[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gdb-patches
Subject:    [rfa] Cross corefile support for ARM
From:       rearnsha () arm ! com (Richard Earnshaw)
Date:       2006-06-23 13:36:00
Message-ID: 1151069710.29989.17.camel () pc960 ! cambridge ! arm ! com
[Download RAW message or body]

On Fri, 2006-06-23 at 13:42, Daniel Jacobowitz wrote:
> On Thu, Jun 08, 2006 at 11:24:32AM -0400, Daniel Jacobowitz wrote:
> > I've reimplemented ARM Linux corefile support to use the new interfaces, and
> > work for cross-debuggers.  I needed this last week to track down a problem
> > in an EABI binary and I didn't have an EABI native GDB available.
> > 
> > I've tested it manually in both cross and native configurations; it actually
> > fixes FPA (NWFPE) support for core files, which had gotten broken at some
> > point in the past when ARM stopped providing FP0_REGNUM.
> > 
> > This whole story is a bit of an embarrassment.  I've implemented this at
> > least twice in the past but never cleaned it up enough to submit it.  And
> > Girish Shilamkar submitted something similar in February, although I didn't
> > notice at the time that (A) he didn't remove the existing core file support
> > in arm-linux-nat.c, which would probably break native ARM gdbs, and (B) the
> > FP support was bogus and didn't match any core dump layout I'm famliar with.
> > And I never heard back about TimeSys's copyright assignment, so that's been
> > sitting in limbo.
> > 
> > Look OK?
> 
> Hi Richard,
> 
> Did you notice:
>   http://sourceware.org/ml/gdb-patches//2006-06/msg00071.html
> 
> (Needed it again today :-)

Sorry, yes this is OK.

I wanted to review this against the problem that there are at least 3
different FPA-flavoured interfaces in linux: NWFPE, NWFPE with extended
precision and genuine FPA11, and that each of these has a slightly
different format for the values returned via the ptrace interface (yes,
this is really a bug in ARM linux, but it ain't going to get fixed
now...).  However, I'm now happy your patch doesn't make things either
better or worse in this regard.

R.


[prev in list] [next in list] [prev in thread] [next in thread]