This is the mail archive of the
mailing list for the GDB project.
Re: [rfa, v3] Fix software-watchpoint failures on ARM by adding epilogue detection
- From: "Ulrich Weigand" <uweigand at de dot ibm dot com>
- To: rearnsha at arm dot com (Richard Earnshaw)
- Cc: dan at codesourcery dot com (Daniel Jacobowitz), gdb-patches at sourceware dot org
- Date: Fri, 8 Oct 2010 15:26:53 +0200 (CEST)
- Subject: Re: [rfa, v3] Fix software-watchpoint failures on ARM by adding epilogue detection
Richard Earnshaw wrote:
> On Thu, 2010-10-07 at 18:12 +0200, Ulrich Weigand wrote:
> > 2010-10-07 Ulrich Weigand <email@example.com>
> > Daniel Jacobowitz <firstname.lastname@example.org>
> > * arm-tdep.c (thumb_in_function_epilogue_p)
> > (arm_in_function_epilogue_p): New.
> > (arm_gdbarch_init): Install arm_in_function_epilogue_p as
> > gdbarch_in_function_epilogue_p callback.
> This is OK. There are some sequences that could be generated that this
> can't handle, and I don't think we'd ever get a stack decrement in an
> epilogue sequence.
> An example I don't think your code can handle is stack adjustments that
> are not valid immediates for ADD SP, #n. So very large stack adjusts my
> fail. These sequences will load or create an immediate in a register
> (particularly on Thumb1) and then add that to SP.
Thanks for the review; I've checked the patch in now. We can always
go back and add more instructions later ...
Dr. Ulrich Weigand
GNU Toolchain for Linux on System z and Cell BE