Commit Graph

17 Commits

Author SHA1 Message Date
Sjoerd Simons
8ca2c80b17 [media] uvcvideo: Remove buffers from the queues when freeing
When freeing memory for the video buffers also remove them from the
irq & main queues.

This fixes an oops when doing the following:

open ("/dev/video", ..)
VIDIOC_REQBUFS
VIDIOC_QBUF
VIDIOC_REQBUFS
close ()

As the second VIDIOC_REQBUFS will cause the list entries of the buffers
to be cleared while they still hang around on the main and irc queues

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: stable@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2011-06-30 23:18:52 -03:00
Bob Liu
7296944759 [media] uvcvideo: Add support for NOMMU arch
Add support to uvc driver for NOMMU arch including add function
uvc_queue_get_unmapped_area() and make some changes in uvc_queue_mmap().
So that uvc camera can be used on nommu arch like blackfin.

Signed-off-by: Bob Liu <lliubbo@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2011-05-20 12:16:45 -03:00
Laurent Pinchart
4aa275975b [media] uvcvideo: Move mmap() handler to uvc_queue.c
The mmap() implementation belongs to the video buffers queue, move it
there.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-12-01 20:10:23 -02:00
Laurent Pinchart
8e815e1763 [media] uvcvideo: Move mutex lock/unlock inside uvc_free_buffers
Callers outside uvc_queue.c should not be forced to lock/unlock the
queue mutex manually. Move the mutex operations inside
uvc_free_buffers().

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-12-01 20:10:22 -02:00
Laurent Pinchart
650b95feee [media] uvcvideo: Generate discontinuous sequence numbers when frames are lost
Increase the sequence number of the v4l2_buffer structure regardless of
any buffer states, so that discontinuous sequence numbers allow
applications to detect lost video frames.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-10-21 01:18:20 -02:00
Laurent Pinchart
11fc5baf1d [media] uvcvideo: Update e-mail address and copyright notices
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-10-21 01:18:18 -02:00
Laurent Pinchart
9bde9f263e V4L/DVB: uvcvideo: Drop corrupted compressed frames
Corrupted video frames are dropped by default by the driver for
uncompressed formats. Data corruption is not less problematic for
compressed formats, so frame drop should be enabled by default for those
formats as well.

Mark buffers as faulty when an isochronous packet loss is detected for
any format, or when the buffer length doesn't match the image size for
uncompressed formats. Drop erroneous buffers regardless of whether the
format is compressed or uncompressed.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-08-08 23:43:01 -03:00
Laurent Pinchart
ed3de6015a V4L/DVB: uvcvideo: Use POLLOUT and POLLWRNORM for output devices
The V4L2 specification requires drivers to use the write events in the
file operations poll handler for output devices. The uvcvideo driver
erroneously used read events for all devices. Fix this.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-05-18 00:52:52 -03:00
Laurent Pinchart
310fe52461 V4L/DVB (13827): uvcvideo: Switch to a monotonic clock for V4L2 buffers timestamps
The realtime clock provided by do_gettimeofday() is affected by time
jumps caused by NTP or DST. Furthermore, preliminary investigation
showed that SMP systems the realtime clock is based on the CPU TSC,
and those could get slightly out of sync, resulting in jitter in the
timestamps depending on which processor handles the USB interrupts.

Instead of the realtime clock, use a monotonic high resolution clock to
timestamp the buffer. As this could in theory introduce a regression
with some userspace applications expecting a realtime clock timestamp,
add a module parameter to switch back to the realtime clock.

Thanks to Paulo Assis for pointing out and investigating the issue.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-02-26 15:10:24 -03:00
Laurent Pinchart
d7c0d43997 V4L/DVB (13831): uvcvideo: Fix oops caused by a race condition in buffer dequeuing
Buffers were marked as done before being removed from the IRQ queue. If
a userspace application dequeued and requeued the buffer fast enough
during that time window, the buffer could end up being deleted twice,
generating an oops in interrupt context.

Add a new state, UVC_BUF_STATE_READY, to mark buffers as ready for reuse
but not yet removed from the queue, and transition to UVC_BUF_STATE_DONE
only when the buffer is removed from the queue.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2010-01-17 11:31:35 -02:00
Laurent Pinchart
23ff6043ad V4L/DVB (11945): uvcvideo: Don't accept to change the format when buffers are allocated.
Setting a new frame format or size will likely change the buffer size required
to store a complete video frame. To avoid a buffer overflow, don't allow
VIDIOC_S_FMT calls when video buffers are already allocated.

