valgrind-developers — Discussion list for developers of valgrind

[Valgrind-developers] Upgrade to GNU General Public License version 3

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

Hi,

We discussed this on this list, with the core valgrind developers and
the FSF Licensing and Compliance Lab what the correct way was to do
this.

The patch itself is a little too big to post to the list, but you can
see it here:
https://code.wildebeest.org/git/user/mjw/valgrind/log/?h=gplv3

There are two patches, one cleanup:

    Update where to get a copy of the GNU General Public License
    
    Several files still said to write to the Free Software Foundation at a
    particular address. Add a reference to <http://www.gnu.org/licenses/>
    instead.
    
    Some files that were imported into valgrind from other places still
    mention the old address. They should be updated when new versions
    get imported.

And the actual update:

    Upgrade to GNU General Public License version 3
    
    - Update COPYING and VEX/LICENSE.GPL to version 3.
    - Update README, NEWS, docs/manual license and contributing text.
    - Update file headers to say either version 3 of the License,
      or (at your option) any later version.
    - Leave tests and perf file headers as is, unless the code is derived
      from Valgrind/VEX.
    - Leave valgrind.h, cachegrind.h, callgrind.h, drd.h, helgrind.h,
      memcheck.h and dhat.h Hybrid-BSD licensed.

I like to merge these on Friday (Oct 17) before releasing 3.26.0-RC1.

Please let me know if you have any questions or concerns about this
update.

Thanks,

Mark

Re: [Valgrind-developers] Preparing for Valgrind 3.26.0 (RC on Oct 17, final Oct 24)

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

On 2025-10-12 18:28, Florian Krohm wrote:
> Hi Mark,
>
> On 09.10.25 22:03, Mark Wielaard wrote:

>> Please let me know what you believe are blockers for the release or
>> fixes you really would like to see in.
>>
>
> I looked at building valgrind on mips (cfarm230 machine) because there 
> are no nightly runs..
> If mips is a supported platform (?), then there is a bit of work there:


It would be good to get a few more architectures covered by the nightly 
tests:

RISCV, MIPS, PPC with a more recent GCC.

As an indication of where things are going, FreeBSD 15 is making quite a 
few changes to its platform support. Getting dropped will be

x86

MIPS (already dropped in FreeBSD 14)

32bit PPC

32bit arm6

Petar Jovanovic does occasionally push MIPS changes. The last one was a 
bit less than a year ago.

> configury:
> Need to detect presence of -lcrypto and -laio otherwise
> memcheck/tests/linux/bug480706.c
> memcheck/tests/linux/bug420682_1.c
> memcheck/tests/linux/bug420682_2.c
> will not build. 

I'll have a look at those since I added all those testcases.

> The GCC on that machine is rather old: 8.3.0. It does not recognise
> -Wno-missing-exception-spec
> -Wno-implicit-exception-spec-mismatch
> -Wno-alloc-size-larger-than
> -Walloc-size-larger-than=
>
They should be in FLAG_FNO_XXX variables in configure.ac and the 
relevant Makefile.am.


> Building:
> There are 126 warnings about:
>   cast increases required alignment of target type
> I do not know whether that is a problem (other than performance 
> degradation) or not. If it is not then those warnings should be 
> suppressed. Back in the days we were striving to have a clean 
> build/check/regtest which I personally find to be a laudable goal.
>
Yes that is still the goal. Most platforms that run the nightly test 
have either 0 or 1 failure. illumos has a bit more with 7 failing testcases.


> Regtest:
>
> == 768 tests, 288 stderr failures, 60 stdout failures, 10 stderrB 
> failures, 11 stdoutB failures, 3 post failures ==
>
> Essentially all memcheck tests are failing this way:
>
> valgrind:  Fatal error at startup: a function redirection
> valgrind:  which is mandatory for this platform-tool combination
> valgrind:  cannot be set up.  Details of the redirection are:

Hmm. Not for the upcoming release, but I'd like to add a configure 
option for mandatory index and strlen redirection.

>
> which renders memcheck unusable. libc6-dbg appears to be installed, 
> though.
> I guess a MAINTAINERS file would be good to find out whom to pester 
> with this :)
>
> For PPC there is https://bugs.kde.org/show_bug.cgi?id=508661 which 
> would be good to get fixed. 

Perhaps Carl Love will take a look?


A+

Paul

Re: [Valgrind-developers] Preparing for Valgrind 3.26.0 (RC on Oct 17, final Oct 24)

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

Hi Mark,

On 09.10.25 22:03, Mark Wielaard wrote:
> 
> There are a couple of things I like to add before RC1:
> 
> - Update to GPLv3+.
>    We discussed this on the list and then with some core developers
>    asking for advice from the FSF Legal and Compliance team.
>    I'll post a patch tomorrow with the full rational, so we can
>    discuss it before applying for RC1.

Oh yes! Can't wait :)

> Please let me know what you believe are blockers for the release or
> fixes you really would like to see in.
> 

I looked at building valgrind on mips (cfarm230 machine) because there are no 
nightly runs..
If mips is a supported platform (?), then there is a bit of work there:

configury:
Need to detect presence of -lcrypto and -laio otherwise
memcheck/tests/linux/bug480706.c
memcheck/tests/linux/bug420682_1.c
memcheck/tests/linux/bug420682_2.c
will not build.

The GCC on that machine is rather old: 8.3.0. It does not recognise
-Wno-missing-exception-spec
-Wno-implicit-exception-spec-mismatch
-Wno-alloc-size-larger-than
-Walloc-size-larger-than=

Building:
There are 126 warnings about:
   cast increases required alignment of target type
I do not know whether that is a problem (other than performance degradation) or 
not. If it is not then those warnings should be suppressed. Back in the days we 
were striving to have a clean build/check/regtest which I personally find to be 
a laudable goal.

Regtest:

== 768 tests, 288 stderr failures, 60 stdout failures, 10 stderrB failures, 11 
stdoutB failures, 3 post failures ==

Essentially all memcheck tests are failing this way:

valgrind:  Fatal error at startup: a function redirection
valgrind:  which is mandatory for this platform-tool combination
valgrind:  cannot be set up.  Details of the redirection are:
....

which renders memcheck unusable. libc6-dbg appears to be installed, though.
I guess a MAINTAINERS file would be good to find out whom to pester with this :)

For PPC there is https://bugs.kde.org/show_bug.cgi?id=508661 which would be good 
to get fixed.

Cheers,
    Florian

[Valgrind-developers] [valgrind] Bug 509406 - FreeBSD 15 issues

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

https://sourceware.org/cgit/valgrind/commit/?id=ea458dade2cb8f42586ef5a3f271b4cdddbf647a

commit ea458dade2cb8f42586ef5a3f271b4cdddbf647a
Author: Paul Floyd <pj...@wa...>
Date:   Sat Oct 4 18:46:58 2025 +0000

    Bug 509406 - FreeBSD 15 issues
    
    This is the last of a series of fixes and bodges.
    
    I now get the following results.
    
    On arm64
    == 793 tests, 1 stderr failure, 0 stdout failures, 0 stderrB failures, 0 stdoutB failures, 0 post failures ==
    memcheck/tests/thread_alloca             (stderr)
    
    (which is not a new failure) and
    
    == 947 tests, 1 stderr failure, 0 stdout failures, 0 stderrB failures, 1 stdoutB failure, 0 post failures ==
    gdbserver_tests/nlvgdbsigqueue           (stderr)
    gdbserver_tests/nlvgdbsigqueue           (stdoutB)
    
    (again a test that fails regularly)

Diff:
---
 NEWS                                      | 3 ++-
 memcheck/tests/freebsd/pdfork_pdkill.supp | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/NEWS b/NEWS
index eebbf4fabd..b1b27e815b 100644
--- a/NEWS
+++ b/NEWS
@@ -70,7 +70,7 @@ There is a workaround. Run the following command (as root).
   "UnsafeZeroSize".
 
   Checks for C23 free_sized and free_aligned_sized have been added to
-  Linux. Almost no libraries support these functions yes, with
+  Linux. Almost no libraries support these functions yet, with
   the exception being Google tcmalloc.
 
 * ==================== FIXED BUGS ====================
@@ -156,6 +156,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 509107  memcheck/tests/duplicate_align_size_errors.cpp fails
 509139  Update BadSize error messages
 509258  FreeBSD: add jail_attach_jd and jail_remove_jd syscall wrappers
+509406  FreeBSD 15 issues
 509517  s390x: Even/odd lane confusion in various vector insns
 509566  Wrap amd64-linux syscall: 442 (mount_setattr)
 509572  s390x: Overhaul BFP testsuite
diff --git a/memcheck/tests/freebsd/pdfork_pdkill.supp b/memcheck/tests/freebsd/pdfork_pdkill.supp
index cd3ad74798..47a836de92 100644
--- a/memcheck/tests/freebsd/pdfork_pdkill.supp
+++ b/memcheck/tests/freebsd/pdfork_pdkill.supp
@@ -7,7 +7,7 @@
 {
    arm64 suppression 2
    Memcheck:Cond
-   fun:__sys_pdfork
+   fun:_*pdfork
 }

[Valgrind-developers] [valgrind] FreeBSD regtest: add a prereq for bug452274

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

https://sourceware.org/cgit/valgrind/commit/?id=0a5028313b6ff0462b77318b250ed31a2b069510

commit 0a5028313b6ff0462b77318b250ed31a2b069510
Author: Paul Floyd <pj...@wa...>
Date:   Sun Oct 12 13:29:45 2025 +0200

    FreeBSD regtest: add a prereq for bug452274
    
    This test tends to hang when running in a VM. The problem is that
    the test spams both syscalls (write) and signals (SIGVTALRM from
    setitimer). The test halts after 100 signals have been handled.
    Under VirtualBox the timer signals get starved by the syscalls
    and the limit of 100 doesn't get reached (at least not in any
    reasonable amount of time).

Diff:
---
 none/tests/freebsd/bug452274.vgtest | 1 +
 1 file changed, 1 insertion(+)

diff --git a/none/tests/freebsd/bug452274.vgtest b/none/tests/freebsd/bug452274.vgtest
index 30b1179a61..23092b1381 100644
--- a/none/tests/freebsd/bug452274.vgtest
+++ b/none/tests/freebsd/bug452274.vgtest
@@ -1,3 +1,4 @@
+prereq: [ `sysctl -n kern.vm_guest` = "none" ]
 prog: bug452274
 vgopts: -q
 stderr_filter: filter_452274

[Valgrind-developers] [valgrind] FreeBSD cleanup: remove an orphaned comment

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

https://sourceware.org/cgit/valgrind/commit/?id=5addc084570ee5fbb2876093415ffefdb8b9645c

commit 5addc084570ee5fbb2876093415ffefdb8b9645c
Author: Paul Floyd <pj...@wa...>
Date:   Sat Oct 11 15:11:10 2025 +0200

    FreeBSD cleanup: remove an orphaned comment
    
    Removed the function, forgot to remove the reminder comment.

