valgrind-developers — Discussion list for developers of valgrind

Re: [Valgrind-developers] [valgrind] regtest: turn off exec stack warning

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

Hi Philippe,

On Sun, Dec 22, 2024 at 12:29:34PM +0100, Philippe Waroquiers via Valgrind-developers wrote:
> The below change causes a build failure on the gcc110 compile farm
> 
> valgrind revision: valgrind-3.25.0.GIT-544d41e525-20241221
> C compiler:        gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)
> GDB:               GNU gdb (GDB) Red Hat Enterprise Linux 7.6.1-120.el7
> Assembler:         GNU assembler version 2.27-44.base.el7_9.1
> C library:         GNU C Library (GNU libc) stable release version 2.17
> uname -mrs:        Linux 3.10.0-1160.105.1.el7.ppc64 ppc64
> Vendor version:    CentOS Linux 7 (AltArch)
> [...]
> /usr/bin/ld: unrecognized option '--no-warn-execstack'

I saw Paul already added a configure check to see if the linker
supports --no-warn-execstack. Thanks Paul.

But this is a really old system. Do we really want to keep supporting
gcc 4.8 and binutils 2.27? Both are 9 years old. CentOS7 and CentOS8
stopped doing updates earlier this year.

Cheers,

Mark

Re: [Valgrind-developers] zstd.h missing in make dist tarball

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

Hi Florian,

On Sun, Dec 22, 2024 at 11:06:04AM +0100, Florian Krohm wrote:
> On 21.12.24 23:01, Mark Wielaard wrote:
> >
> >No objection. It looks correct.  I was just wondering why make
> >distcheck didn't detect it.
> 
> The reason could be that make distcheck is run on demand only. ?
> I think it would be good to run make distcheck as part of
> post_regtest_checks.
> That way we'd catch the problem early on in the nightly runs.
> Doing so implies, though, that you have xml machinery installed
> which is not universally so.

There is at least one buildbot that includes make distcheck.
https://builder.sourceware.org/buildbot/#/builders/valgrind-fedora-x86_64
(But that did have zstd-devel installed so...)

> >But if you have zstd devel headers
> >installed the build will pick up the /usr/include/zstd.h variant.
> 
> Interesting. So, ideally, make distcheck ought to run on a machine
> where these headers are not installed.

Or maybe we should rename the internal copy to vg-zstd.h ?

Cheers,

Mark

[Valgrind-developers] [valgrind] regtest: add check for -Wl, --no-warn-execstack

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=7deeb51a44392f26712f3a429813491423c405e9

commit 7deeb51a44392f26712f3a429813491423c405e9
Author: Paul Floyd <pj...@wa...>
Date:   Sun Dec 22 16:49:11 2024 +0100

    regtest: add check for -Wl,--no-warn-execstack

Diff:
---
 configure.ac           | 16 ++++++++++++++++
 none/tests/Makefile.am |  2 +-
 2 files changed, 17 insertions(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index 7c417eb9d7..cb1dc69d35 100755
--- a/configure.ac
+++ b/configure.ac
@@ -3100,6 +3100,22 @@ AC_LINK_IFELSE(
 CFLAGS=$safe_CFLAGS
 fi
 
+# check if linker accepts --no-warn-execstack
+AC_MSG_CHECKING([if the linker accepts -Wl,--no-warn-execstack])
+safe_CFLAGS=$CFLAGS
+CFLAGS="-Wl,--no-warn-execstack"
+
+AC_LINK_IFELSE(
+[AC_LANG_PROGRAM([ ], [return 0;])],
+[
+  AC_SUBST([FLAG_NO_WARN_EXECSTACK], ["-Wl,--no-warn-execstack"])
+  AC_MSG_RESULT([yes])
+], [
+  AC_SUBST([FLAG_NO_WARN_EXECSTACK], [""])
+  AC_MSG_RESULT([no])
+])
+CFLAGS=$safe_CFLAGS
+
 # does the ppc assembler support "mtocrf" et al?
 AC_MSG_CHECKING([if ppc32/64 as supports mtocrf/mfocrf])
 
diff --git a/none/tests/Makefile.am b/none/tests/Makefile.am
index 76825b9478..3dc8762514 100644
--- a/none/tests/Makefile.am
+++ b/none/tests/Makefile.am
@@ -335,7 +335,7 @@ endif
 
 if HAVE_NESTED_FUNCTIONS
    check_PROGRAMS += nestedfns
-   nestedfns_LDFLAGS = -Wl,--no-warn-execstack
+   nestedfns_LDFLAGS = @FLAG_NO_WARN_EXECSTACK@
 endif
 
