Fixed Powerpc SPE data type conversion failure

All of lore.kernel.org
 help / color / mirror / Atom feed

* Fixed Powerpc SPE data type conversion failure
@ 2010-11-14  4:11 Hai Shan
  2010-11-14  4:11 ` [PATCH] Fix float to unsigned conversion failure with SPE enabled Hai Shan
  0 siblings, 1 reply; 6+ messages in thread
From: Hai Shan @ 2010-11-14  4:11 UTC (permalink / raw)
  To: linux-kernel


The following test case failed on Powerpc e500 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 conversion instruction
in the SPE driver.

---
 arch/powerpc/math-emu/math_efp.c |   17 +++++++++++++----
 1 files changed, 13 insertions(+), 4 deletions(-)


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH] Fix float to unsigned conversion failure with SPE enabled
  2010-11-14  4:11 Fixed Powerpc SPE data type conversion failure Hai Shan
@ 2010-11-14  4:11 ` Hai Shan
  2010-11-16 11:29     ` Josh Boyer
  0 siblings, 1 reply; 6+ messages in thread
From: Hai Shan @ 2010-11-14  4:11 UTC (permalink / raw)
  To: linux-kernel; +Cc: Hai Shan

Fixed the failure on converting minus float to unsigned int with SPE enabled

Signed-off-by: Hai Shan <shan.hai@windriver.com>
---
 arch/powerpc/math-emu/math_efp.c |   17 +++++++++++++----
 1 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/arch/powerpc/math-emu/math_efp.c b/arch/powerpc/math-emu/math_efp.c
index 41f4ef3..338a128 100644
--- a/arch/powerpc/math-emu/math_efp.c
+++ b/arch/powerpc/math-emu/math_efp.c
@@ -320,7 +320,9 @@ 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));
+			/* SB_s: convert from minus float to unsigned int */
+			FP_TO_INT_S(vc.wp[1], SB, 32, 
+					((func & 0x3) != 0) || SB_s);
 			goto update_regs;
 
 		default:
@@ -458,7 +460,11 @@ cmp_s:
 			} else {
 				_FP_ROUND_ZERO(2, DB);
 			}
-			FP_TO_INT_D(vc.wp[1], DB, 32, ((func & 0x3) != 0));
+			/* DB_s: convert from minus long double to 
+			 * unsigned long long
+			 */
+			FP_TO_INT_D(vc.wp[1], DB, 32, 
+					((func & 0x3) != 0) || DB_s);
 			goto update_regs;
 
 		default:
@@ -589,8 +595,11 @@ 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));
+			/* SB*_s: convert from minus float to unsigned int */
+			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:
-- 
1.7.0.4


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix float to unsigned conversion failure with SPE enabled
  2010-11-14  4:11 ` [PATCH] Fix float to unsigned conversion failure with SPE enabled Hai Shan
@ 2010-11-16 11:29     ` Josh Boyer
  0 siblings, 0 replies; 6+ messages in thread
From: Josh Boyer @ 2010-11-16 11:29 UTC (permalink / raw)
  To: Hai Shan; +Cc: Hai Shan, linuxppc-dev, linux-kernel

On Sat, Nov 13, 2010 at 11:11 PM, Hai Shan <haishan.bai@gmail.com> wrote:
> Fixed the failure on converting minus float to unsigned int with SPE enab=
led
>
> Signed-off-by: Hai Shan <shan.hai@windriver.com>

You should make sure to send PowerPC patches to linuxppc-dev.