Diff:
---
 coregrind/m_initimg/initimg-freebsd.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/coregrind/m_initimg/initimg-freebsd.c b/coregrind/m_initimg/initimg-freebsd.c
index 2f8ac82ebb..f187c0e3de 100644
--- a/coregrind/m_initimg/initimg-freebsd.c
+++ b/coregrind/m_initimg/initimg-freebsd.c
@@ -334,10 +334,6 @@ static const struct auxv *find_auxv(const UWord* sp)
    return (const struct auxv *)sp;
 }
 
-/*
- * @todo PJF Make this multi-platform
- */
-
 /* ----------------------------------------------------------------
 
    This sets up the client's initial stack, containing the args,

[Valgrind-developers] [valgrind] NEWS: Add FreeBSD platform section

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

https://sourceware.org/cgit/valgrind/commit/?id=01db0b93a8b9145d26e2e15da1759f1377f7ee91

commit 01db0b93a8b9145d26e2e15da1759f1377f7ee91
Author: Paul Floyd <pj...@wa...>
Date:   Sat Oct 11 15:09:46 2025 +0200

    NEWS: Add FreeBSD platform section

Diff:
---
 NEWS | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/NEWS b/NEWS
index b097e3789f..eebbf4fabd 100644
--- a/NEWS
+++ b/NEWS
@@ -16,6 +16,21 @@ X86/macOS 10.13, AMD64/macOS 10.13 and nanoMIPS/Linux.
 
 * ================== PLATFORM CHANGES =================
 
+FreeBSD 15 (which is expected to ship in December 2025, after
+Valgrind 3.26 is released) contains a change to ptrace that affects
+use of Valgrind with vgdb. This impacts the mechanism that vgdb
+uses to interrupt Valgrind if all threads are blocked and you want
+to get back to the gdb prompt by hitting ctrl-c. This mechanism
+is no longer reliable. On arm64 Valgrind will crash with an assert.
+On amd64 syscalls may give spurious and incorrect return codes.
+
+There is a workaround. Run the following command (as root).
+
+ sysctl debug.ptrace_attach_transparent=0
+
+ See also
+ https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=290008
+
 * ==================== TOOL CHANGES ===================
 
 * There is a new utility script, "vgstack". It has two

[Valgrind-developers] [valgrind] gdbserver doc: add some comments about the invoker implementations.

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

https://sourceware.org/cgit/valgrind/commit/?id=7068b807905af87729a66d192f7c10247c42212f

commit 7068b807905af87729a66d192f7c10247c42212f
Author: Paul Floyd <pj...@wa...>
Date:   Sat Oct 11 13:01:03 2025 +0200

    gdbserver doc: add some comments about the invoker implementations.

Diff:
---
 gdbserver_tests/README_DEVELOPERS | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

diff --git a/gdbserver_tests/README_DEVELOPERS b/gdbserver_tests/README_DEVELOPERS
index 250c14326b..0a9f2846f6 100644
--- a/gdbserver_tests/README_DEVELOPERS
+++ b/gdbserver_tests/README_DEVELOPERS
@@ -30,11 +30,27 @@ if testing with a lower version).
 Test behaviour with gdb < 7.0 is unknown: some might fail,
 some might block or loop for a very long time.
 
-Some tests implies to have a vgdb "ptrace invoker" capable.
+Some tests imply having the vgdb "ptrace invoker" capability.
 
 The prerequisite are established during make regtest (using marker files).
 Each test verifies the prerequisite using the prereq: line.
 
+The invoker is implemented as follows on the currently supported OSes.
+
+------------------------------------------------------------
+|  Linux  |        ptrace         | vgdb-invoker-ptrace.c  |
+| FreeBSD |        ptrace         | vgdb-invoker-freebsd.c |
+| Solaris | process control files | vgdb-invoker-solaris.c |
+| Darwin  |    not implemented    | vgdb-invoker-none.c    |
+------------------------------------------------------------
+
+The source files are all in coregrind/. ptrace is not part of any
+standardised interface like POSIX. That means that the Linux and
+FreeBSD implementations are very similar but there are some
+differences. Darwin not having an invoker means that it is not possible
+to interrupt Valgrind attached to GDB if it hangs. You will need to
+kill Valgrind and start again if that happens.
+
 In case of failing tests
 ------------------------
 When executed with a new gdb version and/or depending on the OS version,

[Valgrind-developers] [valgrind] FreeBSD regtest: improve filter_arg_check

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

https://sourceware.org/cgit/valgrind/commit/?id=5a3aa8c8d13720636df4415ca6bb0aa43b67ef1c

commit 5a3aa8c8d13720636df4415ca6bb0aa43b67ef1c
Author: Paul Floyd <pj...@wa...>
Date:   Sat Oct 11 00:32:23 2025 +0200

    FreeBSD regtest: improve filter_arg_check
    
    This was looking for PIDs with 5 digits so the test failed
    when the PIDs being used have fewer digits.

Diff:
---
 memcheck/tests/amd64-freebsd/filter_arg_check | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/memcheck/tests/amd64-freebsd/filter_arg_check b/memcheck/tests/amd64-freebsd/filter_arg_check
index 334a326694..c1152cf836 100755
--- a/memcheck/tests/amd64-freebsd/filter_arg_check
+++ b/memcheck/tests/amd64-freebsd/filter_arg_check
@@ -18,5 +18,5 @@
 grep "SYSCALL.*sendfile" |
 sed 's/==.*//' |
 awk '{l=length($9);$9="0x"substr($9, l-2, l);print}' |
-sed -E 's/\[[0-9]{5}/[xxxxx/'
+sed -E 's/\[[0-9]+/[xxxxx/'

[Valgrind-developers] [valgrind] FreeBSD regtest: tweak setcred test

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

https://sourceware.org/cgit/valgrind/commit/?id=9640ff697cd9c9be485ea8bffc6381395b3ad1d4

commit 9640ff697cd9c9be485ea8bffc6381395b3ad1d4
Author: Paul Floyd <pj...@wa...>
Date:   Fri Oct 10 20:57:26 2025 +0200

    FreeBSD regtest: tweak setcred test
    
    Should now work on both arm64 and amd64
    (and with a diff on the size of struct setcred for x86).
    
    Still don't see what caused the Conditional jump error.

Diff:
---
 memcheck/tests/freebsd/setcred.cpp        | 33 +++++++++++++++++++++----------
 memcheck/tests/freebsd/setcred.stderr.exp | 26 ++++++++++++------------
 2 files changed, 36 insertions(+), 23 deletions(-)

diff --git a/memcheck/tests/freebsd/setcred.cpp b/memcheck/tests/freebsd/setcred.cpp
index 619d1959f1..1ae28f7f04 100644
--- a/memcheck/tests/freebsd/setcred.cpp
+++ b/memcheck/tests/freebsd/setcred.cpp
@@ -19,30 +19,23 @@ int main()
 
    std::memset(&cred1, 250, sizeof(cred1));
 
-   // needs to be root to work correctly
-   ret = setcred(flags1, &cred1, size1);
-   assert(ret == -1);
-   assert(errno == EPERM);
-
-   // not accessible
-   ret = setcred(flags1, nullptr, size1);
-   assert(ret == -1);
-   assert(errno == EFAULT);
-
    // uninit
    ret = setcred(flags2+x0, (struct setcred*)x0, size1+x0);
    assert(ret == -1);
    assert(errno == EFAULT);
+   errno = 0;
 
    // invalid flags
    ret = setcred(9999+x0, &cred1, sizeof(cred1));
    assert(ret == -1);
    assert(errno == EINVAL);
+   errno = 0;
 
    // invalid size
    ret = setcred(flags1, &cred1, 3+x0);
    assert(ret == -1);
    assert(errno == EINVAL);
+   errno = 0;
 
    cred2 = new struct setcred;
 
@@ -50,6 +43,26 @@ int main()
    ret = setcred(flags1, cred2, size1);
    assert(ret == -1);
    assert(errno == EPERM);
+   errno = 0;
+
+   // PJF these two calls to setcred were before the
+   // uninit one that is now first
+   // that was fine on arm64 but on amd64 the uninit
+   // call generated an extre Conditional jump ... error
+
+   // fairly mysterious, and usually that means that there
+   // is something wrong with the syscall wrapper
+
+   // needs to be root to work correctly
+   ret = setcred(flags1, &cred1, size1);
+   assert(ret == -1);
+   assert(errno == EPERM);
+   errno = 0;
+
+   // not accessible
+   ret = setcred(flags1, nullptr, size1);
+   assert(ret == -1);
+   assert(errno == EFAULT);
 
    delete cred2;
    free(px);
diff --git a/memcheck/tests/freebsd/setcred.stderr.exp b/memcheck/tests/freebsd/setcred.stderr.exp
index 8766361c09..5f84b4f288 100644
--- a/memcheck/tests/freebsd/setcred.stderr.exp
+++ b/memcheck/tests/freebsd/setcred.stderr.exp
@@ -1,37 +1,37 @@
-Syscall param setcred(wcred) points to unaddressable byte(s)
-   at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:28)
- Address 0x........ is not stack'd, malloc'd or (recently) free'd
-
 Syscall param setcred(flags) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:33)
+   by 0x........: main (setcred.cpp:23)
 
 Syscall param setcred(wcred) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:33)
+   by 0x........: main (setcred.cpp:23)
 
 Syscall param setcred(size) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:33)
+   by 0x........: main (setcred.cpp:23)
 
 Syscall param setcred(wcred) points to unaddressable byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:33)
+   by 0x........: main (setcred.cpp:23)
  Address 0x........ is not stack'd, malloc'd or (recently) free'd
 
 Syscall param setcred(flags) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:38)
+   by 0x........: main (setcred.cpp:29)
 
 Syscall param setcred(size) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:43)
+   by 0x........: main (setcred.cpp:35)
 
 Syscall param setcred(wcred) points to uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:50)
+   by 0x........: main (setcred.cpp:43)
  Address 0x........ is 0 bytes inside a block of size 48 alloc'd
    at 0x........: ...operator new... (vg_replace_malloc.c:...)
-   by 0x........: main (setcred.cpp:47)
+   by 0x........: main (setcred.cpp:40)
+
+Syscall param setcred(wcred) points to unaddressable byte(s)
+   at 0x........: setcred (in /...libc...)
+   by 0x........: main (setcred.cpp:63)
+ Address 0x........ is not stack'd, malloc'd or (recently) free'd

[Valgrind-developers] [valgrind] Define __NR_swapon and __NR_swapoff on riscv64

[Martin C. <mc...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=001f7439d6e20f3b117b5032070111fd701195bc

commit 001f7439d6e20f3b117b5032070111fd701195bc
Author: Martin Cermak <mc...@re...>
Date:   Fri Oct 10 18:12:02 2025 +0200

    Define __NR_swapon and __NR_swapoff on riscv64
    
    Update commit c3f8f3ceb4ba6942993762f82911a670fa128789 adding
    a missing define.
    
    https://bugs.kde.org/show_bug.cgi?id=368791

Diff:
---
 include/vki/vki-scnums-riscv64-linux.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/vki/vki-scnums-riscv64-linux.h b/include/vki/vki-scnums-riscv64-linux.h
index 6130ce37f1..d630a19161 100644
--- a/include/vki/vki-scnums-riscv64-linux.h
+++ b/include/vki/vki-scnums-riscv64-linux.h
@@ -254,6 +254,8 @@
 #define __NR_execve 221
 #define __NR3264_mmap 222
 #define __NR3264_fadvise64 223
+#define __NR_swapon 224
+#define __NR_swapoff 225
 #define __NR_mprotect 226
 #define __NR_msync 227
 #define __NR_mlock 228

[Valgrind-developers] [valgrind] Wrap swapon and swapoff syscalls on non-mips arches.

