[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <92ecca66-e0ca-42d8-a54a-f3e5d016e42a@linux.intel.com>
Date: Thu, 27 Feb 2025 14:29:36 +0800
From: "Mi, Dapeng" <dapeng1.mi@...ux.intel.com>
To: Andi Kleen <ak@...ux.intel.com>, Peter Zijlstra <peterz@...radead.org>
Cc: Ingo Molnar <mingo@...hat.com>, Arnaldo Carvalho de Melo
<acme@...nel.org>, Namhyung Kim <namhyung@...nel.org>,
Ian Rogers <irogers@...gle.com>, Adrian Hunter <adrian.hunter@...el.com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Kan Liang <kan.liang@...ux.intel.com>, Eranian Stephane
<eranian@...gle.com>, linux-kernel@...r.kernel.org,
linux-perf-users@...r.kernel.org, Dapeng Mi <dapeng1.mi@...el.com>
Subject: Re: [Patch v2 15/24] perf/x86/intel: Add SSP register support for
arch-PEBS
On 2/26/2025 4:44 AM, Andi Kleen wrote:
> On Tue, Feb 25, 2025 at 12:54:50PM +0100, Peter Zijlstra wrote:
>> On Tue, Feb 18, 2025 at 03:28:09PM +0000, Dapeng Mi wrote:
>>
>>> @@ -651,6 +651,16 @@ int x86_pmu_hw_config(struct perf_event *event)
>>> return -EINVAL;
>>> }
>>>
>>> + /* sample_regs_user never support SSP register. */
>>> + if (unlikely(event->attr.sample_regs_user & BIT_ULL(PERF_REG_X86_SSP)))
>>> + return -EINVAL;
>> We can easily enough read user SSP, no?
> Not for multi record PEBS.
>
> Also technically it may not be precise.
Yes, If need to support to capture user space SSP, then PEBS has to fall
back to only capture single record.
Andi, you said "it may not be precise", could you please give more details?
I didn't get this. Thanks.
>
> -andi
>
Powered by blists - more mailing lists