Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 10 Nov 2010 10:34:08 +0200
From:      Kostik Belousov <kostikbel@gmail.com>
To:        Andriy Gapon <avg@freebsd.org>
Cc:        freebsd-x11@freebsd.org, freebsd-current@freebsd.org
Subject:   Re: radeon_cp_texture: page fault with non-sleepable locks held
Message-ID:  <20101110083408.GC2392@deviant.kiev.zoral.com.ua>
In-Reply-To: <4CDA2B3E.9030402@freebsd.org>
References:  <4CD3B1D2.30003@icyb.net.ua> <4CD7E401.1010206@freebsd.org> <20101108120403.GC2392@deviant.kiev.zoral.com.ua> <4CD9139D.9060302@freebsd.org> <20101109140518.GV2392@deviant.kiev.zoral.com.ua> <4CDA2B3E.9030402@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help

--yOV8wujajMQrgQLM
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Nov 10, 2010 at 07:18:54AM +0200, Andriy Gapon wrote:
> on 09/11/2010 16:05 Kostik Belousov said the following:
> > Easiest would be for DRM to provide wrappers for copyin/copyout that
> > unlock, do operation and lock.
>=20
> I am a little bit worried about this approach in general.
> Driver state may be changed by a process running in parallel while the lo=
ck is
> dropped.  And I don't think that we have any mechanism in DRM to check fo=
r that
> and to restart operations or otherwise account for the state change.  The=
 code
> seems to be written with an assumption that it runs in exclusive mode fro=
m DRM
> ioctl start to its completion.
>=20
> > Where is the reverse order (user map -> drm) ?
>=20
> You mean the following or the opposite?
>=20
> 1st 0xffffff0001b968a0 drmdev (drmdev) @ /usr/src/sys/dev/drm/drm_drv.c:7=
91
> 2nd 0xffffff000a621510 user map (user map) @ /usr/src/sys/vm/vm_map.c:3548
> KDB: stack backtrace:
> db_trace_self_wrapper() at 0xffffffff801b8b3a =3D db_trace_self_wrapper+0=
x2a
> kdb_backtrace() at 0xffffffff803a7a8a =3D kdb_backtrace+0x3a
> _witness_debugger() at 0xffffffff803bd42c =3D _witness_debugger+0x2c
> witness_checkorder() at 0xffffffff803be899 =3D witness_checkorder+0x959
> _sx_slock() at 0xffffffff80378af8 =3D _sx_slock+0x88
> _vm_map_lock_read() at 0xffffffff80510a06 =3D _vm_map_lock_read+0x36
> vm_map_lookup() at 0xffffffff805127d4 =3D vm_map_lookup+0x54
> vm_fault() at 0xffffffff80509819 =3D vm_fault+0xf9
> trap_pfault() at 0xffffffff80545d6f =3D trap_pfault+0x11f
> trap() at 0xffffffff805465f7 =3D trap+0x657
> calltrap() at 0xffffffff80530628 =3D calltrap+0x8
> --- trap 0xc, rip =3D 0xffffffff805440bd, rsp =3D 0xffffff8123fe37f0, rbp=
 =3D
> 0xffffff8123fe3870 ---
> copyin() at 0xffffffff805440bd =3D copyin+0x3d
> radeon_cp_texture() at 0xffffffff8022fbd7 =3D radeon_cp_texture+0x167
> drm_ioctl() at 0xffffffff8020fa38 =3D drm_ioctl+0x318
> devfs_ioctl_f() at 0xffffffff802dd649 =3D devfs_ioctl_f+0x109
> kern_ioctl() at 0xffffffff803c1127 =3D kern_ioctl+0x1f7
> ioctl() at 0xffffffff803c12e8 =3D ioctl+0x168
> syscallenter() at 0xffffffff803b57de =3D syscallenter+0x26e
> syscall() at 0xffffffff80545eb2 =3D syscall+0x42
> Xfast_syscall() at 0xffffffff80530902 =3D Xfast_syscall+0xe2
>=20
> If you meant the opposite order, how can I check it?
> I guess that it could be in a mmap() call for drm cdev.

Explicitely insert the reversed order into the witness array and watch.

--yOV8wujajMQrgQLM
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (FreeBSD)

iEYEARECAAYFAkzaWP8ACgkQC3+MBN1Mb4h4ugCcDp3xrnYTuWHdQZYtdlk1D78A
kPMAnRQ4Kr+WltgIDSyaNiKGb7vExGe3
=VeFf
-----END PGP SIGNATURE-----

--yOV8wujajMQrgQLM--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20101110083408.GC2392>