Sylvestre Ledru wrote:
> Good question... I can try to do that. I was just trying to match the
> current upstream behavior first.
>
Ok did this :
************
int foo(i)
{
return i*3;
}
************
Compiled this with :
gcc -O3 foo.c -S -o foo.s
which yields :
************
.file "foo.c"
.text
.align 4
.globl foo
.type foo, @function
foo:
.LFB0:
.cfi_startproc
lr %r1,%r2
sll %r1,1
ar %r2,%r1
br %r14
.cfi_endproc
.LFE0:
.size foo, .-foo
.ident "GCC: (Debian 4.4.2-5) 4.4.2"
.section .note.GNU-stack,"",@progbits
************
(apparently, GCC optimizer says that 3x=2x+x is the best way..)
That being said, and with the s390 ABI giving parm 1 in R2, returning
R2, R1 being expandable and R14 being the return address there are
several possibilities...
lr %r1,%r2
ar %r2,%r1
ar %r2,%r1
br %r14
would probably also work (3x=x+x+x)
The 64 bit version (s390x) - with gcc -m64 adds a 32 bit truncation..
lr %r1,%r2
sll %r1,1
ar %r1,%r2
lgfr %r2,%r1
br %r14
You will probably also need to tweak your atlas_asm.h - and add a case
for ATL_GAS_S390 - I don't know what it does, but if it is some function
decoration, it *may* be ABI dependent (not sure.. but checking wouldn't
hurt).
--Ivan
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature