Re: memmove() in 2.4.0-test12, alpha platform

2000-12-22 Thread Alexander Zarochentcev 

On Thu, Dec 21, 2000 at 06:40:46PM +0300, Ivan Kokshaysky wrote:
> On Wed, Dec 20, 2000 at 10:03:42PM +0300, Alexander Zarochentcev wrote:
> > New (since test12) optimized memmove function seems to be broken
> > on alpha platform. 
> 
> Indeed it is.
> 
> > If dest and src arguments are misaligned, new memmove does wrong things.
> 
> Actually it broke when dest < src. Incrementing pointers should be used in
> this case.
> 
> This patch seems to work (tested in userspace).
> 
> Ivan.
> 
> --- linux/arch/alpha/lib/memmove.S.orig   Tue Dec 12 00:46:26 2000
> +++ linux/arch/alpha/lib/memmove.SThu Dec 21 18:32:59 2000
> @@ -26,12 +26,16 @@ memmove:
>   bne $1,memcpy
>  
>   and $2,7,$2 /* Test for src/dest co-alignment.  */
> - bne $2,$misaligned
> + and $16,7,$1
> + cmpule $16,$17,$3
> + bne $3,$memmove_up  /* dest < src */
>  
>   and $4,7,$1
> - beq $1,$skip_aligned_byte_loop_head
> + bne $2,$misaligned_dn
> + unop
> + beq $1,$skip_aligned_byte_loop_head_dn
>  
> -$aligned_byte_loop_head:
> +$aligned_byte_loop_head_dn:
>   lda $4,-1($4)
>   lda $5,-1($5)
>   unop
> @@ -48,13 +52,13 @@ $aligned_byte_loop_head:
>   and $4,7,$6
>  
>   stq_u $1,0($4)
> - bne $6,$aligned_byte_loop_head
> + bne $6,$aligned_byte_loop_head_dn
>  
> -$skip_aligned_byte_loop_head:
> +$skip_aligned_byte_loop_head_dn:
>   lda $18,-8($18)
> - blt $18,$skip_aligned_word_loop
> + blt $18,$skip_aligned_word_loop_dn
>  
> -$aligned_word_loop:
> +$aligned_word_loop_dn:
>   ldq $1,-8($5)
>   nop
>   lda $5,-8($5)
> @@ -63,22 +67,22 @@ $aligned_word_loop:
>   stq $1,-8($4)
>   nop
>   lda $4,-8($4)
> - bge $18,$aligned_word_loop
> + bge $18,$aligned_word_loop_dn
>  
> -$skip_aligned_word_loop:
> +$skip_aligned_word_loop_dn:
>   lda $18,8($18)
> - bgt $18,$byte_loop_tail
> + bgt $18,$byte_loop_tail_dn
>   unop
>   ret $31,($26),1
>  
>   .align 4
> -$misaligned:
> +$misaligned_dn:
>   nop
>   fnop
>   unop
>   beq $18,$egress
>  
> -$byte_loop_tail:
> +$byte_loop_tail_dn:
>   ldq_u $3,-1($5)
>   ldq_u $2,-1($4)
>   lda $5,-1($5)
> @@ -91,8 +95,77 @@ $byte_loop_tail:
>  
>   bis $1,$2,$1
>   stq_u $1,0($4)
> + bgt $18,$byte_loop_tail_dn
> + br $egress
> +
> +$memmove_up:
> + mov $16,$4
> + mov $17,$5
> + bne $2,$misaligned_up
> + beq $1,$skip_aligned_byte_loop_head_up
> +
> +$aligned_byte_loop_head_up:
> + unop
> + ble $18,$egress
> + ldq_u $3,0($5)
> + ldq_u $2,0($4)
> +
> + lda $18,-1($18)
> + extbl $3,$5,$1
> + insbl $1,$4,$1
> + mskbl $2,$4,$2
> +
> + bis $1,$2,$1
> + lda $5,1($5)
> + stq_u $1,0($4)
> + lda $4,1($4)
> +
> + and $4,7,$6
> + bne $6,$aligned_byte_loop_head_up
> +
> +$skip_aligned_byte_loop_head_up:
> + lda $18,-8($18)
> + blt $18,$skip_aligned_word_loop_up
> +
> +$aligned_word_loop_up:
> + ldq $1,0($5)
> + nop
> + lda $5,8($5)
> + lda $18,-8($18)
> +
> + stq $1,0($4)
> + nop
> + lda $4,8($4)
> + bge $18,$aligned_word_loop_up
> +
> +$skip_aligned_word_loop_up:
> + lda $18,8($18)
> + bgt $18,$byte_loop_tail_up
> + unop
> + ret $31,($26),1
> +
> + .align 4
> +$misaligned_up:
> + nop
> + fnop
> + unop
> + beq $18,$egress
> +
> +$byte_loop_tail_up:
> + ldq_u $3,0($5)
> + ldq_u $2,0($4)
> + lda $18,-1($18)
> + extbl $3,$5,$1
> +
> + insbl $1,$4,$1
> + mskbl $2,$4,$2
> + bis $1,$2,$1
> + stq_u $1,0($4)
> +
> + lda $5,1($5)
> + lda $4,1($4)
>   nop
> - bgt $18,$byte_loop_tail
> + bgt $18,$byte_loop_tail_up
>  
>  $egress:
>   ret $31,($26),1

Ok. It works.

Thanks.

