valgrind-developers — Discussion list for developers of valgrind

Re: [Valgrind-developers] Valgrind-3.25.0.RC1 is available for testing

[Mark W. <ma...@kl...>]

Hi Florian,

On Tue, 2025-04-22 at 00:29 +0200, Florian Krohm wrote:
> On 18.04.25 15:53, Mark Wielaard wrote:
> > Slightly later than originally planned, but the RC1 is finally out!
> > 
> > An RC1 tarball for 3.25.0 is now available at
> > https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2
> > (md5sum = 2f02fe951278ebde62bba65c3a311a40)
> > (sha1sum = 3679ddc3237455f07de0ae30f21e947868c2218e)
> > https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2.asc
> > 
> > Please give it a try in configurations that are important for you and
> > report any problems you have
> 
> none/tests/bug290061.c
> bug290061.c:1:13: warning: ‘meh’ defined but not used [-Wunused-variable]
> 
> I guess we want a clean build.. Making 'meh' volatile removes the warning.

I am not seeing that locally. meh is used, although maybe the compiler
is smart enough to notice only meh[0] is actually used. What compiler
(version) and flags are you using?

> These files are missing:
> - auxprogs/ltp-tester.sh
> - auxprogs/ltp-excludes.txt
> - auxprogs/ltp-error-patterns.txt

Oops. Added them to EXTRA_DIST so they are in the tar ball for RC2.

> After adding them make ltpchecks took some 45 minutes on my z15 with about 46 
> failing tests. Not too bad..

I assume most of the failures are missing syscall wrappers? On my
x86_64 setup I got 66 FAILs, full log here:
https://bugsfiles.kde.org/attachment.cgi?id=180366

Cheers,

Mark

[Valgrind-developers] [valgrind] auxprogs/Makefile.am (EXTRA_DIST): Add ltpchecks helper files

[Mark W. <ma...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=f1a9c084d59bacac51535922f1f3c77b6fea08d0

commit f1a9c084d59bacac51535922f1f3c77b6fea08d0
Author: Mark Wielaard <ma...@kl...>
Date:   Tue Apr 22 18:24:31 2025 +0200

    auxprogs/Makefile.am (EXTRA_DIST): Add ltpchecks helper files
    
    ltp-tester.sh, ltp-excludes.txt and ltp-error-patterns.txt, breaking
    make ltpchecks.

Diff:
---
 auxprogs/Makefile.am | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/auxprogs/Makefile.am b/auxprogs/Makefile.am
index 00c743bc11..9cec4f222b 100644
--- a/auxprogs/Makefile.am
+++ b/auxprogs/Makefile.am
@@ -16,6 +16,9 @@ EXTRA_DIST = \
 	gsl-1.6.patch \
 	gsl-1.6.supp \
 	gsl-1.6.out.x86.exp \
+	ltp-tester.sh \
+	ltp-excludes.txt \
+	ltp-error-patterns.txt \
 	posixtestsuite-1.5.1-diff.txt \
 	ppcfround.c \
 	ppc64shifts.c \

Re: [Valgrind-developers] Valgrind-3.25.0.RC1 is available for testing

[Paul F. <pj...@wa...>]

On 4/22/25 00:29, Florian Krohm wrote:
> On 18.04.25 15:53, Mark Wielaard wrote:
>> Slightly later than originally planned, but the RC1 is finally out!
>>
>> An RC1 tarball for 3.25.0 is now available at
>> https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2
>> (md5sum = 2f02fe951278ebde62bba65c3a311a40)
>> (sha1sum = 3679ddc3237455f07de0ae30f21e947868c2218e)
>> https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2.asc
>>
>> Please give it a try in configurations that are important for you and
>> report any problems you have
>
>
> none/tests/bug290061.c
> bug290061.c:1:13: warning: ‘meh’ defined but not used [-Wunused-variable]
>
> I guess we want a clean build.. Making 'meh' volatile removes the 
> warning.
>
There was another problem as well with clang. Both should now be fixed.


Thanks

Paul

[Valgrind-developers] [valgrind] Regtest: clean up warning and compilation of bug290061.c

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=866f58b3b5edd4731f210b4f5f48f459f2c600a5

commit 866f58b3b5edd4731f210b4f5f48f459f2c600a5
Author: Paul Floyd <pj...@wa...>
Date:   Tue Apr 22 07:22:42 2025 +0200

    Regtest: clean up warning and compilation of bug290061.c
    
    On some platforms there was a 'defined but not used' warning.
    When I fixed that I got a link error from clang. Using a
    _LDFLAGS option causes automake to split building the test into
    separate compile and link commands and clang was optimizing away
    the unused static 'meh' symbol.

Diff:
---
 none/tests/Makefile.am | 1 +
 none/tests/bug290061.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/none/tests/Makefile.am b/none/tests/Makefile.am
index 8f9018f46a..043454828a 100644
--- a/none/tests/Makefile.am
+++ b/none/tests/Makefile.am
@@ -378,6 +378,7 @@ AM_CXXFLAGS += $(AM_FLAG_M3264_PRI)
 
 # Extra stuff for C tests
 ansi_CFLAGS		= $(AM_CFLAGS) -ansi
+bug290061_CFLAGS	= -fPIC
 bug290061_LDFLAGS	= @FLAG_PIE@
 bug491394_LDADD		= -lc
 bug491394_LDFLAGS	= -nostdlib -static
diff --git a/none/tests/bug290061.c b/none/tests/bug290061.c
index 40d847e8e1..8a77037adf 100644
--- a/none/tests/bug290061.c
+++ b/none/tests/bug290061.c
@@ -1,5 +1,6 @@
 static char meh[3000000]; // ~3mb
 int main(void)
 {
+    return meh[0];
 }

Re: [Valgrind-developers] Valgrind-3.25.0.RC1 is available for testing

[Florian K. <fl...@ei...>]

On 18.04.25 15:53, Mark Wielaard wrote:
> Slightly later than originally planned, but the RC1 is finally out!
> 
> An RC1 tarball for 3.25.0 is now available at
> https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2
> (md5sum = 2f02fe951278ebde62bba65c3a311a40)
> (sha1sum = 3679ddc3237455f07de0ae30f21e947868c2218e)
> https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2.asc
> 
> Please give it a try in configurations that are important for you and
> report any problems you have


none/tests/bug290061.c
bug290061.c:1:13: warning: ‘meh’ defined but not used [-Wunused-variable]

I guess we want a clean build.. Making 'meh' volatile removes the warning.


These files are missing:
- auxprogs/ltp-tester.sh
- auxprogs/ltp-excludes.txt
- auxprogs/ltp-error-patterns.txt

After adding them make ltpchecks took some 45 minutes on my z15 with about 46 
failing tests. Not too bad..

make auxchecks and gsl19test (both with -O2 and -O3) ran without errors.
I'll do some more testing on Tuesday.

   Florian

[Valgrind-developers] [valgrind] Illumos regtest: add an expected for none/tests/fdleak_socketpair_xml.stderr

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=078ba93d0d00c2a3cb62b99af2ba77a34c5212ac

commit 078ba93d0d00c2a3cb62b99af2ba77a34c5212ac
Author: Paul Floyd <pj...@wa...>
Date:   Mon Apr 21 20:44:31 2025 +0200

    Illumos regtest: add an expected for none/tests/fdleak_socketpair_xml.stderr
    
    illumos socketpair doesn't get the next two fds (3 and 4), instead it
    gets 4 and 5. That looks like it's because this is done in two steps in libc.
    so_socket gets called twice returnning fds 3 and 4 the so_socketpair takes
    those and does some rebinding(?) resulting in fds 4 and 5.

Diff:
---
 none/tests/Makefile.am                             |  1 +
 .../tests/fdleak_socketpair_xml.stderr.exp-illumos | 83 ++++++++++++++++++++++
 2 files changed, 84 insertions(+)

