NetBSD Problem Report #59189
From wiz@gatalith.at Mon Mar 17 14:06:37 2025
Return-Path: <wiz@gatalith.at>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256
client-signature RSA-PSS (2048 bits) client-digest SHA256)
(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
by mollari.NetBSD.org (Postfix) with ESMTPS id 445E21A9239
for <gnats-bugs@gnats.NetBSD.org>; Mon, 17 Mar 2025 14:06:37 +0000 (UTC)
Message-Id: <20250317140633.2665E3915ED@gatalith.at>
Date: Mon, 17 Mar 2025 15:06:33 +0100 (CET)
From: wiz@gatalith.at
Reply-To: wiz@gatalith.at
To: gnats-bugs@NetBSD.org
Subject: leak sanitizer broken
X-Send-Pr-Version: 3.95
>Number: 59189
>Category: toolchain
>Synopsis: leak sanitizer broken
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: toolchain-manager
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Mon Mar 17 14:10:00 +0000 2025
>Last-Modified: Mon Mar 24 07:25:01 +0000 2025
>Originator: Thomas Klausner
>Release: NetBSD 10.99.12
>Organization:
>Environment:
Architecture: x86_64
Machine: amd64
>Description:
The leak sanitizer is broken on -current.
On NetBSD 10, the program below (from riastradh) does:
# make lsanthread DBG=-g\ -O2\ -Wall\ -Werror\ -fsanitize=leak LDLIBS=-pthread
cc -g -O2 -Wall -Werror -fsanitize=leak -o lsanthread lsanthread.c -pthread
# ./lsanthread
=================================================================
==9052==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 2 byte(s) in 2 object(s) allocated from:
#0 0x75d53c005d27 in __interceptor_malloc /usr/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cc:53
#1 0x400abb in thread /home/wiz/lsanthread.c:14
#2 0x75d53bc0c89e in pthread__create_tramp /usr/src/lib/libpthread/pthread.c:595
SUMMARY: LeakSanitizer: 2 byte(s) leaked in 2 allocation(s).
while on NetBSD 10.99.12, it does
# make lsanthread DBG=-g\ -O2\ -Wall\ -Werror\ -fsanitize=leak LDLIBS=-pthread
cc -g -O2 -Wall -Werror -fsanitize=leak -o lsanthread lsanthread.c -pthread
ld: /usr/lib/libm.so.0: warning: warning: reference to compatibility cabs()
ld: /usr/lib/libm.so.0: warning: warning: reference to compatibility cabsf()
# ./lsanthread
zsh: segmentation fault (core dumped) ./lsanthread
Here's the program:
/*
* make lsanthread DBG=-g\ -O2\ -Wall\ -Werror\ -fsanitize=leak LDLIBS=-pthread
*/
#include <err.h>
#include <pthread.h>
#include <stdlib.h>
static void *
thread(void *cookie)
{
void *volatile p;
p = malloc(1);
p = NULL;
return p;
}
int
main(void)
{
pthread_t t[2];
void *result;
unsigned i;
int error;
for (i = 0; i < __arraycount(t); i++) {
error = pthread_create(&t[i], NULL, &thread, NULL);
if (error)
errc(1, error, "pthread_create");
}
for (i = 0; i < __arraycount(t); i++) {
error = pthread_join(t[i], &result);
if (error)
errc(1, error, "pthread_join");
}
return 0;
}
>How-To-Repeat:
See above.
>Fix:
Yes, please, and please add tests so this doesn't break again, unnoticed.
>Audit-Trail:
From: Thomas Klausner <wiz@NetBSD.org>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: toolchain/59189: leak sanitizer broken
Date: Tue, 18 Mar 2025 16:15:32 +0100
--Yv63Z8Q90D8MSmcU
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
The breakage is not pthread-related, it also causes a segfault with an
even simpler test case, attached, which on NetBSD 10 prints:
$ ./lsan
All is well.
