On Tue, Nov 18, 2014 at 12:18:43PM -0500, Sasha Levin wrote:
> On 11/18/2014 11:56 AM, Aneesh Kumar K.V wrote:
> >>> 4. Similarly, does the kernel boot properly without without patches?
> >> >
> >> > Yes, the kernel works fine without the patches both with and without fake
> >> > numa.
> >
> >
On Tue, Nov 18, 2014 at 12:18:43PM -0500, Sasha Levin wrote:
On 11/18/2014 11:56 AM, Aneesh Kumar K.V wrote:
4. Similarly, does the kernel boot properly without without patches?
Yes, the kernel works fine without the patches both with and without fake
numa.
Hmm that is
On 11/18/2014 11:56 AM, Aneesh Kumar K.V wrote:
>>> 4. Similarly, does the kernel boot properly without without patches?
>> >
>> > Yes, the kernel works fine without the patches both with and without fake
>> > numa.
>
> Hmm that is interesting. I am not sure how writeback_fid can be
> related. We
On Tue, Nov 18, 2014 at 10:26:41PM +0530, Aneesh Kumar K.V wrote:
> Sasha Levin writes:
>
> > On 11/18/2014 10:42 AM, Mel Gorman wrote:
> >> 1. I'm assuming this is a KVM setup but can you confirm?
> >
> > Yes.
> >
> >> 2. Are you using numa=fake=N?
> >
> > Yes. numa=fake=24, which is probably
On Tue, Nov 18, 2014 at 10:03:30PM +0530, Aneesh Kumar K.V wrote:
> > diff --git a/arch/powerpc/mm/copro_fault.c b/arch/powerpc/mm/copro_fault.c
> > index 5a236f0..46152aa 100644
> > --- a/arch/powerpc/mm/copro_fault.c
> > +++ b/arch/powerpc/mm/copro_fault.c
> > @@ -64,7 +64,12 @@ int
Sasha Levin writes:
> On 11/18/2014 10:42 AM, Mel Gorman wrote:
>> 1. I'm assuming this is a KVM setup but can you confirm?
>
> Yes.
>
>> 2. Are you using numa=fake=N?
>
> Yes. numa=fake=24, which is probably way more nodes on any physical machine
> than the new code was tested on?
>
>> 3. If
On 11/18/2014 10:42 AM, Mel Gorman wrote:
> 1. I'm assuming this is a KVM setup but can you confirm?
Yes.
> 2. Are you using numa=fake=N?
Yes. numa=fake=24, which is probably way more nodes on any physical machine
than the new code was tested on?
> 3. If you are using fake NUMA, what happens
Mel Gorman writes:
> On Mon, Nov 17, 2014 at 01:56:19PM +0530, Aneesh Kumar K.V wrote:
>> Mel Gorman writes:
>>
>> > This is follow up from the "pipe/page fault oddness" thread.
>> >
>> > Automatic NUMA balancing depends on being able to protect PTEs to trap a
>> > fault and gather reference
On Mon, Nov 17, 2014 at 01:56:19PM +0530, Aneesh Kumar K.V wrote:
> Mel Gorman writes:
>
> > This is follow up from the "pipe/page fault oddness" thread.
> >
> > Automatic NUMA balancing depends on being able to protect PTEs to trap a
> > fault and gather reference locality information. Very
On Fri, Nov 14, 2014 at 10:29:41PM -0500, Sasha Levin wrote:
> On 11/14/2014 08:32 AM, Mel Gorman wrote:> This is follow up from the
> "pipe/page fault oddness" thread.
>
> Hi Mel,
>
> Applying this patch series I've started seeing the following straight away:
>
> [ 367.547848]
On Fri, Nov 14, 2014 at 10:29:41PM -0500, Sasha Levin wrote:
On 11/14/2014 08:32 AM, Mel Gorman wrote: This is follow up from the
pipe/page fault oddness thread.
Hi Mel,
Applying this patch series I've started seeing the following straight away:
[ 367.547848] page:ea0003fb7db0
On Mon, Nov 17, 2014 at 01:56:19PM +0530, Aneesh Kumar K.V wrote:
Mel Gorman mgor...@suse.de writes:
This is follow up from the pipe/page fault oddness thread.
