AuxXxilium/linux_dsm_epyc7002
1
0
Fork 0
mirror of https://github.com/AuxXxilium/linux_dsm_epyc7002.git synced 2024-12-24 10:37:45 +07:00
Code Issues Actions Packages Projects Releases Wiki Activity
32902177f7
linux_dsm_epyc7002/arch/powerpc/kvm/e500_mmu_host.c

814 lines
21 KiB
C
Raw Normal View History

KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
/*
* Copyright (C) 2008-2013 Freescale Semiconductor, Inc. All rights reserved.
*
* Author: Yu Liu, yu.liu@freescale.com
* Scott Wood, scottwood@freescale.com
* Ashish Kalra, ashish.kalra@freescale.com
* Varun Sethi, varun.sethi@freescale.com
* Alexander Graf, agraf@suse.de
*
* Description:
* This file is based on arch/powerpc/kvm/44x_tlb.c,
* by Hollis Blanchard <hollisb@us.ibm.com>.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License, version 2, as
* published by the Free Software Foundation.
*/
#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/string.h>
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/highmem.h>
#include <linux/log2.h>
#include <linux/uaccess.h>
#include <linux/sched.h>
#include <linux/rwsem.h>
#include <linux/vmalloc.h>
#include <linux/hugetlb.h>
#include <asm/kvm_ppc.h>
#include "e500.h"
#include "timing.h"
#include "e500_mmu_host.h"
kvm: powerpc: booke: Move booke related tracepoints to separate header Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-10-07 23:47:58 +07:00
#include "trace_booke.h"
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
#define to_htlb1_esel(esel) (host_tlb_params[1].entries - (esel) - 1)
static struct kvmppc_e500_tlb_params host_tlb_params[E500_TLB_NUM];
static inline unsigned int tlb1_max_shadow_size(void)
{
/* reserve one entry for magic page */
return host_tlb_params[1].entries - tlbcam_index - 1;
}
static inline u32 e500_shadow_mas3_attrib(u32 mas3, int usermode)
{
/* Mask off reserved bits. */
mas3 &= MAS3_ATTRIB_MASK;
#ifndef CONFIG_KVM_BOOKE_HV
if (!usermode) {
/* Guest is in supervisor mode,
* so we need to translate guest
* supervisor permissions into user permissions. */
mas3 &= ~E500_TLB_USER_PERM_MASK;
mas3 |= (mas3 & E500_TLB_SUPER_PERM_MASK) << 1;
}
mas3 |= E500_TLB_SUPER_PERM_MASK;
#endif
return mas3;
}
/*
* writing shadow tlb entry to host TLB
*/
static inline void __write_host_tlbe(struct kvm_book3e_206_tlb_entry *stlbe,
KVM: PPC: e500mc: Add support for single threaded vcpus on e6500 core ePAPR represents hardware threads as cpu node properties in device tree. So with existing QEMU, hardware threads are simply exposed as vcpus with one hardware thread. The e6500 core shares TLBs between hardware threads. Without tlb write conditional instruction, the Linux kernel uses per core mechanisms to protect against duplicate TLB entries. The guest is unable to detect real siblings threads, so it can't use the TLB protection mechanism. An alternative solution is to use the hypervisor to allocate different lpids to guest's vcpus that runs simultaneous on real siblings threads. On systems with two threads per core this patch halves the size of the lpid pool that the allocator sees and use two lpids per VM. Use even numbers to speedup vcpu lpid computation with consecutive lpids per VM: vm1 will use lpids 2 and 3, vm2 lpids 4 and 5, and so on. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> [agraf: fix spelling] Signed-off-by: Alexander Graf <agraf@suse.de>
2014-09-01 16:01:58 +07:00
uint32_t mas0,
uint32_t lpid)
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
{
unsigned long flags;
local_irq_save(flags);
mtspr(SPRN_MAS0, mas0);
mtspr(SPRN_MAS1, stlbe->mas1);
mtspr(SPRN_MAS2, (unsigned long)stlbe->mas2);
mtspr(SPRN_MAS3, (u32)stlbe->mas7_3);
mtspr(SPRN_MAS7, (u32)(stlbe->mas7_3 >> 32));
#ifdef CONFIG_KVM_BOOKE_HV
KVM: PPC: e500mc: Add support for single threaded vcpus on e6500 core ePAPR represents hardware threads as cpu node properties in device tree. So with existing QEMU, hardware threads are simply exposed as vcpus with one hardware thread. The e6500 core shares TLBs between hardware threads. Without tlb write conditional instruction, the Linux kernel uses per core mechanisms to protect against duplicate TLB entries. The guest is unable to detect real siblings threads, so it can't use the TLB protection mechanism. An alternative solution is to use the hypervisor to allocate different lpids to guest's vcpus that runs simultaneous on real siblings threads. On systems with two threads per core this patch halves the size of the lpid pool that the allocator sees and use two lpids per VM. Use even numbers to speedup vcpu lpid computation with consecutive lpids per VM: vm1 will use lpids 2 and 3, vm2 lpids 4 and 5, and so on. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> [agraf: fix spelling] Signed-off-by: Alexander Graf <agraf@suse.de>
2014-09-01 16:01:58 +07:00
mtspr(SPRN_MAS8, MAS8_TGS | get_thread_specific_lpid(lpid));
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
#endif
asm volatile("isync; tlbwe" : : : "memory");
#ifdef CONFIG_KVM_BOOKE_HV
/* Must clear mas8 for other host tlbwe's */
mtspr(SPRN_MAS8, 0);
isync();
#endif
local_irq_restore(flags);
trace_kvm_booke206_stlb_write(mas0, stlbe->mas8, stlbe->mas1,
stlbe->mas2, stlbe->mas7_3);
}
/*
* Acquire a mas0 with victim hint, as if we just took a TLB miss.
*
* We don't care about the address we're searching for, other than that it's
* in the right set and is not present in the TLB. Using a zero PID and a
* userspace address means we don't have to set and then restore MAS5, or
* calculate a proper MAS6 value.
*/
static u32 get_host_mas0(unsigned long eaddr)
{
unsigned long flags;
u32 mas0;
KVM: PPC: e500: Fix default tlb for victim hint Tlb search operation used for victim hint relies on the default tlb set by the host. When hardware tablewalk support is enabled in the host, the default tlb is TLB1 which leads KVM to evict the bolted entry. Set and restore the default tlb when searching for victim hint. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-06-30 19:54:58 +07:00
u32 mas4;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
local_irq_save(flags);
mtspr(SPRN_MAS6, 0);
KVM: PPC: e500: Fix default tlb for victim hint Tlb search operation used for victim hint relies on the default tlb set by the host. When hardware tablewalk support is enabled in the host, the default tlb is TLB1 which leads KVM to evict the bolted entry. Set and restore the default tlb when searching for victim hint. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-06-30 19:54:58 +07:00
mas4 = mfspr(SPRN_MAS4);
mtspr(SPRN_MAS4, mas4 & ~MAS4_TLBSEL_MASK);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
asm volatile("tlbsx 0, %0" : : "b" (eaddr & ~CONFIG_PAGE_OFFSET));
mas0 = mfspr(SPRN_MAS0);
KVM: PPC: e500: Fix default tlb for victim hint Tlb search operation used for victim hint relies on the default tlb set by the host. When hardware tablewalk support is enabled in the host, the default tlb is TLB1 which leads KVM to evict the bolted entry. Set and restore the default tlb when searching for victim hint. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-06-30 19:54:58 +07:00
mtspr(SPRN_MAS4, mas4);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
local_irq_restore(flags);
return mas0;
}
/* sesel is for tlb1 only */
static inline void write_host_tlbe(struct kvmppc_vcpu_e500 *vcpu_e500,
int tlbsel, int sesel, struct kvm_book3e_206_tlb_entry *stlbe)
{
u32 mas0;
if (tlbsel == 0) {
mas0 = get_host_mas0(stlbe->mas2);
KVM: PPC: e500mc: Add support for single threaded vcpus on e6500 core ePAPR represents hardware threads as cpu node properties in device tree. So with existing QEMU, hardware threads are simply exposed as vcpus with one hardware thread. The e6500 core shares TLBs between hardware threads. Without tlb write conditional instruction, the Linux kernel uses per core mechanisms to protect against duplicate TLB entries. The guest is unable to detect real siblings threads, so it can't use the TLB protection mechanism. An alternative solution is to use the hypervisor to allocate different lpids to guest's vcpus that runs simultaneous on real siblings threads. On systems with two threads per core this patch halves the size of the lpid pool that the allocator sees and use two lpids per VM. Use even numbers to speedup vcpu lpid computation with consecutive lpids per VM: vm1 will use lpids 2 and 3, vm2 lpids 4 and 5, and so on. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> [agraf: fix spelling] Signed-off-by: Alexander Graf <agraf@suse.de>
2014-09-01 16:01:58 +07:00
__write_host_tlbe(stlbe, mas0, vcpu_e500->vcpu.kvm->arch.lpid);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
} else {
__write_host_tlbe(stlbe,
MAS0_TLBSEL(1) |