josh

> ---
> =A0arch/powerpc/math-emu/math_efp.c | =A0 17 +++++++++++++----
> =A01 files changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/arch/powerpc/math-emu/math_efp.c b/arch/powerpc/math-emu/mat=
h_efp.c
> index 41f4ef3..338a128 100644
> --- a/arch/powerpc/math-emu/math_efp.c
> +++ b/arch/powerpc/math-emu/math_efp.c
> @@ -320,7 +320,9 @@ int do_spe_mathemu(struct pt_regs *regs)
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0} else {
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0_FP_ROUND_=
ZERO(1, SB);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0}
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_S(vc.wp[1], SB, 3=
2, ((func & 0x3) !=3D 0));
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* SB_s: convert from minus=
 float to unsigned int */
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_S(vc.wp[1], SB, 3=
2,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 ((func & 0x3) !=3D 0) || SB_s);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto update_regs;
>
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0default:
> @@ -458,7 +460,11 @@ cmp_s:
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0} else {
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0_FP_ROUND_=
ZERO(2, DB);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0}
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_D(vc.wp[1], DB, 3=
2, ((func & 0x3) !=3D 0));
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* DB_s: convert from minus=
 long double to
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* unsigned long long
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_D(vc.wp[1], DB, 3=
2,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 ((func & 0x3) !=3D 0) || DB_s);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto update_regs;
>
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0default:
> @@ -589,8 +595,11 @@ cmp_d:
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0_FP_ROUND_=
ZERO(1, SB0);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0_FP_ROUND_=
ZERO(1, SB1);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0}
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_S(vc.wp[0], SB0, =
32, ((func & 0x3) !=3D 0));
> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_S(vc.wp[1], SB1, =
32, ((func & 0x3) !=3D 0));
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* SB*_s: convert from minu=
s float to unsigned int */
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_S(vc.wp[0], SB0, =
32,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 ((func & 0x3) !=3D 0) || SB0_s);
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 FP_TO_INT_S(vc.wp[1], SB1, =
32,
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 ((func & 0x3) !=3D 0) || SB1_s);
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto update_regs;
>
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0default:
> --
> 1.7.0.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" i=
n
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at =A0http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at =A0http://www.tux.org/lkml/
>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix float to unsigned conversion failure with SPE enabled
@ 2010-11-16 11:29     ` Josh Boyer
  0 siblings, 0 replies; 6+ messages in thread
From: Josh Boyer @ 2010-11-16 11:29 UTC (permalink / raw)
  To: Hai Shan; +Cc: linux-kernel, Hai Shan, linuxppc-dev

On Sat, Nov 13, 2010 at 11:11 PM, Hai Shan <haishan.bai@gmail.com> wrote:
> Fixed the failure on converting minus float to unsigned int with SPE enabled
>
> Signed-off-by: Hai Shan <shan.hai@windriver.com>

You should make sure to send PowerPC patches to linuxppc-dev.

josh

> ---
>  arch/powerpc/math-emu/math_efp.c |   17 +++++++++++++----
>  1 files changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/arch/powerpc/math-emu/math_efp.c b/arch/powerpc/math-emu/math_efp.c
> index 41f4ef3..338a128 100644
> --- a/arch/powerpc/math-emu/math_efp.c
> +++ b/arch/powerpc/math-emu/math_efp.c
> @@ -320,7 +320,9 @@ 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));
> +                       /* SB_s: convert from minus float to unsigned int */
> +                       FP_TO_INT_S(vc.wp[1], SB, 32,
> +                                       ((func & 0x3) != 0) || SB_s);
>                        goto update_regs;
>
>                default:
> @@ -458,7 +460,11 @@ cmp_s:
>                        } else {
>                                _FP_ROUND_ZERO(2, DB);
>                        }
> -                       FP_TO_INT_D(vc.wp[1], DB, 32, ((func & 0x3) != 0));
> +                       /* DB_s: convert from minus long double to
> +                        * unsigned long long
> +                        */
> +                       FP_TO_INT_D(vc.wp[1], DB, 32,
> +                                       ((func & 0x3) != 0) || DB_s);
>                        goto update_regs;
>
>                default:
> @@ -589,8 +595,11 @@ 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));
> +                       /* SB*_s: convert from minus float to unsigned int */
> +                       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:
> --
> 1.7.0.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix float to unsigned conversion failure with SPE enabled
  2010-11-16 11:29     ` Josh Boyer
@ 2010-11-17  1:39       ` haishan
  -1 siblings, 0 replies; 6+ messages in thread
From: haishan @ 2010-11-17  1:39 UTC (permalink / raw)
  To: Josh Boyer; +Cc: linuxppc-dev, linux-kernel, Hai Shan

Josh Boyer wrote:
> On Sat, Nov 13, 2010 at 11:11 PM, Hai Shan <haishan.bai@gmail.com> wrote:
>   
>> Fixed the failure on converting minus float to unsigned int with SPE enabled
>>
>> Signed-off-by: Hai Shan <shan.hai@windriver.com>
>>     
>
> You should make sure to send PowerPC patches to linuxppc-dev.
>
>   

Thanks for your suggestion, I will send it to linuxppc-dev.

