linux.git/drivers/media, branch v3.9-rc6

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid

2013-04-04T15:40:45+00:00

Pull HID fixes from Jiri Kosina:

 - Workaround for device ID conflict between Masterkit MA901 usb radio
   device and Atmel V-USB devices, to avoid regressions from older
   kernels, by Alexey Klimov

 - fix for possible race during input device registration in magicmouse
   driver, by Benjamin Tissoires

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: magicmouse: fix race between input_register() and probe()
  media: radio-ma901: return ENODEV in probe if usb_device doesn't match
  HID: fix Masterkit MA901 hid quirks

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

2013-04-03T17:48:22+00:00

Pull s390 fixes from Martin Schwidefsky:
 "Just a bunch of bugfixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/mm: provide emtpy check_pgt_cache() function
  s390/uaccess: fix page table walk
  s390/3270: fix minor_start issue
  s390/uaccess: fix clear_user_pt()
  s390/scm_blk: fix error return code in scm_blk_init()
  s390/scm_block: fix printk format string
  drivers/Kconfig: add several missing GENERIC_HARDIRQS dependencies

media: radio-ma901: return ENODEV in probe if usb_device doesn't match

2013-03-29T14:13:40+00:00

Masterkit MA901 usb radio device shares USB ID with Atmel V-USB devices.
This patch adds additional checks in usb_ma901radio_probe() and if
product or manufacturer doesn't match we return -ENODEV and don't
continue. This allows hid drivers to handle not MA901 device.

Signed-off-by: Alexey Klimov 
Acked-by: Mauro Carvalho Chehab 
Signed-off-by: Jiri Kosina

[media] [REGRESSION] bt8xx: Fix too large height in cropcap

2013-03-26T11:37:00+00:00

Since commit a1fd287780c8e91fed4957b30c757b0c93021162:
"[media] bttv-driver: fix two warnings"
cropcap.defrect.height and cropcap.bounds.height for the PAL entry are 32
resp 30 pixels too large, if a userspace app (ie xawtv) actually tries to use
the full advertised height, the resulting image is broken in ways only a
screenshot can describe.
The cause of this is the fix for this warning:
drivers/media/pci/bt8xx/bttv-driver.c:308:3: warning: initialized field overwritten [-Woverride-init]
In this chunk of the commit:
@@ -301,11 +301,10 @@ const struct bttv_tvnorm bttv_tvnorms[] = {
                        /* totalwidth */ 1135,
                        /* sqwidth */ 944,
                        /* vdelay */ 0x20,
-                       /* sheight */ 576,
-                       /* videostart0 */ 23)
                /* bt878 (and bt848?) can capture another
                   line below active video. */
-               .cropcap.bounds.height = (576 + 2) + 0x20 - 2,
+                       /* sheight */ (576 + 2) + 0x20 - 2,
+                       /* videostart0 */ 23)
        },{
                .v4l2_id        = V4L2_STD_NTSC_M | V4L2_STD_NTSC_M_KR,
                .name           = "NTSC",
Which replaces the overriding of cropcap.bounds.height initialization outside
of the CROPCAP macro (which also initializes it), with passing a
different sheight value to the CROPCAP macro.
There are 2 problems with this warning fix:
1) The sheight value is used twice in the CROPCAP macro, and the old code
   only changed one resulting value.
2) The old code increased the .cropcap.bounds.height value (and did not
   touch the .cropcap.defrect.height value at all) by 2, where as the fixed
   code increases it by 32, as the fixed code passes (576 + 2) + 0x20 - 2
   to the CROPCAP macro, but the + 0x20 - 2 is already done by the macro so
   now is done twice for .cropcap.bounds.height, and also is applied to
   .cropcap.defrect.height where it should not be applied at all.
This patch fixes this by adding an extraheight parameter to the CROPCAP entry
and using it for the PAL entry.

