linux-kernel - Re: [PATCH v4] PM: EM: Fix potential division-by-zero error in em_compute

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: <ea6bfcbc-7665-4808-8188-b444d425b516@kylinos.cn>
Date: Mon, 28 Apr 2025 09:45:48 +0800
From: Yaxiong Tian <tianyaxiong@...inos.cn>
To: "Rafael J. Wysocki" <rafael@...nel.org>
Cc: linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org,
 rafael@...nel.org, Yaxiong Tian <iambestgod@...com>,
 Lukasz Luba <lukasz.luba@....com>
Subject: Re: [PATCH v4] PM: EM: Fix potential division-by-zero error in
 em_compute_costs()


在 2025/4/18 09:26, Yaxiong Tian 写道:

> 在 2025/4/17 21:27, Lukasz Luba 写道:
>> I have run with the GPU EM registered in the boot:
>>
>> -------------------------------------------------------
>> [    2.753333] panfrost ff9a0000.gpu: EM: created perf domain
>> [    2.759863] panfrost ff9a0000.gpu: mali-t860 id 0x860 major 0x2 
>> minor 0x0 status 0x0
>> [    2.768530] panfrost ff9a0000.gpu: features: 00000000,00000407, 
>> issues: 00000000,24040400
>> [    2.777678] panfrost ff9a0000.gpu: Features: L2:0x07120206 
>> Shader:0x00000000 Tiler:0x00000809 Mem:0x1 MMU:0x00002830 AS:0xff JS:0x7
>> [    2.780746] mmc_host mmc2: Bus speed (slot 0) = 148500000Hz (slot 
>> req 150000000Hz, actual 148500000HZ div = 0)
>> [    2.790905] panfrost ff9a0000.gpu: shader_present=0xf l2_present=0x1
>>
>> root@arm:~# cat /sys/kernel/debug/energy_model/ff9a0000.gpu/flags
>> 0x1
>> root@arm:~# grep . /sys/kernel/debug/energy_model/ff9a0000.gpu/ps*/*
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:200000/cost:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:200000/frequency:200000
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:200000/inefficient:1
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:200000/performance:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:200000/power:404250
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:300000/cost:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:300000/frequency:300000
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:300000/inefficient:1
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:300000/performance:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:300000/power:606375
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:400000/cost:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:400000/frequency:400000
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:400000/inefficient:1
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:400000/performance:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:400000/power:808500
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:600000/cost:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:600000/frequency:600000
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:600000/inefficient:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:600000/performance:0
>> /sys/kernel/debug/energy_model/ff9a0000.gpu/ps:600000/power:1505790
>>
>> --------------------------------------------------------
>>
>> The EM for the GPU is not modified during the boot like the CPUs'
>> EM are, thus this code is not triggered. Although, the API is
>> open and in theory the GPU EM can be modified at runtime
>> as well and it will reach that em_compute_costs() issue
>> with 'performance' field having value 0.
>>
>> So this v4 patch would be needed in this case.
>>
>> Please re-send this v4 patch as a completely new message.
>>
>> Thanks for looking at that code path and the fix for potential
>> issue.
>>
>> You can also add my:
>>
>> Reviewed-by: Lukasz Luba <lukasz.luba@....com>
>>
>> Regrds,
>> Lukasz
>
> Got it - patch resent with Reviewed-by.
>
> https://lore.kernel.org/all/tencent_7F99ED4767C1AF7889D0D8AD50F34859CE06@qq.com/ 
>

Hi Rafael:

This bug appears to be stalled. Are there any alternative fixes we could 
consider?