2018-08-02 01:39:56

by Jacek Tomaka

[permalink] [raw]
Subject: [PATCH v4] perf/x86/intel: Add support for MISPREDICT bit on Knights Landing cpus

From: Jacek Tomaka <[email protected]>

Problem: perf did not show branch predicted/mispredicted bit in brstack.

Output of perf -F brstack for profile collected

Before:
0x4fdbcd/0x4fdc03/-/-/-/0
0x45f4c1/0x4fdba0/-/-/-/0
0x45f544/0x45f4bb/-/-/-/0
0x45f555/0x45f53c/-/-/-/0
0x7f66901cc24b/0x45f555/-/-/-/0
0x7f66901cc22e/0x7f66901cc23d/-/-/-/0
0x7f66901cc1ff/0x7f66901cc20f/-/-/-/0
0x7f66901cc1e8/0x7f66901cc1fc/-/-/-/0

After:
0x4fdbcd/0x4fdc03/P/-/-/0
0x45f4c1/0x4fdba0/P/-/-/0
0x45f544/0x45f4bb/P/-/-/0
0x45f555/0x45f53c/P/-/-/0
0x7f66901cc24b/0x45f555/P/-/-/0
0x7f66901cc22e/0x7f66901cc23d/P/-/-/0
0x7f66901cc1ff/0x7f66901cc20f/P/-/-/0
0x7f66901cc1e8/0x7f66901cc1fc/P/-/-/0

Cause:
As mentioned in Software Development Manual vol 3, 17.4.8.1,
IA32_PERF_CAPABILITIES[5:0] indicates the format of the address that is
stored in the LBR stack. Knights Landing reports 1 (LBR_FORMAT_LIP) as
its format. Despite that, registers containing FROM address of the branch,
do have MISPREDICT bit but because of the format indicated in
IA32_PERF_CAPABILITIES[5:0], LBR did not read MISPREDICT bit.

Solution:
Teach LBR about above Knights Landing quirk and make it read MISPREDICT bit.

Signed-off-by: Jacek Tomaka <[email protected]>
---
arch/x86/events/intel/lbr.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index cf372b9055..81fe5047c6 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -1230,4 +1230,8 @@ void intel_pmu_lbr_init_knl(void)

x86_pmu.lbr_sel_mask = LBR_SEL_MASK;
x86_pmu.lbr_sel_map = snb_lbr_sel_map;
+
+ /* Knights Landing does have MISPREDICT bit */
+ if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_LIP)
+ x86_pmu.intel_cap.lbr_format = LBR_FORMAT_EIP_FLAGS;
}
--
2.17.0



2018-08-02 10:10:24

by Thomas Gleixner

[permalink] [raw]
Subject: Re: [PATCH v4] perf/x86/intel: Add support for MISPREDICT bit on Knights Landing cpus

On Thu, 2 Aug 2018, Jacek Tomaka wrote:

The actiual purpose of sending V4 which is identical to V3 is?

>
> Signed-off-by: Jacek Tomaka <[email protected]>
> ---

It's good practice to add a

V3 -> V4: changed foo
V2 -> V3: fixed bla
...

section to patches which have more than one version.

Thanks,

tglx

2018-08-27 11:51:55

by Jacek Tomaka

[permalink] [raw]
Subject: Re: [PATCH v4] perf/x86/intel: Add support for MISPREDICT bit on Knights Landing cpus


> On 2 Aug 2018, at 6:07 pm, Thomas Gleixner <[email protected]> wrote:
>
> The actiual purpose of sending V4 which is identical to V3 is?
>
>>
>> Signed-off-by: Jacek Tomaka <[email protected]>
>> ---

Yes, thanks. I missed it initially, sorry.

> It's good practice to add a
>
> V3 -> V4: changed foo
> V2 -> V3: fixed bla
> ...
>
> section to patches which have more than one version.

Sure. Would you like me to send it for this patch as well?

Regards.
Jacek Tomaka

Subject: [tip:perf/urgent] perf/x86/intel: Add support/quirk for the MISPREDICT bit on Knights Landing CPUs

Commit-ID: 16160c1946b702dcfa95ef63389a56deb2f1c7cb
Gitweb: https://git.kernel.org/tip/16160c1946b702dcfa95ef63389a56deb2f1c7cb
Author: Jacek Tomaka <[email protected]>
AuthorDate: Thu, 2 Aug 2018 09:38:30 +0800
Committer: Ingo Molnar <[email protected]>
CommitDate: Mon, 10 Sep 2018 10:03:01 +0200

perf/x86/intel: Add support/quirk for the MISPREDICT bit on Knights Landing CPUs

Problem: perf did not show branch predicted/mispredicted bit in brstack.

Output of perf -F brstack for profile collected

Before:

0x4fdbcd/0x4fdc03/-/-/-/0
0x45f4c1/0x4fdba0/-/-/-/0
0x45f544/0x45f4bb/-/-/-/0
0x45f555/0x45f53c/-/-/-/0
0x7f66901cc24b/0x45f555/-/-/-/0
0x7f66901cc22e/0x7f66901cc23d/-/-/-/0
0x7f66901cc1ff/0x7f66901cc20f/-/-/-/0
0x7f66901cc1e8/0x7f66901cc1fc/-/-/-/0

After:

0x4fdbcd/0x4fdc03/P/-/-/0
0x45f4c1/0x4fdba0/P/-/-/0
0x45f544/0x45f4bb/P/-/-/0
0x45f555/0x45f53c/P/-/-/0
0x7f66901cc24b/0x45f555/P/-/-/0
0x7f66901cc22e/0x7f66901cc23d/P/-/-/0
0x7f66901cc1ff/0x7f66901cc20f/P/-/-/0
0x7f66901cc1e8/0x7f66901cc1fc/P/-/-/0

Cause:

As mentioned in Software Development Manual vol 3, 17.4.8.1,
IA32_PERF_CAPABILITIES[5:0] indicates the format of the address that is
stored in the LBR stack. Knights Landing reports 1 (LBR_FORMAT_LIP) as
its format. Despite that, registers containing FROM address of the branch,
do have MISPREDICT bit but because of the format indicated in
IA32_PERF_CAPABILITIES[5:0], LBR did not read MISPREDICT bit.

Solution:

Teach LBR about above Knights Landing quirk and make it read MISPREDICT bit.

Signed-off-by: Jacek Tomaka <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
---
arch/x86/events/intel/lbr.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index f3e006bed9a7..c88ed39582a1 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -1272,4 +1272,8 @@ void intel_pmu_lbr_init_knl(void)

x86_pmu.lbr_sel_mask = LBR_SEL_MASK;
x86_pmu.lbr_sel_map = snb_lbr_sel_map;
+
+ /* Knights Landing does have MISPREDICT bit */
+ if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_LIP)
+ x86_pmu.intel_cap.lbr_format = LBR_FORMAT_EIP_FLAGS;
}