* Another linux/alpha bug
@ 1997-08-30 5:43 H.J. Lu
1997-09-02 18:24 ` Jim Wilson
0 siblings, 1 reply; 3+ messages in thread
From: H.J. Lu @ 1997-08-30 5:43 UTC (permalink / raw)
To: egcs
Hi,
egcs 970828 for linux/alpha dumps core on this test, but gcc 2.7.2.1
is fine.
# gcc -S -mieee -O2 float.c
gcc: Internal compiler error: program cc1 got fatal signal 6
--
H.J. Lu (hjl@gnu.ai.mit.edu)
---float---
extern __const unsigned short int *__ctype_b;
int cost_of(const char * cp)
{
float number = 0;
for (cp += 2; *cp != '>'; cp++)
{
if (__ctype_b[(int) ( ( *cp ) )])
number = number * 10 + (*cp - '0');
else if (*cp == '.')
number += (*++cp - 10) / 10.0;
}
return (int) (number * 10);
}
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Another linux/alpha bug
1997-08-30 5:43 Another linux/alpha bug H.J. Lu
@ 1997-09-02 18:24 ` Jim Wilson
0 siblings, 0 replies; 3+ messages in thread
From: Jim Wilson @ 1997-09-02 18:24 UTC (permalink / raw)
To: H.J. Lu; +Cc: egcs
The problem is a missing pattern for the -mieee support. The following patch
fixes this.
Tue Sep 2 18:09:39 1997 Jim Wilson <wilson@cygnus.com>
* alpha.md (floatdisf2-1): New pattern.
Index: alpha.md
===================================================================
RCS file: /cvs/cvsfiles/egcs/gcc/config/alpha/alpha.md,v
retrieving revision 1.2
diff -p -r1.2 alpha.md
*** alpha.md 1997/08/15 05:29:33 1.2
--- alpha.md 1997/09/03 01:09:12
***************
*** 1517,1522 ****
--- 1517,1530 ----
"cvt%-qc %R1,%0"
[(set_attr "type" "fadd")])
+ (define_insn ""
+ [(set (match_operand:SF 0 "register_operand" "=&f")
+ (float:SF (match_operand:DI 1 "register_operand" "f")))]
+ "TARGET_FP && alpha_tp == ALPHA_TP_INSN"
+ "cvtq%,%+%& %1,%0"
+ [(set_attr "type" "fadd")
+ (set_attr "trap" "yes")])
+
(define_insn "floatdisf2"
[(set (match_operand:SF 0 "register_operand" "=f")
(float:SF (match_operand:DI 1 "register_operand" "f")))]
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Another linux/alpha bug
@ 1997-08-31 2:07 Richard Henderson
0 siblings, 0 replies; 3+ messages in thread
From: Richard Henderson @ 1997-08-31 2:07 UTC (permalink / raw)
To: egcs
> # gcc -S -mieee -O2 float.c
> gcc: Internal compiler error: program cc1 got fatal signal 6
Hum. I don't get that here. True, I'm not using 970828, but whatever
was in the cvs tree as of this morning, but I don't see any changes that
should have affected it.
r~
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~1997-09-02 18:24 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1997-08-30 5:43 Another linux/alpha bug H.J. Lu
1997-09-02 18:24 ` Jim Wilson
1997-08-31 2:07 Richard Henderson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).