diff --git a/none/tests/Makefile.am b/none/tests/Makefile.am
index 37c94515d6..8f9018f46a 100644
--- a/none/tests/Makefile.am
+++ b/none/tests/Makefile.am
@@ -165,6 +165,7 @@ EXTRA_DIST = \
 	fdleak_pipe_xml.stderr.exp fdleak_pipe_xml.vgtest \
 	fdleak_socketpair.stderr.exp fdleak_socketpair.vgtest \
 	fdleak_socketpair_xml.stderr.exp fdleak_socketpair_xml.vgtest \
+		fdleak_socketpair_xml.stderr.exp-illumos \
 	floored.stderr.exp floored.stdout.exp floored.vgtest \
 	fork.stderr.exp fork.stdout.exp fork.vgtest \
 	fucomip.stderr.exp fucomip.vgtest \
diff --git a/none/tests/fdleak_socketpair_xml.stderr.exp-illumos b/none/tests/fdleak_socketpair_xml.stderr.exp-illumos
new file mode 100644
index 0000000000..f01a80440a
--- /dev/null
+++ b/none/tests/fdleak_socketpair_xml.stderr.exp-illumos
@@ -0,0 +1,83 @@
+<?xml version="1.0"?>
+
+<valgrindoutput>
+
+<protocolversion>5</protocolversion>
+<protocoltool>none</protocoltool>
+
+<preamble>
+  <line>Nulgrind, the minimal Valgrind tool</line>
+  <line>Copyright...</line>
+  <line>Using Valgrind...</line>
+  <line>Command: ./fdleak_socketpair</line>
+</preamble>
+
+<pid>...</pid>
+<ppid>...</ppid>
+<tool>none</tool>
+
+<args>
+  <vargv>
+    <exe>...</exe>
+    <arg>--command-line-only=yes</arg>
+    <arg>--memcheck:leak-check=no</arg>
+    <arg>--tool=none</arg>
+    <arg>--track-fds=yes</arg>
+    <arg>--xml=yes</arg>
+    <arg>--xml-fd=2</arg>
+  </vargv>
+  <argv>
+    <exe>...</exe>
+  </argv>
+</args>
+
+<status>
+  <state>RUNNING</state>
+  <time>...</time>
+</status>
+
+
+<status>
+  <state>FINISHED</state>
+  <time>...</time>
+</status>
+
+<error>
+  <unique>0x........</unique>
+  <tid>...</tid>
+  <kind>FdNotClosed</kind>
+  <fd>5</fd>
+  <what>...</what>
+  <stack>
+    <frame>
+      <ip>0x........</ip>
+      <obj>...</obj>
+      <fn>main</fn>
+      <dir>...</dir>
+      <file>fdleak_socketpair.c</file>
+      <line>13</line>
+    </frame>
+  </stack>
+</error>
+
+<error>
+  <unique>0x........</unique>
+  <tid>...</tid>
+  <kind>FdNotClosed</kind>
+  <fd>4</fd>
+  <what>...</what>
+  <stack>
+    <frame>
+      <ip>0x........</ip>
+      <obj>...</obj>
+      <fn>main</fn>
+      <dir>...</dir>
+      <file>fdleak_socketpair.c</file>
+      <line>13</line>
+    </frame>
+  </stack>
+</error>
+
+
+</valgrindoutput>
+

[Valgrind-developers] [valgrind] Illumos suppression and regtest

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=8464c3cf8fb4dd8138c60fd2f013fd376142153c

commit 8464c3cf8fb4dd8138c60fd2f013fd376142153c
Author: Paul Floyd <pj...@wa...>
Date:   Sun Apr 20 22:14:44 2025 +0200

    Illumos suppression and regtest
    
    Broaden the printf suppression
    Add an expected for helgrind/tests/bar_bad

Diff:
---
 helgrind/tests/Makefile.am                |  1 +
 helgrind/tests/bar_bad.stderr.exp-illumos | 84 +++++++++++++++++++++++++++++++
 solaris11.supp                            |  1 +
 3 files changed, 86 insertions(+)

diff --git a/helgrind/tests/Makefile.am b/helgrind/tests/Makefile.am
index bffd5bf76a..7adc5c6021 100755
--- a/helgrind/tests/Makefile.am
+++ b/helgrind/tests/Makefile.am
@@ -33,6 +33,7 @@ EXTRA_DIST = \
 	bar_bad.vgtest bar_bad.stdout.exp bar_bad.stderr.exp \
 		bar_bad.stderr.exp-destroy-hang \
 		bar_bad.stderr.exp-freebsd \
+		bar_bad.stderr.exp-illumos \
 	bar_trivial.vgtest bar_trivial.stdout.exp bar_trivial.stderr.exp \
 	free_is_write.vgtest free_is_write.stdout.exp \
 		free_is_write.stderr.exp \
diff --git a/helgrind/tests/bar_bad.stderr.exp-illumos b/helgrind/tests/bar_bad.stderr.exp-illumos
new file mode 100644
index 0000000000..86bd8fab77
--- /dev/null
+++ b/helgrind/tests/bar_bad.stderr.exp-illumos
@@ -0,0 +1,84 @@
+
+initialise a barrier with zero count
+---Thread-Announcement------------------------------------------
+
+Thread #x is the program's root thread
+
+----------------------------------------------------------------
+
+Thread #x: pthread_barrier_init: 'count' argument is zero
+   at 0x........: pthread_barrier_init (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:49)
+
+----------------------------------------------------------------
+
+Thread #x's call to pthread_barrier_init failed
+   with error code 22 (EINVAL: Invalid argument)
+   at 0x........: pthread_barrier_init (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:49)
+
+
+initialise a barrier twice
+----------------------------------------------------------------
+
+Thread #x: pthread_barrier_init: barrier is already initialised
+   at 0x........: pthread_barrier_init (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:55)
+
+
+initialise a barrier which has threads waiting on it
+----------------------------------------------------------------
+
+Thread #x: pthread_barrier_init: barrier is already initialised
+   at 0x........: pthread_barrier_init (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:72)
+
+----------------------------------------------------------------
+
+Thread #x: pthread_barrier_init: threads are waiting at barrier
+   at 0x........: pthread_barrier_init (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:72)
+
+
+destroy a barrier that has waiting threads
+----------------------------------------------------------------
+
+Thread #x: pthread_barrier_destroy: threads are waiting at barrier
+   at 0x........: pthread_barrier_destroy (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:90)
+
+----------------------------------------------------------------
+
+Thread #x: pthread_cond_destroy: destruction of condition variable being waited upon
+   at 0x........: pthread_cond_destroy_WRK (hg_intercepts.c:...)
+   by 0x........: pthread_cond_destroy@* (hg_intercepts.c:...)
+   ...
+   by 0x........: pthread_barrier_destroy (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:90)
+
+
+destroy a barrier that was never initialised
+----------------------------------------------------------------
+
+Thread #x: pthread_barrier_destroy: barrier was never initialised
+   at 0x........: pthread_barrier_destroy (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:107)
+
+----------------------------------------------------------------
+
+Thread #x: pthread_mutex_destroy with invalid argument
+   at 0x........: mutex_destroy_WRK (hg_intercepts.c:...)
+   by 0x........: pthread_mutex_destroy (hg_intercepts.c:...)
+   ...
+   by 0x........: pthread_barrier_destroy (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:107)
+
+----------------------------------------------------------------
+
+Thread #x: pthread_cond_destroy: destruction of unknown cond var
+   at 0x........: pthread_cond_destroy_WRK (hg_intercepts.c:...)
+   by 0x........: pthread_cond_destroy@* (hg_intercepts.c:...)
+   ...
+   by 0x........: pthread_barrier_destroy (hg_intercepts.c:...)
+   by 0x........: main (bar_bad.c:107)
+
diff --git a/solaris11.supp b/solaris11.supp
index c772df3955..5d51a6a3f7 100644
--- a/solaris11.supp
+++ b/solaris11.supp
@@ -66,6 +66,7 @@
 {
    Illumos:printf
    Helgrind:Race
+   ...
    fun:printf
 }
 # conflict between pthread_barrier_init and pthread_barrier_wait

[Valgrind-developers] [valgrind] FreeBSD regtest: add auxv_script to dist_noinst_SCRIPTS

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=5d8271e5773da3ea1756a06085271b58e0384106

commit 5d8271e5773da3ea1756a06085271b58e0384106
Author: Paul Floyd <pj...@wa...>
Date:   Sat Apr 19 12:01:26 2025 +0200

    FreeBSD regtest: add auxv_script to dist_noinst_SCRIPTS

Diff:
---
 none/tests/freebsd/Makefile.am | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/none/tests/freebsd/Makefile.am b/none/tests/freebsd/Makefile.am
index 15ea01cc9e..5d209c6d27 100644
--- a/none/tests/freebsd/Makefile.am
+++ b/none/tests/freebsd/Makefile.am
@@ -1,7 +1,7 @@
 
 include $(top_srcdir)/Makefile.tool-tests.am
 
-dist_noinst_SCRIPTS = filter_stderr test.sh filter_452274 filter_auxv filter_fdleak
+dist_noinst_SCRIPTS = filter_stderr test.sh filter_452274 filter_auxv filter_fdleak auxv_script
 
 EXTRA_DIST = \
 	auxv.vgtest \

[Valgrind-developers] [valgrind] Helgrind regtest: use --check-cond-signal-mutex=yes in tc20_verifywrap

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=860e1b488446b7b4c16bece1159321c796f4e40c

commit 860e1b488446b7b4c16bece1159321c796f4e40c
Author: Paul Floyd <pj...@wa...>
Date:   Fri Apr 18 22:14:54 2025 +0200

    Helgrind regtest: use --check-cond-signal-mutex=yes in tc20_verifywrap

Diff:
---
 helgrind/tests/tc20_verifywrap.vgtest | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/helgrind/tests/tc20_verifywrap.vgtest b/helgrind/tests/tc20_verifywrap.vgtest
index cde1b21751..5e41d76d5f 100644
--- a/helgrind/tests/tc20_verifywrap.vgtest
+++ b/helgrind/tests/tc20_verifywrap.vgtest
@@ -1,3 +1,3 @@
 prereq: test -e tc20_verifywrap && ! ../../tests/os_test freebsd
 prog: tc20_verifywrap
-vgopts: --read-var-info=yes --sigill-diagnostics=no
+vgopts: --read-var-info=yes --sigill-diagnostics=no --check-cond-signal-mutex=yes

[Valgrind-developers] [valgrind] Doc: add description of cond signal without mutex lock.

