Moderator Control Panel ]

TBS Driver on Fed x86_64 3.9.11 Breaks existing ngene driver

Re: TBS Driver on Fed x86_64 3.9.11 Breaks existing ngene dr

Postby pinsb » Thu Aug 08, 2013 9:03 pm

cody Wrote:hmm, can you revert back to kernel drivers in which it's working and check if the same firmware file is loaded, i.e. "ngene: Loading firmware file ngene_15.fw.", in the mean time i will compare again the code in 3.9.11 kernel for "ngene" driver with what's inside TBS package - as i mentioned there are almost no differences and that's why i expect it's something very small - maybe even my idea i've just mentioned that different than "ngene_15.fw" firmware file is loaded.


I'll do that tonight.....

Thought you might find this interesting.

[116777.086287] irq 35: nobody cared (try booting with the "irqpoll" option)
[116777.086292] Pid: 0, comm: swapper/0 Tainted: PF O 3.9.11-200.fc18.x86_64 #1
[116777.086293] Call Trace:
[116777.086295] <IRQ> [<ffffffff810ef6cd>] __report_bad_irq+0x3d/0xe0
[116777.086303] [<ffffffff810efb77>] note_interrupt+0x1b7/0x200
[116777.086377] [<ffffffffa035adfd>] ? _nv014716rm+0xad/0x1c2 [nvidia]
[116777.086380] [<ffffffff810ed357>] handle_irq_event_percpu+0xa7/0x1f0
[116777.086382] [<ffffffff810ed4e1>] handle_irq_event+0x41/0x70
[116777.086385] [<ffffffff810f06f9>] handle_fasteoi_irq+0x59/0x100
[116777.086389] [<ffffffff810161af>] handle_irq+0xbf/0x150
[116777.086391] [<ffffffff81067861>] ? irq_enter+0x51/0x90
[116777.086395] [<ffffffff81669bda>] do_IRQ+0x5a/0xe0
[116777.086398] [<ffffffff8165fbed>] common_interrupt+0x6d/0x6d
[116777.086444] [<ffffffffa035adfd>] ? _nv014716rm+0xad/0x1c2 [nvidia]
[116777.086491] [<ffffffffa035adfd>] ? _nv014716rm+0xad/0x1c2 [nvidia]
[116777.086609] [<ffffffffa06f76b5>] ? _nv004049rm+0x48f1/0xaef0 [nvidia]
[116777.086728] [<ffffffffa06d5162>] ? _nv009861rm+0x57/0xa1 [nvidia]
[116777.086846] [<ffffffffa06f7755>] ? _nv004049rm+0x4991/0xaef0 [nvidia]
[116777.086964] [<ffffffffa06f7bbe>] ? _nv004049rm+0x4dfa/0xaef0 [nvidia]
[116777.087062] [<ffffffffa07ad725>] ? _nv013367rm+0xc5/0x4e2 [nvidia]
[116777.087159] [<ffffffffa07cf962>] ? _nv013343rm+0xf2/0x137 [nvidia]
[116777.087219] [<ffffffffa092d700>] ? _nv001110rm+0x143/0x1c2 [nvidia]
[116777.087279] [<ffffffffa09341e1>] ? rm_isr_bh+0x23/0x73 [nvidia]
[116777.087337] [<ffffffffa0952692>] ? nv_kern_isr_bh+0x42/0x70 [nvidia]
[116777.087340] [<ffffffff81067afa>] ? tasklet_action+0x6a/0xe0
[116777.087341] [<ffffffff81067678>] ? __do_softirq+0xe8/0x230
[116777.087344] [<ffffffff81067945>] ? irq_exit+0xa5/0xb0
[116777.087346] [<ffffffff81091be2>] ? scheduler_ipi+0x62/0x120
[116777.087350] [<ffffffff810382ea>] ? smp_reschedule_interrupt+0x2a/0x30
[116777.087352] [<ffffffff81668e9d>] ? reschedule_interrupt+0x6d/0x80
[116777.087353] <EOI> [<ffffffff8150672b>] ? cpuidle_wrap_enter+0x5b/0xa0
[116777.087359] [<ffffffff810f4f2d>] ? rcu_idle_exit+0x6d/0xc0
[116777.087362] [<ffffffff8101d138>] ? cpu_idle+0xb8/0x120
[116777.087365] [<ffffffff81644712>] ? rest_init+0x72/0x80
[116777.087368] [<ffffffff81d02ed6>] ? start_kernel+0x3f2/0x3ff
[116777.087371] [<ffffffff81d028e3>] ? repair_env_string+0x5e/0x5e
[116777.087373] [<ffffffff81d025dc>] ? x86_64_start_reservations+0x2a/0x2c
[116777.087375] [<ffffffff81d026cf>] ? x86_64_start_kernel+0xf1/0x100
[116777.087377] handlers:
[116777.087381] [<ffffffffa0e04940>] saa716x_tbs6280_pci_irq [saa716x_tbs_dvb]
[116777.087382] Disabling IRQ #35
[116777.596028] irq 18: nobody cared (try booting with the "irqpoll" option)
[116777.596035] Pid: 0, comm: swapper/3 Tainted: PF O 3.9.11-200.fc18.x86_64 #1
[116777.596038] Call Trace:
[116777.596040] <IRQ> [<ffffffff810ef6cd>] __report_bad_irq+0x3d/0xe0
[116777.596051] [<ffffffff810efb77>] note_interrupt+0x1b7/0x200
[116777.596057] [<ffffffff81506720>] ? cpuidle_wrap_enter+0x50/0xa0
[116777.596061] [<ffffffff81505dc0>] ? cpufreq_frequency_table_update_policy_cpu+0x80/0x80
[116777.596064] [<ffffffff810ed357>] handle_irq_event_percpu+0xa7/0x1f0
[116777.596068] [<ffffffff81505dc0>] ? cpufreq_frequency_table_update_policy_cpu+0x80/0x80
[116777.596072] [<ffffffff810ed4e1>] handle_irq_event+0x41/0x70
[116777.596076] [<ffffffff810f06f9>] handle_fasteoi_irq+0x59/0x100
[116777.596081] [<ffffffff810161af>] handle_irq+0xbf/0x150
[116777.596088] [<ffffffff81663992>] ? __atomic_notifier_call_chain+0x12/0x20
[116777.596092] [<ffffffff816639b6>] ? atomic_notifier_call_chain+0x16/0x20
[116777.596096] [<ffffffff81669bda>] do_IRQ+0x5a/0xe0
[116777.596100] [<ffffffff8165fbed>] common_interrupt+0x6d/0x6d
[116777.596102] <EOI> [<ffffffff81506720>] ? cpuidle_wrap_enter+0x50/0xa0
[116777.596109] [<ffffffff81506719>] ? cpuidle_wrap_enter+0x49/0xa0
[116777.596112] [<ffffffff81506780>] cpuidle_enter_tk+0x10/0x20
[116777.596116] [<ffffffff81506445>] cpuidle_idle_call+0xa5/0x1d0
[116777.596120] [<ffffffff8101d12f>] cpu_idle+0xaf/0x120
[116777.596124] [<ffffffff8164f156>] start_secondary+0x24f/0x251
[116777.596126] handlers:
[116777.596130] [<ffffffff81470a80>] usb_hcd_irq
[116777.596133] [<ffffffff81470a80>] usb_hcd_irq
[116777.596135] Disabling IRQ #18
pinsb
 
