USB-OHCI + USB broken in 2.4.14/15pre2?
Resume from suspend does not work correctly on my machine. Here's what the log says:
----
Nov 18 04:33:52 oland cardmgr[188]: executing: './network suspend eth1' Nov 18 04:33:52 oland kernel: usb-ohci.c: USB suspend: usb-00:01.2 Nov 18 04:33:52 oland kernel: NETDEV WATCHDOG: eth1: transmit timed out Nov 18 04:33:52 oland kernel: eth1: Transmit timeout. Nov 18 04:33:52 oland kernel: usb-ohci.c: Bus suspended Nov 18 04:33:52 oland kernel: usb-ohci.c: USB suspend: usb-00:01.3 Nov 18 04:33:53 oland kernel: usb-ohci.c: Bus suspended Nov 18 04:33:57 oland logger: Storing ALSA mixer settings...done. Nov 18 04:33:58 oland logger: Shutting down ALSA sound driver (version 0.9.0beta9): done. Nov 18 04:33:59 oland apmd[349]: User Suspend Nov 18 04:34:09 oland kernel: usb-ohci.c: odd PCI resume for usb-00:01.2 Nov 18 04:34:09 oland kernel: usb-ohci.c: odd PCI resume for usb-00:01.3 Nov 18 04:34:09 oland kernel: APIC error on CPU0: 00(40) Nov 18 04:34:10 oland cardmgr[188]: executing: './network resume eth1' Nov 18 04:34:11 oland logger: ALSA driver (version 0.9.0beta9) is already running. Nov 18 04:34:13 oland kernel: PCI: Found IRQ 11 for device 00:01.4 Nov 18 04:34:13 oland kernel: PCI: Sharing IRQ 11 with 00:03.0 Nov 18 04:34:14 oland apmd[349]: Normal Resume after 00:00:15 (99%
unknown) AC power
----
(Don't care about eth1 errors. ALSA doesn't matter either, although I never saw this behavior (driver already running after stopping them?!) before I started using the USB drivers. However, don't bother....)
Anyway: Machine gets up, but - needless to say - my usb mouse (or any other usb device) does not work after resume:
----
Nov 18 04:44:40 oland kernel: hub.c: USB new device connect on bus2/1, assigned device number 2 Nov 18 04:44:43 oland kernel: usb_control/bulk_msg: timeout Nov 18 04:44:43 oland kernel: usb-ohci.c: unlink URB timeout Nov 18 04:44:43 oland kernel: usb.c: USB device not accepting new address=2 (error=-110) Nov 18 04:44:43 oland kernel: hub.c: USB new device connect on bus2/1, assigned device number 3 Nov 18 04:44:46 oland kernel: usb_control/bulk_msg: timeout Nov 18 04:44:46 oland kernel: usb-ohci.c: unlink URB timeout Nov 18 04:44:46 oland kernel: usb.c: USB device not accepting new
address=3 (error=-110)
----
The APIC error (shown in the first log) does not appear if I compile the kernel without APIC (duh!). Situation regarding the USB is the same in either case.
Although the bus gets suspended (ohci_pci_suspend() works correctly, I doublechecked this by simply replacing dbg() with info() in ohci_pci_suspend(), thus the log entry at 04:33:53), it seems that the control flags are for some reason (already) set to OHCI_USB_OPER (as the only value capable of causing "odd PCI resume" messages) when the resume procedure in usb_ohci.c is called.
I tried to "recover" this behavior by temporarily patching ohci_pci_resume() so that it does a brutal hc_restart(ohci) instead of nothing when detecting this "odd PCI resume" situation - without any success.
Configuration (imho relevant parts):
Laptop (Celeron 1Ghz) with SiS630 chipset, including two SiS 7001 USB controllers. USB controllers share IRQ (11) with cardbus and sound (and unused winmodem).
# lspci -vv 00:01.2 USB Controller: Silicon Integrated Systems [SiS] 7001 (rev 07) (prog-if 10 [OHCI]) Subsystem: Silicon Integrated Systems [SiS] 7001 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR+ FastB2B- Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- Latency: 32 (20000ns max), cache line size 08 Interrupt: pin D routed to IRQ 11 Region 0: Memory at dffd0000 (32-bit, non-prefetchable) [size=4K]
00:01.3 USB Controller: Silicon Integrated Systems [SiS] 7001 (rev 07) (prog-if 10 [OHCI])
Subsystem: Silicon Integrated Systems [SiS]: Unknown device 7000
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr-
Stepping- SERR+ FastB2B-
Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
Latency: 32 (20000ns max), cache line size 08 Interrupt: pin D routed to IRQ 11 Region 0: Memory at dffe0000 (32-bit, non-prefetchable) [size=4K]
I am running kernel 2.4.15pre2 but I did not see any relevant changes in usb-ohci since then up to pre6.
Any further information will, of course, be supplied on request.
The Usual Questions(tm): Did I miss anything obvious? Is this a known bug? Any ideas?
Thomas
Signature · 3 lines
Thomas Winischhofer Vienna/Austria Check it out: mailto:[email protected] *** http://www.webit.com/tw
In article <[email protected]> you write:
Quoted 5 lines
I tried to "recover" this behavior by temporarily patching
ohci_pci_resume()so that it does a brutalhc_restart(ohci)instead of nothing when detecting this "odd PCI resume" situation - without any success.
I would suggest trying to do a "pci_enable_device(dev);" at the very top of ohci_pci_resume(). It sounds like your suspend/resume doesn't re-enable PCI interrupt routing, and doing the device enable will make the kernel re-route the interrupt for you.
If that helps, please send me the tested patch, and forward it to the appropriate USB people too.
Linus