tree-into-ssa patch causes regression on ppclinux

Jeffrey A Law law@redhat.com
Wed Jun 9 02:24:00 GMT 2004


On Mon, 2004-05-31 at 14:04, Andreas Tobler wrote:
> Hello,
> 
> 2004-05-19  Jeff Law <law@redhat.com>
> 
>          * tree-into-ssa.c (prepare_operand_for_rename): New argument
> 	is_use.
>          If the operand is for a use, then strip away the SSA_NAME, do
>   	not strip away the SSA_NAME for a set.  Never call
> 	release_ssa_name.
>          (mark_def_sites): Appropriately pass additional argument to
>          prepare_operand_for_rename.  If a VDEF_RESULT is not an
>   	SSA_NAME, then set the VDEF_RESULT to the VDEF_OP.
>          (set_def_block): Strip away any SSA_NAME to get to the real
>          underlying variable.
> 
> 
> this patch causes a regression in the libjava test suite on ppclinux. 
> The famous Throw_2 test case times out.
> 
> I reverted the whole tree just before and after this patch. Before ok, 
> afterwards it hangs. I know, not much help, but I'm stuck atm.
> Does anyone have an idea? What do you need to give me a help?
I can't even get a stage1 build going -- presumably my aging YDL
system has some out-of-date bits which is giving gengtype and friends
indigestion.

I'm downloading a more recent version of YDL and hope to have my
box upgrade tomorrow.  Then I can return to trying to nail down this
problem.

jeff



More information about the Gcc-patches mailing list