Date: Sat, 05 Oct 1996 13:30:58 +0800 From: Peter Wemm <peter@spinner.dialix.com> To: smp@freebsd.org Subject: gdb -k patches for smp kernels Message-ID: <199610050530.NAA00477@spinner.DIALix.COM>
next in thread | raw e-mail | index | archive | help
In order to run "gdb -k /kernel /dev/mem" on a smp kernel, you need the
following [crude, ugly, bogus, etc] patches:
Index: freebsd-nat.c
===================================================================
RCS file: /home/ncvs/src/gnu/usr.bin/gdb/gdb/freebsd-nat.c,v
retrieving revision 1.10
diff -u -r1.10 freebsd-nat.c
--- freebsd-nat.c 1996/06/08 11:03:19 1.10
+++ freebsd-nat.c 1996/07/18 19:34:07
@@ -29,6 +29,8 @@
#include <sys/ptrace.h>
#include "defs.h"
+#include "symtab.h"
+#include "inferior.h"
/* this table must line up with REGISTER_NAMES in tm-i386v.h */
/* symbols like 'tEAX' come from <machine/reg.h> */
@@ -351,6 +383,7 @@
struct stat stb;
CORE_ADDR addr;
int cfd;
+ struct minimal_symbol *sym;
if ((cfd = open(cfile, perm, 0)) < 0)
return (cfd);
@@ -363,7 +396,12 @@
physrd(cfd, ksym_lookup("IdlePTD") - KERNOFF, (char*)&sbr, sizeof sbr);
printf("IdlePTD %x\n", sbr);
- curpcb = ksym_lookup("curpcb") - KERNOFF;
+
+ sym = lookup_minimal_symbol("SMPcurpcb", (struct objfile *)NULL);
+ if (sym != NULL)
+ curpcb = ksym_lookup("SMPcurpcb") - KERNOFF;
+ else
+ curpcb = ksym_lookup("curpcb") - KERNOFF;
physrd(cfd, curpcb, (char*)&curpcb, sizeof curpcb);
kstack = ksym_lookup("kstack");
Index: kcorelow.c
===================================================================
RCS file: /home/ncvs/src/gnu/usr.bin/gdb/gdb/kcorelow.c,v
retrieving revision 1.4
diff -u -r1.4 kcorelow.c
--- kcorelow.c 1995/05/30 04:57:22 1.4
+++ kcorelow.c 1996/07/18 08:44:48
@@ -86,7 +86,14 @@
curProc()
{
struct proc *p;
- CORE_ADDR addr = ksym_lookup("curproc");
+ CORE_ADDR addr;
+ struct minimal_symbol *sym;
+
+ sym = lookup_minimal_symbol("SMPcurproc", (struct objfile *)NULL);
+ if (sym != NULL)
+ addr = ksym_lookup("SMPcurproc");
+ else
+ addr = ksym_lookup("curproc");
if (kvread(addr, &p))
error("cannot read proc pointer at %x\n", addr);
I've been using these hacks for a while, even though they're not exactly
correct, but they work for simple stuff (eg: read/modify data). They
adapt to the style of kernel, so these patches will work on both kernels
if you are flipping back and forward from smp kernels like I am.
-Peter
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610050530.NAA00477>