KVM: PPC: e500mc: Add support for single threaded vcpus on e6500 core ePAPR represents hardware threads as cpu node properties in device tree. So with existing QEMU, hardware threads are simply exposed as vcpus with one hardware thread. The e6500 core shares TLBs between hardware threads. Without tlb write conditional instruction, the Linux kernel uses per core mechanisms to protect against duplicate TLB entries. The guest is unable to detect real siblings threads, so it can't use the TLB protection mechanism. An alternative solution is to use the hypervisor to allocate different lpids to guest's vcpus that runs simultaneous on real siblings threads. On systems with two threads per core this patch halves the size of the lpid pool that the allocator sees and use two lpids per VM. Use even numbers to speedup vcpu lpid computation with consecutive lpids per VM: vm1 will use lpids 2 and 3, vm2 lpids 4 and 5, and so on. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> [agraf: fix spelling] Signed-off-by: Alexander Graf <agraf@suse.de>
2014-09-01 16:01:58 +07:00
MAS0_ESEL(to_htlb1_esel(sesel)),
vcpu_e500->vcpu.kvm->arch.lpid);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
}
}
/* sesel is for tlb1 only */
static void write_stlbe(struct kvmppc_vcpu_e500 *vcpu_e500,
struct kvm_book3e_206_tlb_entry *gtlbe,
struct kvm_book3e_206_tlb_entry *stlbe,
int stlbsel, int sesel)
{
int stid;
preempt_disable();
stid = kvmppc_e500_get_tlb_stid(&vcpu_e500->vcpu, gtlbe);
stlbe->mas1 |= MAS1_TID(stid);
write_host_tlbe(vcpu_e500, stlbsel, sesel, stlbe);
preempt_enable();
}
#ifdef CONFIG_KVM_E500V2
/* XXX should be a hook in the gva2hpa translation */
void kvmppc_map_magic(struct kvm_vcpu *vcpu)
{
struct kvmppc_vcpu_e500 *vcpu_e500 = to_e500(vcpu);
struct kvm_book3e_206_tlb_entry magic;
ulong shared_page = ((ulong)vcpu->arch.shared) & PAGE_MASK;
unsigned int stid;
kvm: rename pfn_t to kvm_pfn_t To date, we have implemented two I/O usage models for persistent memory, PMEM (a persistent "ram disk") and DAX (mmap persistent memory into userspace). This series adds a third, DAX-GUP, that allows DAX mappings to be the target of direct-i/o. It allows userspace to coordinate DMA/RDMA from/to persistent memory. The implementation leverages the ZONE_DEVICE mm-zone that went into 4.3-rc1 (also discussed at kernel summit) to flag pages that are owned and dynamically mapped by a device driver. The pmem driver, after mapping a persistent memory range into the system memmap via devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus page-backed pmem-pfns via flags in the new pfn_t type. The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the resulting pte(s) inserted into the process page tables with a new _PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys off _PAGE_DEVMAP to pin the device hosting the page range active. Finally, get_page() and put_page() are modified to take references against the device driver established page mapping. Finally, this need for "struct page" for persistent memory requires memory capacity to store the memmap array. Given the memmap array for a large pool of persistent may exhaust available DRAM introduce a mechanism to allocate the memmap from persistent memory. The new "struct vmem_altmap *" parameter to devm_memremap_pages() enables arch_add_memory() to use reserved pmem capacity rather than the page allocator. This patch (of 18): The core has developed a need for a "pfn_t" type [1]. Move the existing pfn_t in KVM to kvm_pfn_t [2]. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html [2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.html Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: Christoffer Dall <christoffer.dall@linaro.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 07:56:11 +07:00
kvm_pfn_t pfn;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
kvm: rename pfn_t to kvm_pfn_t To date, we have implemented two I/O usage models for persistent memory, PMEM (a persistent "ram disk") and DAX (mmap persistent memory into userspace). This series adds a third, DAX-GUP, that allows DAX mappings to be the target of direct-i/o. It allows userspace to coordinate DMA/RDMA from/to persistent memory. The implementation leverages the ZONE_DEVICE mm-zone that went into 4.3-rc1 (also discussed at kernel summit) to flag pages that are owned and dynamically mapped by a device driver. The pmem driver, after mapping a persistent memory range into the system memmap via devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus page-backed pmem-pfns via flags in the new pfn_t type. The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the resulting pte(s) inserted into the process page tables with a new _PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys off _PAGE_DEVMAP to pin the device hosting the page range active. Finally, get_page() and put_page() are modified to take references against the device driver established page mapping. Finally, this need for "struct page" for persistent memory requires memory capacity to store the memmap array. Given the memmap array for a large pool of persistent may exhaust available DRAM introduce a mechanism to allocate the memmap from persistent memory. The new "struct vmem_altmap *" parameter to devm_memremap_pages() enables arch_add_memory() to use reserved pmem capacity rather than the page allocator. This patch (of 18): The core has developed a need for a "pfn_t" type [1]. Move the existing pfn_t in KVM to kvm_pfn_t [2]. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html [2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.html Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: Christoffer Dall <christoffer.dall@linaro.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 07:56:11 +07:00
pfn = (kvm_pfn_t)virt_to_phys((void *)shared_page) >> PAGE_SHIFT;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
get_page(pfn_to_page(pfn));
preempt_disable();
stid = kvmppc_e500_get_sid(vcpu_e500, 0, 0, 0, 0);
magic.mas1 = MAS1_VALID | MAS1_TS | MAS1_TID(stid) |
MAS1_TSIZE(BOOK3E_PAGESZ_4K);
magic.mas2 = vcpu->arch.magic_page_ea | MAS2_M;
magic.mas7_3 = ((u64)pfn << PAGE_SHIFT) |
MAS3_SW | MAS3_SR | MAS3_UW | MAS3_UR;
magic.mas8 = 0;
KVM: PPC: e500mc: Add support for single threaded vcpus on e6500 core ePAPR represents hardware threads as cpu node properties in device tree. So with existing QEMU, hardware threads are simply exposed as vcpus with one hardware thread. The e6500 core shares TLBs between hardware threads. Without tlb write conditional instruction, the Linux kernel uses per core mechanisms to protect against duplicate TLB entries. The guest is unable to detect real siblings threads, so it can't use the TLB protection mechanism. An alternative solution is to use the hypervisor to allocate different lpids to guest's vcpus that runs simultaneous on real siblings threads. On systems with two threads per core this patch halves the size of the lpid pool that the allocator sees and use two lpids per VM. Use even numbers to speedup vcpu lpid computation with consecutive lpids per VM: vm1 will use lpids 2 and 3, vm2 lpids 4 and 5, and so on. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> [agraf: fix spelling] Signed-off-by: Alexander Graf <agraf@suse.de>
2014-09-01 16:01:58 +07:00
__write_host_tlbe(&magic, MAS0_TLBSEL(1) | MAS0_ESEL(tlbcam_index), 0);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
preempt_enable();
}
#endif
void inval_gtlbe_on_host(struct kvmppc_vcpu_e500 *vcpu_e500, int tlbsel,
int esel)
{
struct kvm_book3e_206_tlb_entry *gtlbe =
get_entry(vcpu_e500, tlbsel, esel);
struct tlbe_ref *ref = &vcpu_e500->gtlb_priv[tlbsel][esel].ref;
/* Don't bother with unmapped entries */
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
if (!(ref->flags & E500_TLB_VALID)) {
WARN(ref->flags & (E500_TLB_BITMAP | E500_TLB_TLB0),
"%s: flags %x\n", __func__, ref->flags);
WARN_ON(tlbsel == 1 && vcpu_e500->g2h_tlb1_map[esel]);
}
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
if (tlbsel == 1 && ref->flags & E500_TLB_BITMAP) {
u64 tmp = vcpu_e500->g2h_tlb1_map[esel];
int hw_tlb_indx;
unsigned long flags;
local_irq_save(flags);
while (tmp) {
hw_tlb_indx = __ilog2_u64(tmp & -tmp);
mtspr(SPRN_MAS0,
MAS0_TLBSEL(1) |
MAS0_ESEL(to_htlb1_esel(hw_tlb_indx)));
mtspr(SPRN_MAS1, 0);
asm volatile("tlbwe");
vcpu_e500->h2g_tlb1_rmap[hw_tlb_indx] = 0;
tmp &= tmp - 1;
}
mb();
vcpu_e500->g2h_tlb1_map[esel] = 0;
ref->flags &= ~(E500_TLB_BITMAP | E500_TLB_VALID);
local_irq_restore(flags);
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
}
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
if (tlbsel == 1 && ref->flags & E500_TLB_TLB0) {
/*
* TLB1 entry is backed by 4k pages. This should happen
* rarely and is not worth optimizing. Invalidate everything.
