linux-kernel - Re: [RFC V6 2/2] firmware: arm_scmi: Add quirk to bypass SCP fw bug

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: <Z9mUCmQW7L3yx7qt@hovoldconsulting.com>
Date: Tue, 18 Mar 2025 16:40:58 +0100
From: Johan Hovold <johan@...nel.org>
To: Cristian Marussi <cristian.marussi@....com>
Cc: Dan Carpenter <dan.carpenter@...aro.org>,
	Sibi Sankar <quic_sibis@...cinc.com>, sudeep.holla@....com,
	dmitry.baryshkov@...aro.org, maz@...nel.org,
	linux-kernel@...r.kernel.org, arm-scmi@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org, linux-arm-msm@...r.kernel.org,
	konradybcio@...nel.org
Subject: Re: [RFC V6 2/2] firmware: arm_scmi: Add quirk to bypass SCP fw bug

On Tue, Mar 18, 2025 at 01:29:19PM +0000, Cristian Marussi wrote:
> On Tue, Mar 18, 2025 at 09:16:36AM +0100, Johan Hovold wrote:

> > Have you made any progress on the quirk framework prototyping?
> 
> I have not forgot, tried a few things, but nothing really to post as of
> now...dont wnat to rush either .... I was hoping to push something out at
> the end of this next merge window...
> 
> > Do you need any input from Sibi on the protocol versioning for that?
> 
> No I am fine, I am planning anyway for something generic enough to be
> easy then to plug your own quirks separately...

Sounds good, thanks.

> > We'd really like to enable cpufreq on this platform and ideally in 6.15.
> > I think that should be possible given that we now understand in what
> > ways the firmware is broken and what is needed to handle it even if we
> > still need to decide on how best to implement this.
> 
> v6.15 seems hard/impossible even using the original Sibi patch 
> given the usual upstreaming-timeline of the SCMI stack where everything has
> to be usually reviewed and accepted by rc4/rc5.....so both Sibi initial
> patch and my own babbling were alreaady sort of late.

Yes, sorry, I wasn't referring to Sibi's SCMI patches, but the
devicetree changes needed to enable cpufreq on these platforms (that
will go through the qcom tree).

It may be a little late for that too, but with an understanding of the
problem and a quirk implementation around the corner, we could enable
cpufreq as long as we make sure that Sibi's per-channel FC fix (that
addresses the FC warnings) is not merged without the quirk in place (to
avoid the crashes).

Johan