linux-kernel - Re: [PATCH V10] i2c: i2c-qcom-geni: Add shutdown callback for i2c

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, 21 May 2021 12:20:51 -0700
From:   Stephen Boyd <swboyd@...omium.org>
To:     rojay@...eaurora.org
Cc:     wsa@...nel.org, dianders@...omium.org,
        saiprakash.ranjan@...eaurora.org, gregkh@...uxfoundation.org,
        mka@...omium.org, skananth@...eaurora.org,
        msavaliy@....qualcomm.com, skakit@...eaurora.org,
        rnayak@...eaurora.org, agross@...nel.org,
        bjorn.andersson@...aro.org, linux-arm-msm@...r.kernel.org,
        linux-i2c@...r.kernel.org, linux-kernel@...r.kernel.org,
        sumit.semwal@...aro.org, linux-media@...r.kernel.org
Subject: Re: [PATCH V10] i2c: i2c-qcom-geni: Add shutdown callback for i2c

Quoting rojay@...eaurora.org (2021-05-21 09:12:02)
>
> Yes, we cannot mark device as suspended/resumed during
> runtime PM operations. Bus will be resumed during i2c
> transfers and before transfer initiation, in __i2c_transfer()
> from i2c-core-base.c there is a check to see whether the device
> is marked as suspended with "__i2c_check_suspended(adap)" call,
> which is "true" in this case and returning from there.
>
> To mark it only suspended during system wide suspend/resume
> transitions, currently our geni i2c driver has only
> system_suspend implemented (geni_i2c_suspend_noirq()) and
> does not have system_resume implemented, which again causes i2c
> transfers to fail during system_resume after system_suspend.

Got it.

>
> Shall I go ahead with marking device suspended during
> shutdown() only?
>

Yes that sounds good. Can you send two patches, one to mark suspend in
the system PM paths and one in the shutdown path? There are various bugs
to fix in other drivers that are trying to use the i2c bus after it has
been suspended and shutdown, but that doesn't block these patches from
moving forward. I'll take a look and see what's going on with
suspend/resume on my board.