Regards
Shan Hai

> josh
>
>   
>> ---
>>  arch/powerpc/math-emu/math_efp.c |   17 +++++++++++++----
>>  1 files changed, 13 insertions(+), 4 deletions(-)
>>
>> diff --git a/arch/powerpc/math-emu/math_efp.c b/arch/powerpc/math-emu/math_efp.c
>> index 41f4ef3..338a128 100644
>> --- a/arch/powerpc/math-emu/math_efp.c
>> +++ b/arch/powerpc/math-emu/math_efp.c
>> @@ -320,7 +320,9 @@ 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));
>> +                       /* SB_s: convert from minus float to unsigned int */
>> +                       FP_TO_INT_S(vc.wp[1], SB, 32,
>> +                                       ((func & 0x3) != 0) || SB_s);
>>                        goto update_regs;
>>
>>                default:
>> @@ -458,7 +460,11 @@ cmp_s:
>>                        } else {
>>                                _FP_ROUND_ZERO(2, DB);
>>                        }
>> -                       FP_TO_INT_D(vc.wp[1], DB, 32, ((func & 0x3) != 0));
>> +                       /* DB_s: convert from minus long double to
>> +                        * unsigned long long
>> +                        */
>> +                       FP_TO_INT_D(vc.wp[1], DB, 32,
>> +                                       ((func & 0x3) != 0) || DB_s);
>>                        goto update_regs;
>>
>>                default:
>> @@ -589,8 +595,11 @@ 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));
>> +                       /* SB*_s: convert from minus float to unsigned int */
>> +                       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:
>> --
>> 1.7.0.4
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
>>
>>     

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix float to unsigned conversion failure with SPE enabled
@ 2010-11-17  1:39       ` haishan
  0 siblings, 0 replies; 6+ messages in thread
From: haishan @ 2010-11-17  1:39 UTC (permalink / raw)
  To: Josh Boyer; +Cc: Hai Shan, linux-kernel, linuxppc-dev

Josh Boyer wrote:
> On Sat, Nov 13, 2010 at 11:11 PM, Hai Shan <haishan.bai@gmail.com> wrote:
>   
>> Fixed the failure on converting minus float to unsigned int with SPE enabled
>>
>> Signed-off-by: Hai Shan <shan.hai@windriver.com>
>>     
>
> You should make sure to send PowerPC patches to linuxppc-dev.
>
>   

Thanks for your suggestion, I will send it to linuxppc-dev.

Regards
Shan Hai

> josh
>
>   
>> ---
>>  arch/powerpc/math-emu/math_efp.c |   17 +++++++++++++----
>>  1 files changed, 13 insertions(+), 4 deletions(-)
>>
>> diff --git a/arch/powerpc/math-emu/math_efp.c b/arch/powerpc/math-emu/math_efp.c
>> index 41f4ef3..338a128 100644
>> --- a/arch/powerpc/math-emu/math_efp.c
>> +++ b/arch/powerpc/math-emu/math_efp.c
>> @@ -320,7 +320,9 @@ 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));
>> +                       /* SB_s: convert from minus float to unsigned int */
>> +                       FP_TO_INT_S(vc.wp[1], SB, 32,
>> +                                       ((func & 0x3) != 0) || SB_s);
>>                        goto update_regs;
>>
>>                default:
>> @@ -458,7 +460,11 @@ cmp_s:
>>                        } else {
>>                                _FP_ROUND_ZERO(2, DB);
>>                        }
>> -                       FP_TO_INT_D(vc.wp[1], DB, 32, ((func & 0x3) != 0));
>> +                       /* DB_s: convert from minus long double to
>> +                        * unsigned long long
>> +                        */
>> +                       FP_TO_INT_D(vc.wp[1], DB, 32,
>> +                                       ((func & 0x3) != 0) || DB_s);
>>                        goto update_regs;
>>
>>                default:
>> @@ -589,8 +595,11 @@ 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));
>> +                       /* SB*_s: convert from minus float to unsigned int */
>> +                       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:
>> --
>> 1.7.0.4
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
>>
>>     


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2010-11-17  1:34 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-11-14  4:11 Fixed Powerpc SPE data type conversion failure Hai Shan
2010-11-14  4:11 ` [PATCH] Fix float to unsigned conversion failure with SPE enabled Hai Shan
2010-11-16 11:29   ` Josh Boyer
2010-11-16 11:29     ` Josh Boyer
2010-11-17  1:39     ` haishan
2010-11-17  1:39       ` haishan

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.