Re: Correct backtraces for do_IRQ

Tom,

osabi wasn't a problem. osabi was suggested by gdb community. They
didn't like the part that interpreted do_IRQ in a hackish way. It wasn't
quite maintainable and it incorrectly (read smartly) reported do_IRQ
frame information.

We are still working on all the patches I had sent last week. Let's
check them into cvs only when we know that gdb community is going to
accept the corresponding gdb changes.

I guess you'll submit these patches to lkml again when 2.5.15 comes out,
so we'll have to be careful with not breaking anything at this point of
time.
-Amit

> On Thu, Dec 29, 2005 at 05:45:51PM +0530, Amit Kale wrote:
> 
> > Got gdb also working (patch attached). Setting osabi to Linux-kernel
makes gdb 
> > understand stack switches as shown below. First backtrace is with
GNU/Linux 
> > osabi, which prints a corrupt stack error. Second one is after
setting osabi 
> > to Linux-kernel; it doesn't contain any errors.
> 
> I understand that previously the GDB community wasn't fond of the osabi
> linux-kernel change, has the community view changed here?  I'd really
> like to avoid having unmergable GDB changes required for best behavior.
> 
> -- 
> Tom Rini
> http://gate.crashing.org/~trini/
> 
> 

-- 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click