[PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[HAO CHEN GUI]

Hi,
  This patch adds an optab for __builtin_isfinite. The finite check can be
implemented on rs6000 by a single instruction. It needs an optab to be
expanded to the certain sequence of instructions.

  The subsequent patches will implement the expand on rs6000.

  Compared to previous version, the main change is to specify acceptable
modes for the optab.
https://gcc.gnu.org/pipermail/gcc-patches/2024-May/652170.html

  Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no
regressions. Is this OK for trunk?

Thanks
Gui Haochen

ChangeLog
optab: Add isfinite_optab for isfinite builtin

gcc/
	* builtins.cc (interclass_mathfn_icode): Set optab to isfinite_optab
	for isfinite builtin.
	* optabs.def (isfinite_optab): New.
	* doc/md.texi (isfinite): Document.


patch.diff
diff --git a/gcc/builtins.cc b/gcc/builtins.cc
index f8d94c4b435..b8432f84020 100644
--- a/gcc/builtins.cc
+++ b/gcc/builtins.cc
@@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
       errno_set = true; builtin_optab = ilogb_optab; break;
     CASE_FLT_FN (BUILT_IN_ISINF):
       builtin_optab = isinf_optab; break;
-    case BUILT_IN_ISNORMAL:
     case BUILT_IN_ISFINITE:
+      builtin_optab = isfinite_optab; break;
+    case BUILT_IN_ISNORMAL:
     CASE_FLT_FN (BUILT_IN_FINITE):
     case BUILT_IN_FINITED32:
     case BUILT_IN_FINITED64:
diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
index 5730bda80dc..67407fad37d 100644
--- a/gcc/doc/md.texi
+++ b/gcc/doc/md.texi
@@ -8557,6 +8557,15 @@ operand 2, greater than operand 2 or is unordered with operand 2.

 This pattern is not allowed to @code{FAIL}.

+@cindex @code{isfinite@var{m}2} instruction pattern
+@item @samp{isfinite@var{m}2}
+Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
+@code{DFmode}, or @code{TFmode} floating point number and to 0
+otherwise.
+
+If this pattern @code{FAIL}, a call to the library function
+@code{isfinite} is used.
+
 @end table

 @end ifset
diff --git a/gcc/optabs.def b/gcc/optabs.def
index ad14f9328b9..dcd77315c2a 100644
--- a/gcc/optabs.def
+++ b/gcc/optabs.def
@@ -352,6 +352,7 @@ OPTAB_D (fmod_optab, "fmod$a3")
 OPTAB_D (hypot_optab, "hypot$a3")
 OPTAB_D (ilogb_optab, "ilogb$a2")
 OPTAB_D (isinf_optab, "isinf$a2")
+OPTAB_D (isfinite_optab, "isfinite$a2")
 OPTAB_D (issignaling_optab, "issignaling$a2")
 OPTAB_D (ldexp_optab, "ldexp$a3")
 OPTAB_D (log10_optab, "log10$a2")

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Richard Biener]

On Mon, May 27, 2024 at 11:37 AM HAO CHEN GUI <guihaoc@linux.ibm.com> wrote:
>
> Hi,
>   This patch adds an optab for __builtin_isfinite. The finite check can be
> implemented on rs6000 by a single instruction. It needs an optab to be
> expanded to the certain sequence of instructions.
>
>   The subsequent patches will implement the expand on rs6000.
>
>   Compared to previous version, the main change is to specify acceptable
> modes for the optab.
> https://gcc.gnu.org/pipermail/gcc-patches/2024-May/652170.html
>
>   Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no
> regressions. Is this OK for trunk?
>
> Thanks
> Gui Haochen
>
> ChangeLog
> optab: Add isfinite_optab for isfinite builtin
>
> gcc/
>         * builtins.cc (interclass_mathfn_icode): Set optab to isfinite_optab
>         for isfinite builtin.
>         * optabs.def (isfinite_optab): New.
>         * doc/md.texi (isfinite): Document.
>
>
> patch.diff
> diff --git a/gcc/builtins.cc b/gcc/builtins.cc
> index f8d94c4b435..b8432f84020 100644
> --- a/gcc/builtins.cc
> +++ b/gcc/builtins.cc
> @@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
>        errno_set = true; builtin_optab = ilogb_optab; break;
>      CASE_FLT_FN (BUILT_IN_ISINF):
>        builtin_optab = isinf_optab; break;
> -    case BUILT_IN_ISNORMAL:
>      case BUILT_IN_ISFINITE:
> +      builtin_optab = isfinite_optab; break;
> +    case BUILT_IN_ISNORMAL:
>      CASE_FLT_FN (BUILT_IN_FINITE):
>      case BUILT_IN_FINITED32:
>      case BUILT_IN_FINITED64:
> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
> index 5730bda80dc..67407fad37d 100644
> --- a/gcc/doc/md.texi
> +++ b/gcc/doc/md.texi
> @@ -8557,6 +8557,15 @@ operand 2, greater than operand 2 or is unordered with operand 2.
>
>  This pattern is not allowed to @code{FAIL}.
>
> +@cindex @code{isfinite@var{m}2} instruction pattern
> +@item @samp{isfinite@var{m}2}
> +Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
> +@code{DFmode}, or @code{TFmode} floating point number and to 0