*/
kvmppc_e500_tlbil_all(vcpu_e500);
ref->flags &= ~(E500_TLB_TLB0 | E500_TLB_VALID);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
}
kvm: booke: clear host tlb reference flag on guest tlb invalidation On booke, "struct tlbe_ref" contains host tlb mapping information (pfn: for guest-pfn to pfn, flags: attribute associated with this mapping) for a guest tlb entry. So when a guest creates a TLB entry then "struct tlbe_ref" is set to point to valid "pfn" and set attributes in "flags" field of the above said structure. When a guest TLB entry is invalidated then flags field of corresponding "struct tlbe_ref" is updated to point that this is no more valid, also we selectively clear some other attribute bits, example: if E500_TLB_BITMAP was set then we clear E500_TLB_BITMAP, if E500_TLB_TLB0 is set then we clear this. Ideally we should clear complete "flags" as this entry is invalid and does not have anything to re-used. The other part of the problem is that when we use the same entry again then also we do not clear (started doing or-ing etc). So far it was working because the selectively clearing mentioned above actually clears "flags" what was set during TLB mapping. But the problem starts coming when we add more attributes to this then we need to selectively clear them and which is not needed. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-15 12:31:13 +07:00
/*
* If TLB entry is still valid then it's a TLB0 entry, and thus
* backed by at most one host tlbe per shadow pid
*/
if (ref->flags & E500_TLB_VALID)
kvmppc_e500_tlbil_one(vcpu_e500, gtlbe);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
/* Mark the TLB as not backed by the host anymore */
kvm: booke: clear host tlb reference flag on guest tlb invalidation On booke, "struct tlbe_ref" contains host tlb mapping information (pfn: for guest-pfn to pfn, flags: attribute associated with this mapping) for a guest tlb entry. So when a guest creates a TLB entry then "struct tlbe_ref" is set to point to valid "pfn" and set attributes in "flags" field of the above said structure. When a guest TLB entry is invalidated then flags field of corresponding "struct tlbe_ref" is updated to point that this is no more valid, also we selectively clear some other attribute bits, example: if E500_TLB_BITMAP was set then we clear E500_TLB_BITMAP, if E500_TLB_TLB0 is set then we clear this. Ideally we should clear complete "flags" as this entry is invalid and does not have anything to re-used. The other part of the problem is that when we use the same entry again then also we do not clear (started doing or-ing etc). So far it was working because the selectively clearing mentioned above actually clears "flags" what was set during TLB mapping. But the problem starts coming when we add more attributes to this then we need to selectively clear them and which is not needed. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-15 12:31:13 +07:00
ref->flags = 0;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
}
static inline int tlbe_is_writable(struct kvm_book3e_206_tlb_entry *tlbe)
{
return tlbe->mas7_3 & (MAS3_SW|MAS3_UW);
}
static inline void kvmppc_e500_ref_setup(struct tlbe_ref *ref,
struct kvm_book3e_206_tlb_entry *gtlbe,
kvm: rename pfn_t to kvm_pfn_t To date, we have implemented two I/O usage models for persistent memory, PMEM (a persistent "ram disk") and DAX (mmap persistent memory into userspace). This series adds a third, DAX-GUP, that allows DAX mappings to be the target of direct-i/o. It allows userspace to coordinate DMA/RDMA from/to persistent memory. The implementation leverages the ZONE_DEVICE mm-zone that went into 4.3-rc1 (also discussed at kernel summit) to flag pages that are owned and dynamically mapped by a device driver. The pmem driver, after mapping a persistent memory range into the system memmap via devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus page-backed pmem-pfns via flags in the new pfn_t type. The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the resulting pte(s) inserted into the process page tables with a new _PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys off _PAGE_DEVMAP to pin the device hosting the page range active. Finally, get_page() and put_page() are modified to take references against the device driver established page mapping. Finally, this need for "struct page" for persistent memory requires memory capacity to store the memmap array. Given the memmap array for a large pool of persistent may exhaust available DRAM introduce a mechanism to allocate the memmap from persistent memory. The new "struct vmem_altmap *" parameter to devm_memremap_pages() enables arch_add_memory() to use reserved pmem capacity rather than the page allocator. This patch (of 18): The core has developed a need for a "pfn_t" type [1]. Move the existing pfn_t in KVM to kvm_pfn_t [2]. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html [2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.html Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: Christoffer Dall <christoffer.dall@linaro.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 07:56:11 +07:00
kvm_pfn_t pfn, unsigned int wimg)
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
{
ref->pfn = pfn;
kvm: booke: clear host tlb reference flag on guest tlb invalidation On booke, "struct tlbe_ref" contains host tlb mapping information (pfn: for guest-pfn to pfn, flags: attribute associated with this mapping) for a guest tlb entry. So when a guest creates a TLB entry then "struct tlbe_ref" is set to point to valid "pfn" and set attributes in "flags" field of the above said structure. When a guest TLB entry is invalidated then flags field of corresponding "struct tlbe_ref" is updated to point that this is no more valid, also we selectively clear some other attribute bits, example: if E500_TLB_BITMAP was set then we clear E500_TLB_BITMAP, if E500_TLB_TLB0 is set then we clear this. Ideally we should clear complete "flags" as this entry is invalid and does not have anything to re-used. The other part of the problem is that when we use the same entry again then also we do not clear (started doing or-ing etc). So far it was working because the selectively clearing mentioned above actually clears "flags" what was set during TLB mapping. But the problem starts coming when we add more attributes to this then we need to selectively clear them and which is not needed. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-15 12:31:13 +07:00
ref->flags = E500_TLB_VALID;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
/* Use guest supplied MAS2_G and MAS2_E */
ref->flags |= (gtlbe->mas2 & MAS2_ATTRIB_MASK) | wimg;
kvm: powerpc: e500: mark page accessed when mapping a guest page Mark the guest page as accessed so that there is likely less chances of this page getting swap-out. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Acked-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-08-07 17:03:45 +07:00
/* Mark the page accessed */
kvm_set_pfn_accessed(pfn);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
if (tlbe_is_writable(gtlbe))
kvm_set_pfn_dirty(pfn);
}
static inline void kvmppc_e500_ref_release(struct tlbe_ref *ref)
{
if (ref->flags & E500_TLB_VALID) {
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
/* FIXME: don't log bogus pfn for TLB1 */
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
trace_kvm_booke206_ref_release(ref->pfn, ref->flags);
ref->flags = 0;
}
}
KVM: PPC: E500: Make clear_tlb_refs and clear_tlb1_bitmap static Host shadow TLB flushing is logic that the guest TLB code should have no insight about. Declare the internal clear_tlb_refs and clear_tlb1_bitmap functions static to the host TLB handling file. Instead of these, we can use the already exported kvmppc_core_flush_tlb(). This gives us a common API across the board to say "please flush any pending host shadow translation". Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-18 21:13:19 +07:00
static void clear_tlb1_bitmap(struct kvmppc_vcpu_e500 *vcpu_e500)
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
{
if (vcpu_e500->g2h_tlb1_map)
memset(vcpu_e500->g2h_tlb1_map, 0,
sizeof(u64) * vcpu_e500->gtlb_params[1].entries);
if (vcpu_e500->h2g_tlb1_rmap)
memset(vcpu_e500->h2g_tlb1_rmap, 0,
sizeof(unsigned int) * host_tlb_params[1].entries);
}
static void clear_tlb_privs(struct kvmppc_vcpu_e500 *vcpu_e500)
{
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
int tlbsel;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
int i;
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
for (tlbsel = 0; tlbsel <= 1; tlbsel++) {
for (i = 0; i < vcpu_e500->gtlb_params[tlbsel].entries; i++) {
struct tlbe_ref *ref =
&vcpu_e500->gtlb_priv[tlbsel][i].ref;
kvmppc_e500_ref_release(ref);
}
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
}
}
void kvmppc_core_flush_tlb(struct kvm_vcpu *vcpu)
{
struct kvmppc_vcpu_e500 *vcpu_e500 = to_e500(vcpu);
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
kvmppc_e500_tlbil_all(vcpu_e500);
clear_tlb_privs(vcpu_e500);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
clear_tlb1_bitmap(vcpu_e500);
}
/* TID must be supplied by the caller */
static void kvmppc_e500_setup_stlbe(
struct kvm_vcpu *vcpu,
struct kvm_book3e_206_tlb_entry *gtlbe,
int tsize, struct tlbe_ref *ref, u64 gvaddr,
struct kvm_book3e_206_tlb_entry *stlbe)
{
