linux_dsm_epyc7002/arch/powerpc/mm/hash_utils_64.c

776 lines
19 KiB
C
Raw Normal View History

/*
* PowerPC64 port by Mike Corrigan and Dave Engebretsen
* {mikejc|engebret}@us.ibm.com
*
* Copyright (c) 2000 Mike Corrigan <mikejc@us.ibm.com>
*
* SMP scalability work:
* Copyright (C) 2001 Anton Blanchard <anton@au.ibm.com>, IBM
*
* Module name: htab.c
*
* Description:
* PowerPC Hashed Page Table functions
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
#undef DEBUG
#undef DEBUG_LOW
#include <linux/config.h>
#include <linux/spinlock.h>
#include <linux/errno.h>
#include <linux/sched.h>
#include <linux/proc_fs.h>
#include <linux/stat.h>
#include <linux/sysctl.h>
#include <linux/ctype.h>
#include <linux/cache.h>
#include <linux/init.h>
#include <linux/signal.h>
#include <asm/processor.h>
#include <asm/pgtable.h>
#include <asm/mmu.h>
#include <asm/mmu_context.h>
#include <asm/page.h>
#include <asm/types.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/machdep.h>
#include <asm/lmb.h>
#include <asm/abs_addr.h>
#include <asm/tlbflush.h>
#include <asm/io.h>
#include <asm/eeh.h>
#include <asm/tlb.h>
#include <asm/cacheflush.h>
#include <asm/cputable.h>
#include <asm/abs_addr.h>
#include <asm/sections.h>
#ifdef DEBUG
#define DBG(fmt...) udbg_printf(fmt)
#else
#define DBG(fmt...)
#endif
#ifdef DEBUG_LOW
#define DBG_LOW(fmt...) udbg_printf(fmt)
#else
#define DBG_LOW(fmt...)
#endif
#define KB (1024)
#define MB (1024*KB)
/*
* Note: pte --> Linux PTE
* HPTE --> PowerPC Hashed Page Table Entry
*
* Execution context:
* htab_initialize is called with the MMU off (of course), but
* the kernel has been copied down to zero so it can directly
* reference global data. At this point it is very difficult
* to print debug info.
*
*/
#ifdef CONFIG_U3_DART
extern unsigned long dart_tablebase;
#endif /* CONFIG_U3_DART */
static unsigned long _SDR1;
struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT];
hpte_t *htab_address;
unsigned long htab_hash_mask;
int mmu_linear_psize = MMU_PAGE_4K;
int mmu_virtual_psize = MMU_PAGE_4K;
#ifdef CONFIG_HUGETLB_PAGE
int mmu_huge_psize = MMU_PAGE_16M;
unsigned int HPAGE_SHIFT;
#endif
/* There are definitions of page sizes arrays to be used when none
* is provided by the firmware.
*/
/* Pre-POWER4 CPUs (4k pages only)
*/
struct mmu_psize_def mmu_psize_defaults_old[] = {
[MMU_PAGE_4K] = {
.shift = 12,
.sllp = 0,
.penc = 0,
.avpnm = 0,
.tlbiel = 0,
},
};
/* POWER4, GPUL, POWER5
*
* Support for 16Mb large pages
*/
struct mmu_psize_def mmu_psize_defaults_gp[] = {
[MMU_PAGE_4K] = {
.shift = 12,
.sllp = 0,
.penc = 0,
.avpnm = 0,
.tlbiel = 1,
},
[MMU_PAGE_16M] = {
.shift = 24,
.sllp = SLB_VSID_L,
.penc = 0,
.avpnm = 0x1UL,
.tlbiel = 0,
},
};
int htab_bolt_mapping(unsigned long vstart, unsigned long vend,
unsigned long pstart, unsigned long mode, int psize)
{
unsigned long vaddr, paddr;
unsigned int step, shift;
unsigned long tmp_mode;
int ret = 0;
shift = mmu_psize_defs[psize].shift;
step = 1 << shift;
for (vaddr = vstart, paddr = pstart; vaddr < vend;
vaddr += step, paddr += step) {
unsigned long vpn, hash, hpteg;
unsigned long vsid = get_kernel_vsid(vaddr);
unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff);
vpn = va >> shift;
tmp_mode = mode;
/* Make non-kernel text non-executable */
if (!in_kernel_text(vaddr))
tmp_mode = mode | HPTE_R_N;
hash = hpt_hash(va, shift);
hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP);
/* The crap below can be cleaned once ppd_md.probe() can
* set up the hash callbacks, thus we can just used the
* normal insert callback here.
