Skip to contentSkip to conversation

Kexec-tools: Fix physcal address value in vmcore ELF segment header

PATCH Horms 2 participants 2 messages Sep 21, 2006

lore
Share & tools
lore

Horms

author
#0 lore raw
Quoted 8 lines · Zou Nan hai

I used fixed 64M as physical offset in vmcore ELF segment header. This gives trouble to crash utility to analyze vmcore on platform that load kernel to a different address.

This patch will fix it.

Thanks,

I have merged this into kexec-tools-testing

Signature · 3 lines
Horms
  H: http://www.vergenet.net/~horms/
  W: http://www.valinux.co.jp/en/
#1 lore raw
Quoted 13 lines · Zou Nan hai

I used fixed 64M as physical offset in vmcore ELF segment header. This gives trouble to crash utility to analyze vmcore on platform that load kernel to a different address.

This patch will fix it.

diff -Nraup kexec-tools-1.101/kexec/arch/ia64/crashdump-ia64.c kexec-tools-1.101-fix/kexec/arch/ia64/crashdump-ia64.c

kexec/arch/ia64/crashdump-ia64.c line 31

Quoted 5 lines
kexec/arch/ia64/crashdump-ia64.c +1 -1
@@ -28,7 +28,7 @@
 #include "crashdump-ia64.h"
 
 int memory_ranges = 0;
-#define LOAD_OFFSET 	(0xa000000000000000UL + 0x100000000UL - (1UL<<26))
+#define LOAD_OFFSET 	(0xa000000000000000UL + 0x100000000UL - kernel_code_start)

This will work as long as virtual address for which kernel is being compiled remains same. (I guess, 0xa000000000000000UL + 0x100000000UL). If you shift the kernel little bit in virtual address space, this assumption will be broken.

That's why I put a patch to read /proc/kcore and determine for which virtual address the kernel has been built and there is no hardcoding. (Thanks to Eric for the suggestion.)

-Vivek