https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69027
Bug ID: 69027
Summary: SPARC: Missing optimization for fall through functions
Product: gcc
Version: 6.0
Status: UNCONFIRMED
Severity: enhancement
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: [email protected]
Target Milestone: ---
Consider the following test case:
int i(void);
int f(void)
{
return i();
}
int g(int (*j)(void))
{
return (*j)();
}
GCC generates on SPARC something like this:
sparc-rtems4.12-gcc -O2 -S fallthrough.c && cat fallthrough.s
.file "fallthrough.c"
.section ".text"
.align 4
.global f
.type f, #function
.proc 04
f:
or %o7, %g0, %g1
call i, 0
or %g1, %g0, %o7
.size f, .-f
.align 4
.global g
.type g, #function
.proc 04
g:
save %sp, -96, %sp
call %i0, 0
nop
jmp %i7+8
restore %g0, %o0, %o0
.size g, .-g
.ident "GCC: (GNU) 6.0.0 20151221 (experimental)
For g() an superfluous stack frame is generated.
On PowerPC for example this is optimized to:
.file "fallthrough.c"
.machine ppc
.section ".text"
.align 2
.globl f
.type f, @function
f:
b i
.size f, .-f
.align 2
.globl g
.type g, @function
g:
mtctr 3
bctr
.size g, .-g
.ident "GCC: (GNU) 6.0.0 20151126 (experimental)
