HD-PVR Linux Support

[cenwesi]

they really need to port these driver to 64bit. Doesn't make sense at all that we all have 64bit cpu and still using 32bit codes

[bcjenkins]

If you're willing to test you can do a pull from http://linuxtv.org/hg/~hverkuil/v4l-dvb-compat32 using Mercurial and get the updated ioctls. I will outline the steps I took to get it plugged in to the HDPVR driver code in a bit.

B

[trueno]

I filed a support request with Sage, asking if/when a 64 bit native version of Sage Linux would be available... would certainly eliminate a huge swatch of problems..

[bcjenkins]

So, I sent a note to Janne who develops the HDPVR and asked him to sync up to the latest DVB tree and that is done. So we're in business. Below are the steps I took in compiling the driver from a clean start:

hg clone http://hg.jannau.net/hdpvr
cd hdpvr
hg pull http://linuxtv.org/hg/~hverkuil/v4l-dvb-compat32
hg merge
hg commit
nano linux/drivers/media/video/hdpvr/hdpvr.h #Edit header file to support my firmware
make menuconfig
make
sudo make install
reboot

This assumes you have Mercurial installed.

B

PS - Make sure you remove the old ioctl as well! I was bitten by this.

[harrijay]

I am trying to get both of my HDPVR's working on 64bit ubuntu but I am running into issues. it appears that I need to get the io32ctl merged into my hdpvr driver but I am not able to find it anywhere. I think Hans did some reorganizing of his source code. can anyone verify that I should be able ot pull the source from http://linuxtv.org/hg/~hverkuil/v4l-dvb-ctrls. or would the main source for v4l be a better place to get the best stuff

TIA,
Jayson

[bcjenkins]

Here's the issue: Hans has reorganized the ioctl and Janne's tree is out of date. I have sent Janne a note asking for him to sync with linuxtv, but have not seen a response. I also popped into the IRC room as well, with the same request.

If I knew how to make the changes myself, I would. Otherwise, you should be able to build just out of Janne's tree unless you need something newer (like for a newer kernel)

B

[Ikarius]

Hmmm. What's the status of support for using the hdpvr's built-in IR blaster?

I see in recent builds of sagetv that it appears like they mention adding support for using it under linux, but that obviously requires kernel driver level support, and I can't find any documentation anyplace about the kernel driver supporting it.

Any clues to what the real state is?

[_Demo_]

Originally Posted by Ikarius

Hmmm. What's the status of support for using the hdpvr's built-in IR blaster?

I see in recent builds of sagetv that it appears like they mention adding support for using it under linux, but that obviously requires kernel driver level support, and I can't find any documentation anyplace about the kernel driver supporting it.

Any clues to what the real state is?

I have made a patch to the driver to enable i2c support inside the hdpvr driver and sent it to its developer. With the latest version of that patch and the latest SageTV server the ir blaster seems to be working reliably.

_Demo_

[_Demo_]

I have attached a copy of the patch needed for ir blaster support.

_Demo_

[Ikarius]

Okay, and given bcjenkin's comments above, what constitutes appropriate tree to apply this patch to?

I know the HDPVR support is in the tree at http://hg.jannau.net/hdpvr/ - but the 32-bit ioctl is in a different tree, and the trees are out of sync, according to bcjenkin's post.

So, what would I pull today to apply that patch to?

[harrijay]

I did build out of Janne's Tree and I am currently running Ubuntu 8.10 64Bit Kernel 2.6.27-9. I am able to tune the initial channel but if I try to change I get an error in sage about unable to setup software decoder and the following lines show up in dmesg

