NetBSD Problem Report #51460

From www@NetBSD.org  Sat Sep  3 16:12:56 2016
Return-Path: <www@NetBSD.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
	(Client CN "mail.netbsd.org", Issuer "Postmaster NetBSD.org" (verified OK))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 8C61B7A265
	for <gnats-bugs@gnats.NetBSD.org>; Sat,  3 Sep 2016 16:12:56 +0000 (UTC)
Message-Id: <20160903161255.251CD7A2B8@mollari.NetBSD.org>
Date: Sat,  3 Sep 2016 16:12:55 +0000 (UTC)
From: ray.phillips@gmail.com
Reply-To: ray.phillips@gmail.com
To: gnats-bugs@NetBSD.org
Subject: Source code bugs
X-Send-Pr-Version: www-1.0

>Number:         51460
>Category:       kern
>Synopsis:       Source code bugs
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Sep 03 16:15:00 +0000 2016
>Last-Modified:  Mon Sep 05 17:45:00 +0000 2016
>Originator:     Ray Phillips
>Release:        7.0.1
>Organization:
>Environment:
NetBSD aw135.sbs.uq.edu.au 7.0.1 NetBSD 7.0.1 (GENERIC.201605221355Z) amd64
>Description:
I hesitate to submit this report because I've no experience with kernel programming.  Hopefully it's not a bogus report...

I've been attempting to get Hubert Feyrer's g4u 2.6beta1:

http://www.feyrer.de/g4u/g4u-2.6beta1.tgz

to build with NetBSD 7.0.1 source code and found it was necessary to edit two files in the 7.0.1 source to get them to compile:

usr/srcsys/dev/acpi/fujhk_acpi.c
usr/src/sys/kern/core_elf32.c

The errors were (as shown in the "Fix to the problem if known" section of this send-pr):

/usr/src/sys/dev/acpi/fujhk_acpi.c:262:10: error: 'irb' may be used uninitialized in this function

/usr/src/sys/kern/core_elf32.c:513:12: error: conflicting types for 'coredump_elf32'
>How-To-Repeat:
It's probably not necessary, but if you like I can send you my modified g4u-2.6beta1.tgz if you want to follow the same steps I have, but you can see from the error messages why the compile stopped.

They're obviously not critical problems since NetBSD/i386 7.0.1 has been compiled successfully and released.  I suppose the g4u-build script must use different compiler settings to have produced these errors.
>Fix:
Edit  ./sys/dev/acpi/fujhk_acpi.c :

# diff -u ./sys/dev/acpi/fujhk_acpi.c.orig ./sys/dev/acpi/fujhk_acpi.c
--- ./sys/dev/acpi/fujhk_acpi.c.orig    2016-08-31 21:55:27.000000000 +1000
+++ ./sys/dev/acpi/fujhk_acpi.c 2016-08-31 22:13:25.000000000 +1000
@@ -246,6 +246,7 @@
        uint32_t irb;
        int i, index;

+       irb = 0; /* make sure irb has been initialized */
        for (i = 0; i < max_irb_buffer_size; i++) {
                if (fujitsu_hk_get_irb(sc, &irb) || irb == 0)
                        return;
#


Without that change this error occurred:

