atom feed8 messages in org.freebsd.freebsd-amd645.2-beta panics after copyright
FromSent OnAttachments
Adriaan de GrootNov 30, 2003 3:15 pm 
Adriaan de GrootNov 30, 2003 3:59 pm 
David O'BrienNov 30, 2003 8:54 pm 
Adriaan de GrootDec 1, 2003 12:47 am 
David O'BrienDec 1, 2003 2:22 am 
Adriaan de GrootDec 1, 2003 10:28 am 
Adriaan de GrootDec 1, 2003 2:54 pm 
David O'BrienDec 1, 2003 7:58 pm 
Subject:5.2-beta panics after copyright
From:David O'Brien (obr@freebsd.org)
Date:Dec 1, 2003 7:58:31 pm
List:org.freebsd.freebsd-amd64

On Mon, Dec 01, 2003 at 11:53:37PM +0100, Adriaan de Groot wrote:

On Yaum al-Ithnain 06 Shawwal 1424 19:28, Adriaan de Groot wrote:

Let's see what an even newer cvsup yields (dec 1 18:30 UTC) ..

OK, with sources from this evening I have:

1) with "option SMP" it panics right after the copyright message. I compiled with the kernel debugger and all the debugging stuff this time, it tells me: mtx_lock of spin mux (null) at uma_core.c:1388 (paraphrased as usual, the numbers are exact).

Verified on my Solo9. This patch fixes it:

Index: amd64/amd64/mp_machdep.c =================================================================== RCS file: /home/ncvs/src/sys/amd64/amd64/mp_machdep.c,v retrieving revision 1.229 diff -u -r1.229 mp_machdep.c --- amd64/amd64/mp_machdep.c 30 Nov 2003 22:20:40 -0000 1.229 +++ amd64/amd64/mp_machdep.c 2 Dec 2003 01:53:08 -0000 @@ -199,7 +199,7 @@ KASSERT(mp_maxid == 0, ("%s: mp_ncpus is zero, but mp_maxid is not", __func__)); else if (mp_ncpus == 1) - mp_maxid = 0; + mp_maxid = 1; else KASSERT(mp_maxid >= mp_ncpus - 1, ("%s: counters out of sync: max %d, count %d", __func__, @@ -232,7 +232,7 @@ * One CPU was found, so this must be a UP system with * an I/O APIC. */ - mp_maxid = 0; + mp_maxid = 1; return (0); }