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=d497c30fc6398a4b3aee92231749ac4072c7de3b

commit d497c30fc6398a4b3aee92231749ac4072c7de3b
Author: Paul Floyd <pj...@wa...>
Date:   Fri Apr 18 20:37:12 2025 +0200

    Doc: add description of cond signal without mutex lock.

Diff:
---
 helgrind/docs/hg-manual.xml | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/helgrind/docs/hg-manual.xml b/helgrind/docs/hg-manual.xml
index 6ec8caa3f4..398fd73a70 100644
--- a/helgrind/docs/hg-manual.xml
+++ b/helgrind/docs/hg-manual.xml
@@ -141,6 +141,15 @@ are:</para>
 <para>Checks pertaining to the validity of mutexes are generally also
 performed for reader-writer locks.</para>
 
+<para>Signalling or broadcasting a condition variable when the
+associated mutex is unlocked is not strictly an error. The resulting
+thread scheduling may be unpredictable if the mutex is not held.
+The option  <computeroutput>--check-cond-signal-mutex=yes|no</computeroutput>
+turns on checking for this situation. This kind of error is
+categorised as 'dubious'. The check is not turned on by default
+because some standard C and C++ libraries use condition
+signals/broadcasts with the associated mutex unlocked.</para>
+
 <para>Various kinds of this-can't-possibly-happen events are also
 reported.  These usually indicate bugs in the system threading
 library.</para>

[Valgrind-developers] [valgrind] Add 3.25 highlights to NEWS for FreeBSD.

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=e2aa12c84e662cbe1966c61408323f1a4d589f6e

commit e2aa12c84e662cbe1966c61408323f1a4d589f6e
Author: Paul Floyd <pj...@wa...>
Date:   Fri Apr 18 19:32:05 2025 +0200

    Add 3.25 highlights to NEWS for FreeBSD.

Diff:
---
 NEWS | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/NEWS b/NEWS
index b6b822202d..e5be7f53a9 100644
--- a/NEWS
+++ b/NEWS
@@ -19,6 +19,12 @@ AMD64/macOS 10.13 and nanoMIPS/Linux.
 * Numerous bug fixes for Illumos, in particular fixed a Valgrind crash
   whenever a signal handler was called.
 
+* On FreeBSD, a change to the libc code that runs atexit handlers was
+  causing Helgrind to produce an extra error about exiting threads
+  still holding locks for. This applied to every multithreaded application.
+  The extra error is now filtered out. A syscall wrapper had been added
+  for getrlimitusage.
+
 * ==================== TOOL CHANGES ===================
 
 * The --track-fds=yes and --track-fds=all options now treat all

[Valgrind-developers] [valgrind] Bug 502871 - Make Helgrind "pthread_cond_{signal, broadcast}: dubious: associated lock is not held by

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=954eadb149d5165700fc3ed6d47e435d3c381531

commit 954eadb149d5165700fc3ed6d47e435d3c381531
Author: Paul Floyd <pj...@wa...>
Date:   Fri Apr 18 13:35:45 2025 +0200

    Bug 502871 - Make Helgrind "pthread_cond_{signal,broadcast}: dubious: associated lock is not held by any thread" optional

Diff:
---
 NEWS                                                  | 11 +++++++++++
 helgrind/hg_main.c                                    | 17 ++++++++++++-----
 helgrind/tests/bug392331.vgtest                       |  2 +-
 helgrind/tests/tc23_bogus_condwait.stderr.exp         |  3 ---
 helgrind/tests/tc23_bogus_condwait.stderr.exp-freebsd |  3 ---
 helgrind/tests/tc23_bogus_condwait.stderr.exp-mips32  |  3 ---
 helgrind/tests/tc23_bogus_condwait.vgtest             |  2 ++
 7 files changed, 26 insertions(+), 15 deletions(-)

diff --git a/NEWS b/NEWS
index afdbacd331..b6b822202d 100644
--- a/NEWS
+++ b/NEWS
@@ -26,6 +26,15 @@ AMD64/macOS 10.13 and nanoMIPS/Linux.
   And when the stdin/out/err descriptors are reassigned they are
   now treated as normal (non-inherited) file descriptors.
 
+* Helgrind:
+  There is a change to warnings about calls to pthread_cond_signal and
+  pthread_cond_broadcast when the associated mutex is unlocked. Previously
+  Helgrind would always warn about this. Now this error is controlled by
+  a command line option, --check-cond-signal-mutex=yes|no. The default is
+  no. This change has been made because some C and C++ standard libraries
+  use pthread_cond_signal/pthread_cond_broadcast in this way. Users are
+  obliged to use suppressions if they wish to avoid this noise.
+
 * ==================== FIXED BUGS ====================
 
 The following bugs have been fixed or resolved.  Note that "n-i-bz"
@@ -80,6 +89,8 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 502288  s390x: Memcheck false positives with NNPA last tensor dimension
 502324  s390x: Memcheck false positives with TMxx and TM/TMY
 502679  Use LTP for testing valgrind
+502871  Make Helgrind "pthread_cond_{signal,broadcast}: dubious: associated
+        lock is not held by any thread" optional
 
 To see details of a given bug, visit
   https://bugs.kde.org/show_bug.cgi?id=XXXXXX
diff --git a/helgrind/hg_main.c b/helgrind/hg_main.c
index 6a300b83bb..135c4d3b98 100644
--- a/helgrind/hg_main.c
+++ b/helgrind/hg_main.c
@@ -165,6 +165,7 @@ Bool HG_(clo_ignore_thread_creation) = True;
 #else
 Bool HG_(clo_ignore_thread_creation) = False;
 #endif /* VGO_solaris */
+Bool HG_(clo_check_cond_signal_mutex) = False;
 
 static
 ThreadId map_threads_maybe_reverse_lookup_SLOW ( Thread* thr ); /*fwds*/
@@ -2454,7 +2455,7 @@ static void evh__HG_PTHREAD_COND_SIGNAL_PRE ( ThreadId tid, void* cond )
             HG_(record_error_Misc)(thr,
                "pthread_cond_{signal,broadcast}: associated lock is a rwlock");
          }
-         if (lk->heldBy == NULL) {
+         if (HG_(clo_check_cond_signal_mutex) && lk->heldBy == NULL) {
             HG_(record_error_Dubious)(thr,
                "pthread_cond_{signal,broadcast}: dubious: "
                "associated lock is not held by any thread");
@@ -5840,7 +5841,8 @@ static Bool hg_process_cmd_line_option ( const HChar* arg )
                             HG_(clo_check_stack_refs)) {}
    else if VG_BOOL_CLO(arg, "--ignore-thread-creation",
                             HG_(clo_ignore_thread_creation)) {}
-
+   else if VG_BOOL_CLO(arg, "--check-cond-signal-mutex",
+                            HG_(clo_check_cond_signal_mutex)) {}
    else
       return VG_(replacement_malloc_process_cmd_line_option)(arg);
 
@@ -5859,14 +5861,19 @@ static void hg_print_usage ( void )
 "    --history-backtrace-size=<number>  record <number> callers for full\n"
 "        history level [8]\n"
 "    --delta-stacktrace=no|yes [yes on linux amd64/x86]\n"
-"        no : always compute a full history stacktrace from unwind info\n"
-"        yes : derive a stacktrace from the previous stacktrace\n"
+"        no: always compute a full history stacktrace from unwind info\n"
+"        yes: derive a stacktrace from the previous stacktrace\n"
 "          if there was no call/return or similar instruction\n"
 "    --conflict-cache-size=N   size of 'full' history cache [2000000]\n"
 "    --check-stack-refs=no|yes race-check reads and writes on the\n"
 "                              main stack and thread stacks? [yes]\n"
 "    --ignore-thread-creation=yes|no Ignore activities during thread\n"
-"                              creation [%s]\n",
+"                              creation [%s]\n"""
+"    --check-cond-signal-mutex=yes|no [no]\n"
+"        no: do not check that the associated mutex is locked for calls\n"
+"          to pthread_cond_{signal,broadcast}\n"
+"        yes: generate 'dubious' error messages if the associated mutex\n"
+"          is unlocked\n",
 HG_(clo_ignore_thread_creation) ? "yes" : "no"
    );
 }
diff --git a/helgrind/tests/bug392331.vgtest b/helgrind/tests/bug392331.vgtest
index 3889b15f77..f300c5cc13 100644
--- a/helgrind/tests/bug392331.vgtest
+++ b/helgrind/tests/bug392331.vgtest
@@ -1,4 +1,4 @@
 prereq: test -e bug392331
-vgopts: -q
+vgopts: -q --check-cond-signal-mutex=yes
 prog: bug392331
 stderr_filter: filter_bug392331
diff --git a/helgrind/tests/tc23_bogus_condwait.stderr.exp b/helgrind/tests/tc23_bogus_condwait.stderr.exp
index e1611a5d8f..84289f1684 100644
--- a/helgrind/tests/tc23_bogus_condwait.stderr.exp
+++ b/helgrind/tests/tc23_bogus_condwait.stderr.exp
@@ -1,4 +1,3 @@
-
 ---Thread-Announcement------------------------------------------
 
