[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210611213443.ira5gc65jlafz7pu@gupta-dev2.localdomain>
Date: Fri, 11 Jun 2021 14:34:43 -0700
From: Pawan Gupta <pawan.kumar.gupta@...ux.intel.com>
To: Borislav Petkov <bp@...en8.de>
Cc: Thomas Gleixner <tglx@...utronix.de>,
Jonathan Corbet <corbet@....net>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...hat.com>,
Namhyung Kim <namhyung@...nel.org>, x86@...nel.org,
"H. Peter Anvin" <hpa@...or.com>,
"Paul E. McKenney" <paulmck@...nel.org>,
Randy Dunlap <rdunlap@...radead.org>,
Andrew Morton <akpm@...ux-foundation.org>,
"Maciej W. Rozycki" <macro@...am.me.uk>,
Viresh Kumar <viresh.kumar@...aro.org>,
Vlastimil Babka <vbabka@...e.cz>,
Tony Luck <tony.luck@...el.com>,
Paolo Bonzini <pbonzini@...hat.com>,
Sean Christopherson <seanjc@...gle.com>,
Kyung Min Park <kyung.min.park@...el.com>,
Fenghua Yu <fenghua.yu@...el.com>,
Ricardo Neri <ricardo.neri-calderon@...ux.intel.com>,
Tom Lendacky <thomas.lendacky@....com>,
Juergen Gross <jgross@...e.com>,
Krish Sadhukhan <krish.sadhukhan@...cle.com>,
Kan Liang <kan.liang@...ux.intel.com>,
Joerg Roedel <jroedel@...e.de>,
Victor Ding <victording@...gle.com>,
Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
Brijesh Singh <brijesh.singh@....com>,
Dave Hansen <dave.hansen@...el.com>,
Mike Rapoport <rppt@...nel.org>,
Anthony Steinhauser <asteinhauser@...gle.com>,
Anand K Mistry <amistry@...gle.com>,
Andi Kleen <ak@...ux.intel.com>,
Miguel Ojeda <ojeda@...nel.org>, Joe Perches <joe@...ches.com>,
linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-perf-users@...r.kernel.org
Subject: Re: [PATCH 2/4] perf/x86/intel: Do not deploy workaround when TSX is
deprecated
On 11.06.2021 09:50, Borislav Petkov wrote:
>On Wed, Jun 09, 2021 at 02:12:38PM -0700, Pawan Gupta wrote:
>> Earlier workaround added by commit 400816f60c54 ("perf/x86/intel:
>> Implement support for TSX Force Abort") for perf counter interactions
>> [1] are not required on some client systems which received a microcode
>> update that deprecates TSX.
>>
>> Bypass the perf workaround when such microcode is enumerated.
>>
>> [1] Performance Monitoring Impact of IntelĀ® Transactional Synchronization Extension Memory
>> http://cdrdv2.intel.com/v1/dl/getContent/604224
>>
>> Signed-off-by: Pawan Gupta <pawan.kumar.gupta@...ux.intel.com>
>> Reviewed-by: Andi Kleen <ak@...ux.intel.com>
>> Reviewed-by: Tony Luck <tony.luck@...el.com>
>> Tested-by: Neelima Krishnan <neelima.krishnan@...el.com>
>> ---
>> arch/x86/events/intel/core.c | 22 ++++++++++++++++++----
>> 1 file changed, 18 insertions(+), 4 deletions(-)
>>
>> diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
>> index e28892270c58..b5953e1e59a2 100644
>> --- a/arch/x86/events/intel/core.c
>> +++ b/arch/x86/events/intel/core.c
>> @@ -6016,10 +6016,24 @@ __init int intel_pmu_init(void)
>> intel_pmu_pebs_data_source_skl(pmem);
>>
>> if (boot_cpu_has(X86_FEATURE_TSX_FORCE_ABORT)) {
>> - x86_pmu.flags |= PMU_FL_TFA;
>> - x86_pmu.get_event_constraints = tfa_get_event_constraints;
>> - x86_pmu.enable_all = intel_tfa_pmu_enable_all;
>> - x86_pmu.commit_scheduling = intel_tfa_commit_scheduling;
>> + u64 msr;
>> +
>> + rdmsrl(MSR_TSX_FORCE_ABORT, msr);
>> + /* Systems that enumerate CPUID.RTM_ALWAYS_ABORT or
>> + * support MSR_TSX_FORCE_ABORT[SDV_ENABLE_RTM] bit have
>> + * TSX deprecated by default. TSX force abort hooks are
>> + * not required on these systems.
>
>So if they're not required, why aren't you simply disabling the force
>abort "workaround" by clearing the feature flag?
Feature flag also enumerates MSR_TSX_FORCE_ABORT, which is still present
after the microcode update. Patch 3/4 in this series clears the TSX
CPUID bits using MSR_TSX_FORCE_ABORT. So we do need the feature flag
X86_FEATURE_TSX_FORCE_ABORT.
>
> if (boot_cpu_has(X86_FEATURE_TSX_FORCE_ABORT)) {
> if (boot_cpu_has(X86_FEATURE_RTM_ALWAYS_ABORT))
> setup_clear_cpu_cap(X86_FEATURE_TSX_FORCE_ABORT);
> }
>
>so that it doesn't get enabled in the first place?
Feature flag is needed as explained above.
Thanks,
Pawan
Powered by blists - more mailing lists