public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
* Re: c/5091: variables qualified volatile of type va_list will cause a segmentation fault
@ 2001-12-12 20:36 rodrigc
  0 siblings, 0 replies; 3+ messages in thread
From: rodrigc @ 2001-12-12 20:36 UTC (permalink / raw)
  To: nobody; +Cc: gcc-prs

The following reply was made to PR c/5091; it has been noted by GNATS.

From: rodrigc@gcc.gnu.org
To: gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org,
  maierkom@rcs.ei.tum.de, nobody@gcc.gnu.org
Cc:  
Subject: Re: c/5091: variables qualified volatile of type va_list will cause a segmentation fault
Date: 13 Dec 2001 04:28:14 -0000

 Synopsis: variables qualified volatile of type va_list will cause a segmentation fault
 
 State-Changed-From-To: open->analyzed
 State-Changed-By: rodrigc
 State-Changed-When: Wed Dec 12 20:28:14 2001
 State-Changed-Why:
     I can reproduce your segfault.
     I don't think you are doing a very safe thing by
     defining a volatile va_list.  va_list is usually a macro
     which expands to a compiler, platform, and architecture
     specific thing.  So messing around with va_list is not
     safe.
     
     If I compile your code with g++ instead of gcc, I get the
     following errors:
     
     varargtest.c: In function `void foo(char*, ...)':
     varargtest.c:10: could not convert `ap' to `void*&'
     <internal>:10: in passing argument 1 of `void __builtin_stdarg_start(void*&,
        ...)'
     varargtest.c:27: could not convert `ap' to `void*&'
     <internal>:27: in passing argument 1 of `void __builtin_va_end(void*&)
     
     If I were you, I would try to change my code generator
     or generated code.
     
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=5091&database=gcc


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: c/5091: variables qualified volatile of type va_list will cause a segmentation fault
@ 2001-12-12 20:28 rodrigc
  0 siblings, 0 replies; 3+ messages in thread
From: rodrigc @ 2001-12-12 20:28 UTC (permalink / raw)
  To: gcc-bugs, gcc-gnats, gcc-prs, maierkom, nobody

Synopsis: variables qualified volatile of type va_list will cause a segmentation fault

State-Changed-From-To: open->analyzed
State-Changed-By: rodrigc
State-Changed-When: Wed Dec 12 20:28:14 2001
State-Changed-Why:
    I can reproduce your segfault.
    I don't think you are doing a very safe thing by
    defining a volatile va_list.  va_list is usually a macro
    which expands to a compiler, platform, and architecture
    specific thing.  So messing around with va_list is not
    safe.
    
    If I compile your code with g++ instead of gcc, I get the
    following errors:
    
    varargtest.c: In function `void foo(char*, ...)':
    varargtest.c:10: could not convert `ap' to `void*&'
    <internal>:10: in passing argument 1 of `void __builtin_stdarg_start(void*&,
       ...)'
    varargtest.c:27: could not convert `ap' to `void*&'
    <internal>:27: in passing argument 1 of `void __builtin_va_end(void*&)
    
    If I were you, I would try to change my code generator
    or generated code.
    

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=5091&database=gcc


^ permalink raw reply	[flat|nested] 3+ messages in thread

* c/5091: variables qualified volatile of type va_list will cause a segmentation fault
@ 2001-12-12  0:26 maierkom
  0 siblings, 0 replies; 3+ messages in thread
From: maierkom @ 2001-12-12  0:26 UTC (permalink / raw)
  To: gcc-gnats


>Number:         5091
>Category:       c
>Synopsis:       variables qualified volatile of type va_list will cause a segmentation fault
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 12 00:26:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     Thomas Maier-Komor
>Release:        3.0.0, 3.0.1, 3.0.2
>Organization:
>Environment:
Linux 2.4
>Description:
if one qulifies a variable of type va_list (#include <stdarg.h>) with volatile the program will crash. The variable will own an incorrect address (not on the stack).

This happens with all recent gcc 3.x versions - but not with
the older 2.95.x series
>How-To-Repeat:
The attached file is just a copy of the man page example with volatile inserted. just: compile - run - bang...
>Fix:
Workaround: do not use volatile - but I can't as the code, which I would like to get compiled is automatically generated... So a fix in 3.0.3 would be very helpful. Thanks.
>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: text/x-csrc; name="varargtest.c"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="varargtest.c"
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^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2001-12-13  4:36 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-12-12 20:36 c/5091: variables qualified volatile of type va_list will cause a segmentation fault rodrigc
  -- strict thread matches above, loose matches on Subject: below --
2001-12-12 20:28 rodrigc
2001-12-12  0:26 maierkom

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).