new ebuild for 0.4.1 mostly the same as before but with the firmware located under the standard /lib/firmware directory for hotplug and with slightly different names
Created attachment 75632 [details] ivtv-0.4.1.ebuild
I was waiting for this bug to appear after the hordes of people e-mailed me directly about it. ivtv-0.4.1 fails horribly on both my systems. Not being commited to Portage. We'll wait for 0.4.2. Thanks.
*** Bug 117139 has been marked as a duplicate of this bug. ***
Unfortunately, 0.4.0 won't build with kernel 2.6.15, so if you need that kernel (say, for the latest DVB drivers to support the HD-5000 card), then you're hosed without this.
Well, 0.4.1 won't have sound for like half the cards out there.... So it's either force no sound on half of our user base. Or you have to suffer a little bit with "older" drivers from kernel 2.6.14
*** Bug 117319 has been marked as a duplicate of this bug. ***
*** Bug 117757 has been marked as a duplicate of this bug. ***
Created attachment 76368 [details] ivtv-0.4.2.ebuild svn - 3099
I've just re-opened the bug but for a different reason the latest 0.4.1 or the 0.5 development branch won't currently work with the latest 2.6.15 vanilla release of the linux kernel I think some last minuite changes were made before the 2.6.15 official release of the kernel which prevents the ivtv-0.4.1 or earlier release from working the only version that I've discovered that does work is the latest svn 0.4.2 release of ivtv (3099), it's not an official release of ivtv as this is still in the Subversion, so I won't expect this to show up in the portage tree just yet but for anyone interested in getting ivtv to work with the 2.6.15 gentoo sources (even though it's masked) using it as an overlay ebuild the 0.4.2 ebuild attached may be useful note this will download the latest svn release, I wasn't able to find a way to specify which release the tarball comes from in the ebuild, but the one I'm using at the moment which is 3099 seems to be okay for info there's also some mention of a 2.4 fix of some form in the Changelog although I don't know if this fixes all the 2.4 issues
Created attachment 76379 [details] ivtv-0.4.2.ebuild svn - 3099 darn used the wrong type for the file
0.4.1 worked for me just after I started my computer with the PVR-150 in it for the first time. After that, it wouldn't load because it "claimed" it couldn't find the firmware, but yet it was right in /lib/firmware where it was supposed to be, permissions 644.
It's probably best to stick with 0.4.0 if you can until 0.4.2 comes out properly I posted the above ebuild for the svn release in case anyone was using 2.6.15 with a pvr350 (which is what I've got at the moment) as for this card this is the only one I could get to work, but it may be different for different cards something to note is that for 0.4.1 and above the firmware has changed name / location inside /lib/firmware ivtv-fw-enc.bin -> v4l-cx2341x-enc.fw ivtv-fw-dec.bin -> v4l-cx2341x-dec.fw HcwMakoA.ROM -> v4l-cx25840.fw ivtv_init_mpeg.bin -> v4l-cx2341x-init-mpeg.bin I tried to reflect this in the ebuild but perhaps something was missed somewhere
we don't have a working ivtv until upstream releases 0.4.2 which should be in the next few days
Created attachment 77266 [details] IVTV 0.4.2 Ebuild that works! *NOTE* this requires that you copy /usr/portage/media-tv/ivtv/files/ivtv -> /usr/local/portage/media-tv/ivtv/files/ *NOTE* I've used the ivtv-0.4.1 and ivtv-0.4.2-svn ebuilds to create a proper 0.4.2 ebuild now that ivtv-0.4.2 has been properly released. I made two changes that were causing a "cannot stat" error, due to file name miss-matches. I have tested on 2.6.15 w/ a PVR-250 and the module loads. I've not had a chance to test actual video due to being remote to the server. I have also tested on a 2.6.14 kernel (and found a typo in the process), module loads fine.
Created attachment 77293 [details] update ivtv-0.4.2.ebuild that checks for correct kernel config Modified ebuild to check that the correct modules are enabled in the kernel. CONFIG_VIDEO_BT848 enables the required VIDEO_TUNER and VIDEO_TVEEPROM modules. Without VIDEO_TUNER and VIDEO_TVEEPROM the driver will appear to build and install correctly, but you will not be able to change channels.
good catch, doesn't it also require the "build extra video/audio modules" options in the 2.6.15 kernel?
Created attachment 77366 [details] ivtv-0.4.2-r1.ebuild looking at the changelog for 0.4.2 "renamed v4l-cx2341x-init-mpeg.bin to v4l-cx2341x-init.mpg (the name was too long for the 2.4 kernel firmware load, the max name length in 2.4 is 20, for 2.6 it is 29)." I've checked the source and v4l-cx2341x-init.mpg is the one it's using, so I've renamed this back in the ebuild the support for additional audio / video chipsets, will be covered by the "VIDEO_DECODER VIDEO_AUDIO_DECODER" on the config_check line
I used the ebuild in comment #17 and it works (including the kernel config checks!) for my PVR-500 on kernel 2.6.15-gentoo-r1. However, due to all the changes going on, I ended up just deleting all of my firmware and following the directions from http://ivtvdriver.org/index.php/Firmware to get new copies. It seems to load everything correctly and I can make an mpeg file using "cat /dev/video0 >> blah.mpeg". Strangely, tvtime is giving me an error of "videoinput: Card failed to allocate capture buffers: Invalid argument", but I think that's a separate tvtime-specific issue. Thanks to everyone who worked on this! :-)
I used ebuild from #17 and everything works fine (inclusive channels change), even without VIDEO_DECODER VIDEO_AUDIO_DECODER and CONFIG_VIDEO_BT848. Perhaps these switches are not obligatory? With enabling, the modules cs53132a.ko, msp3400.ko, saa7115.ko, saa7127.ko, tda9887.ko, tuner.ko, tveeprom.ko and wm8775.ko are build twice, once as a part of kernel (in drivers/media/video, once by ivtv (in extra). And these modules differ! It should not be, I suppose. My kernel is 2.6.15-gentoo-r1 (for x86_64).
There's been an ebuild for this in Portage for almost 24 hrs now...
*** Bug 127994 has been marked as a duplicate of this bug. ***
