linux-kernel - Re: [PATCH v2] sched/cputime: Fix using smp_processor

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]

Date:   Fri, 7 Jul 2017 20:17:13 +0800
From:   Wanpeng Li <kernellwp@...il.com>
To:     Frederic Weisbecker <fweisbec@...il.com>
Cc:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Ingo Molnar <mingo@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Wanpeng Li <wanpeng.li@...mail.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Luiz Capitulino <lcapitulino@...hat.com>,
        Rik van Riel <riel@...hat.com>
Subject: Re: [PATCH v2] sched/cputime: Fix using smp_processor_id() in preemptible

2017-07-07 20:01 GMT+08:00 Frederic Weisbecker <fweisbec@...il.com>:
> On Fri, Jul 07, 2017 at 02:08:25AM -0700, Wanpeng Li wrote:
>> From: Wanpeng Li <wanpeng.li@...mail.com>
>>
>>  BUG: using smp_processor_id() in preemptible [00000000] code: 99-trinity/181
>>  caller is debug_smp_processor_id+0x17/0x19
>>  CPU: 0 PID: 181 Comm: 99-trinity Not tainted 4.12.0-01059-g2a42eb9 #1
>>  Call Trace:
>>   dump_stack+0x82/0xb8
>>   check_preemption_disabled+0xd1/0xe3
>>   debug_smp_processor_id+0x17/0x19
>>   vtime_delta+0xd/0x2c
>>   task_cputime+0x89/0xdb
>>   thread_group_cputime+0x11b/0x1ed
>>   thread_group_cputime_adjusted+0x1f/0x47
>>   wait_consider_task+0x2a9/0xaf9
>>   ? lock_acquire+0x97/0xa4
>>   do_wait+0xdf/0x1f4
>>   SYSC_wait4+0x8e/0xb5
>>   ? list_add+0x34/0x34
>>   SyS_wait4+0x9/0xb
>>   do_syscall_64+0x70/0x82
>>   entry_SYSCALL64_slow_path+0x25/0x25
>>
>> This patch fixes it by replacing sched_clock_cpu() in vtime_delta() by
>> local_clock() for effectively raw_smp_processor_id().
>
> That's also broken because task_cputime() can be called from a different CPU than
> where the target task is running on, even though there shouldn't be practical effect

Agreed.

> as the clock must be stable but still the code would be confusing.
>
> No I think you can still use sched_clock(), just make sure you also use it on
> arch_vtime_task_switch() and vtime_init_idle().

Is it acceptable to you, Peterz? :)

Regards,
Wanpeng Li