This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: block merging
- To: law at cygnus dot com
- Subject: Re: block merging
- From: Jason Merrill <jason at cygnus dot com>
- Date: 20 Sep 1999 09:52:46 -0700
- Cc: gcc-patches at gcc dot gnu dot org
- References: <1580.937841825@upchuck.cygnus.com>
>>>>> Jeffrey A Law <law@cygnus.com> writes:
> In message <u9d7vdbnyd.fsf@yorick.cygnus.com>you write:
>> >>>>> Jeffrey A Law <law@cygnus.com> writes:
>>
>> > ! For now we avoid the EH issues by not allowing any physical
>> > ! block movement when exception handling is enabled. */
>> > ! if (flag_exceptions)
>> > ! return 0;
>>
>> Please don't do this. If you don't want to take the time to make various
>> optimizations work in the presence of EH, at least allow them to be done in
>> functions with no EH regions.
> I've wanted to do this, but I simply don't have the time and I'm
> unlikely to have the time in the future. The code has been basically
> sitting here for a year, and I don't want to see it continue to languish
> because I don't have the time to make it EH safe.
I'm not asking you to make it EH safe. I'm asking you to enable it in
functions that don't use EH even in the presence of flag_exceptions. That
should be trivial; if there are no EH_REGION_BEG notes in the function, you
don't have to worry about EH. It should be exactly like the -fno-exceptions
case.
> IMHO, this is one of the cases where putting the code out is the right
> thing to do. We derive some benefit from the optimization and hopefully
> someone with the interest and time will take up the task of extending it
> to handle more cases.
Agreed in principle, but I don't think I'm asking for much here.
Jason