 # This doesn't appear to be compilable on Darwin.

Re: [Valgrind-developers] [valgrind] regtest: turn off exec stack warning

[Philippe W. <phi...@sk...>]

The below change causes a build failure on the gcc110 compile farm

valgrind revision: valgrind-3.25.0.GIT-544d41e525-20241221
C compiler:        gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)
GDB:               GNU gdb (GDB) Red Hat Enterprise Linux 7.6.1-120.el7
Assembler:         GNU assembler version 2.27-44.base.el7_9.1
C library:         GNU C Library (GNU libc) stable release version 2.17
uname -mrs:        Linux 3.10.0-1160.105.1.el7.ppc64 ppc64
Vendor version:    CentOS Linux 7 (AltArch)

Nightly build on gcc110 ( \S, ppc64 )
Started at 2024-12-22 03:00:02 UTC
Ended   at 2024-12-22 03:25:19 UTC
Results differ from 24 hours ago

   Checking out Valgrind source tree  ... done
   Configuring valgrind               ... done
   Building valgrind                  ... failed

Last 20 lines of verbose log follow echo
gcc -std=gnu99 -Winline -Wall -Wshadow -Wno-long-long -g -fno-stack-protector   -m64    -o
use_after_close use_after_close.o 
gcc -std=gnu99 -Winline -Wall -Wshadow -Wno-long-long -g -fno-stack-protector   -m64  -
Wl,--no-warn-execstack  -o nestedfns nestedfns-nestedfns.o 
gcc -std=gnu99 -Winline -Wall -Wshadow -Wno-long-long -g -fno-stack-protector   -m64    -o
rlimit64_nofile rlimit64_nofile.o 
gcc -std=gnu99 -Winline -Wall -Wshadow -Wno-long-long -g -fno-stack-protector   -m64    -o
ppoll_alarm ppoll_alarm.o -lpthread
/usr/bin/ld: unrecognized option '--no-warn-execstack'
/usr/bin/ld: use the --help option for usage information
collect2: error: ld returned 1 exit status
make[5]: *** [nestedfns] Error 1
make[5]: *** Waiting for unfinished jobs....
gcc -std=gnu99 -Winline -Wall -Wshadow -Wno-long-long -g -fno-stack-protector   -m64 -ansi
-o ansi ansi-ansi.o 
make[5]: Leaving directory `/home/philippe/valgrind/cron_nightly/nightly/valgrind-
new/none/tests'
make[4]: *** [check-am] Error 2
make[4]: Leaving directory `/home/philippe/valgrind/cron_nightly/nightly/valgrind-
new/none/tests'
make[3]: *** [check-recursive] Error 1
make[3]: Leaving directory `/home/philippe/valgrind/cron_nightly/nightly/valgrind-
new/none/tests'
make[2]: *** [check-recursive] Error 1
make[2]: Leaving directory `/home/philippe/valgrind/cron_nightly/nightly/valgrind-
new/none'
make[1]: *** [check-recursive] Error 1valgrind revision: valgrind-3.25.0.GIT-544d41e525-
20241221


On Sat, 2024-12-21 at 16:21 +0000, Paul Floyd wrote:
> https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=0446f4ec3d11f8be373ef36edbec9818e1e1f4d4
> 
> commit 0446f4ec3d11f8be373ef36edbec9818e1e1f4d4
> Author: Paul Floyd <pj...@wa...>
> Date:   Sat Dec 21 17:18:35 2024 +0100
> 
>     regtest: turn off exec stack warning
>     
>     Rather than turning off exec stack.
> 
> Diff:
> ---
>  none/tests/Makefile.am | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/none/tests/Makefile.am b/none/tests/Makefile.am
> index fa0128d3a8..76825b9478 100644
> --- a/none/tests/Makefile.am
> +++ b/none/tests/Makefile.am
> @@ -335,7 +335,7 @@ endif
>  
>  if HAVE_NESTED_FUNCTIONS
>     check_PROGRAMS += nestedfns
> -   nestedfns_LDFLAGS = -Wl,-z,noexecstack
> +   nestedfns_LDFLAGS = -Wl,--no-warn-execstack
>  endif
>  
>  # This doesn't appear to be compilable on Darwin.
> 
> 
> _______________________________________________
> Valgrind-developers mailing list
> Val...@li...
> https://lists.sourceforge.net/lists/listinfo/valgrind-developers