It should probably say scalar floating-point mode?  But what about the result?
Is any integer mode OK?  That's esp. important if this might be used on
vector modes.

> +otherwise.
> +
> +If this pattern @code{FAIL}, a call to the library function
> +@code{isfinite} is used.

Or it's otherwise inline expanded?  Or does this imply targets
have to make sure to implement the pattern when isfinite is
not available in libc/libm?  I suggest to leave this sentence out,
we usually only say when a pattern may _not_ FAIL (and usually
FAILing isn't different from not providing a pattern).

>  @end table
>
>  @end ifset
> diff --git a/gcc/optabs.def b/gcc/optabs.def
> index ad14f9328b9..dcd77315c2a 100644
> --- a/gcc/optabs.def
> +++ b/gcc/optabs.def
> @@ -352,6 +352,7 @@ OPTAB_D (fmod_optab, "fmod$a3")
>  OPTAB_D (hypot_optab, "hypot$a3")
>  OPTAB_D (ilogb_optab, "ilogb$a2")
>  OPTAB_D (isinf_optab, "isinf$a2")
> +OPTAB_D (isfinite_optab, "isfinite$a2")
>  OPTAB_D (issignaling_optab, "issignaling$a2")
>  OPTAB_D (ldexp_optab, "ldexp$a3")
>  OPTAB_D (log10_optab, "log10$a2")

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Kewen.Lin]

Hi,

on 2024/5/27 20:54, Richard Biener wrote:
> On Mon, May 27, 2024 at 11:37 AM HAO CHEN GUI <guihaoc@linux.ibm.com> wrote:
>>
>> Hi,
>>   This patch adds an optab for __builtin_isfinite. The finite check can be
>> implemented on rs6000 by a single instruction. It needs an optab to be
>> expanded to the certain sequence of instructions.
>>
>>   The subsequent patches will implement the expand on rs6000.
>>
>>   Compared to previous version, the main change is to specify acceptable
>> modes for the optab.
>> https://gcc.gnu.org/pipermail/gcc-patches/2024-May/652170.html
>>
>>   Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no
>> regressions. Is this OK for trunk?
>>
>> Thanks
>> Gui Haochen
>>
>> ChangeLog
>> optab: Add isfinite_optab for isfinite builtin
>>
>> gcc/
>>         * builtins.cc (interclass_mathfn_icode): Set optab to isfinite_optab
>>         for isfinite builtin.
>>         * optabs.def (isfinite_optab): New.
>>         * doc/md.texi (isfinite): Document.
>>
>>
>> patch.diff
>> diff --git a/gcc/builtins.cc b/gcc/builtins.cc
>> index f8d94c4b435..b8432f84020 100644
>> --- a/gcc/builtins.cc
>> +++ b/gcc/builtins.cc
>> @@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
>>        errno_set = true; builtin_optab = ilogb_optab; break;
>>      CASE_FLT_FN (BUILT_IN_ISINF):
>>        builtin_optab = isinf_optab; break;
>> -    case BUILT_IN_ISNORMAL:
>>      case BUILT_IN_ISFINITE:
>> +      builtin_optab = isfinite_optab; break;
>> +    case BUILT_IN_ISNORMAL:
>>      CASE_FLT_FN (BUILT_IN_FINITE):
>>      case BUILT_IN_FINITED32:
>>      case BUILT_IN_FINITED64:
>> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
>> index 5730bda80dc..67407fad37d 100644
>> --- a/gcc/doc/md.texi
>> +++ b/gcc/doc/md.texi
>> @@ -8557,6 +8557,15 @@ operand 2, greater than operand 2 or is unordered with operand 2.
>>
>>  This pattern is not allowed to @code{FAIL}.
>>
>> +@cindex @code{isfinite@var{m}2} instruction pattern
>> +@item @samp{isfinite@var{m}2}
>> +Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
>> +@code{DFmode}, or @code{TFmode} floating point number and to 0
> 
> It should probably say scalar floating-point mode?  But what about the result?
> Is any integer mode OK?  That's esp. important if this might be used on
> vector modes.
> 
>> +otherwise.
>> +
>> +If this pattern @code{FAIL}, a call to the library function
>> +@code{isfinite} is used.
> 
> Or it's otherwise inline expanded?  Or does this imply targets
> have to make sure to implement the pattern when isfinite is
> not available in libc/libm?  I suggest to leave this sentence out,
> we usually only say when a pattern may _not_ FAIL (and usually
> FAILing isn't different from not providing a pattern).

