From: Stephane Eranian <eranian@google.com>
To: linux-kernel@vger.kernel.org
Cc: peterz@infradead.org, andi@firstfloor.org, mingo@elte.hu,
ming.m.lin@intel.com
Subject: [PATCH] perf/x86: check ucode before disabling PEBS on SandyBridge
Date: Thu, 7 Jun 2012 09:15:31 +0200 [thread overview]
Message-ID: <20120607071531.GA4849@quad> (raw)
This patch checks the microcode version before disabling
PEBS on SandyBridge model 42 (desktop, mobile), and 45 (SNB-EP).
PEBS was disabled for both models due to an erratum.
A workaround is implemented by micro-code 0x28. This patch checks
the microcode version and disables PEBS support if version < 0x28.
The check is done each time a PEBS event is created and NOT at boot
time because the micro-code update may only be done after the kernel
has booted.
Go to downloadcenter.intel.com to download microcode updates.
Need microcode update 6/6/2012 or later.
Signed-off-by: Stephane Eranian <eranian@google.com>
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 5ec146c..9c905bd 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -1394,6 +1394,27 @@ static void intel_pebs_aliases_snb(struct perf_event *event)
}
}
+static int check_pebs_quirks(void)
+{
+ int uversion = cpu_data(smp_processor_id()).microcode;
+ int model = cpu_data(smp_processor_id()).x86_model;
+
+ /* do not have PEBS to begin with */
+ if (!x86_pmu.pebs)
+ return 0;
+
+ /*
+ * check ucode version for SNB, SNB-EP
+ */
+ if ((model == 42 || model == 45) && uversion < 0x28) {
+ pr_warn("SandyBridge PEBS unavailable due to CPU erratum, "
+ " update microcode (was 0x%x, needs at least 0x28).\n",
+ uversion);
+ return -ENOTSUPP;
+ }
+ return 0;
+}
+
static int intel_pmu_hw_config(struct perf_event *event)
{
int ret = x86_pmu_hw_config(event);
@@ -1401,8 +1422,14 @@ static int intel_pmu_hw_config(struct perf_event *event)
if (ret)
return ret;
- if (event->attr.precise_ip && x86_pmu.pebs_aliases)
- x86_pmu.pebs_aliases(event);
+ if (event->attr.precise_ip) {
+
+ if (check_pebs_quirks())
+ return -ENOTSUPP;
+
+ if (x86_pmu.pebs_aliases)
+ x86_pmu.pebs_aliases(event);
+ }
if (intel_pmu_needs_lbr_smpl(event)) {
ret = intel_pmu_setup_lbr_filter(event);
@@ -1714,13 +1741,6 @@ static __init void intel_clovertown_quirk(void)
x86_pmu.pebs_constraints = NULL;
}
-static __init void intel_sandybridge_quirk(void)
-{
- pr_warn("PEBS disabled due to CPU errata\n");
- x86_pmu.pebs = 0;
- x86_pmu.pebs_constraints = NULL;
-}
-
static const struct { int id; char *name; } intel_arch_events_map[] __initconst = {
{ PERF_COUNT_HW_CPU_CYCLES, "cpu cycles" },
{ PERF_COUNT_HW_INSTRUCTIONS, "instructions" },
@@ -1914,7 +1934,6 @@ __init int intel_pmu_init(void)
case 42: /* SandyBridge */
case 45: /* SandyBridge, "Romely-EP" */
- x86_add_quirk(intel_sandybridge_quirk);
case 58: /* IvyBridge */
memcpy(hw_cache_event_ids, snb_hw_cache_event_ids,
sizeof(hw_cache_event_ids));
next reply other threads:[~2012-06-07 7:15 UTC|newest]
Thread overview: 107+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-06-07 7:15 Stephane Eranian [this message]
2012-06-07 10:18 ` [PATCH] perf/x86: check ucode before disabling PEBS on SandyBridge Peter Zijlstra
2012-06-07 10:35 ` Stephane Eranian
2012-06-07 10:45 ` Peter Zijlstra
2012-06-07 10:48 ` Stephane Eranian
2012-06-07 11:15 ` Peter Zijlstra
2012-06-08 9:35 ` Ingo Molnar
2012-06-08 10:00 ` Peter Zijlstra
2012-06-08 10:03 ` Stephane Eranian
2012-06-08 13:26 ` Peter Zijlstra
2012-06-08 13:51 ` Borislav Petkov
2012-06-08 13:54 ` Peter Zijlstra
2012-06-08 14:15 ` Borislav Petkov
2012-06-08 14:20 ` Peter Zijlstra
2012-06-08 14:36 ` Borislav Petkov
2012-06-08 14:45 ` Peter Zijlstra