Re: [Valgrind-developers] UPDATE: makefile patch to remove regtest artefacts

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

Hi all

To test this I did the following, starting from a clean repo

1. Modified a sample 5 different kinds of expected files to get some 
failures.
2. Ran 'make regtest'
3. make distclean
4. git status --ignored

The main thing that is not cleaned are all of the Makefile.in that are 
the intermediary step from Makefile.am to Makefile.

There's also a bit of autotools cruft in the root directory, a few 
garbage.filtered.out files and stuff like the gdb and vgdb links in 
gdbserver_tests.

A+
Paul

Re: [Valgrind-developers] zstd.h missing in make dist tarball

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

On 21.12.24 23:01, Mark Wielaard wrote:
> 
> No objection. It looks correct.  I was just wondering why make
> distcheck didn't detect it.  

The reason could be that make distcheck is run on demand only. ?
I think it would be good to run make distcheck as part of
post_regtest_checks. That way we'd catch the problem early on in the nightly runs.
Doing so implies, though, that you have xml machinery installed which is not 
universally so.

> But if you have zstd devel headers
> installed the build will pick up the /usr/include/zstd.h variant.
> 

Interesting. So, ideally, make distcheck ought to run on a machine where these 
headers are not installed.

   Florian

[Valgrind-developers] [valgrind] Add missing zstd.h to coregrind Makefile.am noinst_HEADERS

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=b1ee7bab4cc6879be849e15a50815dc8964642df

commit b1ee7bab4cc6879be849e15a50815dc8964642df
Author: Florian Krohm <fl...@ei...>
Date:   Sun Dec 22 10:53:03 2024 +0100

    Add missing zstd.h to coregrind Makefile.am noinst_HEADERS
    
    The header was missing from dist tarballs

Diff:
---
 coregrind/Makefile.am | 1 +
 1 file changed, 1 insertion(+)

diff --git a/coregrind/Makefile.am b/coregrind/Makefile.am
index 4eeb6e3d29..5491fc6726 100644
--- a/coregrind/Makefile.am
+++ b/coregrind/Makefile.am
@@ -265,6 +265,7 @@ noinst_HEADERS = \
 	m_debuginfo/lzoconf.h		\
 	m_debuginfo/lzodefs.h		\
 	m_debuginfo/minilzo.h		\
+	m_debuginfo/zstd.h              \
 	m_demangle/ansidecl.h	\
 	m_demangle/cp-demangle.h \
 	m_demangle/dyn-string.h	\

[Valgrind-developers] [valgrind] Bug 497723 - tweak awk filter in regtest

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=59cdf3890399ad68d044479fe3c068d2286efa43

commit 59cdf3890399ad68d044479fe3c068d2286efa43
Author: Paul Floyd <pj...@wa...>
Date:   Sun Dec 22 08:37:12 2024 +0100

    Bug 497723 - tweak awk filter in regtest

Diff:
---
 callgrind/tests/bug497723.vgtest | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/callgrind/tests/bug497723.vgtest b/callgrind/tests/bug497723.vgtest
index bdbf239870..92be389b07 100644
--- a/callgrind/tests/bug497723.vgtest
+++ b/callgrind/tests/bug497723.vgtest
@@ -1,4 +1,4 @@
 prog: ../../memcheck/tests/bug445235_ada_demangle
 vgopts: --callgrind-out-file=callgrind.out
-post: awk '/cfn.*ada_/{print $2}' callgrind.out
+post: awk '/fn.*ada_/{print $2}' callgrind.out
 cleanup: rm callgrind.out

