On 17.02.2014 [15:14:06 -0800], David Rientjes wrote:
> On Mon, 17 Feb 2014, Nishanth Aravamudan wrote:
>
> > Here is what I'm running into now:
> >
> > setup_arch ->
> > do_init_bootmem ->
> > cpu_numa_callback ->
> > numa_setup_cpu ->
> >
On Mon, 17 Feb 2014, Nishanth Aravamudan wrote:
> Here is what I'm running into now:
>
> setup_arch ->
> do_init_bootmem ->
> cpu_numa_callback ->
> numa_setup_cpu ->
> map_cpu_to_node ->
>
On Thu, Feb 13, 2014 at 8:32 PM, Nishanth Aravamudan
wrote:
>
> Agreed that for the readahead case the above is probably more than
> sufficient.
>
> Apologies for hijacking the thread, my comments below were purely about
> the memoryless node support, not about readahead specifically.
Ok, no prob
On Thu, Feb 13, 2014 at 11:43 PM, Jan Kara wrote:
>
> max_sane_readahead() is also used for limiting amount of readahead for
> [fm]advice(2) WILLNEED and that is used e.g. by a dynamic linker to preload
> shared libraries into memory. So I'm convinced this usecase *will* notice
> the change - ef
On 14.02.2014 [02:54:06 -0800], David Rientjes wrote:
> On Thu, 13 Feb 2014, Nishanth Aravamudan wrote:
>
> > There is an open issue on powerpc with memoryless nodes (inasmuch as we
> > can have them, but the kernel doesn't support it properly). There is a
> > separate discussion going on on linux
On Thu, 13 Feb 2014, Nishanth Aravamudan wrote:
> There is an open issue on powerpc with memoryless nodes (inasmuch as we
> can have them, but the kernel doesn't support it properly). There is a
> separate discussion going on on linuxppc-dev about what is necessary for
> CONFIG_HAVE_MEMORYLESS_NOD
On Thu 13-02-14 16:37:53, Linus Torvalds wrote:
> Is this whole thread still just for the crazy and pointless
> "max_sane_readahead()"?
>
> Or is there some *real* reason we should care?
>
> Because if it really is just for max_sane_readahead(), then for the
> love of God, let us just do this
>
On 13.02.2014 [14:41:04 -0800], David Rientjes wrote:
> On Thu, 13 Feb 2014, Raghavendra K T wrote:
>
> > Thanks David, unfortunately even after applying that patch, I do not see
> > the improvement.
> >
> > Interestingly numa_mem_id() seem to still return the value of a
> > memoryless node.
> >
Hi Linus,
On 13.02.2014 [16:37:53 -0800], Linus Torvalds wrote:
> Is this whole thread still just for the crazy and pointless
> "max_sane_readahead()"?
>
> Or is there some *real* reason we should care?
There is an open issue on powerpc with memoryless nodes (inasmuch as we
can have them, but th
On Thu, 13 Feb 2014 16:37:53 -0800 Linus Torvalds
wrote:
> unsigned long max_sane_readahead(unsigned long nr)
> {
> return min(nr, 128);
> }
I bet nobody will notice.
It should be 128*4096/PAGE_CACHE_SIZE so that variations in PAGE_SIZE
don't affect readahead behaviour.
--
To unsub
Is this whole thread still just for the crazy and pointless
"max_sane_readahead()"?
Or is there some *real* reason we should care?
Because if it really is just for max_sane_readahead(), then for the
love of God, let us just do this
unsigned long max_sane_readahead(unsigned long nr)
{
r
On 13.02.2014 [14:41:04 -0800], David Rientjes wrote:
> On Thu, 13 Feb 2014, Raghavendra K T wrote:
>
> > Thanks David, unfortunately even after applying that patch, I do not see
> > the improvement.
> >
> > Interestingly numa_mem_id() seem to still return the value of a
> > memoryless node.
> >
On Thu, 13 Feb 2014, Raghavendra K T wrote:
> Thanks David, unfortunately even after applying that patch, I do not see
> the improvement.
>
> Interestingly numa_mem_id() seem to still return the value of a
> memoryless node.
> May be per cpu _numa_mem_ values are not set properly. Need to dig ou
On 13.02.2014 [13:06:43 -0800], Andrew Morton wrote:
> On Thu, 13 Feb 2014 00:05:31 -0800 (PST) David Rientjes
> wrote:
>
> > On Thu, 13 Feb 2014, Raghavendra K T wrote:
> >
> > > I was able to test (1) implementation on the system where readahead
> > > problem
> > > occurred. Unfortunately it
On Thu, 13 Feb 2014 00:05:31 -0800 (PST) David Rientjes
wrote:
> On Thu, 13 Feb 2014, Raghavendra K T wrote:
>
> > I was able to test (1) implementation on the system where readahead problem
> > occurred. Unfortunately it did not help.
> >
> > Reason seem to be that CONFIG_HAVE_MEMORYLESS_NODE
On 02/13/2014 01:35 PM, David Rientjes wrote:
On Thu, 13 Feb 2014, Raghavendra K T wrote:
I was able to test (1) implementation on the system where readahead problem
occurred. Unfortunately it did not help.