=================================================================
==29831==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 1 byte(s) in 1 object(s) allocated from:
#0 0x73aab1405d27 in __interceptor_malloc /usr/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cc:53
#1 0x4009db in main /home/wiz/lsan.c:13
#2 0x4008fc in ___start /usr/src/lib/csu/common/crt0-common.c:350
#3 0x7f7f9900baf7 (/usr/libexec/ld.elf_so+0xbaf7)
SUMMARY: LeakSanitizer: 1 byte(s) leaked in 1 allocation(s).
Thomas
--Yv63Z8Q90D8MSmcU
Content-Type: text/x-c; charset=us-ascii
Content-Disposition: attachment; filename=lsan.c
/*
* make lsan DBG=-g\ -Wall\ -Werror\ -fsanitize=leak
*/
#include <stdio.h>
#include <stdlib.h>
int
main(void)
{
void *p;
p = malloc(1);
p = NULL;
if (p == NULL) {
printf("All is well.\n");
}
return 0;
}
--Yv63Z8Q90D8MSmcU--
From: Christos Zoulas <christos@zoulas.com>
To: gnats-bugs@netbsd.org
Cc: toolchain-manager@netbsd.org,
gnats-admin@netbsd.org,
netbsd-bugs@netbsd.org
Subject: Re: toolchain/59189: leak sanitizer broken
Date: Thu, 20 Mar 2025 09:46:40 -0400
--Apple-Mail=_6B2B9F26-2CF6-49AF-8883-96257DE7A8CE
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
charset=us-ascii
So it dies jumping to 0 when trying to call the real ataxit() from =
__interceptor_atexit() in the call ataxit(cleanup) from crt0.
(gdb) where
#0 0x0000000000000000 in ?? ()
#1 0x00007384ff3f44e2 in __interceptor_atexit (
f=3Df@entry=3D0x7f7ff73384cf <_rtld_exit>)
at =
/usr/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cpp:38=
1
#2 0x00000000004012e7 in ___start (cleanup=3D0x7f7ff73384cf =
<_rtld_exit>,=20
ps_strings=3D0x7f7fff70afe0) at =
/usr/src/lib/csu/common/crt0-common.c:321
#3 0x00007f7ff733ef58 in ?? () from /usr/libexec/ld.elf_so
#4 0x0000000000000001 in ?? ()
#5 0x00007f7fff70a410 in ?? ()
#6 0x0000000000000000 in ?? ()
(gdb) up
#1 0x00007384ff3f44e2 in __interceptor_atexit (
f=3Df@entry=3D0x7f7ff73384cf <_rtld_exit>)
at =
/usr/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cpp:38=
1
381 return REAL(__cxa_atexit)((void (*)(void *a))f, 0, 0);
Current language: auto
The current source language is "auto; currently c++".
(gdb) disassemble
Dump of assembler code for function __interceptor_atexit(void (*)()):
0x00007384ff3f44c4 <+0>: push %rbp
0x00007384ff3f44c5 <+1>: mov %rsp,%rbp
0x00007384ff3f44c8 <+4>: push %rbx
0x00007384ff3f44c9 <+5>: sub $0x8,%rsp
0x00007384ff3f44cd <+9>: mov %rdi,%rbx
0x00007384ff3f44d0 <+12>: call 0x7384ff3f9d53 =
<_ZN6__lsan19DisableInThisThreadEv>
0x00007384ff3f44d5 <+17>: xor %edx,%edx
0x00007384ff3f44d7 <+19>: xor %esi,%esi
0x00007384ff3f44d9 <+21>: mov %rbx,%rdi
0x00007384ff3f44dc <+24>: call *0x3bb3e(%rip) # =
0x7384ff430020 <_ZN14__interception17real___cxa_atexitE>
=3D> 0x00007384ff3f44e2 <+30>: mov %eax,%ebx
0x00007384ff3f44e4 <+32>: call 0x7384ff3f9d68 =
<_ZN6__lsan18EnableInThisThreadEv>
0x00007384ff3f44e9 <+37>: mov %ebx,%eax
0x00007384ff3f44eb <+39>: add $0x8,%rsp
0x00007384ff3f44ef <+43>: pop %rbx
0x00007384ff3f44f0 <+44>: pop %rbp
0x00007384ff3f44f1 <+45>: ret
End of assembler dump.
