* [PATCH] powerpc/mm: Only read faulting instruction when necessary in do_page_fault()
@ 2017-04-24 8:28 Christophe Leroy
2017-04-24 9:15 ` Michael Ellerman
0 siblings, 1 reply; 3+ messages in thread
From: Christophe Leroy @ 2017-04-24 8:28 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Paul Mackerras, Michael Ellerman, Scott Wood
Cc: linux-kernel, linuxppc-dev
Commit a7a9dcd882a67 ("powerpc: Avoid taking a data miss on every
userspace instruction miss") has shown that limiting the read of
faulting instruction to likely cases improves performance.
This patch goes further into this direction by limiting the read
of the faulting instruction to the only cases where it is definitly
needed.
On an MPC885, with the same benchmark app as in the commit referred
above, we see a reduction of 4000 dTLB misses (approx 3%):
Before the patch:
Performance counter stats for './fault 500' (10 runs):
720495838 cpu-cycles ( +- 0.04% )
141769 dTLB-load-misses ( +- 0.02% )
52722 iTLB-load-misses ( +- 0.01% )
19611 faults ( +- 0.02% )
5.750535176 seconds time elapsed ( +- 0.16% )
With the patch:
Performance counter stats for './fault 500' (10 runs):
717669123 cpu-cycles ( +- 0.02% )
137344 dTLB-load-misses ( +- 0.03% )
52731 iTLB-load-misses ( +- 0.01% )
19614 faults ( +- 0.03% )
5.728423115 seconds time elapsed ( +- 0.14% )
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
---
This patch applies after the serie "powerpc/mm: some cleanup of do_page_fault()"
arch/powerpc/mm/fault.c | 26 ++++++++++++++++----------
1 file changed, 16 insertions(+), 10 deletions(-)
diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
index 400f2d0d42f8..3d506589236c 100644
--- a/arch/powerpc/mm/fault.c
+++ b/arch/powerpc/mm/fault.c
@@ -280,14 +280,6 @@ int do_page_fault(struct pt_regs *regs, unsigned long address,
perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
- /*
- * We want to do this outside mmap_sem, because reading code around nip
- * can result in fault, which will cause a deadlock when called with
- * mmap_sem held
- */
- if (is_write && is_user)
- __get_user(inst, (unsigned int __user *)regs->nip);
-
if (is_user)
flags |= FAULT_FLAG_USER;
@@ -356,8 +348,22 @@ int do_page_fault(struct pt_regs *regs, unsigned long address,
* between the last mapped region and the stack will
* expand the stack rather than segfaulting.
*/
- if (address + 2048 < uregs->gpr[1] && !store_updates_sp(inst))
- goto bad_area;
+ if (address + 2048 < uregs->gpr[1]) {
+ if (!inst) {
+ /*
+ * We want to do this outside mmap_sem, because
+ * reading code around nip can result in fault,
+ * which will cause a deadlock when called with
+ * mmap_sem held
+ */
+ up_read(&mm->mmap_sem);
+ __get_user(inst,
+ (unsigned int __user *)regs->nip);
+ if (!store_updates_sp(inst))
+ goto bad_area_nosemaphore;
+ goto retry;
+ }
+ }
}
if (expand_stack(vma, address))
goto bad_area;
--
2.12.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] powerpc/mm: Only read faulting instruction when necessary in do_page_fault()
2017-04-24 8:28 [PATCH] powerpc/mm: Only read faulting instruction when necessary in do_page_fault() Christophe Leroy
@ 2017-04-24 9:15 ` Michael Ellerman
2017-04-24 10:26 ` Christophe LEROY
0 siblings, 1 reply; 3+ messages in thread
From: Michael Ellerman @ 2017-04-24 9:15 UTC (permalink / raw)
To: Christophe Leroy, Benjamin Herrenschmidt, Paul Mackerras, Scott Wood
Cc: linux-kernel, linuxppc-dev
Christophe Leroy <christophe.leroy@c-s.fr> writes:
> diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
> index 400f2d0d42f8..3d506589236c 100644
> --- a/arch/powerpc/mm/fault.c
> +++ b/arch/powerpc/mm/fault.c
> @@ -356,8 +348,22 @@ int do_page_fault(struct pt_regs *regs, unsigned long address,
> * between the last mapped region and the stack will
> * expand the stack rather than segfaulting.
> */
> - if (address + 2048 < uregs->gpr[1] && !store_updates_sp(inst))
> - goto bad_area;
> + if (address + 2048 < uregs->gpr[1]) {
> + if (!inst) {
That looks like it could lead to an infinite loop, if the instruction
that caused the fault has been overwritten with NULL by another thread.
Or did I miss something? (I only read the diff)
> + /*
> + * We want to do this outside mmap_sem, because
> + * reading code around nip can result in fault,
> + * which will cause a deadlock when called with
> + * mmap_sem held
> + */
> + up_read(&mm->mmap_sem);
> + __get_user(inst,
> + (unsigned int __user *)regs->nip);
> + if (!store_updates_sp(inst))
> + goto bad_area_nosemaphore;
> + goto retry;
> + }
cheers
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] powerpc/mm: Only read faulting instruction when necessary in do_page_fault()
2017-04-24 9:15 ` Michael Ellerman
@ 2017-04-24 10:26 ` Christophe LEROY
0 siblings, 0 replies; 3+ messages in thread
From: Christophe LEROY @ 2017-04-24 10:26 UTC (permalink / raw)
To: Michael Ellerman, Benjamin Herrenschmidt, Paul Mackerras, Scott Wood
Cc: linux-kernel, linuxppc-dev
Le 24/04/2017 à 11:15, Michael Ellerman a écrit :
> Christophe Leroy <christophe.leroy@c-s.fr> writes:
>
>> diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
>> index 400f2d0d42f8..3d506589236c 100644
>> --- a/arch/powerpc/mm/fault.c
>> +++ b/arch/powerpc/mm/fault.c
>> @@ -356,8 +348,22 @@ int do_page_fault(struct pt_regs *regs, unsigned long address,
>> * between the last mapped region and the stack will
>> * expand the stack rather than segfaulting.
>> */
>> - if (address + 2048 < uregs->gpr[1] && !store_updates_sp(inst))
>> - goto bad_area;
>> + if (address + 2048 < uregs->gpr[1]) {
>> + if (!inst) {
>
> That looks like it could lead to an infinite loop, if the instruction
> that caused the fault has been overwritten with NULL by another thread.
If the inst is NULL, store_updates_sp(inst) returns false so it jumps to
bad_area_nosemaphore
Christophe
>
> Or did I miss something? (I only read the diff)
>
>> + /*
>> + * We want to do this outside mmap_sem, because
>> + * reading code around nip can result in fault,
>> + * which will cause a deadlock when called with
>> + * mmap_sem held
>> + */
>> + up_read(&mm->mmap_sem);
>> + __get_user(inst,
>> + (unsigned int __user *)regs->nip);
>> + if (!store_updates_sp(inst))
>> + goto bad_area_nosemaphore;
>> + goto retry;
>> + }
>
>
> cheers
>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2017-04-24 10:26 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-04-24 8:28 [PATCH] powerpc/mm: Only read faulting instruction when necessary in do_page_fault() Christophe Leroy
2017-04-24 9:15 ` Michael Ellerman
2017-04-24 10:26 ` Christophe LEROY
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.