2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* 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
|
2005-11-07 08:06:55 +08:00
|
|
|
#undef DEBUG_LOW
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#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/ppcdebug.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
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
#ifdef DEBUG_LOW
|
|
|
|
#define DBG_LOW(fmt...) udbg_printf(fmt)
|
|
|
|
#else
|
|
|
|
#define DBG_LOW(fmt...)
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define KB (1024)
|
|
|
|
#define MB (1024*KB)
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* 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 */
|
|
|
|
|
2005-07-13 16:11:42 +08:00
|
|
|
hpte_t *htab_address;
|
|
|
|
unsigned long htab_hash_mask;
|
2005-10-10 19:58:35 +08:00
|
|
|
unsigned long _SDR1;
|
2005-11-07 08:06:55 +08:00
|
|
|
struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT];
|
|
|
|
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
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* There are definitions of page sizes arrays to be used when none
|
|
|
|
* is provided by the firmware.
|
|
|
|
*/
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* 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)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2005-11-07 08:06:55 +08:00
|
|
|
unsigned long vaddr, paddr;
|
|
|
|
unsigned int step, shift;
|
2005-04-17 06:20:36 +08:00
|
|
|
unsigned long tmp_mode;
|
2005-11-07 08:06:55 +08:00
|
|
|
int ret = 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
shift = mmu_psize_defs[psize].shift;
|
|
|
|
step = 1 << shift;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
for (vaddr = vstart, paddr = pstart; vaddr < vend;
|
|
|
|
vaddr += step, paddr += step) {
|
2005-04-17 06:20:36 +08:00
|
|
|
unsigned long vpn, hash, hpteg;
|
2005-11-07 08:06:55 +08:00
|
|
|
unsigned long vsid = get_kernel_vsid(vaddr);
|
|
|
|
unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
vpn = va >> shift;
|
2005-04-17 06:20:36 +08:00
|
|
|
tmp_mode = mode;
|
|
|
|
|
|
|
|
/* Make non-kernel text non-executable */
|
2005-11-07 08:06:55 +08:00
|
|
|
if (!in_kernel_text(vaddr))
|
|
|
|
tmp_mode = mode | HPTE_R_N;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
hash = hpt_hash(va, shift);
|
2005-04-17 06:20:36 +08:00
|
|
|
hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP);
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* 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.
|
|
|
|
*/
|
2005-09-23 12:47:58 +08:00
|
|
|
#ifdef CONFIG_PPC_ISERIES
|
2005-11-07 08:06:55 +08:00
|
|
|
if (systemcfg->platform == PLATFORM_ISERIES_LPAR)
|
|
|
|
ret = iSeries_hpte_insert(hpteg, va,
|
|
|
|
virt_to_abs(paddr),
|
|
|
|
tmp_mode,
|
|
|
|
HPTE_V_BOLTED,
|
|
|
|
psize);
|
2005-09-23 12:47:58 +08:00
|
|
|
else
|
|
|
|
#endif
|
2005-04-17 06:20:36 +08:00
|
|
|
#ifdef CONFIG_PPC_PSERIES
|
|
|
|
if (systemcfg->platform & PLATFORM_LPAR)
|
|
|
|
ret = pSeries_lpar_hpte_insert(hpteg, va,
|
2005-11-07 08:06:55 +08:00
|
|
|
virt_to_abs(paddr),
|
|
|
|
tmp_mode,
|
|
|
|
HPTE_V_BOLTED,
|
|
|
|
psize);
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
2005-09-23 12:47:58 +08:00
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_PPC_MULTIPLATFORM
|
2005-04-17 06:20:36 +08:00
|
|
|
ret = native_hpte_insert(hpteg, va,
|
2005-11-07 08:06:55 +08:00
|
|
|
virt_to_abs(paddr),
|
|
|
|
tmp_mode, HPTE_V_BOLTED,
|
|
|
|
psize);
|
2005-09-23 12:47:58 +08:00
|
|
|
#endif
|
2005-11-07 08:06:55 +08:00
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return ret < 0 ? ret : 0;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
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);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2005-11-07 08:06:55 +08:00
|
|
|
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 ((systemcfg->platform != 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;
|
|
|
|