-- 
Alex.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
Please read the FAQ at http://www.tux.org/lkml/

Re: memmove() in 2.4.0-test12, alpha platform

2000-12-21 Thread Ivan Kokshaysky 

On Wed, Dec 20, 2000 at 10:03:42PM +0300, Alexander Zarochentcev wrote:
> New (since test12) optimized memmove function seems to be broken
> on alpha platform. 

Indeed it is.

> If dest and src arguments are misaligned, new memmove does wrong things.

Actually it broke when dest < src. Incrementing pointers should be used in
this case.

This patch seems to work (tested in userspace).

Ivan.

--- linux/arch/alpha/lib/memmove.S.orig Tue Dec 12 00:46:26 2000
+++ linux/arch/alpha/lib/memmove.S  Thu Dec 21 18:32:59 2000
@@ -26,12 +26,16 @@ memmove:
bne $1,memcpy
 
and $2,7,$2 /* Test for src/dest co-alignment.  */
-   bne $2,$misaligned
+   and $16,7,$1
+   cmpule $16,$17,$3
+   bne $3,$memmove_up  /* dest < src */
 
and $4,7,$1
-   beq $1,$skip_aligned_byte_loop_head
+   bne $2,$misaligned_dn
+   unop
+   beq $1,$skip_aligned_byte_loop_head_dn
 
-$aligned_byte_loop_head:
+$aligned_byte_loop_head_dn:
lda $4,-1($4)
lda $5,-1($5)
unop
@@ -48,13 +52,13 @@ $aligned_byte_loop_head:
and $4,7,$6
 
stq_u $1,0($4)
-   bne $6,$aligned_byte_loop_head
+   bne $6,$aligned_byte_loop_head_dn
 
-$skip_aligned_byte_loop_head:
+$skip_aligned_byte_loop_head_dn:
lda $18,-8($18)
-   blt $18,$skip_aligned_word_loop
+   blt $18,$skip_aligned_word_loop_dn
 
-$aligned_word_loop:
+$aligned_word_loop_dn:
ldq $1,-8($5)
nop
lda $5,-8($5)
@@ -63,22 +67,22 @@ $aligned_word_loop:
stq $1,-8($4)
nop
lda $4,-8($4)
-   bge $18,$aligned_word_loop
+   bge $18,$aligned_word_loop_dn
 
-$skip_aligned_word_loop:
+$skip_aligned_word_loop_dn:
lda $18,8($18)
-   bgt $18,$byte_loop_tail
+   bgt $18,$byte_loop_tail_dn
unop
ret $31,($26),1
 
.align 4
-$misaligned:
+$misaligned_dn:
nop
fnop
unop
beq $18,$egress
 
-$byte_loop_tail:
+$byte_loop_tail_dn:
ldq_u $3,-1($5)
ldq_u $2,-1($4)
lda $5,-1($5)
@@ -91,8 +95,77 @@ $byte_loop_tail:
 
bis $1,$2,$1
stq_u $1,0($4)
+   bgt $18,$byte_loop_tail_dn
+   br $egress
+
+$memmove_up:
+   mov $16,$4
+   mov $17,$5
+   bne $2,$misaligned_up
+   beq $1,$skip_aligned_byte_loop_head_up
+
+$aligned_byte_loop_head_up:
+   unop
+   ble $18,$egress
+   ldq_u $3,0($5)
+   ldq_u $2,0($4)
+
+   lda $18,-1($18)
+   extbl $3,$5,$1
+   insbl $1,$4,$1
+   mskbl $2,$4,$2
+
+   bis $1,$2,$1
+   lda $5,1($5)
+   stq_u $1,0($4)
+   lda $4,1($4)
+
+   and $4,7,$6
+   bne $6,$aligned_byte_loop_head_up
+
+$skip_aligned_byte_loop_head_up:
+   lda $18,-8($18)
+   blt $18,$skip_aligned_word_loop_up
+
+$aligned_word_loop_up:
+   ldq $1,0($5)
+   nop
+   lda $5,8($5)
+   lda $18,-8($18)
+
+   stq $1,0($4)
+   nop
+   lda $4,8($4)
+   bge $18,$aligned_word_loop_up
+
+$skip_aligned_word_loop_up:
+   lda $18,8($18)
+   bgt $18,$byte_loop_tail_up
+   unop
+   ret $31,($26),1
+
+   .align 4
+$misaligned_up:
+   nop
+   fnop
+   unop
+   beq $18,$egress
+
+$byte_loop_tail_up:
+   ldq_u $3,0($5)
+   ldq_u $2,0($4)
+   lda $18,-1($18)
+   extbl $3,$5,$1
+
+   insbl $1,$4,$1
+   mskbl $2,$4,$2
+   bis $1,$2,$1
+   stq_u $1,0($4)
+
+   lda $5,1($5)
+   lda $4,1($4)
nop
-   bgt $18,$byte_loop_tail
+   bgt $18,$byte_loop_tail_up
 
 $egress:
ret $31,($26),1

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
Please read the FAQ at http://www.tux.org/lkml/

memmove() in 2.4.0-test12, alpha platform

2000-12-20 Thread Alexander Zarochentcev 

Hello !

New (since test12) optimized memmove function seems to be broken
on alpha platform. 

If dest and src arguments are misaligned, new memmove does wrong things.


example:
 
   static char p[] = "abcdefghijklmnopkrstuvwxyz01234567890";
   memmove(p + 2, p + 13, 17);
   printk ("DEBUG: memmove test: %s\n", p);

produces:

   DEBUG: memmove test: abyz0123tuvwxyz0123tuvwxyz01234567890


Old memmove variant didn't have this problem.

Thanks,
Alex.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
Please read the FAQ at http://www.tux.org/lkml/