This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [4.1] Fix PR rtl-optimization/29329


Eric Botcazou <ebotcazou@libertysurf.fr> writes:
>> I'm probably misunderstanding something here, but doesn't that mean you
>> end up with a REG_DEAD note for register 107 attached to an insn that
>> doesn't actually use register 107?
>
> No, distribute_notes scans the insn chain and punts.

Duh, right, sorry.  I was definitely misunderstanding.  I saw the
"tem = i3" in the patch and somehow got the impression you were
attaching the note to i3.  That would have been "place = i3",
of course.

I still need a bit of hand-holding though.  Why is this case any
different from a case in which I3 _did_ use the original rhs of I2,
the original rhs of I2 mentions ELIM_I2, but the simplified form of
I3 doesn't?  The comments give an example of when this might happen.
Wouldn't we need to propagate the death note then too?

Richard


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]