Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1418636 > unrolled thread
| Started by | Andi Kleen <andi@firstfloor.org> |
|---|---|
| First post | 2016-06-09 23:00 +0200 |
| Last post | 2016-06-10 12:10 +0200 |
| Articles | 3 — 3 participants |
Back to article view | Back to linux.kernel
[PATCH] x86/microcode/intel: Quieten down microcode updates on large systems Andi Kleen <andi@firstfloor.org> - 2016-06-09 23:00 +0200
Re: [PATCH] x86/microcode/intel: Quieten down microcode updates on large systems Henrique de Moraes Holschuh <hmh@hmh.eng.br> - 2016-06-10 02:00 +0200
Re: [PATCH] x86/microcode/intel: Quieten down microcode updates on large systems Borislav Petkov <bp@alien8.de> - 2016-06-10 12:10 +0200
| From | Andi Kleen <andi@firstfloor.org> |
|---|---|
| Date | 2016-06-09 23:00 +0200 |
| Subject | [PATCH] x86/microcode/intel: Quieten down microcode updates on large systems |
| Message-ID | <rIfdv-360-5@gated-at.bofh.it> |
From: Andi Kleen <ak@linux.intel.com>
On large systems the microcode driver is very noisy, because it prints
a line for each CPU. The lines are redundant because because usually
all CPUs are updated to the same microcode revision.
All other subsystems have been patched previously to not print
a line for each CPU. Only the microcode driver is left.
Only print an microcode revision update when something changed. This results
in typically only a single line being printed.
v2: Change message to "One or more CPUs"
Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
arch/x86/kernel/cpu/microcode/intel.c | 25 ++++++++++++++++++-------
1 file changed, 18 insertions(+), 7 deletions(-)
diff --git a/arch/x86/kernel/cpu/microcode/intel.c b/arch/x86/kernel/cpu/microcode/intel.c
index cbb3cf0..54f5f6c 100644
--- a/arch/x86/kernel/cpu/microcode/intel.c
+++ b/arch/x86/kernel/cpu/microcode/intel.c
@@ -794,6 +794,7 @@ void reload_ucode_intel(void)
static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
{
+ static struct cpu_signature prev;
struct cpuinfo_x86 *c = &cpu_data(cpu_num);
unsigned int val[2];
@@ -808,8 +809,14 @@ static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
}
csig->rev = c->microcode;
- pr_info("CPU%d sig=0x%x, pf=0x%x, revision=0x%x\n",
- cpu_num, csig->sig, csig->pf, csig->rev);
+
+ /* No extra locking on prev, races are harmless. */
+ if (csig->sig != prev.sig || csig->pf != prev.pf ||
+ csig->rev != prev.rev) {
+ pr_info("One or more CPUs sig=0x%x, pf=0x%x, revision=0x%x\n",
+ csig->sig, csig->pf, csig->rev);
+ prev = *csig;
+ }
return 0;
}
@@ -838,6 +845,7 @@ static int apply_microcode_intel(int cpu)
struct ucode_cpu_info *uci;
struct cpuinfo_x86 *c;
unsigned int val[2];
+ static int prev_rev;
/* We should bind the task to the CPU */
if (WARN_ON(raw_smp_processor_id() != cpu))
@@ -872,11 +880,14 @@ static int apply_microcode_intel(int cpu)
return -1;
}
- pr_info("CPU%d updated to revision 0x%x, date = %04x-%02x-%02x\n",
- cpu, val[1],
- mc->hdr.date & 0xffff,
- mc->hdr.date >> 24,
- (mc->hdr.date >> 16) & 0xff);
+ if (val[1] != prev_rev) {
+ pr_info("One or more CPUs updated to revision 0x%x, date = %04x-%02x-%02x\n",
+ val[1],
+ mc->hdr.date & 0xffff,
+ mc->hdr.date >> 24,
+ (mc->hdr.date >> 16) & 0xff);
+ prev_rev = val[1];
+ }
c = &cpu_data(cpu);
--
2.8.3
[toc] | [next] | [standalone]
| From | Henrique de Moraes Holschuh <hmh@hmh.eng.br> |
|---|---|
| Date | 2016-06-10 02:00 +0200 |
| Subject | Re: [PATCH] x86/microcode/intel: Quieten down microcode updates on large systems |
| Message-ID | <rIi1H-5bB-1@gated-at.bofh.it> |
| In reply to | #1418636 |
On Thu, 09 Jun 2016, Andi Kleen wrote:
> From: Andi Kleen <ak@linux.intel.com>
>
> On large systems the microcode driver is very noisy, because it prints
> a line for each CPU. The lines are redundant because because usually
> all CPUs are updated to the same microcode revision.
>
> All other subsystems have been patched previously to not print
> a line for each CPU. Only the microcode driver is left.
>
> Only print an microcode revision update when something changed. This results
> in typically only a single line being printed.