[Martin C. <mc...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=c3f8f3ceb4ba6942993762f82911a670fa128789

commit c3f8f3ceb4ba6942993762f82911a670fa128789
Author: Martin Cermak <mc...@re...>
Date:   Wed Oct 8 10:07:30 2025 +0200

    Wrap swapon and swapoff syscalls on non-mips arches.
    
    Drop the mips-specific syswraps for swapon and swapoff syscalls.
    Instead, wrap these syscalls on all the arches.
    
    https://bugs.kde.org/show_bug.cgi?id=368791

Diff:
---
 NEWS                                         |  1 +
 coregrind/m_syswrap/priv_syswrap-linux.h     |  3 +++
 coregrind/m_syswrap/syswrap-amd64-linux.c    |  4 ++--
 coregrind/m_syswrap/syswrap-arm-linux.c      |  4 ++--
 coregrind/m_syswrap/syswrap-arm64-linux.c    |  4 ++--
 coregrind/m_syswrap/syswrap-linux.c          | 13 +++++++++++++
 coregrind/m_syswrap/syswrap-mips32-linux.c   |  4 ++--
 coregrind/m_syswrap/syswrap-mips64-linux.c   | 18 ++----------------
 coregrind/m_syswrap/syswrap-nanomips-linux.c | 18 ++----------------
 coregrind/m_syswrap/syswrap-ppc32-linux.c    |  4 ++--
 coregrind/m_syswrap/syswrap-ppc64-linux.c    |  4 ++--
 coregrind/m_syswrap/syswrap-riscv64-linux.c  |  2 ++
 coregrind/m_syswrap/syswrap-s390x-linux.c    |  4 ++--
 coregrind/m_syswrap/syswrap-x86-linux.c      |  4 ++--
 14 files changed, 39 insertions(+), 48 deletions(-)

diff --git a/NEWS b/NEWS
index ca3f9fca4a..b097e3789f 100644
--- a/NEWS
+++ b/NEWS
@@ -72,6 +72,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 309554  Wrap syscall remap_file_pages (216)
 331311  Valgrind shows open files in /proc/self/fd that don't work for the process
 338803  Handling of dwz debug alt files or cross-CU is broken
+368791  Handle swapon and swapoff syscalls as linux generic
 369030  Wrap linux syscall: 171 (setdomainname)
 388526  Inconsistent severity in message text: "WARNING: Serious error"
 418756  MAP_FIXED_NOREPLACE mmap flag unsupported
diff --git a/coregrind/m_syswrap/priv_syswrap-linux.h b/coregrind/m_syswrap/priv_syswrap-linux.h
index 53aa1b582d..ca462e896f 100644
--- a/coregrind/m_syswrap/priv_syswrap-linux.h
+++ b/coregrind/m_syswrap/priv_syswrap-linux.h
@@ -288,6 +288,9 @@ DECL_TEMPLATE(linux, sys_init_module);
 DECL_TEMPLATE(linux, sys_finit_module);
 DECL_TEMPLATE(linux, sys_delete_module);
 
+DECL_TEMPLATE(linux, sys_swapon);
+DECL_TEMPLATE(linux, sys_swapoff);
+
 // Linux-specific (oprofile-related)
 DECL_TEMPLATE(linux, sys_lookup_dcookie);        // (*/32/64) L
 
diff --git a/coregrind/m_syswrap/syswrap-amd64-linux.c b/coregrind/m_syswrap/syswrap-amd64-linux.c
index 4a3bad55f4..838bf5e848 100644
--- a/coregrind/m_syswrap/syswrap-amd64-linux.c
+++ b/coregrind/m_syswrap/syswrap-amd64-linux.c
@@ -670,8 +670,8 @@ static SyscallTableEntry syscall_table[] = {
 
    LINX_(__NR_mount,             sys_mount),          // 165
    LINX_(__NR_umount2,           sys_umount),         // 166 
-   //   (__NR_swapon,            sys_swapon),         // 167 
-   //   (__NR_swapoff,           sys_swapoff),        // 168 
+   LINX_(__NR_swapon,            sys_swapon),         // 167 
+   LINX_(__NR_swapoff,           sys_swapoff),        // 168 
    //   (__NR_reboot,            sys_reboot),         // 169 
 
    GENX_(__NR_sethostname,       sys_sethostname),    // 170 
diff --git a/coregrind/m_syswrap/syswrap-arm-linux.c b/coregrind/m_syswrap/syswrap-arm-linux.c
index f55ed05e61..1fda7ba8a0 100644
--- a/coregrind/m_syswrap/syswrap-arm-linux.c
+++ b/coregrind/m_syswrap/syswrap-arm-linux.c
@@ -653,7 +653,7 @@ static SyscallTableEntry syscall_main_table[] = {
 //zz 
    GENXY(__NR_readlink,          sys_readlink),       // 85
 //zz    //   (__NR_uselib,            sys_uselib),         // 86 */Linux
-//zz    //   (__NR_swapon,            sys_swapon),         // 87 */Linux
+   LINX_(__NR_swapon,            sys_swapon),         // 87 */Linux
 //zz    //   (__NR_reboot,            sys_reboot),         // 88 */Linux
 //zz    //   (__NR_readdir,           old_readdir),        // 89 -- superseded
 //zz 
@@ -687,7 +687,7 @@ static SyscallTableEntry syscall_main_table[] = {
 // PLAXY(__NR_vm86old,           sys_vm86old),        // 113 __NR_syscall... weird
    GENXY(__NR_wait4,             sys_wait4),          // 114
 //zz 
-//zz    //   (__NR_swapoff,           sys_swapoff),        // 115 */Linux 
+   LINX_(__NR_swapoff,           sys_swapoff),        // 115 */Linux 
    LINXY(__NR_sysinfo,           sys_sysinfo),        // 116
 //   _____(__NR_ipc,               sys_ipc),            // 117
    GENX_(__NR_fsync,             sys_fsync),          // 118
diff --git a/coregrind/m_syswrap/syswrap-arm64-linux.c b/coregrind/m_syswrap/syswrap-arm64-linux.c
index b28b4598c0..175002e6ba 100644
--- a/coregrind/m_syswrap/syswrap-arm64-linux.c
+++ b/coregrind/m_syswrap/syswrap-arm64-linux.c
@@ -772,8 +772,8 @@ static SyscallTableEntry syscall_main_table[] = {
    GENX_(__NR_execve,            sys_execve),            // 221
    PLAX_(__NR_mmap,              sys_mmap),              // 222
    PLAX_(__NR_fadvise64,         sys_fadvise64),         // 223
-   //   (__NR_swapon,            sys_swapon),            // 224
-   //   (__NR_swapoff,           sys_swapoff),           // 225
+   LINX_(__NR_swapon,            sys_swapon),            // 224
+   LINX_(__NR_swapoff,           sys_swapoff),           // 225
    GENXY(__NR_mprotect,          sys_mprotect),          // 226
    GENX_(__NR_msync,             sys_msync),             // 227
    GENX_(__NR_mlock,             sys_mlock),             // 228
diff --git a/coregrind/m_syswrap/syswrap-linux.c b/coregrind/m_syswrap/syswrap-linux.c
index d891ac6daa..1402c3d1f5 100644
--- a/coregrind/m_syswrap/syswrap-linux.c
+++ b/coregrind/m_syswrap/syswrap-linux.c
@@ -2133,6 +2133,19 @@ POST(sys_ppoll_time64)
    ppoll_post_helper (tid, arrghs, status);
 }
 
+PRE(sys_swapon)
+{
+   PRINT("sys_swapon ( %#lx, %#lx )", ARG1, ARG2);
+   PRE_REG_READ2(long, "swapon", const void *, path, int, flags);
+   PRE_MEM_RASCIIZ( "swapon(path)", ARG1);
+}
+
+PRE(sys_swapoff)
+{
+   PRINT("sys_swapoff ( %#lx )", ARG1);
+   PRE_REG_READ1(long, "swapoff", const void *, path);
+   PRE_MEM_RASCIIZ( "swapoff(path)", ARG1);
+}
 
 /* ---------------------------------------------------------------------
    epoll_* wrappers
diff --git a/coregrind/m_syswrap/syswrap-mips32-linux.c b/coregrind/m_syswrap/syswrap-mips32-linux.c
index c4d7a66206..4edfe8a703 100644
--- a/coregrind/m_syswrap/syswrap-mips32-linux.c
+++ b/coregrind/m_syswrap/syswrap-mips32-linux.c
@@ -852,7 +852,7 @@ static SyscallTableEntry syscall_main_table[] = {
    //..    //   (__NR_oldlstat,          sys_lstat),             // 84
    GENXY (__NR_readlink,               sys_readlink),                // 85
    //..    //   (__NR_uselib,            sys_uselib),            // 86
-   //..    //   (__NR_swapon,            sys_swapon),            // 87
+   LINX_ (__NR_swapon,                 sys_swapon),                  // 87
    //..    //   (__NR_reboot,            sys_reboot),            // 88
    //..    //   (__NR_readdir,           old_readdir),           // 89
    PLAX_ (__NR_mmap,                   sys_mmap),                    // 90
@@ -880,7 +880,7 @@ static SyscallTableEntry syscall_main_table[] = {
    //..    GENX_(__NR_idle,              sys_ni_syscall),        // 112
    //..    //   (__NR_vm86old,           sys_vm86old),           // 113
    GENXY (__NR_wait4,                  sys_wait4),                   // 114
-   //..    //   (__NR_swapoff,           sys_swapoff),           // 115
+   LINX_ (__NR_swapoff,                sys_swapoff),                 // 115
    LINXY (__NR_sysinfo,                sys_sysinfo),                 // 116
    LINXY (__NR_ipc,                    sys_ipc),                     // 117
    GENX_ (__NR_fsync,                  sys_fsync),                   // 118
diff --git a/coregrind/m_syswrap/syswrap-mips64-linux.c b/coregrind/m_syswrap/syswrap-mips64-linux.c
index 45af0a3fd9..4fb6f060e3 100644
--- a/coregrind/m_syswrap/syswrap-mips64-linux.c
+++ b/coregrind/m_syswrap/syswrap-mips64-linux.c
@@ -215,8 +215,6 @@ SysRes sys_set_tls ( ThreadId tid, Addr tlsptr )
    file, but that requires even more macro magic. */
 
