linux/arch/mips/mm/tlb-r4k.c

538 lines
12 KiB
C
Raw Normal View History

/*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*
* Copyright (C) 1996 David S. Miller (davem@davemloft.net)
* Copyright (C) 1997, 1998, 1999, 2000 Ralf Baechle ralf@gnu.org
* Carsten Langgaard, carstenl@mips.com
* Copyright (C) 2002 MIPS Technologies, Inc. All rights reserved.
*/
#include <linux/cpu_pm.h>
#include <linux/init.h>
#include <linux/sched.h>
#include <linux/smp.h>
#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/module.h>
#include <asm/cpu.h>
#include <asm/cpu-type.h>
#include <asm/bootinfo.h>
#include <asm/mmu_context.h>
#include <asm/pgtable.h>
#include <asm/tlb.h>
#include <asm/tlbmisc.h>
extern void build_tlb_refill_handler(void);
/*
* LOONGSON2/3 has a 4 entry itlb which is a subset of dtlb,
* unfortunately, itlb is not totally transparent to software.
*/
static inline void flush_itlb(void)
{
switch (current_cpu_type()) {
case CPU_LOONGSON2:
case CPU_LOONGSON3:
write_c0_diag(4);
break;
default:
break;
}
}
static inline void flush_itlb_vm(struct vm_area_struct *vma)
{
if (vma->vm_flags & VM_EXEC)
flush_itlb();
}
void local_flush_tlb_all(void)
{
unsigned long flags;
unsigned long old_ctx;
int entry, ftlbhighset;
local_irq_save(flags);
/* Save old context and create impossible VPN2 value */
old_ctx = read_c0_entryhi();
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_stop();
write_c0_entrylo0(0);
write_c0_entrylo1(0);
entry = read_c0_wired();
/* Blast 'em all away. */
if (cpu_has_tlbinv) {
if (current_cpu_data.tlbsizevtlb) {
write_c0_index(0);
mtc0_tlbw_hazard();
tlbinvf(); /* invalidate VTLB */
}
ftlbhighset = current_cpu_data.tlbsizevtlb +
current_cpu_data.tlbsizeftlbsets;
for (entry = current_cpu_data.tlbsizevtlb;
entry < ftlbhighset;
entry++) {
write_c0_index(entry);
mtc0_tlbw_hazard();
tlbinvf(); /* invalidate one FTLB set */
}
} else {
while (entry < current_cpu_data.tlbsize) {
/* Make sure all entries differ. */
write_c0_entryhi(UNIQUE_ENTRYHI(entry));
write_c0_index(entry);
mtc0_tlbw_hazard();
tlb_write_indexed();
entry++;
}
}
tlbw_use_hazard();
write_c0_entryhi(old_ctx);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_start();
flush_itlb();
local_irq_restore(flags);
}
EXPORT_SYMBOL(local_flush_tlb_all);
/* All entries common to a mm share an asid. To effectively flush
these entries, we just bump the asid. */
void local_flush_tlb_mm(struct mm_struct *mm)
{
int cpu;
preempt_disable();
cpu = smp_processor_id();
if (cpu_context(cpu, mm) != 0) {
drop_mmu_context(mm, cpu);
}
preempt_enable();
}
void local_flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
unsigned long end)
{
struct mm_struct *mm = vma->vm_mm;
int cpu = smp_processor_id();
if (cpu_context(cpu, mm) != 0) {
unsigned long size, flags;
local_irq_save(flags);
MIPS: Avoid mcheck by flushing page range in huge_ptep_set_access_flags() Problem: 1) Huge page mapping of anonymous memory is initially invalid. Will be faulted in by copy-on-write mechanism. 2) Userspace attempts store at the end of the huge mapping. 3) TLB Refill exception handler fill TLB with a normal (4K sized) invalid page at the end of the huge mapping virtual address range. 4) Userspace restarted, and re-attempts the store at the end of the huge mapping. 5) Page from #3 is invalid, we get a fault and go to the hugepage fault handler. This tries to map a huge page and calls huge_ptep_set_access_flags() to install the mapping. 6) We just call the generic ptep_set_access_flags() to set up the page tables, but the flush there assumes a normal (4K sized) page and only tries to flush the first part of the huge page virtual address out of the TLB, since the existing entry from step #3 doesn't conflict, nothing is flushed. 7) We attempt to load the mapping into the TLB, but because it conflicts with the entry from step #3, we get a Machine Check exception. The fix: Flush the entire rage covered by the huge page in huge_ptep_set_access_flags(), and remove the optimization in local_flush_tlb_range() so that the flush actually does the correct thing. Signed-off-by: David Daney <david.daney@cavium.com> Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Cc: Hillf Danton <dhillf@gmail.com> Patchwork: https://patchwork.linux-mips.org/patch/4661/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org> (cherry picked from commit dd617f258cc39d36be26afee9912624a2d23112c)
2012-12-03 20:44:26 +00:00
start = round_down(start, PAGE_SIZE << 1);
end = round_up(end, PAGE_SIZE << 1);
size = (end - start) >> (PAGE_SHIFT + 1);
if (size <= (current_cpu_data.tlbsizeftlbsets ?