Posts: 13
Joined: Mon Aug 05, 2013 7:30 am

Re: TBS Driver on Fed x86_64 3.9.11 Breaks existing ngene dr

Postby cody » Fri Aug 09, 2013 6:43 pm

pinsb Wrote:Thought you might find this interesting.

[116777.086287] irq 35: nobody cared (try booting with the "irqpoll" option)
snip....


that's not new, full details and solution can be found here:
viewtopic.php?f=52&t=7631#p24903

however, the case with 2400i it's really confusing - if you compare the "ngene" source code inside TBS Linux driver package and in kernel 3.9.11 you will see that they are practically identical - there are few different lines, but those are mainly due to different/many kernels that TBS Linux driver package supports, while code in kernel 3.9.11 more or less needs to support just that kernel. so, i still can't spot the small difference that make the difference - in fact i can't spot any difference related to 2400i settings at all - they are exactly the same. so, it still continue to seems like firmware problem - maybe there are different versions of "ngene_15.fw", no matter that the file-name is the same. so, i believe "md5sum" of "ngene_15.fw" in working and not working configuration should be checked too.

[EDIT] please, collect logs with the following:

http://www.basicupload.com/5dopscbcai0g

i guess it would be to send them via email. thanks

so, i don't expect it will solve the problem, but i add to some debug messages, as well enable "ngene" driver debugging to double confirm i understand correct the current error, which seems like "ngene" firmware can't read from "drxd" chip - it really seems like firmware problem rather than driver problem and that is what makes sense, because driver code is the same.
cody
 
Posts: 627
Joined: Tue Apr 13, 2010 11:20 pm

Re: TBS Driver on Fed x86_64 3.9.11 Breaks existing ngene dr

Postby cody » Tue Aug 13, 2013 4:15 am

thanks for the new sets of logs you provided via email - from them it seems more strange than i was expecting - it looks like the first read via the "ngene"firmware passes and then every consecutive one fails. so, since there is no relevant differences that can spot between the code in kernel 3.9.11 and TBS driver package, i start to believe that maybe Fedora include some special patch for the problem to their kernel build. so, that's why i'm suggesting and please do the following test (as 'root' user):

# git clone git://linuxtv.org/media_build.git
# cd media_build
# make download
# make untar
# make
# make install

that will download, build and install latest V4L code - what i want to confirm is if that code works with 2400i for you or not - as usual you can provide logs from it via email. thank you.
cody
 
Posts: 627
Joined: Tue Apr 13, 2010 11:20 pm

Previous

Return to Linux

Who is online

Users browsing this forum: No registered users and 3 guests