(gdb) info reg
rax 0xffffffffffffffe0 -32
rbx 0x7f7ff73384cf 140187584922831
rcx 0x4033b8 4207544
rdx 0x0 0
rsi 0x0 0
rdi 0x7f7ff73384cf 140187584922831
rbp 0x7f7fff709d50 0x7f7fff709d50
rsp 0x7f7fff709d40 0x7f7fff709d40
r8 0x10 16
r9 0x50654 329300
r10 0x7384ff984438 127015061046328
r11 0x246 582
r12 0x7f7fff70afe0 140187723149280
r13 0x403480 4207744
r14 0x7f7fff709d98 140187723144600
r15 0x104aa88 17083016
rip 0x7384ff3f44e2 0x7384ff3f44e2 =
<__interceptor_atexit(void (*)())+30>
eflags 0x10246 [ PF ZF IF RF ]
cs 0x47 71
ss 0x3f 63
ds 0x23 35
es 0x23 35
fs 0x0 0
gs 0x0 0
fs_base <unavailable>
gs_base <unavailable>
(gdb) print *(char **)(0x7384ff3f44e2+0x3bb3e)
$1 =3D 0x0
Something seems to be wrong with the relocation record of ataxit() but =
it is supposed to work
because the dynamic linking has finished. Adding -Wl,-z,now does not =
help either.
Trying to compile statically does not work at all because we are missing =
libsanitizer.spec:
[9:41am] 182>gcc -static -fsanitize=3Dleak -g -O3 lsan.c
gcc: fatal error: cannot read spec file 'libsanitizer.spec': No such =
file or directory
compilation terminated.
Providing libsanitizer.spec in /usr/lib containing:
[9:42am] 156#cat /usr/lib/libsanitizer.spec=20
# This spec file is read by gcc when linking. It is used to specify the
# standard libraries we need in order to link with various sanitizer =
libs.
*link_libasan: -lrt -lpthread -lm
*link_libhwasan: -lrt -lpthread -lm
*link_libtsan: -lrt -lpthread -lm
*link_libubsan: -lrt -lpthread -lm
*link_liblsan: -lrt -lpthread -lm
We get link errors:
[9:43am] 184>gcc -static -fsanitize=3Dleak -g -O3 lsan.c
ld: /usr/lib/libpthread.a(libpthread.a.o): in function =
`__libc_thr_create':
(.text+0x589e): multiple definition of `pthread_create'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x35b=
): first defined here
ld: /usr/lib/libpthread.a(libpthread.a.o): in function `pthread_detach':
(.text+0x6050): multiple definition of `pthread_detach'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x1b7=
): first defined here
ld: /usr/lib/libpthread.a(libpthread.a.o): in function `pthread_join':
(.text+0x5e56): multiple definition of `pthread_join'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x129=
): first defined here
ld: /usr/lib/libc.a(atexit.o): in function `__cxa_atexit':
atexit.c:(.text+0xe4): multiple definition of `__cxa_atexit'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x56)=
: first defined here
ld: /usr/lib/libc.a(atexit.o): in function `atexit':
atexit.c:(.text+0x290): multiple definition of `atexit'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x94)=
: first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `malloc':
jemalloc.c:(.text+0x921f): multiple definition of `malloc'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x1e0=
8): first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `posix_memalign':
jemalloc.c:(.text+0x92f1): multiple definition of `posix_memalign'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x7bb=
): first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `aligned_alloc':
jemalloc.c:(.text+0x997a): multiple definition of `aligned_alloc'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x5bb=
): first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `calloc':
jemalloc.c:(.text+0x9fa0): multiple definition of `calloc'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x19b=
3): first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `free':
jemalloc.c:(.text+0xabd0): multiple definition of `free'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x18d=
9): first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `valloc':
jemalloc.c:(.text+0xac86): multiple definition of `valloc'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x4c3=
): first defined here
ld: /usr/lib/libc.a(jemalloc.o): in function `realloc':
jemalloc.c:(.text+0xc53f): multiple definition of `realloc'; =
/usr/lib/liblsan.a(lsan_interceptors.o):lsan_interceptors.cpp:(.text+0x1b2=
d): first defined here
Exit 1
Should we fix them (by putting the methods in separate files? I don't =
think that it will work if we do.
