: missed-optimization
Severity: normal
Priority: P3
Component: middle-end
Assignee: unassigned at gcc dot gnu.org
Reporter: pdj at knaldgas dot dk
Target Milestone: ---
Host: x86_64
Target: x86_64
Build: x86_64
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91131
--- Comment #12 from Per Dalgas Jakobsen ---
Created attachment 46591
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=46591&action=edit
Added more combinations of bitfield sets.
Now includes all 8 combinations of zero/non-zero assignments t
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91131
--- Comment #11 from Per Dalgas Jakobsen ---
(In reply to Richard Biener from comment #10)
> (In reply to Per Dalgas Jakobsen from comment #9)
> > avr-gcc -O3 (and -Os, -O1, -O2):
> > sts 0x0064, r1 ; 0x800064
> > ldi r24, 0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91131
--- Comment #9 from Per Dalgas Jakobsen ---
(In reply to Richard Biener from comment #8)
> Fixed on trunk sofar.
>
> Note the non-optimal code-gen probably was a side-effect of us making
> three volatile accesses out of one. On x86 I now see
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91131
--- Comment #6 from Per Dalgas Jakobsen ---
(In reply to Richard Biener from comment #3)
> note your use of packed might end up doing more than one store depending
> on the architecture.
If you mean that a packed structure beyond the data-width
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91131
--- Comment #5 from Per Dalgas Jakobsen ---
(In reply to Richard Biener from comment #2)
> So your complaint is that
>
> struct Reg_T {
> unsigned int a : 3;
> unsigned int b : 1;
> unsigned int c : 4;
> } __attribute__((
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91131
--- Comment #1 from Per Dalgas Jakobsen ---
Created attachment 46585
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=46585&action=edit
MWE: Preprocessed C-file
Priority: P3
Component: middle-end
Assignee: unassigned at gcc dot gnu.org
Reporter: pdj at knaldgas dot dk
Target Milestone: ---
In both C and Ada, bitfield coalescing does not work right, or at least not
optimal. Since the issue manifests itself in the same way in
