Re: [PATCH v6 2/8] hugetlb: restructure pool allocations

[Konrad Dybcio]

On 10/10/23 23:45, Mike Kravetz wrote:
> On 10/10/23 23:30, Konrad Dybcio wrote:
> > On 10/10/23 02:07, Andrew Morton wrote:
> > > On Mon, 9 Oct 2023 08:15:13 -0700 Mike Kravetz <mike.kravetz@xxxxxxxxxx> wrote:
> > >
> > > > > This should also be empty and a noop.
> > > > >
> > > > > Is it possible that the misaligned kernel image could make these lists
> > > > > appear as non-empty?
> > > >
> > > > Actually, just saw this:
> > > >
> > > > https://lore.kernel.org/linux-mm/20231009145605.2150897-1-usama.arif@xxxxxxxxxxxxx/
> > > >
> > > > Will take a look, although as mentioned above prep_and_add_bootmem_folios on
> > > > an empty list should be a noop.
> > >
> > > Konrad, are you able to test Usama's patch?  Thanks.
> > I legitimately spent a sad amount of time trying to regain access to the
> > remote board farm. Previously I could hit the bug on SM8550, but now I can't
> > do it on SM8450, SM8350 and SM8250 (previous gens), with the same config.. I
> > have no idea when I'll be able to get access to SM8550 again.
> >
> > I did test it on the QCM6490 Fairphone 5 that I initially reported this on,
> > and neither booting next-20231010 (with your patchset applied) nor adding
> > the below patch on top of it seems to work. I can't get serial output from
> > this device though to find out what it's unhappy about :/
>
> Sorry for causing you to spend so much time on this.
No worries, that was my explanation for why it took me so long to 
respond again..

> As mentioned in the reply to Usama's patch, the root cause seems to be
> the locking.  So, the real change to test is the locking changes in
> that thread; not Usama's patch.
Ack

Konrad