Signed-off-by: Laurent Pinchart <laurent.pinchart@skynet.be>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2009-06-16 19:07:30 -03:00
Laurent Pinchart
2c2d264bb9 V4L/DVB (10197): uvcvideo: Whitespace and comments cleanup, copyright updates.
Signed-off-by: Laurent Pinchart <laurent.pinchart@skynet.be>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2009-01-29 08:35:36 -02:00
Huang Weiyi
16e68d7cb8 V4L/DVB (10193): removed unused #include <version.h>'s
Removed unused #include <version.h>'s in files below,
  drivers/media/video/cs5345.c
  drivers/media/video/pwc/pwc-if.c
  drivers/media/video/saa717x.c
  drivers/media/video/upd64031a.c
  drivers/media/video/upd64083.c
  drivers/media/video/uvc/uvc_ctrl.c
  drivers/media/video/uvc/uvc_driver.c
  drivers/media/video/uvc/uvc_queue.c
  drivers/media/video/uvc/uvc_video.c
  drivers/media/video/uvc/uvc_status.c

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Acked-by: Laurent Pinchart <laurent.pinchart@skynet.be>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2009-01-29 08:35:36 -02:00
Laurent Pinchart
ff924203c9 V4L/DVB (10104): uvcvideo: Add support for video output devices
Extend the range of supported UVC devices by allowing video output devices
matching the following structure:

TT_STREAMING -> VC_PROCESSING_UNIT -> VC_EXTENSION_UNIT{0,n} -> OTT_*

Video output devices are reported with the V4L2_CAP_VIDEO_OUTPUT capability
flag and are subject to the same restrictions as video input devices.

Signed-off-by: Laurent Pinchart <laurent.pinchart@skynet.be>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2008-12-30 09:40:32 -02:00
Andrea Righi
27ac792ca0 PAGE_ALIGN(): correctly handle 64-bit values on 32-bit architectures
On 32-bit architectures PAGE_ALIGN() truncates 64-bit values to the 32-bit
boundary. For example:

	u64 val = PAGE_ALIGN(size);

always returns a value < 4GB even if size is greater than 4GB.

The problem resides in PAGE_MASK definition (from include/asm-x86/page.h for
example):

#define PAGE_SHIFT      12
#define PAGE_SIZE       (_AC(1,UL) << PAGE_SHIFT)
#define PAGE_MASK       (~(PAGE_SIZE-1))
...
#define PAGE_ALIGN(addr)       (((addr)+PAGE_SIZE-1)&PAGE_MASK)

The "~" is performed on a 32-bit value, so everything in "and" with
PAGE_MASK greater than 4GB will be truncated to the 32-bit boundary.
Using the ALIGN() macro seems to be the right way, because it uses
typeof(addr) for the mask.

Also move the PAGE_ALIGN() definitions out of include/asm-*/page.h in
include/linux/mm.h.

See also lkml discussion: http://lkml.org/lkml/2008/6/11/237

[akpm@linux-foundation.org: fix drivers/media/video/uvc/uvc_queue.c]
[akpm@linux-foundation.org: fix v850]
[akpm@linux-foundation.org: fix powerpc]
[akpm@linux-foundation.org: fix arm]
[akpm@linux-foundation.org: fix mips]
[akpm@linux-foundation.org: fix drivers/media/video/pvrusb2/pvrusb2-dvb.c]
[akpm@linux-foundation.org: fix drivers/mtd/maps/uclinux.c]
[akpm@linux-foundation.org: fix powerpc]
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-24 10:47:21 -07:00
Hans Verkuil
f87086e302 v4l-dvb: remove legacy checks to allow support for kernels < 2.6.10
Also remove some blank lines that were used to split compat code at -devel
tree.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2008-07-20 07:17:52 -03:00
Laurent Pinchart
c0efd23292 V4L/DVB (8145a): USB Video Class driver
This driver supports video input devices compliant with the USB Video Class
specification. This means lots of currently manufactured webcams, and probably
most of the future ones.

Signed-off-by: Laurent Pinchart <laurent.pinchart@skynet.be>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2008-06-30 17:31:13 -03:00