As Haochen's previous reply, I think there are three cases:
  1) no optab defined, fold in a generic way;
  2) optab defined, SUCC, expand as what it defines;
  3) optab defined, FAIL, generate a library call;

From above, I had the concern that ports may assume FAILing can
fall back with the generic folding, but it's not actually.
Does your comment imply ports usually don't make such assumption
(or they just check what happens for FAIL)?

BR,
Kewen

> 
>>  @end table
>>
>>  @end ifset
>> diff --git a/gcc/optabs.def b/gcc/optabs.def
>> index ad14f9328b9..dcd77315c2a 100644
>> --- a/gcc/optabs.def
>> +++ b/gcc/optabs.def
>> @@ -352,6 +352,7 @@ OPTAB_D (fmod_optab, "fmod$a3")
>>  OPTAB_D (hypot_optab, "hypot$a3")
>>  OPTAB_D (ilogb_optab, "ilogb$a2")
>>  OPTAB_D (isinf_optab, "isinf$a2")
>> +OPTAB_D (isfinite_optab, "isfinite$a2")
>>  OPTAB_D (issignaling_optab, "issignaling$a2")
>>  OPTAB_D (ldexp_optab, "ldexp$a3")
>>  OPTAB_D (log10_optab, "log10$a2")

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Richard Biener]

On Tue, May 28, 2024 at 9:09 AM Kewen.Lin <linkw@linux.ibm.com> wrote:
>
> Hi,
>
> on 2024/5/27 20:54, Richard Biener wrote:
> > On Mon, May 27, 2024 at 11:37 AM HAO CHEN GUI <guihaoc@linux.ibm.com> wrote:
> >>
> >> Hi,
> >>   This patch adds an optab for __builtin_isfinite. The finite check can be
> >> implemented on rs6000 by a single instruction. It needs an optab to be
> >> expanded to the certain sequence of instructions.
> >>
> >>   The subsequent patches will implement the expand on rs6000.
> >>
> >>   Compared to previous version, the main change is to specify acceptable
> >> modes for the optab.
> >> https://gcc.gnu.org/pipermail/gcc-patches/2024-May/652170.html
> >>
> >>   Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no
> >> regressions. Is this OK for trunk?
> >>
> >> Thanks
> >> Gui Haochen
> >>
> >> ChangeLog
> >> optab: Add isfinite_optab for isfinite builtin
> >>
> >> gcc/
> >>         * builtins.cc (interclass_mathfn_icode): Set optab to isfinite_optab
> >>         for isfinite builtin.
> >>         * optabs.def (isfinite_optab): New.
> >>         * doc/md.texi (isfinite): Document.
> >>
> >>
> >> patch.diff
> >> diff --git a/gcc/builtins.cc b/gcc/builtins.cc
> >> index f8d94c4b435..b8432f84020 100644
> >> --- a/gcc/builtins.cc
> >> +++ b/gcc/builtins.cc
> >> @@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
> >>        errno_set = true; builtin_optab = ilogb_optab; break;
> >>      CASE_FLT_FN (BUILT_IN_ISINF):
> >>        builtin_optab = isinf_optab; break;
> >> -    case BUILT_IN_ISNORMAL:
> >>      case BUILT_IN_ISFINITE:
> >> +      builtin_optab = isfinite_optab; break;
> >> +    case BUILT_IN_ISNORMAL:
> >>      CASE_FLT_FN (BUILT_IN_FINITE):
> >>      case BUILT_IN_FINITED32:
> >>      case BUILT_IN_FINITED64:
> >> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
> >> index 5730bda80dc..67407fad37d 100644
> >> --- a/gcc/doc/md.texi
> >> +++ b/gcc/doc/md.texi
> >> @@ -8557,6 +8557,15 @@ operand 2, greater than operand 2 or is unordered with operand 2.
> >>
> >>  This pattern is not allowed to @code{FAIL}.
> >>
> >> +@cindex @code{isfinite@var{m}2} instruction pattern
> >> +@item @samp{isfinite@var{m}2}
> >> +Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
> >> +@code{DFmode}, or @code{TFmode} floating point number and to 0
> >
> > It should probably say scalar floating-point mode?  But what about the result?
> > Is any integer mode OK?  That's esp. important if this might be used on
> > vector modes.
> >
> >> +otherwise.
> >> +
> >> +If this pattern @code{FAIL}, a call to the library function
> >> +@code{isfinite} is used.
> >
> > Or it's otherwise inline expanded?  Or does this imply targets
> > have to make sure to implement the pattern when isfinite is
> > not available in libc/libm?  I suggest to leave this sentence out,
> > we usually only say when a pattern may _not_ FAIL (and usually
> > FAILing isn't different from not providing a pattern).
>
> As Haochen's previous reply, I think there are three cases:
>   1) no optab defined, fold in a generic way;
>   2) optab defined, SUCC, expand as what it defines;
>   3) optab defined, FAIL, generate a library call;
>
> From above, I had the concern that ports may assume FAILing can
> fall back with the generic folding, but it's not actually.