kvm: rename pfn_t to kvm_pfn_t To date, we have implemented two I/O usage models for persistent memory, PMEM (a persistent "ram disk") and DAX (mmap persistent memory into userspace). This series adds a third, DAX-GUP, that allows DAX mappings to be the target of direct-i/o. It allows userspace to coordinate DMA/RDMA from/to persistent memory. The implementation leverages the ZONE_DEVICE mm-zone that went into 4.3-rc1 (also discussed at kernel summit) to flag pages that are owned and dynamically mapped by a device driver. The pmem driver, after mapping a persistent memory range into the system memmap via devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus page-backed pmem-pfns via flags in the new pfn_t type. The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the resulting pte(s) inserted into the process page tables with a new _PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys off _PAGE_DEVMAP to pin the device hosting the page range active. Finally, get_page() and put_page() are modified to take references against the device driver established page mapping. Finally, this need for "struct page" for persistent memory requires memory capacity to store the memmap array. Given the memmap array for a large pool of persistent may exhaust available DRAM introduce a mechanism to allocate the memmap from persistent memory. The new "struct vmem_altmap *" parameter to devm_memremap_pages() enables arch_add_memory() to use reserved pmem capacity rather than the page allocator. This patch (of 18): The core has developed a need for a "pfn_t" type [1]. Move the existing pfn_t in KVM to kvm_pfn_t [2]. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html [2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.html Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: Christoffer Dall <christoffer.dall@linaro.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 07:56:11 +07:00
kvm_pfn_t pfn = ref->pfn;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
u32 pr = vcpu->arch.shared->msr & MSR_PR;
BUG_ON(!(ref->flags & E500_TLB_VALID));
/* Force IPROT=0 for all guest mappings. */
stlbe->mas1 = MAS1_TSIZE(tsize) | get_tlb_sts(gtlbe) | MAS1_VALID;
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
stlbe->mas2 = (gvaddr & MAS2_EPN) | (ref->flags & E500_TLB_MAS2_ATTR);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
stlbe->mas7_3 = ((u64)pfn << PAGE_SHIFT) |
e500_shadow_mas3_attrib(gtlbe->mas7_3, pr);
}
static inline int kvmppc_e500_shadow_map(struct kvmppc_vcpu_e500 *vcpu_e500,
u64 gvaddr, gfn_t gfn, struct kvm_book3e_206_tlb_entry *gtlbe,
int tlbsel, struct kvm_book3e_206_tlb_entry *stlbe,
struct tlbe_ref *ref)
{
struct kvm_memory_slot *slot;
unsigned long pfn = 0; /* silence GCC warning */
unsigned long hva;
int pfnmap = 0;
int tsize = BOOK3E_PAGESZ_4K;
kvm: ppc: booke: check range page invalidation progress on page setup When the MM code is invalidating a range of pages, it calls the KVM kvm_mmu_notifier_invalidate_range_start() notifier function, which calls kvm_unmap_hva_range(), which arranges to flush all the TLBs for guest pages. However, the Linux PTEs for the range being flushed are still valid at that point. We are not supposed to establish any new references to pages in the range until the ...range_end() notifier gets called. The PPC-specific KVM code doesn't get any explicit notification of that; instead, we are supposed to use mmu_notifier_retry() to test whether we are or have been inside a range flush notifier pair while we have been referencing a page. This patch calls the mmu_notifier_retry() while mapping the guest page to ensure we are not referencing a page when in range invalidation. This call is inside a region locked with kvm->mmu_lock, which is the same lock that is called by the KVM MMU notifier functions, thus ensuring that no new notification can proceed while we are in the locked region. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Acked-by: Alexander Graf <agraf@suse.de> [Backported to 3.12 - Paolo] Reviewed-by: Bharat Bhushan <bharat.bhushan@freescale.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-08-07 17:03:46 +07:00
int ret = 0;
unsigned long mmu_seq;
struct kvm *kvm = vcpu_e500->vcpu.kvm;
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
unsigned long tsize_pages = 0;
pte_t *ptep;
unsigned int wimg = 0;
pgd_t *pgdir;
powerpc/mm/thp: Make page table walk safe against thp split/collapse We can disable a THP split or a hugepage collapse by disabling irq. We do send IPI to all the cpus in the early part of split/collapse, and disabling local irq ensure we don't make progress with split/collapse. If the THP is getting split we return NULL from find_linux_pte_or_hugepte(). For all the current callers it should be ok. We need to be careful if we want to use returned pte_t pointer outside the irq disabled region. W.r.t to THP split, the pfn remains the same, but then a hugepage collapse will result in a pfn change. There are few steps we can take to avoid a hugepage collapse.One way is to take page reference inside the irq disable region. Other option is to take mmap_sem so that a parallel collapse will not happen. We can also disable collapse by taking pmd_lock. Another method used by kvm subsystem is to check whether we had a mmu_notifer update in between using mmu_notifier_retry(). Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:11:03 +07:00
unsigned long flags;
kvm: ppc: booke: check range page invalidation progress on page setup When the MM code is invalidating a range of pages, it calls the KVM kvm_mmu_notifier_invalidate_range_start() notifier function, which calls kvm_unmap_hva_range(), which arranges to flush all the TLBs for guest pages. However, the Linux PTEs for the range being flushed are still valid at that point. We are not supposed to establish any new references to pages in the range until the ...range_end() notifier gets called. The PPC-specific KVM code doesn't get any explicit notification of that; instead, we are supposed to use mmu_notifier_retry() to test whether we are or have been inside a range flush notifier pair while we have been referencing a page. This patch calls the mmu_notifier_retry() while mapping the guest page to ensure we are not referencing a page when in range invalidation. This call is inside a region locked with kvm->mmu_lock, which is the same lock that is called by the KVM MMU notifier functions, thus ensuring that no new notification can proceed while we are in the locked region. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Acked-by: Alexander Graf <agraf@suse.de> [Backported to 3.12 - Paolo] Reviewed-by: Bharat Bhushan <bharat.bhushan@freescale.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-08-07 17:03:46 +07:00
/* used to check for invalidations in progress */
mmu_seq = kvm->mmu_notifier_seq;
smp_rmb();
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
/*
* Translate guest physical to true physical, acquiring
* a page reference if it is normal, non-reserved memory.
*
* gfn_to_memslot() must succeed because otherwise we wouldn't
* have gotten this far. Eventually we should just pass the slot
* pointer through from the first lookup.
*/
slot = gfn_to_memslot(vcpu_e500->vcpu.kvm, gfn);
hva = gfn_to_hva_memslot(slot, gfn);
if (tlbsel == 1) {
struct vm_area_struct *vma;
down_read(&current->mm->mmap_sem);
vma = find_vma(current->mm, hva);
if (vma && hva >= vma->vm_start &&
(vma->vm_flags & VM_PFNMAP)) {
/*
* This VMA is a physically contiguous region (e.g.
* /dev/mem) that bypasses normal Linux page
* management. Find the overlap between the
* vma and the memslot.
*/
unsigned long start, end;
unsigned long slot_start, slot_end;
pfnmap = 1;
start = vma->vm_pgoff;
end = start +
((vma->vm_end - vma->vm_start) >> PAGE_SHIFT);
pfn = start + ((hva - vma->vm_start) >> PAGE_SHIFT);
slot_start = pfn - (gfn - slot->base_gfn);
slot_end = slot_start + slot->npages;
if (start < slot_start)
start = slot_start;
if (end > slot_end)
end = slot_end;
tsize = (gtlbe->mas1 & MAS1_TSIZE_MASK) >>
MAS1_TSIZE_SHIFT;
/*
* e500 doesn't implement the lowest tsize bit,
* or 1K pages.
*/
tsize = max(BOOK3E_PAGESZ_4K, tsize & ~1);
/*
* Now find the largest tsize (up to what the guest
* requested) that will cover gfn, stay within the
* range, and for which gfn and pfn are mutually
* aligned.
*/
for (; tsize > BOOK3E_PAGESZ_4K; tsize -= 2) {
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
unsigned long gfn_start, gfn_end;
KVM: PPC: e500: fix couple of shift operations on 64 bits Fix couple of cases where we shift left a 32-bit value thus might get truncated results on 64-bit targets. Signed-off-by: Laurentiu Tudor <Laurentiu.Tudor@freescale.com> Suggested-by: Scott Wood <scotttwood@freescale.com> Acked-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2015-10-01 19:58:03 +07:00
tsize_pages = 1UL << (tsize - 2);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
gfn_start = gfn & ~(tsize_pages - 1);
gfn_end = gfn_start + tsize_pages;
if (gfn_start + pfn - gfn < start)
continue;
if (gfn_end + pfn - gfn > end)
continue;
if ((gfn & (tsize_pages - 1)) !=
(pfn & (tsize_pages - 1)))
continue;
gvaddr &= ~((tsize_pages << PAGE_SHIFT) - 1);
pfn &= ~(tsize_pages - 1);
break;
}
} else if (vma && hva >= vma->vm_start &&
(vma->vm_flags & VM_HUGETLB)) {
unsigned long psize = vma_kernel_pagesize(vma);
tsize = (gtlbe->mas1 & MAS1_TSIZE_MASK) >>
MAS1_TSIZE_SHIFT;
/*
* Take the largest page size that satisfies both host
* and guest mapping
*/
tsize = min(__ilog2(psize) - 10, tsize);
/*
* e500 doesn't implement the lowest tsize bit,
* or 1K pages.