#   compile  G4U-KERNEL/fujhk_acpi.o
/usr/src/obj.i386/tooldir-NetBSD/bin/i486--netbsdelf-gcc -msoft-float -mno-mmx -
mno-sse -mno-avx -ffreestanding -fno-zero-initialized-in-bss -g -Os -fstack-prot
ector -Wstack-protector --param ssp-buffer-size=1 -fno-strict-aliasing -fno-comm
on -std=gnu99 -Werror -Wall -Wno-main -Wno-format-zero-length -Wpointer-arith -W
missing-prototypes -Wstrict-prototypes -Wold-style-definition -Wswitch -Wshadow
-Wcast-qual -Wwrite-strings -Wno-unreachable-code -Wno-pointer-sign -Wno-attribu
tes -Wextra -Wno-unused-parameter -Wold-style-definition -Wno-sign-compare -marc
h=i486 -mtune=pentiumpro --sysroot=/usr/src/obj.i386/destdir -Di386 -I. -I/usr/s
rc/sys/../common/include -I/usr/src/sys/arch -I/usr/src/sys -nostdinc -DCHILD_MA
X=1024 -DOPEN_MAX=1024 -DMSGBUFSIZE=50000 -DHALT_ON_PANIC -DMAXUSERS=2 -D_KERNEL
 -D_KERNEL_OPT -std=gnu99 -I/usr/src/sys/lib/libkern/../../../common/lib/libc/qu
ad -I/usr/src/sys/lib/libkern/../../../common/lib/libc/string -I/usr/src/sys/lib
/libkern/../../../common/lib/libc/arch/i386/string -D_FORTIFY_SOURCE=2 -I/usr/sr
c/sys/external/isc/atheros_hal/dist -I/usr/src/sys/external/isc/atheros_hal/ic -
I/usr/src/sys/external/bsd/acpica/dist/include -c /usr/src/sys/dev/acpi/fujhk_ac
pi.c
/usr/src/sys/dev/acpi/fujhk_acpi.c: In function 'fujitsu_hk_event_callback':
/usr/src/sys/dev/acpi/fujhk_acpi.c:262:10: error: 'irb' may be used uninitialized in this function [-Werror=maybe-uninitialized]
    index = FUJITSU_HK_IRB_HOTKEY_INDEX(irb);
          ^
cc1: all warnings being treated as errors

*** Failed target:  fujhk_acpi.o


Edit  ./sys/kern/core_elf32.c  :

# diff -u ./sys/kern/core_elf32.c.orig ./sys/kern/core_elf32.c
--- ./sys/kern/core_elf32.c.orig        2016-08-31 15:24:19.000000000 +1000
+++ ./sys/kern/core_elf32.c     2016-08-31 17:32:59.000000000 +1000
@@ -510,7 +510,7 @@
 #else  /* COREDUMP */

 int
-ELFNAMEEND(coredump)(struct lwp *l, void *cookie)
+ELFNAMEEND(coredump)(struct lwp *l, struct coredump_iostate *cookie)
 {

        return ENOSYS;
#


Without that change this error occurred:

#   compile  G4U-KERNEL/core_elf32.o
/usr/src/obj.i386/tooldir-NetBSD/bin/i486--netbsdelf-gcc -msoft-float -mno-mmx -mno-sse -mno-avx -ffreestanding -fno-zero-initialized-in-bss -g -Os -fstack-protector -Wstack-protector --param ssp-buffer-size=1 -fno-strict-aliasing -fno-common -std=gnu99 -Werror -Wall -Wno-main -Wno-format-zero-length -Wpointer-arith -Wmissing-prototypes -Wstrict-prototypes -Wold-style-definition -Wswitch -Wshadow -Wcast-qual -Wwrite-strings -Wno-unreachable-code -Wno-pointer-sign -Wno-attributes -Wextra -Wno-unused-parameter -Wold-style-definition -Wno-sign-compare -march=i486 -mtune=pentiumpro --sysroot=/usr/src/obj.i386/destdir -Di386 -I. -I/usr/src/sys/../common/include -I/usr/src/sys/arch -I/usr/src/sys -nostdinc -DCHILD_MAX=1024 -DOPEN_MAX=1024 -DMSGBUFSIZE=50000 -DHALT_ON_PANIC -DMAXUSERS=2 -D_KERNEL -D_KERNEL_OPT -std=gnu99 -I/usr/src/sys/lib/libkern/../../../common/lib/libc/quad -I/usr/src/sys/lib/libkern/../../../common/lib/libc/string -I/usr/src/sys/lib/libkern/../../../common/lib/libc/arch
 /i386/string -D_FORTIFY_SOURCE=2 -I/usr/src/sys/external/isc/atheros_hal/dist -I/usr/src/sys/external/isc/atheros_hal/ic -I/usr/src/sys/external/bsd/acpica/dist/include -c /usr/src/sys/kern/core_elf32.c
In file included from /usr/src/sys/kern/core_elf32.c:42:0:
/usr/src/sys/kern/core_elf32.c:513:12: error: conflicting types for 'coredump_elf32'
 ELFNAMEEND(coredump)(struct lwp *l, void *cookie)
            ^
/usr/src/sys/sys/cdefs.h:93:23: note: in definition of macro '__CONCAT'
 #define __CONCAT(x,y) x ## y
                       ^
/usr/src/sys/sys/exec_elf.h:1029:23: note: in expansion of macro 'CONCAT'
 #define ELFNAMEEND(x) CONCAT(x,CONCAT(_elf,ELFSIZE))
                       ^
/usr/src/sys/kern/core_elf32.c:513:1: note: in expansion of macro 'ELFNAMEEND'
 ELFNAMEEND(coredump)(struct lwp *l, void *cookie)
 ^
In file included from /usr/src/sys/kern/core_elf32.c:58:0:
/usr/src/sys/sys/exec_elf.h:1286:5: note: previous declaration of 'coredump_elf32' was here
 int coredump_elf32(struct lwp *, struct coredump_iostate *);
     ^