Re: [Valgrind-developers] zstd.h missing in make dist tarball

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

Hi Paul,

On Sat, Dec 21, 2024 at 10:02:12PM +0100, Paul Floyd via Valgrind-developers wrote:
> 
> On 20-12-24 19:19, Florian Krohm wrote:
> >diff --git a/coregrind/Makefile.am b/coregrind/Makefile.am
> >index 4eeb6e3d2..63bbc5ef5 100644
> >--- a/coregrind/Makefile.am
> >+++ b/coregrind/Makefile.am
> >@@ -265,6 +265,7 @@ noinst_HEADERS = \
> >         m_debuginfo/lzoconf.h           \
> >         m_debuginfo/lzodefs.h           \
> >         m_debuginfo/minilzo.h           \
> >+       m_debuginfo/zstd.h              \
> >         m_demangle/ansidecl.h   \
> >         m_demangle/cp-demangle.h \
> >         m_demangle/dyn-string.h \
> 
> I can reproduce this. Unless there are any objections I'll push this
> change tomorrow.

No objection. It looks correct.  I was just wondering why make
distcheck didn't detect it.  But if you have zstd devel headers
installed the build will pick up the /usr/include/zstd.h variant.

Cheers,

Mark

Re: [Valgrind-developers] zstd.h missing in make dist tarball

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

On 20-12-24 19:19, Florian Krohm wrote:
> diff --git a/coregrind/Makefile.am b/coregrind/Makefile.am
> index 4eeb6e3d2..63bbc5ef5 100644
> --- a/coregrind/Makefile.am
> +++ b/coregrind/Makefile.am
> @@ -265,6 +265,7 @@ noinst_HEADERS = \
>          m_debuginfo/lzoconf.h           \
>          m_debuginfo/lzodefs.h           \
>          m_debuginfo/minilzo.h           \
> +       m_debuginfo/zstd.h              \
>          m_demangle/ansidecl.h   \
>          m_demangle/cp-demangle.h \
>          m_demangle/dyn-string.h \

I can reproduce this. Unless there are any objections I'll push this 
change tomorrow.

A+
Paul

[Valgrind-developers] [valgrind] Bug 497723 - forgot to restore callgrind output cleanup

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=544d41e525879f868e6f4d7c4872d46c666827c4

commit 544d41e525879f868e6f4d7c4872d46c666827c4
Author: Paul Floyd <pj...@wa...>
Date:   Sat Dec 21 21:32:19 2024 +0100

    Bug 497723 - forgot to restore callgrind output cleanup

Diff:
---
 callgrind/tests/bug497723.vgtest | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/callgrind/tests/bug497723.vgtest b/callgrind/tests/bug497723.vgtest
index 002dfb1920..bdbf239870 100644
--- a/callgrind/tests/bug497723.vgtest
+++ b/callgrind/tests/bug497723.vgtest
@@ -1,4 +1,4 @@
 prog: ../../memcheck/tests/bug445235_ada_demangle
 vgopts: --callgrind-out-file=callgrind.out
 post: awk '/cfn.*ada_/{print $2}' callgrind.out
-#cleanup: rm callgrind.out
+cleanup: rm callgrind.out

[Valgrind-developers] [valgrind] Command line help - demangling isn't just for C++

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=476f2558457e2c02c202709d699d5ef73987c432

commit 476f2558457e2c02c202709d699d5ef73987c432
Author: Paul Floyd <pj...@wa...>
Date:   Sat Dec 21 21:30:45 2024 +0100

    Command line help - demangling isn't just for C++
    
    Make the first line generic and add a list of languages
    that are suppoted.

