From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from Chamillionaire.breakpoint.cc (Chamillionaire.breakpoint.cc [IPv6:2001:4d88:1ffa:82:880:aa0:9009:64ae]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3qZC6q6jcZzDq5s for ; Wed, 30 Mar 2016 00:58:43 +1100 (AEDT) Date: Tue, 29 Mar 2016 15:40:24 +0200 From: Sebastian Andrzej Siewior To: "Aneesh Kumar K.V" , Benjamin Herrenschmidt , Scott Wood , Michael Ellerman Cc: Sebastian Andrzej Siewior , Paul Mackerras , Christoph Lameter , linuxppc-dev@lists.ozlabs.org Subject: Re: [PATCH v2] powerpc: mm: fixup preempt undefflow with huge pages Message-ID: <20160329134024.GA25057@breakpoint.cc> References: <20160307135506.GA12595@linutronix.de> <877fheti8i.fsf@linux.vnet.ibm.com> <1457394104.2943.69.camel@kernel.crashing.org> <20160308090356.GA21842@linutronix.de> <87bn6n1m4f.fsf@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 In-Reply-To: <87bn6n1m4f.fsf@linux.vnet.ibm.com> List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 2016-03-10 01:04:24 [+0530], Aneesh Kumar K.V wrote: > Sebastian Andrzej Siewior writes: *ping* http://patchwork.ozlabs.org/patch/593943/ > > [ text/plain ] > > hugepd_free() used __get_cpu_var() once. Nothing ensured that the code > > accessing the variable did not migrate from one CPU to another and soon > > this was noticed by Tiejun Chen in 94b09d755462 ("powerpc/hugetlb: > > Replace __get_cpu_var with get_cpu_var"). So we had it fixed. > > > > Christoph Lameter was doing his __get_cpu_var() replaces and forgot > > PowerPC. Then he noticed this and sent his fixed up batch again which > > got applied as 69111bac42f5 ("powerpc: Replace __get_cpu_var uses"). > > > > The careful reader will noticed one little detail: get_cpu_var() got > > replaced with this_cpu_ptr(). So now we have a put_cpu_var() which does > > a preempt_enable() and nothing that does preempt_disable() so we > > underflow the preempt counter. > > >=20 > Reviewed-by: Aneesh Kumar K.V >=20 > > Cc: Benjamin Herrenschmidt > > Cc: Christoph Lameter > > Cc: Michael Ellerman > > Cc: > > Signed-off-by: Sebastian Andrzej Siewior > > --- > > v1=E2=80=A6v2: - use get_cpu_var() instead of get_cpu_ptr() > > - correct indentation of put_cpu_var() > > > > arch/powerpc/mm/hugetlbpage.c | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpag= e.c > > index 744e24bcb85c..4a811ca7ac9d 100644 > > --- a/arch/powerpc/mm/hugetlbpage.c > > +++ b/arch/powerpc/mm/hugetlbpage.c > > @@ -414,13 +414,13 @@ static void hugepd_free(struct mmu_gather *tlb, v= oid *hugepte) > > { > > struct hugepd_freelist **batchp; > > =20 > > - batchp =3D this_cpu_ptr(&hugepd_freelist_cur); > > + batchp =3D &get_cpu_var(hugepd_freelist_cur); > > =20 > > if (atomic_read(&tlb->mm->mm_users) < 2 || > > cpumask_equal(mm_cpumask(tlb->mm), > > cpumask_of(smp_processor_id()))) { > > kmem_cache_free(hugepte_cache, hugepte); > > - put_cpu_var(hugepd_freelist_cur); > > + put_cpu_var(hugepd_freelist_cur); > > return; > > } > > =20 Sebastian