*** Failed target:  core_elf32.o

>Audit-Trail:
From: Martin Husemann <martin@duskware.de>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/51460: Source code bugs
Date: Sat, 3 Sep 2016 21:22:47 +0200

 On Sat, Sep 03, 2016 at 04:15:00PM +0000, ray.phillips@gmail.com wrote:
 > They're obviously not critical problems since NetBSD/i386 7.0.1 has
 > been compiled successfully and released.  I suppose the g4u-build
 > script must use different compiler settings to have produced these
 > errors.

 The -Os is likely causing this. This should be fixed (maybe we should
 add a ALL-Os kernel config to verify this permanently)

 Martin

From: David Holland <dholland-bugs@netbsd.org>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/51460: Source code bugs
Date: Sat, 3 Sep 2016 20:23:46 +0000

 On Sat, Sep 03, 2016 at 07:25:01PM +0000, Martin Husemann wrote:
  >  On Sat, Sep 03, 2016 at 04:15:00PM +0000, ray.phillips@gmail.com wrote:
  >  > They're obviously not critical problems since NetBSD/i386 7.0.1 has
  >  > been compiled successfully and released.  I suppose the g4u-build
  >  > script must use different compiler settings to have produced these
  >  > errors.
  >  
  >  The -Os is likely causing this. This should be fixed (maybe we should
  >  add a ALL-Os kernel config to verify this permanently)

 The uninitialized warning, certainly. The type error doesn't seem so
 likely...

 -- 
 David A. Holland
 dholland@netbsd.org

From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@NetBSD.org
Cc: kern-bug-people@netbsd.org, gnats-admin@netbsd.org,
    netbsd-bugs@netbsd.org
Subject: re: kern/51460: Source code bugs
Date: Mon, 05 Sep 2016 16:03:42 +1000

 > # diff -u ./sys/dev/acpi/fujhk_acpi.c.orig ./sys/dev/acpi/fujhk_acpi.c
 > --- ./sys/dev/acpi/fujhk_acpi.c.orig    2016-08-31 21:55:27.000000000 +1000
 > +++ ./sys/dev/acpi/fujhk_acpi.c 2016-08-31 22:13:25.000000000 +1000
 > @@ -246,6 +246,7 @@
 >         uint32_t irb;
 >         int i, index;
 > 
 > +       irb = 0; /* make sure irb has been initialized */
 >         for (i = 0; i < max_irb_buffer_size; i++) {
 >                 if (fujitsu_hk_get_irb(sc, &irb) || irb == 0)
 >                         return;
 > #

 this is a bug in GCC.  i've stopped patching our sources
 to build with -O3 / -Os because of these sorts of issues.

 eg, i haven't checked, but i fully expect that in the above
 code that fujitsu_hk_get_irb(), which is local, always
 sets irb if it returns true, but for some reason GCC thinks
 it is a problem.  i've seen this pattern dozens of times.

 i don't agree with Martin - i don't think we should "fix"
 these cases, but simply advise not using '-O[3s] -Werror'.


 .mrg.

From: "David A. Holland" <dholland@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/51460 CVS commit: src/sys/kern
Date: Mon, 5 Sep 2016 17:42:57 +0000

 Module Name:	src
 Committed By:	dholland
 Date:		Mon Sep  5 17:42:57 UTC 2016

 Modified Files:
 	src/sys/kern: core_elf32.c

 Log Message:
 Build fix for when COREDUMP is turned off, from Ray Phillips in PR 51460.


 To generate a diff of this commit:
 cvs rdiff -u -r1.47 -r1.48 src/sys/kern/core_elf32.c

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

Home	PR Database Search