From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Rapoport Subject: Re: [PATCH -next] arm64/mm: fix a bogus GFP flag in pgd_alloc() Date: Tue, 11 Jun 2019 15:41:19 +0300 Message-ID: <20190611124118.GA4761@rapoport-lnx> References: <1559656836-24940-1-git-send-email-cai@lca.pw> <20190604142338.GC24467@lakrids.cambridge.arm.com> <20190610114326.GF15979@fuggles.cambridge.arm.com> <1560187575.6132.70.camel@lca.pw> <20190611100348.GB26409@lakrids.cambridge.arm.com> Mime-Version: 1.0 Content-Transfer-Encoding: quoted-printable Return-path: Content-Disposition: inline In-Reply-To: <20190611100348.GB26409@lakrids.cambridge.arm.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="iso-8859-1" To: Mark Rutland Cc: Qian Cai , Will Deacon , akpm@linux-foundation.org, catalin.marinas@arm.com, linux-kernel@vger.kernel.org, mhocko@kernel.org, linux-mm@kvack.org, vdavydov.dev@gmail.com, hannes@cmpxchg.org, cgroups@vger.kernel.org, linux-arm-kernel@lists.infradead.org On Tue, Jun 11, 2019 at 11:03:49AM +0100, Mark Rutland wrote: > On Mon, Jun 10, 2019 at 01:26:15PM -0400, Qian Cai wrote: > > On Mon, 2019-06-10 at 12:43 +0100, Will Deacon wrote: > > > On Tue, Jun 04, 2019 at 03:23:38PM +0100, Mark Rutland wrote: > > > > On Tue, Jun 04, 2019 at 10:00:36AM -0400, Qian Cai wrote: > > > > > The commit "arm64: switch to generic version of pte allocation" > > > > > introduced endless failures during boot like, > > > > >=20 > > > > > kobject_add_internal failed for pgd_cache(285:chronyd.service) (e= rror: > > > > > -2 parent: cgroup) > > > > >=20 > > > > > It turns out __GFP_ACCOUNT is passed to kernel page table allocat= ions > > > > > and then later memcg finds out those don't belong to any cgroup. > > > >=20 > > > > Mike, I understood from [1] that this wasn't expected to be a probl= em, > > > > as the accounting should bypass kernel threads. > > > >=20 > > > > Was that assumption wrong, or is something different happening here? > > > >=20 > > > > >=20 > > > > > backtrace: > > > > > =A0 kobject_add_internal > > > > > =A0 kobject_init_and_add > > > > > =A0 sysfs_slab_add+0x1a8 > > > > > =A0 __kmem_cache_create > > > > > =A0 create_cache > > > > > =A0 memcg_create_kmem_cache > > > > > =A0 memcg_kmem_cache_create_func > > > > > =A0 process_one_work > > > > > =A0 worker_thread > > > > > =A0 kthread > > > > >=20 > > > > > Signed-off-by: Qian Cai > > > > > --- > > > > > =A0arch/arm64/mm/pgd.c | 2 +- > > > > > =A01 file changed, 1 insertion(+), 1 deletion(-) > > > > >=20 > > > > > diff --git a/arch/arm64/mm/pgd.c b/arch/arm64/mm/pgd.c > > > > > index 769516cb6677..53c48f5c8765 100644 > > > > > --- a/arch/arm64/mm/pgd.c > > > > > +++ b/arch/arm64/mm/pgd.c > > > > > @@ -38,7 +38,7 @@ pgd_t *pgd_alloc(struct mm_struct *mm) > > > > > =A0 if (PGD_SIZE =3D=3D PAGE_SIZE) > > > > > =A0 return (pgd_t *)__get_free_page(gfp); > > > > > =A0 else > > > > > - return kmem_cache_alloc(pgd_cache, gfp); > > > > > + return kmem_cache_alloc(pgd_cache, GFP_PGTABLE_KERNEL); > > > >=20 > > > > This is used to allocate PGDs for both user and kernel pagetables (= e.g. > > > > for the efi runtime services), so while this may fix the regression= , I'm > > > > not sure it's the right fix. > > > >=20 > > > > Do we need a separate pgd_alloc_kernel()? > > >=20 > > > So can I take the above for -rc5, or is somebody else working on a di= fferent > > > fix to implement pgd_alloc_kernel()? > >=20 > > The offensive commit "arm64: switch to generic version of pte allocatio= n" is not > > yet in the mainline, but only in the Andrew's tree and linux-next, and = I doubt > > Andrew will push this out any time sooner given it is broken. >=20 > I'd assumed that Mike would respin these patches to implement and use > pgd_alloc_kernel() (or take gfp flags) and the updated patches would > replace these in akpm's tree. >=20 > Mike, could you confirm what your plan is? I'm happy to review/test > updated patches for arm64. Sorry for the delay, I'm mostly offline these days. I wanted to understand first what is the reason for the failure. I've tried to reproduce it with qemu, but I failed to find a bootable configuration that will have PGD_SIZE !=3D PAGE_SIZE :( Qian Cai, can you share what is your environment and the kernel config? =20 > Thanks, > Mark. >=20 --=20 Sincerely yours, Mike.