 Thread #x is the program's root thread
@@ -52,5 +51,3 @@ Thread #x: pthread_cond_{timed}wait: cond is associated with a different mutex
    by 0x........: pthread_cond_wait@* (hg_intercepts.c:...)
    by 0x........: main (tc23_bogus_condwait.c:85)
 
-
-ERROR SUMMARY: 7 errors from 7 contexts (suppressed: 0 from 0)
diff --git a/helgrind/tests/tc23_bogus_condwait.stderr.exp-freebsd b/helgrind/tests/tc23_bogus_condwait.stderr.exp-freebsd
index 1eee38a90e..562bf71ebe 100644
--- a/helgrind/tests/tc23_bogus_condwait.stderr.exp-freebsd
+++ b/helgrind/tests/tc23_bogus_condwait.stderr.exp-freebsd
@@ -1,4 +1,3 @@
-
 ---Thread-Announcement------------------------------------------
 
 Thread #x is the program's root thread
@@ -96,5 +95,3 @@ Thread #x: pthread_cond_{signal,broadcast}: dubious: associated lock is not held
    by 0x........: mythread_wrapper (hg_intercepts.c:...)
    ...
 
-
-ERROR SUMMARY: 11 errors from 11 contexts (suppressed: 0 from 0)
diff --git a/helgrind/tests/tc23_bogus_condwait.stderr.exp-mips32 b/helgrind/tests/tc23_bogus_condwait.stderr.exp-mips32
index 6ed9ef8239..8ce9de66c2 100644
--- a/helgrind/tests/tc23_bogus_condwait.stderr.exp-mips32
+++ b/helgrind/tests/tc23_bogus_condwait.stderr.exp-mips32
@@ -1,4 +1,3 @@
-
 ---Thread-Announcement------------------------------------------
 
 Thread #x is the program's root thread
@@ -60,5 +59,3 @@ Thread #x: pthread_cond_{timed}wait: cond is associated with a different mutex
    by 0x........: pthread_cond_wait@* (hg_intercepts.c:...)
    by 0x........: main (tc23_bogus_condwait.c:78)
 
-
-ERROR SUMMARY: 8 errors from 8 contexts (suppressed: 0 from 0)
diff --git a/helgrind/tests/tc23_bogus_condwait.vgtest b/helgrind/tests/tc23_bogus_condwait.vgtest
index 31c3ed8b9a..918cd86e4c 100644
--- a/helgrind/tests/tc23_bogus_condwait.vgtest
+++ b/helgrind/tests/tc23_bogus_condwait.vgtest
@@ -1 +1,3 @@
 prog: tc23_bogus_condwait
+vgopts: -q  --check-cond-signal-mutex=yes
+

Re: [Valgrind-developers] [PATCH] Linux: Inhibit tail calls in cancellable system calls

[Mark W. <ma...@kl...>]

Hi all,

On Mon, 2025-04-14 at 14:06 +0200, Mark Wielaard wrote:
> On Sun, Apr 06, 2025 at 03:23:54PM +0200, Mark Wielaard wrote:
> > On Mon, Mar 31, 2025 at 11:29:41AM +0200, Mark Wielaard wrote:
> > > On Fri, Mar 28, 2025 at 07:02:28PM +0100, Mark Wielaard wrote:
> > > > On Fri, 2025-03-21 at 14:01 +0100, Florian Weimer wrote:
> > > > > Without this change, the system call wrapper function is not visible
> > > > > on the stack at the time of the system call, which causes problems
> > > > > for interception tools such as valgrind.
> > > > > 
> > > > > Enhances commit 89b53077d2a58f00e7debdfe58afabe953dac60d ("nptl: Fix
> > > > > Race conditions in pthread cancellation [BZ#12683]").
> > > > > 
> > > > > Tested on i686-linux-gnu, powerpc64le-linux-gnu, x86_64-linux-gnu.
> > > > > (We're still discussing if valgrind needs this, but if it does, here's a
> > > > > patch.)
> > > > 
> > > > I implemented the valgrind part of skipping the syscall_cancel frames
> > > > here: https://bugs.kde.org/show_bug.cgi?id=502126#c2
> > > > And there is a valgrind package build for fedora rawhide:
> > > > https://koji.fedoraproject.org/koji/buildinfo?buildID=2687393
> > > > 
> > > > For ppc64le, s390x and x86_64 that patch seems enough.
> > > > 
> > > > For i686 and aarch64 there does seem to be an issue with missing the
> > > > glibc calling function because of a tail call.
> > > > 
> > > > Also on i686 there is another extra frame on top __libc_do_syscall.
> > > 
> > > I extended the patch to cover some extra sycall wrapper function
> > > symbols on i386 and armhf and pushed it to valgrind trunk and
> > > VALGRIND_3_24_BRANCH. There are builds for fedora rawhide and
> > > f42. This does seem to show that only on arm64 the tail calls
> > > obscure observing the full call stack.
> > 
> > This has now landed in fedora rawhide and f42. Test results look good,
> > except for some if the arm64 tests where the tail calls obscure
> > observing the full call stack. Please let me know if you need any more
> > input from us to get this fix in glibc.
> 
> Please let me know. Valgrind test results for syscall backtraces on
> anything except arm64 look good. We are working on valgrind 3.25.0
> now, to be released around April 24.

valgrind 3.25.0-RC1 has been released and test results look good on
most arches. arm64 does show the issue described above where the tail
calls obscure observing the full call stack when doing system calls.

Let me know what would be needed to get the above patch reviewed.

Thanks,

Mark

[Valgrind-developers] Valgrind-3.25.0.RC1 is available for testing

[Mark W. <ma...@kl...>]

Slightly later than originally planned, but the RC1 is finally out!

An RC1 tarball for 3.25.0 is now available at
https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2
(md5sum = 2f02fe951278ebde62bba65c3a311a40)
(sha1sum = 3679ddc3237455f07de0ae30f21e947868c2218e)
https://sourceware.org/pub/valgrind/valgrind-3.25.0.RC1.tar.bz2.asc

Please give it a try in configurations that are important for you and
report any problems you have, either on this mailing list, or
(preferably) via our bug tracker at
https://bugs.kde.org/enter_bug.cgi?product=valgrind

The NEWS file isn't complete up to date yet, but some highlights:

- Initial RISCV64/Linux support.
- Valgrind gdbserver supports 'x' packets.
- Numerous bug fixes for Illumos.
- --track-fds=yes now treats all inherited file descriptors like
  stdin/out/err (0,1,2) and there is a --modify-fds=high option.
- s390x support for various new instructions (BPP, BPRP and NIAI)
- Various new linux syscalls are supported (landlock*, open_tree,
  move_mount, fsopen, fsconfig, fsmount, fspick, userfaultfd)
- The Linux Test Project (ltp) is integrated in the testsuite
  try 'make ltpchecks' (this will take a while and will point out
  various missing syscalls and valgrind crashes!)

Since this RC1 is slightly later than planned and it is a long Easter
weekend for those that celebrate, lets do the RC2 on Wed Apr 25, with
the 3.25.0 final on Fri Apr 27.

[Valgrind-developers] [valgrind] Set version to 3.25.0-RC1

[Mark W. <ma...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=7279362391e1b53f7b7ac1a92544e92ada0247e9

commit 7279362391e1b53f7b7ac1a92544e92ada0247e9
Author: Mark Wielaard <ma...@kl...>
Date:   Fri Apr 18 14:22:31 2025 +0200

    Set version to 3.25.0-RC1
    
    Also update vg-lifespan which is used for documentation copyright years.

Diff:
---
 NEWS                        | 3 ++-
 configure.ac                | 4 ++--
 docs/xml/vg-entities.xml.in | 2 +-
 3 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/NEWS b/NEWS
index 89d450a8b4..afdbacd331 100644
--- a/NEWS
+++ b/NEWS
@@ -1,4 +1,4 @@
-Release 3.25.0 (?? Apr 2025)
+Release 3.25.0.RC1 (18 Apr 2025)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 This release supports X86/Linux, AMD64/Linux, ARM32/Linux, ARM64/Linux,
@@ -85,6 +85,7 @@ To see details of a given bug, visit
   https://bugs.kde.org/show_bug.cgi?id=XXXXXX
 where XXXXXX is the bug number as listed above.
 
