Re: PATCH: MIPS support for libjava (version 2).

[David Daney <ddaney at avtrex dot com>]

Andrew Haley wrote:

> David Daney writes:
>
> > 2003-09-10  David Daney <ddaney@avtrex.com>
> > 
> > 	MIPS support for libjava.
> > 	* libjava/Makefile.am: Support for extra .c and .S files.
> > 	* libjava/configure.host: Added mips*el*-linux* host.
> > 	* libjava/configure.in: Added --with-broken-dladdr,
> > 	--with-libgcj-mips-xgot and mips*-*-linux* host.
> > 	* libjava/include/mips-signal.h: New file.
> > 	* libjava/sysdep/mips/mipsel-backtracer.c: New file.
> > 	* libjava/sysdep/mips/mipsel-backtracer.h: New file.
> > 	* libjava/sysdep/mips/mipsel-bthelper.S: New file.
>
> Please post your libgcj test results.
Unfortunatly I currently have no way to run the testsuite as I have it 
configured as a cross compiler, and the testsuite seems to be intended 
for native builds.

> If you have a clean test run, this is OK with Anthony's suggested
> changes.  Your MIPS stacktrace looks very hairy, but I take your word
> that there really is no better way to do it.
I am using glibc-2.2.4 and its backtrace does not work at all.  Is there 
a working version in later glibcS?  I have no idea, and am stuck using 
2.2.4 so there is no way for me to find out.

There is a small problem with the backtracer I posted.  I will post a 
corrected version.

> Once you have checked in the MD_FALLBACK_FRAME_STATE_FOR changes,
> please check this in.
>  
I will post a new patch with Anthony's suggested changes and a new 
backtracer.  I think I will change the file names from mipsel-* to just 
mips-* because I don't think there are any endian dependencies in the code.

David.