*/
#ifdef CONFIG_PPC_ISERIES
if (_machine == PLATFORM_ISERIES_LPAR)
ret = iSeries_hpte_insert(hpteg, va,
virt_to_abs(paddr),
tmp_mode,
HPTE_V_BOLTED,
psize);
else
#endif
#ifdef CONFIG_PPC_PSERIES
if (_machine & PLATFORM_LPAR)
ret = pSeries_lpar_hpte_insert(hpteg, va,
virt_to_abs(paddr),
tmp_mode,
HPTE_V_BOLTED,
psize);
else
#endif
#ifdef CONFIG_PPC_MULTIPLATFORM
ret = native_hpte_insert(hpteg, va,
virt_to_abs(paddr),
tmp_mode, HPTE_V_BOLTED,
psize);
#endif
if (ret < 0)
break;
}
return ret < 0 ? ret : 0;
}
static int __init htab_dt_scan_page_sizes(unsigned long node,
const char *uname, int depth,
void *data)
{
char *type = of_get_flat_dt_prop(node, "device_type", NULL);
u32 *prop;
unsigned long size = 0;
/* We are scanning "cpu" nodes only */
if (type == NULL || strcmp(type, "cpu") != 0)
return 0;
prop = (u32 *)of_get_flat_dt_prop(node,
"ibm,segment-page-sizes", &size);
if (prop != NULL) {
DBG("Page sizes from device-tree:\n");
size /= 4;
cur_cpu_spec->cpu_features &= ~(CPU_FTR_16M_PAGE);
while(size > 0) {
unsigned int shift = prop[0];
unsigned int slbenc = prop[1];
unsigned int lpnum = prop[2];
unsigned int lpenc = 0;
struct mmu_psize_def *def;
int idx = -1;
size -= 3; prop += 3;
while(size > 0 && lpnum) {
if (prop[0] == shift)
lpenc = prop[1];
prop += 2; size -= 2;
lpnum--;
}
switch(shift) {
case 0xc:
idx = MMU_PAGE_4K;
break;
case 0x10:
idx = MMU_PAGE_64K;
break;
case 0x14:
idx = MMU_PAGE_1M;
break;
case 0x18:
idx = MMU_PAGE_16M;
cur_cpu_spec->cpu_features |= CPU_FTR_16M_PAGE;
break;
case 0x22:
idx = MMU_PAGE_16G;
break;
}
if (idx < 0)
continue;
def = &mmu_psize_defs[idx];
def->shift = shift;
if (shift <= 23)
def->avpnm = 0;
else
def->avpnm = (1 << (shift - 23)) - 1;
def->sllp = slbenc;
def->penc = lpenc;
/* We don't know for sure what's up with tlbiel, so
* for now we only set it for 4K and 64K pages
*/
if (idx == MMU_PAGE_4K || idx == MMU_PAGE_64K)
def->tlbiel = 1;
else
def->tlbiel = 0;
DBG(" %d: shift=%02x, sllp=%04x, avpnm=%08x, "
"tlbiel=%d, penc=%d\n",
idx, shift, def->sllp, def->avpnm, def->tlbiel,
def->penc);
}
return 1;
}
return 0;
}
static void __init htab_init_page_sizes(void)
{
int rc;
/* Default to 4K pages only */
memcpy(mmu_psize_defs, mmu_psize_defaults_old,
sizeof(mmu_psize_defaults_old));
/*
* Try to find the available page sizes in the device-tree
*/
rc = of_scan_flat_dt(htab_dt_scan_page_sizes, NULL);
if (rc != 0) /* Found */
goto found;
/*
* Not in the device-tree, let's fallback on known size
* list for 16M capable GP & GR
*/
if ((_machine != PLATFORM_ISERIES_LPAR) &&
cpu_has_feature(CPU_FTR_16M_PAGE))
memcpy(mmu_psize_defs, mmu_psize_defaults_gp,
sizeof(mmu_psize_defaults_gp));
found:
/*
* Pick a size for the linear mapping. Currently, we only support
* 16M, 1M and 4K which is the default
*/
if (mmu_psize_defs[MMU_PAGE_16M].shift)
mmu_linear_psize = MMU_PAGE_16M;
else if (mmu_psize_defs[MMU_PAGE_1M].shift)
mmu_linear_psize = MMU_PAGE_1M;
/*
* Pick a size for the ordinary pages. Default is 4K, we support
* 64K if cache inhibited large pages are supported by the
* processor
*/
#ifdef CONFIG_PPC_64K_PAGES
if (mmu_psize_defs[MMU_PAGE_64K].shift &&
cpu_has_feature(CPU_FTR_CI_LARGE_PAGE))
mmu_virtual_psize = MMU_PAGE_64K;
#endif
printk(KERN_INFO "Page orders: linear mapping = %d, others = %d\n",
mmu_psize_defs[mmu_linear_psize].shift,
mmu_psize_defs[mmu_virtual_psize].shift);
#ifdef CONFIG_HUGETLB_PAGE
/* Init large page size. Currently, we pick 16M or 1M depending
