linux-kernel - Re: [PATCH v2 06/19] PM / devfreq: tegra: Fix missed error checking on devfreq initialization failure

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]

Date:   Tue, 16 Apr 2019 17:29:18 +0300
From:   Dmitry Osipenko <digetx@...il.com>
To:     Chanwoo Choi <cw00.choi@...sung.com>,
        Thierry Reding <thierry.reding@...il.com>,
        Jonathan Hunter <jonathanh@...dia.com>,
        MyungJoo Ham <myungjoo.ham@...sung.com>,
        Kyungmin Park <kyungmin.park@...sung.com>,
        Tomeu Vizoso <tomeu.vizoso@...labora.com>
Cc:     linux-tegra@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-pm@...r.kernel.org
Subject: Re: [PATCH v2 06/19] PM / devfreq: tegra: Fix missed error checking
 on devfreq initialization failure

16.04.2019 5:32, Chanwoo Choi пишет:
> Hi,
> 
> patch6/7/8/9 are for handling of exception handling in probe() function.
> Actually, I'm not sure that there are special reason to split out
> the patches. I think that you can squash patch6/7/8/9 to only one patch.

Indeed, I was rebasing and reordering patches multiple times and looks like there is no reason not to squash these patches now.

> Also, even if patch6/7/8/9 handle the exception handling in probe(),
> the tegra_devfreq_probe() doesn't support the restoring sequence
> when fail happen. I think that if you want to fix the fail case of probe(),
> please add the restoring sequence about following function.
> - clk_disable_unprepare()
> - clk_notifier_unregister()
> - dev_pm_opp_remove()

When all of 6/7/8/9 patches are applied, the clk_notifier_register() becomes the last invocation of the probe function and clk_enable() is kept at the first place of the probe order. Hence the sequence you're suggesting is already incorrect because error-unwinding order usually should be opposite to the probe order. It looks to me that the current final result of these patches is already correct.