sparc32: Kill off software 32-bit multiply/divide routines.
For the explicit calls to .udiv/.umul in assembler, I made a
mechanical (read as: safe) transformation. I didn't attempt
to make any simplifications.
In particular, __ndelay and __udelay can be simplified significantly.
Some of the %y reads are unnecessary and these routines have no need
any longer for allocating a register window, they can be leaf
functions.
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/arch/sparc/kernel/entry.S b/arch/sparc/kernel/entry.S
index 773f3f0..3f3976e 100644
--- a/arch/sparc/kernel/entry.S
+++ b/arch/sparc/kernel/entry.S
@@ -1161,11 +1161,13 @@
.globl __ndelay
__ndelay:
save %sp, -STACKFRAME_SZ, %sp
- mov %i0, %o0
- call .umul ! round multiplier up so large ns ok
- mov 0x1ae, %o1 ! 2**32 / (1 000 000 000 / HZ)
- call .umul
- mov %i1, %o1 ! udelay_val
+ mov %i0, %o0 ! round multiplier up so large ns ok
+ mov 0x1ae, %o1 ! 2**32 / (1 000 000 000 / HZ)
+ umul %o0, %o1, %o0
+ rd %y, %o1
+ mov %i1, %o1 ! udelay_val
+ umul %o0, %o1, %o0
+ rd %y, %o1
ba delay_continue
mov %o1, %o0 ! >>32 later for better resolution
@@ -1174,18 +1176,21 @@
save %sp, -STACKFRAME_SZ, %sp
mov %i0, %o0
sethi %hi(0x10c7), %o1 ! round multiplier up so large us ok
- call .umul
- or %o1, %lo(0x10c7), %o1 ! 2**32 / 1 000 000
- call .umul
- mov %i1, %o1 ! udelay_val
+ or %o1, %lo(0x10c7), %o1 ! 2**32 / 1 000 000
+ umul %o0, %o1, %o0
+ rd %y, %o1
+ mov %i1, %o1 ! udelay_val
+ umul %o0, %o1, %o0
+ rd %y, %o1
sethi %hi(0x028f4b62), %l0 ! Add in rounding constant * 2**32,
or %g0, %lo(0x028f4b62), %l0
addcc %o0, %l0, %o0 ! 2**32 * 0.009 999
bcs,a 3f
add %o1, 0x01, %o1
3:
- call .umul
- mov HZ, %o0 ! >>32 earlier for wider range
+ mov HZ, %o0 ! >>32 earlier for wider range
+ umul %o0, %o1, %o0
+ rd %y, %o1
delay_continue:
cmp %o0, 0x0