This is really incredible. This is an Intel mainboard and there is something going on that locks out hardware interrupts for 2-5 SECONDS while running FreeBSD. I think the user made the right choice: buy a different mainboard.
Another poster reports up to 1 second delays for the BIOS USB stack (mainboard not specified in that case).
Finally, I had another report of a different Intel motherboard with an incorrect _MP_ table, which the vendor refused to fix. The fix was the same as the first poster: buy a different mainboard.
These are not old mainboards. Every time I start to wonder if coreboot is going to get overtaken by proprietary BIOSes I hear a story of this kind. At that point, I have to wonder why anyone would want a motherboard with a proprietary BIOS, given how bad they are, even today.
ron
---------- Forwarded message ---------- From: Lyndon Nerenberg (VE6BBM/VE7TFX) lyndon@orthanc.ca Date: Wed, Dec 30, 2009 at 7:17 PM Subject: Re: [9fans] Broken Hardware List To: 9fans@9fans.net
i think it would be more valuable to explain exactly what's not working and point to some of the workarounds, if they exist.
What's not working is the ACPI component of the BIOS. The P9 boot fails very early on (right after E820 I think). FreeBSD runs, but something in the ACPI code wakes up every couple of seconds and leaps into the ACPI BIOS code. While it's in there it locks out all hardware interrupts for 2-5 seconds, which makes the box pretty much unusable for anything close to interactive work.
As to workarounds, I worked around to the computer store and bought a non-Intel motherboard that actually implements Intel's ACPI spec somewhat correctly. I'm not going to spend time I could be billing out to try to fix a completely braindead motherboard that I can replace for what I earn in an hour or two. Shitty hardware like this deserves to die, not get fixed.
--lyndon
-
ron minnich