+(3.25.0.RC1: 18 Apr 2025)
 
 Release 3.24.0 (31 Oct 2024)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
diff --git a/configure.ac b/configure.ac
index 862e6feed2..873d065dc4 100755
--- a/configure.ac
+++ b/configure.ac
@@ -18,8 +18,8 @@ AC_PREREQ(2.69)
 m4_define([v_major_ver], [3])
 m4_define([v_minor_ver], [25])
 m4_define([v_micro_ver], [0])
-m4_define([v_suffix_ver], [GIT])
-m4_define([v_rel_date], ["?? Apr 2025"])
+m4_define([v_suffix_ver], [RC1])
+m4_define([v_rel_date], ["18 Apr 2025"])
 m4_define([v_version],
 	  m4_if(v_suffix_ver, [],
 		[v_major_ver.v_minor_ver.v_micro_ver],
diff --git a/docs/xml/vg-entities.xml.in b/docs/xml/vg-entities.xml.in
index 2feaeec2f2..9cf75536db 100644
--- a/docs/xml/vg-entities.xml.in
+++ b/docs/xml/vg-entities.xml.in
@@ -2,7 +2,7 @@
 <!ENTITY vg-jemail     "ju...@va...">
 <!ENTITY vg-vemail     "val...@va...">
 <!ENTITY cl-email      "Jos...@gm...">
-<!ENTITY vg-lifespan   "2000-2024">
+<!ENTITY vg-lifespan   "2000-2025">
 
 <!-- valgrind release + version stuff -->
 <!ENTITY rel-type    "Release">

[Valgrind-developers] [valgrind] Update libiberty demangler

[Mark W. <ma...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=ab770bba01d5585748de8d790e7da36ed9475301

commit ab770bba01d5585748de8d790e7da36ed9475301
Author: Mark Wielaard <ma...@kl...>
Date:   Fri Apr 18 14:53:44 2025 +0200

    Update libiberty demangler
    
    Update the libiberty demangler using the auxprogs/update-demangler
    script to gcc git commit 94bea5dd6c9a06ddb6244be1e5196ff5fbe2b186.
    
    This update includes:
    
        * libiberty: Restore build with CP_DEMANGLE_DEBUG defined
        * non-gcc: Remove trailing whitespace
        * libiberty: Fix comment typos

Diff:
---
 auxprogs/update-demangler          | 4 ++--
 coregrind/m_demangle/cp-demangle.c | 4 ++--
 coregrind/m_demangle/demangle.h    | 6 +++---
 coregrind/m_demangle/dyn-string.h  | 2 +-
 coregrind/m_demangle/safe-ctype.h  | 1 -
 5 files changed, 8 insertions(+), 9 deletions(-)

diff --git a/auxprogs/update-demangler b/auxprogs/update-demangler
index 9b62fffca5..3968239759 100755
--- a/auxprogs/update-demangler
+++ b/auxprogs/update-demangler
@@ -17,8 +17,8 @@ set -e
 #---------------------------------------------------------------------
 
 # You need to modify these revision numbers for your update.
-old_gcc_revision=1719fa40c4ee4def60a2ce2f27e17f8168cf28ba # the revision of the previous update
-new_gcc_revision=ca2f7c84927f85b95f0f48f82b93f1460c372db4 # the revision for this update
+old_gcc_revision=ca2f7c84927f85b95f0f48f82b93f1460c372db4 # the revision of the previous update
+new_gcc_revision=94bea5dd6c9a06ddb6244be1e5196ff5fbe2b186 # the revision for this update
 
 # Unless the organization of demangler related files has changed, no
 # changes below this line should be necessary.
diff --git a/coregrind/m_demangle/cp-demangle.c b/coregrind/m_demangle/cp-demangle.c
index b1a8d87e59..a3d1b8810b 100644
--- a/coregrind/m_demangle/cp-demangle.c
+++ b/coregrind/m_demangle/cp-demangle.c
@@ -672,9 +672,9 @@ d_dump (struct demangle_component *dc, int indent)
       return;
     case DEMANGLE_COMPONENT_EXTENDED_BUILTIN_TYPE:
       {
-	char suffix[2] = { dc->u.s_extended_builtin.type->suffix, 0 };
+	char suffix[2] = { dc->u.s_extended_builtin.suffix, 0 };
 	printf ("builtin type %s%d%s\n", dc->u.s_extended_builtin.type->name,
-		dc->u.s_extended_builtin.type->arg, suffix);
+		dc->u.s_extended_builtin.arg, suffix);
       }
       return;
     case DEMANGLE_COMPONENT_OPERATOR:
diff --git a/coregrind/m_demangle/demangle.h b/coregrind/m_demangle/demangle.h
index a518153472..8b7dd705ac 100644
--- a/coregrind/m_demangle/demangle.h
+++ b/coregrind/m_demangle/demangle.h
@@ -66,14 +66,14 @@ extern "C" {
 /* Disable a limit on the depth of recursion in mangled strings.
    Note if this limit is disabled then stack exhaustion is possible when
    demangling pathologically complicated strings.  Bug reports about stack
-   exhaustion when the option is enabled will be rejected.  */  
-#define DMGL_NO_RECURSE_LIMIT (1 << 18)	
+   exhaustion when the option is enabled will be rejected.  */
+#define DMGL_NO_RECURSE_LIMIT (1 << 18)
 
 /* If DMGL_NO_RECURSE_LIMIT is not enabled, then this is the value used as
    the maximum depth of recursion allowed.  It should be enough for any
    real-world mangled name.  */
 #define DEMANGLE_RECURSION_LIMIT 2048
-  
+
 /* Enumeration of possible demangling styles.
 
    Lucid and ARM styles are still kept logically distinct, even though
diff --git a/coregrind/m_demangle/dyn-string.h b/coregrind/m_demangle/dyn-string.h
index 9052780c96..b0a108d9fa 100644
--- a/coregrind/m_demangle/dyn-string.h
+++ b/coregrind/m_demangle/dyn-string.h
@@ -3,7 +3,7 @@
    Contributed by Mark Mitchell (ma...@ma...).
 
 This file is part of GCC.
-   
+
 GCC is free software; you can redistribute it and/or modify
 it under the terms of the GNU General Public License as published by
 the Free Software Foundation; either version 2, or (at your option)
diff --git a/coregrind/m_demangle/safe-ctype.h b/coregrind/m_demangle/safe-ctype.h
index fa9c85af4b..d3e5214d58 100644
--- a/coregrind/m_demangle/safe-ctype.h
+++ b/coregrind/m_demangle/safe-ctype.h
@@ -122,7 +122,6 @@ extern const unsigned char  _sch_tolower[256];
 #if 0 /* in valgrind */
 #include <ctype.h>
 #endif /* ! in valgrind */
-
 #undef isalpha
 #define isalpha(c) do_not_use_isalpha_with_safe_ctype
 #undef isalnum

[Valgrind-developers] [valgrind] Add a platform note for Illumos to NEWS.

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=15198bde3544cd01edfe17580b136d72aa2595cc

commit 15198bde3544cd01edfe17580b136d72aa2595cc
Author: Paul Floyd <pj...@wa...>
Date:   Fri Apr 18 13:24:44 2025 +0200

    Add a platform note for Illumos to NEWS.

Diff:
---
 NEWS | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/NEWS b/NEWS
index e77c62456b..89d450a8b4 100644
--- a/NEWS
+++ b/NEWS
@@ -16,6 +16,9 @@ AMD64/macOS 10.13 and nanoMIPS/Linux.
 
 * ================== PLATFORM CHANGES =================
 
+* Numerous bug fixes for Illumos, in particular fixed a Valgrind crash
+  whenever a signal handler was called.
+
 * ==================== TOOL CHANGES ===================
 
 * The --track-fds=yes and --track-fds=all options now treat all

[Valgrind-developers] [valgrind] Regtest: gdbserver filter another auto-load warning

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=13f422b816ec4f6cf3ede836f782e21a67f01124

commit 13f422b816ec4f6cf3ede836f782e21a67f01124
Author: Paul Floyd <pj...@wa...>
Date:   Fri Apr 18 12:42:55 2025 +0200

    Regtest: gdbserver filter another auto-load warning
    
    Seen on arm64 with GDB 16

Diff:
---
 gdbserver_tests/filter_vgdb.in | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gdbserver_tests/filter_vgdb.in b/gdbserver_tests/filter_vgdb.in
index b32be68973..50d7a9ff6d 100755
--- a/gdbserver_tests/filter_vgdb.in
+++ b/gdbserver_tests/filter_vgdb.in
@@ -42,6 +42,7 @@ $SED -e '/warning: File .*valgrind-monitor.py.*declined.*/d' |
 # and remove loading error messages when python scripts haven't
 # been installed yet (running against vg-in-place)
 $SED -e '/warning: Missing auto-load script.*/d' |
+$SED -e '/warning: Unsupported auto-load script.*/d' |
 $SED -e '/of file .*vgpreload.*so.*/d' |
 $SED -e '/Use.*info auto-load python-scripts.*/d' |
 $SED -e '/warning: Invalid entry in .debug_gdb_scripts section/d' |

[Valgrind-developers] [valgrind] Add auxprogs/ltp-excludes.txt

[Mark W. <ma...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=ca5a8e303b8bef7baccfc28ff8b46b65ceb50708

commit ca5a8e303b8bef7baccfc28ff8b46b65ceb50708
Author: Mark Wielaard <ma...@kl...>
Date:   Fri Apr 18 12:22:29 2025 +0200

    Add auxprogs/ltp-excludes.txt
    
    There are a couple of ltp testcases that take a very long time to run
    (under valgrind). Add a file auxprogs/ltp-excludes.txt that is used to
    exclude them from a make ltpchecks run, containing 10 tests:
    
    bind06
    epoll-ltp
    inotify09
    msgstress01
    sendmsg03
    setsockopt06
    setsockopt07
    signal05
    signal06
    timerfd_settime02
    
    Excluding these 10 tests brings the execution time of make ltpchecks
    down to ~45 minutes.

