Em Sun, 27 Jul 2014 10:44:53 -0300 Mauro Carvalho Chehab <m.chehab@xxxxxxxxxxx> escreveu: > Em Sun, 27 Jul 2014 11:32:26 +0200 > Matthias Schwarzott <zzam@xxxxxxxxxx> escreveu: > > > > > Hi Mauro. > > > > On 26.07.2014 21:27, Mauro Carvalho Chehab wrote: > > > Tried to apply your patch series, but there's something wrong on it. > > > > > > See the enclosed logs. I suspect that you missed a patch adding the > > > proper tuner for this device. > > > > well, tuner_type is set to TUNER_ABSENT, and for me the oops does not > > happen. > > > > > [76185.463359] BUG: unable to handle kernel NULL pointer dereference at (null) > > > [76185.465295] IP: [<ffffffffa048dec8>] cx231xx_usb_probe+0x5a8/0xb20 [cx231xx] > > > [76185.465654] cx231xx #0: can't change interface 5 alt no. to 0 (err=-22) > > > [76185.465705] cx231xx #0: called cx231xx_uninit_vbi_isoc > > > [76185.465707] cx231xx #0: cx231xx_stop_stream():: ep_mask = 10 > > > [76185.465872] cx231xx #0: can't change interface 5 alt no. to 0 (err=-22) > > > [76185.474837] PGD 21c0fa067 PUD 2182b4067 PMD 0 > > > [76185.476593] Oops: 0000 [#1] SMP > > > [76185.478285] Modules linked in: cx25840 cx231xx(+) cx2341x tveeprom videobuf_vmalloc videobuf_core rc_core v4l2_common videodev media ip6table_filter ip6_tables bnep nouveau x86_pkg_temp_thermal coretemp i915 binfmt_misc kvm_intel kvm ttm i2c_algo_bit drm_kms_helper vfat fat drm arc4 iwldvm mac80211 snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic snd_hda_intel snd_hda_controller snd_hda_codec iwlwifi crct10dif_pclmul snd_hwdep crc32_pclmul cfg80211 snd_seq snd_seq_device snd_pcm crc32c_intel snd_timer btusb bluetooth iTCO_wdt iTCO_vendor_support mxm_wmi ghash_clmulni_intel lpc_ich i2c_i801 snd mei_me soundcore rfkill joydev serio_raw shpchp mei wmi mfd_core i2c_core video microcode r8169 mii [last unloaded: dvb_core] > > > [76185.487219] CPU: 0 PID: 12079 Comm: modprobe Not tainted 3.16.0-rc6+ #12 > > > [76185.488936] Hardware name: SAMSUNG ELECTRONICS CO., LTD. 550P5C/550P7C/SAMSUNG_NP1234567890, BIOS P05ABI.016.130917.dg 09/17/2013 > > > [76185.490710] task: ffff880211172ee0 ti: ffff88021866c000 task.ti: ffff88021866c000 > > > [76185.492502] RIP: 0010:[<ffffffffa048dec8>] [<ffffffffa048dec8>] cx231xx_usb_probe+0x5a8/0xb20 [cx231xx] > > > [76185.494321] RSP: 0018:ffff88021866fb50 EFLAGS: 00010246 > > > [76185.496037] RAX: ffff88021c34fc00 RBX: ffff8800c4f70000 RCX: 0000000000000006 > > > [76185.497781] RDX: 0000000000000004 RSI: ffff8800c4f70000 RDI: ffffffffa04a7508 > > > [76185.499548] RBP: ffff88021866fb90 R08: 0000000000000000 R09: 00000000000011d3 > > > [76185.501323] R10: 0000000000000000 R11: ffff88021866f85e R12: ffff88021bc20800 > > > [76185.503119] R13: ffff88021bc27c00 R14: ffff8800c4f70168 R15: ffff8802229ca000 > > > [76185.504934] FS: 00007f6c13b2e740(0000) GS:ffff88022f200000(0000) knlGS:0000000000000000 > > > [76185.506662] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > > > [76185.508406] CR2: 0000000000000000 CR3: 00000002222b8000 CR4: 00000000001407f0 > > > [76185.510169] Stack: > > > [76185.511921] ffff88020000b131 ffffffff00000004 ffff880000000001 ffff8802229ca090 > > > [76185.513725] ffff8802229ca000 ffffffffa04ab568 ffff88021bc27c30 ffffffffa04ab7a0 > > > [76185.515517] ffff88021866fbd8 ffffffff814eefcf 000000001866fbb0 ffff88021bc27c00 > > > [76185.517248] Call Trace: > > > [76185.518963] [<ffffffff814eefcf>] usb_probe_interface+0x1df/0x330 > > > [76185.520714] [<ffffffff8146605d>] driver_probe_device+0x12d/0x3d0 > > > [76185.522452] [<ffffffff814663d3>] __driver_attach+0x93/0xa0 > > > [76185.524203] [<ffffffff81466340>] ? __device_attach+0x40/0x40 > > > [76185.525959] [<ffffffff81463f03>] bus_for_each_dev+0x73/0xc0 > > > [76185.527599] [<ffffffff81465a5e>] driver_attach+0x1e/0x20 > > > [76185.529240] [<ffffffff81465640>] bus_add_driver+0x180/0x250 > > > [76185.537278] [<ffffffff81002144>] do_one_initcall+0xd4/0x210 > > > [76185.538809] [<ffffffff811b5fd4>] ? __vunmap+0x94/0x100 > > > [76185.540330] [<ffffffff81106767>] load_module+0x1ea7/0x24f0 > > > [76185.541806] [<ffffffff81102370>] ? store_uevent+0x70/0x70 > > > [76185.543255] [<ffffffff811f3200>] ? kernel_read+0x50/0x80 > > > [76185.544661] [<ffffffff81106f66>] SyS_finit_module+0xa6/0xd0 > > > [76185.546047] [<ffffffff816fc869>] system_call_fastpath+0x16/0x1b > > > [76185.547354] Code: 1d 00 00 85 ff 0f 8f 94 04 00 00 0f b6 93 bb 21 00 00 49 8b 87 68 03 00 00 48 89 de 48 c7 c7 08 75 4a a0 4c 8b 84 d0 a0 00 00 00 <49> 8b 00 4c 89 45 d0 48 8b 40 18 0f b6 40 02 0f b6 d0 66 89 93 > > > [76185.550208] RIP [<ffffffffa048dec8>] cx231xx_usb_probe+0x5a8/0xb20 [cx231xx] > > > [76185.551613] RSP <ffff88021866fb50> > > > [76185.553006] CR2: 0000000000000000 > > > [76185.559997] ---[ end trace 1e58ef345d6b3f24 ]--- > > > > > > > Is the order of the printk messages and of the oops message fixed? Are > > all messages printed before the oops also before the oops message? > > > > It seems to oops is at cx231xx_usb_probe+0x5a8/0xb20. > > > > > $ gdb drivers/media/usb/cx231xx/cx231xx.ko > > > GNU gdb (GDB) Fedora 7.7.1-15.fc20 > > > Copyright (C) 2014 Free Software Foundation, Inc. > > > License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> > > > This is free software: you are free to change and redistribute it. > > > There is NO WARRANTY, to the extent permitted by law. Type "show copying" > > [...] > > > Type "apropos word" to search for commands related to "word"... > > > Reading symbols from drivers/media/usb/cx231xx/cx231xx.ko...done. > > > (gdb) list *cx231xx_usb_driver_init+0x1e > > > > Could you retry this with the offset from above (0x5a8). > > (gdb) list *cx231xx_usb_probe+0x5a8 > 0x4ec8 is in cx231xx_usb_probe (drivers/media/usb/cx231xx/cx231xx-cards.c:1432). > 1427 uif = udev->actconfig->interface[dev->current_pcb_config. > 1428 hs_config_info[0].interface_info. > 1429 vanc_index + 1]; > 1430 > 1431 dev->vbi_mode.end_point_addr = > 1432 uif->altsetting[0].endpoint[isoc_pipe].desc. > 1433 bEndpointAddress; > 1434 > 1435 dev->vbi_mode.num_alt = uif->num_altsetting; > 1436 cx231xx_info("EndPoint Addr 0x%x, Alternate settings: %i\n", > > > I normally use nm to get the base addr of the function, then add the > > offset and then run "addr2line -ispf -e cx231xx.ko -a 0x....." > > Well, gdb is quicker, as it is just one command ;) I found the issue. there are just two interfaces on this device, but vanc_index is equal to 4 for this PCB. So, it tries to set VBI on interface 5, with obviously fails. It seems that the code is identifying a wrong PCB for this device. That's said, the probe logic is doing a crap job by allowing to go past the array. Regards, Mauro -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
