[rfc/rft] [0/5] Untangle register_addr et. al.

[rfc/rft] [0/5] Untangle register_addr et. al.

[Ulrich Weigand]

Hello,

the following series of 5 patches attempts to untangle everything related
to "traditional" ptrace-based register access.  Overall, the patch set:

 - Removes every definition of the global register_addr function.
 - Removes most NM file definitions of KERNEL_U_SIZE, U_REGS_OFFSET,
   KERNEL_U_ADDR, and REGISTER_U_ADDR (except on OSF and AIX).
 - Removes all NM file definitions of CANNOT_STORE_REGISTER and
   CANNOT_FETCH_REGISTER.
 - Completely removes core-aout.c
 - Adds a linux_trad_target () function to simplify use of traditional
   register access in a Linux target.
 - Gets rid of nearly all platform nm-linux.h files (except i386).

While not all portions of this are completely mechanical, the patch set
should not actually change the behaviour of any target (except core file
handling on alpha-linux, which should in fact improve ...).

However, as I don't have access to the majority of the affected native
targets, I'd appreciate any volunteers to test test patch set, in 
particular on mips-linux and alpha-linux.

I'd also appreciate any comments on the patch set: is this the right
think to do at this point; am I overlooking anything here?

Thanks,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com

Re: [rfc/rft] [0/5] Untangle register_addr et. al.

[Daniel Jacobowitz]

On Sat, Apr 14, 2007 at 11:07:27PM +0200, Ulrich Weigand wrote:
> However, as I don't have access to the majority of the affected native
> targets, I'd appreciate any volunteers to test test patch set, in 
> particular on mips-linux and alpha-linux.

I haven't looked through the patches yet, but after a bit of a
struggle I've tested them; the differences on mips-linux are in the
noise.  (mips-linux test results are not great at the moment, but not
awful either; the signal and threading tests are doing poorly.)


-- 
Daniel Jacobowitz
CodeSourcery

Re: [rfc/rft] [0/5] Untangle register_addr et. al.

[Ulrich Weigand]

Daniel Jacobowitz wrote:
> On Sat, Apr 14, 2007 at 11:07:27PM +0200, Ulrich Weigand wrote:
> > However, as I don't have access to the majority of the affected native
> > targets, I'd appreciate any volunteers to test test patch set, in 
> > particular on mips-linux and alpha-linux.
> 
> I haven't looked through the patches yet, but after a bit of a
> struggle I've tested them; the differences on mips-linux are in the
> noise.  (mips-linux test results are not great at the moment, but not
> awful either; the signal and threading tests are doing poorly.)

Thanks for testing!

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com