If you're among those who have voiced concerns with the FireWire support in Fedora, particularly in the audio/video area, please do try to test out the latest rawhide kernels (2.6.24-0.149.fc7.git2.fc9 or later) and/or the 2.6.23.13 kernel working its way through koji right now[1].
For the most part, dvgrab should be functional with these kernels, on all but Via VT630x chipsets in OHCI 1.0 mode[2]. IIDC video streaming (such as coriander + libdc1394 v2) should be fully functional on all OHCI 1.0 and 1.1 FireWire controllers.
As yet untested with this latest kernel is FireWire video capture off a cable box[3], but that's on my todo list for the weekend. (There's reason to believe this might be working now with the addition of dynamic buffer allocation).
Additional feedback on other open FireWire bugs would be appreciated as well[4][5]. Note that most of the enhancements added to the latest kernels are on the a/v side, but I plan to start looking at storage more soon as well...
[1] http://koji.fedoraproject.org/koji/taskinfo?taskID=343219 [2] https://bugzilla.redhat.com/show_bug.cgi?id=415841 [3] https://bugzilla.redhat.com/show_bug.cgi?id=240774 [4] https://bugzilla.redhat.com/show_bug.cgi?id=243081 [5] https://bugzilla.redhat.com/show_bug.cgi?id=370931
Jarod Wilson wrote:
If you're among those who have voiced concerns with the FireWire support in Fedora, particularly in the audio/video area, please do try to test out the latest rawhide kernels (2.6.24-0.149.fc7.git2.fc9 or later) and/or the 2.6.23.13 kernel working its way through koji right now[1].
For the most part, dvgrab should be functional with these kernels, on all but Via VT630x chipsets in OHCI 1.0 mode[2]. IIDC video streaming (such as coriander + libdc1394 v2) should be fully functional on all OHCI 1.0 and 1.1 FireWire controllers.
As yet untested with this latest kernel is FireWire video capture off a cable box[3], but that's on my todo list for the weekend. (There's reason to believe this might be working now with the addition of dynamic buffer allocation).
Additional feedback on other open FireWire bugs would be appreciated as well[4][5]. Note that most of the enhancements added to the latest kernels are on the a/v side, but I plan to start looking at storage more soon as well...
As a perhaps odd case, is it likely to be possible to build an initrd containing the firewire driver and have firewire drives included in md devices recognized and matched correctly during boot-up?
Les Mikesell wrote:
Jarod Wilson wrote:
If you're among those who have voiced concerns with the FireWire support in Fedora, particularly in the audio/video area, please do try to test out the latest rawhide kernels (2.6.24-0.149.fc7.git2.fc9 or later) and/or the 2.6.23.13 kernel working its way through koji right now[1].
Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
As yet untested with this latest kernel is FireWire video capture off a cable box[3], but that's on my todo list for the weekend. (There's reason to believe this might be working now with the addition of dynamic buffer allocation).
Nope, still no dice. Will poke more soonish, I'm inclined to think its a userspace issue...
As a perhaps odd case, is it likely to be possible to build an initrd containing the firewire driver and have firewire drives included in md devices recognized and matched correctly during boot-up?
Should be doable, yes. If you've got a line in modprobe.conf of 'alias scsi_hostadapterX firewire-sbp2' (for some value of X) and the array running when you create your initrd, it *should* automagically include the necessary modules in the initrd. I'd probably run mkinitrd -v by hand to see for sure.
Jarod Wilson wrote:
Les Mikesell wrote:
Jarod Wilson wrote:
If you're among those who have voiced concerns with the FireWire support in Fedora, particularly in the audio/video area, please do try to test out the latest rawhide kernels (2.6.24-0.149.fc7.git2.fc9 or later) and/or the 2.6.23.13 kernel working its way through koji right now[1].
Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
As yet untested with this latest kernel is FireWire video capture off a cable box[3], but that's on my todo list for the weekend. (There's reason to believe this might be working now with the addition of dynamic buffer allocation).
Nope, still no dice. Will poke more soonish, I'm inclined to think its a userspace issue...
As a perhaps odd case, is it likely to be possible to build an initrd containing the firewire driver and have firewire drives included in md devices recognized and matched correctly during boot-up?
Should be doable, yes. If you've got a line in modprobe.conf of 'alias scsi_hostadapterX firewire-sbp2' (for some value of X) and the array running when you create your initrd, it *should* automagically include the necessary modules in the initrd. I'd probably run mkinitrd -v by hand to see for sure.
I can confirm that the firewire-sbp2 driver will be included in initrd by kernel installs with that alias present, and it works at least for my single external firewire drive... I don't know about md devices though. It doesn't work flawlessly, I have a problem with my drive causing some kernels not fail booting (https://bugzilla.redhat.com/show_bug.cgi?id=349921) but I'm not sure if its the hardware's fault or not on that one. But as the bug says once the kernel is running access to the drive is fine.
