[PATCH 01/19 v2] avr32/mm/fault.c: Port OOM changes to do_page_fault

[PATCH 01/19 v2] avr32/mm/fault.c: Port OOM changes to do_page_fault

[Kautuk Consul]

Commit d065bd810b6deb67d4897a14bfe21f8eb526ba99
(mm: retry page fault when blocking on disk transfer) and
commit 37b23e0525d393d48a7d59f870b3bc061a30ccdb
(x86,mm: make pagefault killable)

The above commits introduced changes into the x86 pagefault handler
for making the page fault handler retryable as well as killable.

These changes reduce the mmap_sem hold time, which is crucial
during OOM killer invocation.

Port these changes to AVR32.

Signed-off-by: Mohd. Faris <mohdfarisq2010@gmail.com>
Signed-off-by: Kautuk Consul <consul.kautuk@gmail.com>
---
 arch/avr32/mm/fault.c |   30 +++++++++++++++++++++++++-----
 1 files changed, 25 insertions(+), 5 deletions(-)

diff --git a/arch/avr32/mm/fault.c b/arch/avr32/mm/fault.c
index f7040a1..0362d03 100644
--- a/arch/avr32/mm/fault.c
+++ b/arch/avr32/mm/fault.c
@@ -65,6 +65,7 @@ asmlinkage void do_page_fault(unsigned long ecr, struct pt_regs *regs)
 	long signr;
 	int code;
 	int fault;
+	unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
 
 	if (notify_page_fault(regs, ecr))
 		return;
@@ -86,6 +87,7 @@ asmlinkage void do_page_fault(unsigned long ecr, struct pt_regs *regs)
 
 	local_irq_enable();
 
+retry:
 	down_read(&mm->mmap_sem);
 
 	vma = find_vma(mm, address);
@@ -122,6 +124,7 @@ good_area:
 		if (!(vma->vm_flags & VM_WRITE))
 			goto bad_area;
 		writeaccess = 1;
+		flags |= FAULT_FLAG_WRITE;
 		break;
 	default:
 		panic("Unhandled case %lu in do_page_fault!", ecr);
@@ -132,7 +135,11 @@ good_area:
 	 * sure we exit gracefully rather than endlessly redo the
 	 * fault.
 	 */
-	fault = handle_mm_fault(mm, vma, address, writeaccess ? FAULT_FLAG_WRITE : 0);
+	fault = handle_mm_fault(mm, vma, address, flags);
+
+	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
+		return;
+
 	if (unlikely(fault & VM_FAULT_ERROR)) {
 		if (fault & VM_FAULT_OOM)
 			goto out_of_memory;
@@ -140,10 +147,23 @@ good_area:
 			goto do_sigbus;
 		BUG();
 	}
-	if (fault & VM_FAULT_MAJOR)
-		tsk->maj_flt++;
-	else
-		tsk->min_flt++;
+
+	if (flags & FAULT_FLAG_ALLOW_RETRY) {
+		if (fault & VM_FAULT_MAJOR)
+			tsk->maj_flt++;
+		else
+			tsk->min_flt++;
+		if (fault & VM_FAULT_RETRY) {
+			flags &= ~FAULT_FLAG_ALLOW_RETRY;
+
+			 /* No need to up_read(&mm->mmap_sem) as we would
+			 * have already released it in __lock_page_or_retry
+			 * in mm/filemap.c.
+			 */
+
+			goto retry;
+		}
+	}
 
 	up_read(&mm->mmap_sem);
 	return;
-- 
1.7.5.4

Re: [PATCH 01/19 v2] avr32/mm/fault.c: Port OOM changes to do_page_fault

[Håvard Skinnemoen]

On Sat, Mar 31, 2012 at 4:58 AM, Kautuk Consul <consul.kautuk@gmail.com> wrote:
> Commit d065bd810b6deb67d4897a14bfe21f8eb526ba99
> (mm: retry page fault when blocking on disk transfer) and
> commit 37b23e0525d393d48a7d59f870b3bc061a30ccdb
> (x86,mm: make pagefault killable)
>
> The above commits introduced changes into the x86 pagefault handler
> for making the page fault handler retryable as well as killable.
>
> These changes reduce the mmap_sem hold time, which is crucial
> during OOM killer invocation.
>
> Port these changes to AVR32.
>
> Signed-off-by: Mohd. Faris <mohdfarisq2010@gmail.com>
> Signed-off-by: Kautuk Consul <consul.kautuk@gmail.com>

Looks very good, thanks.

> @@ -122,6 +124,7 @@ good_area:
>                if (!(vma->vm_flags & VM_WRITE))
>                        goto bad_area;
>                writeaccess = 1;
> +               flags |= FAULT_FLAG_WRITE;

Could you remove the 'writeaccess' variable as well? It looks unused
after applying your patch.

Ug...the avr32 arch is in a terrible state. Someone appears to have
removed the atmel_nand driver...

Havard

Re: [PATCH 01/19 v2] avr32/mm/fault.c: Port OOM changes to do_page_fault

[Håvard Skinnemoen]

On Sun, Apr 1, 2012 at 7:08 PM, Håvard Skinnemoen <hskinnemoen@gmail.com> wrote:
> Ug...the avr32 arch is in a terrible state. Someone appears to have
> removed the atmel_nand driver...

Never mind; Paul Gortmaker already fixed it :)

Your patch doesn't cause anything to blow up in any obvious way, so it
should be good to go as far as I can tell (although it would be great
if you could do that tiny cleanup I mentioned).

Havard

Re: [PATCH 01/19 v2] avr32/mm/fault.c: Port OOM changes to do_page_fault

[Kautuk Consul]

>
> Looks very good, thanks.
>
>> @@ -122,6 +124,7 @@ good_area:
>>                if (!(vma->vm_flags & VM_WRITE))
>>                        goto bad_area;
>>                writeaccess = 1;
>> +               flags |= FAULT_FLAG_WRITE;
>
> Could you remove the 'writeaccess' variable as well? It looks unused
> after applying your patch.

Sure, Ill do that and send v2.1 of this patch to you.

>
> Ug...the avr32 arch is in a terrible state. Someone appears to have
> removed the atmel_nand driver...

I don't know anything about that. :)

>
> Havard

Re: [PATCH 01/19 v2] avr32/mm/fault.c: Port OOM changes to do_page_fault

[Kautuk Consul]

On Mon, Apr 2, 2012 at 8:28 AM, Håvard Skinnemoen <hskinnemoen@gmail.com> wrote:
> On Sun, Apr 1, 2012 at 7:08 PM, Håvard Skinnemoen <hskinnemoen@gmail.com> wrote:
>> Ug...the avr32 arch is in a terrible state. Someone appears to have
>> removed the atmel_nand driver...
>
> Never mind; Paul Gortmaker already fixed it :)
>

Great.

> Your patch doesn't cause anything to blow up in any obvious way, so it
> should be good to go as far as I can tell (although it would be great
> if you could do that tiny cleanup I mentioned).

Ok, Ill make that change and resend v2.1 of this to the mailing list.
Can you please ACK that if you find it to be okay ?
Thanks !

>
> Havard