* [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT
@ 2013-12-04 21:09 Toshi Kani
2013-12-05 10:25 ` Yasuaki Ishimatsu
0 siblings, 1 reply; 6+ messages in thread
From: Toshi Kani @ 2013-12-04 21:09 UTC (permalink / raw)
To: akpm, mingo, hpa, tglx; +Cc: linux-mm, linux-kernel, x86, Toshi Kani
When ACPI SLIT table has an I/O locality (i.e. a locality unique
to an I/O device), numa_set_distance() emits the warning message
below.
NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
acpi_numa_slit_init() calls numa_set_distance() with pxm_to_node(),
which assumes that all localities have been parsed with SRAT previously.
SRAT does not list I/O localities, where as SLIT lists all localities
including I/Os. Hence, pxm_to_node() returns NUMA_NO_NODE (-1) for
an I/O locality. I/O localities are not supported and are ignored
today, but emitting such warning message leads unnecessary confusion.
Change acpi_numa_slit_init() to avoid calling numa_set_distance()
with NUMA_NO_NODE.
Signed-off-by: Toshi Kani <toshi.kani@hp.com>
---
arch/x86/mm/srat.c | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/arch/x86/mm/srat.c b/arch/x86/mm/srat.c
index 266ca91..29a2ced 100644
--- a/arch/x86/mm/srat.c
+++ b/arch/x86/mm/srat.c
@@ -47,10 +47,16 @@ void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
{
int i, j;
- for (i = 0; i < slit->locality_count; i++)
- for (j = 0; j < slit->locality_count; j++)
+ for (i = 0; i < slit->locality_count; i++) {
+ if (pxm_to_node(i) == NUMA_NO_NODE)
+ continue;
+ for (j = 0; j < slit->locality_count; j++) {
+ if (pxm_to_node(j) == NUMA_NO_NODE)
+ continue;
numa_set_distance(pxm_to_node(i), pxm_to_node(j),
slit->entry[slit->locality_count * i + j]);
+ }
+ }
}
/* Callback for Proximity Domain -> x2APIC mapping */
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT
2013-12-04 21:09 [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT Toshi Kani
@ 2013-12-05 10:25 ` Yasuaki Ishimatsu
2013-12-05 15:11 ` Toshi Kani
2013-12-05 23:21 ` David Rientjes
0 siblings, 2 replies; 6+ messages in thread
From: Yasuaki Ishimatsu @ 2013-12-05 10:25 UTC (permalink / raw)
To: Toshi Kani; +Cc: akpm, mingo, hpa, tglx, linux-mm, linux-kernel, x86
(2013/12/05 6:09), Toshi Kani wrote:
> When ACPI SLIT table has an I/O locality (i.e. a locality unique
> to an I/O device), numa_set_distance() emits the warning message
> below.
>
> NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
>
> acpi_numa_slit_init() calls numa_set_distance() with pxm_to_node(),
> which assumes that all localities have been parsed with SRAT previously.
> SRAT does not list I/O localities, where as SLIT lists all localities
> including I/Os. Hence, pxm_to_node() returns NUMA_NO_NODE (-1) for
> an I/O locality. I/O localities are not supported and are ignored
> today, but emitting such warning message leads unnecessary confusion.
In this case, the warning message should not be shown. But if SLIT table
is really broken, the message should be shown. Your patch seems to not care
for second case.
Thanks,
Yasuaki Ishimatsu
>
> Change acpi_numa_slit_init() to avoid calling numa_set_distance()
> with NUMA_NO_NODE.
