linux-kernel - Re: [PATCH resend] arm64: dts: mediatek: mt8192: Mark scp

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:   Thu, 1 Dec 2022 10:02:24 +0100
From:   AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...labora.com>
To:     Chen-Yu Tsai <wenst@...omium.org>
Cc:     Matthias Brugger <matthias.bgg@...il.com>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Nícolas F . R . A . Prado 
        <nfraprado@...labora.com>, devicetree@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org,
        "allen-kh.cheng" <allen-kh.cheng@...iatek.com>
Subject: Re: [PATCH resend] arm64: dts: mediatek: mt8192: Mark scp_adsp clock
 as broken

Il 01/12/22 09:56, Chen-Yu Tsai ha scritto:
> On Wed, Nov 30, 2022 at 7:10 PM AngeloGioacchino Del Regno
> <angelogioacchino.delregno@...labora.com> wrote:
>>
>> Il 30/11/22 04:17, Chen-Yu Tsai ha scritto:
>>> The scp_adsp clock controller is under the SCP_ADSP power domain. This
>>> power domain is currently not supported nor defined.
>>>
>>> Mark the clock controller as broken for now, to avoid the system from
>>> trying to access it, and causing the CPU or bus to stall.
>>>
>>> Fixes: 5d2b897bc6f5 ("arm64: dts: mediatek: Add mt8192 clock controllers")
>>> Signed-off-by: Chen-Yu Tsai <wenst@...omium.org>
>>
>> ....or we can add the ADSP power domain to actually fix this properly, which looks
>> like being a generally good idea :-)
> 
> Sure, but that and any driver changes have to be backported, or anything
> touching the clocks will still break the system.
> 

I agree.

> There's no reason we can't have both. I think having this one merged and
> backported to stable first, then adding the SCP_ADSP power domain, and tying
> it to the clock controller as a follow up addition works best.
> 
> What do you think?
> 

I think that one reason to not have both is that we'd have to revert this commit
after the SCP_ADSP power domain is added (with the appropriate Fixes tags and/or
Cc stable)...

I'd expect that entire addition to be no more than 3 commits, including the dtsi
one... and if it comes out as I expect, we'd be solving that issue on all of the
affected older versions of the kernel - the right way.

Can we wait for... let's say, a day or two to check how that works, before taking
a final decision on this commit?

Cheers,
Angelo

> ChenYu
> 
>> Allen, can you please take care of that?
>>
>> Thank you,
>> Angelo
>>
>>> ---
>>>    arch/arm64/boot/dts/mediatek/mt8192.dtsi | 2 ++
>>>    1 file changed, 2 insertions(+)
>>>
>>> diff --git a/arch/arm64/boot/dts/mediatek/mt8192.dtsi b/arch/arm64/boot/dts/mediatek/mt8192.dtsi
>>> index 6b20376191a7..ef91941848ae 100644
>>> --- a/arch/arm64/boot/dts/mediatek/mt8192.dtsi
>>> +++ b/arch/arm64/boot/dts/mediatek/mt8192.dtsi
>>> @@ -575,6 +575,8 @@ scp_adsp: clock-controller@...20000 {
>>>                        compatible = "mediatek,mt8192-scp_adsp";
>>>                        reg = <0 0x10720000 0 0x1000>;
>>>                        #clock-cells = <1>;
>>> +                     /* power domain dependency not upstreamed */
>>> +                     status = "broken";
>>>                };
>>>
>>>                uart0: serial@...02000 {
>>>
>>