Distribution: Debian 3.1 (sarge) Hardware Environment: Laptop, AMD Mobile Athlon 1400XP (1200 Mhz), VIA chipset, 256Mb RAM. Tascam US-122 USB Audio/Midi interface. Software Environment: jackd 0.99.0 Problem Description: Attempting to run jackd when CPU frequency scaling is operative results in jackd aborting with a "killed by watchdog" error and sometimes this causes the USB host controller to halt with the following message on the system console/syslog: Mar 26 18:36:04 alfroid kernel: ep=8 stalled with status=-63 Mar 26 18:36:28 alfroid kernel: Badness in kref_get at lib/kref.c:32 Mar 26 18:36:28 alfroid kernel: [kref_get+68/80] kref_get+0x44/0x50 Mar 26 18:36:28 alfroid kernel: [pg0+252059655/1069745152] usb_get_urb+0x17/0x30 [usbcore] Mar 26 18:36:28 alfroid kernel: [pg0+252057143/1069745152] hcd_submit_urb+0xb7/0x1e0 [usbcore] Mar 26 18:36:28 alfroid kernel: [autoremove_wake_function+47/96] autoremove_wake_function+0x2f/0x60 Mar 26 18:36:28 alfroid kernel: [pg0+252060209/1069745152] usb_submit_urb+0x211/0x390 [usbcore] Mar 26 18:36:28 alfroid kernel: [pg0+252334758/1069745152] usX2Y_urb_submit+0x46/0x70 [snd_usb_usx2y] Mar 26 18:36:28 alfroid kernel: [pg0+252335459/1069745152] i_usX2Y_urb_complete+0x133/0x1f0 [snd_usb_usx2y] Mar 26 18:36:28 alfroid kernel: [pg0+252058805/1069745152] usb_hcd_giveback_urb+0x25/0x60 [usbcore] Mar 26 18:36:28 alfroid kernel: [pg0+251882225/1069745152] uhci_finish_urb+0x51/0x70 [uhci_hcd] Mar 26 18:36:28 alfroid kernel: [pg0+251882325/1069745152] uhci_finish_completion+0x45/0x60 [uhci_hcd] Mar 26 18:36:28 alfroid kernel: [pg0+251882839/1069745152] uhci_irq+0x197/0x1f0 [uhci_hcd] Mar 26 18:36:28 alfroid kernel: [pg0+252058925/1069745152] usb_hcd_irq+0x3d/0x70 [usbcore] Mar 26 18:36:28 alfroid kernel: [handle_IRQ_event+48/112] handle_IRQ_event+0x30/0x70 Mar 26 18:36:28 alfroid kernel: [__do_IRQ+222/352] __do_IRQ+0xde/0x160 Mar 26 18:36:28 alfroid kernel: [__do_softirq+120/144] __do_softirq+0x78/0x90 Mar 26 18:36:28 alfroid kernel: [do_IRQ+35/64] do_IRQ+0x23/0x40 Mar 26 18:36:28 alfroid kernel: [common_interrupt+26/32] common_interrupt+0x1a/0x20 Mar 26 18:36:28 alfroid kernel: uhci_hcd 0000:00:11.2: host system error, PCI problems? Mar 26 18:36:28 alfroid kernel: uhci_hcd 0000:00:11.2: host controller halted, very bad! Mar 26 18:36:58 alfroid kernel: hub 1-0:1.0: state 5 ports 2 chg 0000 evt 0004 Mar 26 18:36:58 alfroid kernel: uhci_hcd 0000:00:11.2: port 2 portsc 008a,00 Mar 26 18:36:58 alfroid kernel: hub 1-0:1.0: port 2, status 0100, change 0003, 12 Mb/s Mar 26 18:36:58 alfroid kernel: usb 1-2: USB disconnect, address 3 Mar 26 18:36:58 alfroid kernel: usb 1-2: usb_disable_device nuking all URBs Mar 26 18:36:58 alfroid kernel: uhci_hcd 0000:00:11.2: shutdown urb c9c8d560 pipe 40020380 ep4in-intr Mar 26 18:36:59 alfroid kernel: uhci_hcd 0000:00:11.2: suspend_hc Mar 26 18:37:13 alfroid kernel: uhci_hcd 0000:00:11.2: wakeup_hc Mar 26 18:37:14 alfroid kernel: uhci_hcd 0000:00:11.2: host system error, PCI problems? Mar 26 18:37:14 alfroid kernel: uhci_hcd 0000:00:11.2: host controller halted, very bad! After this it appears the kernel tries to re-start the controller but without success as a series of messages get repeated: Mar 26 18:37:27 alfroid kernel: uhci_hcd 0000:00:11.2: suspend_hc Mar 26 18:37:28 alfroid kernel: uhci_hcd 0000:00:11.2: wakeup_hc Mar 26 18:37:28 alfroid kernel: uhci_hcd 0000:00:11.2: host system error, PCI problems? Mar 26 18:37:28 alfroid kernel: uhci_hcd 0000:00:11.2: host controller halted, very bad! Mar 26 18:37:28 alfroid kernel: uhci_hcd 0000:00:11.3: wakeup_hc Mar 26 18:37:30 alfroid kernel: uhci_hcd 0000:00:11.2: suspend_hc Mar 26 18:37:30 alfroid kernel: uhci_hcd 0000:00:11.2: wakeup_hc Mar 26 18:37:30 alfroid kernel: uhci_hcd 0000:00:11.2: host system error, PCI problems? Mar 26 18:37:30 alfroid kernel: uhci_hcd 0000:00:11.2: host controller halted, very bad! The only way out of this seems to be to reboot. I have since learned that the design of jackd means it will not work with CPU frequency scaling but still the messages above suggest a bug within the USB subsystem of the kernel. Steps to reproduce: Run jackd on above hardware with CPU frequence scaling enabled and wait.
Mar 26 18:37:14 alfroid kernel: uhci_hcd 0000:00:11.2: host system error, PCI problems? Mar 26 18:37:14 alfroid kernel: uhci_hcd 0000:00:11.2: host controller halted, very bad! -- I believe these are mobo issues. Could just be hardware having problems once cpufreq was screwed with. Don't know. Try with cpu frequency disabled, same problems? Thanks, Nish
Can you please test 2.6.12-rc5 and update this bug? Thanks.
If this is still a problem, please reopen with the requested info.