Diff:
---
 coregrind/m_main.c                       | 3 ++-
 none/tests/cmdline1.stdout.exp           | 3 ++-
 none/tests/cmdline1.stdout.exp-non-linux | 3 ++-
 none/tests/cmdline2.stdout.exp           | 3 ++-
 none/tests/cmdline2.stdout.exp-non-linux | 3 ++-
 5 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/coregrind/m_main.c b/coregrind/m_main.c
index 85e744d157..3eba026cc5 100644
--- a/coregrind/m_main.c
+++ b/coregrind/m_main.c
@@ -130,7 +130,8 @@ static void usage_NORETURN ( int need_help )
 "    --xml-file=<file>         XML output to <file>\n"
 "    --xml-socket=ipaddr:port  XML output to socket ipaddr:port\n"
 "    --xml-user-comment=STR    copy STR verbatim into XML output\n"
-"    --demangle=no|yes         automatically demangle C++ names? [yes]\n"
+"    --demangle=no|yes         automatically demangle decorated names? [yes]\n"
+"                              supported languages: C++, D, Rust, Java, Ada\n"
 "    --num-callers=<number>    show <number> callers in stack traces [12]\n"
 "    --error-limit=no|yes      stop showing new errors if too many? [yes]\n"
 "    --exit-on-first-error=no|yes exit code on the first error found? [no]\n"
diff --git a/none/tests/cmdline1.stdout.exp b/none/tests/cmdline1.stdout.exp
index f88e63d24d..c7b7536856 100644
--- a/none/tests/cmdline1.stdout.exp
+++ b/none/tests/cmdline1.stdout.exp
@@ -43,7 +43,8 @@ usage: valgrind [options] prog-and-args
     --xml-file=<file>         XML output to <file>
     --xml-socket=ipaddr:port  XML output to socket ipaddr:port
     --xml-user-comment=STR    copy STR verbatim into XML output
-    --demangle=no|yes         automatically demangle C++ names? [yes]
+    --demangle=no|yes         automatically demangle decorated names? [yes]
+                              supported languages: C++, D, Rust, Java, Ada
     --num-callers=<number>    show <number> callers in stack traces [12]
     --error-limit=no|yes      stop showing new errors if too many? [yes]
     --exit-on-first-error=no|yes exit code on the first error found? [no]
diff --git a/none/tests/cmdline1.stdout.exp-non-linux b/none/tests/cmdline1.stdout.exp-non-linux
index 6449dc15b6..6650cc5b90 100644
--- a/none/tests/cmdline1.stdout.exp-non-linux
+++ b/none/tests/cmdline1.stdout.exp-non-linux
@@ -41,7 +41,8 @@ usage: valgrind [options] prog-and-args
     --xml-file=<file>         XML output to <file>
     --xml-socket=ipaddr:port  XML output to socket ipaddr:port
     --xml-user-comment=STR    copy STR verbatim into XML output
-    --demangle=no|yes         automatically demangle C++ names? [yes]
+    --demangle=no|yes         automatically demangle decorated names? [yes]
+                              supported languages: C++, D, Rust, Java, Ada
     --num-callers=<number>    show <number> callers in stack traces [12]
     --error-limit=no|yes      stop showing new errors if too many? [yes]
     --exit-on-first-error=no|yes exit code on the first error found? [no]
diff --git a/none/tests/cmdline2.stdout.exp b/none/tests/cmdline2.stdout.exp
index c4dc05f187..24ca407f80 100644
--- a/none/tests/cmdline2.stdout.exp
+++ b/none/tests/cmdline2.stdout.exp
@@ -43,7 +43,8 @@ usage: valgrind [options] prog-and-args
     --xml-file=<file>         XML output to <file>
     --xml-socket=ipaddr:port  XML output to socket ipaddr:port
     --xml-user-comment=STR    copy STR verbatim into XML output
-    --demangle=no|yes         automatically demangle C++ names? [yes]
+    --demangle=no|yes         automatically demangle decorated names? [yes]
+                              supported languages: C++, D, Rust, Java, Ada
     --num-callers=<number>    show <number> callers in stack traces [12]
     --error-limit=no|yes      stop showing new errors if too many? [yes]
     --exit-on-first-error=no|yes exit code on the first error found? [no]
diff --git a/none/tests/cmdline2.stdout.exp-non-linux b/none/tests/cmdline2.stdout.exp-non-linux
index 034f63d580..e13d196954 100644
--- a/none/tests/cmdline2.stdout.exp-non-linux
+++ b/none/tests/cmdline2.stdout.exp-non-linux
@@ -41,7 +41,8 @@ usage: valgrind [options] prog-and-args
     --xml-file=<file>         XML output to <file>
     --xml-socket=ipaddr:port  XML output to socket ipaddr:port
     --xml-user-comment=STR    copy STR verbatim into XML output
-    --demangle=no|yes         automatically demangle C++ names? [yes]
+    --demangle=no|yes         automatically demangle decorated names? [yes]
+                              supported languages: C++, D, Rust, Java, Ada
     --num-callers=<number>    show <number> callers in stack traces [12]
     --error-limit=no|yes      stop showing new errors if too many? [yes]
     --exit-on-first-error=no|yes exit code on the first error found? [no]