>
> v2: Change message to "One or more CPUs"
> Signed-off-by: Andi Kleen <ak@linux.intel.com>
Thanks, for WLIW,
Reviewed-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
> ---
> arch/x86/kernel/cpu/microcode/intel.c | 25 ++++++++++++++++++-------
> 1 file changed, 18 insertions(+), 7 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/microcode/intel.c b/arch/x86/kernel/cpu/microcode/intel.c
> index cbb3cf0..54f5f6c 100644
> --- a/arch/x86/kernel/cpu/microcode/intel.c
> +++ b/arch/x86/kernel/cpu/microcode/intel.c
> @@ -794,6 +794,7 @@ void reload_ucode_intel(void)
>
> static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
> {
> + static struct cpu_signature prev;
> struct cpuinfo_x86 *c = &cpu_data(cpu_num);
> unsigned int val[2];
>
> @@ -808,8 +809,14 @@ static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
> }
>
> csig->rev = c->microcode;
> - pr_info("CPU%d sig=0x%x, pf=0x%x, revision=0x%x\n",
> - cpu_num, csig->sig, csig->pf, csig->rev);
> +
> + /* No extra locking on prev, races are harmless. */
> + if (csig->sig != prev.sig || csig->pf != prev.pf ||
> + csig->rev != prev.rev) {
> + pr_info("One or more CPUs sig=0x%x, pf=0x%x, revision=0x%x\n",
> + csig->sig, csig->pf, csig->rev);
> + prev = *csig;
> + }
>
> return 0;
> }
> @@ -838,6 +845,7 @@ static int apply_microcode_intel(int cpu)
> struct ucode_cpu_info *uci;
> struct cpuinfo_x86 *c;
> unsigned int val[2];
> + static int prev_rev;
>
> /* We should bind the task to the CPU */
> if (WARN_ON(raw_smp_processor_id() != cpu))
> @@ -872,11 +880,14 @@ static int apply_microcode_intel(int cpu)
> return -1;
> }
>
> - pr_info("CPU%d updated to revision 0x%x, date = %04x-%02x-%02x\n",
> - cpu, val[1],
> - mc->hdr.date & 0xffff,
> - mc->hdr.date >> 24,
> - (mc->hdr.date >> 16) & 0xff);
> + if (val[1] != prev_rev) {
> + pr_info("One or more CPUs updated to revision 0x%x, date = %04x-%02x-%02x\n",
> + val[1],
> + mc->hdr.date & 0xffff,
> + mc->hdr.date >> 24,
> + (mc->hdr.date >> 16) & 0xff);
> + prev_rev = val[1];
> + }
>
> c = &cpu_data(cpu);
>
--
"One disk to rule them all, One disk to find them. One disk to bring
them all and in the darkness grind them. In the Land of Redmond
where the shadows lie." -- The Silicon Valley Tarot
Henrique Holschuh
[toc] | [prev] | [next] | [standalone]
| From | Borislav Petkov <bp@alien8.de> |
|---|---|
| Date | 2016-06-10 12:10 +0200 |
| Subject | Re: [PATCH] x86/microcode/intel: Quieten down microcode updates on large systems |
| Message-ID | <rIry1-3er-1@gated-at.bofh.it> |
| In reply to | #1418636 |
On Thu, Jun 09, 2016 at 06:41:41AM -0700, Andi Kleen wrote:
> From: Andi Kleen <ak@linux.intel.com>
>
> On large systems the microcode driver is very noisy, because it prints
> a line for each CPU. The lines are redundant because because usually
> all CPUs are updated to the same microcode revision.
>
> All other subsystems have been patched previously to not print
> a line for each CPU. Only the microcode driver is left.
>
> Only print an microcode revision update when something changed. This results
> in typically only a single line being printed.
>
> v2: Change message to "One or more CPUs"
> Signed-off-by: Andi Kleen <ak@linux.intel.com>
> ---
> arch/x86/kernel/cpu/microcode/intel.c | 25 ++++++++++++++++++-------
> 1 file changed, 18 insertions(+), 7 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/microcode/intel.c b/arch/x86/kernel/cpu/microcode/intel.c
> index cbb3cf0..54f5f6c 100644
> --- a/arch/x86/kernel/cpu/microcode/intel.c
> +++ b/arch/x86/kernel/cpu/microcode/intel.c
> @@ -794,6 +794,7 @@ void reload_ucode_intel(void)
>
> static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
> {
> + static struct cpu_signature prev;
> struct cpuinfo_x86 *c = &cpu_data(cpu_num);
> unsigned int val[2];
>
> @@ -808,8 +809,14 @@ static int collect_cpu_info(int cpu_num, struct cpu_signature *csig)
> }
>
> csig->rev = c->microcode;
> - pr_info("CPU%d sig=0x%x, pf=0x%x, revision=0x%x\n",
> - cpu_num, csig->sig, csig->pf, csig->rev);
> +
> + /* No extra locking on prev, races are harmless. */
> + if (csig->sig != prev.sig || csig->pf != prev.pf ||
> + csig->rev != prev.rev) {
> + pr_info("One or more CPUs sig=0x%x, pf=0x%x, revision=0x%x\n",
This "One or more CPUs" is just silly. I've removed it while applying.
This way, there's no mentioning of CPUs and people can check
/proc/cpuinfo for that.
Thanks.
--
Regards/Gruss,
Boris.
ECO tip #101: Trim your mails when you reply.
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web