*/
tsize = max(BOOK3E_PAGESZ_4K, tsize & ~1);
}
up_read(&current->mm->mmap_sem);
}
if (likely(!pfnmap)) {
KVM: PPC: e500: fix couple of shift operations on 64 bits Fix couple of cases where we shift left a 32-bit value thus might get truncated results on 64-bit targets. Signed-off-by: Laurentiu Tudor <Laurentiu.Tudor@freescale.com> Suggested-by: Scott Wood <scotttwood@freescale.com> Acked-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2015-10-01 19:58:03 +07:00
tsize_pages = 1UL << (tsize + 10 - PAGE_SHIFT);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
pfn = gfn_to_pfn_memslot(slot, gfn);
if (is_error_noslot_pfn(pfn)) {
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
if (printk_ratelimit())
pr_err("%s: real page not found for gfn %lx\n",
__func__, (long)gfn);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
return -EINVAL;
}
/* Align guest and physical address to page map boundaries */
pfn &= ~(tsize_pages - 1);
gvaddr &= ~((tsize_pages << PAGE_SHIFT) - 1);
}
kvm: ppc: booke: check range page invalidation progress on page setup When the MM code is invalidating a range of pages, it calls the KVM kvm_mmu_notifier_invalidate_range_start() notifier function, which calls kvm_unmap_hva_range(), which arranges to flush all the TLBs for guest pages. However, the Linux PTEs for the range being flushed are still valid at that point. We are not supposed to establish any new references to pages in the range until the ...range_end() notifier gets called. The PPC-specific KVM code doesn't get any explicit notification of that; instead, we are supposed to use mmu_notifier_retry() to test whether we are or have been inside a range flush notifier pair while we have been referencing a page. This patch calls the mmu_notifier_retry() while mapping the guest page to ensure we are not referencing a page when in range invalidation. This call is inside a region locked with kvm->mmu_lock, which is the same lock that is called by the KVM MMU notifier functions, thus ensuring that no new notification can proceed while we are in the locked region. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Acked-by: Alexander Graf <agraf@suse.de> [Backported to 3.12 - Paolo] Reviewed-by: Bharat Bhushan <bharat.bhushan@freescale.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-08-07 17:03:46 +07:00
spin_lock(&kvm->mmu_lock);
if (mmu_notifier_retry(kvm, mmu_seq)) {
ret = -EAGAIN;
goto out;
}
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
pgdir = vcpu_e500->vcpu.arch.pgdir;
powerpc/mm/thp: Make page table walk safe against thp split/collapse We can disable a THP split or a hugepage collapse by disabling irq. We do send IPI to all the cpus in the early part of split/collapse, and disabling local irq ensure we don't make progress with split/collapse. If the THP is getting split we return NULL from find_linux_pte_or_hugepte(). For all the current callers it should be ok. We need to be careful if we want to use returned pte_t pointer outside the irq disabled region. W.r.t to THP split, the pfn remains the same, but then a hugepage collapse will result in a pfn change. There are few steps we can take to avoid a hugepage collapse.One way is to take page reference inside the irq disable region. Other option is to take mmap_sem so that a parallel collapse will not happen. We can also disable collapse by taking pmd_lock. Another method used by kvm subsystem is to check whether we had a mmu_notifer update in between using mmu_notifier_retry(). Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:11:03 +07:00
/*
* We are just looking at the wimg bits, so we don't
* care much about the trans splitting bit.
* We are holding kvm->mmu_lock so a notifier invalidate
* can't run hence pfn won't change.
*/
local_irq_save(flags);
powerpc/mm: Differentiate between hugetlb and THP during page walk We need to properly identify whether a hugepage is an explicit or a transparent hugepage in follow_huge_addr(). We used to depend on hugepage shift argument to do that. But in some case that can result in wrong results. For ex: On finding a transparent hugepage we set hugepage shift to PMD_SHIFT. But we can end up clearing the thp pte, via pmdp_huge_get_and_clear. We do prevent reusing the pfn page via the usage of kick_all_cpus_sync(). But that happens after we updated the pte to 0. Hence in follow_huge_addr() we can find hugepage shift set, but transparent huge page check fail for a thp pte. NOTE: We fixed a variant of this race against thp split in commit 691e95fd7396905a38d98919e9c150dbc3ea21a3 ("powerpc/mm/thp: Make page table walk safe against thp split/collapse") Without this patch, we may hit the BUG_ON(flags & FOLL_GET) in follow_page_mask occasionally. In the long term, we may want to switch ppc64 64k page size config to enable CONFIG_ARCH_WANT_GENERAL_HUGETLB Reported-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-10-09 10:02:21 +07:00
ptep = find_linux_pte_or_hugepte(pgdir, hva, NULL, NULL);
KVM: PPC: Use READ_ONCE when dereferencing pte_t pointer pte can get updated from other CPUs as part of multiple activities like THP split, huge page collapse, unmap. We need to make sure we don't reload the pte value again and again for different checks. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:09:12 +07:00
if (ptep) {
pte_t pte = READ_ONCE(*ptep);
powerpc/mm/thp: Make page table walk safe against thp split/collapse We can disable a THP split or a hugepage collapse by disabling irq. We do send IPI to all the cpus in the early part of split/collapse, and disabling local irq ensure we don't make progress with split/collapse. If the THP is getting split we return NULL from find_linux_pte_or_hugepte(). For all the current callers it should be ok. We need to be careful if we want to use returned pte_t pointer outside the irq disabled region. W.r.t to THP split, the pfn remains the same, but then a hugepage collapse will result in a pfn change. There are few steps we can take to avoid a hugepage collapse.One way is to take page reference inside the irq disable region. Other option is to take mmap_sem so that a parallel collapse will not happen. We can also disable collapse by taking pmd_lock. Another method used by kvm subsystem is to check whether we had a mmu_notifer update in between using mmu_notifier_retry(). Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:11:03 +07:00
if (pte_present(pte)) {
KVM: PPC: Use READ_ONCE when dereferencing pte_t pointer pte can get updated from other CPUs as part of multiple activities like THP split, huge page collapse, unmap. We need to make sure we don't reload the pte value again and again for different checks. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:09:12 +07:00
wimg = (pte_val(pte) >> PTE_WIMGE_SHIFT) &
MAS2_WIMGE_MASK;
powerpc/mm/thp: Make page table walk safe against thp split/collapse We can disable a THP split or a hugepage collapse by disabling irq. We do send IPI to all the cpus in the early part of split/collapse, and disabling local irq ensure we don't make progress with split/collapse. If the THP is getting split we return NULL from find_linux_pte_or_hugepte(). For all the current callers it should be ok. We need to be careful if we want to use returned pte_t pointer outside the irq disabled region. W.r.t to THP split, the pfn remains the same, but then a hugepage collapse will result in a pfn change. There are few steps we can take to avoid a hugepage collapse.One way is to take page reference inside the irq disable region. Other option is to take mmap_sem so that a parallel collapse will not happen. We can also disable collapse by taking pmd_lock. Another method used by kvm subsystem is to check whether we had a mmu_notifer update in between using mmu_notifier_retry(). Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:11:03 +07:00
local_irq_restore(flags);
} else {
local_irq_restore(flags);
KVM: PPC: Use READ_ONCE when dereferencing pte_t pointer pte can get updated from other CPUs as part of multiple activities like THP split, huge page collapse, unmap. We need to make sure we don't reload the pte value again and again for different checks. Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-03-30 12:09:12 +07:00
pr_err_ratelimited("%s: pte not present: gfn %lx,pfn %lx\n",
__func__, (long)gfn, pfn);
ret = -EINVAL;
goto out;
}
kvm: powerpc: use caching attributes as per linux pte KVM uses same WIM tlb attributes as the corresponding qemu pte. For this we now search the linux pte for the requested page and get these cache caching/coherency attributes from pte. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Reviewed-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-11-18 12:48:54 +07:00
}
kvmppc_e500_ref_setup(ref, gtlbe, pfn, wimg);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
kvmppc_e500_setup_stlbe(&vcpu_e500->vcpu, gtlbe, tsize,
ref, gvaddr, stlbe);
/* Clear i-cache for new pages */
kvmppc_mmu_flush_icache(pfn);