Cc: stable@kernel.org	# For Kernel 3.8
Signed-off-by: Hans de Goede 
Signed-off-by: Mauro Carvalho Chehab

[media] fix compilation with both V4L2 and I2C as 'm'

2013-03-25T18:09:38+00:00

When config options are:
	CONFIG_VIDEO_DEV=y
	CONFIG_VIDEO_V4L2=m
	CONFIG_I2C=m

Compilation breaks, as reported by:
	https://bugzilla.kernel.org/show_bug.cgi?id=55681

Before changeset 7b34be71db533f3e0cf93d53cf62d036cdb5418a,
no compilation errors occurred. However, the I2C code there at
v4l2-device was incorrectly disabled.

Signed-off-by: Mauro Carvalho Chehab

[media] m5mols: Fix bug in stream on handler

2013-03-21T21:25:09+00:00

Due to improper condition check streaming start for some pixel
formats was prevent and the s_stream just reatuned -EINVAL.
This fixes regression introduced in commit 5565a2ad47cdd8e697
[media] m5mols: Protect driver data with a mutex.

Signed-off-by: Andrzej Hajda 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-fimc: Do not attempt to disable not enabled media pipeline

2013-03-21T21:24:57+00:00

This fixes following warnings when all links are being disconnected:
[   20.080000] WARNING: at drivers/media/platform/s5p-fimc/fimc-mdevice.c:1269 __fimc_md_set_camclk+0x208/0x20c()
[   20.090000] Modules linked in:
[   20.095000] [] (unwind_backtrace+0x0/0x13c) from [] (warn_slowpath_common+0x54/0x64)
[   20.105000] [] (warn_slowpath_common+0x54/0x64) from [] (warn_slowpath_null+0x1c/0x24)
[   20.115000] [] (warn_slowpath_null+0x1c/0x24) from [] (__fimc_md_set_camclk+0x208/0x20c)
[   20.125000] [] (__fimc_md_set_camclk+0x208/0x20c) from [] (__fimc_pipeline_close+0x38/0x48)
[   20.135000] [] (__fimc_pipeline_close+0x38/0x48) from [] (fimc_md_link_notify+0x10c/0x198)
[   20.145000] [] (fimc_md_link_notify+0x10c/0x198) from [] (__media_entity_setup_link+0x1c0/0x1e8)
[   20.155000] [] (__media_entity_setup_link+0x1c0/0x1e8) from [] (media_device_ioctl+0x2c0/0x41c)
[   20.165000] [] (media_device_ioctl+0x2c0/0x41c) from [] (media_ioctl+0x30/0x34)
[   20.175000] [] (media_ioctl+0x30/0x34) from [] (do_vfs_ioctl+0x84/0x5e8)
[   20.185000] [] (do_vfs_ioctl+0x84/0x5e8) from [] (sys_ioctl+0x3c/0x60)
[   20.190000] [] (sys_ioctl+0x3c/0x60) from [] (ret_fast_syscall+0x0/0x30)

Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Kyungmin Park 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-mfc: Fix encoder control 15 issue

2013-03-21T21:24:38+00:00

mfc-encoder is not working in the latest kernel giving the
erorr "Adding control (15) failed". Adding the missing step
parameter in this control to fix the issue.

Signed-off-by: Arun Kumar K 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-mfc: Fix frame skip bug

2013-03-21T21:24:07+00:00

The issue was seen in VP8 decoding where the last frame was
skipped by the driver. This patch gets the correct frame_type value
to fix this bug.

Signed-off-by: Arun Kumar K 
Signed-off-by: Arun Mankuzhi 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab

[media] s5p-fimc: send valid m2m ctx to fimc_m2m_job_finish

2013-03-21T21:23:21+00:00

fimc_m2m_job_finish() has to be called with the m2m context for the necessary
cleanup while resume. But currently fimc_m2m_job_finish() always passes m2m
context as NULL.
This patch preserves the context before making it null, for necessary cleanup.

Signed-off-by: Shaik Ameer Basha 
Signed-off-by: Sylwester Nawrocki 
Signed-off-by: Mauro Carvalho Chehab