* on what is available
*/
if (mmu_psize_defs[MMU_PAGE_16M].shift)
mmu_huge_psize = MMU_PAGE_16M;
[PATCH] ppc64: Fix bug in SLB miss handler for hugepages This patch, however, should be applied on top of the 64k-page-size patch to fix some problems with hugepage (some pre-existing, another introduced by this patch). The patch fixes a bug in the SLB miss handler for hugepages on ppc64 introduced by the dynamic hugepage patch (commit id c594adad5653491813959277fb87a2fef54c4e05) due to a misunderstanding of the srd instruction's behaviour (mea culpa). The problem arises when a 64-bit process maps some hugepages in the low 4GB of the address space (unusual). In this case, as well as the 256M segment in question being marked for hugepages, other segments at 32G intervals will be incorrectly marked for hugepages. In the process, this patch tweaks the semantics of the hugepage bitmaps to be more sensible. Previously, an address below 4G was marked for hugepages if the appropriate segment bit in the "low areas" bitmask was set *or* if the low bit in the "high areas" bitmap was set (which would mark all addresses below 1TB for hugepage). With this patch, any given address is governed by a single bitmap. Addresses below 4GB are marked for hugepage if and only if their bit is set in the "low areas" bitmap (256M granularity). Addresses between 4GB and 1TB are marked for hugepage iff the low bit in the "high areas" bitmap is set. Higher addresses are marked for hugepage iff their bit in the "high areas" bitmap is set (1TB granularity). To avoid conflicts, this patch must be applied on top of BenH's pending patch for 64k base page size [0]. As such, this patch also addresses a hugepage problem introduced by that patch. That patch allows hugepages of 1MB in size on hardware which supports it, however, that won't work when using 4k pages (4 level pagetable), because in that case hugepage PTEs are stored at the PMD level, and each PMD entry maps 2MB. This patch simply disallows hugepages in that case (we can do something cleverer to re-enable them some other day). Built, booted, and a handful of hugepage related tests passed on POWER5 LPAR (both ARCH=powerpc and ARCH=ppc64). [0] http://gate.crashing.org/~benh/ppc64-64k-pages.diff Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 15:57:52 +07:00
/* With 4k/4level pagetables, we can't (for now) cope with a
* huge page size < PMD_SIZE */
else if (mmu_psize_defs[MMU_PAGE_1M].shift)
mmu_huge_psize = MMU_PAGE_1M;
/* Calculate HPAGE_SHIFT and sanity check it */
[PATCH] ppc64: Fix bug in SLB miss handler for hugepages This patch, however, should be applied on top of the 64k-page-size patch to fix some problems with hugepage (some pre-existing, another introduced by this patch). The patch fixes a bug in the SLB miss handler for hugepages on ppc64 introduced by the dynamic hugepage patch (commit id c594adad5653491813959277fb87a2fef54c4e05) due to a misunderstanding of the srd instruction's behaviour (mea culpa). The problem arises when a 64-bit process maps some hugepages in the low 4GB of the address space (unusual). In this case, as well as the 256M segment in question being marked for hugepages, other segments at 32G intervals will be incorrectly marked for hugepages. In the process, this patch tweaks the semantics of the hugepage bitmaps to be more sensible. Previously, an address below 4G was marked for