[ 46.347818] compat_ioctl32: VIDIOC_S_AUDIO
[ 46.347838] ioctl32(java:6266): Unknown cmd fd(43) cmd(40345622){t:'V';sz:52} arg(df15cb34) on /dev/video1
[ 46.828954] compat_ioctl32: VIDIOC_S_AUDIO
[ 46.829139] ioctl32(java:6266): Unknown cmd fd(47) cmd(40345622){t:'V';sz:52} arg(df15cb34) on /dev/video2
[ 47.025506] compat_ioctl32: VIDIOC_S_AUDIO
[ 47.025525] ioctl32(java:6266): Unknown cmd fd(48) cmd(40345622){t:'V';sz:52} arg(df15cb34) on /dev/video0
[ 57.881933] mtrr: base(0xc0000000) is not aligned on a size(0x7f0000) boundary
[ 93.177225] compat_ioctl32: VIDIOC_S_AUDIO
[ 93.177245] ioctl32(java:6681): Unknown cmd fd(43) cmd(40345622){t:'V';sz:52} arg(dd472bc4) on /dev/video1
[ 93.273482] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[ 93.273816] ioctl32(java:6681): Unknown cmd fd(43) cmd(c0185648){t:'V';sz:24} arg(dd4727a0) on /dev/video1
[ 93.274439] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[ 93.274716] ioctl32(java:6681): Unknown cmd fd(43) cmd(c0185648){t:'V';sz:24} arg(dd4727a0) on /dev/video1
[ 93.275319] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[ 93.275593] ioctl32(java:6681): Unknown cmd fd(43) cmd(c0185648){t:'V';sz:24} arg(dd4727a0) on /dev/video1
[ 785.616878] compat_ioctl32: VIDIOC_ENCODER_CMD
[ 785.616895] ioctl32(java:6684): Unknown cmd fd(95) cmd(c028564d){t:'V';sz:40} arg(de909ee4) on /dev/video1
[ 798.709669] compat_ioctl32: VIDIOC_S_AUDIO
[ 798.709686] ioctl32(java:6758): Unknown cmd fd(48) cmd(40345622){t:'V';sz:52} arg(de909924) on /dev/video0
[ 798.798699] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[ 798.799060] ioctl32(java:6758): Unknown cmd fd(48) cmd(c0185648){t:'V';sz:24} arg(de909510) on /dev/video0
[ 798.799807] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[ 798.800150] ioctl32(java:6758): Unknown cmd fd(48) cmd(c0185648){t:'V';sz:24} arg(de909510) on /dev/video0
[ 798.800891] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[ 798.801235] ioctl32(java:6758): Unknown cmd fd(48) cmd(c0185648){t:'V';sz:24} arg(de909510) on /dev/video0
[ 811.425585] compat_ioctl32: VIDIOC_ENCODER_CMD
[ 811.425603] ioctl32(java:6761): Unknown cmd fd(95) cmd(c028564d){t:'V';sz:40} arg(dcf5cce4) on /dev/video0


This is causing me problems with both my PVR150 and both HDPVR's and just based on the info I assume it is the ioctl32 that is causing the problem. thanks for looking into this for us.

Jayson

[bcjenkins]

If you have compat_ioctl32 and a v4l2_compat_ioctl32 then remove (delete) this first one.

B

[harrijay]

I deleted the compat_ioctl32 and it did not help.

Jayson

[Ikarius]

latest hdpvr driver, I'm getting the same sort of errors as harrijay-

Following shows up in my dmesg:

[12876821.979515] compat_ioctl32: VIDIOC_S_AUDIO
[12876821.979528] ioctl32(java:31744): Unknown cmd fd(86) cmd(40345622){t:'V';sz:52} arg(e6fbdff0) on /dev/video0
[12876823.177699] compat_ioctl32: VIDIOC_S_AUDIO
[12876823.177711] ioctl32(java:31744): Unknown cmd fd(86) cmd(40345622){t:'V';sz:52} arg(e6fbdff0) on /dev/video0
[12876823.259482] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[12876823.259494] ioctl32(java:31744): Unknown cmd fd(86) cmd(c0185648){t:'V';sz:24} arg(e6fbdb7c) on /dev/video0
[12876823.259575] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[12876823.259581] ioctl32(java:31744): Unknown cmd fd(86) cmd(c0185648){t:'V';sz:24} arg(e6fbdb7c) on /dev/video0
[12876823.259661] compat_ioctl32: VIDIOC_S_EXT_CTRLS
[12876823.259667] ioctl32(java:31744): Unknown cmd fd(86) cmd(c0185648){t:'V';sz:24} arg(e6fbdb7c) on /dev/video0
[12876847.636301] compat_ioctl32: VIDIOC_ENCODER_CMD
[12876847.636314] ioctl32(java:31745): Unknown cmd fd(87) cmd(c028564d){t:'V';sz:40} arg(e66b9fe0) on /dev/video0

