Re: [PATCH] PPC - Stepping off breakpoints in non-stop mode

[Jonathan Larmour <jifl@eCosCentric.com>]

Luis Machado wrote:
> [...]
> Checked in the following.
> [...]
> 2008-06-30  Luis Machado  <luisgpm@br.ibm.com>
> 
> 	* rs6000-tdep.c (ppc_displaced_step_fixup): New function.
> 	(deal_with_atomic_sequence): Update BC masks.
> 	(rs6000_gdbarch_init): Init displaced stepping infra-structure.
> 	Define BRANCH_MASK, B_INSN, BC_INSN, BXL_INSN, BP_MASK and BP_INSN.

Unfortunately this fails to build due to this warning (in combination with
-Werror):
.../rs6000-tdep.c: In function ‘ppc_displaced_step_fixup’:
.../rs6000-tdep.c:885: warning: passing argument 3 of
‘regcache_cooked_read_unsigned’ from incompatible pointer type
make[2]: *** [rs6000-tdep.o] Error 1

This is because in the following line, current_pc is CORE_ADDR, but
regcache_cooked_read_unsigned's last arg is meant to be a ULONGEST *.

+      regcache_cooked_read_unsigned (regs, gdbarch_pc_regnum (gdbarch),
+				      &current_pc);

I was thinking of just casting it, but in fact would it be better to be
calling regcache_read_pc() here as well as regcache_write_pc() in the rest
of the patch when operating on the pc? From earlier in this thread, there
was a little discussion about address sizes, so I thought I'd best ask
before submitting a patch, as there could be portability implications and
it might have been done this way for a reason.

Jifl
-- 
eCosCentric Limited      http://www.eCosCentric.com/     The eCos experts
Barnwell House, Barnwell Drive, Cambridge, UK.       Tel: +44 1223 245571
Registered in England and Wales: Reg No 4422071.
------["Si fractum non sit, noli id reficere"]------       Opinions==mine