Skip to contentSkip to conversation

2.6.9-rc3: USB OHCI failure on suspend on AMD64

DISCUSSION Rafael J. Wysocki 3 participants 4 messages Sep 30, 2004 – Oct 1, 2004

lore
Share & tools
lore
#0 lore raw

Hi,

It seems there's a problem with USB OHCI driver that causes these traces to appear on suspend on an AMD64-based box:

..<7>PM: Image restored successfully.
PCI: Setting latency timer of device 0000:00:02.0 to 64 ohci_hcd 0000:00:02.0: HC died; cleaning up
usb 1-2: USB disconnect, address 3
Badness in hcd_endpoint_disable at drivers/usb/core/hcd.c:1310

 Call Trace:<7>Losing some ticks... checking if CPU frequency changed.
 <ffffffff803455fb>{hcd_endpoint_disable+107} 
<ffffffff80346729>{usb_disable_endpoint+41}
        <ffffffff803468aa>{usb_disable_device+26} 
<ffffffff803422fc>{usb_disconnect+188}
        <ffffffff803441b0>{hcd_panic+0} <ffffffff803441fa>{hcd_panic+74}
        <ffffffff8015824d>{worker_thread+733} 
<ffffffff80137c00>{default_wake_function+0}
        <ffffffff80137c00>{default_wake_function+0} 
<ffffffff80157f70>{worker_thread+0}
        <ffffffff8015f51d>{kthread+205} <ffffffff80111b43>{child_rip+8}
        <ffffffff8015f450>{kthread+0} <ffffffff80111b3b>{child_rip+0}

[-- several times the above --]

 Call Trace:<ffffffff80345c3e>{hcd_unlink_urb+494} 
<ffffffff803463dc>{usb_kill_urb+380}
        <ffffffff80345a37>{hcd_endpoint_disable+1191} 
<ffffffff8019ed6e>{invalidate_inode_buffers+14}
        <ffffffff80346729>{usb_disable_endpoint+41} 
<ffffffffa022374c>{:usbhid:hid_disconnect+44}
        <ffffffff8033f4e9>{usb_unbind_interface+73} 
<ffffffff8030b95e>{device_release_driver+94}
        <ffffffff8030baf4>{bus_remove_device+164} 
<ffffffff8030a898>{device_del+88}
        <ffffffff8034690b>{usb_disable_device+123} 
<ffffffff803422fc>{usb_disconnect+188}
        <ffffffff803441b0>{hcd_panic+0} <ffffffff803441fa>{hcd_panic+74}
        <ffffffff8015824d>{worker_thread+733} 
<ffffffff80137c00>{default_wake_function+0}
        <ffffffff80137c00>{default_wake_function+0} 
<ffffffff80157f70>{worker_thread+0}
        <ffffffff8015f51d>{kthread+205} <ffffffff80111b43>{child_rip+8}
        <ffffffff8015f450>{kthread+0} <ffffffff80111b3b>{child_rip+0}

[-- once again the above --]

The workaround is to explicitly unload the driver before suspend and reload it after resume. It is possible that it's fixed in -mm kernels, as I haven't seen such symptoms there, recently (since 2.6.9-rc2-mm1 at least).

Greets,
RJW

Signature · 3 lines
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
		-- Lewis Carroll "Alice's Adventures in Wonderland"
#1 lore raw

Hi!

Quoted 2 lines

It seems there's a problem with USB OHCI driver that causes these traces to appear on suspend on an AMD64-based box:

Check if OHCI's suspend/resume method is called correctly and that OHCI really acts on it.
Pavel

Signature · 2 lines
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!
#2 lore raw
In reply to Rafael J. Wysocki
Quoted 4 lines · Rafael J. Wysocki

Hi,

It seems there's a problem with USB OHCI driver that causes these traces to appear on suspend on an AMD64-based box:

Not all parts of power management integrate well yet with each other, though it's acting better. Plus there are some OHCI updates in the USB patches (you'll have those in the MM tree) that help address some of the more perverse ways that suspend-to-disk kicks USB around, and a few similar usbcore changes that need re-integration with the RC3 swsusp/pmdisk updates.

Keep using the "rmmod" workaround for now. And don't combine CONFIG_USB_SUSPEND with suspend-to-{disk,ram} unless you're feeling like debugging the results.

- Dave

#3 lore raw
Quoted 4 lines · David Brownell
Quoted 3 lines · Rafael J. Wysocki

Hi,

It seems there's a problem with USB OHCI driver that causes these traces

to

Quoted 9 lines
Quoted 1 line

appear on suspend on an AMD64-based box:

Not all parts of power management integrate well yet with each other, though it's acting better. Plus there are some OHCI updates in the USB patches (you'll have those in the MM tree) that help address some of the more perverse ways that suspend-to-disk kicks USB around, and a few similar usbcore changes that need re-integration with the RC3 swsusp/pmdisk updates.

Quoted 3 lines

Keep using the "rmmod" workaround for now. And don't combine CONFIG_USB_SUSPEND with suspend-to-{disk,ram} unless you're feeling like debugging the results.

I'm doing just this. And reporting. ;-)

Greets,
RJW

Signature · 3 lines
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
		-- Lewis Carroll "Alice's Adventures in Wonderland"