Automatic NUMA balancing depends on being able to protect PTEs to trap a
fault and gather reference locality information. Very
Mel Gorman mgor...@suse.de writes:
On Mon, Nov 17, 2014 at 01:56:19PM +0530, Aneesh Kumar K.V wrote:
Mel Gorman mgor...@suse.de writes:
This is follow up from the pipe/page fault oddness thread.
Automatic NUMA balancing depends on being able to protect PTEs to trap a
fault and gather
On 11/18/2014 10:42 AM, Mel Gorman wrote:
1. I'm assuming this is a KVM setup but can you confirm?
Yes.
2. Are you using numa=fake=N?
Yes. numa=fake=24, which is probably way more nodes on any physical machine
than the new code was tested on?
3. If you are using fake NUMA, what happens if
Sasha Levin sasha.le...@oracle.com writes:
On 11/18/2014 10:42 AM, Mel Gorman wrote:
1. I'm assuming this is a KVM setup but can you confirm?
Yes.
2. Are you using numa=fake=N?
Yes. numa=fake=24, which is probably way more nodes on any physical machine
than the new code was tested on?
On Tue, Nov 18, 2014 at 10:03:30PM +0530, Aneesh Kumar K.V wrote:
diff --git a/arch/powerpc/mm/copro_fault.c b/arch/powerpc/mm/copro_fault.c
index 5a236f0..46152aa 100644
--- a/arch/powerpc/mm/copro_fault.c
+++ b/arch/powerpc/mm/copro_fault.c
@@ -64,7 +64,12 @@ int
On Tue, Nov 18, 2014 at 10:26:41PM +0530, Aneesh Kumar K.V wrote:
Sasha Levin sasha.le...@oracle.com writes:
On 11/18/2014 10:42 AM, Mel Gorman wrote:
1. I'm assuming this is a KVM setup but can you confirm?
Yes.
2. Are you using numa=fake=N?
Yes. numa=fake=24, which is probably
On 11/18/2014 11:56 AM, Aneesh Kumar K.V wrote:
4. Similarly, does the kernel boot properly without without patches?
Yes, the kernel works fine without the patches both with and without fake
numa.
Hmm that is interesting. I am not sure how writeback_fid can be
related. We use writeback
Mel Gorman writes:
> This is follow up from the "pipe/page fault oddness" thread.
>
> Automatic NUMA balancing depends on being able to protect PTEs to trap a
> fault and gather reference locality information. Very broadly speaking it
> would mark PTEs as not present and use another bit to
Mel Gorman mgor...@suse.de writes:
This is follow up from the pipe/page fault oddness thread.
Automatic NUMA balancing depends on being able to protect PTEs to trap a
fault and gather reference locality information. Very broadly speaking it
would mark PTEs as not present and use another bit
On 11/14/2014 08:32 AM, Mel Gorman wrote:> This is follow up from the
"pipe/page fault oddness" thread.
Hi Mel,
Applying this patch series I've started seeing the following straight away:
[ 367.547848] page:ea0003fb7db0 count:1007 mapcount:1005
mapping:8800691f2f58 index:0x37
[
On Fri, Nov 14, 2014 at 5:32 AM, Mel Gorman wrote:
>
> This series is very heavily based on patches from Linus and Aneesh to
> replace the existing PTE/PMD NUMA helper functions with normal change
> protections. I did alter and add parts of it but I consider them relatively
> minor contributions.
This is follow up from the "pipe/page fault oddness" thread.
Automatic NUMA balancing depends on being able to protect PTEs to trap a
fault and gather reference locality information. Very broadly speaking it
would mark PTEs as not present and use another bit to distinguish between
NUMA hinting
This is follow up from the pipe/page fault oddness thread.
Automatic NUMA balancing depends on being able to protect PTEs to trap a
fault and gather reference locality information. Very broadly speaking it
would mark PTEs as not present and use another bit to distinguish between
NUMA hinting
On Fri, Nov 14, 2014 at 5:32 AM, Mel Gorman mgor...@suse.de wrote:
This series is very heavily based on patches from Linus and Aneesh to
replace the existing PTE/PMD NUMA helper functions with normal change
protections. I did alter and add parts of it but I consider them relatively
minor
On 11/14/2014 08:32 AM, Mel Gorman wrote: This is follow up from the
pipe/page fault oddness thread.
Hi Mel,
Applying this patch series I've started seeing the following straight away:
[ 367.547848] page:ea0003fb7db0 count:1007 mapcount:1005
mapping:8800691f2f58 index:0x37
[
26 matches
Mail list logo