[Valgrind-developers] [valgrind] Bug 497723 - improve filter

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=0ef63235518b15b97ecfcab5216c99637ad77205

commit 0ef63235518b15b97ecfcab5216c99637ad77205
Author: Paul Floyd <pj...@wa...>
Date:   Sat Dec 21 21:27:23 2024 +0100

    Bug 497723 - improve filter
    
    Just print out the function names. Previously it was unstable.

Diff:
---
 callgrind/tests/bug497723.post.exp | 4 ++--
 callgrind/tests/bug497723.vgtest   | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/callgrind/tests/bug497723.post.exp b/callgrind/tests/bug497723.post.exp
index 7be1e7e38c..095e3b850b 100644
--- a/callgrind/tests/bug497723.post.exp
+++ b/callgrind/tests/bug497723.post.exp
@@ -1,2 +1,2 @@
-cfn=(5164) ada__text_io__put_line
-cfn=(5160) _ada_main
+ada__text_io__put_line
+_ada_main
diff --git a/callgrind/tests/bug497723.vgtest b/callgrind/tests/bug497723.vgtest
index 722ee33d05..002dfb1920 100644
--- a/callgrind/tests/bug497723.vgtest
+++ b/callgrind/tests/bug497723.vgtest
@@ -1,4 +1,4 @@
 prog: ../../memcheck/tests/bug445235_ada_demangle
 vgopts: --callgrind-out-file=callgrind.out
-post: grep "cfn.*ada_" callgrind.out
-cleanup: rm callgrind.out
+post: awk '/cfn.*ada_/{print $2}' callgrind.out
+#cleanup: rm callgrind.out

[Valgrind-developers] [valgrind] Bug 497723 - Enabling Ada demangling breaks callgrind differentiation between overloaded functions a

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=2e029beff537aab841ea7a3b4207b8b5ff33c86f

commit 2e029beff537aab841ea7a3b4207b8b5ff33c86f
Author: Paul Floyd <pj...@wa...>
Date:   Sat Dec 21 20:46:40 2024 +0100

    Bug 497723 - Enabling Ada demangling breaks callgrind differentiation between overloaded functions and procedures

Diff:
---
 NEWS                                 | 2 ++
 callgrind/tests/Makefile.am          | 1 +
 callgrind/tests/bug497723.post.exp   | 2 ++
 callgrind/tests/bug497723.stderr.exp | 6 ++++++
 callgrind/tests/bug497723.vgtest     | 4 ++++
 coregrind/m_redir.c                  | 7 ++++++-
 6 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/NEWS b/NEWS
index b668ad2e4b..0be2ff0339 100644
--- a/NEWS
+++ b/NEWS
@@ -36,6 +36,8 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 469782  Valgrind does not support zstd-compressed debug sections
 497130  Recognize new DWARF5 DW_LANG constants
 497455  Update drd/scripts/download-and-build-gcc
+497723  Enabling Ada demangling breaks callgrind differentiation between
+        overloaded functions and procedures
 
