libgcc: Fix a bug in _BitInt -> dfp conversions

Message ID ZcdIFVDl6Bby8QZG@tucnak
State New
Headers
Series libgcc: Fix a bug in _BitInt -> dfp conversions |

Checks

Context Check Description
linaro-tcwg-bot/tcwg_gcc_build--master-arm fail Patch failed to apply
linaro-tcwg-bot/tcwg_gcc_build--master-aarch64 fail Patch failed to apply

Commit Message

Jakub Jelinek Feb. 10, 2024, 9:55 a.m. UTC
  Hi!

The ia32 _BitInt support revealed a bug in floatbitint?d.c.
As can be even guessed from how the code is written in the loop,
the intention was to set inexact to non-zero whenever the remainder
after division wasn't zero, but I've ended up just checking whether
the 2 least significant limbs of the remainder were non-zero.
Now, in the dfp/bitint-4.c test in one case the remainder happens
to have least significant 64 bits zero and then the higher limbs are
non-zero; with 32-bit limbs that means 2 least significant limbs are zero
and so the code acted as if it was exactly divisible.

Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux
(including the _BitInt enablement there), ok for trunk?

2024-02-10  Jakub Jelinek  <jakub@redhat.com>

	* soft-fp/floatbitintdd.c (__bid_floatbitintdd): Or in all remainder
	limbs into inexact rather than just first two.
	* soft-fp/floatbitintsd.c (__bid_floatbitintsd): Likewise.
	* soft-fp/floatbitinttd.c (__bid_floatbitinttd): Likewise.


	Jakub
  

Comments

Richard Biener Feb. 10, 2024, 11:43 a.m. UTC | #1
> Am 10.02.2024 um 10:56 schrieb Jakub Jelinek <jakub@redhat.com>:
> 
> Hi!
> 
> The ia32 _BitInt support revealed a bug in floatbitint?d.c.
> As can be even guessed from how the code is written in the loop,
> the intention was to set inexact to non-zero whenever the remainder
> after division wasn't zero, but I've ended up just checking whether
> the 2 least significant limbs of the remainder were non-zero.
> Now, in the dfp/bitint-4.c test in one case the remainder happens
> to have least significant 64 bits zero and then the higher limbs are
> non-zero; with 32-bit limbs that means 2 least significant limbs are zero
> and so the code acted as if it was exactly divisible.
> 
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux
> (including the _BitInt enablement there), ok for trunk?

Ok

Richard 


> 2024-02-10  Jakub Jelinek  <jakub@redhat.com>
> 
>    * soft-fp/floatbitintdd.c (__bid_floatbitintdd): Or in all remainder
>    limbs into inexact rather than just first two.
>    * soft-fp/floatbitintsd.c (__bid_floatbitintsd): Likewise.
>    * soft-fp/floatbitinttd.c (__bid_floatbitinttd): Likewise.
> 
> --- libgcc/soft-fp/floatbitintdd.c.jj    2024-02-09 17:11:47.095675716 +0100
> +++ libgcc/soft-fp/floatbitintdd.c    2024-02-09 18:22:25.786674139 +0100
> @@ -114,7 +114,7 @@ __bid_floatbitintdd (const UBILtype *i,
>               buf + BITINT_END (q_limbs - 1, 0), q_limbs);
>      inexact = buf[q_limbs + pow10_limbs];
>      for (j = 1; j < pow10_limbs; ++j)
> -        inexact |= buf[q_limbs + pow10_limbs + 1];
> +        inexact |= buf[q_limbs + pow10_limbs + j];
>    }
>       else
>    {
> --- libgcc/soft-fp/floatbitintsd.c.jj    2023-09-06 17:34:04.978510447 +0200
> +++ libgcc/soft-fp/floatbitintsd.c    2024-02-09 18:21:50.949158569 +0100
> @@ -113,7 +113,7 @@ __bid_floatbitintsd (const UBILtype *i,
>               buf + BITINT_END (q_limbs - 1, 0), q_limbs);
>      inexact = buf[q_limbs + pow10_limbs];
>      for (j = 1; j < pow10_limbs; ++j)
> -        inexact |= buf[q_limbs + pow10_limbs + 1];
> +        inexact |= buf[q_limbs + pow10_limbs + j];
>    }
>       else
>    {
> --- libgcc/soft-fp/floatbitinttd.c.jj    2024-02-09 17:14:17.485578556 +0100
> +++ libgcc/soft-fp/floatbitinttd.c    2024-02-09 18:21:34.409388566 +0100
> @@ -113,7 +113,7 @@ __bid_floatbitinttd (const UBILtype *i,
>               buf + BITINT_END (q_limbs - 1, 0), q_limbs);
>      inexact = buf[q_limbs + pow10_limbs];
>      for (j = 1; j < pow10_limbs; ++j)
> -        inexact |= buf[q_limbs + pow10_limbs + 1];
> +        inexact |= buf[q_limbs + pow10_limbs + j];
>    }
>       else
>    {
> 
>    Jakub
>
  

Patch

--- libgcc/soft-fp/floatbitintdd.c.jj	2024-02-09 17:11:47.095675716 +0100
+++ libgcc/soft-fp/floatbitintdd.c	2024-02-09 18:22:25.786674139 +0100
@@ -114,7 +114,7 @@  __bid_floatbitintdd (const UBILtype *i,
 			   buf + BITINT_END (q_limbs - 1, 0), q_limbs);
 	  inexact = buf[q_limbs + pow10_limbs];
 	  for (j = 1; j < pow10_limbs; ++j)
-	    inexact |= buf[q_limbs + pow10_limbs + 1];
+	    inexact |= buf[q_limbs + pow10_limbs + j];
 	}
       else
 	{
--- libgcc/soft-fp/floatbitintsd.c.jj	2023-09-06 17:34:04.978510447 +0200
+++ libgcc/soft-fp/floatbitintsd.c	2024-02-09 18:21:50.949158569 +0100
@@ -113,7 +113,7 @@  __bid_floatbitintsd (const UBILtype *i,
 			   buf + BITINT_END (q_limbs - 1, 0), q_limbs);
 	  inexact = buf[q_limbs + pow10_limbs];
 	  for (j = 1; j < pow10_limbs; ++j)
-	    inexact |= buf[q_limbs + pow10_limbs + 1];
+	    inexact |= buf[q_limbs + pow10_limbs + j];
 	}
       else
 	{
--- libgcc/soft-fp/floatbitinttd.c.jj	2024-02-09 17:14:17.485578556 +0100
+++ libgcc/soft-fp/floatbitinttd.c	2024-02-09 18:21:34.409388566 +0100
@@ -113,7 +113,7 @@  __bid_floatbitinttd (const UBILtype *i,
 			   buf + BITINT_END (q_limbs - 1, 0), q_limbs);
 	  inexact = buf[q_limbs + pow10_limbs];
 	  for (j = 1; j < pow10_limbs; ++j)
-	    inexact |= buf[q_limbs + pow10_limbs + 1];
+	    inexact |= buf[q_limbs + pow10_limbs + j];
 	}
       else
 	{