lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Thu, 13 Jun 2019 17:31:08 -0700
From:   Ira Weiny <ira.weiny@...el.com>
To:     Jason Gunthorpe <jgg@...lanox.com>
Cc:     Dan Williams <dan.j.williams@...el.com>,
        linux-nvdimm <linux-nvdimm@...ts.01.org>,
        "nouveau@...ts.freedesktop.org" <nouveau@...ts.freedesktop.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Maling list - DRI developers 
        <dri-devel@...ts.freedesktop.org>, Linux MM <linux-mm@...ck.org>,
        Jérôme Glisse <jglisse@...hat.com>,
        Ben Skeggs <bskeggs@...hat.com>,
        "linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Christoph Hellwig <hch@....de>
Subject: Re: dev_pagemap related cleanups

On Thu, Jun 13, 2019 at 08:40:46PM +0000, Jason Gunthorpe wrote:
> On Thu, Jun 13, 2019 at 11:27:39AM -0700, Dan Williams wrote:
> > On Thu, Jun 13, 2019 at 2:43 AM Christoph Hellwig <hch@....de> wrote:
> > >
> > > Hi Dan, Jérôme and Jason,
> > >
> > > below is a series that cleans up the dev_pagemap interface so that
> > > it is more easily usable, which removes the need to wrap it in hmm
> > > and thus allowing to kill a lot of code
> > >
> > > Diffstat:
> > >
> > >  22 files changed, 245 insertions(+), 802 deletions(-)
> > 
> > Hooray!
> > 
> > > Git tree:
> > >
> > >     git://git.infradead.org/users/hch/misc.git hmm-devmem-cleanup
> > 
> > I just realized this collides with the dev_pagemap release rework in
> > Andrew's tree (commit ids below are from next.git and are not stable)
> > 
> > 4422ee8476f0 mm/devm_memremap_pages: fix final page put race
> > 771f0714d0dc PCI/P2PDMA: track pgmap references per resource, not globally
> > af37085de906 lib/genalloc: introduce chunk owners
> > e0047ff8aa77 PCI/P2PDMA: fix the gen_pool_add_virt() failure path
> > 0315d47d6ae9 mm/devm_memremap_pages: introduce devm_memunmap_pages
> > 216475c7eaa8 drivers/base/devres: introduce devm_release_action()
> > 
> > CONFLICT (content): Merge conflict in tools/testing/nvdimm/test/iomap.c
> > CONFLICT (content): Merge conflict in mm/hmm.c
> > CONFLICT (content): Merge conflict in kernel/memremap.c
> > CONFLICT (content): Merge conflict in include/linux/memremap.h
> > CONFLICT (content): Merge conflict in drivers/pci/p2pdma.c
> > CONFLICT (content): Merge conflict in drivers/nvdimm/pmem.c
> > CONFLICT (content): Merge conflict in drivers/dax/device.c
> > CONFLICT (content): Merge conflict in drivers/dax/dax-private.h
> > 
> > Perhaps we should pull those out and resend them through hmm.git?
> 
> It could be done - but how bad is the conflict resolution?
> 
> I'd be more comfortable to take a PR from you to merge into hmm.git,
> rather than raw patches, then apply CH's series on top. I think.
> 
> That way if something goes wrong you can send your PR to Linus
> directly.
> 
> > It also turns out the nvdimm unit tests crash with this signature on
> > that branch where base v5.2-rc3 passes:
> > 
> >     BUG: kernel NULL pointer dereference, address: 0000000000000008
> >     [..]
> >     CPU: 15 PID: 1414 Comm: lt-libndctl Tainted: G           OE
> > 5.2.0-rc3+ #3399
> >     Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 0.0.0 02/06/2015
> >     RIP: 0010:percpu_ref_kill_and_confirm+0x1e/0x180
> >     [..]
> >     Call Trace:
> >      release_nodes+0x234/0x280
> >      device_release_driver_internal+0xe8/0x1b0
> >      bus_remove_device+0xf2/0x160
> >      device_del+0x166/0x370
> >      unregister_dev_dax+0x23/0x50
> >      release_nodes+0x234/0x280
> >      device_release_driver_internal+0xe8/0x1b0
> >      unbind_store+0x94/0x120
> >      kernfs_fop_write+0xf0/0x1a0
> >      vfs_write+0xb7/0x1b0
> >      ksys_write+0x5c/0xd0
> >      do_syscall_64+0x60/0x240
> 
> Too bad the trace didn't say which devm cleanup triggered it.. Did
> dev_pagemap_percpu_exit get called with a NULL pgmap->ref ?

I would guess something like that.  I did not fully wrap my head around the ref
counting there but I don't think the patch is correct.  See my review.

Ira

> 
> Jason
> _______________________________________________
> Linux-nvdimm mailing list
> Linux-nvdimm@...ts.01.org
> https://lists.01.org/mailman/listinfo/linux-nvdimm

Powered by blists - more mailing lists