Andrew Farris wrote:
Jarod Wilson wrote:
Les Mikesell wrote:
Jarod Wilson wrote:
If you're among those who have voiced concerns with the FireWire support in Fedora, particularly in the audio/video area, please do try to test out the latest rawhide kernels (2.6.24-0.149.fc7.git2.fc9 or later) and/or the 2.6.23.13 kernel working its way through koji right now[1].
Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
I tried the koji kernel linked above and this boot panic problem I was having seems to be fixed with that new kernel. If anyone has had issues with firewire drives when booting kernels in F8 take a look at:
https://bugzilla.redhat.com/show_bug.cgi?id=428554
On Sat, 2008-01-12 at 23:43 -0800, Andrew Farris wrote:
Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
I tried the koji kernel linked above and this boot panic problem I was having seems to be fixed with that new kernel. If anyone has had issues with firewire drives when booting kernels in F8 take a look at:
I tried that kernel, it booted fine, but when it starting the x server, going into gnome or whatever, it shows a weird jibberish display and locks up hard, nothing works. You can boot to run level 3 and startx locks it up. Hope this update isn't coming through testing/updates for F8.
Mike Chambers wrote:
On Sat, 2008-01-12 at 23:43 -0800, Andrew Farris wrote:
Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
I tried the koji kernel linked above and this boot panic problem I was having seems to be fixed with that new kernel. If anyone has had issues with firewire drives when booting kernels in F8 take a look at:
I tried that kernel, it booted fine, but when it starting the x server, going into gnome or whatever, it shows a weird jibberish display and locks up hard, nothing works. You can boot to run level 3 and startx locks it up. Hope this update isn't coming through testing/updates for F8.
Hrm... Nothing I added to that kernel should have any impact whatsoever on X... I believe 104.fc8 or 105.fc8 did go into updates-testing though. Would certainly be curious to hear if those *don't* cause the problem that 106.fc8 does, but I'd be quite surprised if the firewire update was the case.
FWIW, 106.fc8 runs without a problem on all of my boxes I've installed it on thus far.
Jarod Wilson wrote:
Mike Chambers wrote:
On Sat, 2008-01-12 at 23:43 -0800, Andrew Farris wrote:
Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
I tried the koji kernel linked above and this boot panic problem I was having seems to be fixed with that new kernel. If anyone has had issues with firewire drives when booting kernels in F8 take a look at:
I tried that kernel, it booted fine, but when it starting the x server, going into gnome or whatever, it shows a weird jibberish display and locks up hard, nothing works. You can boot to run level 3 and startx locks it up. Hope this update isn't coming through testing/updates for F8.
Hrm... Nothing I added to that kernel should have any impact whatsoever on X... I believe 104.fc8 or 105.fc8 did go into updates-testing though. Would certainly be curious to hear if those *don't* cause the problem that 106.fc8 does, but I'd be quite surprised if the firewire update was the case.
FWIW, 106.fc8 runs without a problem on all of my boxes I've installed it on thus far.
I haven't seen 104 or 105 in updates-testing, and no kernel at all is in the repo today. I will see if 104 and 105 both fix my firewire drive issue though.
Jarod Wilson wrote: Its built.
http://koji.fedoraproject.org/packages/kernel/2.6.23.13/106.fc8/
Works like like a charm over here!
Regards,
Hans
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Jarod Wilson wrote:
If you're among those who have voiced concerns with the FireWire support in Fedora, particularly in the audio/video area, please do try to test out the latest rawhide kernels (2.6.24-0.149.fc7.git2.fc9 or later) and/or the 2.6.23.13 kernel working its way through koji right now[1].
For the most part, dvgrab should be functional with these kernels, on all but Via VT630x chipsets in OHCI 1.0 mode[2]. IIDC video streaming (such as coriander + libdc1394 v2) should be fully functional on all OHCI 1.0 and 1.1 FireWire controllers.
As yet untested with this latest kernel is FireWire video capture off a cable box[3], but that's on my todo list for the weekend. (There's reason to believe this might be working now with the addition of dynamic buffer allocation).
Additional feedback on other open FireWire bugs would be appreciated as well[4][5]. Note that most of the enhancements added to the latest kernels are on the a/v side, but I plan to start looking at storage more soon as well...
[1] http://koji.fedoraproject.org/koji/taskinfo?taskID=343219 [2] https://bugzilla.redhat.com/show_bug.cgi?id=415841 [3] https://bugzilla.redhat.com/show_bug.cgi?id=240774 [4] https://bugzilla.redhat.com/show_bug.cgi?id=243081 [5] https://bugzilla.redhat.com/show_bug.cgi?id=370931
Okay, we've made some significant progress on the storage front in the past week and a half now as well. The fixes aren't all yet into rawhide, but locally, I've now got multiple drives that used to hit either 'failed to login[6]' or 'giving up on config rom[7]' (both of which render the drive useless) now working perfectly.
[6] https://bugzilla.redhat.com/show_bug.cgi?id=238606 [7] https://bugzilla.redhat.com/show_bug.cgi?id=429598
Stay tuned for test kernels if you've got storage issues...
- -- Jarod Wilson jwilson@redhat.com
-
Andrew Farris -
Hans de Goede -
Jarod Wilson -
Les Mikesell -
Mike Chambers