Should we provide libsanitizer.spec?
christos
--Apple-Mail=_6B2B9F26-2CF6-49AF-8883-96257DE7A8CE
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename=signature.asc
Content-Type: application/pgp-signature;
name=signature.asc
Content-Description: Message signed with OpenPGP
-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org
iF0EARECAB0WIQS+BJlbqPkO0MDBdsRxESqxbLM7OgUCZ9wcQAAKCRBxESqxbLM7
OsEhAKCfXto9aWuS2vqltlQ5EY3y0QAICACgom/u7u4hhhimVL5sRg/4Rl5ELCo=
=bNdP
-----END PGP SIGNATURE-----
--Apple-Mail=_6B2B9F26-2CF6-49AF-8883-96257DE7A8CE--
From: RVP <rvp@SDF.ORG>
To: gnats-bugs@netbsd.org
Cc: Christos Zoulas <christos@zoulas.com>, toolchain-manager@netbsd.org
Subject: Re: toolchain/59189: leak sanitizer broken
Date: Sat, 22 Mar 2025 09:46:37 +0000 (UTC)
On Thu, 20 Mar 2025, Christos Zoulas wrote:
> So it dies jumping to 0 when trying to call the real ataxit() from __interceptor_atexit() in the call ataxit(cleanup) from crt0.
> [...]
> Something seems to be wrong with the relocation record of ataxit() but it is supposed to work
> because the dynamic linking has finished. Adding -Wl,-z,now does not help either.
>
Yeah. I think what's happening on NetBSD-HEAD is:
1. src/lib/csu/common/crt0-common.c calls atexit() before calling _preinit()--which
would've called libsanitizer/lsan/lsan_preinit.cpp:__lsan_init() to fill in
the address of the real atexit().
2. Because the LSAN atexit() isn't defined as a weak symbol, _that_ gets called
instead of the libc one--which then crashes when it calls NULL.
(I don't know why this works on NetBSD 10.x: something must be causing __lsan_init()
to be called there; or, it's fluke that it works...)
Anyway, I've come up with 2 patches. The first, just calls __lsan_init() in the
atexit() and __cxa_atexit() LSAN wrappers. (I tried using `ENSURE_LSAN_INITED'
similar to `ENSURE_ASAN_INITED', but that doesn't work for some reason...)
This is a bit of hack, but, seems to work.
