kern/123140: SMP boot causes slow KB, ATA drives not detected

[Gavin Atkinson]

The following reply was made to PR kern/123140; it has been noted by GNATS.

From: Gavin Atkinson <gavin at FreeBSD.org>
To: bug-followup at FreeBSD.org
Cc:  
Subject: Re: kern/123140: SMP boot causes slow KB, ATA drives not detected
Date: Thu, 17 Jul 2008 15:47:22 +0100

 -------- Forwarded Message --------
 From: Bob Frazier <bobf at mrp3.com>
 Subject: Re: kern/123140: SMP boot causes slow KB, ATA drives not
 detected
 Date: Wed, 16 Jul 2008 19:22:08 -0700
 
 > Sorry for the lack of response, somehow procmail seems to have filtered
 > a handful of my emails into a folder I never check, so I missed all of
 > your responses.  I guess you're still seeing these issues?  If so, I'll
 > have a close look at the information you've provided, and probably get
 > somebody else involved who knows more about the low level
 > interrupt-related stuff.
 
 thanks - FYI I'm down to the point where I need to look at wiring 
 diagrams and chipset docs if I were to continue doing things myself. 
 Last month or so I've had no time so no progress either.  I sent a 
 support request to Asus for grins, got the 'door slam' (not even a 
 'reject' e-mail) when they simply closed the request.  At this point I 
 think the ACPI configuration for the APIC is wrong, maybe even int 2. 
 Without a wiring diagram there's no way to tell.  Hopefully someone has 
 contacts at ASUS for that info.  It could also be a unique protocol 
 between APIC and LAPIC, requiring additional setup and/or additional 
 steps during ISRs.  Without the docs that's not possible to determine 
 either.
 
 In any case I read the docs on setting up a "substitute ACPI BIOS" to 
 include with the kernel during startup and may go that route first with 
 my own attempts at testing.
 
 The LAPIC timer happening 4 times more frequently than it should 
 (meaning all 4 LAPICs are generating timer interrupts but the CPUs still 
 'round robin' the handling?) seems a little unusual to me.  Do you think 
 this is at the core of the problem?
 
 > What might be of use is if you could attempt to boot one of the recent
 > 8.0 snapshot CDs (from
 > ftp://ftp.freebsd.org/pub/FreeBSD/snapshots/200806/ ) and see if you
 > still see the issue.  Also knowing if a 6.3 bootonly CD shows the same
 > issue would probably be useful.
 
 I'll check it out.  It can't hurt to download and test-boot a CD as long 
 as it doean't destroy my file system.  If 8.0 works I'll see if I can't 
 port it over to 7 myself for testing, since there would be sample code 
 of an "it works" solution.