hugepages if the appropriate segment bit in the "low areas" bitmask was set *or* if the low bit in the "high areas" bitmap was set (which would mark all addresses below 1TB for hugepage). With this patch, any given address is governed by a single bitmap. Addresses below 4GB are marked for hugepage if and only if their bit is set in the "low areas" bitmap (256M granularity). Addresses between 4GB and 1TB are marked for hugepage iff the low bit in the "high areas" bitmap is set. Higher addresses are marked for hugepage iff their bit in the "high areas" bitmap is set (1TB granularity). To avoid conflicts, this patch must be applied on top of BenH's pending patch for 64k base page size [0]. As such, this patch also addresses a hugepage problem introduced by that patch. That patch allows hugepages of 1MB in size on hardware which supports it, however, that won't work when using 4k pages (4 level pagetable), because in that case hugepage PTEs are stored at the PMD level, and each PMD entry maps 2MB. This patch simply disallows hugepages in that case (we can do something cleverer to re-enable them some other day). Built, booted, and a handful of hugepage related tests passed on POWER5 LPAR (both ARCH=powerpc and ARCH=ppc64). [0] http://gate.crashing.org/~benh/ppc64-64k-pages.diff Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-07 15:57:52 +07:00
if (mmu_psize_defs[mmu_huge_psize].shift > MIN_HUGEPTE_SHIFT &&
mmu_psize_defs[mmu_huge_psize].shift < SID_SHIFT)
HPAGE_SHIFT = mmu_psize_defs[mmu_huge_psize].shift;
else
HPAGE_SHIFT = 0; /* No huge pages dude ! */
#endif /* CONFIG_HUGETLB_PAGE */
}
static int __init htab_dt_scan_pftsize(unsigned long node,
const char *uname, int depth,
void *data)
{
char *type = of_get_flat_dt_prop(node, "device_type", NULL);
u32 *prop;
/* We are scanning "cpu" nodes only */
if (type == NULL || strcmp(type, "cpu") != 0)
return 0;
prop = (u32 *)of_get_flat_dt_prop(node, "ibm,pft-size", NULL);
if (prop != NULL) {
/* pft_size[0] is the NUMA CEC cookie */
ppc64_pft_size = prop[1];
return 1;
}
return 0;
}
static unsigned long __init htab_get_table_size(void)
{
unsigned long mem_size, rnd_mem_size, pteg_count;
/* If hash size isn't already provided by the platform, we try to
* retreive it from the device-tree. If it's not there neither, we
* calculate it now based on the total RAM size
*/
if (ppc64_pft_size == 0)
of_scan_flat_dt(htab_dt_scan_pftsize, NULL);
if (ppc64_pft_size)
return 1UL << ppc64_pft_size;
/* round mem_size up to next power of 2 */
mem_size = lmb_phys_mem_size();
rnd_mem_size = 1UL << __ilog2(mem_size);
if (rnd_mem_size < mem_size)
rnd_mem_size <<= 1;
/* # pages / 2 */
pteg_count = max(rnd_mem_size >> (12 + 1), 1UL << 11);
return pteg_count << 7;
}
#ifdef CONFIG_MEMORY_HOTPLUG
void create_section_mapping(unsigned long start, unsigned long end)
{
BUG_ON(htab_bolt_mapping(start, end, start,
_PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_COHERENT | PP_RWXX,
mmu_linear_psize));
}
#endif /* CONFIG_MEMORY_HOTPLUG */
void __init htab_initialize(void)
{
unsigned long table, htab_size_bytes;
unsigned long pteg_count;
unsigned long mode_rw;
unsigned long base = 0, size = 0;
int i;
extern unsigned long tce_alloc_start, tce_alloc_end;
DBG(" -> htab_initialize()\n");
/* Initialize page sizes */
htab_init_page_sizes();
/*
* Calculate the required size of the htab. We want the number of
* PTEGs to equal one half the number of real pages.
