optimization/6398: faulty code when inlining in a var-arged function

optimization/6398: faulty code when inlining in a var-arged function

[lucho]

>Number:         6398
>Category:       optimization
>Synopsis:       faulty code when inlining in a var-arged function
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Sun Apr 21 13:16:01 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     Luchezar Belev
>Release:        gcc-3.0
>Organization:
>Environment:
pentium III; linux-2.4.6; glibc-2.2 (does it matter?)
>Description:
This is some combination of:
 1) function inlining (-O3)
 2) no frame pointer (-fomit-frame-pointer)
 2) var-args
and it seems like this is somehow confusing GCC 3.0
so it generates incorrect code
(haven't tested other GCC versions)
>How-To-Repeat:
compile this source with -O3 and -fomit-frame-pointer
and see the comment

----begin----
#include <stdarg.h>
#include <stdio.h>

/*
	compile with these options: -O3 -fomit-frame-pointer
	to let the bug to appear
	
	The expected output of this code is "AAAAxyz\n" printed out on the console
	and it does so when compiled with -O2 (or lower) or without -fomit-frami-pointer

	One fact i've discovered that may be a hint:
	in the assembly output, near the beginning of f2() function
	one can notice that %ebp is used for addressing ("movl 4(%ebp), %ecx")
	before it is loaded with some meaningful value for this purpose
*/

static void f1(unsigned p, char *fmt, va_list args)
{
	char *h[2] = {"AAAA", "BBBB"};
	char buf[128];
	int l;
	
	if (p > 1) return;
	l = strlen(h[p]);
	memcpy(buf, h[p], l);
	vsnprintf(buf + l, 128 - l, fmt, args);
	puts(buf);
}

void f2(int p, char *fmt, ...)
{
	va_list args;
	va_start(args, fmt);
	f1(p, fmt, args);
	va_end(args);
}

int main()
{
	f2(0, "xyz\n");
	return 0;
}
----end----
>Fix:
don't know how to fix, but found a clue that may help: see the comment in the source
>Release-Note:
>Audit-Trail:
>Unformatted:

Re: optimization/6398: faulty code when inlining in a var-arged function

[reichelt]

Synopsis: faulty code when inlining in a var-arged function

State-Changed-From-To: feedback->closed
State-Changed-By: reichelt
State-Changed-When: Fri Dec  6 15:30:49 2002
State-Changed-Why:
    Seems to be fixed.
    
    I could reproduce the bug with gcc 3.0, but not with 3.0.4,
    3.1, 3.2.1 and 3.3-20021118.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6398

Re: optimization/6398: faulty code when inlining in a var-arged function

[lucho]

The following reply was made to PR optimization/6398; it has been noted by GNATS.

From: lucho <lucho@haemimont.bg>
To: jakub@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org,
	lucho@haemimont.bg, nobody@gcc.gnu.org, gcc-gnats@gcc.gnu.org
Cc:  
Subject: Re: optimization/6398: faulty code when inlining in a var-arged function
Date: Mon, 22 Apr 2002 18:45:29 +0300

 This is a multi-part message in MIME format.
 --------------010302030908040209010702
 Content-Type: text/plain; charset=us-ascii; format=flowed
 Content-Transfer-Encoding: 7bit
 
 here is the preprocessed source (as attachment)
 i haven't installed other gcc than 3.0 yet and so cannot test other versions
 
 --------------010302030908040209010702
 Content-Type: text/plain;
  name="bug.i"
 Content-Transfer-Encoding: 7bit
 Content-Disposition: inline;
  filename="bug.i"
 
 # 1 "bug.c"
 # 1 "/usr/lib/gcc-lib/i686-pc-linux-gnu/3.0/include/stdarg.h" 1 3
 # 43 "/usr/lib/gcc-lib/i686-pc-linux-gnu/3.0/include/stdarg.h" 3
 typedef __builtin_va_list __gnuc_va_list;
 # 110 "/usr/lib/gcc-lib/i686-pc-linux-gnu/3.0/include/stdarg.h" 3
 typedef __gnuc_va_list va_list;
 # 2 "bug.c" 2
 
 
 static void f1(int p, char *fmt, va_list args)
 {
         char *hdr[2] = {"AAAA", "BBBB"};
         char buf[128];
         int l;
 
         if (p > 2) return;
         l = strlen(hdr[p]);
         memcpy(buf, hdr[p], l);
         vsnprintf(buf + l, 128 - l, fmt, args);
         puts(buf);
 }
 
 void f2(int p, char *fmt, ...)
 {
         va_list args;
         __builtin_stdarg_start((args),fmt);
         f1(p, fmt, args);
         __builtin_va_end(args);
 }
 
 int main()
 {
         f2(0, "xyz");
         return 0;
 }
 
 --------------010302030908040209010702--
 

Re: optimization/6398: faulty code when inlining in a var-arged function

[jakub]

Synopsis: faulty code when inlining in a var-arged function

State-Changed-From-To: open->feedback
State-Changed-By: jakub
State-Changed-When: Sun Apr 21 14:45:03 2002
State-Changed-Why:
    Hard to say because the testcase is not preprocessed
    and thus it matters a lot what exact glibc headers it has.
    I couldn't reproduce this with 3.1 20020417 (prerelease),
    nor 3.0.2 20011002, nor 3.0.4 (glibc 2.2.5 installed).
    Can you please provide preprocessed source for the testcase?

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6398

Re: optimization/6398: faulty code when inlining in a var-arged function

[Andrew Pinski]

The following reply was made to PR optimization/6398; it has been noted by GNATS.

From: Andrew Pinski <pinskia@physics.uc.edu>
To: lucho@haemimont.bg
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: optimization/6398: faulty code when inlining in a var-arged function
Date: Sun, 21 Apr 2002 16:41:45 -0400

 This has been fixed in gcc version 3.2 20020421 (experimental).
 Could you test 3.1 prelease?
 
 Thanks,
 Andrew Pinski