Re: ahci module throws "Internal error: synchronous external abort: 0000000096000210 [#1] SMP"

[Niklas Cassel <cassel@kernel.org>]

On Tue, Apr 02, 2024 at 08:15:05PM +0200, Ruben Van Boxem wrote:
> Hi Niklas and others,
> 
> Op ma 26 feb 2024 om 21:41 schreef Niklas Cassel <cassel@kernel.org>:
> >
> > Hello Ruben,
> >
> > I would recommend you to test with the latest kernel, v6.8-rc6.
> >
> > I don't know which device tree you are using, but very often custom
> > add on cards/HATs require specific device tree overlays to enable
> > certain regulators etc.
> >
> > I suspect that this is something not related to AHCI, but rather
> > some regulator or similar that is not enabled.
> >
> > How does your kernel command line look on both kernels?
> >
> > Perhaps try to supply:
> > " clk_ignore_unused pd_ignore_unused regulator_ignore_unused"
> > on the kernel command line.
> >
> > (regulator_ignore_unused requires v6.8-rc1 or newer.)
> 
> I tried v6.8-rc6 and it wouldn't boot at the time.
> 
> It seems I ran into an old issue which may have started in version 6.5:
> https://patchwork.kernel.org/project/linux-rockchip/patch/20230509153912.515218-1-vincenzopalazzodev@gmail.com/
> 
> This describes pretty much what I see.
> I understand the solution presented there isn't ideal, but as some
> major ARM distributions are using it, I assume it will do for me for
> now.
> I'll try to use the variant that goes through the device tree flag and
> see if I can get a recent kernel working.
> I'm also unsure if this mailing list is still the relevant place for
> this, as I can't readily find where I found the e-mail address
> anymore.
> Guess I was born in a different age :).
> 
> If you have any more pointers given the above or an alternative
> solution that has more merit, I'm all ears.

rk3399 should boot fine on recent kernels (not sure which kernel I used,
but surely newer than v6.8-rc1).
If it doesn't boot for you, then there is probably something wrong with
your kernel config.

If you still have issues with a recent kernel, I guess you could try the
patch that you linked to, but you could also try:
https://lore.kernel.org/linux-pci/20240330035043.1546087-1-dlemoal@kernel.org/


Kind regards,
Niklas