```
diff -urN a/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cpp b/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cpp
--- a/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cpp 2023-07-31 01:44:55.000000000 +0000
+++ b/src/external/gpl3/gcc/dist/libsanitizer/lsan/lsan_interceptors.cpp 2025-03-22 08:04:16.230083864 +0000
@@ -367,6 +367,8 @@
#if SANITIZER_INTERCEPT___CXA_ATEXIT
INTERCEPTOR(int, __cxa_atexit, void (*func)(void *), void *arg,
void *dso_handle) {
+ if (!lsan_inited && !lsan_init_is_running)
+ __lsan_init();
__lsan::ScopedInterceptorDisabler disabler;
return REAL(__cxa_atexit)(func, arg, dso_handle);
}
@@ -377,6 +379,8 @@
#if SANITIZER_INTERCEPT_ATEXIT
INTERCEPTOR(int, atexit, void (*f)()) {
+ if (!lsan_inited && !lsan_init_is_running)
+ __lsan_init();
__lsan::ScopedInterceptorDisabler disabler;
return REAL(__cxa_atexit)((void (*)(void *a))f, 0, 0);
}
diff -urN a/src/external/gpl3/gcc/dist/libsanitizer/sanitizer_common/sanitizer_flags.inc b/src/external/gpl3/gcc/dist/libsanitizer/sanitizer_common/sanitizer_flags.inc
--- a/src/external/gpl3/gcc/dist/libsanitizer/sanitizer_common/sanitizer_flags.inc 2023-07-31 01:44:56.000000000 +0000
+++ b/src/external/gpl3/gcc/dist/libsanitizer/sanitizer_common/sanitizer_flags.inc 2025-03-22 08:13:00.083831970 +0000
@@ -28,7 +28,7 @@
"Path to external symbolizer. If empty, the tool will search $PATH for "
"the symbolizer.")
COMMON_FLAG(
- bool, allow_addr2line, false,
+ bool, allow_addr2line, true,
"If set, allows online symbolizer to run addr2line binary to symbolize "
"stack traces (addr2line will only be used if llvm-symbolizer binary is "
"unavailable.")
```
The second one is to do things the Linux way: call the `.preinit_array'
functions first, _then_ call atexit(); and provide a spec for GCC.
(Note: I haven't tested this spec...):
```
diff -urN a/src/lib/csu/common/crt0-common.c b/src/lib/csu/common/crt0-common.c
--- a/src/lib/csu/common/crt0-common.c 2024-01-19 19:22:17.000000000 +0000
+++ b/src/lib/csu/common/crt0-common.c 2025-03-22 07:51:24.617231789 +0000
@@ -317,9 +317,6 @@
__progname = empty_string;
}
- if (cleanup != NULL)
- atexit(cleanup);
-
_libc_init();
if (&rtld_DYNAMIC == NULL) {
@@ -333,6 +330,9 @@
_preinit();
+ if (cleanup != NULL)
+ atexit(cleanup);
+
#ifdef MCRT0
atexit(_mcleanup);
monstartup((u_long)&__eprol, (u_long)&__etext);
diff -urN a/src/external/gpl3/gcc/dist/gcc/config/netbsd.h b/src/external/gpl3/gcc/dist/gcc/config/netbsd.h
--- a/src/external/gpl3/gcc/dist/gcc/config/netbsd.h 2023-07-31 01:46:31.000000000 +0000
+++ b/src/external/gpl3/gcc/dist/gcc/config/netbsd.h 2025-03-22 08:18:31.573486823 +0000
@@ -204,8 +204,29 @@
netbsd_patch_builtins (); \
} while(0)
+/* Link -lasan early on the command line. For -static-libasan, don't link
+ it for -shared link, the executable should be compiled with -static-libasan
+ in that case, and for executable link with --{,no-}whole-archive around
+ it to force everything into the executable. And similarly for -ltsan,
+ -lhwasan, and -llsan. */
#if defined(HAVE_LD_STATIC_DYNAMIC)
+#undef LIBASAN_EARLY_SPEC
+#define LIBASAN_EARLY_SPEC "%{!shared:libasan_preinit%O%s} " \
+ "%{static-libasan:%{!shared:" \
+ LD_STATIC_OPTION " --whole-archive -lasan --no-whole-archive " \
+ LD_DYNAMIC_OPTION "}}%{!static-libasan:-lasan}"
+#undef LIBHWASAN_EARLY_SPEC
+#define LIBHWASAN_EARLY_SPEC "%{static-libhwasan:%{!shared:" \
+ LD_STATIC_OPTION " --whole-archive -lhwasan --no-whole-archive " \