kvm: ppc: booke: check range page invalidation progress on page setup When the MM code is invalidating a range of pages, it calls the KVM kvm_mmu_notifier_invalidate_range_start() notifier function, which calls kvm_unmap_hva_range(), which arranges to flush all the TLBs for guest pages. However, the Linux PTEs for the range being flushed are still valid at that point. We are not supposed to establish any new references to pages in the range until the ...range_end() notifier gets called. The PPC-specific KVM code doesn't get any explicit notification of that; instead, we are supposed to use mmu_notifier_retry() to test whether we are or have been inside a range flush notifier pair while we have been referencing a page. This patch calls the mmu_notifier_retry() while mapping the guest page to ensure we are not referencing a page when in range invalidation. This call is inside a region locked with kvm->mmu_lock, which is the same lock that is called by the KVM MMU notifier functions, thus ensuring that no new notification can proceed while we are in the locked region. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Acked-by: Alexander Graf <agraf@suse.de> [Backported to 3.12 - Paolo] Reviewed-by: Bharat Bhushan <bharat.bhushan@freescale.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-08-07 17:03:46 +07:00
out:
spin_unlock(&kvm->mmu_lock);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
/* Drop refcount on page, so that mmu notifiers can clear it */
kvm_release_pfn_clean(pfn);
kvm: ppc: booke: check range page invalidation progress on page setup When the MM code is invalidating a range of pages, it calls the KVM kvm_mmu_notifier_invalidate_range_start() notifier function, which calls kvm_unmap_hva_range(), which arranges to flush all the TLBs for guest pages. However, the Linux PTEs for the range being flushed are still valid at that point. We are not supposed to establish any new references to pages in the range until the ...range_end() notifier gets called. The PPC-specific KVM code doesn't get any explicit notification of that; instead, we are supposed to use mmu_notifier_retry() to test whether we are or have been inside a range flush notifier pair while we have been referencing a page. This patch calls the mmu_notifier_retry() while mapping the guest page to ensure we are not referencing a page when in range invalidation. This call is inside a region locked with kvm->mmu_lock, which is the same lock that is called by the KVM MMU notifier functions, thus ensuring that no new notification can proceed while we are in the locked region. Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com> Acked-by: Alexander Graf <agraf@suse.de> [Backported to 3.12 - Paolo] Reviewed-by: Bharat Bhushan <bharat.bhushan@freescale.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-08-07 17:03:46 +07:00
return ret;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
}
/* XXX only map the one-one case, for now use TLB0 */
static int kvmppc_e500_tlb0_map(struct kvmppc_vcpu_e500 *vcpu_e500, int esel,
struct kvm_book3e_206_tlb_entry *stlbe)
{
struct kvm_book3e_206_tlb_entry *gtlbe;
struct tlbe_ref *ref;
int stlbsel = 0;
int sesel = 0;
int r;
gtlbe = get_entry(vcpu_e500, 0, esel);
ref = &vcpu_e500->gtlb_priv[0][esel].ref;
r = kvmppc_e500_shadow_map(vcpu_e500, get_tlb_eaddr(gtlbe),
get_tlb_raddr(gtlbe) >> PAGE_SHIFT,
gtlbe, 0, stlbe, ref);
if (r)
return r;
write_stlbe(vcpu_e500, gtlbe, stlbe, stlbsel, sesel);
return 0;
}
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
static int kvmppc_e500_tlb1_map_tlb1(struct kvmppc_vcpu_e500 *vcpu_e500,
struct tlbe_ref *ref,
int esel)
{
unsigned int sesel = vcpu_e500->host_tlb1_nv++;
if (unlikely(vcpu_e500->host_tlb1_nv >= tlb1_max_shadow_size()))
vcpu_e500->host_tlb1_nv = 0;
if (vcpu_e500->h2g_tlb1_rmap[sesel]) {
kvm/ppc/e500: h2g_tlb1_rmap: esel 0 is valid Add one to esel values in h2g_tlb1_rmap, so that "no mapping" can be distinguished from "esel 0". Note that we're not saved by the fact that host esel 0 is reserved for non-KVM use, because KVM host esel numbering is not the raw host numbering (see to_htlb1_esel). Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-02-14 02:37:48 +07:00
unsigned int idx = vcpu_e500->h2g_tlb1_rmap[sesel] - 1;
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
vcpu_e500->g2h_tlb1_map[idx] &= ~(1ULL << sesel);
}
kvm/ppc/e500: g2h_tlb1_map: clear old bit before setting new bit It's possible that we're using the same host TLB1 slot to map (a presumably different portion of) the same guest TLB1 entry. Clear the bit in the map before setting it, so that if the esels are the same the bit will remain set. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-02-14 02:37:49 +07:00
vcpu_e500->gtlb_priv[1][esel].ref.flags |= E500_TLB_BITMAP;
vcpu_e500->g2h_tlb1_map[esel] |= (u64)1 << sesel;
kvm/ppc/e500: h2g_tlb1_rmap: esel 0 is valid Add one to esel values in h2g_tlb1_rmap, so that "no mapping" can be distinguished from "esel 0". Note that we're not saved by the fact that host esel 0 is reserved for non-KVM use, because KVM host esel numbering is not the raw host numbering (see to_htlb1_esel). Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-02-14 02:37:48 +07:00
vcpu_e500->h2g_tlb1_rmap[sesel] = esel + 1;
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
WARN_ON(!(ref->flags & E500_TLB_VALID));
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
return sesel;
}
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
/* Caller must ensure that the specified guest TLB entry is safe to insert into
* the shadow TLB. */
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
/* For both one-one and one-to-many */
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
static int kvmppc_e500_tlb1_map(struct kvmppc_vcpu_e500 *vcpu_e500,
u64 gvaddr, gfn_t gfn, struct kvm_book3e_206_tlb_entry *gtlbe,
struct kvm_book3e_206_tlb_entry *stlbe, int esel)
{
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
struct tlbe_ref *ref = &vcpu_e500->gtlb_priv[1][esel].ref;
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
int sesel;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
int r;
r = kvmppc_e500_shadow_map(vcpu_e500, gvaddr, gfn, gtlbe, 1, stlbe,
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
ref);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
if (r)
return r;
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
/* Use TLB0 when we can only map a page with 4k */
if (get_tlb_tsize(stlbe) == BOOK3E_PAGESZ_4K) {
vcpu_e500->gtlb_priv[1][esel].ref.flags |= E500_TLB_TLB0;
write_stlbe(vcpu_e500, gtlbe, stlbe, 0, 0);
return 0;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
}
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
/* Otherwise map into TLB1 */
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
sesel = kvmppc_e500_tlb1_map_tlb1(vcpu_e500, ref, esel);
KVM: PPC: e500: Implement TLB1-in-TLB0 mapping When a host mapping fault happens in a guest TLB1 entry today, we map the translated guest entry into the host's TLB1. This isn't particularly clever when the guest is mapped by normal 4k pages, since these would be a lot better to put into TLB0 instead. This patch adds the required logic to map 4k TLB1 shadow maps into the host's TLB0. Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-17 23:54:36 +07:00
write_stlbe(vcpu_e500, gtlbe, stlbe, 1, sesel);
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
return 0;
}
void kvmppc_mmu_map(struct kvm_vcpu *vcpu, u64 eaddr, gpa_t gpaddr,
unsigned int index)
{
struct kvmppc_vcpu_e500 *vcpu_e500 = to_e500(vcpu);
struct tlbe_priv *priv;
struct kvm_book3e_206_tlb_entry *gtlbe, stlbe;
int tlbsel = tlbsel_of(index);
int esel = esel_of(index);
gtlbe = get_entry(vcpu_e500, tlbsel, esel);
switch (tlbsel) {
case 0:
priv = &vcpu_e500->gtlb_priv[tlbsel][esel];
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
/* Triggers after clear_tlb_privs or on initial mapping */
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
if (!(priv->ref.flags & E500_TLB_VALID)) {
kvmppc_e500_tlb0_map(vcpu_e500, esel, &stlbe);
} else {
kvmppc_e500_setup_stlbe(vcpu, gtlbe, BOOK3E_PAGESZ_4K,
&priv->ref, eaddr, &stlbe);
write_stlbe(vcpu_e500, gtlbe, &stlbe, 0, 0);
}
break;
case 1: {
gfn_t gfn = gpaddr >> PAGE_SHIFT;
kvmppc_e500_tlb1_map(vcpu_e500, eaddr, gfn, gtlbe, &stlbe,
esel);
break;
}
default:
BUG();
break;
}
}
KVM: PPC: Bookehv: Get vcpu's last instruction for emulation On book3e, KVM uses load external pid (lwepx) dedicated instruction to read guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI and LRAT), generated by loading a guest address, needs to be handled by KVM. These exceptions are generated in a substituted guest translation context (EPLC[EGS] = 1) from host context (MSR[GS] = 0). Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1), doing minimal checks on the fast path to avoid host performance degradation. lwepx exceptions originate from host state (MSR[GS] = 0) which implies additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious too intrusive for the host. Read guest last instruction from kvmppc_load_last_inst() by searching for the physical address and kmap it. This address the TODO for TLB eviction and execute-but-not-read entries, and allow us to get rid of lwepx until we are able to handle failures. A simple stress benchmark shows a 1% sys performance degradation compared with previous approach (lwepx without failure handling): time for i in `seq 1 10000`; do /bin/echo > /dev/null; done real 0m 8.85s user 0m 4.34s sys 0m 4.48s vs real 0m 8.84s user 0m 4.36s sys 0m 4.44s A solution to use lwepx and to handle its exceptions in KVM would be to temporary highjack the interrupt vector from host. This imposes additional synchronizations for cores like FSL e6500 that shares host IVOR registers between hardware threads. This optimized solution can be later developed on top of this patch. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:22 +07:00
