lkml.org 
[lkml]   [2008]   [Jan]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subject[patch 0/8] Series to improve the x86 backtracing code
Hi,

this patch series improves the x86 backtracing code in the following ways:
1) Fix a bad bug in x86 (32 bit) FRAME_POINTER backtracing (2.6.24 material)
2) Add the capability to mark a backtrace entry as reliable / unreliable
3) Change the x86 (32 bit) FRAME_POINTER backtracing to use the normal backtrace but use the frames
as guidance to the reliability of the backtrace entries
4) Patch to capture the EBP register earlier in the backtrace callchain to get a better FRAME_POINTER
based backtrace
5) Split the x86 (64 bit) backtrace code up similar to the 32 bit code by turning a macro into a function
6) Enable FRAME_POINTER backtracing on 64 bit similar to the 32 bit patch in number 3
7) Add a simple self-test module for backtraces (will move to tests/ later when that gets merged)
8) Add the feature to x86 64bit to print a set of bytes prior to the current EIP in the Code: line
(already present in 32 bit)

I have not yet implemented 2 requested items
1) To give dump_stack() and co a parameter towards the KERN_ log level; this is a patch that touches 125 files
and while I like the idea I'm not sure it's worth THAT.
2) Linus' request to only print the first 5 entries of a backtrace with a important loglevel, and do the rest
as KERN_WARNING or so. This realistically wants 1) and I've not gotten to it yet (and it's independent functionality
anyway)


Andrew: You already have patch 1 in your tree, and it's not modified since that, so no need to refresh

--
If you want to reach me at my work email, use arjan@linux.intel.com
For development, discussion and tips for power savings,
visit http://www.lesswatts.org


\
 
 \ /
  Last update: 2008-01-12 06:29    [W:0.043 / U:0.488 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site