 DECL_TEMPLATE (mips_linux, sys_set_thread_area);
-DECL_TEMPLATE (mips_linux, sys_swapon);
-DECL_TEMPLATE (mips_linux, sys_swapoff);
 DECL_TEMPLATE (mips_linux, sys_sethostname);
 DECL_TEMPLATE (mips_linux, sys_reboot);
 DECL_TEMPLATE (mips_linux, sys_cacheflush);
@@ -236,18 +234,6 @@ PRE(sys_sched_rr_get_interval)
    *flags |= SfMayBlock;
 }
 
-PRE(sys_swapon)
-{
-   PRINT("sys_swapon ( %#" FMT_REGWORD "x, %#" FMT_REGWORD "x )", ARG1, ARG2);
-   PRE_REG_READ2(long, "swapon", const void *, path, int, flags);
-}
-
-PRE(sys_swapoff)
-{
-   PRINT("sys_swapoff ( %#" FMT_REGWORD "x )", ARG1);
-   PRE_REG_READ1(long, "swapoff", const void *, path);
-}
-
 /* Very much MIPS specific */
 PRE(sys_cacheflush)
 {
@@ -648,8 +634,8 @@ static SyscallTableEntry syscall_main_table[] = {
    LINX_ (__NR_mount, sys_mount),
    LINX_ (__NR_mount_setattr, sys_mount_setattr),
    LINX_ (__NR_umount2, sys_umount),
-   PLAX_ (__NR_swapon, sys_swapon),
-   PLAX_ (__NR_swapoff, sys_swapoff),
+   LINX_ (__NR_swapon, sys_swapon),
+   LINX_ (__NR_swapoff, sys_swapoff),
    PLAX_ (__NR_reboot, sys_reboot),
    PLAX_ (__NR_sethostname, sys_sethostname),
    LINX_ (__NR_setdomainname, sys_setdomainname),
diff --git a/coregrind/m_syswrap/syswrap-nanomips-linux.c b/coregrind/m_syswrap/syswrap-nanomips-linux.c
index 94e4ce4d43..f1a0b3c593 100644
--- a/coregrind/m_syswrap/syswrap-nanomips-linux.c
+++ b/coregrind/m_syswrap/syswrap-nanomips-linux.c
@@ -383,8 +383,6 @@ DECL_TEMPLATE (mips_linux, sys_ptrace);
 DECL_TEMPLATE (mips_linux, sys_unshare);
 DECL_TEMPLATE (mips_linux, sys_reboot);
 DECL_TEMPLATE (mips_linux, sys_sethostname);
-DECL_TEMPLATE (mips_linux, sys_swapon);
-DECL_TEMPLATE (mips_linux, sys_swapoff);
 
 PRE(sys_mmap2)
 {
@@ -519,18 +517,6 @@ PRE(sys_sethostname)
    PRE_REG_READ2 (long, "sethostname", const void *, name, int, len);
 }
 
-PRE(sys_swapon)
-{
-   PRINT("sys_swapon ( %#lx, %#lx )", ARG1, ARG2);
-   PRE_REG_READ2(long, "swapon", const void *, path, int, flags);
-}
-
-PRE(sys_swapoff)
-{
-   PRINT("sys_swapoff ( %#lx )", ARG1);
-   PRE_REG_READ1(long, "swapoff", const void *, path);
-}
-
 #undef PRE
 #undef POST
 
@@ -765,8 +751,8 @@ static SyscallTableEntry syscall_main_table[] = {
    GENX_ (__NR_execve,                 sys_execve),
    PLAX_ (__NR_mmap2,                  sys_mmap2),
    LINX_ (__NR_fadvise64_64,           sys_fadvise64_64),
-   PLAX_ (__NR_swapon,                 sys_swapon),
-   PLAX_ (__NR_swapoff,                sys_swapoff),
+   LINX_ (__NR_swapon,                 sys_swapon),
+   LINX_ (__NR_swapoff,                sys_swapoff),
    GENXY (__NR_mprotect,               sys_mprotect),
    GENX_ (__NR_msync,                  sys_msync),
    GENX_ (__NR_mlock,                  sys_mlock),
diff --git a/coregrind/m_syswrap/syswrap-ppc32-linux.c b/coregrind/m_syswrap/syswrap-ppc32-linux.c
index 8bb0a04f56..b1390f3aba 100644
--- a/coregrind/m_syswrap/syswrap-ppc32-linux.c
+++ b/coregrind/m_syswrap/syswrap-ppc32-linux.c
@@ -716,7 +716,7 @@ static SyscallTableEntry syscall_table[] = {
 //.. 
    GENXY(__NR_readlink,          sys_readlink),          // 85
 //..    //   (__NR_uselib,            sys_uselib),            // 86 */Linux
-//..    //   (__NR_swapon,            sys_swapon),            // 87 */Linux
+   LINX_(__NR_swapon,            sys_swapon),            // 87 */Linux
 //..    //   (__NR_reboot,            sys_reboot),            // 88 */Linux
 //..    //   (__NR_readdir,           old_readdir),           // 89 -- superseded
 
@@ -750,7 +750,7 @@ static SyscallTableEntry syscall_table[] = {
 //..    //   (__NR_vm86old,           sys_vm86old),           // 113 x86/Linux-only
    GENXY(__NR_wait4,             sys_wait4),             // 114
 //.. 
-//..    //   (__NR_swapoff,           sys_swapoff),           // 115 */Linux 
+   LINX_(__NR_swapoff,           sys_swapoff),           // 115 */Linux 
    LINXY(__NR_sysinfo,           sys_sysinfo),           // 116
    LINXY(__NR_ipc,               sys_ipc),               // 117
    GENX_(__NR_fsync,             sys_fsync),             // 118
diff --git a/coregrind/m_syswrap/syswrap-ppc64-linux.c b/coregrind/m_syswrap/syswrap-ppc64-linux.c
index 0f51117289..4a08650846 100644
--- a/coregrind/m_syswrap/syswrap-ppc64-linux.c
+++ b/coregrind/m_syswrap/syswrap-ppc64-linux.c
@@ -704,7 +704,7 @@ static SyscallTableEntry syscall_table[] = {
 
    GENXY(__NR_readlink,          sys_readlink),           //  85
 // _____(__NR_uselib,            sys_uselib),             //  86
-// _____(__NR_swapon,            sys_swapon),             //  87
+   LINX_(__NR_swapon,            sys_swapon),             //  87
 // _____(__NR_reboot,            sys_reboot),             //  88
 // _____(__NR_readdir,           sys_readdir),            //  89
 
@@ -738,7 +738,7 @@ static SyscallTableEntry syscall_table[] = {
 // _____(__NR_vm86,              sys_vm86),               // 113
    GENXY(__NR_wait4,             sys_wait4),              // 114
 
-// _____(__NR_swapoff,           sys_swapoff),            // 115
+   LINX_(__NR_swapoff,           sys_swapoff),            // 115
    LINXY(__NR_sysinfo,           sys_sysinfo),            // 116
    LINXY(__NR_ipc,               sys_ipc),                // 117
    GENX_(__NR_fsync,             sys_fsync),              // 118
diff --git a/coregrind/m_syswrap/syswrap-riscv64-linux.c b/coregrind/m_syswrap/syswrap-riscv64-linux.c
index d806b92b8f..5a1ea2553d 100644
--- a/coregrind/m_syswrap/syswrap-riscv64-linux.c
+++ b/coregrind/m_syswrap/syswrap-riscv64-linux.c
@@ -530,6 +530,8 @@ static SyscallTableEntry syscall_main_table[] = {
    GENX_(__NR_execve, sys_execve),                                 /* 221 */
    PLAX_(__NR_mmap, sys_mmap),                                     /* 222 */
    GENX_(__NR_fadvise64, sys_ni_syscall),                          /* 223 */
+   LINX_(__NR_swapon, sys_swapon),                                 /* 224 */
+   LINX_(__NR_swapoff, sys_swapoff),                               /* 225 */
    GENXY(__NR_mprotect, sys_mprotect),                             /* 226 */
    GENX_(__NR_msync, sys_msync),                                   /* 227 */
    GENX_(__NR_mlock, sys_mlock),                                   /* 228 */
diff --git a/coregrind/m_syswrap/syswrap-s390x-linux.c b/coregrind/m_syswrap/syswrap-s390x-linux.c
index 643549c64f..acb4aefee8 100644
--- a/coregrind/m_syswrap/syswrap-s390x-linux.c
+++ b/coregrind/m_syswrap/syswrap-s390x-linux.c
@@ -515,7 +515,7 @@ static SyscallTableEntry syscall_table[] = {
 
    GENXY(__NR_readlink,  sys_readlink),                               // 85
 // ?????(__NR_uselib, ),                                              // 86
-// ?????(__NR_swapon, ),                                              // 87
+   LINX_(__NR_swapon,  sys_swapon),                                   // 87
 // ?????(__NR_reboot, ),                                              // 88
    GENX_(89, sys_ni_syscall), /* unimplemented (by the kernel) */     // 89
 
@@ -549,7 +549,7 @@ static SyscallTableEntry syscall_table[] = {
    GENX_(113, sys_ni_syscall), /* unimplemented (by the kernel) */    // 113
    GENXY(__NR_wait4,  sys_wait4),                                     // 114
 
-// ?????(__NR_swapoff, ),                                             // 115
+   LINX_(__NR_swapoff, sys_swapoff),                                  // 115
    LINXY(__NR_sysinfo,  sys_sysinfo),                                 // 116
    LINXY(__NR_ipc, sys_ipc),                                          // 117
    GENX_(__NR_fsync,  sys_fsync),                                     // 118
diff --git a/coregrind/m_syswrap/syswrap-x86-linux.c b/coregrind/m_syswrap/syswrap-x86-linux.c
index 42a69cb965..f697a9e19d 100644
--- a/coregrind/m_syswrap/syswrap-x86-linux.c
+++ b/coregrind/m_syswrap/syswrap-x86-linux.c
@@ -1261,7 +1261,7 @@ static SyscallTableEntry syscall_table[] = {
 //zz 
    GENXY(__NR_readlink,          sys_readlink),       // 85
 //zz    //   (__NR_uselib,            sys_uselib),         // 86 */Linux
-//zz    //   (__NR_swapon,            sys_swapon),         // 87 */Linux
+   LINX_(__NR_swapon,            sys_swapon),         // 87 */Linux
 //zz    //   (__NR_reboot,            sys_reboot),         // 88 */Linux
 //zz    //   (__NR_readdir,           old_readdir),        // 89 -- superseded
 //zz 
@@ -1295,7 +1295,7 @@ static SyscallTableEntry syscall_table[] = {
    PLAXY(__NR_vm86old,           sys_vm86old),        // 113 x86/Linux-only
    GENXY(__NR_wait4,             sys_wait4),          // 114
 //zz 
-//zz    //   (__NR_swapoff,           sys_swapoff),        // 115 */Linux 
+   LINX_(__NR_swapoff,           sys_swapoff),        // 115 */Linux 
    LINXY(__NR_sysinfo,           sys_sysinfo),        // 116
    LINXY(__NR_ipc,               sys_ipc),            // 117
    GENX_(__NR_fsync,             sys_fsync),          // 118

