Re: 5.10 backport request: ASoC: Intel: sof_rt5682: shrink platform_id names below 20 characters

[Greg KH <gregkh@linuxfoundation.org>]

On Sun, Aug 31, 2025 at 08:25:07PM +0200, Michał Górny wrote:
> On Fri, 2025-08-29 at 15:55 +0200, Greg KH wrote:
> > On Fri, Aug 29, 2025 at 09:09:33AM +0200, Michał Górny wrote:
> > > Hello,
> > > 
> > > On Fri, 2025-08-29 at 08:03 +0200, Michał Górny wrote:
> > > > Hello,
> > > > 
> > > > I would like to request backporting the following patch to 5.10 series:
> > > > 
> > > >   590cfb082837cc6c0c595adf1711330197c86a58
> > > >   ASoC: Intel: sof_rt5682: shrink platform_id names below 20 characters
> > > > 
> > > > The patch seems to be already present in 5.15 and newer branches, and
> > > > its lack seems to be causing out-of-bounds read.  I've hit it in the
> > > > wild while trying to install 5.10.241 on i686:
> > > > 
> > > >   sh /var/tmp/portage/sys-kernel/gentoo-kernel-5.10.241/work/linux-5.10/scripts/depmod.sh depmod 5.10.241-gentoo-dist
> > > > depmod: FATAL: Module index: bad character '�'=0x80 - only 7-bit ASCII is supported:
> > > > platform:jsl_rt5682_max98360ax�
> > > > 
> > > 
> > > I'm sorry, I should've waited for my test results first.  Looks like
> > > this patch alone is insufficient.  Looking at 5.15 stable branch, I see
> > > that we probably need:
> > > 
> > >     ASoC: Intel: bxt_da7219_max98357a: shrink platform_id below 20 characters
> > >     ASoC: Intel: sof_rt5682: shrink platform_id names below 20 characters
> > >     ASoC: Intel: glk_rt5682_max98357a: shrink platform_id below 20 characters
> > >     ASoC: Intel: kbl_da7219_max98357a: shrink platform_id below 20 characters
> > >     ASoC: Intel: sof_da7219_max98373: shrink platform_id below 20 characters
> > >     ASoC: Intel: sof_da7219_mx98360a: fail to initialize soundcard
> > >     ASoC: Intel: Fix platform ID matching
> > > 
> > > Unless I'm mistaken, the firt series are part of the merge commit
> > > 98c69fcc9f5902b0c340acdbbfa365464efc52d2.  The followup fixes are:
> > > 
> > >     0f32d9eb38c13c32895b5bf695eac639cee02d6c
> > >     f4eeaed04e861b95f1f2c911263f2fcaa959c078
> > > 
> > > I didn't find anything else that seemed obviously elevant, but I didn't
> > > dug deep.  With these backports, I can build 5.10.241 fine -- but I
> > > don't have any hardware to test these drivers.
> > 
> > So what exact commits are needed and in what order?  Can you just send
> > tested backports to us so that we know we got it right?
> 
> Would it be okay if I cherry-picked them from 5.15.y rather than from
> master?

We need the commit id from Linus's tree in them, not the id from 5.15.y,
for obvious reasons.

thanks,

greg k-h