 To see details of a given bug, visit
   https://bugs.kde.org/show_bug.cgi?id=XXXXXX
diff --git a/callgrind/tests/Makefile.am b/callgrind/tests/Makefile.am
index 9b68679908..4a49ea7de4 100644
--- a/callgrind/tests/Makefile.am
+++ b/callgrind/tests/Makefile.am
@@ -10,6 +10,7 @@ EXTRA_DIST = \
 	ann1.post.exp ann1.stderr.exp ann1.vgtest \
 	ann2.post.exp ann2.stderr.exp ann2.vgtest \
 	clreq.vgtest clreq.stderr.exp \
+	bug497723.stderr.exp bug497723.post.exp bug497723.vgtest \
 	simwork1.vgtest simwork1.stdout.exp simwork1.stderr.exp \
 	simwork2.vgtest simwork2.stdout.exp simwork2.stderr.exp \
 	simwork3.vgtest simwork3.stdout.exp simwork3.stderr.exp \
diff --git a/callgrind/tests/bug497723.post.exp b/callgrind/tests/bug497723.post.exp
new file mode 100644
index 0000000000..7be1e7e38c
--- /dev/null
+++ b/callgrind/tests/bug497723.post.exp
@@ -0,0 +1,2 @@
+cfn=(5164) ada__text_io__put_line
+cfn=(5160) _ada_main
diff --git a/callgrind/tests/bug497723.stderr.exp b/callgrind/tests/bug497723.stderr.exp
new file mode 100644
index 0000000000..d0b7820ae3
--- /dev/null
+++ b/callgrind/tests/bug497723.stderr.exp
@@ -0,0 +1,6 @@
+
+
+Events    : Ir
+Collected :
+
+I   refs:
diff --git a/callgrind/tests/bug497723.vgtest b/callgrind/tests/bug497723.vgtest
new file mode 100644
index 0000000000..722ee33d05
--- /dev/null
+++ b/callgrind/tests/bug497723.vgtest
@@ -0,0 +1,4 @@
+prog: ../../memcheck/tests/bug445235_ada_demangle
+vgopts: --callgrind-out-file=callgrind.out
+post: grep "cfn.*ada_" callgrind.out
+cleanup: rm callgrind.out
diff --git a/coregrind/m_redir.c b/coregrind/m_redir.c
index f9e8d8801a..6c360d59ce 100644
--- a/coregrind/m_redir.c
+++ b/coregrind/m_redir.c
@@ -537,8 +537,13 @@ void VG_(redir_notify_new_DebugInfo)( DebugInfo* newdi )
           * For Ada demangling, the language doesn't use a regular
           * prefix like _Z or _R, so look for a common symbol and
           * set a global flag.
+          *
+          * https://bugs.kde.org/show_bug.cgi?id=497723 but not for
+          * callgrind because demangled overloaded manes get
+          * incorrectly counted together.
           */
-         if (!isText && VG_(strcmp)(*names, "__gnat_ada_main_program_name") == 0) {
+         if (!isText && VG_(strcmp)(*names, "__gnat_ada_main_program_name") == 0 &&
+             VG_(strcmp)(VG_(clo_toolname), "callgrind") != 0)  {
             VG_(lang_is_ada) = True;
          }
          isGlobal = False;

Re: [Valgrind-developers] UPDATE: makefile patch to remove regtest artefacts

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

On 20-12-24 19:22, Florian Krohm wrote:
> Any takers ?

Hi Florian

I'm planning to do some tests over the next few days. I'll compare make 
clean-local with git clean -fdx.

A+
Paul

[Valgrind-developers] [valgrind] regtest: turn off exec stack warning

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=0446f4ec3d11f8be373ef36edbec9818e1e1f4d4

commit 0446f4ec3d11f8be373ef36edbec9818e1e1f4d4
Author: Paul Floyd <pj...@wa...>
Date:   Sat Dec 21 17:18:35 2024 +0100

    regtest: turn off exec stack warning
    
    Rather than turning off exec stack.

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

diff --git a/none/tests/Makefile.am b/none/tests/Makefile.am
index fa0128d3a8..76825b9478 100644
--- a/none/tests/Makefile.am
+++ b/none/tests/Makefile.am
@@ -335,7 +335,7 @@ endif
 
 if HAVE_NESTED_FUNCTIONS
    check_PROGRAMS += nestedfns
-   nestedfns_LDFLAGS = -Wl,-z,noexecstack
+   nestedfns_LDFLAGS = -Wl,--no-warn-execstack
 endif
 
 # This doesn't appear to be compilable on Darwin.

[Valgrind-developers] [valgrind] Implement the new GDB 16.0 'x addr, len' packet.

[Philippe W. <phi...@so...>]

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=706c5dd0bae49be20ce15c2eeacbbf2c9d49a852

commit 706c5dd0bae49be20ce15c2eeacbbf2c9d49a852
Author: Philippe Waroquiers <phi...@sk...>
Date:   Sat Dec 21 15:45:20 2024 +0100

    Implement the new GDB 16.0 'x addr,len' packet.
    
    GDB can read memory faster with this packet than with the 'm addr,len' packet.