Hmm, but it should.  Can you make that work?

> Does your comment imply ports usually don't make such assumption
> (or they just check what happens for FAIL)?
>
> BR,
> Kewen
>
> >
> >>  @end table
> >>
> >>  @end ifset
> >> diff --git a/gcc/optabs.def b/gcc/optabs.def
> >> index ad14f9328b9..dcd77315c2a 100644
> >> --- a/gcc/optabs.def
> >> +++ b/gcc/optabs.def
> >> @@ -352,6 +352,7 @@ OPTAB_D (fmod_optab, "fmod$a3")
> >>  OPTAB_D (hypot_optab, "hypot$a3")
> >>  OPTAB_D (ilogb_optab, "ilogb$a2")
> >>  OPTAB_D (isinf_optab, "isinf$a2")
> >> +OPTAB_D (isfinite_optab, "isfinite$a2")
> >>  OPTAB_D (issignaling_optab, "issignaling$a2")
> >>  OPTAB_D (ldexp_optab, "ldexp$a3")
> >>  OPTAB_D (log10_optab, "log10$a2")
>
>
>

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Richard Sandiford]

HAO CHEN GUI <guihaoc@linux.ibm.com> writes:
> Hi,
>   This patch adds an optab for __builtin_isfinite. The finite check can be
> implemented on rs6000 by a single instruction. It needs an optab to be
> expanded to the certain sequence of instructions.
>
>   The subsequent patches will implement the expand on rs6000.
>
>   Compared to previous version, the main change is to specify acceptable
> modes for the optab.
> https://gcc.gnu.org/pipermail/gcc-patches/2024-May/652170.html
>
>   Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no
> regressions. Is this OK for trunk?
>
> Thanks
> Gui Haochen
>
> ChangeLog
> optab: Add isfinite_optab for isfinite builtin
>
> gcc/
> 	* builtins.cc (interclass_mathfn_icode): Set optab to isfinite_optab
> 	for isfinite builtin.
> 	* optabs.def (isfinite_optab): New.
> 	* doc/md.texi (isfinite): Document.
>
>
> patch.diff
> diff --git a/gcc/builtins.cc b/gcc/builtins.cc
> index f8d94c4b435..b8432f84020 100644
> --- a/gcc/builtins.cc
> +++ b/gcc/builtins.cc
> @@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
>        errno_set = true; builtin_optab = ilogb_optab; break;
>      CASE_FLT_FN (BUILT_IN_ISINF):
>        builtin_optab = isinf_optab; break;
> -    case BUILT_IN_ISNORMAL:
>      case BUILT_IN_ISFINITE:
> +      builtin_optab = isfinite_optab; break;
> +    case BUILT_IN_ISNORMAL:
>      CASE_FLT_FN (BUILT_IN_FINITE):
>      case BUILT_IN_FINITED32:
>      case BUILT_IN_FINITED64:
> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
> index 5730bda80dc..67407fad37d 100644
> --- a/gcc/doc/md.texi
> +++ b/gcc/doc/md.texi
> @@ -8557,6 +8557,15 @@ operand 2, greater than operand 2 or is unordered with operand 2.
>
>  This pattern is not allowed to @code{FAIL}.
>
> +@cindex @code{isfinite@var{m}2} instruction pattern
> +@item @samp{isfinite@var{m}2}
> +Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
> +@code{DFmode}, or @code{TFmode} floating point number and to 0
> +otherwise.

