Skip to contentSkip to conversation

Regression: VIDIOCGMBUF ioctl hangs on bttv driver (2.6.24-rc6)

DISCUSSION Gregor Jasny 5 participants 9 messages Jan 5, 2008 – Jan 7, 2008 Media (V4L/DVB)

lore
Share & tools
lore
#0 lore raw

Hi,

During some tests I've noticed that the VIDIOCGMBUF ioctl hangs on my bttv video device. It simply does not return and the process is stuck in the D+ state. With Kernel 2.6.22.9 the (attached) testcase works like a charm. The V4L2 interface is working with 2.6.24, too (at least displays xawtv the usual pixel snow).

Hard and Software:
64 bit Linux 2.6.24-rc6 (todays pull) on a Intel Core 2 Duo

WinTV PCI:
05:02.0 0400: 109e:036e (rev 02)
05:02.1 0480: 109e:0878 (rev 02)

dmesg:

Rincewind:~# dmesg |grep bttv
bttv: driver version 0.9.17 loaded
bttv: using 8 buffers with 2080k (520 pages) each for capture
bttv: Bt8xx card found (0).
bttv0: Bt878 (rev 2) at 0000:05:02.0, irq: 23, latency: 64, mmio: 0xdfefe000
bttv0: detected: Hauppauge WinTV [card=10], PCI subsystem ID is 0070:13eb
bttv0: using: Hauppauge (bt878) [card=10,autodetected]
bttv0: gpio: en=00000000, out=00000000 in=00ffffdb [init]
bttv0: Hauppauge/Voodoo msp34xx: reset line init [5]
bttv0: Hauppauge eeprom indicates model#61324
bttv0: tuner type=14
bttv0: i2c: checking for MSP34xx @ 0x80... found
bttv0: i2c: checking for TDA9875 @ 0xb0... not found
bttv0: i2c: checking for TDA7432 @ 0x8a... not found
bttv0: registered device video0
bttv0: registered device vbi0
bttv0: registered device radio0
bttv0: PLL: 28636363 => 35468950 .. ok

from /proc/interrupts:
23: 55900 0 IO-APIC-fasteoi ehci_hcd:usb2, uhci_hcd:usb5, bttv0

If I could help with more logs, a bisection, some debug printks, please contact me.

Thanks!

Gregor

#include <stdio.h>
#include <stdlib.h>
#include <assert.h>
#include <errno.h>

#include <sys/ioctl.h>
#include <sys/mman.h>
#include <fcntl.h>
#include <unistd.h>

#include <linux/ioctl.h>
#include <linux/videodev.h>

int main( int argc, char *argv[])

{
	const char device[] = "/dev/video0";
	int r;

	int fd = open (device, O_RDWR);
	assert(fd);

	struct video_mbuf mbuf = { 0 };
	r = ioctl (fd, VIDIOCGMBUF, &mbuf);
	assert (r == 0);

	printf ("Frames: %i\n", mbuf.frames);	

	close (fd);

	return 0;
}
#1 lore raw
Quoted 7 lines · Gregor Jasny

Hi,

During some tests I've noticed that the VIDIOCGMBUF ioctl hangs on my bttv video device. It simply does not return and the process is stuck in the D+ state. With Kernel 2.6.22.9 the (attached) testcase works like a charm. The V4L2 interface is working with 2.6.24, too (at least displays xawtv the usual pixel snow).

Do the 2.6.23.x kernels work?

Rafael

#2 lore raw
In reply to Gregor Jasny

From: Gregor Jasny <[email protected]>

Fix bttv VIDIOCGMBUF locking like done in commit
820eacd84cff23b76693f4be1e28feb672f4488f.

---

drivers/media/video/bt8xx/bttv-driver.c +1 -3
index c02d92d..581a3c9 100644
@@ -3063,11 +3063,10 @@ static int bttv_do_ioctl(struct inode *inode, struct file *file,
 		struct video_mbuf *mbuf = arg;
 		unsigned int i;
 
-		mutex_lock(&fh->cap.lock);
 		retval = videobuf_mmap_setup(&fh->cap,gbuffers,gbufsize,
 					     V4L2_MEMORY_MMAP);
 		if (retval < 0)
-			goto fh_unlock_and_return;
+			return retval;
 
 		gbuffers = retval;
 		memset(mbuf,0,sizeof(*mbuf));
@@ -3075,7 +3074,6 @@ static int bttv_do_ioctl(struct inode *inode, struct file *file,
 		mbuf->size   = gbuffers * gbufsize;
 		for (i = 0; i < gbuffers; i++)
 			mbuf->offsets[i] = i * gbufsize;
-		mutex_unlock(&fh->cap.lock);
 		return 0;
 	}
 	case VIDIOCMCAPTURE:
#3 lore raw
Quoted 4 lines · Gregor Jasny

From: Gregor Jasny <[email protected]>

Fix bttv VIDIOCGMBUF locking like done in commit
820eacd84cff23b76693f4be1e28feb672f4488f.

i've double-checked that there are no other video drivers affected by this problem:

 $ cd drivers/media/
 $ grep -C 10 videobuf_mmap_setup `grep -l mutex_lock */*.c */*/*.c */*/*/*.c`

video/videobuf-core.c- mutex_lock(&q->lock);
video/videobuf-core.c- mutex_lock(&q->lock);
video/bt8xx/bttv-driver.c- mutex_lock(&fh->cap.lock);

so only bttv-driver.c was affected, at first sight.

