Most recent kernel where this bug did *NOT* occur: 2.6.21-rc3-mm2 Distribution: Fedora devel Hardware Environment: AMD X2 + CK804 + E-MU 1820m + PVR150 Software Environment: N/A Problem Description: snd_emu10k & ivtv are not finding their firmware files anymore Mar 20 20:28:35 rousalka kernel: ivtv: ==================== START INIT IVTV ==================== Mar 20 20:28:35 rousalka kernel: ivtv: version 1.0.0 (2.6.21-0.23.rc4.mm1.fc7 SMP mod_unload ) loading Mar 20 20:28:35 rousalka kernel: sr0: scsi3-mmc drive: 40x/40x writer cd/rw xa/form2 cdda tray Mar 20 20:28:35 rousalka kernel: Uniform CD-ROM driver Revision: 3.20 Mar 20 20:28:35 rousalka kernel: eth0: forcedeth.c: subsystem: 01458:e000 bound to 0000:00:0a.0 Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt 0000:01:08.2[B] -> Link [APC2] -> GSI 17 (level, low) -> IRQ 17 Mar 20 20:28:35 rousalka kernel: fw_ohci: Added fw-ohci device 0000:01:08.2, OHCI version 1.10 Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt Link [APC3] enabled at IRQ 18 Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt 0000:01:0a.0[A] -> Link [APC3] -> GSI 18 (level, low) -> IRQ 18 Mar 20 20:28:35 rousalka kernel: fw_ohci: Added fw-ohci device 0000:01:0a.0, OHCI version 1.10 Mar 20 20:28:35 rousalka kernel: ivtv0: Autodetected Hauppauge card (cx23416 based) Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt Link [APC4] enabled at IRQ 19 Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt 0000:01:07.0[A] -> Link [APC4] -> GSI 19 (level, low) -> IRQ 19 Mar 20 20:28:35 rousalka kernel: fw_core: created new fw device fw0 (0 config rom retries) Mar 20 20:28:35 rousalka kernel: fw_core: created new fw device fw1 (0 config rom retries) Mar 20 20:28:35 rousalka kernel: firmware_loading_store: unexpected value (0) Mar 20 20:28:35 rousalka kernel: ivtv0: unable to open firmware v4l-cx2341x-enc.fw (must be 376836 bytes) Mar 20 20:28:35 rousalka kernel: ivtv0: did you put the firmware in the hotplug firmware directory? Mar 20 20:28:35 rousalka kernel: ivtv0: Retry loading firmware Mar 20 20:28:35 rousalka kernel: firmware_loading_store: unexpected value (0) Mar 20 20:28:35 rousalka kernel: ivtv0: unable to open firmware v4l-cx2341x-enc.fw (must be 376836 bytes) Mar 20 20:28:35 rousalka kernel: ivtv0: did you put the firmware in the hotplug firmware directory? Mar 20 20:28:35 rousalka kernel: ivtv0: Error initializing firmware Mar 20 20:28:35 rousalka kernel: ivtv0: Error -19 on initialization Mar 20 20:28:35 rousalka kernel: ivtv: ==================== END INIT IVTV ==================== Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt 0000:02:00.0[A] -> Link [APC2] -> GSI 17 (level, low) -> IRQ 17 Mar 20 20:28:35 rousalka kernel: sky2 0000:02:00.0: v1.13 addr 0xe5000000 irq 17 Yukon-EC (0xb6) rev 1 Mar 20 20:28:35 rousalka kernel: sky2 eth1: addr 00:14:85:0c:3e:a4 Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt Link [APC1] enabled at IRQ 16 Mar 20 20:28:35 rousalka kernel: ACPI: PCI Interrupt 0000:01:08.0[A] -> Link [APC1] -> GSI 16 (level, low) -> IRQ 16 Mar 20 20:28:35 rousalka kernel: emu1010: Special config. Mar 20 20:28:35 rousalka kernel: emu1010: EMU_HANA_ID=0x1 Mar 20 20:28:35 rousalka kernel: firmware_loading_store: unexpected value (0) Mar 20 20:28:35 rousalka kernel: firmware: emu/hana.fw not found. Err=-2 Steps to reproduce: boot
It's broken for ivtv (v4l) and E-MU (alsa) so looks like it's general
Just to be clear : 1. it's not a userspace problem (rebooting on previous -mm works) 2. it's not a v4l problem : the two drivers on system that need firmware fail the same way : ivtv (v4) and emu1010 (alsa) So it's a pure 2.6.21-rc4-mm1 bug
I'm hit by this bug too. I don't know how but I even managed to get the same use-after-free bug : http://lkml.org/lkml/2007/03/22/62 I don't have time to test further but it may be interesting to try reverting this patch : hack-to-make-wireless-work.patch
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.21-rc4/2.6.21-rc4-mm1/hot-fixes/revert-driver-core-handles-kobject_uevent-failure-while-device_add.patch does not fix the problem (both for ivtv and emu1010)
Created attachment 10909 [details] kernel config
Created attachment 10910 [details] /var/log/dmesg
Reply-To: akpm@linux-foundation.org crap, ipw2200 is the only thing I have formware for, and it works OK. Can you please turn on DEBUG_KOBJECT, see if we can get any interesting messages. Or, better, see the thread on linux-kernel today (Subject: Re: 2.6.21-rc4-mm1) in which Cornelia Huck has generated a debug patch. If you can continue this investigation in that email thread it would be better, thanks.
This combo works: Patch1: ftp://ftp.kernel.org/pub/linux/kernel/v2.6/testing/patch-2.6.21-rc4.bz2 Patch2: ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.21-rc4/2.6.21-rc4-mm1/2.6.21-rc4-mm1.bz2 Patch3: ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.21-rc4/2.6.21-rc4-mm1/hot-fixes/mm-debug-check-for-the-fault-vs-invalidate-race.patch Patch4: ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.21-rc4/2.6.21-rc4-mm1/hot-fixes/mm-fix-fault-vs-invalidate-race-for-linear-mappings-fix.patc h Patch5: ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.21-rc4/2.6.21-rc4-mm1/hot-fixes/mm-fix-fault-vs-invalidate-race-for-linear-mappings-fix-2.pa tch #http://marc.info/?l=linux-kernel&m=117464447600932&w=2 Patch6: firmware-kobjadd.patch Patch7: ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.21-rc4/2.6.21-rc4-mm1/hot-fixes/tty-in-tiocsctty-when-we-steal-a-tty-hang-it-up-fix.patch Patch8: http://ck.kolivas.org/patches/staircase-deadline/2.6.21-rc4-mm1-rsdl-0.33.patch
Reply-To: akpm@linux-foundation.org On Fri, 23 Mar 2007 12:58:07 -0700 bugme-daemon@bugzilla.kernel.org wrote: > This combo works: I'm lost. What can we conclude from your observation??
I conclude that the patch Cornelia posted here http://marc.info/?l=linux-kernel&m=117464447600932&w=2 works on my system (with latest RSDL version, could test without it if you want) I just wanted to be clear what exact patch combo I tested, since I removed one of the hot fixes and they've been changing a lot lately
fixed in 2.6.21-rc5-mm2