+ LD_DYNAMIC_OPTION "}}%{!static-libhwasan:-lhwasan}"
#undef LIBTSAN_EARLY_SPEC
-#define LIBTSAN_EARLY_SPEC "%{static-libtsan:" LD_STATIC_OPTION "}" \
- " -ltsan %{static-libtsan:" LD_DYNAMIC_OPTION "} -lpthread"
+#define LIBTSAN_EARLY_SPEC "%{!shared:libtsan_preinit%O%s} " \
+ "%{static-libtsan:%{!shared:" \
+ LD_STATIC_OPTION " --whole-archive -ltsan --no-whole-archive " \
+ LD_DYNAMIC_OPTION "}}%{!static-libtsan:-ltsan}"
+#undef LIBLSAN_EARLY_SPEC
+#define LIBLSAN_EARLY_SPEC "%{!shared:liblsan_preinit%O%s} " \
+ "%{static-liblsan:%{!shared:" \
+ LD_STATIC_OPTION " --whole-archive -llsan --no-whole-archive " \
+ LD_DYNAMIC_OPTION "}}%{!static-liblsan:-llsan}"
#endif
```
The spec will rely on there now being files called libasan_preinit.o,
liblsan_preinit.o, etc., in the compiler-specific lib. directory--like on Linux:
```
$ find /usr/ -name lib*_preinit.o 2>/dev/null
/usr/lib/gcc/x86_64-linux-gnu/13/libhwasan_preinit.o
/usr/lib/gcc/x86_64-linux-gnu/13/libtsan_preinit.o
/usr/lib/gcc/x86_64-linux-gnu/13/libasan_preinit.o
/usr/lib/gcc/x86_64-linux-gnu/13/liblsan_preinit.o
$
```
(Some Makefile magic is surely needed for this... :))
HTH,
-RVP
PS. My test code follows:
```base64 mylsan.tar.bz2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```
From: Christos Zoulas <christos@zoulas.com>
To: RVP <rvp@sdf.org>
Cc: gnats-bugs@netbsd.org,
toolchain-manager@netbsd.org
Subject: Re: toolchain/59189: leak sanitizer broken
Date: Sat, 22 Mar 2025 12:20:34 -0400
--Apple-Mail=_01562FF4-1AC2-412C-8422-64E1435F114D
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
charset=us-ascii
Thanks for working on this. No more coredumps, but it does not detect =
leaks either. Static linking still does not work.
I tried the spec and it works. To install preinit files:
$ cat /usr/src/external/gpl3/gcc/lib/preinit.mk
# $NetBSD$
FILE=3Dlib${LIB}_preinit.o
FILES+=3D${FILE}
FILESDIR=3D/usr/lib
CLEANFILES+=3D${FILE} ${LIB}_preinit.o
${FILE}: ${LIB}_preinit.o
${_MKTARGET_CREATE}
cp ${.ALLSRC} ${.TARGET}
.include <bsd.files.mk>
$ cvs diff Makefile
[12:16pm] 18>cvs diff Makefile=20
Index: Makefile
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
RCS file: /cvsroot/src/external/gpl3/gcc/lib/liblsan/Makefile,v
retrieving revision 1.10
diff -u -p -r1.10 Makefile
--- Makefile 30 Sep 2024 00:32:12 -0000 1.10
+++ Makefile 22 Mar 2025 16:16:44 -0000
@@ -28,4 +28,6 @@ LIBDPLIBS+=3D pthread ${NETBSDSRCDIR}/lib/
=20
CWARNFLAGS.clang+=3D -Wno-error=3Dunused-private-field
=20
+.include "../preinit.mk"
+
.include <bsd.lib.mk>
christos
--Apple-Mail=_01562FF4-1AC2-412C-8422-64E1435F114D
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename=signature.asc
Content-Type: application/pgp-signature;
name=signature.asc
Content-Description: Message signed with OpenPGP
-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org
iF0EARECAB0WIQS+BJlbqPkO0MDBdsRxESqxbLM7OgUCZ97jUgAKCRBxESqxbLM7
OoMPAJ0S7dSJAxIiVdif6aKK1dkhbuYyzgCg4FY5B8kBrq+LVZ5ZQs54oiU0ckY=
=A6gd
-----END PGP SIGNATURE-----
--Apple-Mail=_01562FF4-1AC2-412C-8422-64E1435F114D--
From: Christos Zoulas <christos@zoulas.com>
To: RVP <rvp@sdf.org>
Cc: gnats-bugs@netbsd.org,
toolchain-manager@netbsd.org
Subject: Re: toolchain/59189: leak sanitizer broken
Date: Sat, 22 Mar 2025 12:31:43 -0400
--Apple-Mail=_43AFE604-3279-4242-925A-D1A533D37934
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
charset=us-ascii
Sorry, it does detect leaks. The optimizer took out the malloc code.