*/
htab_size_bytes = htab_get_table_size();
pteg_count = htab_size_bytes >> 7;
htab_hash_mask = pteg_count - 1;
if (platform_is_lpar()) {
/* Using a hypervisor which owns the htab */
htab_address = NULL;
_SDR1 = 0;
} else {
/* Find storage for the HPT. Must be contiguous in
* the absolute address space.
*/
table = lmb_alloc(htab_size_bytes, htab_size_bytes);
BUG_ON(table == 0);
DBG("Hash table allocated at %lx, size: %lx\n", table,
htab_size_bytes);
htab_address = abs_to_virt(table);
/* htab absolute addr + encoded htabsize */
_SDR1 = table + __ilog2(pteg_count) - 11;
/* Initialize the HPT with no entries */
memset((void *)table, 0, htab_size_bytes);
/* Set SDR1 */
mtspr(SPRN_SDR1, _SDR1);
}
mode_rw = _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_COHERENT | PP_RWXX;
/* On U3 based machines, we need to reserve the DART area and
* _NOT_ map it to avoid cache paradoxes as it's remapped non
* cacheable later on
*/
/* create bolted the linear mapping in the hash table */
for (i=0; i < lmb.memory.cnt; i++) {
base = lmb.memory.region[i].base + KERNELBASE;
size = lmb.memory.region[i].size;
DBG("creating mapping for region: %lx : %lx\n", base, size);
#ifdef CONFIG_U3_DART
/* Do not map the DART space. Fortunately, it will be aligned
* in such a way that it will not cross two lmb regions and
* will fit within a single 16Mb page.
* The DART space is assumed to be a full 16Mb region even if
* we only use 2Mb of that space. We will use more of it later
* for AGP GART. We have to use a full 16Mb large page.
*/
DBG("DART base: %lx\n", dart_tablebase);
if (dart_tablebase != 0 && dart_tablebase >= base
&& dart_tablebase < (base + size)) {
if (base != dart_tablebase)
BUG_ON(htab_bolt_mapping(base, dart_tablebase,
base, mode_rw,
mmu_linear_psize));
if ((base + size) > (dart_tablebase + 16*MB))
BUG_ON(htab_bolt_mapping(dart_tablebase+16*MB,
base + size,
dart_tablebase+16*MB,
mode_rw,
mmu_linear_psize));
continue;
}
#endif /* CONFIG_U3_DART */
BUG_ON(htab_bolt_mapping(base, base + size, base,
mode_rw, mmu_linear_psize));
}
/*
* If we have a memory_limit and we've allocated TCEs then we need to
* explicitly map the TCE area at the top of RAM. We also cope with the
* case that the TCEs start below memory_limit.
* tce_alloc_start/end are 16MB aligned so the mapping should work
* for either 4K or 16MB pages.