Diff:
---
 auxprogs/ltp-excludes.txt | 10 ++++++++++
 auxprogs/ltp-tester.sh    |  3 ++-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/auxprogs/ltp-excludes.txt b/auxprogs/ltp-excludes.txt
new file mode 100644
index 0000000000..5b1d898c80
--- /dev/null
+++ b/auxprogs/ltp-excludes.txt
@@ -0,0 +1,10 @@
+bind06
+epoll-ltp
+inotify09
+msgstress01
+sendmsg03
+setsockopt06
+setsockopt07
+signal05
+signal06
+timerfd_settime02
diff --git a/auxprogs/ltp-tester.sh b/auxprogs/ltp-tester.sh
index a3c2157c56..000cfaa7f3 100755
--- a/auxprogs/ltp-tester.sh
+++ b/auxprogs/ltp-tester.sh
@@ -26,7 +26,8 @@ myLog ()
 
 cd $LTP_SRC_DIR
 
-mapfile -t files < <(find testcases/kernel/syscalls -executable -and -type f | sort)
+mapfile -t files < <(find testcases/kernel/syscalls -executable -and -type f \
+                     | sort | grep -v -f $ORIG_PWD/ltp-excludes.txt)
 c=${#files[@]}; i=0
 
 for test in "${files[@]}"; do

[Valgrind-developers] [valgrind] Use LTP for testing valgrind

[Mark W. <ma...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=944a8c864c5e1df6388bf4b47900eb9c001ceb73

commit 944a8c864c5e1df6388bf4b47900eb9c001ceb73
Author: Martin Cermak <mc...@re...>
Date:   Thu Apr 17 16:14:19 2025 +0200

    Use LTP for testing valgrind
    
    Add a new top level make target ltpchecks which will fetch the latest
    linux test project (ltp) release as defined by the LTP_VERSION and
    LTP_SHA256 variables in auxprogs/Makefile.am (update those when a new
    version of ltp is released). If the ltp tar.xz has already been
    downloaded, or it has already been unpacked and build, the (cached)
    file and build will be reused.
    
    The actual testing is done through the auxprogs/ltp-tester.sh script.
    It takes all executable tests from the ltp testcases under
    kernel/syscalls and runs them 3 times. Once directly, not under
    valgrind, once with -q --tool=none and once with -q
    --tool=memcheck. It then checks that valgrind didn't produce any
    messages with the none tool, that there were no fatal errors produced
    (as defined in auxprogs/ltp-error-patterns.txt) and that the ltp
    results are the same with and without valgrind.
    
    Currently there are 1472 test binaries and running them all (serially)
    takes more than three hours and detects various missing or incomplete
    syscall handlers in valgrind, plus various crashers.
    
    https://bugs.kde.org/show_bug.cgi?id=502679

Diff:
---
 Makefile.am                     |  3 ++
 NEWS                            |  1 +
 auxprogs/Makefile.am            | 35 +++++++++++++++++-
 auxprogs/ltp-error-patterns.txt |  6 ++++
 auxprogs/ltp-tester.sh          | 78 +++++++++++++++++++++++++++++++++++++++++
 5 files changed, 122 insertions(+), 1 deletion(-)

diff --git a/Makefile.am b/Makefile.am
index 14309dacfa..e67356b5a0 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -103,6 +103,9 @@ perf: check
 auxchecks: all
 	$(MAKE) -C auxprogs auxchecks
 
+ltpchecks: all
+	$(MAKE) -C auxprogs ltpchecks
+
 # Nb: no need to include any Makefile.am files here, or files included from
 # them, as automake includes them automatically.  Also not COPYING, README
 # or NEWS.
diff --git a/NEWS b/NEWS
index fac28e7984..e77c62456b 100644
--- a/NEWS
+++ b/NEWS
@@ -76,6 +76,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 502126  glibc 2.41 extra syscall_cancel frames
 502288  s390x: Memcheck false positives with NNPA last tensor dimension
 502324  s390x: Memcheck false positives with TMxx and TM/TMY
+502679  Use LTP for testing valgrind
 
 To see details of a given bug, visit
   https://bugs.kde.org/show_bug.cgi?id=XXXXXX
diff --git a/auxprogs/Makefile.am b/auxprogs/Makefile.am
index 3a9709da6f..00c743bc11 100644
--- a/auxprogs/Makefile.am
+++ b/auxprogs/Makefile.am
@@ -124,6 +124,9 @@ endif
 auxchecks: gsl-check
 auxclean: gsl-clean
 
+ltpchecks: ltp-check
+ltpclean: ltp-clean
+
 # You can override AUX_CHECK_DIR to use a previous download/build.
 # Must be an absolute path.
 # e.g. make auxchecks AUX_CHECK_DIR=$HOME/valgrind-auxtests
@@ -148,6 +151,14 @@ GSL_CFLAGS+=-mfpmath=sse -msse2
 endif
 endif
 
+# Linux Test Project
+LTP_VERSION=20250130
+LTP_SHA256_SUM=02e4ec326be54c3fd92968229a468c02c665d168a8a673edc38a891f7395ae10
+LTP_TAR_NAME=ltp-full-$(LTP_VERSION).tar.xz
+LTP_URL=https://github.com/linux-test-project/ltp/releases/download/$(LTP_VERSION)/$(LTP_TAR_NAME)
+LTP_TAR=$(AUX_CHECK_DIR)/$(LTP_TAR_NAME)
+LTP_SRC_DIR=$(AUX_CHECK_DIR)/ltp-full-$(LTP_VERSION)
+
 # Trick to get a literal space to use in substitutions
 sp := $(subst ,, )
 
@@ -166,6 +177,10 @@ $(GSL_TAR):
 	mkdir -p $(AUX_CHECK_DIR)
 	wget -q -O $(GSL_TAR) $(GSL_URL)
 
+$(LTP_TAR):
+	mkdir -p $(AUX_CHECK_DIR)
+	wget -q -O $(LTP_TAR) $(LTP_URL)
+
 # We need to autoreconf to make sure to get config.guess, config.sub
 # and libtool for newer architectures.
 $(GSL_SRC_DIR)/gsl-patched: $(GSL_TAR)
@@ -177,6 +192,15 @@ $(GSL_SRC_DIR)/gsl-patched: $(GSL_TAR)
 	 autoreconf -f -i -Wnone)
 	touch $@
 
+$(LTP_SRC_DIR): $(LTP_TAR)
+	echo "$(LTP_SHA256_SUM)  $(LTP_TAR)" | @SHA256SUM@ --check -
+	(cd $(AUX_CHECK_DIR) && \
+	 tar Jxf $(LTP_TAR_NAME) && \
+	 cd $(LTP_SRC_DIR) && \
+	 ./configure CC="${CC}" CXX="${CXX}" CFLAGS="$(LTP_CFLAGS)" && \
+	 ${MAKE} -j $(nproc) -C testcases/kernel/syscalls)
+	touch $@
+
 # We need make check -k because
 # some tests might fail even native (only on i386 though).
 # make check doesn't work reliably with -j.
@@ -204,17 +228,26 @@ gsl-check: $(GSL_BUILD_DIR)/gsl-build
 	diff -u $(abs_top_srcdir)/auxprogs/gsl-1.6.out.x86.exp \
 		$(GSL_BUILD_DIR)/valgrind-gsl.out
 
+ltp-check: $(LTP_SRC_DIR)
+	LTP_SRC_DIR=$(LTP_SRC_DIR) \
+	VALGRIND=$(abs_top_builddir)/vg-in-place \
+	  $(abs_top_srcdir)/auxprogs/ltp-tester.sh
+
+
 # We keep the tarball but remove the unpacked sources and build
 gsl-clean:
 	rm -rf $(GSL_SRC_NAME) $(GSL_BUILD_DIR)
 
