2004-06-17 15:40:51

by Carsten Rietzschel

[permalink] [raw]
Subject: ACPI S3 - USB resume problem (kernel 2.6.7)

Hello,

I tried to get S3 "Suspend to RAM" working. Almost everything works now, but
USB doesn't respond after resuming (kernel 2.6.7).

The log shows (log appended; lspci -v output also):

usb 1-1: control timeout on ep0out
uhci_hcd 0000:00:10.0: Unlink after no-IRQ? Different ACPI or APIC settings
may help.
usb 1-1: control timeout on ep0out
usb 1-1: device not accepting address 2, error -110
usb 1-1: new low speed USB device using address 3
usb 1-1: control timeout on ep0out
usb 1-1: control timeout on ep0out
usb 1-1: device not accepting address 3, error -110

Noticed that in /proc/interrupts the values for uhci_hcd are not incremented
after resume. So are no IRQs where received (is that right ?).
What could be reason ?

So far I tried:
- set CONFIG_USB_DEBUG = y
- disable/enable APIC as suggested
("Unlink after no-IRQ? Different ACPI or APIC settings" goes away, timeout
stays)
- disabled USB 1.1 Controller or USB 2.0 Controller in BIOS
- disabled preemption
- compiled USB-stuff into kernel instead using as modules
( even tried to unload before suspend and reloaded after resume )
- ensured CONFIG_PM is on and resume/suspend in is
called /drivers/usb/host/uhci-hcd.c
- forced reset_hc(uhci) in uhci_suspend() instead of suspend_hc()
( comment says it's for some broken motherboards )
- I saw CONFIG_USB_SUSPEND in some source files under host and core.
I enabled it directly in these files and recompiled/restarted. No success.
(doesn't harm, but seems not ready to use:
- it's not configureable ;)
- usb_suspend/resume_device() functions are nowhere)

Nothing helped :/
Still timeout messages and no USB device works after resume.

Anybody any idea where to look next or what to search for, or even a fix ?
If you want more infos I will tell you :)

Regards,
Carsten


Attachments:
(No filename) (1.78 kB)
usb_messages.log (14.48 kB)
lspci.log (3.83 kB)
Download all attachments

2004-06-17 19:34:22

by Alexander Gran

[permalink] [raw]
Subject: Re: ACPI S3 - USB resume problem (kernel 2.6.7)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am Donnerstag, 17. Juni 2004 17:44 schrieb Carsten Rietzschel:
> Noticed that in /proc/interrupts the values for uhci_hcd are not
> incremented after resume. So are no IRQs where received (is that right ?).
> What could be reason ?

No Idea. I also tried to get this working, without success (And no more time
at the moment to dig deeper). The e1000 driver has the same problem. No
Interrupts on RX. Someone suggested to "Hook the driver to the timer
interrupt and see if that works at least somehow", however I'm unsure how to
do that ;)

regards
Alex

- --
Encrypted Mails welcome.
PGP-Key at http://zodiac.dnsalias.org/misc/pgpkey.asc | Key-ID: 0x6D7DD291
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA0e+4/aHb+2190pERAseiAJ9vtdew4kLYFdeGUu735Cb7As8afwCff9qN
VVtwNWQpVIWz3yLR5P/O7m8=
=BA0Z
-----END PGP SIGNATURE-----

2004-06-18 11:37:46

by Carsten Rietzschel

[permalink] [raw]
Subject: Re: ACPI S3 - USB resume problem (kernel 2.6.7)

Hello,

thanks Alex. You're right - the e1000 doesn't work too.
Also all other PCI-devices like firewire, USB and soundcard do not.
So it's not a problem of USB, but of ACPI-PCI.

After reading a few mails of the acpi-mailinglist and some bug reports, I know
now, I'm not alone with this problem :/

I'll have a closer look to acpi-list & bug reports.
Maybe someone here, has found a solution or has some hints what to do next ???

Regards,
Carsten


Am Donnerstag, 17. Juni 2004 21:23 schrieb Alexander Gran:
> Am Donnerstag, 17. Juni 2004 17:44 schrieb Carsten Rietzschel:
> > Noticed that in /proc/interrupts the values for uhci_hcd are not
> > incremented after resume. So are no IRQs where received (is that right
> > ?). What could be reason ?
>
> No Idea. I also tried to get this working, without success (And no more
> time at the moment to dig deeper). The e1000 driver has the same problem.
> No Interrupts on RX. Someone suggested to "Hook the driver to the timer
> interrupt and see if that works at least somehow", however I'm unsure how
> to do that ;)
>
> regards
> Alex

2004-06-18 11:58:20

by Alexander Gran

[permalink] [raw]
Subject: Re: ACPI S3 - USB resume problem (kernel 2.6.7)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

My ACPI additionall want's to put the devices to state 2 instead of 3 when
going to STR. I have hacked up the pcicode to convert state 2 to state 3
transitions, but this cannot be the solution, i suppose.
However acpi still fucks up the interrupts. Perhaps it's an dsdt problem.

Alex

Am Freitag, 18. Juni 2004 13:41 schrieb Carsten Rietzschel:
> Hello,
>
> thanks Alex. You're right - the e1000 doesn't work too.
> Also all other PCI-devices like firewire, USB and soundcard do not.
> So it's not a problem of USB, but of ACPI-PCI.
>
> After reading a few mails of the acpi-mailinglist and some bug reports, I
> know now, I'm not alone with this problem :/
>
> I'll have a closer look to acpi-list & bug reports.
> Maybe someone here, has found a solution or has some hints what to do next
> ???
>
> Regards,
> Carsten
>
> Am Donnerstag, 17. Juni 2004 21:23 schrieb Alexander Gran:
> > Am Donnerstag, 17. Juni 2004 17:44 schrieb Carsten Rietzschel:
> > > Noticed that in /proc/interrupts the values for uhci_hcd are not
> > > incremented after resume. So are no IRQs where received (is that right
> > > ?). What could be reason ?
> >
> > No Idea. I also tried to get this working, without success (And no more
> > time at the moment to dig deeper). The e1000 driver has the same problem.
> > No Interrupts on RX. Someone suggested to "Hook the driver to the timer
> > interrupt and see if that works at least somehow", however I'm unsure how
> > to do that ;)
> >
> > regards
> > Alex
>
> -
> 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/

- --
Encrypted Mails welcome.
PGP-Key at http://zodiac.dnsalias.org/misc/pgpkey.asc | Key-ID: 0x6D7DD291
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA0tcS/aHb+2190pERAh4YAKCZwzHKhU19deLn16eRpJdj4kyiwACfVnHn
tGDWuxzBPCb0H5Ipf00wMAI=
=bNOL
-----END PGP SIGNATURE-----