[RFC/PATCH v2 0/7] locking fixes for cx88

[Jonathan Nieder]

Hi again,

Jonathan Nieder wrote:
> Huber Andreas wrote[1]:

>> Processes that try to open a cx88-blackbird driven MPEG device will hang up.
>
> Here's a possible fix based on a patch by Ben Hutchings and
> corrections from Andi Huber.  Warning: probably full of mistakes (my
> fault) since I'm not familiar with any of this stuff.  Untested.
> Review and testing would be welcome.

A reroll.  As before, the goals are: (1) eliminate deadlock, (2)
eliminate races, (3) introduce some clarity.  The same caveats as
last time apply --- this is only compile-tested.  Thanks again to Andi
for testing the previous series and for other useful feedback.

Patch 1 is meant to protect dev->drvlist against data races.
Since v1, I removed some clutter in the patch itself and clarified the
change description to match.

Patch 2 addresses the original deadlock.  The only changes are the
description and declared authorship of the patch (at Ben's request).

Patch 3 is new.  It fixes the reference count breakage Andi noticed
(another race previously protected against by the BKL).

Patch 4 fixes a data race noticed by Ben (also from his patch).  It's
unchanged.

Patches 5, 6, and 7 are cleanups.

Bugs?  Thoughts?
Jonathan Nieder (7):
  [media] cx88: protect per-device driver list with device lock
  [media] cx88: fix locking of sub-driver operations
  [media] cx88: hold device lock during sub-driver initialization
  [media] cx88: use a mutex to protect cx8802_devlist
  [media] cx88: handle attempts to use unregistered cx88-blackbird
    driver
  [media] cx88: don't use atomic_t for core->mpeg_users
  [media] cx88: don't use atomic_t for core->users

 drivers/media/video/cx88/cx88-blackbird.c |   41 +++++++++++++++-------------
 drivers/media/video/cx88/cx88-dvb.c       |    2 +
 drivers/media/video/cx88/cx88-mpeg.c      |   40 ++++++++++++++++++----------
 drivers/media/video/cx88/cx88-video.c     |    5 ++-
 drivers/media/video/cx88/cx88.h           |   11 +++++--
 5 files changed, 61 insertions(+), 38 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/