[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZLbaBpfaAqigFzIT@ziepe.ca>
Date: Tue, 18 Jul 2023 15:29:26 -0300
From: Jason Gunthorpe <jgg@...pe.ca>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: Alistair Popple <apopple@...dia.com>, ajd@...ux.ibm.com,
catalin.marinas@....com, fbarrat@...ux.ibm.com,
iommu@...ts.linux.dev, jhubbard@...dia.com, kevin.tian@...el.com,
kvm@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, linux-mm@...ck.org,
linuxppc-dev@...ts.ozlabs.org, mpe@...erman.id.au,
nicolinc@...dia.com, npiggin@...il.com, robin.murphy@....com,
seanjc@...gle.com, will@...nel.org, x86@...nel.org,
zhi.wang.linux@...il.com
Subject: Re: [PATCH 3/4] mmu_notifiers: Call arch_invalidate_secondary_tlbs()
when invalidating TLBs
On Tue, Jul 18, 2023 at 11:17:59AM -0700, Andrew Morton wrote:
> On Tue, 18 Jul 2023 17:56:17 +1000 Alistair Popple <apopple@...dia.com> wrote:
>
> > The arch_invalidate_secondary_tlbs() is an architecture specific mmu
> > notifier used to keep the TLB of secondary MMUs such as an IOMMU in
> > sync with the CPU page tables. Currently it is called from separate
> > code paths to the main CPU TLB invalidations. This can lead to a
> > secondary TLB not getting invalidated when required and makes it hard
> > to reason about when exactly the secondary TLB is invalidated.
> >
> > To fix this move the notifier call to the architecture specific TLB
> > maintenance functions for architectures that have secondary MMUs
> > requiring explicit software invalidations.
> >
> > This fixes a SMMU bug on ARM64. On ARM64 PTE permission upgrades
> > require a TLB invalidation. This invalidation is done by the
> > architecutre specific ptep_set_access_flags() which calls
> > flush_tlb_page() if required. However this doesn't call the notifier
> > resulting in infinite faults being generated by devices using the SMMU
> > if it has previously cached a read-only PTE in it's TLB.
>
> This sounds like a pretty serious bug. Can it happen in current
> released kernels? If so, is a -stable backport needed?
There are currently no in-kernel drivers using the IOMMU SVA API, so
the impact for -stable is sort of muted. But it is serious if you are
unlucky to hit it.
Jason
Powered by blists - more mailing lists