[Bug rtl-optimization/43616] Extra register move
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=43616 Andrew Pinski changed: What|Removed |Added Target Milestone|--- |5.0 Status|NEW |RESOLVED Resolution|--- |FIXED Known to work||5.1.0, 9.1.0 --- Comment #3 from Andrew Pinski --- The arm code generation was fixed in GCC 5 for the testcase in comment #0. The x86_64 code generation was fixed in GCC 9 for the testcase in comment #2. So fixed a long time ago for both.
[Bug rtl-optimization/43616] Extra register move
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43616 Uros Bizjak changed: What|Removed |Added Target|arm-eabi|arm-eabi, x86 Known to fail|| --- Comment #2 from Uros Bizjak 2012-06-17 11:03:47 UTC --- The same problem happens on x86_64: --cut here-- void foo (void); float test (float a, float b) { foo (); return b; } --cut here-- gcc -O2 test: .LFB0: subq$24, %rsp movss %xmm1, (%rsp) callfoo A movss (%rsp), %xmm1 addq$24, %rsp B movaps %xmm1, %xmm0 ret We could move from stack directly to xmm0
[Bug rtl-optimization/43616] Extra register move
-- rguenth at gcc dot gnu dot org changed: What|Removed |Added Severity|normal |enhancement Summary|[4.5, trunk] Extra register |Extra register move |move| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43616