[PATCH RFC] DWARF2 CFI exploitation for Linux on S/390

[Michal Ludvig]

Gerhard Tonn wrote:
> 1) The sequence and number of DWARF2 registers are different from the gdb
> registers on s390 for some reason.
> 2) The DWARF2 CFA value is different from the frame_base value.
> 
> In order to adapt the code to the first item, I have introduced #defines
> for DWARF2 registers and a REGNUM_TO_DWARF2_REG macro and its
> implementation. See the attached patch for details.

It makes sense. This wasn't yet noticed because the numbering of most 
dwarf2 and x86-64 registres is intentionally the same.

> The following code fragment in dwarf2cfi.c around line 1316 seems to be
> intel specific, at least on s390 it doesn't make sense. Should I move it to
> a gdbarch function?
> 
> if (i == SP_REGNUM)
>      {
>         context->reg[i].how = REG_CTX_VALUE;
>         context->reg[i].loc.addr = cfa;
>       }

Not only Intel specific. Also AMD specific ;-)

> Finally I am interested in how signal frame and dummy frame handling is
> supposed to work with DWARF2 CFI support. Does anybody have done already
> work in this area?

That's the question I'm solving too. The first approach (for x86-64) is 
here: http://sources.redhat.com/ml/gdb-patches/2002-09/msg00384.html
Basically I don't set_gdbarch_*() directly to cfi_*() functions but 
instead to corresponding x86_64_*() functions, that eventually call 
cfi_*() themselves. For sighandler caller frames I'm afraid I'll have to 
manually fill appropriate structures in struct context (probably in 
those x86_64_*() functions).

Michal Ludvig
-- 
* SuSE CR, s.r.o     * mludvig@suse.cz
* (+420) 296.545.373 * http://www.suse.cz