math_efp.c: Fixed SPE data type conversion failure
The following test case failed on Powerpc sbc8548 with CONFIG_SPE
static float fm;
static signed int si_min = (-2147483647 - 1);
static unsigned int ui;
int main()
{
fm = (float) si_min; ;
ui = (unsigned int)fm;
printf("ui=%d, should be %d\n", ui, si_min);
return 0;
}Result: ui=-1, should be -2147483648
The reason is failure to emulate the minus float to unsigned integer conversion instruction in the SPE driver.
PATCH 1/1 Fix SPE float to integer conversion failure
Conversion from float to integer should based on both the instruction encoding and the sign of the operand.
Signed-off-by Shan Hai <[email protected]>
--- arch/powerpc/math-emu/math_efp.c | 12 ++++++++---- 1 files changed, 8 insertions(+), 4 deletions(-)
arch/powerpc/math-emu/math_efp.c +8 -4
@@ -320,7 +320,8 @@ int do_spe_mathemu(struct pt_regs *regs) }else {_FP_ROUND_ZERO (1 ,SB ); } -FP_TO_INT_S (vc.wp[1 ],SB ,32 , ((func &0x3 ) !=0 )); +FP_TO_INT_S (vc.wp[1 ],SB ,32 , + (((func &0x3 ) !=0 ) || SB_s)); goto update_regs;default :
@@ -458,7 +459,8 @@ cmp_s: }else {_FP_ROUND_ZERO (2 ,DB ); } -FP_TO_INT_D (vc.wp[1 ],DB ,32 , ((func &0x3 ) !=0 )); +FP_TO_INT_D (vc.wp[1 ],DB ,32 , + (((func &0x3 ) !=0 ) || DB_s)); goto update_regs;default :
@@ -589,8 +591,10 @@ cmp_d:_FP_ROUND_ZERO (1 ,SB0 );_FP_ROUND_ZERO (1 ,SB1 ); } -FP_TO_INT_S (vc.wp[0 ],SB0 ,32 , ((func &0x3 ) !=0 )); -FP_TO_INT_S (vc.wp[1 ],SB1 ,32 , ((func &0x3 ) !=0 )); +FP_TO_INT_S (vc.wp[0 ],SB0 ,32 , + (((func &0x3 ) !=0 ) || SB0_s)); +FP_TO_INT_S (vc.wp[1 ],SB1 ,32 , + (((func &0x3 ) !=0 ) || SB1_s)); goto update_regs;default :
Signature · 1 line
1.7.0.2
Kumar Gala
AppliedRe: [PATCH] Fix SPE float to integer conversion failure
Quoted 7 lines · Shan Hai
Conversion from float to integer should based on both the instruction encoding and the sign of the operand.
Signed-off-by Shan Hai <[email protected]>---
arch/powerpc/math-emu/math_efp.c| 12 ++++++++----1 files changed, 8 insertions(+), 4 deletions(-)
applied
- k