2012-06-08 15:09 ` Borislav Petkov
2012-06-08 13:56 ` Peter Zijlstra
2012-06-08 14:03 ` Borislav Petkov
2012-06-08 13:59 ` Peter Zijlstra
2012-06-08 14:03 ` Stephane Eranian
2012-06-08 14:17 ` Borislav Petkov
2012-06-08 16:02 ` H. Peter Anvin
2012-06-08 21:02 ` Henrique de Moraes Holschuh
2012-06-08 21:16 ` Borislav Petkov
2012-06-08 14:07 ` Stephane Eranian
2012-06-08 14:13 ` Peter Zijlstra
2012-06-08 14:19 ` Borislav Petkov
2012-06-08 14:23 ` Peter Zijlstra
2012-06-08 14:26 ` Stephane Eranian
2012-06-08 14:25 ` Peter Zijlstra
2012-06-08 14:26 ` Stephane Eranian
2012-06-12 8:14 ` Stephane Eranian
2012-06-08 16:28 ` Stephane Eranian
2012-06-08 18:49 ` Peter Zijlstra
2012-06-08 16:50 ` Andi Kleen
2012-06-08 18:05 ` Borislav Petkov
2012-06-08 18:52 ` Peter Zijlstra
2012-06-08 20:38 ` Borislav Petkov
2012-06-12 17:07 ` Robert Richter
2012-06-12 17:09 ` Stephane Eranian
2012-06-12 17:13 ` Peter Zijlstra
2012-06-12 17:17 ` Borislav Petkov
2012-06-12 17:18 ` Peter Zijlstra
2012-06-12 17:23 ` Borislav Petkov
2012-06-12 17:26 ` Peter Zijlstra
2012-06-12 17:35 ` Borislav Petkov
2012-06-12 17:40 ` Peter Zijlstra
2012-06-12 18:04 ` Andi Kleen
2012-06-12 20:10 ` Borislav Petkov
2012-06-13 1:04 ` Henrique de Moraes Holschuh
2012-06-13 6:51 ` Borislav Petkov
2012-06-13 12:36 ` Henrique de Moraes Holschuh
2012-06-13 16:11 ` Borislav Petkov
2012-06-13 21:41 ` Henrique de Moraes Holschuh
2012-06-15 12:37 ` Borislav Petkov
2012-06-15 12:42 ` Peter Zijlstra
2012-06-15 12:52 ` Borislav Petkov
2012-06-15 16:52 ` [PATCH] x86, microcode: Make reload interface per system Borislav Petkov
2012-06-15 17:16 ` Peter Zijlstra
2012-06-15 17:49 ` Borislav Petkov
2012-06-19 2:46 ` Henrique de Moraes Holschuh
2012-06-19 3:31 ` H. Peter Anvin
2012-06-19 5:11 ` Borislav Petkov
2012-06-19 8:17 ` Peter Zijlstra
2012-06-19 10:22 ` Borislav Petkov
2012-06-19 10:26 ` Peter Zijlstra
2012-06-19 15:06 ` Borislav Petkov
2012-06-19 15:57 ` H. Peter Anvin
2012-06-19 18:22 ` Henrique de Moraes Holschuh
2012-06-19 20:22 ` H. Peter Anvin
2012-06-20 23:46 ` Henrique de Moraes Holschuh
2012-06-20 23:49 ` H. Peter Anvin
2012-06-21 0:06 ` Henrique de Moraes Holschuh
2012-06-21 0:18 ` H. Peter Anvin
2012-06-21 2:33 ` Henrique de Moraes Holschuh
2012-06-19 5:03 ` Borislav Petkov
2012-06-19 18:57 ` Henrique de Moraes Holschuh
2012-06-19 22:19 ` Borislav Petkov
2012-06-19 2:42 ` [PATCH] perf/x86: check ucode before disabling PEBS on SandyBridge Henrique de Moraes Holschuh
2012-06-12 17:23 ` Robert Richter
2012-06-12 17:15 ` Borislav Petkov
2012-06-07 13:28 ` Andi Kleen
2012-06-07 13:27 ` Andi Kleen
2012-06-12 18:35 ` H. Peter Anvin
2012-06-12 19:07 ` Borislav Petkov
2012-06-12 19:33 ` Peter Zijlstra
2012-06-12 19:35 ` Andi Kleen
2012-06-12 19:49 ` Borislav Petkov
2012-06-12 20:28 ` H. Peter Anvin
2012-06-12 20:37 ` Borislav Petkov
2012-06-12 20:42 ` H. Peter Anvin
2012-06-12 20:56 ` Borislav Petkov
2012-06-12 20:58 ` H. Peter Anvin
2012-06-12 21:04 ` Borislav Petkov
2012-06-13 8:38 ` Peter Zijlstra
2012-06-13 14:00 ` H. Peter Anvin
2012-06-13 15:37 ` Peter Zijlstra
2012-06-13 15:46 ` H. Peter Anvin
2012-06-13 12:39 ` Peter Zijlstra
2012-06-13 13:22 ` H. Peter Anvin
2012-06-13 14:01 ` Henrique de Moraes Holschuh
2012-06-13 8:32 ` Peter Zijlstra
2012-06-13 13:59 ` H. Peter Anvin
2012-06-13 15:32 ` Peter Zijlstra
2012-06-13 15:38 ` H. Peter Anvin
2012-06-12 20:27 ` H. Peter Anvin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20120607071531.GA4849@quad \
--to=eranian@google.com \
--cc=andi@firstfloor.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ming.m.lin@intel.com \
--cc=mingo@elte.hu \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).