current_cpu_data.tlbsize / 8 :
current_cpu_data.tlbsize / 2)) {
int oldpid = read_c0_entryhi();
int newpid = cpu_asid(cpu, mm);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_stop();
while (start < end) {
int idx;
write_c0_entryhi(start | newpid);
MIPS: Avoid mcheck by flushing page range in huge_ptep_set_access_flags() Problem: 1) Huge page mapping of anonymous memory is initially invalid. Will be faulted in by copy-on-write mechanism. 2) Userspace attempts store at the end of the huge mapping. 3) TLB Refill exception handler fill TLB with a normal (4K sized) invalid page at the end of the huge mapping virtual address range. 4) Userspace restarted, and re-attempts the store at the end of the huge mapping. 5) Page from #3 is invalid, we get a fault and go to the hugepage fault handler. This tries to map a huge page and calls huge_ptep_set_access_flags() to install the mapping. 6) We just call the generic ptep_set_access_flags() to set up the page tables, but the flush there assumes a normal (4K sized) page and only tries to flush the first part of the huge page virtual address out of the TLB, since the existing entry from step #3 doesn't conflict, nothing is flushed. 7) We attempt to load the mapping into the TLB, but because it conflicts with the entry from step #3, we get a Machine Check exception. The fix: Flush the entire rage covered by the huge page in huge_ptep_set_access_flags(), and remove the optimization in local_flush_tlb_range() so that the flush actually does the correct thing. Signed-off-by: David Daney <david.daney@cavium.com> Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Cc: Hillf Danton <dhillf@gmail.com> Patchwork: https://patchwork.linux-mips.org/patch/4661/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org> (cherry picked from commit dd617f258cc39d36be26afee9912624a2d23112c)
2012-12-03 20:44:26 +00:00
start += (PAGE_SIZE << 1);
mtc0_tlbw_hazard();
tlb_probe();
tlb_probe_hazard();
idx = read_c0_index();
write_c0_entrylo0(0);
write_c0_entrylo1(0);
if (idx < 0)
continue;
/* Make sure all entries differ. */
write_c0_entryhi(UNIQUE_ENTRYHI(idx));
mtc0_tlbw_hazard();
tlb_write_indexed();
}
tlbw_use_hazard();
write_c0_entryhi(oldpid);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_start();
} else {
drop_mmu_context(mm, cpu);
}
flush_itlb();
local_irq_restore(flags);
}
}
void local_flush_tlb_kernel_range(unsigned long start, unsigned long end)
{
unsigned long size, flags;
local_irq_save(flags);
size = (end - start + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
size = (size + 1) >> 1;
if (size <= (current_cpu_data.tlbsizeftlbsets ?
