Re: [PATCH] Revert "mm: fix MAX_FOLIO_ORDER on powerpc configs with hugetlb"

["David Hildenbrand (Red Hat)" <david@kernel.org>]

On 12/4/25 22:57, Andrew Morton wrote:
> On Wed,  3 Dec 2025 19:33:56 -0700 Shuah Khan <skhan@linuxfoundation.org> wrote:
> 
>> This reverts commit 39231e8d6ba7f794b566fd91ebd88c0834a23b98.
>>
>> Enabling HAVE_GIGANTIC_FOLIOS broke kernel build and git clone on two
>> systems. git fetch-pack fails when cloning large repos and make hangs
>> or errors out of Makefile.build with Error: 139. These failures are
>> random with git clone failing after fetching 1% of the objects, and
>> make hangs while compiling random files.
>>
>> The blow is is one of the git clone failures:
>>
>> git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git linux_6.19
>> Cloning into 'linux_6.19'...
>> remote: Enumerating objects: 11173575, done.
>> remote: Counting objects: 100% (785/785), done.
>> remote: Compressing objects: 100% (373/373), done.
>> remote: Total 11173575 (delta 534), reused 505 (delta 411), pack-reused 11172790 (from 1)
>> Receiving objects: 100% (11173575/11173575), 3.00 GiB | 7.08 MiB/s, done.
>> Resolving deltas: 100% (9195212/9195212), done.
>> fatal: did not receive expected object 0002003e951b5057c16de5a39140abcbf6e44e50
>> fatal: fetch-pack: invalid index-pack output
> 
> 39231e8d6ba7 simply shuffles ifdefs and Kconfig items, so I assume it
> exposed a pre-existing bug.
> 
> Reverting 39231e8d6ba7 will re-hide that bug.
> 
> And that isn't a bad thing.  If we re-hide the bug in 6.18.x and in
> mainline then that relieves the people who are hitting this and it
> takes the pressure off David, Mike and yourself to get the underlying
> bug fixed in a hurry.
> 
> So I think I'll queue this as a hotfix, plan to send it Linuswards in a
> couple of days.
> 
> Or Linus may choose to apply it directly or to do a local revert of
> 39231e8d6ba7.  But I don't see how a local revert will get communicated
> to the 6.18.x maintainers.
> 
> David, Linus, opinions please?

I have so far no indication that this patch here would change anything 
relevant to the problem we are seeing, all it does is changing 
MAX_FOLIO_ORDER that does not affect any logic we would really care 
about here (safety checks and snapshot_page()).

Can we please wait a bit so we have confirmation that it's not a 
leftover from the huge-zero-folio thingy or something different?

As Mike reports, he found ways to reproduce something similar even with 
39231e8d6ba7 reverted.

-- 
Cheers

David