At the same time, I get the following in sagetv_0.txt:

Mon 1/19 20:43:37.317 V4L: getV4LCardType /dev/video0
Mon 1/19 20:43:37.317 V4L: getV4LInputName /dev/video0 0
Mon 1/19 20:43:37.317 V4L: setInput0 0 0 1 1
Mon 1/19 20:43:37.317 V4L: setting standard to 0x3000
Mon 1/19 20:43:37.334 V4L: Failed setting the input on the capture device


This is with linux server version 6.4.8.1

With version 6.5.7.1, I get a Java exception on trying to deal with the hdpvr:

Mon 1/19 20:37:15.018 Seeker.finishWatch(/172.31.5.108:1497@@SAGETV_PROCESS_LOCAL_UI)
Mon 1/19 20:37:17.385 Called Seeker.requestWatch(FA[BAD AIRING]) hostname=/172.31.5.108:1497@@SAGETV_PROCESS_LOCAL_UI
Mon 1/19 20:37:17.386 findBestEncoderForNow(FA[BAD AIRING] record=false host=/172.31.5.108:1497@@SAGETV_PROCESS_LOCAL_UI)
Mon 1/19 20:37:17.386 ERROR in message processing: java.lang.NullPointerException
Mon 1/19 20:37:17.386 java.lang.NullPointerException
Mon 1/19 20:37:17.386 at sage.cb$c.access$1400(Unknown Source)
Mon 1/19 20:37:17.386 at sage.cb.a(Unknown Source)
Mon 1/19 20:37:17.386 at sage.cb.a(Unknown Source)
Mon 1/19 20:37:17.386 at sage.w.h(Unknown Source)
Mon 1/19 20:37:17.386 at sage.w.d(Unknown Source)
Mon 1/19 20:37:17.386 at sage.w.access$100(Unknown Source)
Mon 1/19 20:37:17.386 at sage.w$3.run(Unknown Source)


So.... looks like the current hdpvr drivers aren't playing nice with SageTV.


I can confirm that I get an i2c device, and I can record off /dev/video0 via the "cat" program, so this looks like it's on SageTV's side.


Any ideas?

[bcjenkins]

Originally Posted by harrijay

I deleted the compat_ioctl32 and it did not help.

Jayson

Did you restart?

B

[harrijay]

Yes. I did a search for *ioctl32* on my system and the only place it showed up is in the source for the HDPVR driver and the /lib/modules folder. I removed them out of both locations. rebuilt the driver, reinstalled it and then restarted. same results. If I should be doing something different let me know.

Thanks,
Jayson

[Ikarius]

I'm 100% certain that module is gone on my system too- I've verified that through lsmod. While what I pulled from jannau.net took the irblasting support patch fine, compiles loads and appears to work at a system level, it doesn't appear to be playing nice with sagetv.

I'm still hoping someone can answer if there's any place I can pull a copy of the hdpvr module known to work with sagetv from. This may be the same issue that Harrijay is having.

[bcjenkins]

I do have a working hd-pvr right now, but I can't upgrade a kernel or my cx18 drivers as it breaks the driver compilation.

B

[bcjenkins]

Janne has updated his mercurial today. Aside from:

CC [M]  /root/drivers/hdpvr/v4l/hdpvr-video.o
/root/drivers/hdpvr/v4l/hdpvr-video.c:374: warning: 'hdpvr_flush' defined but not used

It has compiled cleanly for me. I won't be able to test until tomorrow in fear of mutiny if I ruin Thursday night prime time.

B

[Ikarius]

Cool,
I'll give that a go tonight, and see where I get with it. Thanks