target/5985: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

target/5985: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

[hp]

>Number:         5985
>Category:       target
>Synopsis:       fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Mar 17 05:46:01 PST 2002
>Closed-Date:
>Last-Modified:
>Originator:     Hans-Peter Nilsson
>Release:        trunk and 3.1 branch and everywhere else
>Organization:
>Environment:

>Description:
>From <URL:http://gcc.gnu.org/ml/gcc-patches/2002-03/msg01023.html>, edited:
While checking on the floating point operations and comparing
with the manual and other targets, it seems several targets
(briefly: a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64,
m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa) have fix_truncMN2
patterns that are incorrectly expressed.  For example for M=df
N=di, the input operand should be (fix:DI (fix:DF (DF-operand))
not just (fix:DI (DF-operand)), since the latter expression
requires/promises that the operand is already truncated to an
integer in a float-mode, which is the input operand for the
fixdfdi2 pattern.  Similarly for fixuns_truncMN2.
>How-To-Repeat:
Read .md for target, specifically fix_truncMN2 and fixuns_truncMN2
patterns.  Compare with docs and other targets, like sparc.md, m68k.md.
Marvel at how it doesn't seem to affect anything.
>Fix:
See sparc.md, m68k.md, mmix.md.
>Release-Note:
>Audit-Trail:
>Unformatted:

Re: target/5985: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

[Hans-Peter Nilsson]

The following reply was made to PR target/5985; it has been noted by GNATS.

From: Hans-Peter Nilsson <hp@bitrange.com>
To: neroden@gcc.gnu.org,  <gcc-bugs@gcc.gnu.org>,  <gcc-prs@gcc.gnu.org>, 
     <hp@bitrange.com>,  <nobody@gcc.gnu.org>,  <gcc-gnats@gcc.gnu.org>
Cc:  
Subject: Re: target/5985: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2
 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips,
 ns32k, pj, romp, rs6000, s390, sh, xtensa
Date: Thu, 2 Jan 2003 15:18:02 -0500 (EST)

 On 2 Jan 2003 neroden@gcc.gnu.org wrote:
 > Synopsis: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa
 >     Theoretically annoyance which appears to not be real; in feedback 9 months.
 
 The FP bits have changed, IIRC by RTH himself making changes in
 a direction against his earlier comment.  I'll revisit some day.
 The issue is real enough for people writing new ports.  Sorry I
 haven't persisted and still don't.
 
 brgds, H-P
 

Re: target/5985: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

[neroden]

Synopsis: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

State-Changed-From-To: feedback->closed
State-Changed-By: neroden
State-Changed-When: Thu Jan  2 10:33:04 2003
State-Changed-Why:
    Theoretically annoyance which appears to not be real; in feedback 9 months.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=5985

Re: target/5985: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

[rth]

Synopsis: fix_truncMN2 (e.g. fix_truncdfdi) and fixuns_truncMN2 incorrect on a29k, alpha, arm, c4x, dsp16xx, i370, i386, ia64, m88k, mips, ns32k, pj, romp, rs6000, s390, sh, xtensa

State-Changed-From-To: open->feedback
State-Changed-By: rth
State-Changed-When: Sat Mar 30 16:01:26 2002
State-Changed-Why:
    I don't think this is really true.  The documentation for
    these rtx codes only mentions FIX and UNSIGNED_FIX as being
    defined for fixed point modes.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=5985