {
pr_err("kvm_htab_write ret %ld i=%ld v=%lx "
"r=%lx\n", ret, i, v, r);
goto out;
}
if (!mmu_ready && is_vrma_hpte(v)) {
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
unsigned long senc, lpcr;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
senc = slb_pgsize_encoding(1ul << pshift);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
kvm->arch.vrma_slb_v = senc | SLB_VSID_B_1T |
(VRMA_VSID << SLB_VSID_SHIFT_1T);
KVM: PPC: Book3S HV: Fix migration and HPT resizing of HPT guests on radix hosts This fixes two errors that prevent a guest using the HPT MMU from successfully migrating to a POWER9 host in radix MMU mode, or resizing its HPT when running on a radix host. The first bug was that commit 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information", 2017-09-11) missed two uses of hpte_base_page_size(), one in the HPT rehashing code and one in kvm_htab_write() (which is used on the destination side in migrating a HPT guest). Instead we use kvmppc_hpte_base_page_shift(). Having the shift count means that we can use left and right shifts instead of multiplication and division in a few places. Along the way, this adds a check in kvm_htab_write() to ensure that the page size encoding in the incoming HPTEs is recognized, and if not return an EINVAL error to userspace. The second bug was that kvm_htab_write was performing some but not all of the functions of kvmhv_setup_mmu(), resulting in the destination VM being left in radix mode as far as the hardware is concerned. The simplest fix for now is make kvm_htab_write() call kvmppc_setup_partition_table() like kvmppc_hv_setup_htab_rma() does. In future it would be better to refactor the code more extensively to remove the duplication. Fixes: 8dc6cca556e4 ("KVM: PPC: Book3S HV: Don't rely on host's page size information") Fixes: 7a84084c6054 ("KVM: PPC: Book3S HV: Set partition table rather than SDR1 on POWER9") Reported-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Tested-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
2017-11-22 10:38:53 +07:00
if (!cpu_has_feature(CPU_FTR_ARCH_300)) {
lpcr = senc << (LPCR_VRMASD_SH - 4);
kvmppc_update_lpcr(kvm, lpcr,
LPCR_VRMASD);
} else {
kvmppc_setup_partition_table(kvm);
}
mmu_ready = 1;
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
}
++i;
hptp += 2;
}
for (j = 0; j < hdr.n_invalid; ++j) {
if (be64_to_cpu(hptp[0]) & (HPTE_V_VALID | HPTE_V_ABSENT))
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
kvmppc_do_h_remove(kvm, 0, i, 0, tmp);
++i;
hptp += 2;
}
err = 0;
}
out:
/* Order HPTE updates vs. mmu_ready */
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
smp_wmb();
kvm->arch.mmu_ready = mmu_ready;
mutex_unlock(&kvm->arch.mmu_setup_lock);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
if (err)
return err;
return nb;
}
static int kvm_htab_release(struct inode *inode, struct file *filp)
{
struct kvm_htab_ctx *ctx = filp->private_data;
filp->private_data = NULL;
if (!(ctx->flags & KVM_GET_HTAB_WRITE))
atomic_dec(&ctx->kvm->arch.hpte_mod_interest);
kvm_put_kvm(ctx->kvm);
kfree(ctx);
return 0;
}
static const struct file_operations kvm_htab_fops = {
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
.read = kvm_htab_read,
.write = kvm_htab_write,
.llseek = default_llseek,
.release = kvm_htab_release,
};
int kvm_vm_ioctl_get_htab_fd(struct kvm *kvm, struct kvm_get_htab_fd *ghf)
{
int ret;
struct kvm_htab_ctx *ctx;
int rwflag;
/* reject flags we don't recognize */
if (ghf->flags & ~(KVM_GET_HTAB_BOLTED_ONLY | KVM_GET_HTAB_WRITE))
return -EINVAL;
ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
if (!ctx)
return -ENOMEM;
kvm_get_kvm(kvm);
ctx->kvm = kvm;
ctx->index = ghf->start_index;
ctx->flags = ghf->flags;
ctx->first_pass = 1;
rwflag = (ghf->flags & KVM_GET_HTAB_WRITE) ? O_WRONLY : O_RDONLY;
ret = anon_inode_getfd("kvm-htab", &kvm_htab_fops, ctx, rwflag | O_CLOEXEC);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
if (ret < 0) {
kfree(ctx);
kvm_put_kvm_no_destroy(kvm);
