[mythtv-users] HD-PVR polling errors -- possible solution
Larry.Finger at lwfinger.net
Fri Sep 21 18:01:56 UTC 2012
On 09/20/2012 11:30 PM, Steven Adeff wrote:
> On Wed, Sep 19, 2012 at 9:15 PM, John P Poet <jppoet at gmail.com> wrote:
>> Christopher Neufeld has been trying to track down a serious reliability
>> issue with his HD-PVR. After much research, he discovered that his problems
>> were due to a bug in the driver. This bug was fixed in the 3.2.11, but he
>> runs LinHES which is based on Arch Linux, which has an older kernel.
>> So, if you are having problems with your HD-PVR, and your kernel predates
>> 3.2.11, you probably want to upgrade.
>> Christopher is not a member of this mailing list, and he asked me to forward
>> I've been having some HD-PVR discussions on IRC, and it was suggested that
>> I post some stuff to the mailing list as it's of general interest.
>> First, I have been having terrible reliability problems with my HD-PVR, it
>> would get caught in a polling loop and fail to start up. I tracked this
>> down to a race condition in the hdpvr module for the kernel I was using
>> (2.6.39). While this race condition has been fixed in recent kernels, if
>> you find yourself with failed HD-PVR recordings on an older kernel, and log
>> messages that look like this:
>> 2012-09-17T00:00:19.945718-04:00 mythtv mythbackend: E RecThread
>> mpegrecorder.cpp:1010 (run) MPEGRec(/dev/hdpvr): Device error detected
>> 2012-09-17T00:00:23.347878-04:00 mythtv mythbackend: E
>> DeviceReadBuffer DeviceReadBuffer.cpp:513 (Poll) DevRdB(/dev/hdpvr): Poll
>> giving up 2
>> repeating every 3 or 4 seconds, check your module. In hdpvr-video.c, the
>> function hdpvr_start_streaming(), dev->status must be set to
>> STATUS_STREAMING before INIT_WORK sets up hdpvr_transmit_buffers. If the
>> order is reversed, fix it and recompile.
>> The second note is informational. If you have a free encoder on your
>> system, it is now possible to make recordings with the HD-PVR with
>> closed-captions available, capable of being toggled at playback time,
>> subject to certain conditions. See
>> for details.
>> Thank you.
> Some more info on exactly what change to hdpvr-video.c needs to be
> made would be helpful.
The change that is most likely is as follows:
Author: Janne Grunau <j at jannau.net>
Date: Thu Feb 2 13:35:21 2012 -0300
[media] hdpvr: fix race conditon during start of streaming
status has to be set to STREAMING before the streaming worker is
queued. hdpvr_transmit_buffers() will exit immediately otherwise.
Reported-by: Joerg Desch <vvd.joede at googlemail.com>
CC: stable at kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab at redhat.com>
diff --git a/drivers/media/video/hdpvr/hdpvr-video.c
index 087f7c0..41fd57b 100644
@@ -283,12 +283,13 @@ static int hdpvr_start_streaming(struct hdpvr_device *dev)
hdpvr_config_call(dev, CTRL_START_STREAMING_VALUE, 0x00);
+ dev->status = STATUS_STREAMING;
v4l2_dbg(MSG_BUFFER, hdpvr_debug, &dev->v4l2_dev,
- dev->status = STATUS_STREAMING;
This patch to move the STATUS_STREAMING line was back ported in 3.2.9.
More information about the mythtv-users