current_cpu_data.tlbsize / 8 :
current_cpu_data.tlbsize / 2)) {
int pid = read_c0_entryhi();
start &= (PAGE_MASK << 1);
end += ((PAGE_SIZE << 1) - 1);
end &= (PAGE_MASK << 1);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_stop();
while (start < end) {
int idx;
write_c0_entryhi(start);
start += (PAGE_SIZE << 1);
mtc0_tlbw_hazard();
tlb_probe();
tlb_probe_hazard();
idx = read_c0_index();
write_c0_entrylo0(0);
write_c0_entrylo1(0);
if (idx < 0)
continue;
/* Make sure all entries differ. */
write_c0_entryhi(UNIQUE_ENTRYHI(idx));
mtc0_tlbw_hazard();
tlb_write_indexed();
}
tlbw_use_hazard();
write_c0_entryhi(pid);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_start();
} else {
local_flush_tlb_all();
}
flush_itlb();
local_irq_restore(flags);
}
void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
int cpu = smp_processor_id();
if (cpu_context(cpu, vma->vm_mm) != 0) {
unsigned long flags;
int oldpid, newpid, idx;
newpid = cpu_asid(cpu, vma->vm_mm);
page &= (PAGE_MASK << 1);
local_irq_save(flags);
oldpid = read_c0_entryhi();
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_stop();
write_c0_entryhi(page | newpid);
mtc0_tlbw_hazard();
tlb_probe();
tlb_probe_hazard();
idx = read_c0_index();
write_c0_entrylo0(0);
write_c0_entrylo1(0);
if (idx < 0)
goto finish;
/* Make sure all entries differ. */
write_c0_entryhi(UNIQUE_ENTRYHI(idx));
mtc0_tlbw_hazard();
tlb_write_indexed();
tlbw_use_hazard();
finish:
write_c0_entryhi(oldpid);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_start();
flush_itlb_vm(vma);
local_irq_restore(flags);
}
}
/*
* This one is only used for pages with the global bit set so we don't care
* much about the ASID.
*/
void local_flush_tlb_one(unsigned long page)
{
unsigned long flags;
int oldpid, idx;
local_irq_save(flags);
oldpid = read_c0_entryhi();
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_stop();
page &= (PAGE_MASK << 1);
write_c0_entryhi(page);
mtc0_tlbw_hazard();
tlb_probe();
tlb_probe_hazard();
idx = read_c0_index();
write_c0_entrylo0(0);
write_c0_entrylo1(0);
if (idx >= 0) {
/* Make sure all entries differ. */
write_c0_entryhi(UNIQUE_ENTRYHI(idx));
mtc0_tlbw_hazard();
tlb_write_indexed();
tlbw_use_hazard();
}
write_c0_entryhi(oldpid);
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_start();
flush_itlb();
local_irq_restore(flags);
}
/*
* We will need multiple versions of update_mmu_cache(), one that just
* updates the TLB with the new pte(s), and another which also checks
* for the R4k "end of page" hardware bug and does the needy.
*/
void __update_tlb(struct vm_area_struct * vma, unsigned long address, pte_t pte)
{
unsigned long flags;
pgd_t *pgdp;
pud_t *pudp;
pmd_t *pmdp;
pte_t *ptep;
int idx, pid;
/*
* Handle debugger faulting in for debugee.
*/
if (current->active_mm != vma->vm_mm)
return;
local_irq_save(flags);
pid = read_c0_entryhi() & ASID_MASK;
address &= (PAGE_MASK << 1);
write_c0_entryhi(address | pid);
pgdp = pgd_offset(vma->vm_mm, address);
mtc0_tlbw_hazard();
tlb_probe();
tlb_probe_hazard();
pudp = pud_offset(pgdp, address);
pmdp = pmd_offset(pudp, address);
idx = read_c0_index();
#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT
/* this could be a huge page */
if (pmd_huge(*pmdp)) {
unsigned long lo;
write_c0_pagemask(PM_HUGE_MASK);
ptep = (pte_t *)pmdp;
lo = pte_to_entrylo(pte_val(*ptep));
write_c0_entrylo0(lo);
write_c0_entrylo1(lo + (HPAGE_SIZE >> 7));
mtc0_tlbw_hazard();
if (idx < 0)
tlb_write_random();
else
tlb_write_indexed();
tlbw_use_hazard();
write_c0_pagemask(PM_DEFAULT_MASK);
} else
#endif
{
ptep = pte_offset_map(pmdp, address);
#if defined(CONFIG_64BIT_PHYS_ADDR) && defined(CONFIG_CPU_MIPS32)
write_c0_entrylo0(ptep->pte_high);
ptep++;
write_c0_entrylo1(ptep->pte_high);
#else