KVM: PPC: Book3S HV: Provide a method for userspace to read and write the HPT A new ioctl, KVM_PPC_GET_HTAB_FD, returns a file descriptor. Reads on this fd return the contents of the HPT (hashed page table), writes create and/or remove entries in the HPT. There is a new capability, KVM_CAP_PPC_HTAB_FD, to indicate the presence of the ioctl. The ioctl takes an argument structure with the index of the first HPT entry to read out and a set of flags. The flags indicate whether the user is intending to read or write the HPT, and whether to return all entries or only the "bolted" entries (those with the bolted bit, 0x10, set in the first doubleword). This is intended for use in implementing qemu's savevm/loadvm and for live migration. Therefore, on reads, the first pass returns information about all HPTEs (or all bolted HPTEs). When the first pass reaches the end of the HPT, it returns from the read. Subsequent reads only return information about HPTEs that have changed since they were last read. A read that finds no changed HPTEs in the HPT following where the last read finished will return 0 bytes. The format of the data provides a simple run-length compression of the invalid entries. Each block of data starts with a header that indicates the index (position in the HPT, which is just an array), the number of valid entries starting at that index (may be zero), and the number of invalid entries following those valid entries. The valid entries, 16 bytes each, follow the header. The invalid entries are not explicitly represented. Signed-off-by: Paul Mackerras <paulus@samba.org> [agraf: fix documentation] Signed-off-by: Alexander Graf <agraf@suse.de>
2012-11-20 05:57:20 +07:00
return ret;
}
if (rwflag == O_RDONLY) {
mutex_lock(&kvm->slots_lock);
atomic_inc(&kvm->arch.hpte_mod_interest);
/* make sure kvmppc_do_h_enter etc. see the increment */
synchronize_srcu_expedited(&kvm->srcu);
mutex_unlock(&kvm->slots_lock);
}
return ret;
}
struct debugfs_htab_state {
struct kvm *kvm;
struct mutex mutex;
unsigned long hpt_index;
int chars_left;
int buf_index;
char buf[64];
};
static int debugfs_htab_open(struct inode *inode, struct file *file)
{
struct kvm *kvm = inode->i_private;
struct debugfs_htab_state *p;
p = kzalloc(sizeof(*p), GFP_KERNEL);
if (!p)
return -ENOMEM;
kvm_get_kvm(kvm);
p->kvm = kvm;
mutex_init(&p->mutex);
file->private_data = p;
return nonseekable_open(inode, file);
}
static int debugfs_htab_release(struct inode *inode, struct file *file)
{
struct debugfs_htab_state *p = file->private_data;
kvm_put_kvm(p->kvm);
kfree(p);
return 0;
}
static ssize_t debugfs_htab_read(struct file *file, char __user *buf,
size_t len, loff_t *ppos)
{
struct debugfs_htab_state *p = file->private_data;
ssize_t ret, r;
unsigned long i, n;
unsigned long v, hr, gr;
struct kvm *kvm;
__be64 *hptp;
kvm = p->kvm;
if (kvm_is_radix(kvm))
return 0;
ret = mutex_lock_interruptible(&p->mutex);
if (ret)
return ret;
if (p->chars_left) {
n = p->chars_left;
if (n > len)
n = len;
r = copy_to_user(buf, p->buf + p->buf_index, n);
n -= r;
p->chars_left -= n;
p->buf_index += n;
buf += n;
len -= n;
ret = n;
if (r) {
if (!n)
ret = -EFAULT;
goto out;
}
}
i = p->hpt_index;
hptp = (__be64 *)(kvm->arch.hpt.virt + (i * HPTE_SIZE));
for (; len != 0 && i < kvmppc_hpt_npte(&kvm->arch.hpt);
++i, hptp += 2) {
if (!(be64_to_cpu(hptp[0]) & (HPTE_V_VALID | HPTE_V_ABSENT)))
continue;
/* lock the HPTE so it's stable and read it */
preempt_disable();
while (!try_lock_hpte(hptp, HPTE_V_HVLOCK))
cpu_relax();
v = be64_to_cpu(hptp[0]) & ~HPTE_V_HVLOCK;
hr = be64_to_cpu(hptp[1]);
gr = kvm->arch.hpt.rev[i].guest_rpte;
unlock_hpte(hptp, v);
preempt_enable();
if (!(v & (HPTE_V_VALID | HPTE_V_ABSENT)))
continue;
n = scnprintf(p->buf, sizeof(p->buf),
"%6lx %.16lx %.16lx %.16lx\n",
i, v, hr, gr);
p->chars_left = n;
if (n > len)
n = len;
r = copy_to_user(buf, p->buf, n);
n -= r;
p->chars_left -= n;
p->buf_index = n;
buf += n;
len -= n;
ret += n;
if (r) {
if (!ret)
ret = -EFAULT;
goto out;
}
}
p->hpt_index = i;
out:
mutex_unlock(&p->mutex);
return ret;
}
static ssize_t debugfs_htab_write(struct file *file, const char __user *buf,
size_t len, loff_t *ppos)
{
return -EACCES;
}