>
> Signed-off-by: Toshi Kani <toshi.kani@hp.com>
> ---
> arch/x86/mm/srat.c | 10 ++++++++--
> 1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/mm/srat.c b/arch/x86/mm/srat.c
> index 266ca91..29a2ced 100644
> --- a/arch/x86/mm/srat.c
> +++ b/arch/x86/mm/srat.c
> @@ -47,10 +47,16 @@ void __init acpi_numa_slit_init(struct acpi_table_slit *slit)
> {
> int i, j;
>
> - for (i = 0; i < slit->locality_count; i++)
> - for (j = 0; j < slit->locality_count; j++)
> + for (i = 0; i < slit->locality_count; i++) {
> + if (pxm_to_node(i) == NUMA_NO_NODE)
> + continue;
> + for (j = 0; j < slit->locality_count; j++) {
> + if (pxm_to_node(j) == NUMA_NO_NODE)
> + continue;
> numa_set_distance(pxm_to_node(i), pxm_to_node(j),
> slit->entry[slit->locality_count * i + j]);
> + }
> + }
> }
>
> /* Callback for Proximity Domain -> x2APIC mapping */
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org. For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT
2013-12-05 10:25 ` Yasuaki Ishimatsu
@ 2013-12-05 15:11 ` Toshi Kani
2013-12-06 7:23 ` Yasuaki Ishimatsu
2013-12-05 23:21 ` David Rientjes
1 sibling, 1 reply; 6+ messages in thread
From: Toshi Kani @ 2013-12-05 15:11 UTC (permalink / raw)
To: Yasuaki Ishimatsu; +Cc: akpm, mingo, hpa, tglx, linux-mm, linux-kernel, x86
On Thu, 2013-12-05 at 19:25 +0900, Yasuaki Ishimatsu wrote:
> (2013/12/05 6:09), Toshi Kani wrote:
> > When ACPI SLIT table has an I/O locality (i.e. a locality unique
> > to an I/O device), numa_set_distance() emits the warning message
> > below.
> >
> > NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
> >
> > acpi_numa_slit_init() calls numa_set_distance() with pxm_to_node(),
> > which assumes that all localities have been parsed with SRAT previously.
> > SRAT does not list I/O localities, where as SLIT lists all localities
>
> > including I/Os. Hence, pxm_to_node() returns NUMA_NO_NODE (-1) for
> > an I/O locality. I/O localities are not supported and are ignored
> > today, but emitting such warning message leads unnecessary confusion.
>
> In this case, the warning message should not be shown. But if SLIT table
> is really broken, the message should be shown. Your patch seems to not care
> for second case.
In the second case, I assume you are worrying about the case of SLIT
table with bad locality numbers. Since SLIT is a matrix of the number
of localities, it is only possible by making the table bigger than
necessary. Such excessive localities are safe to ignore (as they are
ignored today) and regular users have nothing to concern about them.
The warning message in this case may be helpful for platform vendors to
test their firmware, but they have plenty of other methods to verify
their SLIT table.
Thanks,
-Toshi
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT
2013-12-05 15:11 ` Toshi Kani
@ 2013-12-06 7:23 ` Yasuaki Ishimatsu
2013-12-06 14:52 ` Toshi Kani
0 siblings, 1 reply; 6+ messages in thread
From: Yasuaki Ishimatsu @ 2013-12-06 7:23 UTC (permalink / raw)
To: Toshi Kani; +Cc: akpm, mingo, hpa, tglx, linux-mm, linux-kernel, x86
(2013/12/06 0:11), Toshi Kani wrote:
> On Thu, 2013-12-05 at 19:25 +0900, Yasuaki Ishimatsu wrote:
>> (2013/12/05 6:09), Toshi Kani wrote:
>>> When ACPI SLIT table has an I/O locality (i.e. a locality unique
>>> to an I/O device), numa_set_distance() emits the warning message
>>> below.
>>>
>>> NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
>>>
>>> acpi_numa_slit_init() calls numa_set_distance() with pxm_to_node(),
>>> which assumes that all localities have been parsed with SRAT previously.
>>> SRAT does not list I/O localities, where as SLIT lists all localities
>>
>>> including I/Os. Hence, pxm_to_node() returns NUMA_NO_NODE (-1) for
>>> an I/O locality. I/O localities are not supported and are ignored
>>> today, but emitting such warning message leads unnecessary confusion.
>>
>> In this case, the warning message should not be shown. But if SLIT table
>> is really broken, the message should be shown. Your patch seems to not care
>> for second case.
>
> In the second case, I assume you are worrying about the case of SLIT
> table with bad locality numbers. Since SLIT is a matrix of the number
> of localities, it is only possible by making the table bigger than
> necessary. Such excessive localities are safe to ignore (as they are
> ignored today) and regular users have nothing to concern about them.
> The warning message in this case may be helpful for platform vendors to
> test their firmware, but they have plenty of other methods to verify
> their SLIT table.
