https://bugzilla.novell.com/show_bug.cgi?id=403346
User stefan.neis@kobil.com added comment
https://bugzilla.novell.com/show_bug.cgi?id=403346#c49
Stefan Neis changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEEDINFO |ASSIGNED
Info Provider|stefan.neis@kobil.com |
--- Comment #49 from Stefan Neis 2008-07-24 11:07:33 MDT ---
Okay, here's a patch against 2.6.25.9 which reverts part of aa16ca... and which
solves my problem. Trying to find an explanation for what's going on (and why
we are apparently the first to be bitten by this), I guess that trying to be
clever about "short-circuiting the scan" (as the comment in the code calls it)
is maybe only working for devices with a single interface, but causes confusion
in our case (storage-interface _and_ HID-interface) when both interfaces are
used at the same time. That's just a vague guess, as I don't understand the
code at all, but maybe it could be a hint for those that do understand the
code.
--- ehci-sched.c.orig 2008-07-21 18:03:29.000000000 +0200
+++ ehci-sched.c 2008-07-24 20:43:40.000000000 +0200
@@ -2228,14 +2228,8 @@
}
/* assume completion callbacks modify the queue */
- if (unlikely (modified)) {
- if (likely(ehci->periodic_sched > 0))
- goto restart;
- /* maybe we can short-circuit this scan! */
- disable_periodic(ehci);
- now_uframe = clock;
- break;
- }
+ if (unlikely (modified))
+ goto restart;
}
/* If we can tell we caught up to the hardware, stop now.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.