#ifdef CONFIG_KVM_BOOKE_HV
int kvmppc_load_last_inst(struct kvm_vcpu *vcpu, enum instruction_type type,
u32 *instr)
{
gva_t geaddr;
hpa_t addr;
hfn_t pfn;
hva_t eaddr;
u32 mas1, mas2, mas3;
u64 mas7_mas3;
struct page *page;
unsigned int addr_space, psize_shift;
bool pr;
unsigned long flags;
/* Search TLB for guest pc to get the real address */
geaddr = kvmppc_get_pc(vcpu);
addr_space = (vcpu->arch.shared->msr & MSR_IS) >> MSR_IR_LG;
local_irq_save(flags);
mtspr(SPRN_MAS6, (vcpu->arch.pid << MAS6_SPID_SHIFT) | addr_space);
KVM: PPC: e500mc: Add support for single threaded vcpus on e6500 core ePAPR represents hardware threads as cpu node properties in device tree. So with existing QEMU, hardware threads are simply exposed as vcpus with one hardware thread. The e6500 core shares TLBs between hardware threads. Without tlb write conditional instruction, the Linux kernel uses per core mechanisms to protect against duplicate TLB entries. The guest is unable to detect real siblings threads, so it can't use the TLB protection mechanism. An alternative solution is to use the hypervisor to allocate different lpids to guest's vcpus that runs simultaneous on real siblings threads. On systems with two threads per core this patch halves the size of the lpid pool that the allocator sees and use two lpids per VM. Use even numbers to speedup vcpu lpid computation with consecutive lpids per VM: vm1 will use lpids 2 and 3, vm2 lpids 4 and 5, and so on. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> [agraf: fix spelling] Signed-off-by: Alexander Graf <agraf@suse.de>
2014-09-01 16:01:58 +07:00
mtspr(SPRN_MAS5, MAS5_SGS | get_lpid(vcpu));
KVM: PPC: Bookehv: Get vcpu's last instruction for emulation On book3e, KVM uses load external pid (lwepx) dedicated instruction to read guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI and LRAT), generated by loading a guest address, needs to be handled by KVM. These exceptions are generated in a substituted guest translation context (EPLC[EGS] = 1) from host context (MSR[GS] = 0). Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1), doing minimal checks on the fast path to avoid host performance degradation. lwepx exceptions originate from host state (MSR[GS] = 0) which implies additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious too intrusive for the host. Read guest last instruction from kvmppc_load_last_inst() by searching for the physical address and kmap it. This address the TODO for TLB eviction and execute-but-not-read entries, and allow us to get rid of lwepx until we are able to handle failures. A simple stress benchmark shows a 1% sys performance degradation compared with previous approach (lwepx without failure handling): time for i in `seq 1 10000`; do /bin/echo > /dev/null; done real 0m 8.85s user 0m 4.34s sys 0m 4.48s vs real 0m 8.84s user 0m 4.36s sys 0m 4.44s A solution to use lwepx and to handle its exceptions in KVM would be to temporary highjack the interrupt vector from host. This imposes additional synchronizations for cores like FSL e6500 that shares host IVOR registers between hardware threads. This optimized solution can be later developed on top of this patch. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:22 +07:00
asm volatile("tlbsx 0, %[geaddr]\n" : :
[geaddr] "r" (geaddr));
mtspr(SPRN_MAS5, 0);
mtspr(SPRN_MAS8, 0);
mas1 = mfspr(SPRN_MAS1);
mas2 = mfspr(SPRN_MAS2);
mas3 = mfspr(SPRN_MAS3);
#ifdef CONFIG_64BIT
mas7_mas3 = mfspr(SPRN_MAS7_MAS3);
#else
mas7_mas3 = ((u64)mfspr(SPRN_MAS7) << 32) | mas3;
#endif
local_irq_restore(flags);
/*
* If the TLB entry for guest pc was evicted, return to the guest.
* There are high chances to find a valid TLB entry next time.
*/
if (!(mas1 & MAS1_VALID))
return EMULATE_AGAIN;
/*
* Another thread may rewrite the TLB entry in parallel, don't
* execute from the address if the execute permission is not set
*/
pr = vcpu->arch.shared->msr & MSR_PR;
if (unlikely((pr && !(mas3 & MAS3_UX)) ||
(!pr && !(mas3 & MAS3_SX)))) {
pr_err_ratelimited(
treewide: fix typo in printk and Kconfig This patch fix spelling typo in printk and Kconfig within various part of kernel sources. Signed-off-by: Masanari Iida <standby24x7@gmail.com> Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2014-11-05 20:26:48 +07:00
"%s: Instruction emulation from guest address %08lx without execute permission\n",
KVM: PPC: Bookehv: Get vcpu's last instruction for emulation On book3e, KVM uses load external pid (lwepx) dedicated instruction to read guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI and LRAT), generated by loading a guest address, needs to be handled by KVM. These exceptions are generated in a substituted guest translation context (EPLC[EGS] = 1) from host context (MSR[GS] = 0). Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1), doing minimal checks on the fast path to avoid host performance degradation. lwepx exceptions originate from host state (MSR[GS] = 0) which implies additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious too intrusive for the host. Read guest last instruction from kvmppc_load_last_inst() by searching for the physical address and kmap it. This address the TODO for TLB eviction and execute-but-not-read entries, and allow us to get rid of lwepx until we are able to handle failures. A simple stress benchmark shows a 1% sys performance degradation compared with previous approach (lwepx without failure handling): time for i in `seq 1 10000`; do /bin/echo > /dev/null; done real 0m 8.85s user 0m 4.34s sys 0m 4.48s vs real 0m 8.84s user 0m 4.36s sys 0m 4.44s A solution to use lwepx and to handle its exceptions in KVM would be to temporary highjack the interrupt vector from host. This imposes additional synchronizations for cores like FSL e6500 that shares host IVOR registers between hardware threads. This optimized solution can be later developed on top of this patch. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:22 +07:00
__func__, geaddr);
return EMULATE_AGAIN;
}
/*
* The real address will be mapped by a cacheable, memory coherent,
* write-back page. Check for mismatches when LRAT is used.
*/
if (has_feature(vcpu, VCPU_FTR_MMU_V2) &&
unlikely((mas2 & MAS2_I) || (mas2 & MAS2_W) || !(mas2 & MAS2_M))) {
pr_err_ratelimited(
treewide: fix typo in printk and Kconfig This patch fix spelling typo in printk and Kconfig within various part of kernel sources. Signed-off-by: Masanari Iida <standby24x7@gmail.com> Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2014-11-05 20:26:48 +07:00
"%s: Instruction emulation from guest address %08lx mismatches storage attributes\n",
KVM: PPC: Bookehv: Get vcpu's last instruction for emulation On book3e, KVM uses load external pid (lwepx) dedicated instruction to read guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI and LRAT), generated by loading a guest address, needs to be handled by KVM. These exceptions are generated in a substituted guest translation context (EPLC[EGS] = 1) from host context (MSR[GS] = 0). Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1), doing minimal checks on the fast path to avoid host performance degradation. lwepx exceptions originate from host state (MSR[GS] = 0) which implies additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious too intrusive for the host. Read guest last instruction from kvmppc_load_last_inst() by searching for the physical address and kmap it. This address the TODO for TLB eviction and execute-but-not-read entries, and allow us to get rid of lwepx until we are able to handle failures. A simple stress benchmark shows a 1% sys performance degradation compared with previous approach (lwepx without failure handling): time for i in `seq 1 10000`; do /bin/echo > /dev/null; done real 0m 8.85s user 0m 4.34s sys 0m 4.48s vs real 0m 8.84s user 0m 4.36s sys 0m 4.44s A solution to use lwepx and to handle its exceptions in KVM would be to temporary highjack the interrupt vector from host. This imposes additional synchronizations for cores like FSL e6500 that shares host IVOR registers between hardware threads. This optimized solution can be later developed on top of this patch. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:22 +07:00
__func__, geaddr);
return EMULATE_AGAIN;
}
/* Get pfn */
psize_shift = MAS1_GET_TSIZE(mas1) + 10;
addr = (mas7_mas3 & (~0ULL << psize_shift)) |
(geaddr & ((1ULL << psize_shift) - 1ULL));
pfn = addr >> PAGE_SHIFT;
/* Guard against emulation from devices area */
if (unlikely(!page_is_ram(pfn))) {
treewide: fix typo in printk and Kconfig This patch fix spelling typo in printk and Kconfig within various part of kernel sources. Signed-off-by: Masanari Iida <standby24x7@gmail.com> Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2014-11-05 20:26:48 +07:00
pr_err_ratelimited("%s: Instruction emulation from non-RAM host address %08llx is not supported\n",
