linux-kernel - Re: [PATCH v2 2/5] media: qcom: camss: Convert to per-VFE pointer for power-domain linkages

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]

Message-ID: <9fec3a90-b3dd-4b2e-bb7f-27890ad2b4e0@linaro.org>
Date:   Fri, 27 Oct 2023 10:10:23 +0100
From:   Bryan O'Donoghue <bryan.odonoghue@...aro.org>
To:     Konrad Dybcio <konrad.dybcio@...aro.org>, hverkuil-cisco@...all.nl,
        laurent.pinchart@...asonboard.com, rfoss@...nel.org,
        todor.too@...il.com, andersson@...nel.org, mchehab@...nel.org
Cc:     linux-media@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 2/5] media: qcom: camss: Convert to per-VFE pointer for
 power-domain linkages

On 26/10/2023 21:08, Konrad Dybcio wrote:
>> +    if (vfe->id >= camss->res->vfe_num)
>>           return 0;
> P.S. this seems better suited for some warning, I think

Noo this indicates VFE lite !

power-domains = <VFE_0>,
                 <VFE_1>,
                 <TITAN_TOP>; // the controller pd

vfe-set = <VFE_0>, // has its own PD vfe->id = 0
           <VFE_1>, // has its own PD vfe->id = 1
           <VFE_LITE_N>; // has no PD vfe->id = 2

The basic problem this series fixes is magic indexing.

In the first instance, using named power-domains so that the ordering of 
declaration doesn't matter and we don't have funky code inferring if a 
power-domain belongs to the TOP or not.

Secondly though, which is what the first patch in the series does - is 
as I rebased I realised the VFE/VFE Lite thing was still there.

what vfe->id >= camss->res->vfe_num does is checks to see if the vfe->id 
<= a VFE not a VFE Lite id.

in other words we have yet another magic indexing problem requiring 
VFE_LITE_N to always be declared after VFE.

The solution here is

1. Make the driver support not caring about indexes any more
    This series.
2. Name the power-domains in the various dtsis
    Populating the struct resources in CAMSS to match
    Next series
3. Gate new SoCs to _require_ named pds
    Deprecate the legacy indexing support of 'n' kernel releases
4. Profit

So yeah the check above is I'm sorry to say not an error at all it 
implies VFE Lite...

---
bod