[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20190503114817.GB15275@unicorn.suse.cz>
Date: Fri, 3 May 2019 13:48:17 +0200
From: Michal Kubecek <mkubecek@...e.cz>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: linux-kernel@...r.kernel.org, stable@...r.kernel.org,
Andrea Arcangeli <aarcange@...hat.com>,
Jason Gunthorpe <jgg@...lanox.com>,
Leon Romanovsky <leonro@...lanox.com>
Subject: Re: [PATCH 5.0 25/89] RDMA/ucontext: Fix regression with disassociate
On Fri, May 03, 2019 at 01:47:16PM +0200, Michal Kubecek wrote:
> On Tue, Apr 30, 2019 at 01:38:16PM +0200, Greg Kroah-Hartman wrote:
> > From: Jason Gunthorpe <jgg@...lanox.com>
> >
> > commit 67f269b37f9b4d52c5e7f97acea26c0852e9b8a1 upstream.
> >
> > When this code was consolidated the intention was that the VMA would
> > become backed by anonymous zero pages after the zap_vma_pte - however this
> > very subtly relied on setting the vm_ops = NULL and clearing the VM_SHARED
> > bits to transform the VMA into an anonymous VMA. Since the vm_ops was
> > removed this broke.
> >
> > Now userspace gets a SIGBUS if it touches the vma after disassociation.
> >
> > Instead of converting the VMA to anonymous provide a fault handler that
> > puts a zero'd page into the VMA when user-space touches it after
> > disassociation.
> >
> > Cc: stable@...r.kernel.org
> > Suggested-by: Andrea Arcangeli <aarcange@...hat.com>
> > Fixes: 5f9794dc94f5 ("RDMA/ucontext: Add a core API for mmaping driver IO memory")
> > Signed-off-by: Jason Gunthorpe <jgg@...lanox.com>
> > Signed-off-by: Leon Romanovsky <leonro@...lanox.com>
> > Signed-off-by: Jason Gunthorpe <jgg@...lanox.com>
> > Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
> >
> > ---
>
> This commit breaks build on s390 and mips, please pick also commit
> 6a5c5d26c4c6 ("rdma: fix build errors on s390 and MIPS due to bad
> ZERO_PAGE use").
Oops, it's already there as 71/89, I managed to overlook it. Sorry for
the noise.
Michal Kubecek
Powered by blists - more mailing lists