[Valgrind-developers] [valgrind] Update the LTP version in valgrind testsuite to v20250930

[Martin C. <mc...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=a04974ddcb51fbb24b1e44b1f9c98d0a11356cca

commit a04974ddcb51fbb24b1e44b1f9c98d0a11356cca
Author: Martin Cermak <mc...@re...>
Date:   Thu Oct 2 17:11:50 2025 +0200

    Update the LTP version in valgrind testsuite to v20250930
    
    Update the LTP version in valgrind testsuite to v20250930.
    All patches from auxprogs/ltp-patches were accepted by LTP
    upstream and included in the release, so these can now be
    dropped locally.
    
    https://bugs.kde.org/show_bug.cgi?id=510169

Diff:
---
 NEWS                                               |  1 +
 auxprogs/Makefile.am                               |  9 +--
 ...2-bit-powerpc-syscall-defs-don-t-leak-to-.patch | 41 ------------
 .../0002-Introduce-LTP_QUIET-env-var.patch         | 73 ----------------------
 ...-sure-the-scanf-address-format-is-at-leas.patch | 39 ------------
 5 files changed, 4 insertions(+), 159 deletions(-)

diff --git a/NEWS b/NEWS
index 7335f05b33..ca3f9fca4a 100644
--- a/NEWS
+++ b/NEWS
@@ -147,6 +147,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 509567  unhandled amd64-linux syscall: 443 (quotactl_fd)
 509642  Add missing ppc64-linux syswraps
 509643  Add missing s390x-linux syswraps
+510169  Update the LTP version in valgrind testsuite to 20250930
 
 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 0ce348ef54..d96e7fd0a9 100644
--- a/auxprogs/Makefile.am
+++ b/auxprogs/Makefile.am
@@ -20,10 +20,7 @@ LTP_FILTERS = \
 	filters/prctl10 \
 	filters/select03
 
-LTP_PATCHES = \
-	ltp-patches/0001-Make-sure-32-bit-powerpc-syscall-defs-don-t-leak-to-.patch \
-	ltp-patches/0002-Introduce-LTP_QUIET-env-var.patch \
-	ltp-patches/0003-mmap04-Make-sure-the-scanf-address-format-is-at-leas.patch
+LTP_PATCHES =
 
 EXTRA_DIST = \
 	docs/valgrind-listener-manpage.xml \
@@ -173,8 +170,8 @@ endif
 endif
 
 # Linux Test Project
-LTP_VERSION=20250530
-LTP_SHA256_SUM=27586ba78eac1e40cd422add2842f1ad70f09fea55da3bd6a25e10feb786d4f2
+LTP_VERSION=20250930
+LTP_SHA256_SUM=048fa4d69ddbe8a94aa15da9bdc85713ab07a0abbc3de2b8bdd9757644aef1e4
 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)
diff --git a/auxprogs/ltp-patches/0001-Make-sure-32-bit-powerpc-syscall-defs-don-t-leak-to-.patch b/auxprogs/ltp-patches/0001-Make-sure-32-bit-powerpc-syscall-defs-don-t-leak-to-.patch
deleted file mode 100644
index bd1fb1dd63..0000000000
--- a/auxprogs/ltp-patches/0001-Make-sure-32-bit-powerpc-syscall-defs-don-t-leak-to-.patch
+++ /dev/null
@@ -1,41 +0,0 @@
-From a90b2aac69028bd6b9e0fcc1e36760639b937b99 Mon Sep 17 00:00:00 2001
-From: Martin Cermak <mc...@re...>
-Date: Mon, 4 Aug 2025 21:46:52 +0200
-Subject: [PATCH] Make sure 32-bit powerpc syscall defs don't leak to 64-bit
- powerpc systems
-
-generate_syscalls.sh generates the syscalls.h header at the configure
-time.  At the moment, that header has a set of 32-bit syscalls defined
-with the __powerpc__ gate, plus another set of 64-bit syscalls defined
-with the __powerpc64__ gate.  For 32-bit powerpc systems that's fine.
-But for a 64-bit powerpc system this means that both sets of syscalls
-become defined, which isn't right.
-
-Thing is that on a 64-bit powerpc system, both __powerpc__ and
-__powerpc64__ are defined compiler macros, while on a 32-bit powerpc
-system, only the former is defined while the latter is not.
-
-That said, the correct gate for a 32-bit only powerpc code is:
-  #if defined(__powerpc__) && !defined(__powerpc64__)
-
-Without this patch, e.g. __NR_clock_gettime64 def leaks to
-64-bit powerpc systems, which is wrong.  This patch fixes it.
----
- include/lapi/syscalls/generate_syscalls.sh | 1 +
- 1 file changed, 1 insertion(+)
-
-diff --git a/include/lapi/syscalls/generate_syscalls.sh b/include/lapi/syscalls/generate_syscalls.sh
-index b17c72ddf..19f280dfb 100755
---- a/include/lapi/syscalls/generate_syscalls.sh
-+++ b/include/lapi/syscalls/generate_syscalls.sh
-@@ -78,6 +78,7 @@ while IFS= read -r arch; do
- 		parisc) echo "#ifdef __hppa__" ;;
- 		loongarch64) echo "#ifdef __loongarch__" ;;
- 		arm64) echo "#ifdef __aarch64__" ;;
-+		powerpc) echo "#if defined(__powerpc__) && !defined(__powerpc64__)" ;;
- 		*) echo "#ifdef __${arch}__" ;;
- 		esac
- 
--- 
-2.48.1
-
diff --git a/auxprogs/ltp-patches/0002-Introduce-LTP_QUIET-env-var.patch b/auxprogs/ltp-patches/0002-Introduce-LTP_QUIET-env-var.patch
deleted file mode 100644
index a77162bfca..0000000000
--- a/auxprogs/ltp-patches/0002-Introduce-LTP_QUIET-env-var.patch
+++ /dev/null
@@ -1,73 +0,0 @@
-From 183df3240f8e7ca38fbe2fd472c31c9417ae7eb2 Mon Sep 17 00:00:00 2001
-From: Martin Cermak <mc...@re...>
-Date: Tue, 16 Sep 2025 15:46:40 +0200
-Subject: [PATCH] Introduce LTP_QUIET env var
-
-Introduce LTP_QUIET env variable.  When set to  1 or y, it will
-suppress printing TCONF, TWARN, TINFO, and TDEBUG messages.
----
- lib/tst_test.c | 11 +++++++++++
- 1 file changed, 11 insertions(+)
-
-diff --git a/lib/tst_test.c b/lib/tst_test.c
-index 92872cc89..609a7b075 100644
---- a/lib/tst_test.c
-+++ b/lib/tst_test.c
-@@ -68,6 +68,7 @@ static int iterations = 1;
- static float duration = -1;
- static float timeout_mul = -1;
- static int reproducible_output;
-+static int quiet_output;
- 
- struct context {
- 	int32_t lib_pid;
-@@ -307,15 +308,19 @@ static void print_result(const char *file, const int lineno, int ttype,
- 		res = "TBROK";
- 	break;
- 	case TCONF:
-+		if (quiet_output) return;
- 		res = "TCONF";
- 	break;
- 	case TWARN:
-+		if (quiet_output) return;
- 		res = "TWARN";
- 	break;
- 	case TINFO:
-+		if (quiet_output) return;
- 		res = "TINFO";
- 	break;
- 	case TDEBUG:
-+		if (quiet_output) return;
- 		res = "TDEBUG";
- 	break;
- 	default:
-@@ -670,6 +675,7 @@ static void print_help(void)
- 	fprintf(stderr, "LTP_DEV_FS_TYPE          Filesystem used for testing (default: %s)\n", DEFAULT_FS_TYPE);
- 	fprintf(stderr, "LTP_ENABLE_DEBUG         Print debug messages (set 1 or y)\n");
- 	fprintf(stderr, "LTP_REPRODUCIBLE_OUTPUT  Values 1 or y discard the actual content of the messages printed by the test\n");
-+	fprintf(stderr, "LTP_QUIET                Values 1 or y will suppress printing TCONF, TWARN, TINFO, and TDEBUG messages\n");
- 	fprintf(stderr, "LTP_SINGLE_FS_TYPE       Specifies filesystem instead all supported (for .all_filesystems)\n");
- 	fprintf(stderr, "LTP_FORCE_SINGLE_FS_TYPE Testing only. The same as LTP_SINGLE_FS_TYPE but ignores test skiplist.\n");
- 	fprintf(stderr, "LTP_TIMEOUT_MUL          Timeout multiplier (must be a number >=1)\n");
-@@ -1361,6 +1367,7 @@ static void do_setup(int argc, char *argv[])
- {
- 	char *tdebug_env = getenv("LTP_ENABLE_DEBUG");
- 	char *reproducible_env = getenv("LTP_REPRODUCIBLE_OUTPUT");
-+	char *quiet_env = getenv("LTP_QUIET");
- 
- 	if (!tst_test)
- 		tst_brk(TBROK, "No tests to run");
-@@ -1391,6 +1398,10 @@ static void do_setup(int argc, char *argv[])
- 	    (!strcmp(reproducible_env, "1") || !strcmp(reproducible_env, "y")))
- 		reproducible_output = 1;
- 
-+	if (quiet_env &&
-+	    (!strcmp(quiet_env, "1") || !strcmp(quiet_env, "y")))
-+		quiet_output = 1;
-+
- 	assert_test_fn();
- 
- 	TCID = tcid = get_tcid(argv);
--- 
-2.48.1
-
diff --git a/auxprogs/ltp-patches/0003-mmap04-Make-sure-the-scanf-address-format-is-at-leas.patch b/auxprogs/ltp-patches/0003-mmap04-Make-sure-the-scanf-address-format-is-at-leas.patch
deleted file mode 100644
index 7956999b33..0000000000
--- a/auxprogs/ltp-patches/0003-mmap04-Make-sure-the-scanf-address-format-is-at-leas.patch
+++ /dev/null
@@ -1,39 +0,0 @@
-From 6c3a6a6f625b58e8dc611cc12bc6015dc8dd5b32 Mon Sep 17 00:00:00 2001
-From: Mark Wielaard <ma...@kl...>
-Date: Thu, 18 Sep 2025 17:16:05 +0200
-Subject: [PATCH] mmap04: Make sure the scanf address format is at least 8 hex
- chars
-
-The addresses in /proc/self/maps are at least 8 hex chars. Zeros are
-added to the front of the address when shorter (both on 32bit and
-64bit systems).
-
-Under valgrind the mmaps used in kernel/syscalls/mmap/mmap04.c come
-out very low in the address space and might be shorter than 8 hex
-chars. This causes the scanf to fail:
-mmap04.c:62: TBROK: Expected 1 conversions got 0 FILE '/proc/self/maps'
-
-Fix this by using "%08" PRIxPTR when creating the fmt used.
-
-Signed-off-by: Mark Wielaard <ma...@kl...>
----
- testcases/kernel/syscalls/mmap/mmap04.c | 3 ++-
- 1 file changed, 2 insertions(+), 1 deletion(-)
-
-diff --git a/testcases/kernel/syscalls/mmap/mmap04.c b/testcases/kernel/syscalls/mmap/mmap04.c
-index 4a050b7b50da..5b28180df29b 100644
---- a/testcases/kernel/syscalls/mmap/mmap04.c
-+++ b/testcases/kernel/syscalls/mmap/mmap04.c
-@@ -58,7 +58,8 @@ static void run(unsigned int i)
- 
- 	addr2 = SAFE_MMAP(addr1 + pagesize, pagesize, tc->prot, tc->flags | MAP_FIXED, -1, 0);
- 
--	sprintf(fmt, "%" PRIxPTR "-%%*x %%s", (uintptr_t)addr2);
-+	/* A /proc/self/maps address is at least 8 hex (left zero padded) */
-+	sprintf(fmt, "%08" PRIxPTR "-%%*x %%s", (uintptr_t)addr2);
- 	SAFE_FILE_LINES_SCANF("/proc/self/maps", fmt, perms);
- 
- 	if (!strcmp(perms, tc->exp_perms)) {
--- 
-2.51.0
-

