Skip to contentSkip to conversation

Linux 2.6.10-rc3

DISCUSSION David Brownell 3 participants 3 messages Dec 5, 2004

lore
Share & tools
lore
#0 lore raw
Quoted 11 lines
From:       Martin Josefsson <gandalf () wlug ! westbo ! se>
Date:       2004-12-04 21:42:11
Quoted 4 lines · Alex Romosan

thank you. the laptop wakes up now but i get the following when it resumes (this is the output from dmesg):

scheduling while atomic: sleepbtn.sh/0x00000001/3201

...
That's an usb2.0 bug, the ehci driver sleeps when it can't sleep.

Who changed it so that context was no longer allowed to sleep???

That's a very recent change ... I've done a fair amount of testing in previous kernels and _never_ got that message on that path.

Why was that changed? Are you sure it's not just a bug higher up in the call stack? Classically(*), both suspend() and resume() methods are called in contexts that can sleep, so that's a big change I'd expect to impact other drivers too. In fact that'd explain a lot of other messages I saw reported on the list...

- Dave

(*) Since APM days if not before.

#1 lore raw
Quoted 6 lines

Why was that changed? Are you sure it's not just a bug higher up in the call stack? Classically(*), both suspend() and resume() methods are called in contexts that can sleep, so that's a big change I'd expect to impact other drivers too. In fact that'd explain a lot of other messages I saw reported on the list...

Yes, lots of drivers are expected to be able to schedule in suspend and resume requests. If that was broken upstream, then this is a big BUG.

Quoted 8 lines

- Dave

(*) Since APM days if not before.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 1 line
Benjamin Herrenschmidt <[email protected]>
#2 lore raw
In reply to David Brownell
Quoted 6 lines · David Brownell
Quoted 4 lines
From:       Martin Josefsson <gandalf () wlug ! westbo ! se>
Date:       2004-12-04 21:42:11

That's an usb2.0 bug, the ehci driver sleeps when it can't sleep.

Who changed it so that context was no longer allowed to sleep???

suspend and resume methods must be able to sleep because you may need to talk to external hardware, wait for queues to drain, etc which may in turn require kernel threads to run.

We must be able to sleep in suspend/resume methods. PCMCIA requires it, as do other subsystems.

Signature · 4 lines
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                 2.6 Serial core