linux-kernel - Re: [PATCH 02/33] iris: vidc: add core functions

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

Message-ID: <a113f6cf-db27-67f1-6996-39fd3912c0c8@kernel.org>
Date:   Mon, 14 Aug 2023 22:04:40 +0200
From:   Krzysztof Kozlowski <krzk@...nel.org>
To:     Dikshita Agarwal <quic_dikshita@...cinc.com>,
        Vikash Garodia <quic_vgarodia@...cinc.com>,
        stanimir.k.varbanov@...il.com, agross@...nel.org,
        andersson@...nel.org, konrad.dybcio@...aro.org, mchehab@...nel.org,
        hans.verkuil@...co.com, linux-kernel@...r.kernel.org,
        linux-media@...r.kernel.org, linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH 02/33] iris: vidc: add core functions

On 14/08/2023 20:54, Dikshita Agarwal wrote:
>>> +
>>> +	core = devm_kzalloc(&pdev->dev, sizeof(struct msm_vidc_core), GFP_KERNEL);
>>> +	if (!core) {
>>> +		d_vpr_e("%s: failed to alloc memory for core\n", __func__);
>>
>> Ooops, this for sure did not pass any checks by tools. Sorry, please run
>> basic checks like coccinelle, smatch, sparse, W=1 builds.
>>
> we ran check patch and smatch on this code but no errors were reported.

coccinelle is missing

> please elaborate what issue you see with above code?
>>> +		return -ENOMEM;

We did quite a lot of cleanups long, long time ago removing all unneeded
error messages from memory allocations failures.

Maybe Your use of custom printks() confuses coccinelle, which is a proof
that your code is here an anti-pattern.

...

>>
> Sure, will remove these custom wrapper for locks and use standard
> mutex_lock/unlock APIs directly.
>>> +	allow = msm_vidc_allow_pm_suspend(core);
>>> +
>>> +	if (allow == MSM_VIDC_IGNORE) {
>>> +		d_vpr_h("%s: pm already suspended\n", __func__);
>>
>> So you have bug in PM runtime code? Runtime PM does not suspend devices
>> twice.
>> core is power collapsed in case of idle state when there is no processing
> happening by hardware, this can also change the core state to power_disable.
> this check is ensuring if the core is already in power disable state due to
> that, then no need to suspend it again.

No, you just re-implemented runtime PM.


Best regards,
Krzysztof