[Valgrind-developers] Preparing for Valgrind 3.26.0 (RC on Oct 17, final Oct 24)

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

Hi Valgrind hackers,

Normally we do a major releases in April and October. We did do a minor
3.25.1 release in May. And the VALGRIND_3_25_BRANCH does contain a
couple more backported fixes. But the main branch has 350+ more commits
so it is time to prepare for a new major release.

I am proposing to do a release on Friday October 24. With an RC1 a
week before, Friday October 17. And possibly an RC2 in between if we
find some big issue for which we like more testing.

There are a couple of things I like to add before RC1:

- Update to GPLv3+.
  We discussed this on the list and then with some core developers
  asking for advice from the FSF Legal and Compliance team.
  I'll post a patch tomorrow with the full rational, so we can
  discuss it before applying for RC1.

- Don't warn about fcntl F_GETFD with --track-fds
  https://bugs.kde.org/show_bug.cgi?id=510436
  --track-fds=bad is great, but Paul correctly pointed out
  that there has to be a way for an application to explicitly
  check for bad fds that doesn't generate an error.

- Removing glibc extra syscall frames might crash
  https://bugs.kde.org/show_bug.cgi?id=507188
  This seems to only happen on i386 debian, but it a generic issue
  with syscall frame filtering.

  And if I have time I also like to finish the removal of artificial
  frames from the backtraces. There is already support for tracking
  these artificial frams since commit f7dccaab11b8 "Rewrite DWARF
  inlined subroutine handling to work cross CU". But they aren't yet
  filtered.

- On Fedora 43 (beta) for amd64 we get some make regtest failures
  I haven't really investigated yet, but on f42 we get zero failures,
  so it would be nice to see what updates caused the new failures.

Please let me know what you believe are blockers for the release or
fixes you really would like to see in.

Thanks,

Mark

[Valgrind-developers] [valgrind] s390x: Make IBM z17 known to Valgrind

[Andreas A. <ar...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=32fcdbfafd623569df800d2a1fa351c11f10a128

commit 32fcdbfafd623569df800d2a1fa351c11f10a128
Author: Andreas Arnez <ar...@li...>
Date:   Thu Oct 9 13:42:39 2025 +0200

    s390x: Make IBM z17 known to Valgrind
    
    Make the IBM z17 machine model 9175 known to Valgrind.  Also add the
    expected output of the s390x-specific "ecag" test case on an IBM z17, so
    the test case succeeds on that system.

Diff:
---
 VEX/pub/libvex.h                     |  3 ++-
 coregrind/m_machine.c                |  1 +
 none/tests/s390x/Makefile.am         |  2 +-
 none/tests/s390x/ecag.stdout.exp-z17 | 21 +++++++++++++++++++++
 tests/s390x_features.c               |  1 +
 5 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/VEX/pub/libvex.h b/VEX/pub/libvex.h
index e580160d30..bfa94d033f 100644
--- a/VEX/pub/libvex.h
+++ b/VEX/pub/libvex.h
@@ -158,7 +158,8 @@ typedef
 #define VEX_S390X_MODEL_Z14_ZR1  15
 #define VEX_S390X_MODEL_Z15      16
 #define VEX_S390X_MODEL_Z16      17
-#define VEX_S390X_MODEL_UNKNOWN  18     /* always last in list */
+#define VEX_S390X_MODEL_Z17      18
+#define VEX_S390X_MODEL_UNKNOWN  19     /* always last in list */
 #define VEX_S390X_MODEL_MASK     0x3F
 
 #define VEX_HWCAPS_S390X_LDISP (1<<6)   /* Long-displacement facility */
diff --git a/coregrind/m_machine.c b/coregrind/m_machine.c
index eaa18b0b61..bba498f989 100644
--- a/coregrind/m_machine.c
+++ b/coregrind/m_machine.c
@@ -624,6 +624,7 @@ static UInt VG_(get_machine_model)(void)
       { "8562", VEX_S390X_MODEL_Z15 },
       { "3931", VEX_S390X_MODEL_Z16 },
       { "3932", VEX_S390X_MODEL_Z16 },
+      { "9175", VEX_S390X_MODEL_Z17 },
    };
 
    Int    model, n, fh;
diff --git a/none/tests/s390x/Makefile.am b/none/tests/s390x/Makefile.am
index 22b8128bba..dda923acc6 100644
--- a/none/tests/s390x/Makefile.am
+++ b/none/tests/s390x/Makefile.am
@@ -48,7 +48,7 @@ EXTRA_DIST = \
 	bfp-XxC.vgtest bfp-XxC.stderr.exp bfp-XxC.post.exp \
 	ecag.stdout.exp-z10ec ecag.stdout.exp-z196 ecag.stdout.exp-zec12 \
 	ecag.stdout.exp-z13 ecag.stdout.exp-z14 ecag.stdout.exp-z15 \
-	ecag.stdout.exp-z16 \
+	ecag.stdout.exp-z16 ecag.stdout.exp-z17 \
 	op00.stderr.exp op00.vgtest \
 	dfp-XxC.vgtest dfp-XxC.stderr.exp dfp-XxC.post.exp \
 	dfp-XiC.vgtest dfp-XiC.stderr.exp dfp-XiC.post.exp \
diff --git a/none/tests/s390x/ecag.stdout.exp-z17 b/none/tests/s390x/ecag.stdout.exp-z17
new file mode 100644
index 0000000000..4708b2a835
--- /dev/null
+++ b/none/tests/s390x/ecag.stdout.exp-z17
@@ -0,0 +1,21 @@
+L1 topology: separate data and instruction; private
+L1 cache line size data: 256
+L1 cache line size insn: 256
+L1 total cachesize data: 131072
+L1 total cachesize insn: 131072
+L1 set. assoc.     data: 8
+L1 set. assoc.     insn: 8
+L2 topology: unified data and instruction; private
+L2 cache line size data: 256
+L2 cache line size insn: 256
+L2 total cachesize data: 37748736
+L2 total cachesize insn: 37748736
+L2 set. assoc.     data: 18
+L2 set. assoc.     insn: 18
+L3 topology: unified data and instruction; shared
+L3 cache line size data: 256
+L3 cache line size insn: 256
+L3 total cachesize data: 377487360
+L3 total cachesize insn: 377487360
+L3 set. assoc.     data: 180
+L3 set. assoc.     insn: 180
diff --git a/tests/s390x_features.c b/tests/s390x_features.c
index 507f3ab2f8..e0d4de76ce 100644
--- a/tests/s390x_features.c
+++ b/tests/s390x_features.c
@@ -120,6 +120,7 @@ model_info models[] = {
    { "8562", "z15"    },
    { "3931", "z16"    },
    { "3932", "z16"    },
+   { "9175", "z17"    },
 };

[Valgrind-developers] [valgrind] Document that --vex-guest-chase=not ought to be used when doing IR injection.

[Florian K. <fk...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=0329ab2021727242f31adfd665f602a8a6163aa8

commit 0329ab2021727242f31adfd665f602a8a6163aa8
Author: Florian Krohm <fl...@ei...>
Date:   Tue Oct 7 12:39:38 2025 +0000

    Document that --vex-guest-chase=not ought to be used when doing IR injection.
    
    Fixes https://bugs.kde.org/show_bug.cgi?id=506453

Diff:
---
 NEWS                                          | 1 +
 VEX/priv/ir_inject.c                          | 4 ++++
 memcheck/tests/vbit-test/vbit-test-sec.vgtest | 2 +-
 memcheck/tests/vbit-test/vbit-test.vgtest     | 2 +-
 4 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/NEWS b/NEWS
index bdf175eac8..7335f05b33 100644
--- a/NEWS
+++ b/NEWS
@@ -987,6 +987,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 451843  valgrind fails to start on a FreeBSD system which enforces W^X
 495483  Control building documentation via BUILD_DOCS
 506211  Constant folding improvements
+506453  Unexpected behaviour with IR injection and vex-guest-chase=yes
 509157  riscv64: Shift instructions can behave wrong
 
 To see details of a given bug, visit
diff --git a/VEX/priv/ir_inject.c b/VEX/priv/ir_inject.c
index 04228538f1..b447f6e946 100644
--- a/VEX/priv/ir_inject.c
+++ b/VEX/priv/ir_inject.c
@@ -27,6 +27,10 @@
    The GNU General Public License is contained in the file COPYING.
 */
 
+/* !!! When running valgrind on applications that use IR injection
+   !!! --vex-guest-chase=no should be given on the command line. This
+   !!! avoids that vex_inject_ir is called speculatively.
+*/
 #include "libvex_basictypes.h"
 #include "libvex_ir.h"
 #include "libvex.h"
diff --git a/memcheck/tests/vbit-test/vbit-test-sec.vgtest b/memcheck/tests/vbit-test/vbit-test-sec.vgtest
index 2d3c9389da..8d84e7e90b 100644
--- a/memcheck/tests/vbit-test/vbit-test-sec.vgtest
+++ b/memcheck/tests/vbit-test/vbit-test-sec.vgtest
@@ -1,3 +1,3 @@
 prog: vbit-test-sec
 prereq: test -x vbit-test-sec
-vgopts: -q --expensive-definedness-checks=yes
+vgopts: -q --expensive-definedness-checks=yes --vex-guest-chase=no
diff --git a/memcheck/tests/vbit-test/vbit-test.vgtest b/memcheck/tests/vbit-test/vbit-test.vgtest
index a05890566b..198c62351c 100644
--- a/memcheck/tests/vbit-test/vbit-test.vgtest
+++ b/memcheck/tests/vbit-test/vbit-test.vgtest
@@ -1,2 +1,2 @@
 prog: vbit-test
-vgopts: -q --expensive-definedness-checks=yes
+vgopts: -q --expensive-definedness-checks=yes --vex-guest-chase=no

