[PING] Re: [Patch] Fix the allocation of jump register for RTL sequence abstraction

Gabor Loki loki@gcc.gnu.org
Fri Jun 9 06:52:00 GMT 2006


Bernd Schmidt wrote:
> Gabor Loki wrote:
> 
>> I would like to fix the following critical bug for 4.2:
>>
>> http://gcc.gnu.org/ml/gcc-patches/2006-05/msg00890.html
>> (idle for two week)
> 
> Please resubmit without the formatting changes, without the new 
> parameter (which seems unrelated to the critical bugfix), and with the 
> patch regenerated with -p added to the diff arguments.

I've reduced the patch. Now, it contains only the bugfix.
(Bootstrapped and regtested on arm-elf, i686-linux, ppc-elf, sh-elf)

Is there any comment or suggestion about the fix?

I would like to commit it before 4.2 release.


br,
      Gabor Loki

2006-06-09  Gabor Loki <loki@gcc.gnu.org>

       * gcc/rtl-factoring.c (gate_rtl_seqabstr): Only execute the optimization
	when optimize_size is turned on.
         (compute_hash): Use a parameter to control the number of hashed insn.
         (collect_pattern_seq, recompute_gain_for_pattern_seq, split_pattern_seq,
         erase_matching_seq, abstract_best_seq, dump_pattern_seq,fill_hash_bucket,
         seq_make_jump_insn_raw, seq_make_insn_raw, get_indirect_jump_insn_no,
         get_insn_op_reg_class, get_jump_reg, compute_init_costs, rtl_seqabstr):
	Fix the allocation of jump register.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: rtl-fact.patch
Type: text/x-patch
Size: 15125 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20060609/37e1feb7/attachment.bin>


More information about the Gcc-patches mailing list