Date: Sun, 01 Dec 1996 14:33:40 -0700 From: Steve Passe <smp@csn.net> To: Terje Normann Marthinussen <Terje.N.Marthinussen@cc.uit.no> Cc: freebsd-smp@freebsd.org Subject: Re: SMP-current status Message-ID: <199612012133.OAA15619@clem.systemsix.com> In-Reply-To: Your message of "Sun, 01 Dec 1996 21:36:41 %2B0100." <199612012036.VAA10981@slibo.cc.uit.no>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi,
>This might be of a bit more help...
>
>I let it stay there for a couple of minutes the first time, but it seems
>like that wasn't enough, came to reboot again a bit later. Thought I
>had another kernel in /, went out of the room before it came up, and when
>I came back a lot later it said my kermit window (serial console is nice :)):
>
>changing root device to sd0a
>Enabled INTs: 1, 2, 4, 6, 7, 8, 10, 11, 13, imen: 0x00ffd229
>sd0(ahc0:0:0): timed out while idle, LASTPHASE == 0x1, SCSISIGI == 0x0
smells like a PCI/EISA/level translation problem,
I really am growing to *hate* the MP spec!!!
-
from the output:
>ahc0 <Adaptec aic7870 SCSI host adapter> rev 3 int a irq 7 on pci1:13
^^^^^
-
from your mptable output:
I/O Ints: Type Polarity Trigger Bus ID IRQ APIC ID INT#
ExtINT active-hi edge 19 0 14 0
INT active-hi edge 19 1 14 1
...
INT conforms level 19 7 14 7
^^^^^^^^ ^^^^^ ^^ ^
---
without getting into the issues here (different boards do different things
and its all a very ugly $#%^&*^*$% mess)
looking at your previous mail I'm confused:
>FreeBSD 3.0-SMP #0: Sun Dec 1 18:25:33 MET 1996
> terjem@quattro:/usr/src/sys/compile/netserver
>FreeBSD/SMP: Multiprocessor motherboard
> cpu0 (BSP): apic id: 0, version: 0x00030010
> cpu1 (AP): apic id: 2, version: 0x00030010
> cpu2 (AP): apic id: 3, version: 0x00030010
> cpu3 (AP): apic id: 4, version: 0x00030010
> io0 (APIC): apic id: 14, version: 0x000f0011
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
the above line should ONLY come out of an APIC_IO enabled kernel, yet you
inply this one isn't when you go on to say:
>It then continues to boot, puts #2 into the scheduler and comes up just fine.
>I haven't any other OS on that machine right now, but I moved the disk
>to another identical machine (btw. it had stepping 11 on CPU 3 and 4)
>and got the same results, so I don't think it's the hardware thats
>faulty either.
>
>Tried to compile it with options APIC_IO for the first time as well:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
---
what is important to establish here is:
can you run a 2 CPU system with APIC_IO OFF ("options APIC_IO" NOT in config)?
my guess is you can...
can you run a 2 CPU system with APIC_IO ON ("options APIC_IO" IS in config)?
my guess is you *can't*...
--
the safest way to test 2 CPU version for now is to add line to mp_machdep.c:
/* start each AP */
+ mp_naps = 2;
for ( x = mp_naps; x >= 1; --x ) {
--
Steve Passe | powered by
smp@csn.net | FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199612012133.OAA15619>