Ingo

#4 lore raw
In reply to Gregor Jasny

Hi guys,

Gregor, we have converted bttv driver to use vidioc_ioctl2 some days ago.
Could you check and create your patch against v4l development tree? Bttv driver does not have anymore bttv_do_ioctl().

Cheers,
Douglas

Quoted 39 lines · Gregor Jasny

From: Gregor Jasny <[email protected]>

Fix bttv VIDIOCGMBUF locking like done in commit
820eacd84cff23b76693f4be1e28feb672f4488f.

---

drivers/media/video/bt8xx/bttv-driver.c +1 -3
index c02d92d..581a3c9 100644
@@ -3063,11 +3063,10 @@ static int bttv_do_ioctl(struct inode *inode, struct file *file,
                struct video_mbuf *mbuf = arg;
                unsigned int i;
 
-               mutex_lock(&fh->cap.lock);
                retval = videobuf_mmap_setup(&fh->cap,gbuffers,gbufsize,
                                             V4L2_MEMORY_MMAP);
                if (retval < 0)
-                       goto fh_unlock_and_return;
+                       return retval;
 
                gbuffers = retval;
                memset(mbuf,0,sizeof(*mbuf));
@@ -3075,7 +3074,6 @@ static int bttv_do_ioctl(struct inode *inode, struct file *file,
                mbuf->size   = gbuffers * gbufsize;
                for (i = 0; i < gbuffers; i++)
                        mbuf->offsets[i] = i * gbufsize;
-               mutex_unlock(&fh->cap.lock);
                return 0;
        }
        case VIDIOCMCAPTURE:

_______________________________________________
v4l-dvb-maintainer mailing list
[email protected]
http://www.linuxtv.org/cgi-bin/mailman/listinfo/v4l-dvb-maintainer

#5 lore raw

I mean video_ioctl2 not vidioc_ioctl2, sorry.

Cheers,
Douglas

Quoted 51 lines · Douglas Landgraf

Hi guys,

Gregor, we have converted bttv driver to use vidioc_ioctl2 some days ago.
Could you check and create your patch against v4l development tree? Bttv driver does not have anymore bttv_do_ioctl().

Cheers,
Douglas

Quoted 39 lines · Gregor Jasny

From: Gregor Jasny <[email protected]>

Fix bttv VIDIOCGMBUF locking like done in commit
820eacd84cff23b76693f4be1e28feb672f4488f.

---

drivers/media/video/bt8xx/bttv-driver.c +1 -3
index c02d92d..581a3c9 100644
@@ -3063,11 +3063,10 @@ static int bttv_do_ioctl(struct inode *inode, struct file *file,
                struct video_mbuf *mbuf = arg;
                unsigned int i;
 
-               mutex_lock(&fh->cap.lock);
                retval = videobuf_mmap_setup(&fh->cap,gbuffers,gbufsize,
                                             V4L2_MEMORY_MMAP);
                if (retval < 0)
-                       goto fh_unlock_and_return;
+                       return retval;
 
                gbuffers = retval;
                memset(mbuf,0,sizeof(*mbuf));
@@ -3075,7 +3074,6 @@ static int bttv_do_ioctl(struct inode *inode, struct file *file,
                mbuf->size   = gbuffers * gbufsize;
                for (i = 0; i < gbuffers; i++)
                        mbuf->offsets[i] = i * gbufsize;
-               mutex_unlock(&fh->cap.lock);
                return 0;
        }
        case VIDIOCMCAPTURE:

_______________________________________________
v4l-dvb-maintainer mailing list
[email protected]
http://www.linuxtv.org/cgi-bin/mailman/listinfo/v4l-dvb-maintainer

#6 lore raw
In reply to Douglas Landgraf
Quoted 5 lines · Douglas Landgraf

Hi guys,

Gregor, we have converted bttv driver to use vidioc_ioctl2 some
days ago. Could you check and create your patch against v4l development tree? Bttv driver does not have anymore bttv_do_ioctl().

this is about the bttv driver in 2.6.24-rc6/(rc7?) hanging. So whatever the devel tree does, this fix (if it's a correct fix) needs to be commited upstream ASAP.

Ingo

ps. http://www.zip.com.au/~akpm/linux/patches/stuff/top-posting.txt

#7 lore raw

Hi Ingo,

Quoted 3 lines

this is about the bttv driver in 2.6.24-rc6/(rc7?) hanging. So whatever the devel tree does, this fix (if it's a correct fix) needs to be commited upstream ASAP.

You are right, my misunderstood. Sorry guys!

Cheers,
Douglas

#8 lore raw
In reply to Ingo Molnar
Quoted 5 lines · On Mon, 7 Jan 2008 00:22:34 +0100 Ingo Molnar <[email protected]> wrote:
Quoted 2 lines · Douglas Landgraf

Hi guys,

Quoted 7 lines
Quoted 3 lines

Gregor, we have converted bttv driver to use vidioc_ioctl2 some
days ago. Could you check and create your patch against v4l development tree? Bttv driver does not have anymore bttv_do_ioctl().

this is about the bttv driver in 2.6.24-rc6/(rc7?) hanging. So whatever the devel tree does, this fix (if it's a correct fix) needs to be commited upstream ASAP.

For sure. I've just sent Linus a pull request for the fix.

On the devel tree, all this code got removed, replaced by a simple function call at videobuf. So, I don't expect similar issues for 2.6.25.

I've already fixed the patch conflicts at devel branch for v4l/dvb -git tree.

Cheers,
Mauro