[linux-next:master] BUILD REGRESSION a54df7622717a40ddec95fd98086aff8ba7839a6

[Mike Kravetz]

On 01/24/23 12:28, Andrew Morton wrote:
> On Wed, 25 Jan 2023 00:37:05 +0800 kernel test robot <lkp at intel.com> wrote:
> 
> > tree/branch: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> > branch HEAD: a54df7622717a40ddec95fd98086aff8ba7839a6  Add linux-next specific files for 20230124
> > 
> > Error/Warning: (recently discovered and may have been fixed)
> > 
> > ERROR: modpost: "devm_platform_ioremap_resource" [drivers/dma/fsl-edma.ko] undefined!
> > ERROR: modpost: "devm_platform_ioremap_resource" [drivers/dma/idma64.ko] undefined!
> > drivers/gpu/drm/amd/amdgpu/../display/dc/link/link_dp_training.c:1585:38: warning: variable 'result' set but not used [-Wunused-but-set-variable]
> > 
> > Unverified Error/Warning (likely false positive, please contact us if interested):
> > 
> > ...
> >
> > mm/hugetlb.c:3100 alloc_hugetlb_folio() error: uninitialized symbol 'h_cg'.
> 
> 	hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h), h_cg, folio);
> 
> The warning looks to be bogus.  I guess we could put a "= NULL" in
> there to keep the compiler quiet?

I took a brief look at that yesterday.

h_cg will be initialized, but not in this routine.  Address of h_cg is passed
to hugetlb_cgroup_charge_cgroup which will ensure h_cg is initialized.  TBH,
I am not sure how the compiler would know this as hugetlb_cgroup_charge_cgroup
is in another file.

IMO, this was not introduced by Sid's changes.  However, I can not explain
why it is showing up now.  Neither can I reproduce with my compiler, otherwise
I would have sent a patch yesterday.

I will send a patch to initialize to NULL to keep compiler quiet.
-- 
Mike Kravetz