write_c0_entrylo0(pte_to_entrylo(pte_val(*ptep++)));
write_c0_entrylo1(pte_to_entrylo(pte_val(*ptep)));
#endif
mtc0_tlbw_hazard();
if (idx < 0)
tlb_write_random();
else
tlb_write_indexed();
}
tlbw_use_hazard();
flush_itlb_vm(vma);
local_irq_restore(flags);
}
void add_wired_entry(unsigned long entrylo0, unsigned long entrylo1,
unsigned long entryhi, unsigned long pagemask)
{
unsigned long flags;
unsigned long wired;
unsigned long old_pagemask;
unsigned long old_ctx;
local_irq_save(flags);
/* Save old context and create impossible VPN2 value */
old_ctx = read_c0_entryhi();
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_stop();
old_pagemask = read_c0_pagemask();
wired = read_c0_wired();
write_c0_wired(wired + 1);
write_c0_index(wired);
tlbw_use_hazard(); /* What is the hazard here? */
write_c0_pagemask(pagemask);
write_c0_entryhi(entryhi);
write_c0_entrylo0(entrylo0);
write_c0_entrylo1(entrylo1);
mtc0_tlbw_hazard();
tlb_write_indexed();
tlbw_use_hazard();
write_c0_entryhi(old_ctx);
tlbw_use_hazard(); /* What is the hazard here? */
MIPS: mm: Use the Hardware Page Table Walker if the core supports it The Hardware Page Table Walker aims to speed up TLB refill exceptions by handling them in the hardware level instead of having a software TLB refill handler. However, a TLB refill exception can still be thrown in certain cases such as, synchronus exceptions, or address translation or memory errors during the HTW operation. As a result of which, HTW must not be considered a complete replacement for the TLB refill software handler, but rather a fast-path for it. For HTW to work, the PWBase register must contain the task's page global directory address so the HTW will kick in on TLB refill exceptions. Due to HTW being a separate engine embedded deep in the CPU pipeline, we need to restart the HTW everytime a PTE changes to avoid HTW fetching a old entry from the page tables. It's also necessary to restart the HTW on context switches to prevent it from fetching a page from the previous process. Finally, since HTW is using the entryhi register to write the translations to the TLB, it's necessary to stop the HTW whenever the entryhi changes (eg for tlb probe perations) and enable it back afterwards. == Performance == The following trivial test was used to measure the performance of the HTW. Using the same root filesystem, the following command was used to measure the number of tlb refill handler executions with and without (using 'nohtw' kernel parameter) HTW support. The kernel was modified to use a scratch register as a counter for the TLB refill exceptions. find /usr -type f -exec ls -lh {} \; HTW Enabled: TLB refill exceptions: 12306 HTW Disabled: TLB refill exceptions: 17805 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Markos Chandras <markos.chandras@imgtec.com> Patchwork: https://patchwork.linux-mips.org/patch/7336/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2014-07-14 11:47:09 +00:00
htw_start();
write_c0_pagemask(old_pagemask);
local_flush_tlb_all();
local_irq_restore(flags);
}
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
int __init has_transparent_hugepage(void)
{
unsigned int mask;
unsigned long flags;
local_irq_save(flags);
write_c0_pagemask(PM_HUGE_MASK);
back_to_back_c0_hazard();
mask = read_c0_pagemask();
write_c0_pagemask(PM_DEFAULT_MASK);
local_irq_restore(flags);
return mask == PM_HUGE_MASK;
}
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
/*
* Used for loading TLB entries before trap_init() has started, when we
* don't actually want to add a wired entry which remains throughout the
* lifetime of the system
*/
int temp_tlb_entry __cpuinitdata;
__init int add_temporary_entry(unsigned long entrylo0, unsigned long entrylo1,
unsigned long entryhi, unsigned long pagemask)