+ltp-clean:
+	rm -rf $(LTP_SRC_DIR)
+
 #----------------------------------------------------------------------------
 # General stuff
 #----------------------------------------------------------------------------
 
 all-local: inplace-noinst_PROGRAMS inplace-noinst_DSYMS
 
-clean-local: clean-noinst_DSYMS auxclean
+clean-local: clean-noinst_DSYMS auxclean ltpclean
 
 install-exec-local: install-noinst_PROGRAMS install-noinst_DSYMS
 
diff --git a/auxprogs/ltp-error-patterns.txt b/auxprogs/ltp-error-patterns.txt
new file mode 100644
index 0000000000..ee5048bf42
--- /dev/null
+++ b/auxprogs/ltp-error-patterns.txt
@@ -0,0 +1,6 @@
+VALGRIND INTERNAL ERROR
+^valgrind:
+WARNING: unhandled
+unhandled
+Valgrind does not support
+Valgrind abort
diff --git a/auxprogs/ltp-tester.sh b/auxprogs/ltp-tester.sh
new file mode 100755
index 0000000000..a3c2157c56
--- /dev/null
+++ b/auxprogs/ltp-tester.sh
@@ -0,0 +1,78 @@
+#!/bin/bash
+
+set -e
+
+if [ -z "${LTP_SRC_DIR:-}" ]; then
+    echo "ERROR: LTP_SRC_DIR needs to be set.  Dying, bye bye ..."
+    exit 1
+fi
+
+ORIG_PATH=$PATH
+ORIG_PWD=$PWD
+LOGDIR=${LOGDIR:-$LTP_SRC_DIR/valgrind-ltp-logs}
+SUMMARY_LOG="$LOGDIR/summary.log"
+DIFFCMD="diff -u"
+VALGRIND="${VALGRIND:-$LTP_SRC_DIR/../../../vg-in-place}"
+
+# Initialize LOGDIR
+mkdir -p $LOGDIR; rm -rf $LOGDIR/*
+
+myLog ()
+{
+    msg="$1"
+    echo "$msg"
+    echo -e "FAIL: $msg" >> $SUMMARY_LOG
+}
+
+cd $LTP_SRC_DIR
+
+mapfile -t files < <(find testcases/kernel/syscalls -executable -and -type f | sort)
+c=${#files[@]}; i=0
+
+for test in "${files[@]}"; do
+    dir=$(dirname $test)
+    exe=$(basename $test)
+    l="$LOGDIR/$exe"
+    mkdir -p $l
+    i=$((++i))
+    pushd $dir >/dev/null
+        echo "[$i/$c] Testing $exe ..." | tee -a $SUMMARY_LOG
+        PATH="$ORIG_PATH:$PWD"
+        ./$exe >$l/log1std 2>$l/log1err ||:
+        $VALGRIND -q --tool=none --log-file=$l/log2 ./$exe >$l/log2std 2>$l/log2err ||:
+        $VALGRIND -q --tool=memcheck --log-file=$l/log3 ./$exe >$l/log3std 2>$l/log3err ||:
+
+        # We want to make sure that LTP syscall tests give identical
+        # results with and without valgrind.  The test logs go to the
+        # stderr.  They aren't identical across individual runs. The
+        # differences include port numbers, temporary files, test
+        # output ordering changes and more. They aren't trivially
+        # comparable.  We resort to comparing at least the final
+        # summary of individual test results
+        tail -10 $l/log1err | grep -E "^(passed|failed|broken|skipped|warnings):" > $l/log1summary ||:
+        tail -10 $l/log2err | grep -E "^(passed|failed|broken|skipped|warnings):" > $l/log2summary ||:
+        tail -10 $l/log3err | grep -E "^(passed|failed|broken|skipped|warnings):" > $l/log3summary ||:
+
+        # Check logs, report errors
+        pushd $l >/dev/null
+            if test -s log2; then
+                myLog "${exe}: unempty log2:\n$(cat log2)"
+            fi
+
+            if grep -f $ORIG_PWD/ltp-error-patterns.txt * > error-patterns-found.txt; then
+                myLog "${exe}: error string found:\n$(cat error-patterns-found.txt)"
+            fi
+
+            if ! ${DIFFCMD} log1summary log2summary >/dev/null; then
+                myLog "${exe}: ${DIFFCMD} log1summary log2summary:\n$(${DIFFCMD} log1summary log2summary)"
+            fi
+
+            if ! ${DIFFCMD} log2summary log3summary >/dev/null; then
+                myLog "${exe}: ${DIFFCMD} log2summary log3summary:\n$(${DIFFCMD} log2summary log3summary)"
+            fi
+        popd >/dev/null
+    popd >/dev/null
+done
+
+echo "TESTING FINISHED, logs in $LOGDIR"
+

[Valgrind-developers] [valgrind] FreeBSD: increase coverage of --modify-mds syscalls

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=b921a5fb8706d21efe9723b27ada290f1bba946e

commit b921a5fb8706d21efe9723b27ada290f1bba946e
Author: Paul Floyd <pj...@wa...>
Date:   Thu Apr 17 21:54:43 2025 +0200

    FreeBSD: increase coverage of --modify-mds syscalls
    
    As illumos, still missing a couple.

Diff:
---
 coregrind/m_syswrap/syswrap-freebsd.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/coregrind/m_syswrap/syswrap-freebsd.c b/coregrind/m_syswrap/syswrap-freebsd.c
index 12b18f4e6e..7ea04fdbc6 100644
--- a/coregrind/m_syswrap/syswrap-freebsd.c
+++ b/coregrind/m_syswrap/syswrap-freebsd.c
@@ -1410,6 +1410,7 @@ POST(sys_fcntl)
 {
    vg_assert(SUCCESS);
    if (ARG2 == VKI_F_DUPFD) {
+      POST_newFd_RES;
       if (!ML_(fd_allowed)(RES, "fcntl(DUPFD)", tid, True)) {
          VG_(close)(RES);
          SET_STATUS_Failure( VKI_EMFILE );
@@ -1419,6 +1420,7 @@ POST(sys_fcntl)
          }
       }
    } else if (ARG2 == VKI_F_DUPFD_CLOEXEC) {
+      POST_newFd_RES;
       if (!ML_(fd_allowed)(RES, "fcntl(DUPFD_CLOEXEC)", tid, True)) {
          VG_(close)(RES);
          SET_STATUS_Failure( VKI_EMFILE );
@@ -2728,6 +2730,7 @@ PRE(sys_fhopen)
 POST(sys_fhopen)
 {
    vg_assert(SUCCESS);
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "fhopen", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );
@@ -3537,6 +3540,7 @@ PRE(sys_kqueue)
 
 POST(sys_kqueue)
 {
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "kqueue", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );
@@ -4701,6 +4705,7 @@ PRE(sys_kmq_open)
 POST(sys_kmq_open)
 {
    vg_assert(SUCCESS);
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "mq_open", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );
@@ -5052,6 +5057,7 @@ PRE(sys_shm_open)
 POST(sys_shm_open)
 {
    vg_assert(SUCCESS);
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "shm_open", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );
@@ -6564,6 +6570,7 @@ PRE(sys_shm_open2)
 POST(sys_shm_open2)
 {
    vg_assert(SUCCESS);
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "shm_open2", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );
@@ -6825,6 +6832,7 @@ PRE(sys_kqueuex)
 
 POST(sys_kqueuex)
 {
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "kqueuex", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure(VKI_EMFILE);
@@ -6857,6 +6865,7 @@ PRE(sys_timerfd_create)
 
 POST(sys_timerfd_create)
 {
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "timerfd_create", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );

[Valgrind-developers] [valgrind] Illumos: increase coverage of --modify-fds syscalls

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=b13ec72918a1e9d11827bdde5d92c97222d79ea0

commit b13ec72918a1e9d11827bdde5d92c97222d79ea0
Author: Paul Floyd <pj...@wa...>
Date:   Thu Apr 17 21:26:24 2025 +0200

    Illumos: increase coverage of --modify-fds syscalls
    
    It looks like Solaris/Illumos is missing some F_DUP* coverage
    and we aren't handling syscalls that reaturn 2 fds (pipe, socketpair).
    Otherwise this should cover most Illumos cases at least.

Diff:
---
 coregrind/m_syswrap/syswrap-generic.c     |  1 +
 coregrind/m_syswrap/syswrap-solaris.c     | 15 ++++++++++++++-
 coregrind/m_syswrap/syswrap-x86-solaris.c |  1 +
 3 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/coregrind/m_syswrap/syswrap-generic.c b/coregrind/m_syswrap/syswrap-generic.c
index 1ab494c840..82a682a5ce 100644
--- a/coregrind/m_syswrap/syswrap-generic.c
+++ b/coregrind/m_syswrap/syswrap-generic.c
@@ -1814,6 +1814,7 @@ ML_(generic_POST_sys_socketpair) ( ThreadId tid,
    Int fd1 = ((Int*)arg3)[0];
    Int fd2 = ((Int*)arg3)[1];
    vg_assert(!sr_isError(res)); /* guaranteed by caller */
+   // @todo PJF this needs something like POST_newFd_RES for the two fds?
    POST_MEM_WRITE( arg3, 2*sizeof(int) );
    if (!ML_(fd_allowed)(fd1, "socketcall.socketpair", tid, True) ||
        !ML_(fd_allowed)(fd2, "socketcall.socketpair", tid, True)) {
diff --git a/coregrind/m_syswrap/syswrap-solaris.c b/coregrind/m_syswrap/syswrap-solaris.c
index 6b61a0e0dc..a5d34bc5d5 100644
--- a/coregrind/m_syswrap/syswrap-solaris.c
+++ b/coregrind/m_syswrap/syswrap-solaris.c
@@ -2492,6 +2492,7 @@ PRE(sys_pipe)
 POST(sys_pipe)
 {
    Int p0, p1;
+   // @todo PJF this needs something like POST_newFd_RES for the two fds?
 