else if (mmu_psize_defs[MMU_PAGE_1M].shift)
|
|
|
|
mmu_huge_psize = MMU_PAGE_1M;
|
|
|
|
|
|
|
|
/* Calculate HPAGE_SHIFT and sanity check it */
|
|
|
|
if (mmu_psize_defs[mmu_huge_psize].shift > 16 &&
|
|
|
|
mmu_psize_defs[mmu_huge_psize].shift < 28)
|
|
|
|
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;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2005-11-07 08:06:55 +08:00
|
|
|
return 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
static unsigned long __init htab_get_table_size(void)
|
2005-10-12 14:58:53 +08:00
|
|
|
{
|
|
|
|
unsigned long rnd_mem_size, pteg_count;
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* 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
|
2005-10-12 14:58:53 +08:00
|
|
|
*/
|
2005-11-07 08:06:55 +08:00
|
|
|
if (ppc64_pft_size == 0)
|
|
|
|
of_scan_flat_dt(htab_dt_scan_pftsize, NULL);
|
2005-10-12 14:58:53 +08:00
|
|
|
if (ppc64_pft_size)
|
|
|
|
return 1UL << ppc64_pft_size;
|
|
|
|
|
|
|
|
/* round mem_size up to next power of 2 */
|
|
|
|
rnd_mem_size = 1UL << __ilog2(systemcfg->physicalMemorySize);
|
|
|
|
if (rnd_mem_size < systemcfg->physicalMemorySize)
|
|
|
|
rnd_mem_size <<= 1;
|
|
|
|
|
|
|
|
/* # pages / 2 */
|
|
|
|
pteg_count = max(rnd_mem_size >> (12 + 1), 1UL << 11);
|
|
|
|
|
|
|
|
return pteg_count << 7;
|
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
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;
|
2005-11-07 08:06:55 +08:00
|
|
|
int i;
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
extern unsigned long tce_alloc_start, tce_alloc_end;
|
|
|
|
|
|
|
|
DBG(" -> htab_initialize()\n");
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* Initialize page sizes */
|
|
|
|
htab_init_page_sizes();
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Calculate the required size of the htab. We want the number of
|
|
|
|
* PTEGs to equal one half the number of real pages.
|
|
|
|
*/
|
2005-11-07 08:06:55 +08:00
|
|
|
htab_size_bytes = htab_get_table_size();
|
2005-04-17 06:20:36 +08:00
|
|
|
pteg_count = htab_size_bytes >> 7;
|
|
|
|
|
|
|
|
/* For debug, make the HTAB 1/8 as big as it normally would be. */
|
|
|
|
ifppcdebug(PPCDBG_HTABSIZE) {
|
|
|
|
pteg_count >>= 3;
|
|
|
|
htab_size_bytes = pteg_count << 7;
|
|
|
|
}
|
|
|
|
|
|
|
|
htab_hash_mask = pteg_count - 1;
|
|
|
|
|
|
|
|
if (systemcfg->platform & PLATFORM_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);
|
2005-11-07 08:06:55 +08:00
|
|
|
BUG_ON(table == 0);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
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);
|
|
|
|
}
|
|
|
|
|
2005-06-22 08:15:55 +08:00
|
|
|
mode_rw = _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_COHERENT | PP_RWXX;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/* 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++) {
|
2005-08-03 18:21:26 +08:00
|
|
|
base = lmb.memory.region[i].base + KERNELBASE;
|
2005-04-17 06:20:36 +08:00
|
|
|
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
|
2005-11-07 08:06:55 +08:00
|
|
|
* 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.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
DBG("DART base: %lx\n", dart_tablebase);
|
|
|
|
|
|
|
|
if (dart_tablebase != 0 && dart_tablebase >= base
|
|
|
|
&& dart_tablebase < (base + size)) {
|
|
|
|
if (base != dart_tablebase)
|
2005-11-07 08:06:55 +08:00
|
|
|
BUG_ON(htab_bolt_mapping(base, dart_tablebase,
|
|
|
|
base, mode_rw,
|
|
|
|
mmu_linear_psize));
|
2005-04-17 06:20:36 +08:00
|
|
|
if ((base + size) > (dart_tablebase + 16*MB))
|
2005-11-07 08:06:55 +08:00
|
|
|
BUG_ON(htab_bolt_mapping(dart_tablebase+16*MB,
|
|
|
|
base + size,
|
|
|
|
dart_tablebase+16*MB,
|
|
|
|
mode_rw,
|
|
|
|
mmu_linear_psize));
|
2005-04-17 06:20:36 +08:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_U3_DART */
|
2005-11-07 08:06:55 +08:00
|
|
|
BUG_ON(htab_bolt_mapping(base, base + size, base,
|
|
|
|
mode_rw, mmu_linear_psize));
