linux-kernel - Re: [RESEND] clk: qcom: rcg: Update rcg configuration before enabling it

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: <20230717212309.GA4176673@hu-bjorande-lv.qualcomm.com>
Date:   Mon, 17 Jul 2023 14:23:09 -0700
From:   Bjorn Andersson <quic_bjorande@...cinc.com>
To:     Satya Priya Kakitapalli <quic_skakitap@...cinc.com>
CC:     Konrad Dybcio <konrad.dybcio@...aro.org>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd <sboyd@...nel.org>,
        Andy Gross <agross@...nel.org>,
        Bjorn Andersson <andersson@...nel.org>,
        <linux-arm-msm@...r.kernel.org>, <linux-clk@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>,
        Jagadeesh Kona <quic_jkona@...cinc.com>,
        Taniya Das <quic_tdas@...cinc.com>
Subject: Re: [RESEND] clk: qcom: rcg: Update rcg configuration before
 enabling it

On Wed, Jul 12, 2023 at 07:18:12AM +0530, Satya Priya Kakitapalli wrote:
> From: Taniya Das <quic_tdas@...cinc.com>
> 
> If rcg is in disabled state when clk_rcg2_shared_set_rate is called, the
> new configuration is written to the configuration register but it won't be
> effective in h/w yet because update bit won't be set if rcg is in disabled
> state.

Does this take commit '703db1f5da1e ("clk: qcom: rcg2: Cache CFG
register updates for parked RCGs")', which was merged in v5.19, into
consideration?

> Since the new configuration is not yet updated in h/w, dirty bit of
> configuration register will be set in such case. Clear the dirty bit and
> update the rcg to proper new configuration by setting the update bit before
> enabling the rcg.
> 

For a shared rcg2, which was updated while disabled, updates will be
carried in the "parked_cfg" variable and the RCG_CFG will be stale so
invoking update_config() should lead to exactly the problem you describe
fixing here.

Perhaps I'm missing something here, can you please confirm that this has
been validated on a recent upstream kernel?

> Signed-off-by: Taniya Das <quic_tdas@...cinc.com>
> ---
> Resending this patch as there is no review for 2 months.
> 

Thanks for bumping the discussion.

Regards,
Bjorn