{
int ret = 0;
unsigned long flags;
unsigned long wired;
unsigned long old_pagemask;
unsigned long old_ctx;
local_irq_save(flags);
/* Save old context and create impossible VPN2 value */
old_ctx = read_c0_entryhi();
old_pagemask = read_c0_pagemask();
wired = read_c0_wired();
if (--temp_tlb_entry < wired) {
printk(KERN_WARNING
"No TLB space left for add_temporary_entry\n");
ret = -ENOSPC;
goto out;
}
write_c0_index(temp_tlb_entry);
write_c0_pagemask(pagemask);
write_c0_entryhi(entryhi);
write_c0_entrylo0(entrylo0);
write_c0_entrylo1(entrylo1);
mtc0_tlbw_hazard();
tlb_write_indexed();
tlbw_use_hazard();
write_c0_entryhi(old_ctx);
write_c0_pagemask(old_pagemask);
out:
local_irq_restore(flags);
return ret;
}
MIPS: Delete __cpuinit/__CPUINIT usage from MIPS code commit 3747069b25e419f6b51395f48127e9812abc3596 upstream. The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. Note that some harmless section mismatch warnings may result, since notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c) and are flagged as __cpuinit -- so if we remove the __cpuinit from the arch specific callers, we will also get section mismatch warnings. As an intermediate step, we intend to turn the linux/init.h cpuinit related content into no-ops as early as possible, since that will get rid of these warnings. In any case, they are temporary and harmless. Here, we remove all the MIPS __cpuinit from C code and __CPUINIT from asm files. MIPS is interesting in this respect, because there are also uasm users hiding behind their own renamed versions of the __cpuinit macros. [1] https://lkml.org/lkml/2013/5/20/589 [ralf@linux-mips.org: Folded in Paul's followup fix.] Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/5494/ Patchwork: https://patchwork.linux-mips.org/patch/5495/ Patchwork: https://patchwork.linux-mips.org/patch/5509/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2013-06-18 13:38:59 +00:00
static int ntlb;
static int __init set_ntlb(char *str)
{
get_option(&str, &ntlb);
return 1;
}
__setup("ntlb=", set_ntlb);
/*
* Configure TLB (for init or after a CPU has been powered off).
*/
static void r4k_tlb_configure(void)
{
/*
* You should never change this register:
* - On R4600 1.7 the tlbp never hits for pages smaller than
* the value in the c0_pagemask register.
* - The entire mm handling assumes the c0_pagemask register to
* be set to fixed-size pages.
*/
write_c0_pagemask(PM_DEFAULT_MASK);
write_c0_wired(0);
if (current_cpu_type() == CPU_R10000 ||
current_cpu_type() == CPU_R12000 ||
current_cpu_type() == CPU_R14000)
write_c0_framemask(0);
if (cpu_has_rixi) {
/*
* Enable the no read, no exec bits, and enable large virtual
* address.
*/
u32 pg = PG_RIE | PG_XIE;
#ifdef CONFIG_64BIT
pg |= PG_ELPA;
#endif
write_c0_pagegrain(pg);
}
temp_tlb_entry = current_cpu_data.tlbsize - 1;
/* From this point on the ARC firmware is dead. */
local_flush_tlb_all();
/* Did I tell you that ARC SUCKS? */
}
void tlb_init(void)
{
r4k_tlb_configure();
if (ntlb) {
if (ntlb > 1 && ntlb <= current_cpu_data.tlbsize) {
int wired = current_cpu_data.tlbsize - ntlb;
write_c0_wired(wired);
write_c0_index(wired-1);
printk("Restricting TLB to %d entries\n", ntlb);
} else
printk("Ignoring invalid argument ntlb=%d\n", ntlb);
}
build_tlb_refill_handler();
}
static int r4k_tlb_pm_notifier(struct notifier_block *self, unsigned long cmd,
void *v)
{
switch (cmd) {
case CPU_PM_ENTER_FAILED:
case CPU_PM_EXIT:
r4k_tlb_configure();
break;
}
return NOTIFY_OK;
}
static struct notifier_block r4k_tlb_pm_notifier_block = {
.notifier_call = r4k_tlb_pm_notifier,
};
static int __init r4k_tlb_init_pm(void)
{
return cpu_pm_register_notifier(&r4k_tlb_pm_notifier_block);
}
arch_initcall(r4k_tlb_init_pm);