From lev@ftp.translate.ru  Fri May 23 19:26:22 2008
Return-Path: <lev@ftp.translate.ru>
Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34])
	by hub.freebsd.org (Postfix) with ESMTP id 3C13E1065674
	for <FreeBSD-gnats-submit@freebsd.org>; Fri, 23 May 2008 19:26:22 +0000 (UTC)
	(envelope-from lev@ftp.translate.ru)
Received: from ftp.translate.ru (ftp.translate.ru [195.131.4.140])
	by mx1.freebsd.org (Postfix) with ESMTP id DD5FD8FC0A
	for <FreeBSD-gnats-submit@freebsd.org>; Fri, 23 May 2008 19:26:21 +0000 (UTC)
	(envelope-from lev@ftp.translate.ru)
Received: by ftp.translate.ru (Postfix, from userid 1000)
	id C4EC013DF58; Fri, 23 May 2008 23:26:20 +0400 (MSD)
Message-Id: <20080523192620.C4EC013DF58@ftp.translate.ru>
Date: Fri, 23 May 2008 23:26:20 +0400 (MSD)
From: Lev A. Serebryakov <lev@FreeBSD.org>
Reply-To: Lev A. Serebryakov <lev@FreeBSD.org>
To: FreeBSD-gnats-submit@freebsd.org
Cc:
Subject: FreeBSD 7-STABLE buildworld fails with internal gcc error (no HW problem!)
X-Send-Pr-Version: 3.113
X-GNATS-Notify:

>Number:         123940
>Category:       kern
>Synopsis:       [build] FreeBSD 7-STABLE buildworld fails with internal gcc error (no HW problem!)
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri May 23 19:30:01 UTC 2008
>Closed-Date:    Sat May 24 18:28:46 UTC 2008
>Last-Modified:  Sat May 24 18:28:46 UTC 2008
>Originator:     Lev A. Serebryakov
>Release:        FreeBSD 7.0-STABLE amd64
>Organization:
>Environment:
System: FreeBSD blob.home.serebryakov.spb.ru 7.0-STABLE FreeBSD 7.0-STABLE #0: Mon May 19 21:26:08 MSD 2008 lev@blob.home.serebryakov.spb.ru:/usr/obj/usr/src/sys/BLOB amd64
CPU: E4500
MEM: 6Gb
HDD: Samsung 320Gb as system.
Sources are cvsupped 23 May 2008.

>Description:

`make buildworld' failed about every third time with:

---------------------------------------------------------
### stage 3: cross tools
....
===> gnu/usr.bin/cc/cc_int (all)
....
cc -O2 -fno-strict-aliasing -pipe -DIN_GCC -DHAVE_CONFIG_H -DPREFIX=\"/usr/obj/usr/src/tmp/usr\" -I/usr/obj/usr/src/tmp/usr/src/gnu/usr.bin/cc/cc_int/../cc_tools -I/usr/src/gnu/usr.bin/cc/cc_int/../cc_tools -I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc -I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc/config -I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcclibs/include -I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcclibs/libcpp/include -I/usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcclibs/libdecnumber  -I/usr/obj/usr/src/tmp/legacy/usr/include -c ../cc_tools/insn-attrtab.c
/usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcc/config/i386/i386.md: In function 'insn_default_latency':
/usr/src/gnu/usr.bin/cc/cc_tools/../../../../contrib/gcc/config/i386/i386.md:202: internal compiler error: Segmentation fault: 11
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
*** Error code 1

Stop in /usr/src/gnu/usr.bin/cc/cc_int.
*** Error code 1

Stop in /usr/src/gnu/usr.bin/cc.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
---------------------------------------------------------

Hardware is Ok: it can build world 10 time in loop with 8 jobs (-j8) and MAKEOBJDIRPREFIX set to non-/usr filesystem.
When MAKEOBJDIRPREFIX is empty (is not set) every third build is failed with this error.
/usr filesystem is clean and HDD with it is in good condition.
Here are wnought space on /usr (about 250Gb) and /tmp (450Mb).

Re-run `buildworld' with -DNO_CELAN typically finish build after that.

>How-To-Repeat:

Build world several times with default (not set) MAKEOBJDIRPREFIX 

>Fix:
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->feedback 
State-Changed-By: vwe 
State-Changed-When: Fri May 23 21:20:27 UTC 2008 
State-Changed-Why:  

Lev, please show us your build options (from /etc/make.conf and - if any - 
makeoptions from your kernel config). Also please tell us about the cc 
version you're using. 
Please check, if removing all make options will work for you and check if 
you've got another c compiler installed.  
Also please check your filesystem (esp. /usr/obj). 
You may also want to make sure if your hardware is really fine (memory). 

http://www.freebsd.org/cgi/query-pr.cgi?pr=123940 

From: Kris Kennaway <kris@FreeBSD.org>
To: "Lev A. Serebryakov" <lev@FreeBSD.org>
Cc: FreeBSD-gnats-submit@FreeBSD.org
Subject: Re: bin/123940: FreeBSD 7-STABLE buildworld fails with internal gcc error (no HW problem!)
Date: Fri, 23 May 2008 22:47:59 +0000

 On Fri, May 23, 2008 at 11:26:20PM +0400, Lev A. Serebryakov wrote:
 
 > Hardware is Ok: it can build world 10 time in loop with 8 jobs (-j8) and MAKEOBJDIRPREFIX set to non-/usr filesystem.
 > When MAKEOBJDIRPREFIX is empty (is not set) every third build is failed with this error.
 