[Valgrind-developers] [valgrind] FreeBSD regtest: update setcred expected from arm64

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

https://sourceware.org/cgit/valgrind/commit/?id=e55d0db975c69f0f4c0feae78d64fc8a7196351f

commit e55d0db975c69f0f4c0feae78d64fc8a7196351f
Author: Paul Floyd <pj...@wa...>
Date:   Sun Sep 28 03:39:45 2025 +0000

    FreeBSD regtest: update setcred expected from arm64
    
    Still have an unexpected extra error on amd64

Diff:
---
 memcheck/tests/freebsd/setcred.stderr.exp | 22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/memcheck/tests/freebsd/setcred.stderr.exp b/memcheck/tests/freebsd/setcred.stderr.exp
index 629ceacf84..8766361c09 100644
--- a/memcheck/tests/freebsd/setcred.stderr.exp
+++ b/memcheck/tests/freebsd/setcred.stderr.exp
@@ -1,29 +1,37 @@
 Syscall param setcred(wcred) points to unaddressable byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:23)
+   by 0x........: main (setcred.cpp:28)
  Address 0x........ is not stack'd, malloc'd or (recently) free'd
 
 Syscall param setcred(flags) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:26)
+   by 0x........: main (setcred.cpp:33)
 
 Syscall param setcred(wcred) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:26)
+   by 0x........: main (setcred.cpp:33)
 
 Syscall param setcred(size) contains uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:26)
+   by 0x........: main (setcred.cpp:33)
 
 Syscall param setcred(wcred) points to unaddressable byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:26)
+   by 0x........: main (setcred.cpp:33)
  Address 0x........ is not stack'd, malloc'd or (recently) free'd
 
+Syscall param setcred(flags) contains uninitialised byte(s)
+   at 0x........: setcred (in /...libc...)
+   by 0x........: main (setcred.cpp:38)
+
+Syscall param setcred(size) contains uninitialised byte(s)
+   at 0x........: setcred (in /...libc...)
+   by 0x........: main (setcred.cpp:43)
+
 Syscall param setcred(wcred) points to uninitialised byte(s)
    at 0x........: setcred (in /...libc...)
-   by 0x........: main (setcred.cpp:31)
+   by 0x........: main (setcred.cpp:50)
  Address 0x........ is 0 bytes inside a block of size 48 alloc'd
    at 0x........: ...operator new... (vg_replace_malloc.c:...)
-   by 0x........: main (setcred.cpp:28)
+   by 0x........: main (setcred.cpp:47)

[Valgrind-developers] [valgrind] FreeBSD setcred syscall: use arg3 for size

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

https://sourceware.org/cgit/valgrind/commit/?id=ea1ec47ce1a8f29bf01412869b6a07e5ee6a2976

commit ea1ec47ce1a8f29bf01412869b6a07e5ee6a2976
Author: Paul Floyd <pj...@wa...>
Date:   Mon Oct 6 10:55:34 2025 +0200

    FreeBSD setcred syscall: use arg3 for size
    
    Not really our job to check that thwe size is right.
    Also revamp the secred testcase.

Diff:
---
 coregrind/m_syswrap/syswrap-freebsd.c |  2 +-
 memcheck/tests/freebsd/setcred.cpp    | 33 +++++++++++++++++++++++++++------
 2 files changed, 28 insertions(+), 7 deletions(-)

diff --git a/coregrind/m_syswrap/syswrap-freebsd.c b/coregrind/m_syswrap/syswrap-freebsd.c
index eb8a160586..2c3bffa1c2 100644
--- a/coregrind/m_syswrap/syswrap-freebsd.c
+++ b/coregrind/m_syswrap/syswrap-freebsd.c
@@ -7154,7 +7154,7 @@ PRE(sys_setcred)
 {
    PRINT("sys_setcred(%" FMT_REGWORD "u, %#" FMT_REGWORD "x, %" FMT_REGWORD "u)", ARG1, ARG2, ARG3);
    PRE_REG_READ3(int, "setcred", u_int, flags, const struct setcred*, wcred, size_t, size);
-   PRE_MEM_READ("setcred(wcred)", ARG2, sizeof(struct vki_setcred));
+   PRE_MEM_READ("setcred(wcred)", ARG2, ARG3);
 }
 
 // SYS_exterrctl 592
diff --git a/memcheck/tests/freebsd/setcred.cpp b/memcheck/tests/freebsd/setcred.cpp
index 464de08efe..619d1959f1 100644
--- a/memcheck/tests/freebsd/setcred.cpp
+++ b/memcheck/tests/freebsd/setcred.cpp
@@ -1,6 +1,8 @@
 #include <sys/ucred.h>
 #include <cstring>
 #include <cstdlib>
+#include <cerrno>
+#include <cassert>
 
 static long x0;
 
@@ -10,25 +12,44 @@ int main()
    x0 = px[0];
    struct setcred cred1;
    struct setcred* cred2;
-   int flags1{0};
-   int flags2;
+   int flags1{SETCREDF_RUID};
+   int flags2{SETCREDF_SUPP_GROUPS};
    size_t size1{sizeof(cred1)};
+   int ret;
 
    std::memset(&cred1, 250, sizeof(cred1));
 
    // needs to be root to work correctly
-   setcred(flags1, &cred1, size1);
+   ret = setcred(flags1, &cred1, size1);
+   assert(ret == -1);
+   assert(errno == EPERM);
 
    // not accessible
-   setcred(flags1, nullptr, size1);
+   ret = setcred(flags1, nullptr, size1);
+   assert(ret == -1);
+   assert(errno == EFAULT);
 
    // uninit
-   setcred(flags2, (struct setcred*)x0, size1+x0);
+   ret = setcred(flags2+x0, (struct setcred*)x0, size1+x0);
+   assert(ret == -1);
+   assert(errno == EFAULT);
+
+   // invalid flags
+   ret = setcred(9999+x0, &cred1, sizeof(cred1));
+   assert(ret == -1);
+   assert(errno == EINVAL);
+
+   // invalid size
+   ret = setcred(flags1, &cred1, 3+x0);
+   assert(ret == -1);
+   assert(errno == EINVAL);
 
    cred2 = new struct setcred;
 
    // uninit memory
-   setcred(flags1, cred2, size1);
+   ret = setcred(flags1, cred2, size1);
+   assert(ret == -1);
+   assert(errno == EPERM);
 
    delete cred2;
    free(px);

[Valgrind-developers] [valgrind] FreeBSD vgdb: ensure stack is 16byte aligned for invoker

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

https://sourceware.org/cgit/valgrind/commit/?id=c6092d4230fd2533374555213632e77b95a33a5d

commit c6092d4230fd2533374555213632e77b95a33a5d
Author: Paul Floyd <pj...@wa...>
Date:   Sun Sep 28 02:11:58 2025 +0000

    FreeBSD vgdb: ensure stack is 16byte aligned for invoker
    
    I haven't seen any issues but there's no harm in being certain.

Diff:
---
 coregrind/vgdb-invoker-freebsd.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/coregrind/vgdb-invoker-freebsd.c b/coregrind/vgdb-invoker-freebsd.c
index 607e059919..4a89fb37dd 100644
--- a/coregrind/vgdb-invoker-freebsd.c
+++ b/coregrind/vgdb-invoker-freebsd.c
@@ -561,6 +561,7 @@ Bool invoker_invoke_gdbserver (pid_t pid)
       reg_mod.r_rip = shared64->invoke_gdbserver;
 #elif defined(VGA_arm64)
       reg_mod.x[0] = check;
+      sp &= ~0xf; // keep the stack aligned on 16 bytes ...
       reg_mod.sp = sp;
       reg_mod.elr = shared64->invoke_gdbserver;
       /* put NULL return address in Link Register */

[Valgrind-developers] [valgrind] getdents* syswrap: only check proc fd/fdinfo on Linux and Solaris

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

https://sourceware.org/cgit/valgrind/commit/?id=3bcda36cf8f02c87ff900a0e06984b7e46a34e4f

commit 3bcda36cf8f02c87ff900a0e06984b7e46a34e4f
Author: Paul Floyd <pj...@wa...>
Date:   Fri Oct 3 20:54:15 2025 +0200

    getdents* syswrap: only check proc fd/fdinfo on Linux and Solaris
    
    Keep check for __NR_getdents and use a check for __NR_getdents64

Diff:
---
 coregrind/m_syswrap/syswrap-generic.c | 24 ++++++++++--------------
 1 file changed, 10 insertions(+), 14 deletions(-)

diff --git a/coregrind/m_syswrap/syswrap-generic.c b/coregrind/m_syswrap/syswrap-generic.c
index 296ebdd910..0d13cea5e9 100644
--- a/coregrind/m_syswrap/syswrap-generic.c
+++ b/coregrind/m_syswrap/syswrap-generic.c
@@ -3994,7 +3994,7 @@ PRE(sys_getdents)
    PRE_MEM_WRITE( "getdents(dirp)", ARG2, ARG3 );
 }
 
-#if !defined(VGO_freebsd) // Darwin as well?
+#if defined(VGO_linux) || defined(VGO_solaris)
 
 /* Check if the given file descriptor points to a /proc/PID/fd or /proc/PID/fdinfo directory.
    Returns True if it's a directory we should filter Valgrind FDs from. */
@@ -4049,7 +4049,7 @@ static Bool should_keep_fd_entry(const HChar *name)
 
 /* Make sure we really need the proc filtering using (32bit) getdents,
    which not every linux arch implements.  */
-#if defined(VGO_linux) && defined(__NR_getdents)
+#if defined(__NR_getdents)
 
 /* Filter and compact dirent entries */
 static SizeT filter_dirent_entries(struct vki_dirent *dirp, SizeT orig_size)
@@ -4073,7 +4073,7 @@ static SizeT filter_dirent_entries(struct vki_dirent *dirp, SizeT orig_size)
 
    return new_size;
 }
-#endif /* defined(VGO_linux) && defined(__NR_getdents) */
+#endif /* defined(__NR_getdents) */
 
 /* Filter and compact dirent64 entries */
 static SizeT filter_dirent64_entries(struct vki_dirent64 *dirp, SizeT orig_size)
@@ -4098,11 +4098,9 @@ static SizeT filter_dirent64_entries(struct vki_dirent64 *dirp, SizeT orig_size)
    return new_size;
 }
 
-#endif
-
 /* Make sure we really need the proc filtering using (32bit) getdents,
    which not every linux arch implements.  */
-#if defined(VGO_linux) && defined(__NR_getdents)
+#if defined(__NR_getdents)
 
 /* Filter out Valgrind's internal file descriptors from getdents results with refill capability.
    When entries are filtered out, attempts to read more entries to avoid empty results.
@@ -4132,9 +4130,7 @@ static SizeT filter_valgrind_fds_from_getdents_with_refill(Int fd, struct vki_di
 
    return new_size;
 }
-#endif /* defined(VGO_linux) && defined(__NR_getdents) */
-
-#if !defined(VGO_freebsd) // Darwin as well?
+#endif /* defined(__NR_getdents) */
 