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* 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;
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
BUG_ON(htab_bolt_mapping(tce_alloc_start, tce_alloc_end,
|
|
|
|
tce_alloc_start, mode_rw,
|
|
|
|
mmu_linear_psize));
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
DBG(" <- htab_initialize()\n");
|
|
|
|
}
|
|
|
|
#undef KB
|
|
|
|
#undef MB
|
|
|
|
|
|
|
|
/*
|
|
|
|
* 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;
|
|
|
|
|
|
|
|
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
|
2005-11-07 08:06:55 +08:00
|
|
|
pp |= HPTE_R_N;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
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;
|
2005-11-07 08:06:55 +08:00
|
|
|
int rc, user_region = 0, local = 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
DBG_LOW("hash_page(ea=%016lx, access=%lx, trap=%lx\n",
|
|
|
|
ea, access, trap);
|
2005-05-06 07:15:13 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
if ((ea & ~REGION_MASK) >= PGTABLE_RANGE) {
|
|
|
|
DBG_LOW(" out of pgtable range !\n");
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Get region & vsid */
|
2005-04-17 06:20:36 +08:00
|
|
|
switch (REGION_ID(ea)) {
|
|
|
|
case USER_REGION_ID:
|
|
|
|
user_region = 1;
|
|
|
|
mm = current->mm;
|
2005-11-07 08:06:55 +08:00
|
|
|
if (! mm) {
|
|
|
|
DBG_LOW(" user region with no mm !\n");
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
2005-11-07 08:06:55 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
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;
|
|
|
|
}
|
2005-11-07 08:06:55 +08:00
|
|
|
DBG_LOW(" mm=%p, mm->pgdir=%p, vsid=%016lx\n", mm, mm->pgd, vsid);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* Get pgdir */
|
2005-04-17 06:20:36 +08:00
|
|
|
pgdir = mm->pgd;
|
|
|
|
if (pgdir == NULL)
|
|
|
|
return 1;
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* Check CPU locality */
|
2005-04-17 06:20:36 +08:00
|
|
|
tmp = cpumask_of_cpu(smp_processor_id());
|
|
|
|
if (user_region && cpus_equal(mm->cpu_vm_mask, tmp))
|
|
|
|
local = 1;
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* 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);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* 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;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* 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;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
void hash_preload(struct mm_struct *mm, unsigned long ea,
|
|
|
|
unsigned long access, unsigned long trap)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2005-11-07 08:06:55 +08:00
|
|
|
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);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
/* 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);
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
2005-11-07 08:06:55 +08:00
|
|
|
__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();
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-09-20 11:52:50 +08:00
|
|
|
void flush_hash_range(unsigned long number, int local)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2005-11-07 08:06:55 +08:00
|
|
|
if (ppc_md.flush_hash_range)
|
2005-09-20 11:52:50 +08:00
|
|
|
ppc_md.flush_hash_range(number, local);
|
2005-11-07 08:06:55 +08:00
|
|
|
else {
|
2005-04-17 06:20:36 +08:00
|
|
|
int i;
|
2005-09-20 11:52:50 +08:00
|
|
|
struct ppc64_tlb_batch *batch =
|
|
|
|
&__get_cpu_var(ppc64_tlb_batch);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
for (i = 0; i < number; i++)
|
2005-11-07 08:06:55 +08:00
|
|
|
flush_hash_page(batch->vaddr[i], batch->pte[i],
|
|
|
|
batch->psize, local);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
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;
|
|
|
|
|
2005-11-07 08:06:55 +08:00
|
|
|
#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 */
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
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);
|
|
|
|
}
|