public inbox for linux-media@vger.kernel.org
 help / color / mirror / Atom feed
* bttv problem
@ 2009-07-06 23:15 Linos
  0 siblings, 0 replies; 6+ messages in thread
From: Linos @ 2009-07-06 23:15 UTC (permalink / raw)
  To: linux-media

Hello all,
	some time ago (2008-10-20) i sent a post to the list 
video4linux-list@redhat.com explaining a problem that still seems to be in the 
last kernel, Daniel Glöckner helped me with a workaround, i will paste here the 
problem and the fix from Daniel.

------------------------------------------------------------------------------------------------ 
Original mail
Hello, i have upgraded a debian machine from kernel 2.6.24 to 2.6.26 and now i 
have this error when try to launch helix producer on the capture input.

producer -vc /dev/video0 -ad 128k -vp "0" -dt -vm sharp -o /tmp/test.rm
Helix DNA(TM) Producer 11.0 Build number: 11.0.0.2013
Info: Starting encode

Error: Could not set image size to 352x288 for color format I420 (15) 
(VIDIOCMCAPTURE: buffer 0)

Warning: Capture Buffer is empty at 445090329ms for last 61 times
Warning: Capture Buffer is empty at 445091549ms for last 61 times


exactly the same producer version with the same command line works ok in 2.6.24, 
previously i have saw this error when i was trying to use from 2 different 
capture programs the same video input but this is not the case, it is the only 
program using /dev/video0, what can be happening?

------------------------------------------------------------------------------------------------ 
Daniel's reply
On Mon, Oct 20, 2008 at 03:56:01PM +0200, Linos wrote:
 > Error: Could not set image size to 352x288 for color format I420 (15)
 > (VIDIOCMCAPTURE: buffer 0)

The problem is that the v4l1-compat code for VIDIOCMCAPTURE calls
VIDIOC_S_FMT. At the beginning of bttv_s_fmt_vid_cap the call to
bttv_switch_type fails because the buffers have already been mmap'ed
by the application. I'd say this is a bug in bttv.

In which case does the videobuf_queue_is_busy test prevent bad things
from happening?


A workaround is to set the resolution and image format before the
buffers are mapped, f.ex. with this small program:
--------------
#include <sys/ioctl.h>
#include <fcntl.h>
#include <unistd.h>
#include <linux/videodev.h>

void main()
{
   struct video_mmap vmm;
   vmm.width=352;
   vmm.height=288;
   vmm.format=VIDEO_PALETTE_YUV420P;
   vmm.frame=0;
   ioctl(open("/dev/video",O_RDWR),VIDIOCMCAPTURE,&vmm);
}
--------------

   Daniel

------------------------------------------------------------------------------------------------ 



I have upgraded the same machine now to 2.6.30 and still have the same problem 
and i have to use the code Daniel sent the list to make it work, i am missing 
something or still have not been fixed in mainstream kernel? Thanks.

Regards,
Miguel Angel.

^ permalink raw reply	[flat|nested] 6+ messages in thread
* bttv problem
@ 2010-12-19 12:54 Linos
  0 siblings, 0 replies; 6+ messages in thread
From: Linos @ 2010-12-19 12:54 UTC (permalink / raw)
  To: linux-media

Hi,
	i have 2 Provideo PV150 multi-capture PCI board video cards  (4 bt878 chipsets 
in every board), i have the four ports of the first multi-capture boards used in 
a xorg session with 4 instances of tvtime application showing the realtime of 
the four ports, and the other one recording with Helix producer.

Many times a day i have this errors in kernel:

[447547.296022] bttv7: timeout: drop=3276 irq=1957469/35501901, risc=204bb9ac, 
bits: HSYNC OFLOW
[447547.836020] bttv7: timeout: drop=3287 irq=1957508/35501981, risc=2a9a89ac, 
bits: HSYNC OFLOW
[448114.536016] bttv6: timeout: drop=4747 irq=26301665/59788770, risc=2050c998, 
bits: HSYNC OFLOW
[449923.196018] bttv1: timeout: drop=57206 irq=33728690/35680137, risc=19d83acc, 
bits: HSYNC OFLOW FBUS
[449923.736024] bttv1: timeout: drop=57217 irq=33728729/35680176, risc=344faf20, 
bits: HSYNC OFLOW FBUS

Tvtime seems to survive but the helix producer processess die with this error, 
if i don't load the xorg session with tvtime and at the same time I disable acpi 
in linux i don't get this erros, i have tried to disable acpi because i have 
read about other people having the same problem and fixing this way, but if i 
use tvtime at the same time the problem remains regardless of the acpi status, i 
have now gbuffers at 16 but still the same result.

My system it is using debian squeeze with 2.6.32 kernel in a Asus P5E-V HDMI 
with an intel G-35 chipset.

Any ideas on how i could fix this problem please?

Regards,
Miguel Angel.

^ permalink raw reply	[flat|nested] 6+ messages in thread
* bttv problem
@ 2009-06-17  8:49 Linos
  2009-07-05 22:29 ` Linos
  0 siblings, 1 reply; 6+ messages in thread
From: Linos @ 2009-06-17  8:49 UTC (permalink / raw)
  To: video4linux-list

Hello,
	i posted this problem any time ago ( 
https://www.redhat.com/mailman/private/video4linux-list/2008-October/msg00246.html 
), but still it seems to be in the last kernel, for details of the bug please 
read the link, i have upgraded this week to 2.6.30 and still i have to execute 
the code Daniel send the list to fix the bug in the bttv driver in 2.6.26, i am 
missing something or still have not been fixed in mainstream kernel?

Regards,
Miguel Angel.

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2010-12-19 12:54 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-07-06 23:15 bttv problem Linos
  -- strict thread matches above, loose matches on Subject: below --
2010-12-19 12:54 Linos
2009-06-17  8:49 Linos
2009-07-05 22:29 ` Linos
2009-07-06  1:14   ` hermann pitton
2009-07-06 23:03     ` Linos

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox