[PATCH] mmc: dw_mmc: take SWIOTLB memory size limitation into account

[Emil Renner Berthing]

Ulf Hansson wrote:
> + Adam, Arnd, Shawn-Lin, sydarn
>
>
> On Sun, 20 Oct 2024 at 16:30, Aurelien Jarno <aurelien at aurel32.net> wrote:
> >
> > The Synopsys DesignWare mmc controller on the JH7110 SoC
> > (dw_mmc-starfive.c driver) is using a 32-bit IDMAC address bus width,
> > and thus requires the use of SWIOTLB.
> >
> > The commit 8396c793ffdf ("mmc: dw_mmc: Fix IDMAC operation with pages
> > bigger than 4K") increased the max_seq_size, even for 4K pages, causing
> > "swiotlb buffer is full" to happen because swiotlb can only handle a
> > memory size up to 256kB only.
> >
> > Fix the issue, by making sure the dw_mmc driver doesn't use segments
> > bigger than what SWIOTLB can handle.
> >
> > Reported-by: Ron Economos <re at w6rz.net>
> > Reported-by: Jing Luo <jing at jing.rocks>
> > Fixes: 8396c793ffdf ("mmc: dw_mmc: Fix IDMAC operation with pages bigger than 4K")
> > Cc: stable at vger.kernel.org
> > Signed-off-by: Aurelien Jarno <aurelien at aurel32.net>
>
> Thanks for working on this!

+1

> Looks like we have managed to mess things
> up. Besides the issue that you have been working on to fix, apparently
> there seems to be another one too [1].
>
> Unfortunately, $subject patch doesn't seem to fix the problem in [1],
> as has been reported by Adam.
>
> I have looped in some more people to this thread, hopefully we agree
> on how this should be fixed properly. Otherwise, I tend to say that we
> should simply revert the offending commit and start over.

Yes, unfortunately this patch also doesn't fix MMC when running 6.12-rc4 on the
StarFive VisionFive V1 (JH7100 SoC).

/Emil