Reason seem to be that CONFIG_HAVE_MEMORYLESS_NODES dependency of
numa_mem_id(). The PP
On Thu, 13 Feb 2014, Raghavendra K T wrote:
> I was able to test (1) implementation on the system where readahead problem
> occurred. Unfortunately it did not help.
>
> Reason seem to be that CONFIG_HAVE_MEMORYLESS_NODES dependency of
> numa_mem_id(). The PPC machine I am facing problem has topol
On 02/11/2014 03:05 AM, David Rientjes wrote:
On Mon, 10 Feb 2014, Raghavendra K T wrote:
So I understood that you are suggesting implementations like below
1) I do not have problem with the below approach, I could post this in
next version.
( But this did not include 4k limit Linus mentioned
On Mon, 10 Feb 2014, Raghavendra K T wrote:
> So I understood that you are suggesting implementations like below
>
> 1) I do not have problem with the below approach, I could post this in
> next version.
> ( But this did not include 4k limit Linus mentioned to apply)
>
> unsigned long max_sane_r
On 02/10/2014 03:35 PM, David Rientjes wrote:
On Mon, 10 Feb 2014, Raghavendra K T wrote:
As you rightly pointed , I 'll drop remote memory term and use
something like :
"* Ensure readahead success on a memoryless node cpu. But we limit
* the readahead to 4k pages to avoid trashing page cac
On Mon, 10 Feb 2014, Raghavendra K T wrote:
> As you rightly pointed , I 'll drop remote memory term and use
> something like :
>
> "* Ensure readahead success on a memoryless node cpu. But we limit
> * the readahead to 4k pages to avoid trashing page cache." ..
>
I don't know how to proceed
On 02/08/2014 02:11 AM, David Rientjes wrote:
On Fri, 7 Feb 2014, Raghavendra K T wrote:
3) Change the "readahead into remote memory" part of the documentation
which is misleading.
( I feel no need to add numa_mem_id() since we would specifically limit
the readahead with MAX_REMOTE_READAHEAD in
On Fri, 7 Feb 2014, Raghavendra K T wrote:
> So following discussion TODO for my patch is:
>
> 1) Update the changelog with user visible impact of the patch.
> (Andrew's suggestion)
> 2) Add ACCESS_ONCE to numa_node_id().
> 3) Change the "readahead into remote memory" part of the documentation
>
On 02/07/2014 05:28 AM, David Rientjes wrote:
On Thu, 6 Feb 2014, David Rientjes wrote:
+#define MAX_REMOTE_READAHEAD 4096UL
Normally it wouldn't matter because there's no significant downside to it
racing, things like mempolicies which use numa_node_id() extensively would
result in, oops,
On Thu, 6 Feb 2014, David Rientjes wrote:
> > > > > +#define MAX_REMOTE_READAHEAD 4096UL
> Normally it wouldn't matter because there's no significant downside to it
> racing, things like mempolicies which use numa_node_id() extensively would
> result in, oops, a page allocation on the wrong n
On Thu, 6 Feb 2014, Andrew Morton wrote:
> On Thu, 6 Feb 2014 14:58:21 -0800 (PST) David Rientjes
> wrote:
>
> > > > +#define MAX_REMOTE_READAHEAD 4096UL
> > > > /*
> > > > * Given a desired number of PAGE_CACHE_SIZE readahead pages, return a
> > > > * sensible upper limit.
> > > > */
On Thu, 6 Feb 2014 14:58:21 -0800 (PST) David Rientjes
wrote:
> > > +#define MAX_REMOTE_READAHEAD 4096UL
> > > /*
> > > * Given a desired number of PAGE_CACHE_SIZE readahead pages, return a
> > > * sensible upper limit.
> > > */
> > > unsigned long max_sane_readahead(unsigned long nr)
On Thu, 6 Feb 2014, Andrew Morton wrote:
> > --- a/mm/readahead.c
> > +++ b/mm/readahead.c
> > @@ -237,14 +237,32 @@ int force_page_cache_readahead(struct address_space
> > *mapping, struct file *filp,
> > return ret;
> > }
> >
> > +#define MAX_REMOTE_READAHEAD 4096UL
> > /*
> > * Giv
On Wed, 22 Jan 2014 16:23:45 +0530 Raghavendra K T
wrote:
> max_sane_readahead returns zero on the cpu having no local memory
> node. Fix that by returning a sanitized number of pages viz.,
> minimum of (requested pages, 4k)
um, fix what? The changelog should describe the user-visible impact o
On 01/22/2014 04:23 PM, Raghavendra K T wrote:
max_sane_readahead returns zero on the cpu having no local memory
node. Fix that by returning a sanitized number of pages viz.,
minimum of (requested pages, 4k)
Result:
fadvise experiment with FADV_WILLNEED on a x240 machine with 1GB testfile
32GB*
max_sane_readahead returns zero on the cpu having no local memory
node. Fix that by returning a sanitized number of pages viz.,
minimum of (requested pages, 4k)
Result:
fadvise experiment with FADV_WILLNEED on a x240 machine with 1GB testfile
32GB* 4G RAM numa machine ( 12 iterations) yielded
Ke
31 matches
Mail list logo
