This patch removes a check that prevents correct CPU topology setup
when cpuid extensions for topology enumeration are not supported and
the number of processors reported to the OS is smaller than
smp_num_siblings
Signed-off-by: Nikhil P Rao <[email protected]>
Acked-by: Suresh Siddha <[email protected]>
CC: <[email protected]>
---
arch/x86/kernel/cpu/common.c | 7 -------
1 files changed, 0 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index e2ced007..6547084 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -458,13 +458,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
if (smp_num_siblings <= 1)
goto out;
- if (smp_num_siblings > nr_cpu_ids) {
- pr_warning("CPU: Unsupported number of siblings %d",
- smp_num_siblings);
- smp_num_siblings = 1;
- return;
- }
-
index_msb = get_count_order(smp_num_siblings);
c->phys_proc_id = apic->phys_pkg_id(c->initial_apicid, index_msb);
--
1.7.4.2
????{.n?+???????+%?????ݶ??w??{.n?+????{??G?????{ay? ʇڙ?,j??f???h?????????z_??(?階?ݢj"???m??????G????????????&???~???iO???z??v?^?m???? ????????I?
* Rao, Nikhil <[email protected]> wrote:
> This patch removes a check that prevents correct CPU topology setup
> when cpuid extensions for topology enumeration are not supported and
> the number of processors reported to the OS is smaller than
> smp_num_siblings
What bad effect was observed, exactly? An annoying warning in the bootlog and
incorrect scheduler domain setup (SMP instead of SMT) - or something else as
well?
Thanks,
Ingo
Nothing apart from what you mention - bootlog message and SMP instead of SMT.
Nikhil
> -----Original Message-----
> From: Ingo Molnar [mailto:[email protected]]
> Sent: Wednesday, May 18, 2011 2:08 PM
> To: Rao, Nikhil
> Cc: [email protected]; [email protected]; [email protected];
> [email protected]; Siddha, Suresh B; [email protected]
> Subject: Re: [PATCH] x86: Removing unnecessary check in detect_ht
>
>
> * Rao, Nikhil <[email protected]> wrote:
>
> > This patch removes a check that prevents correct CPU topology setup
> > when cpuid extensions for topology enumeration are not supported and
> > the number of processors reported to the OS is smaller than
> > smp_num_siblings
>
> What bad effect was observed, exactly? An annoying warning in the
> bootlog and
> incorrect scheduler domain setup (SMP instead of SMT) - or something
> else as
> well?
>
> Thanks,
>
> Ingo
* Rao, Nikhil <[email protected]> wrote:
> Nothing apart from what you mention - bootlog message and SMP instead of SMT.
Mind putting this information into the changelog and repost the patch? It's
arguably the kind of practical information that most readers of changelogs are
interested in.
Thanks,
Ingo
Hi Ingo,
Reposting as per your feedback.
Thanks,
Nikhil
---
This patch removes a check that causes incorrect scheduler domain
setup (SMP instead of SMT) and bootlog warning messages when cpuid
extensions for topology enumeration are not supported and the number
of processors reported to the OS is smaller than smp_num_siblings
Signed-off-by: Nikhil P Rao <[email protected]>
Acked-by: Suresh Siddha <[email protected]>
---
arch/x86/kernel/cpu/common.c | 7 -------
1 files changed, 0 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index e2ced007..6547084 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -458,13 +458,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
if (smp_num_siblings <= 1)
goto out;
- if (smp_num_siblings > nr_cpu_ids) {
- pr_warning("CPU: Unsupported number of siblings %d",
- smp_num_siblings);
- smp_num_siblings = 1;
- return;
- }
-
index_msb = get_count_order(smp_num_siblings);
c->phys_proc_id = apic->phys_pkg_id(c->initial_apicid, index_msb);
--
1.7.4.2
> -----Original Message-----
> From: Ingo Molnar [mailto:[email protected]]
> Sent: Wednesday, May 18, 2011 2:24 PM
> To: Rao, Nikhil
> Cc: [email protected]; [email protected]; [email protected];
> [email protected]; Siddha, Suresh B; [email protected]
> Subject: Re: [PATCH] x86: Removing unnecessary check in detect_ht
>
>
> * Rao, Nikhil <[email protected]> wrote:
>
> > Nothing apart from what you mention - bootlog message and SMP instead
> of SMT.
>
> Mind putting this information into the changelog and repost the patch?
> It's
> arguably the kind of practical information that most readers of
> changelogs are
> interested in.
>
> Thanks,
>
> Ingo
* Rao, Nikhil <[email protected]> wrote:
> Hi Ingo,
>
> Reposting as per your feedback.