I understood it. So,
Reviewed-by : Yasuaki Ishimatsu
Thanks,
Yasuaki Ishimatsu
> Thanks,
> -Toshi
>
>
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org. For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT
2013-12-06 7:23 ` Yasuaki Ishimatsu
@ 2013-12-06 14:52 ` Toshi Kani
0 siblings, 0 replies; 6+ messages in thread
From: Toshi Kani @ 2013-12-06 14:52 UTC (permalink / raw)
To: Yasuaki Ishimatsu; +Cc: akpm, mingo, hpa, tglx, linux-mm, linux-kernel, x86
On Fri, 2013-12-06 at 16:23 +0900, Yasuaki Ishimatsu wrote:
> (2013/12/06 0:11), Toshi Kani wrote:
> > On Thu, 2013-12-05 at 19:25 +0900, Yasuaki Ishimatsu wrote:
> >> (2013/12/05 6:09), Toshi Kani wrote:
> >>> When ACPI SLIT table has an I/O locality (i.e. a locality unique
> >>> to an I/O device), numa_set_distance() emits the warning message
> >>> below.
> >>>
> >>> NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
> >>>
> >>> acpi_numa_slit_init() calls numa_set_distance() with pxm_to_node(),
> >>> which assumes that all localities have been parsed with SRAT previously.
> >>> SRAT does not list I/O localities, where as SLIT lists all localities
> >>
> >>> including I/Os. Hence, pxm_to_node() returns NUMA_NO_NODE (-1) for
> >>> an I/O locality. I/O localities are not supported and are ignored
> >>> today, but emitting such warning message leads unnecessary confusion.
> >>
> >> In this case, the warning message should not be shown. But if SLIT table
> >> is really broken, the message should be shown. Your patch seems to not care
> >> for second case.
> >
> > In the second case, I assume you are worrying about the case of SLIT
> > table with bad locality numbers. Since SLIT is a matrix of the number
> > of localities, it is only possible by making the table bigger than
> > necessary. Such excessive localities are safe to ignore (as they are
> > ignored today) and regular users have nothing to concern about them.
> > The warning message in this case may be helpful for platform vendors to
> > test their firmware, but they have plenty of other methods to verify
> > their SLIT table.
>
> I understood it. So,
>
> Reviewed-by : Yasuaki Ishimatsu
Great. Thanks Yasuaki!
-Toshi
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT
2013-12-05 10:25 ` Yasuaki Ishimatsu
2013-12-05 15:11 ` Toshi Kani
@ 2013-12-05 23:21 ` David Rientjes
1 sibling, 0 replies; 6+ messages in thread
From: David Rientjes @ 2013-12-05 23:21 UTC (permalink / raw)
To: Yasuaki Ishimatsu
Cc: Toshi Kani, akpm, mingo, hpa, tglx, linux-mm, linux-kernel, x86
On Thu, 5 Dec 2013, Yasuaki Ishimatsu wrote:
> (2013/12/05 6:09), Toshi Kani wrote:
> > When ACPI SLIT table has an I/O locality (i.e. a locality unique
> > to an I/O device), numa_set_distance() emits the warning message
> > below.
> >
> > NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
> >
> > acpi_numa_slit_init() calls numa_set_distance() with pxm_to_node(),
> > which assumes that all localities have been parsed with SRAT previously.
> > SRAT does not list I/O localities, where as SLIT lists all localities
>
> > including I/Os. Hence, pxm_to_node() returns NUMA_NO_NODE (-1) for
> > an I/O locality. I/O localities are not supported and are ignored
> > today, but emitting such warning message leads unnecessary confusion.
>
> In this case, the warning message should not be shown. But if SLIT table
> is really broken, the message should be shown. Your patch seems to not care
> for second case.
>
It's a subtle problem of the difference in the definition of a "NUMA node"
between the ACPI specification and how it is defined in the kernel. The
specification allows I/O buses to define a NUMA node and the kernel
doesn't setup a separate node id for them, so there's no way to
distinguish between an erroneous SLIT and system localities that only
include I/O devices. If that were to change in the future we could remove
this limitation since pxm_to_node() wouldn't return NUMA_NO_NODE for
Toshi's config. A follow-up patch that adds the comment about why this is
done has been proposed to be folded into this patch.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2013-12-06 14:57 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-12-04 21:09 [PATCH] mm, x86: Skip NUMA_NO_NODE while parsing SLIT Toshi Kani
2013-12-05 10:25 ` Yasuaki Ishimatsu
2013-12-05 15:11 ` Toshi Kani
2013-12-06 7:23 ` Yasuaki Ishimatsu
2013-12-06 14:52 ` Toshi Kani
2013-12-05 23:21 ` David Rientjes
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).