 /* Filter out Valgrind's internal file descriptors from getdents64 results with refill capability.
    Same logic as getdents version but for 64-bit dirent structures.
@@ -4165,7 +4161,7 @@ static SizeT filter_valgrind_fds_from_getdents64_with_refill(Int fd, struct vki_
    return new_size;
 }
 
-#endif
+#endif /* defined(VGO_linux) || defined(VGO_solaris) */
 
 POST(sys_getdents)
 {
@@ -4175,7 +4171,7 @@ POST(sys_getdents)
       
       /* Make sure we really need the proc filtering using (32bit) getdents,
          which not every linux arch implements.  */
-#if defined(VGO_linux) && defined(__NR_getdents)
+#if (defined(VGO_linux) || defined(VGO_solaris)) && defined(__NR_getdents)
 
       /* Only filter Valgrind FDs when listing /proc/PID/fd or /proc/PID/fdinfo directories */
       if (is_proc_fd_directory(ARG1)) {
@@ -4188,13 +4184,13 @@ POST(sys_getdents)
          if (result_size != RES)
             SET_STATUS_Success(result_size);
       }
-#endif /* defined(VGO_linux) && defined(__NR_getdents) */
+#endif /* (defined(VGO_linux) || defined(VGO_solaris)) && defined(__NR_getdents) */
       
       POST_MEM_WRITE( ARG2, result_size );
    }
 }
 
-#if !defined(VGO_freebsd)
+#if defined(__NR_getdents64)
 
 PRE(sys_getdents64)
 {
@@ -4229,7 +4225,7 @@ POST(sys_getdents64)
    }
 }
 
-#endif
+#endif /* defined(__NR_getdents64) */
 
 PRE(sys_getgroups)
 {

[Valgrind-developers] [valgrind] Add BZ 495483 to NEWS.

[Florian K. <fk...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=c0e022d1a8a973c4872f497b79ccb8e0a51adae9

commit c0e022d1a8a973c4872f497b79ccb8e0a51adae9
Author: Florian Krohm <fl...@ei...>
Date:   Fri Oct 3 16:31:59 2025 +0000

    Add BZ 495483 to NEWS.
    
    Darn I keep forgetting this....

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

diff --git a/NEWS b/NEWS
index e4722a7293..bdf175eac8 100644
--- a/NEWS
+++ b/NEWS
@@ -985,6 +985,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 451626  Syscall param bpf(attr->raw_tracepoint.name) points to unaddressable byte(s)
 451827  [ppc64le] VEX temporary storage exhausted with several vbpermq instructions
 451843  valgrind fails to start on a FreeBSD system which enforces W^X
+495483  Control building documentation via BUILD_DOCS
 506211  Constant folding improvements
 509157  riscv64: Shift instructions can behave wrong

[Valgrind-developers] [valgrind] Control building documentation (BZ 495483)

[Florian K. <fk...@so...>]

https://sourceware.org/cgit/valgrind/commit/?id=9c0c0e41d8b72825606da0db897a419c59658a56

commit 9c0c0e41d8b72825606da0db897a419c59658a56
Author: Florian Krohm <fl...@ei...>
Date:   Fri Oct 3 16:26:55 2025 +0000

    Control building documentation   (BZ 495483)
    
    Introduce Makefile variable BUILD_DOCS with these possible values:
    none - does not build any documentation
    all  - builds all documentation
    html - builds HTML docs but skips building PDFs
    
    BUILD_ALL_DOCS is still recognised for backward compatibility and
    is mapped to BUILD_DOCS like so:
    If not specified   --> BUILD_DOCS=all
    BUILD_ALL_DOCS=yes --> BUILD_DOCS=all
    BUILD_ALL_DOCS=no  --> BUILD_DOCS=html
    
    README_DEVELOPERS adjusted.
    
    Fixes https://bugs.kde.org/show_bug.cgi?id=495483

Diff:
---
 README_DEVELOPERS |  2 +-
 docs/Makefile.am  | 35 ++++++++++++++++++++++++++++-------
 2 files changed, 29 insertions(+), 8 deletions(-)

diff --git a/README_DEVELOPERS b/README_DEVELOPERS
index 22e229bd49..5d9fb823f9 100644
--- a/README_DEVELOPERS
+++ b/README_DEVELOPERS
@@ -37,7 +37,7 @@ will also attempt to build the documentation.
 If you only want to test whether the generated tarball is complete and runs
 regression tests successfully, building documentation is not needed.
 
-  make dist BUILD_ALL_DOCS=no
+  make dist BUILD_DOCS=none
 
 If you insist on building documentation some embarrassing instructions
 can be found in docs/README.
diff --git a/docs/Makefile.am b/docs/Makefile.am
index 8672e8dd21..7e5ed389c5 100644
--- a/docs/Makefile.am
+++ b/docs/Makefile.am
@@ -4,10 +4,26 @@
 ## tools to do so is impractical / too difficult
 ##-------------------------------------------------------------
 
-# Comment out the next line to skip building print docs.  The default
-# is not to skip building print docs.  Note, after changing it
-# you need to re-run autogen.sh and configure to make it take effect.
-BUILD_ALL_DOCS=yes
+# Building documentation is a pain. It takes longish and requires certain
+# tools you may not have installed.
+# The variable BUILD_DOCS allows you to control what documentation gets
+# built:
+#
+#  BUILD_DOCS=all     builds all documentation
+#  BUILD_DOCS=html    builds HTML docs but skips building PDFs
+#  BUILD_DOCS=none    does not build any documentation
+#
+# Use it on the command line. E.g.:  make dist BUILD_DOCS=none
+#
+ifndef BUILD_ALL_DOCS
+   BUILD_DOCS ?= all
+ else
+   ifeq "$(BUILD_ALL_DOCS)" "yes"
+     BUILD_DOCS ?= all
+   else
+     BUILD_DOCS ?= html
+   endif
+ endif
 
 # Whether to run xmlto pdf --with-fop
 # This is de fault, set to empty to use the default fo to pdf
@@ -281,7 +297,7 @@ install-data-hook:
 			cp $$f $(DESTDIR)$(mandir)/man1; \
 		fi \
 	done
- ifeq ($(BUILD_ALL_DOCS),yes)
+ ifeq ($(BUILD_DOCS),all)
 	set -e; \
 	if test -r index.pdf ; then \
 		mkdir -p $(DESTDIR)$(datadir)/doc/valgrind; \
@@ -297,18 +313,23 @@ install-data-hook:
 # This is done at 'make dist' time.  It builds the html docs, print
 # docs and man pages and copies them into the docs/ directory in the
 # tarball.
- ifeq ($(BUILD_ALL_DOCS),yes)
+ ifeq ($(BUILD_DOCS),none)
+dist-hook:
+
+ else
+   ifeq ($(BUILD_DOCS),all)
 dist-hook: FAQ.txt html-docs man-pages print-docs
 	cp -r html $(distdir)
 	cp FAQ.txt $(distdir)/..
 	cp *.1 $(distdir)
 	cp print/index.pdf $(distdir)
 	cp print/index.ps $(distdir)
- else
+   else
 dist-hook: FAQ.txt html-docs man-pages
 	cp -r html $(distdir)
 	cp FAQ.txt $(distdir)/..
 	cp *.1 $(distdir)
+   endif
  endif
 
 distclean-local:

Re: [Valgrind-developers] [valgrind] none/tests/Makefile.am: Add s390x/disasm-test to DIST_SUBDIRS

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

Hi Mark,

thanks for taking the time.

On 02.10.25 22:53, Mark Wielaard wrote:
> 
> That looks nice (I do have a wording nitpick, but feel free to ignore
> if you feel no-print is more clear than html).

Agreed.

> I wouldn't integrate the make dist with the post-regtest-checks
> though.  It wouldn't have caught the issue because it only generates
> the tar ball.  make distcheck would have caught it, but I think that
> make distcheck should be a separate step/check (it does multiple
> builds and takes a long time even with BUILD_DOCS=none).

True enough. I am going to check this in with your suggested changes.

    Florian

Re: [Valgrind-developers] [valgrind] FreeBSD: fix build, FreeBSD has no dirent64 or getdent64

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

On 10/3/25 08:36, Paul Floyd via Valgrind-developers wrote:
>
>
> Neither FreeBSD nor macOS have dirent64/getdents64. Furthemore neither 
> has /proc and so they don't need any fd filtering. FreeBSD does have 
> freebsd11_dirent/freebsd11_getdents (which is also the kernel 
> interface) just to confuse matters slightly.
>
> I'm trying to work out what illumos and Solaris have. Certainly they 
> have /proc and also need fd filtering. illumos has dirent (native), 
> dirent32 (x86) and dirent64 (kernel amd64 and x86 
> _LARGEFILE64_SOURCE). illumos headers have syscall numbers for 
> getdents and getdents64 but getdents64 seems to produce a SIGSYS. I'll 
> see what Solaris 11.4 does and try to work out why illumos has 
> getdents64 in the headers.
>

On Illumos, I see 3 combinations

amd64 -> getdents uses syscall 81 SYS_getdents

x86 -> getdents uses syscall 81 again

x86 largefile -> getdents uses syscall 213 SYS_getdents64

I guess Solaris 11.4 is the same, but for amd64 only.


A+

Paul

Re: [Valgrind-developers] [valgrind] FreeBSD: fix build, FreeBSD has no dirent64 or getdent64

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

On 10/2/25 22:57, Mark Wielaard wrote:
> Hi,
>
> On Thu, Oct 02, 2025 at 09:30:11AM -0700, Bart Van Assche via Valgrind-developers wrote:
>> On 10/1/25 10:48 PM, Paul Floyd wrote:
>>> +#if !defined(VGO_freebsd) // Darwin as well?
>> Isn't adding this type of #ifdefs a software anti-pattern? Please
>> consider adding a configure test for dirent64() and getdent64() and to
>> test the appropriate HAVE_* macro here.
> Note that the other checks for this feature use:
>
> #if defined(VGO_linux) && defined(__NR_getdents)
>
> It would be good to use the same guards everywhere.
>
> Were these guards placed wrongly? What was the actual build failure on
> freebsd?


Neither FreeBSD nor macOS have dirent64/getdents64. Furthemore neither 
has /proc and so they don't need any fd filtering. FreeBSD does have 
freebsd11_dirent/freebsd11_getdents (which is also the kernel interface) 
just to confuse matters slightly.

I'm trying to work out what illumos and Solaris have. Certainly they 
have /proc and also need fd filtering. illumos has dirent (native), 
dirent32 (x86) and dirent64 (kernel amd64 and x86 _LARGEFILE64_SOURCE). 
illumos headers have syscall numbers for getdents and getdents64 but 
getdents64 seems to produce a SIGSYS. I'll see what Solaris 11.4 does 
and try to work out why illumos has getdents64 in the headers.

A+

Paul