This has probably already been discussed, sorry, but how about defining
the optab to return a strict 0/1 result, rather than just zero/nonzero?
I realise that's stricter than the underlying math.h routines, but it
would in principle avoid the need to expand extra instructions in
a setcc-like operation.

Richard

> +
> +If this pattern @code{FAIL}, a call to the library function
> +@code{isfinite} is used.
> +
>  @end table
>
>  @end ifset
> diff --git a/gcc/optabs.def b/gcc/optabs.def
> index ad14f9328b9..dcd77315c2a 100644
> --- a/gcc/optabs.def
> +++ b/gcc/optabs.def
> @@ -352,6 +352,7 @@ OPTAB_D (fmod_optab, "fmod$a3")
>  OPTAB_D (hypot_optab, "hypot$a3")
>  OPTAB_D (ilogb_optab, "ilogb$a2")
>  OPTAB_D (isinf_optab, "isinf$a2")
> +OPTAB_D (isfinite_optab, "isfinite$a2")
>  OPTAB_D (issignaling_optab, "issignaling$a2")
>  OPTAB_D (ldexp_optab, "ldexp$a3")
>  OPTAB_D (log10_optab, "log10$a2")

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Segher Boessenkool]

On Tue, May 28, 2024 at 02:09:50PM +0200, Richard Biener wrote:
> On Tue, May 28, 2024 at 9:09 AM Kewen.Lin <linkw@linux.ibm.com> wrote:
> > As Haochen's previous reply, I think there are three cases:
> >   1) no optab defined, fold in a generic way;
> >   2) optab defined, SUCC, expand as what it defines;
> >   3) optab defined, FAIL, generate a library call;
> >
> > From above, I had the concern that ports may assume FAILing can
> > fall back with the generic folding, but it's not actually.
> 
> Hmm, but it should.  Can you make that work?

That certainly would be the least surprising!


Segher

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Segher Boessenkool]

Hi!

On Mon, May 27, 2024 at 05:37:23PM +0800, HAO CHEN GUI wrote:
> --- a/gcc/builtins.cc
> +++ b/gcc/builtins.cc
> @@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
>        errno_set = true; builtin_optab = ilogb_optab; break;
>      CASE_FLT_FN (BUILT_IN_ISINF):
>        builtin_optab = isinf_optab; break;
> -    case BUILT_IN_ISNORMAL:
>      case BUILT_IN_ISFINITE:
> +      builtin_optab = isfinite_optab; break;

This needs a line break after the first ; (like after *any* semicolon
in C).  It is rather important that every "break;" stands out :-)

> +@cindex @code{isfinite@var{m}2} instruction pattern
> +@item @samp{isfinite@var{m}2}
> +Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
> +@code{DFmode}, or @code{TFmode} floating point number and to 0
> +otherwise.

operand 0 is the output of the builtin, right?  So write that instead?
"Return 1 if the operand (a scalar floating poiint number) is finite",
or such?


Segher

Re: [PATCHv3] Optab: add isfinite_optab for __builtin_isfinite

[Kewen.Lin]