*/
if (tce_alloc_start) {
tce_alloc_start += KERNELBASE;
tce_alloc_end += KERNELBASE;
if (base + size >= tce_alloc_start)
tce_alloc_start = base + size + 1;
BUG_ON(htab_bolt_mapping(tce_alloc_start, tce_alloc_end,
tce_alloc_start, mode_rw,
mmu_linear_psize));
}
DBG(" <- htab_initialize()\n");
}
#undef KB
#undef MB
void __init htab_initialize_secondary(void)
{
if (!platform_is_lpar())
mtspr(SPRN_SDR1, _SDR1);
}
/*
* Called by asm hashtable.S for doing lazy icache flush
*/
unsigned int hash_page_do_lazy_icache(unsigned int pp, pte_t pte, int trap)
{
struct page *page;
if (!pfn_valid(pte_pfn(pte)))
return pp;
page = pte_page(pte);
/* page is dirty */
if (!test_bit(PG_arch_1, &page->flags) && !PageReserved(page)) {
if (trap == 0x400) {
__flush_dcache_icache(page_address(page));
set_bit(PG_arch_1, &page->flags);
} else
pp |= HPTE_R_N;
}
return pp;
}
/* Result code is:
* 0 - handled
* 1 - normal page fault
* -1 - critical hash insertion error
*/
int hash_page(unsigned long ea, unsigned long access, unsigned long trap)
{
void *pgdir;
unsigned long vsid;
struct mm_struct *mm;
pte_t *ptep;
cpumask_t tmp;
int rc, user_region = 0, local = 0;
DBG_LOW("hash_page(ea=%016lx, access=%lx, trap=%lx\n",
ea, access, trap);
if ((ea & ~REGION_MASK) >= PGTABLE_RANGE) {
DBG_LOW(" out of pgtable range !\n");
return 1;
}
/* Get region & vsid */
switch (REGION_ID(ea)) {
case USER_REGION_ID:
user_region = 1;
mm = current->mm;
if (! mm) {
DBG_LOW(" user region with no mm !\n");
return 1;
}
vsid = get_vsid(mm->context.id, ea);
break;
case VMALLOC_REGION_ID:
mm = &init_mm;
vsid = get_kernel_vsid(ea);
break;
default:
/* Not a valid range
* Send the problem up to do_page_fault
*/
return 1;
}
DBG_LOW(" mm=%p, mm->pgdir=%p, vsid=%016lx\n", mm, mm->pgd, vsid);
/* Get pgdir */
pgdir = mm->pgd;
if (pgdir == NULL)
return 1;
/* Check CPU locality */
tmp = cpumask_of_cpu(smp_processor_id());
if (user_region && cpus_equal(mm->cpu_vm_mask, tmp))
local = 1;
/* Handle hugepage regions */
if (unlikely(in_hugepage_area(mm->context, ea))) {
DBG_LOW(" -> huge page !\n");
return hash_huge_page(mm, access, ea, vsid, local, trap);
}
/* Get PTE and page size from page tables */
ptep = find_linux_pte(pgdir, ea);
if (ptep == NULL || !pte_present(*ptep)) {
DBG_LOW(" no PTE !\n");
return 1;
}
#ifndef CONFIG_PPC_64K_PAGES
DBG_LOW(" i-pte: %016lx\n", pte_val(*ptep));
#else
DBG_LOW(" i-pte: %016lx %016lx\n", pte_val(*ptep),
pte_val(*(ptep + PTRS_PER_PTE)));
#endif
/* Pre-check access permissions (will be re-checked atomically
* in __hash_page_XX but this pre-check is a fast path
*/
if (access & ~pte_val(*ptep)) {
DBG_LOW(" no access !\n");
return 1;
}
/* Do actual hashing */
#ifndef CONFIG_PPC_64K_PAGES
rc = __hash_page_4K(ea, access, vsid, ptep, trap, local);
#else
if (mmu_virtual_psize == MMU_PAGE_64K)
rc = __hash_page_64K(ea, access, vsid, ptep, trap, local);
else
rc = __hash_page_4K(ea, access, vsid, ptep, trap, local);
#endif /* CONFIG_PPC_64K_PAGES */
#ifndef CONFIG_PPC_64K_PAGES
DBG_LOW(" o-pte: %016lx\n", pte_val(*ptep));
#else
DBG_LOW(" o-pte: %016lx %016lx\n", pte_val(*ptep),
pte_val(*(ptep + PTRS_PER_PTE)));
#endif
DBG_LOW(" -> rc=%d\n", rc);
return rc;
}
void hash_preload(struct mm_struct *mm, unsigned long ea,
unsigned long access, unsigned long trap)
{
unsigned long vsid;
void *pgdir;
pte_t *ptep;
cpumask_t mask;