Please send a patch that applies to the x86 tree cleanly:
http://people.redhat.com/mingo/tip.git/README
Thanks,
Ingo
This patch removes a check that causes incorrect scheduler domain setup (SMP instead of SMT) and bootlog warning messages when cpuid extensions for topology enumeration are not supported and the number of processors reported to the OS is smaller than smp_num_siblings
Signed-off-by: Nikhil P Rao <[email protected]>
Acked-by: Suresh Siddha <[email protected]>
---
arch/x86/kernel/cpu/common.c | 7 -------
1 files changed, 0 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index cbc70a2..d2bcd54 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -477,13 +477,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
if (smp_num_siblings <= 1)
goto out;
- if (smp_num_siblings > nr_cpu_ids) {
- pr_warning("CPU: Unsupported number of siblings %d",
- smp_num_siblings);
- smp_num_siblings = 1;
- return;
- }
-
index_msb = get_count_order(smp_num_siblings);
c->phys_proc_id = apic->phys_pkg_id(c->initial_apicid, index_msb);
--
1.7.4.2
> -----Original Message-----
> From: Ingo Molnar [mailto:[email protected]]
> Sent: Thursday, May 19, 2011 11:10 AM
> To: Rao, Nikhil
> Cc: [email protected]; [email protected]; [email protected];
> [email protected]; Siddha, Suresh B; [email protected]
> Subject: Re: [PATCH] x86: Removing unnecessary check in detect_ht
>
> * Rao, Nikhil <[email protected]> wrote:
>
> > Hi Ingo,
> >
> > Reposting as per your feedback.
>
> Please send a patch that applies to the x86 tree cleanly:
>
> http://people.redhat.com/mingo/tip.git/README
>
> Thanks,
>
> Ingo
* Rao, Nikhil <[email protected]> wrote:
> This patch removes a check that causes incorrect scheduler domain setup (SMP instead of SMT) and bootlog warning messages when cpuid extensions for topology enumeration are not supported and the number of processors reported to the OS is smaller than smp_num_siblings
>
> Signed-off-by: Nikhil P Rao <[email protected]>
> Acked-by: Suresh Siddha <[email protected]>
> ---
> arch/x86/kernel/cpu/common.c | 7 -------
> 1 files changed, 0 insertions(+), 7 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index cbc70a2..d2bcd54 100644
> --- a/arch/x86/kernel/cpu/common.c
> +++ b/arch/x86/kernel/cpu/common.c
> @@ -477,13 +477,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
> if (smp_num_siblings <= 1)
> goto out;
>
> - if (smp_num_siblings > nr_cpu_ids) {
> - pr_warning("CPU: Unsupported number of siblings %d",
> - smp_num_siblings);
> - smp_num_siblings = 1;
> - return;
> - }
> -
Your patch does not apply because it's whitespace damaged (all tabs got
converted to spaces). Please see Documentation/email-clients.txt.
Thanks,
Ingo
This patch removes a check that causes incorrect scheduler domain
setup (SMP instead of SMT) and bootlog warning messages when cpuid
extensions for topology enumeration are not supported and the number
of processors reported to the OS is smaller than smp_num_siblings
Acked-by: Suresh Siddha <[email protected]>
Signed-off-by: Nikhil P Rao <[email protected]>
---
arch/x86/kernel/cpu/common.c | 7 -------
1 files changed, 0 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index cbc70a2..d2bcd54 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -477,13 +477,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
if (smp_num_siblings <= 1)
goto out;
- if (smp_num_siblings > nr_cpu_ids) {
- pr_warning("CPU: Unsupported number of siblings %d",
- smp_num_siblings);
- smp_num_siblings = 1;
- return;
- }
-
index_msb = get_count_order(smp_num_siblings);
c->phys_proc_id = apic->phys_pkg_id(c->initial_apicid, index_msb);
--
1.7.4.4
On Fri, 2011-05-20 at 13:46 +0200, Ingo Molnar wrote:
> Your patch does not apply because it's whitespace damaged (all tabs got
> converted to spaces). Please see Documentation/email-clients.txt.
>
> Thanks,
>
> Ingo
Commit-ID: 8b27f2ff7a24f0735c96055e676872f05398d99b
Gitweb: http://git.kernel.org/tip/8b27f2ff7a24f0735c96055e676872f05398d99b
Author: Nikhil P Rao <[email protected]>
AuthorDate: Wed, 25 May 2011 10:18:41 -0700
Committer: Ingo Molnar <[email protected]>
CommitDate: Wed, 25 May 2011 23:01:08 +0200
x86: Remove unnecessary check in detect_ht()
This patch removes a check that causes incorrect scheduler
domain setup (SMP instead of SMT) and bootlog warning messages
when cpuid extensions for topology enumeration are not supported
and the number of processors reported to the OS is smaller than
smp_num_siblings.
Acked-by: Suresh Siddha <[email protected]>
Signed-off-by: Nikhil P Rao <[email protected]>
Link: http://lkml.kernel.org/r/1306343921.19325.1.camel@fedora13
Signed-off-by: Ingo Molnar <[email protected]>
---
arch/x86/kernel/cpu/common.c | 7 -------
1 files changed, 0 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index c8b4162..53f02f5 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -477,13 +477,6 @@ void __cpuinit detect_ht(struct cpuinfo_x86 *c)
if (smp_num_siblings <= 1)
goto out;
- if (smp_num_siblings > nr_cpu_ids) {
- pr_warning("CPU: Unsupported number of siblings %d",
- smp_num_siblings);
- smp_num_siblings = 1;
- return;
- }
-
index_msb = get_count_order(smp_num_siblings);
c->phys_proc_id = apic->phys_pkg_id(c->initial_apicid, index_msb);