From: "Steinar H. Gunderson" <sesse@xxxxxxxxxxxx> dvb_usercopy(), which is called on all ioctls, not only copies data to and from userspace, but also takes a lock on the file descriptor, which means that only one ioctl can run at a time. This means that if one thread of mumudvb is busy trying to get, say, the SNR from the frontend (which can hang due to the issue above), the CAM thread's ioctl(fd, CA_GET_SLOT_INFO, ...) will hang, even though it doesn't need to communicate with the hardware at all. This obviously requires a better fix, but I don't know the generic DVB layer well enough to say what it is. Maybe it's some sort of remnant of from when all ioctl()s took the BKL. Note that on UMP kernels without preemption, mutex_lock is to the best of my knowledge a no-op, so these delay issues would not show up on non-SMP. Signed-off-by: Steinar H. Gunderson <sesse@xxxxxxxxxxxx> --- drivers/media/dvb/dvb-core/dvbdev.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/media/dvb/dvb-core/dvbdev.c b/drivers/media/dvb/dvb-core/dvbdev.c index 00a6732..e1217f6 100644 --- a/drivers/media/dvb/dvb-core/dvbdev.c +++ b/drivers/media/dvb/dvb-core/dvbdev.c @@ -417,10 +417,8 @@ int dvb_usercopy(struct file *file, } /* call driver */ - mutex_lock(&dvbdev_mutex); if ((err = func(file, cmd, parg)) == -ENOIOCTLCMD) err = -EINVAL; - mutex_unlock(&dvbdev_mutex); if (err < 0) goto out; -- 1.7.9.5 -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
