[PATCH] Don't fail on SS_ABS

[PATCH] Don't fail on SS_ABS

[Jie Zhang]

[-- Attachment #1: Type: text/plain, Size: 140 bytes --]

SS_ABS is not handled in simplify_const_unary_operation and ss_abs is 
not documented. This patch fixes these. Is it OK for mainline?


Jie

[-- Attachment #2: gcc-fix-ss_abs.diff --]
[-- Type: text/x-patch, Size: 975 bytes --]


	* simplify-rtx.c (simplify_const_unary_operation): Handle
	SS_ABS.
	* doc/rtl.texi: Document ss_abs.

Index: doc/rtl.texi
===================================================================
--- doc/rtl.texi	(revision 152895)
+++ doc/rtl.texi	(working copy)
@@ -2364,9 +2364,14 @@ Similar but represent left and right rot
 use @code{rotate}.
 
 @findex abs
+@findex ss_abs
 @cindex absolute value
 @item (abs:@var{m} @var{x})
+@item (ss_abs:@var{m} @var{x})
 Represents the absolute value of @var{x}, computed in mode @var{m}.
+@code{ss_abs} ensures that an out-of-bounds result saturates to the
+maximum signed value.
+
 
 @findex sqrt
 @cindex square root
Index: simplify-rtx.c
===================================================================
--- simplify-rtx.c	(revision 152895)
+++ simplify-rtx.c	(working copy)
@@ -1245,6 +1245,7 @@ simplify_const_unary_operation (enum rtx
 	case US_TRUNCATE:
 	case SS_NEG:
 	case US_NEG:
+	case SS_ABS:
 	  return 0;
 
 	default:

Re: [PATCH] Don't fail on SS_ABS

[Bernd Schmidt]

Jie Zhang wrote:
> SS_ABS is not handled in simplify_const_unary_operation and ss_abs is
> not documented. This patch fixes these. Is it OK for mainline?

> +@code{ss_abs} ensures that an out-of-bounds result saturates to the
> +maximum signed value.
> +

Maybe clarify that as "maximum positive number representable as a signed
value"?  Otherwise ok.


Bernd
-- 
This footer brought to you by insane German lawmakers.
Analog Devices GmbH      Wilhelm-Wagenfeld-Str. 6      80807 Muenchen
Sitz der Gesellschaft Muenchen, Registergericht Muenchen HRB 40368
Geschaeftsfuehrer Thomas Wessel, William A. Martin, Margaret Seif

Re: [PATCH] Don't fail on SS_ABS

[Jie Zhang]

Hi Bernd,

Bernd Schmidt wrote:
> Jie Zhang wrote:
>> SS_ABS is not handled in simplify_const_unary_operation and ss_abs is
>> not documented. This patch fixes these. Is it OK for mainline?
> 
>> +@code{ss_abs} ensures that an out-of-bounds result saturates to the
>> +maximum signed value.
>> +
> 
> Maybe clarify that as "maximum positive number representable as a signed
> value"?  Otherwise ok.
> 
But "maximum signed value" is the same meaning of "maximum positive 
number representable as a signed value", isn't? I prefer the former, 
which is simple and short. Is it OK if I commit the original patch?


Jie

Re: [PATCH] Don't fail on SS_ABS

[Bernd Schmidt]

Jie Zhang wrote:
> Hi Bernd,
> 
> Bernd Schmidt wrote:
>> Jie Zhang wrote:
>>> SS_ABS is not handled in simplify_const_unary_operation and ss_abs is
>>> not documented. This patch fixes these. Is it OK for mainline?
>>
>>> +@code{ss_abs} ensures that an out-of-bounds result saturates to the
>>> +maximum signed value.
>>> +
>>
>> Maybe clarify that as "maximum positive number representable as a signed
>> value"?  Otherwise ok.
>>
> But "maximum signed value" is the same meaning of "maximum positive
> number representable as a signed value", isn't? I prefer the former,
> which is simple and short. Is it OK if I commit the original patch?

Go on then.  If there's a native speaker reading this who prefers one
form or the other we should probably let them choose, but I guess it's
not that important.


Bernd
-- 
This footer brought to you by insane German lawmakers.
Analog Devices GmbH      Wilhelm-Wagenfeld-Str. 6      80807 Muenchen
Sitz der Gesellschaft Muenchen, Registergericht Muenchen HRB 40368
Geschaeftsfuehrer Thomas Wessel, William A. Martin, Margaret Seif

Re: [PATCH] Don't fail on SS_ABS

[Jie Zhang]

Bernd Schmidt wrote:
> Jie Zhang wrote:
>> Hi Bernd,
>>
>> Bernd Schmidt wrote:
>>> Jie Zhang wrote:
>>>> SS_ABS is not handled in simplify_const_unary_operation and ss_abs is
>>>> not documented. This patch fixes these. Is it OK for mainline?
>>>> +@code{ss_abs} ensures that an out-of-bounds result saturates to the
>>>> +maximum signed value.
>>>> +
>>> Maybe clarify that as "maximum positive number representable as a signed
>>> value"?  Otherwise ok.
>>>
>> But "maximum signed value" is the same meaning of "maximum positive
>> number representable as a signed value", isn't? I prefer the former,
>> which is simple and short. Is it OK if I commit the original patch?
> 
> Go on then.  If there's a native speaker reading this who prefers one
> form or the other we should probably let them choose, but I guess it's
> not that important.
> 
Thanks. Committed on trunk.

Jie