The following reply was made to PR kern/22325; it has been noted by GNATS.

From: Chapman Flack <nblists@anastigmatix.net>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/22325 ehci lockups
Date: Sun, 18 Jun 2006 14:26:16 -0400

 I have finally had the chance to try the latest ehci code on my
 Acer Labs M5239 rev 0x01 PCI ehci/ohci card. The news is still bad.
 This seems to be the most appropriate existing PR, as the BBB reset
 and stall clear messages are among those that can appear. I think
 the trouble may be more pervasive than this PR originally suggested
 though.
 
 Transfers will hang with both USB 2.0 devices I have at hand to test:
 a Lexar JDA128-00-501 flash, and an Epson GT-9300 (a/k/a 2400 photo)
 scanner. Both work flawlessly, over the same ehci/ohci card, if ehci
 is disabled in userconf. So I am pretty sure ehci is the problem.
 
 Hangs are repeatable (that is, I can get a hang every time I try to
 do something, not that I can guarantee all hangs are identical in
 details).  See http://www.anastigmatix.net/stuff/ehci/ for gzipped
 dmesgs and klogs of a DIAGNOSTIC+EHCIDEBUG+ehcidebug=20 kernel
 hanging.
 
 The logs labelled "with-dropped-intr-fix" were generated after I
 edited ehci_pci.c to apply the DROPPED_INTR_WORKAROUND unconditionally,
 just to see if it would help with this chipset as well as the one it
 was testing for. It does seem to /change/ the problem, but not to resolve
 it.
 
 The command for the umass test was 'dd if=/dev/sd0a >/dev/null' and for
 the uscanner test was 'scanimage --mode Color --depth 8 --resolution 300
 >/dev/null'.
 
 I cannot boast very much ehci clue, so I am not sure what more to try.
 
 -Chap