christos
--Apple-Mail=_43AFE604-3279-4242-925A-D1A533D37934
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename=signature.asc
Content-Type: application/pgp-signature;
name=signature.asc
Content-Description: Message signed with OpenPGP
-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org
iF0EARECAB0WIQS+BJlbqPkO0MDBdsRxESqxbLM7OgUCZ97l7wAKCRBxESqxbLM7
OkIXAJ4iCeVXetQlrC6+hd50xnb0lbSqwgCgyr8tP06pu4afqnj+zJf4f8kyvCM=
=Hyb0
-----END PGP SIGNATURE-----
--Apple-Mail=_43AFE604-3279-4242-925A-D1A533D37934--
From: "Christos Zoulas" <christos@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59189 CVS commit: src/lib/csu/common
Date: Sat, 22 Mar 2025 12:59:02 -0400
Module Name: src
Committed By: christos
Date: Sat Mar 22 16:59:02 UTC 2025
Modified Files:
src/lib/csu/common: crt0-common.c
Log Message:
PR/59189: RVP: move atexit(cleanup) after _preinit() to allow sanitizer
initialization during preinit as well as atexit interception.
To generate a diff of this commit:
cvs rdiff -u -r1.28 -r1.29 src/lib/csu/common/crt0-common.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: "Christos Zoulas" <christos@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/59189 CVS commit: src/external/gpl3/gcc
Date: Sat, 22 Mar 2025 13:05:48 -0400
Module Name: src
Committed By: christos
Date: Sat Mar 22 17:05:48 UTC 2025
Modified Files:
src/external/gpl3/gcc/dist/gcc/config: netbsd.h
src/external/gpl3/gcc/lib/libasan: Makefile
src/external/gpl3/gcc/lib/liblsan: Makefile
src/external/gpl3/gcc/lib/libtsan: Makefile
Added Files:
src/external/gpl3/gcc/lib: preinit.mk
Log Message:
PR/59189: RVP: Like Linux and FreeBSD use .preinit_array to initialize the
sanitizers.
To generate a diff of this commit:
cvs rdiff -u -r1.40 -r1.41 src/external/gpl3/gcc/dist/gcc/config/netbsd.h
cvs rdiff -u -r0 -r1.1 src/external/gpl3/gcc/lib/preinit.mk
cvs rdiff -u -r1.37 -r1.38 src/external/gpl3/gcc/lib/libasan/Makefile
cvs rdiff -u -r1.10 -r1.11 src/external/gpl3/gcc/lib/liblsan/Makefile
cvs rdiff -u -r1.17 -r1.18 src/external/gpl3/gcc/lib/libtsan/Makefile
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: Thomas Klausner <wiz@NetBSD.org>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: toolchain/59189: leak sanitizer broken
Date: Mon, 24 Mar 2025 08:22:44 +0100
The committed fixes make both test programs work for me, thank you!
Can someone(TM) please also add a test?
Thomas
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.47 2022/09/11 19:34:41 kim Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2025
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.
Home