lkml.org 
[lkml]   [2014]   [Jan]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v2] x86, mm: Avoid duplicated pxm_to_node() calling.
From
Date
On Sun, 2014-01-26 at 13:01 -0800, Yinghai Lu wrote:
> In slit init code, more pxm_to_node() calling are added.
>
> We can store from_node/to_node instead of keep calling pxm_to_node().
>
> After this patch: pxm_to_node() is called n*(1+n)
> Before this patch: pxm_to_node() is called n*(1+n*3)
>
> for 8 socket, it will be 72 instead of 200.
> for 32 socket, it will be 1056 instead of 3104.
>
> -v2: update title and change log according to Ingo.
> move from_node/to_node in loop and change to const according to
> David Rientjes.
>
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>

In my original patch, I added the following comment to the function to
address David's comment (which he acked). So, can you add this comment?

-/* Callback for SLIT parsing */
+/*
+ * Callback for SLIT parsing. pxm_to_node() returns NUMA_NO_NODE for
+ * I/O localities since SRAT does not list them. I/O localities are
+ * not supported at this point.
+ */

Otherwise, the change looks good to me.

Acked-by: Toshi Kani <toshi.kani@hp.com>


Thanks,
-Toshi



\
 
 \ /
  Last update: 2014-01-27 16:42    [W:0.094 / U:0.032 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site