linux-kernel - Re: [PATCH 2/6] clk: qcom: ipq5018: mark XO clock as critical

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: <31008c59-afed-405a-a330-e42d89a62790@oss.qualcomm.com>
Date: Tue, 6 May 2025 02:59:56 +0200
From: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
To: George Moussalem <george.moussalem@...look.com>,
        Konrad Dybcio <konrad.dybcio@....qualcomm.com>,
        Bjorn Andersson <andersson@...nel.org>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd
 <sboyd@...nel.org>, Rob Herring <robh@...nel.org>,
        Krzysztof Kozlowski <krzk+dt@...nel.org>,
        Conor Dooley
 <conor+dt@...nel.org>, Luo Jie <quic_luoj@...cinc.com>,
        Lee Jones <lee@...nel.org>, Konrad Dybcio <konradybcio@...nel.org>
Cc: linux-arm-msm@...r.kernel.org, linux-clk@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/6] clk: qcom: ipq5018: mark XO clock as critical

On 5/4/25 8:59 AM, George Moussalem wrote:
> 
> 
> On 5/2/25 16:45, George Moussalem wrote:
>>
>>
>> On 5/2/25 14:29, Konrad Dybcio wrote:
>>> On 5/2/25 12:15 PM, George Moussalem via B4 Relay wrote:
>>>> From: George Moussalem <george.moussalem@...look.com>
>>>>
>>>> The XO clock must not be disabled, so let's add the CLK_IS_CRITICAL
>>>> flag to avoid the kernel trying to disable the XO clock (when parenting
>>>> it under the CMN PLL reference clock), else the kernel will panic and
>>>> the following message will appear in the kernel logs:
>>>
>>> Remove the struct definition for this clock (and the assignment in
>>> blah_blah_clks[]) and replace it with:
>>>
>>> qcom_branch_set_clk_en(regmap, 0x30030); /* GCC_XO_CLK */
>>
>> understood, thanks for the quick turnaround!
> 
> Tested it, but then then the issue is still there. This time fixable by setting the CLK_IS_CRITICAL flag on gcc_xo_clk_src. I was looking at removing the struct for gcc_xo_clk_src too and use qcom_branch_set_clk_en, but there are clocks that refer to the gcc_xo_clk_src as their parent. I'm a bit hesitant to tinker with the GCC driver without access to the datasheet. The downstream driver actually has the CLK_IS_CRITICAL flag set too on gcc_xo_clk as initially proposed in this patch:
> 
> https://git.codelinaro.org/clo/qsdk/oss/kernel/linux-ipq-5.4/-/blob/NHSS.QSDK.12.5/drivers/clk/qcom/gcc-ipq5018.c#L1457
> 
> Are you okay with this suggested approach?

Since turning off XO means the CPU (and nothing else on the soc for that
matter) clock will not tick, just unregister the RCG along with it

you can remove the .parent_hws (dont forget .num_parents along with it)
from the affected clocks, this is effectively cosmetic

Konrad