unsigned long flags;
int local = 0;
/* We don't want huge pages prefaulted for now
*/
if (unlikely(in_hugepage_area(mm->context, ea)))
return;
DBG_LOW("hash_preload(mm=%p, mm->pgdir=%p, ea=%016lx, access=%lx,"
" trap=%lx\n", mm, mm->pgd, ea, access, trap);
/* Get PTE, VSID, access mask */
pgdir = mm->pgd;
if (pgdir == NULL)
return;
ptep = find_linux_pte(pgdir, ea);
if (!ptep)
return;
vsid = get_vsid(mm->context.id, ea);
/* Hash it in */
local_irq_save(flags);
mask = cpumask_of_cpu(smp_processor_id());
if (cpus_equal(mm->cpu_vm_mask, mask))
local = 1;
#ifndef CONFIG_PPC_64K_PAGES
__hash_page_4K(ea, access, vsid, ptep, trap, local);
#else
if (mmu_virtual_psize == MMU_PAGE_64K)
__hash_page_64K(ea, access, vsid, ptep, trap, local);
else
__hash_page_4K(ea, access, vsid, ptep, trap, local);
#endif /* CONFIG_PPC_64K_PAGES */
local_irq_restore(flags);
}
void flush_hash_page(unsigned long va, real_pte_t pte, int psize, int local)
{
unsigned long hash, index, shift, hidx, slot;
DBG_LOW("flush_hash_page(va=%016x)\n", va);
pte_iterate_hashed_subpages(pte, psize, va, index, shift) {
hash = hpt_hash(va, shift);
hidx = __rpte_to_hidx(pte, index);
if (hidx & _PTEIDX_SECONDARY)
hash = ~hash;
slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
slot += hidx & _PTEIDX_GROUP_IX;
DBG_LOW(" sub %d: hash=%x, hidx=%x\n", index, slot, hidx);
ppc_md.hpte_invalidate(slot, va, psize, local);
} pte_iterate_hashed_end();
}
void flush_hash_range(unsigned long number, int local)
{
if (ppc_md.flush_hash_range)
ppc_md.flush_hash_range(number, local);
else {
int i;
struct ppc64_tlb_batch *batch =
&__get_cpu_var(ppc64_tlb_batch);
for (i = 0; i < number; i++)
flush_hash_page(batch->vaddr[i], batch->pte[i],
batch->psize, local);
}
}
static inline void make_bl(unsigned int *insn_addr, void *func)
{
unsigned long funcp = *((unsigned long *)func);
int offset = funcp - (unsigned long)insn_addr;
*insn_addr = (unsigned int)(0x48000001 | (offset & 0x03fffffc));
flush_icache_range((unsigned long)insn_addr, 4+
(unsigned long)insn_addr);
}
/*
* low_hash_fault is called when we the low level hash code failed
* to instert a PTE due to an hypervisor error
*/
void low_hash_fault(struct pt_regs *regs, unsigned long address)
{
if (user_mode(regs)) {
siginfo_t info;
info.si_signo = SIGBUS;
info.si_errno = 0;
info.si_code = BUS_ADRERR;
info.si_addr = (void __user *)address;
force_sig_info(SIGBUS, &info, current);
return;
}
bad_page_fault(regs, address, SIGBUS);
}
void __init htab_finish_init(void)
{
extern unsigned int *htab_call_hpte_insert1;
extern unsigned int *htab_call_hpte_insert2;
extern unsigned int *htab_call_hpte_remove;
extern unsigned int *htab_call_hpte_updatepp;
#ifdef CONFIG_PPC_64K_PAGES
extern unsigned int *ht64_call_hpte_insert1;
extern unsigned int *ht64_call_hpte_insert2;
extern unsigned int *ht64_call_hpte_remove;
extern unsigned int *ht64_call_hpte_updatepp;
make_bl(ht64_call_hpte_insert1, ppc_md.hpte_insert);
make_bl(ht64_call_hpte_insert2, ppc_md.hpte_insert);
make_bl(ht64_call_hpte_remove, ppc_md.hpte_remove);
make_bl(ht64_call_hpte_updatepp, ppc_md.hpte_updatepp);
#endif /* CONFIG_PPC_64K_PAGES */
make_bl(htab_call_hpte_insert1, ppc_md.hpte_insert);
make_bl(htab_call_hpte_insert2, ppc_md.hpte_insert);
make_bl(htab_call_hpte_remove, ppc_md.hpte_remove);
make_bl(htab_call_hpte_updatepp, ppc_md.hpte_updatepp);
}