linux-kernel - Re: [External] Re: [PATCH v3 3/5] perf/core: Don't need event_filter_match in merge_sched

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <40a084b1-b9ca-77a0-2b1d-56b9c2ec660f@bytedance.com>
Date:   Fri, 25 Mar 2022 23:45:06 +0800
From:   Chengming Zhou <zhouchengming@...edance.com>
To:     "Liang, Kan" <kan.liang@...ux.intel.com>, peterz@...radead.org,
        mingo@...hat.com, acme@...nel.org, mark.rutland@....com,
        alexander.shishkin@...ux.intel.com, jolsa@...nel.org,
        namhyung@...nel.org, eranian@...gle.com
Cc:     linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
        duanxiongchun@...edance.com, songmuchun@...edance.com
Subject: Re: [External] Re: [PATCH v3 3/5] perf/core: Don't need
 event_filter_match in merge_sched_in()

On 2022/3/25 23:11, Liang, Kan wrote:
> 
> 
> On 3/24/2022 11:53 PM, Chengming Zhou wrote:
>> There is one obselete comment in perf_cgroup_switch(), since
>> we don't use event_filter_match() when event_sched_out().
>>
>> Then found we needn't to use event_filter_match() in
>> merge_sched_in() too. Because now we use the perf_event groups
>> RB-tree to get the exact matched perf_events, don't need to
>> go through the event_filter_match() to check if matched again.
>>
>> Signed-off-by: Chengming Zhou <zhouchengming@...edance.com>
>> ---
>>   kernel/events/core.c | 6 ++----
>>   1 file changed, 2 insertions(+), 4 deletions(-)
>>
>> diff --git a/kernel/events/core.c b/kernel/events/core.c
>> index dd985c77bc37..225d408deb1a 100644
>> --- a/kernel/events/core.c
>> +++ b/kernel/events/core.c
>> @@ -856,7 +856,8 @@ static void perf_cgroup_switch(struct task_struct *task, int mode)
>>               cpu_ctx_sched_out(cpuctx, EVENT_ALL);
>>               /*
>>                * must not be done before ctxswout due
>> -             * to event_filter_match() in event_sched_out()
>> +             * to update_cgrp_time_from_cpuctx() in
>> +             * ctx_sched_out()
>>                */
>>               cpuctx->cgrp = NULL;
>>           }
>> @@ -3804,9 +3805,6 @@ static int merge_sched_in(struct perf_event *event, void *data)
>>       if (event->state <= PERF_EVENT_STATE_OFF)
>>           return 0;
>>   -    if (!event_filter_match(event))
>> -        return 0;
>> -
> 
> Both X86 and Arm implemented PMU specific filter_match callback for the hybrid system. I think the check is still required at least for the hybrid system.
> 

Ah, I ignored this point, will fix it.

Thanks.

> 
> Thanks,
> Kan
>>       if (group_can_go_on(event, cpuctx, *can_add_hw)) {
>>           if (!group_sched_in(event, cpuctx, ctx))
>>               list_add_tail(&event->active_list, get_event_list(event));