 Sorry, but this error is strong evidence that your hardware is *not* OK.
 
 Kris

From: Lev Serebryakov <lev@FreeBSD.org>
To: bug-followup@FreeBSD.org, lev@FreeBSD.org, Kris Kennaway <kris@FreeBSD.org>, 
	<vwe@FreeBSD.org>
Cc:  
Subject: Re: kern/123940: [build] FreeBSD 7-STABLE buildworld fails with internal gcc error (no HW problem!)
Date: Sat, 24 May 2008 09:38:47 +0400

 Hello, bug-followup.
 
 gcc version is system one:
 
 blob# gcc -v
 Using built-in specs.
 Target: amd64-undermydesk-freebsd
 Configured with: FreeBSD/amd64 system compiler
 Thread model: posix
 gcc version 4.2.1 20070719  [FreeBSD]
 blob#
 
 There are NO gcc-related ports installed.
 
 ============== /etc/make.conf
 # For local ports building
 WRKDIRPREFIX=/usr/local/build-area/ports
 WITHOUT_X=YES
 WITHOUT_X11=YES
 WITHOUT_GUI=YES
 
 # Global
 CPUTYPE?=core2
 
 # added by use.perl 2008-05-14 23:54:55
 PERL_VER=5.8.8
 PERL_VERSION=5.8.8
 ============== /etc/make.conf
 
 Here is no /etc/src.conf at all.
 
 Only `makeoptions' in kernel config is:
 
 =======================================
 makeoptions DEBUG=-g
 =======================================
 
 memory was tested with 24 hours run of memtest86+ (no errors)
 disk was tested with 5 runs of fullscan of MHDD + full scan of
 manufacturer utility, there is NO errors in S.M.A.R.T. attributes and
 NO old errors in S.M.A.R.T. error log (checked with smartmontools).
 
 filesystem is clean on boot, and shutdown/boot cycle doesn't show any
 problems with it (/usr/obj is same FS, as /usr)
 
 Temperature of CPU is contorlled with coretemp(4) and is about
 40(+/-2) DegC on both cores at `hardest' momemnts, temperature of HDD
 is controleld with smartd from smartmontools and never raised more than 35 DegC.
 
 Now I'm trying, does setting MAKEOBJDIRPREFIX to non-default value, but
 on same FS, help.
 
 -- 
 // Black Lion AKA Lev Serebryakov <lev@FreeBSD.org>
 

From: Kris Kennaway <kris@FreeBSD.org>
To: Lev Serebryakov <lev@FreeBSD.org>
Cc: bug-followup@FreeBSD.org, Kris Kennaway <kris@FreeBSD.org>,
	vwe@FreeBSD.org
Subject: Re: kern/123940: [build] FreeBSD 7-STABLE buildworld fails with internal gcc error (no HW problem!)
Date: Sat, 24 May 2008 08:26:28 +0000

 Hardware problems by their very nature can be subtle and hard to track
 down.  Sometimes it might take a combination of things to occur before
 the fault happens (e.g. to push power load over the level at which
 your power supply can maintain it), and testing individual components
 will not show the fault.
 
 > Now I'm trying, does setting MAKEOBJDIRPREFIX to non-default value, but
 > on same FS, help.
 
 MAKEOBJDIRPREFIX is used by many people, it does not cause problems of
 this nature.
 
 Kris
 
 --
 In God we Trust -- all others must submit an X.509 certificate.
     -- Charles Forsythe <forsythe@alum.mit.edu>

From: Lev Serebryakov <lev@FreeBSD.org>
To: bug-followup@FreeBSD.org, lev@FreeBSD.org
Cc:  
Subject: Re: kern/123940: [build] FreeBSD 7-STABLE buildworld fails with internal gcc error (no HW problem!)
Date: Sat, 24 May 2008 21:51:51 +0400

 Hello, bug-followup.
 
  Yep, it seems to be hardware problems. I've remove half a memory and
 could not reproduce it. 10 new passes of memtest86+ doesn't show
 problem with this memory, but it seems, that 6Gb is too much for this
 MoBo or somerthing like this.
 
 -- 
 // Black Lion AKA Lev Serebryakov <lev@FreeBSD.org>
 
State-Changed-From-To: feedback->closed 
State-Changed-By: lev 
State-Changed-When: Sat May 24 18:28:25 UTC 2008 
State-Changed-Why:  

Local HW failure. 


http://www.freebsd.org/cgi/query-pr.cgi?pr=123940 
>Unformatted:
