egcs-2.92.02 19980905 (m68k-next-nextstep3): Address of hoisted load clobbered

Mark Mitchell
Wed Sep 9 20:50:00 GMT 1998

>>>>> "Richard" == Richard Henderson <> writes:

    Richard> On Wed, Sep 09, 1998 at 08:30:36AM -0700, Mark Mitchell
    Richard> wrote:
    >> That's right; reload does stomp all over MEMs, and they should
    >> not be shared.  I forgot about this.  The bit to hoist the
    >> loads has the same problem, actually.

    Richard> I don't think so.  The mem you create was installed only
    Richard> two places -- at the beginning of the loop for the load,
    Richard> and the end of the loop for the store.  So the one copy
    Richard> is all we should need.

But, when I reread the code, it didn't look to me like I "created" the
MEM at all!  See insert_loop_mem; it just takes a the MEM it's been
handed, and stuffs it into the loop_mems array.  Or are you reading my
code better than me?  (Which is entirely possible... :-( ) 

    Richard> r~

Mark Mitchell
Mark Mitchell Consulting

More information about the Gcc mailing list