 #if defined(SOLARIS_NEW_PIPE_SYSCALL)
    int *fds = (int*)ARG1;
@@ -4074,8 +4075,11 @@ PRE(sys_fcntl)
 
 POST(sys_fcntl)
 {
+   // @todo PJF we're missing
+   // F_DUP2FD_CLOEXEC F_DUP2FD_CLOFORK F_DUPFD_CLOFORK F_DUP3FD
    switch (ARG2 /*cmd*/) {
    case VKI_F_DUPFD:
+      POST_newFd_RES;
       if (!ML_(fd_allowed)(RES, "fcntl(F_DUPFD)", tid, True)) {
          VG_(close)(RES);
          SET_STATUS_Failure(VKI_EMFILE);
@@ -4084,6 +4088,7 @@ POST(sys_fcntl)
       break;
 
    case VKI_F_DUPFD_CLOEXEC:
+      POST_newFd_RES;
       if (!ML_(fd_allowed)(RES, "fcntl(F_DUPFD_CLOEXEC)", tid, True)) {
          VG_(close)(RES);
          SET_STATUS_Failure(VKI_EMFILE);
@@ -4092,6 +4097,7 @@ POST(sys_fcntl)
       break;
 
    case VKI_F_DUP2FD:
+      POST_newFd_RES;
       if (!ML_(fd_allowed)(RES, "fcntl(F_DUP2FD)", tid, True)) {
          VG_(close)(RES);
          SET_STATUS_Failure(VKI_EMFILE);
@@ -4258,6 +4264,7 @@ PRE(sys_openat)
 
 POST(sys_openat)
 {
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "openat", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure(VKI_EMFILE);
@@ -9555,7 +9562,13 @@ POST(sys_door)
 
    switch (ARG6 /*subcode*/) {
    case VKI_DOOR_CREATE:
-      door_record_server(tid, ARG1, RES);
+      POST_newFd_RES;
+      if (!ML_(fd_allowed)(RES, "door_create", tid, True)) {
+         VG_(close)(RES);
+         SET_STATUS_Failure( VKI_EMFILE );
+      } else {
+         door_record_server(tid, ARG1, RES);
+      }
       break;
    case VKI_DOOR_REVOKE:
       door_record_revoke(tid, ARG1);
diff --git a/coregrind/m_syswrap/syswrap-x86-solaris.c b/coregrind/m_syswrap/syswrap-x86-solaris.c
index be36625112..59c36c1852 100644
--- a/coregrind/m_syswrap/syswrap-x86-solaris.c
+++ b/coregrind/m_syswrap/syswrap-x86-solaris.c
@@ -971,6 +971,7 @@ PRE(sys_open64)
 
 POST(sys_open64)
 {
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "open64", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure(VKI_EMFILE);

[Valgrind-developers] [valgrind] Linux regtest 32bit: remove track_new.stderr.exp.debian32

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=992fefc9d1629da7f9eceaf514ac85a2929873da

commit 992fefc9d1629da7f9eceaf514ac85a2929873da
Author: Paul Floyd <pj...@wa...>
Date:   Thu Apr 17 13:00:49 2025 +0100

    Linux regtest 32bit: remove track_new.stderr.exp.debian32
    
    Not needed after switching from dprintf to write to trigger
    the error.

Diff:
---
 none/tests/Makefile.am                   |  2 +-
 none/tests/track_new.stderr.exp.debian32 | 10 ----------
 2 files changed, 1 insertion(+), 11 deletions(-)

diff --git a/none/tests/Makefile.am b/none/tests/Makefile.am
index 2275407738..37c94515d6 100644
--- a/none/tests/Makefile.am
+++ b/none/tests/Makefile.am
@@ -269,7 +269,7 @@ EXTRA_DIST = \
 	fdbaduse.stderr.exp fdbaduse.vgtest \
 	use_after_close.stderr.exp use_after_close.vgtest \
 	track_new.stderr.exp track_new.stdout.exp \
-	track_new.stderr.exp.debian32 track_new.vgtest track_new.stderr.exp-illumos
+	track_new.vgtest track_new.stderr.exp-illumos
 
 
 check_PROGRAMS = \
diff --git a/none/tests/track_new.stderr.exp.debian32 b/none/tests/track_new.stderr.exp.debian32
deleted file mode 100644
index eb86871743..0000000000
--- a/none/tests/track_new.stderr.exp.debian32
+++ /dev/null
@@ -1,10 +0,0 @@
-File descriptor was closed already
-   at 0x........: llseek (in /...libc...)
-   by 0x........: dprintf (in /...libc...)
-   by 0x........: main
- Previously closed
-   at 0x........: close (in /...libc...)
-   by 0x........: main
- Originally opened
-   at 0x........: creat (in /...libc...)
-   by 0x........: main

[Valgrind-developers] [valgrind] Illumos regtest: update none/tests/track_new

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=b6346abc4896b3d50882e6c45dfccbf2872ce3b7

commit b6346abc4896b3d50882e6c45dfccbf2872ce3b7
Author: Paul Floyd <pj...@wa...>
Date:   Thu Apr 17 13:22:34 2025 +0200

    Illumos regtest: update none/tests/track_new
    
    Not sure why I get file/line with Illumos but not
    Linux or FreeBSD.

Diff:
---
 none/tests/track_new.stderr.exp-illumos | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/none/tests/track_new.stderr.exp-illumos b/none/tests/track_new.stderr.exp-illumos
index 670530b603..ff036ca011 100644
--- a/none/tests/track_new.stderr.exp-illumos
+++ b/none/tests/track_new.stderr.exp-illumos
@@ -1,10 +1,8 @@
 File descriptor was closed already
-   at 0x........: fdopen (in /...libc...)
-   by 0x........: vdprintf (in /...libc...)
-   by 0x........: dprintf (in /...libc...)
+   at 0x........: write (in /...libc...)
    by 0x........: main
  Previously closed
    at 0x........: close (in /...libc...)
    by 0x........: main
  Originally opened
-   at 0x........: main (track_new.c:11)
+   at 0x........: main (track_new.c:8)

[Valgrind-developers] [valgrind] FreeBSD: changes for --modify-fds

[Paul F. <pa...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=972e1878bfa03fba4d29f8f05f1c65fbb91b8b2a

commit 972e1878bfa03fba4d29f8f05f1c65fbb91b8b2a
Author: Paul Floyd <pj...@wa...>
Date:   Thu Apr 17 13:00:04 2025 +0200

    FreeBSD: changes for --modify-fds
    
    Again this is just the minumum to get the track_new test to pass.
    Other syscalls later.

Diff:
---
 coregrind/m_syswrap/syswrap-freebsd.c | 1 +
 none/tests/track_new.c                | 5 +----
 none/tests/track_new.stderr.exp       | 1 -
 3 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/coregrind/m_syswrap/syswrap-freebsd.c b/coregrind/m_syswrap/syswrap-freebsd.c
index 6fd4367a83..12b18f4e6e 100644
--- a/coregrind/m_syswrap/syswrap-freebsd.c
+++ b/coregrind/m_syswrap/syswrap-freebsd.c
@@ -5317,6 +5317,7 @@ PRE(sys_openat)
 POST(sys_openat)
 {
    vg_assert(SUCCESS);
+   POST_newFd_RES;
    if (!ML_(fd_allowed)(RES, "openat", tid, True)) {
       VG_(close)(RES);
       SET_STATUS_Failure( VKI_EMFILE );
diff --git a/none/tests/track_new.c b/none/tests/track_new.c
index e4bb546fb4..718fa3a374 100644
--- a/none/tests/track_new.c
+++ b/none/tests/track_new.c
@@ -1,7 +1,4 @@
-#define _XOPEN_SOURCE 700
-
 #include <fcntl.h>
-#include <stdio.h>
 #include <unistd.h>
 #include <sys/stat.h>
 
@@ -15,7 +12,7 @@ main (void)
   /* Lets open another file... */
   int newfd = open ("foobad.txt", O_RDWR|O_CREAT, S_IRUSR | S_IWUSR);
   /* ... oops we are using the wrong fd (but same number...) */
-  dprintf (oldfd, "some new text\n");
+  write(oldfd, "some new text\n", 14);
 
   close (newfd);
   return 0;
diff --git a/none/tests/track_new.stderr.exp b/none/tests/track_new.stderr.exp
index 23dec73bbc..cd443c51a9 100644
--- a/none/tests/track_new.stderr.exp
+++ b/none/tests/track_new.stderr.exp
@@ -1,6 +1,5 @@
 File descriptor was closed already
    at 0x........: write (in /...libc...)
-   by 0x........: dprintf (in /...libc...)
    by 0x........: main
  Previously closed
    at 0x........: close (in /...libc...)