linux-kernel - Re: [PATCH] drm: assure aux_dev is nonzero before using it

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:   Fri, 24 May 2019 14:58:29 +0300
From:   Ville Syrjälä <ville.syrjala@...ux.intel.com>
To:     tony camuso <tcamuso@...hat.com>
Cc:     Jani Nikula <jani.nikula@...ux.intel.com>,
        dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
        airlied@...ux.ie, dkwon@...hat.com
Subject: Re: [PATCH] drm: assure aux_dev is nonzero before using it

On Fri, May 24, 2019 at 06:48:32AM -0400, tony camuso wrote:
> On 5/24/19 4:36 AM, Jani Nikula wrote:
> > On Thu, 23 May 2019, tcamuso <tcamuso@...hat.com> wrote:
> >>  From Daniel Kwon <dkwon@...hat.com>
> >>
> >> The system was crashed due to invalid memory access while trying to access
> >> auxiliary device.
> >>
> >> crash> bt
> >> PID: 9863   TASK: ffff89d1bdf11040  CPU: 1   COMMAND: "ipmitool"
> >>   #0 [ffff89cedd7f3868] machine_kexec at ffffffffb0663674
> >>   #1 [ffff89cedd7f38c8] __crash_kexec at ffffffffb071cf62
> >>   #2 [ffff89cedd7f3998] crash_kexec at ffffffffb071d050
> >>   #3 [ffff89cedd7f39b0] oops_end at ffffffffb0d6d758
> >>   #4 [ffff89cedd7f39d8] no_context at ffffffffb0d5bcde
> >>   #5 [ffff89cedd7f3a28] __bad_area_nosemaphore at ffffffffb0d5bd75
> >>   #6 [ffff89cedd7f3a78] bad_area at ffffffffb0d5c085
> >>   #7 [ffff89cedd7f3aa0] __do_page_fault at ffffffffb0d7080c
> >>   #8 [ffff89cedd7f3b10] do_page_fault at ffffffffb0d70905
> >>   #9 [ffff89cedd7f3b40] page_fault at ffffffffb0d6c758
> >>      [exception RIP: drm_dp_aux_dev_get_by_minor+0x3d]
> >>      RIP: ffffffffc0a589bd  RSP: ffff89cedd7f3bf0  RFLAGS: 00010246
> >>      RAX: 0000000000000000  RBX: 0000000000000000  RCX: ffff89cedd7f3fd8
> >>      RDX: 0000000000000000  RSI: 0000000000000000  RDI: ffffffffc0a613e0
> >>      RBP: ffff89cedd7f3bf8   R8: ffff89f1bcbabbd0   R9: 0000000000000000
> >>      R10: ffff89f1be7a1cc0  R11: 0000000000000000  R12: 0000000000000000
> >>      R13: ffff89f1b32a2830  R14: ffff89d18fadfa00  R15: 0000000000000000
> >>      ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
> >>      RIP: 00002b45f0d80d30  RSP: 00007ffc416066a0  RFLAGS: 00010246
> >>      RAX: 0000000000000002  RBX: 000056062e212d80  RCX: 00007ffc41606810
> >>      RDX: 0000000000000000  RSI: 0000000000000002  RDI: 00007ffc41606ec0
> >>      RBP: 0000000000000000   R8: 000056062dfed229   R9: 00002b45f0cdf14d
> >>      R10: 0000000000000002  R11: 0000000000000246  R12: 00007ffc41606ec0
> >>      R13: 00007ffc41606ed0  R14: 00007ffc41606ee0  R15: 0000000000000000
> >>      ORIG_RAX: 0000000000000002  CS: 0033  SS: 002b
> >>
> >> ----------------------------------------------------------------------------
> >>
> >> It was trying to open '/dev/ipmi0', but as no entry in aux_dir, it returned
> >> NULL from 'idr_find()'. This drm_dp_aux_dev_get_by_minor() should have done a
> >> check on this, but had failed to do it.
> > 
> > I think the better question is, *why* does the idr_find() return NULL? I
> > don't think it should, under any circumstances. I fear adding the check
> > here papers over some other problem, taking us further away from the
> > root cause.
> 
> That's a very good question.
> 
> > Also, can you reproduce this on a recent upstream kernel? The aux device
> > nodes were introduced in kernel v4.6. Whatever you reproduced on v3.10
> > is pretty much irrelevant for upstream.
> 
> I will look into this deeper, using the upstream kernel.

Should be trivial to reproduce with mknod. I wonder if we should stick a
test like that into igt actually. Not sure how happy people would be if
igt creates new device nodes...

-- 
Ville Syrjälä
Intel