KVM: PPC: Bookehv: Get vcpu's last instruction for emulation On book3e, KVM uses load external pid (lwepx) dedicated instruction to read guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI and LRAT), generated by loading a guest address, needs to be handled by KVM. These exceptions are generated in a substituted guest translation context (EPLC[EGS] = 1) from host context (MSR[GS] = 0). Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1), doing minimal checks on the fast path to avoid host performance degradation. lwepx exceptions originate from host state (MSR[GS] = 0) which implies additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious too intrusive for the host. Read guest last instruction from kvmppc_load_last_inst() by searching for the physical address and kmap it. This address the TODO for TLB eviction and execute-but-not-read entries, and allow us to get rid of lwepx until we are able to handle failures. A simple stress benchmark shows a 1% sys performance degradation compared with previous approach (lwepx without failure handling): time for i in `seq 1 10000`; do /bin/echo > /dev/null; done real 0m 8.85s user 0m 4.34s sys 0m 4.48s vs real 0m 8.84s user 0m 4.36s sys 0m 4.44s A solution to use lwepx and to handle its exceptions in KVM would be to temporary highjack the interrupt vector from host. This imposes additional synchronizations for cores like FSL e6500 that shares host IVOR registers between hardware threads. This optimized solution can be later developed on top of this patch. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:22 +07:00
__func__, addr);
return EMULATE_AGAIN;
}
/* Map a page and get guest's instruction */
page = pfn_to_page(pfn);
eaddr = (unsigned long)kmap_atomic(page);
*instr = *(u32 *)(eaddr | (unsigned long)(addr & ~PAGE_MASK));
kunmap_atomic((u32 *)eaddr);
return EMULATE_DONE;
}
#else
KVM: PPC: Allow kvmppc_get_last_inst() to fail On book3e, guest last instruction is read on the exit path using load external pid (lwepx) dedicated instruction. This load operation may fail due to TLB eviction and execute-but-not-read entries. This patch lay down the path for an alternative solution to read the guest last instruction, by allowing kvmppc_get_lat_inst() function to fail. Architecture specific implmentations of kvmppc_load_last_inst() may read last guest instruction and instruct the emulation layer to re-execute the guest in case of failure. Make kvmppc_get_last_inst() definition common between architectures. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:21 +07:00
int kvmppc_load_last_inst(struct kvm_vcpu *vcpu, enum instruction_type type,
u32 *instr)
{
return EMULATE_AGAIN;
}
KVM: PPC: Bookehv: Get vcpu's last instruction for emulation On book3e, KVM uses load external pid (lwepx) dedicated instruction to read guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI and LRAT), generated by loading a guest address, needs to be handled by KVM. These exceptions are generated in a substituted guest translation context (EPLC[EGS] = 1) from host context (MSR[GS] = 0). Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1), doing minimal checks on the fast path to avoid host performance degradation. lwepx exceptions originate from host state (MSR[GS] = 0) which implies additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious too intrusive for the host. Read guest last instruction from kvmppc_load_last_inst() by searching for the physical address and kmap it. This address the TODO for TLB eviction and execute-but-not-read entries, and allow us to get rid of lwepx until we are able to handle failures. A simple stress benchmark shows a 1% sys performance degradation compared with previous approach (lwepx without failure handling): time for i in `seq 1 10000`; do /bin/echo > /dev/null; done real 0m 8.85s user 0m 4.34s sys 0m 4.48s vs real 0m 8.84s user 0m 4.36s sys 0m 4.44s A solution to use lwepx and to handle its exceptions in KVM would be to temporary highjack the interrupt vector from host. This imposes additional synchronizations for cores like FSL e6500 that shares host IVOR registers between hardware threads. This optimized solution can be later developed on top of this patch. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:22 +07:00
#endif
KVM: PPC: Allow kvmppc_get_last_inst() to fail On book3e, guest last instruction is read on the exit path using load external pid (lwepx) dedicated instruction. This load operation may fail due to TLB eviction and execute-but-not-read entries. This patch lay down the path for an alternative solution to read the guest last instruction, by allowing kvmppc_get_lat_inst() function to fail. Architecture specific implmentations of kvmppc_load_last_inst() may read last guest instruction and instruct the emulation layer to re-execute the guest in case of failure. Make kvmppc_get_last_inst() definition common between architectures. Signed-off-by: Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2014-07-23 23:06:21 +07:00
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
/************* MMU Notifiers *************/
int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
{
trace_kvm_unmap_hva(hva);
/*
* Flush all shadow tlb entries everywhere. This is slow, but
* we are 100% sure that we catch the to be unmapped page
*/
kvm_flush_remote_tlbs(kvm);
return 0;
}
int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end)
{
/* kvm_unmap_hva flushes everything anyways */
kvm_unmap_hva(kvm, start);
return 0;
}
kvm: Fix page ageing bugs 1. We were calling clear_flush_young_notify in unmap_one, but we are within an mmu notifier invalidate range scope. The spte exists no more (due to range_start) and the accessed bit info has already been propagated (due to kvm_pfn_set_accessed). Simply call clear_flush_young. 2. We clear_flush_young on a primary MMU PMD, but this may be mapped as a collection of PTEs by the secondary MMU (e.g. during log-dirty). This required expanding the interface of the clear_flush_young mmu notifier, so a lot of code has been trivially touched. 3. In the absence of shadow_accessed_mask (e.g. EPT A bit), we emulate the access bit by blowing the spte. This requires proper synchronizing with MMU notifier consumers, like every other removal of spte's does. Signed-off-by: Andres Lagar-Cavilla <andreslc@google.com> Acked-by: Rik van Riel <riel@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2014-09-23 04:54:42 +07:00
int kvm_age_hva(struct kvm *kvm, unsigned long start, unsigned long end)
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
{
/* XXX could be more clever ;) */
return 0;
}
int kvm_test_age_hva(struct kvm *kvm, unsigned long hva)
{
/* XXX could be more clever ;) */
return 0;
}
void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte)
{
/* The page will get remapped properly on its next fault */
kvm_unmap_hva(kvm, hva);
}
/*****************************************/
int e500_mmu_host_init(struct kvmppc_vcpu_e500 *vcpu_e500)
{
host_tlb_params[0].entries = mfspr(SPRN_TLB0CFG) & TLBnCFG_N_ENTRY;
host_tlb_params[1].entries = mfspr(SPRN_TLB1CFG) & TLBnCFG_N_ENTRY;
/*
* This should never happen on real e500 hardware, but is
* architecturally possible -- e.g. in some weird nested
* virtualization case.
*/
if (host_tlb_params[0].entries == 0 ||
host_tlb_params[1].entries == 0) {
pr_err("%s: need to know host tlb size\n", __func__);
return -ENODEV;
}
host_tlb_params[0].ways = (mfspr(SPRN_TLB0CFG) & TLBnCFG_ASSOC) >>
TLBnCFG_ASSOC_SHIFT;
host_tlb_params[1].ways = host_tlb_params[1].entries;
if (!is_power_of_2(host_tlb_params[0].entries) ||
!is_power_of_2(host_tlb_params[0].ways) ||
host_tlb_params[0].entries < host_tlb_params[0].ways ||
host_tlb_params[0].ways == 0) {
pr_err("%s: bad tlb0 host config: %u entries %u ways\n",
__func__, host_tlb_params[0].entries,
host_tlb_params[0].ways);
return -ENODEV;
}
host_tlb_params[0].sets =
host_tlb_params[0].entries / host_tlb_params[0].ways;
host_tlb_params[1].sets = 1;
vcpu_e500->h2g_tlb1_rmap = kzalloc(sizeof(unsigned int) *
host_tlb_params[1].entries,
GFP_KERNEL);
if (!vcpu_e500->h2g_tlb1_rmap)
kvm/ppc/e500: eliminate tlb_refs Commit 523f0e5421c12610527c620b983b443f329e3a32 ("KVM: PPC: E500: Explicitly mark shadow maps invalid") began using E500_TLB_VALID for guest TLB1 entries, and skipping invalidations if it's not set. However, when E500_TLB_VALID was set for such entries, it was on a fake local ref, and so the invalidations never happen. gtlb_privs is documented as being only for guest TLB0, though we already violate that with E500_TLB_BITMAP. Now that we have MMU notifiers, and thus don't need to actually retain a reference to the mapped pages, get rid of tlb_refs, and use gtlb_privs for E500_TLB_VALID in TLB1. Since we can have more than one host TLB entry for a given tlbe_ref, be careful not to clear existing flags that are relevant to other host TLB entries when preparing a new host TLB entry. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Alexander Graf <agraf@suse.de>
2013-03-06 23:02:49 +07:00
return -EINVAL;
KVM: PPC: E500: Split host and guest MMU parts This patch splits the file e500_tlb.c into e500_mmu.c (guest TLB handling) and e500_mmu_host.c (host TLB handling). The main benefit of this split is readability and maintainability. It's just a lot harder to write dirty code :). Signed-off-by: Alexander Graf <agraf@suse.de>
2013-01-11 21:22:45 +07:00
return 0;
}
void e500_mmu_host_uninit(struct kvmppc_vcpu_e500 *vcpu_e500)
{
kfree(vcpu_e500->h2g_tlb1_rmap);
}
Reference in New Issue Copy Permalink