static const struct file_operations debugfs_htab_fops = {
.owner = THIS_MODULE,
.open = debugfs_htab_open,
.release = debugfs_htab_release,
.read = debugfs_htab_read,
.write = debugfs_htab_write,
.llseek = generic_file_llseek,
};
void kvmppc_mmu_debugfs_init(struct kvm *kvm)
{
kvm->arch.htab_dentry = debugfs_create_file("htab", 0400,
kvm->arch.debugfs_dir, kvm,
&debugfs_htab_fops);
}
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
void kvmppc_mmu_book3s_hv_init(struct kvm_vcpu *vcpu)
{
struct kvmppc_mmu *mmu = &vcpu->arch.mmu;
vcpu->arch.slb_nr = 32; /* POWER7/POWER8 */
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
mmu->xlate = kvmppc_mmu_book3s_64_hv_xlate;
KVM: PPC: Add support for Book3S processors in hypervisor mode This adds support for KVM running on 64-bit Book 3S processors, specifically POWER7, in hypervisor mode. Using hypervisor mode means that the guest can use the processor's supervisor mode. That means that the guest can execute privileged instructions and access privileged registers itself without trapping to the host. This gives excellent performance, but does mean that KVM cannot emulate a processor architecture other than the one that the hardware implements. This code assumes that the guest is running paravirtualized using the PAPR (Power Architecture Platform Requirements) interface, which is the interface that IBM's PowerVM hypervisor uses. That means that existing Linux distributions that run on IBM pSeries machines will also run under KVM without modification. In order to communicate the PAPR hypercalls to qemu, this adds a new KVM_EXIT_PAPR_HCALL exit code to include/linux/kvm.h. Currently the choice between book3s_hv support and book3s_pr support (i.e. the existing code, which runs the guest in user mode) has to be made at kernel configuration time, so a given kernel binary can only do one or the other. This new book3s_hv code doesn't support MMIO emulation at present. Since we are running paravirtualized guests, this isn't a serious restriction. With the guest running in supervisor mode, most exceptions go straight to the guest. We will never get data or instruction storage or segment interrupts, alignment interrupts, decrementer interrupts, program interrupts, single-step interrupts, etc., coming to the hypervisor from the guest. Therefore this introduces a new KVMTEST_NONHV macro for the exception entry path so that we don't have to do the KVM test on entry to those exception handlers. We do however get hypervisor decrementer, hypervisor data storage, hypervisor instruction storage, and hypervisor emulation assist interrupts, so we have to handle those. In hypervisor mode, real-mode accesses can access all of RAM, not just a limited amount. Therefore we put all the guest state in the vcpu.arch and use the shadow_vcpu in the PACA only for temporary scratch space. We allocate the vcpu with kzalloc rather than vzalloc, and we don't use anything in the kvmppc_vcpu_book3s struct, so we don't allocate it. We don't have a shared page with the guest, but we still need a kvm_vcpu_arch_shared struct to store the values of various registers, so we include one in the vcpu_arch struct. The POWER7 processor has a restriction that all threads in a core have to be in the same partition. MMU-on kernel code counts as a partition (partition 0), so we have to do a partition switch on every entry to and exit from the guest. At present we require the host and guest to run in single-thread mode because of this hardware restriction. This code allocates a hashed page table for the guest and initializes it with HPTEs for the guest's Virtual Real Memory Area (VRMA). We require that the guest memory is allocated using 16MB huge pages, in order to simplify the low-level memory management. This also means that we can get away without tracking paging activity in the host for now, since huge pages can't be paged or swapped. This also adds a few new exports needed by the book3s_hv code. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Alexander Graf <agraf@suse.de>
2011-06-29 07:21:34 +07:00
vcpu->arch.hflags |= BOOK3S_HFLAG_SLB;
}