[coreboot] [SeaBIOS] USB Problems geode lx800

Kevin O'Connor kevin at koconnor.net
Thu Oct 11 03:57:38 CEST 2012


On Tue, Oct 09, 2012 at 02:31:05PM +0200, Christian Gmeiner wrote:
> 2012/10/9 Kevin O'Connor <kevin at koconnor.net>:
> > On Mon, Oct 08, 2012 at 02:14:03PM +0200, Christian Gmeiner wrote:
> >> HI all
> >>
> >> I am running into some usb problems with coreboot & seabios:
> > Can you set the debug level to 8 and post the whole log?  Also, for
> > timeout issues, having timestamps (via tools/readserial.py tool)
> > sometimes helps.
> 
> Attached

Hrmm, I guess you've disabled threads in your config?

>From your log, I see:

[14:23:00.5] ehci_send_bulk qh=0x000ef680 dir=0 data=0x00006d55 size=31^M
[14:23:00.5] ehci_send_bulk qh=0x000ef700 dir=128 data=0x00006e3c size=36^M
[14:23:00.5] ehci_send_bulk qh=0x000ef700 dir=128 data=0x00006d8f size=13^M
[14:23:05.5] WARNING - Timeout at ehci_wait_td:542!^M
[14:23:05.5] ehci pipe=0x000ef700 cur=00000000 tok=00000000 next=6b80 td=0x00006b80
+status=d0d80^M
[14:23:05.5] USB transmission failed^M
[14:23:05.5] Unable to configure USB MSC drive.^M

This looks like the transfer is stalling in the cdb_get_inquiry() call
as part of scsi_init_drive().  The drive is identified okay and at
least some transfers look okay, so it doesn't appear to be a
controller issue.

I've been told some USB sticks are very picky about init order, but
I'm surprised to see that this is occuring for you on a variety of USB
sticks.

It's possible to verify that a stall has occurred - print out
tds[i].token for all tds in ehci_send_bulk() on the error path.

Does this device still have problems if you soft-reset after a failed
boot?

-Kevin




More information about the coreboot mailing list