on 2024/5/28 20:09, Richard Biener wrote:
> On Tue, May 28, 2024 at 9:09 AM Kewen.Lin <linkw@linux.ibm.com> wrote:
>>
>> Hi,
>>
>> on 2024/5/27 20:54, Richard Biener wrote:
>>> On Mon, May 27, 2024 at 11:37 AM HAO CHEN GUI <guihaoc@linux.ibm.com> wrote:
>>>>
>>>> Hi,
>>>>   This patch adds an optab for __builtin_isfinite. The finite check can be
>>>> implemented on rs6000 by a single instruction. It needs an optab to be
>>>> expanded to the certain sequence of instructions.
>>>>
>>>>   The subsequent patches will implement the expand on rs6000.
>>>>
>>>>   Compared to previous version, the main change is to specify acceptable
>>>> modes for the optab.
>>>> https://gcc.gnu.org/pipermail/gcc-patches/2024-May/652170.html
>>>>
>>>>   Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no
>>>> regressions. Is this OK for trunk?
>>>>
>>>> Thanks
>>>> Gui Haochen
>>>>
>>>> ChangeLog
>>>> optab: Add isfinite_optab for isfinite builtin
>>>>
>>>> gcc/
>>>>         * builtins.cc (interclass_mathfn_icode): Set optab to isfinite_optab
>>>>         for isfinite builtin.
>>>>         * optabs.def (isfinite_optab): New.
>>>>         * doc/md.texi (isfinite): Document.
>>>>
>>>>
>>>> patch.diff
>>>> diff --git a/gcc/builtins.cc b/gcc/builtins.cc
>>>> index f8d94c4b435..b8432f84020 100644
>>>> --- a/gcc/builtins.cc
>>>> +++ b/gcc/builtins.cc
>>>> @@ -2459,8 +2459,9 @@ interclass_mathfn_icode (tree arg, tree fndecl)
>>>>        errno_set = true; builtin_optab = ilogb_optab; break;
>>>>      CASE_FLT_FN (BUILT_IN_ISINF):
>>>>        builtin_optab = isinf_optab; break;
>>>> -    case BUILT_IN_ISNORMAL:
>>>>      case BUILT_IN_ISFINITE:
>>>> +      builtin_optab = isfinite_optab; break;
>>>> +    case BUILT_IN_ISNORMAL:
>>>>      CASE_FLT_FN (BUILT_IN_FINITE):
>>>>      case BUILT_IN_FINITED32:
>>>>      case BUILT_IN_FINITED64:
>>>> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
>>>> index 5730bda80dc..67407fad37d 100644
>>>> --- a/gcc/doc/md.texi
>>>> +++ b/gcc/doc/md.texi
>>>> @@ -8557,6 +8557,15 @@ operand 2, greater than operand 2 or is unordered with operand 2.
>>>>
>>>>  This pattern is not allowed to @code{FAIL}.
>>>>
>>>> +@cindex @code{isfinite@var{m}2} instruction pattern
>>>> +@item @samp{isfinite@var{m}2}
>>>> +Set operand 0 to nonzero if operand 1 is a finite @code{SFmode},
>>>> +@code{DFmode}, or @code{TFmode} floating point number and to 0
>>>
>>> It should probably say scalar floating-point mode?  But what about the result?
>>> Is any integer mode OK?  That's esp. important if this might be used on
>>> vector modes.
>>>
>>>> +otherwise.
>>>> +
>>>> +If this pattern @code{FAIL}, a call to the library function
>>>> +@code{isfinite} is used.
>>>
>>> Or it's otherwise inline expanded?  Or does this imply targets
>>> have to make sure to implement the pattern when isfinite is
>>> not available in libc/libm?  I suggest to leave this sentence out,
>>> we usually only say when a pattern may _not_ FAIL (and usually
>>> FAILing isn't different from not providing a pattern).
>>
>> As Haochen's previous reply, I think there are three cases:
>>   1) no optab defined, fold in a generic way;
>>   2) optab defined, SUCC, expand as what it defines;
>>   3) optab defined, FAIL, generate a library call;
>>
>> From above, I had the concern that ports may assume FAILing can
>> fall back with the generic folding, but it's not actually.
> 
> Hmm, but it should.  Can you make that work?

Good point, sure, I'll follow up this.

BR,
Kewen

> 
>> Does your comment imply ports usually don't make such assumption
>> (or they just check what happens for FAIL)?
>>
>> BR,
>> Kewen
>>
>>>
>>>>  @end table
>>>>
>>>>  @end ifset
>>>> diff --git a/gcc/optabs.def b/gcc/optabs.def
>>>> index ad14f9328b9..dcd77315c2a 100644
>>>> --- a/gcc/optabs.def
>>>> +++ b/gcc/optabs.def
>>>> @@ -352,6 +352,7 @@ OPTAB_D (fmod_optab, "fmod$a3")
>>>>  OPTAB_D (hypot_optab, "hypot$a3")
>>>>  OPTAB_D (ilogb_optab, "ilogb$a2")
>>>>  OPTAB_D (isinf_optab, "isinf$a2")
>>>> +OPTAB_D (isfinite_optab, "isfinite$a2")
>>>>  OPTAB_D (issignaling_optab, "issignaling$a2")
>>>>  OPTAB_D (ldexp_optab, "ldexp$a3")
>>>>  OPTAB_D (log10_optab, "log10$a2")
>>
>>
>>