This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH, rtl-optimization]: Fix PR rtl-optimization/33638
> Eric, i am not arguing that my code in dse should stay the way that it
> is, i was just hoping somehow i could salvage the ability to do better
> things for accesses off of the frame pointer.
No disagreement. :-) I'm still trying to figure out whether flow.c would
handle this correctly with the help of CALL_INSN_FUNCTION_USAGE.
Note that dse.c also uses the alias.c machinery (canon_rtx), which is global.
I'm under the impression that by enhancing a little this machinery, we could
match the ability of gcse.c to detect that the stack pointer is invariant in
the function; if we had this, canon_rtx would return the right thing.
> Another issue that this brings up is the code that deletes stores into
> the frame pointer that die when the function returns. I am now a little
> worried that code may be also compromised. If i am able to see the
> store as being relative to the frame but there is some load whose
> address has been stored in a register, i am screwed.
This should be more or less an equivalent problem.