From: John Willis (john.willis_at_charter.net)
Date: Thu Nov 02 2006 - 08:03:42 CET
Message-ID: <4549984E.4020200@charter.net> Date: Thu, 02 Nov 2006 01:03:42 -0600 From: John Willis <john.willis@charter.net> Subject: [suse-sles-e] SLES 9.1 - usb-handoff
Has anyone experience (outside of IBM HS20 blade servers) where
usb-handoff helped avoid kernel panics?
I'd be really interested to know.
I read some kernel notes a while back that indicated a 50 msec window
during which a BIOS legacy supported USB keyboard or other pre-boot BIOS
supported USB device might generate an SMI interrupt and corrupt a
register while loading the kernel.
It would seem unlikely that many platforms would share the same BIOS
code and this might be very rare.
But the same discussions seemed to indicate the feature was going into
the kernel proper.. perhaps to make it possible to shutdown or
transition legacy USB support to kernel mode support quicker than is
possible than using say a kernel module?
The other old adage seems to be disable "legacy USB support" in your
BIOS, or disable APIC, both undesirable for many reasons.
If anyone is interested, I thought the rational went something like,
BIOS legacy USB support is implemented in a "way" such that a USB
keyboard or other supported USB device was generating an interrupt which
got translated into an i8042 interrupt the older BIOS code could handle
as if it were not a USB device, but a PS/2 keyboard. At least that was
the gist of what I read.. I could have it all wrong.
Thanks very much for your comments on this matter.
- John
---------------------------------------------------------------------
To unsubscribe, e-mail: suse-sles-e-unsubscribe@suse.com
For additional commands, e-mail: suse-sles-e-help@suse.com
This archive was generated by hypermail 2.1.7 : Thu Nov 02 2006 - 08:06:32 CET