Diff:
---
 NEWS                           |  4 ++++
 coregrind/m_gdbserver/server.c | 14 ++++++++++++++
 2 files changed, 18 insertions(+)

diff --git a/NEWS b/NEWS
index d78d9e9a5d..b668ad2e4b 100644
--- a/NEWS
+++ b/NEWS
@@ -10,6 +10,10 @@ AMD64/macOS 10.13 and nanoMIPS/Linux.
 
 * ==================== CORE CHANGES ===================
 
+* The valgrind gdbserver now supports the GDB remote protocol packet
+  'x addr,len' (available in GDB release >= 16).
+  The x packet can reduce the time taken by GDB to read memory from valgrind.
+
 * ================== PLATFORM CHANGES =================
 
 * ==================== TOOL CHANGES ===================
diff --git a/coregrind/m_gdbserver/server.c b/coregrind/m_gdbserver/server.c
index 19a360c7a7..939b7f2931 100644
--- a/coregrind/m_gdbserver/server.c
+++ b/coregrind/m_gdbserver/server.c
@@ -1376,6 +1376,20 @@ void server_main (void)
          else
             write_enn (own_buf);
          break;
+      case 'x':
+         decode_m_packet (&own_buf[1], &mem_addr, &len);
+         if (valgrind_read_memory (mem_addr, mem_buf, len) == 0) {
+            // Read memory is successful.
+            // Complete the reply packet and indicate its length.
+            int out_len;
+            own_buf[0] = 'b';
+            new_packet_len
+               = 1 + remote_escape_output(mem_buf, len,
+                                          (unsigned char *) &own_buf[1], &out_len,
+                                          PBUFSIZ - POVERHSIZ - 1);
+         } else
+            write_enn (own_buf);
+         break;
       case 'X':
          if (decode_X_packet (&own_buf[1], packet_len - 1,
                               &mem_addr, &len, mem_buf) < 0

[Valgrind-developers] [valgrind] VEX/priv/ir_opt.c: Also fold Iop_And1 expressions when possible

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=1f8a81cc2493670a82334bd5b4bc1ffa6d02174b

commit 1f8a81cc2493670a82334bd5b4bc1ffa6d02174b
Author: Mark Wielaard <ma...@kl...>
Date:   Sun Dec 15 21:26:21 2024 +0100

    VEX/priv/ir_opt.c: Also fold Iop_And1 expressions when possible
    
    Treat Iop_And1 just like And16/And32/And64. Fold fully constant And1
    expressions and handle And1(x,True), And1(True,x), And1(x,False),
    And1(False,x) and And1(x,x).
    
    Make sure isOnesU handles Ico_U1 (isZerosU and sameIRExprs already
    did).
    
    https://bugs.kde.org/show_bug.cgi?id=468575#c52

Diff:
---
 VEX/priv/ir_opt.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/VEX/priv/ir_opt.c b/VEX/priv/ir_opt.c
index 6453f4fdf5..6565378654 100644
--- a/VEX/priv/ir_opt.c
+++ b/VEX/priv/ir_opt.c
@@ -1239,6 +1239,7 @@ static Bool isOnesU ( IRExpr* e )
 {
    if (e->tag != Iex_Const) return False;
    switch (e->Iex.Const.con->tag) {
+      case Ico_U1:    return toBool( e->Iex.Const.con->Ico.U1  == True);
       case Ico_U8:    return toBool( e->Iex.Const.con->Ico.U8  == 0xFF);
       case Ico_U16:   return toBool( e->Iex.Const.con->Ico.U16 == 0xFFFF);
       case Ico_U32:   return toBool( e->Iex.Const.con->Ico.U32
@@ -1832,6 +1833,11 @@ static IRExpr* fold_Expr_WRK ( IRExpr** env, IRExpr* e )
                break;
 
             /* -- And -- */
+            case Iop_And1:
+               e2 = IRExpr_Const(IRConst_U1(toBool(
+                       (e->Iex.Binop.arg1->Iex.Const.con->Ico.U1
+                        & e->Iex.Binop.arg2->Iex.Const.con->Ico.U1))));
+               break;
             case Iop_And8:
                e2 = IRExpr_Const(IRConst_U8(toUChar( 
                        (e->Iex.Binop.arg1->Iex.Const.con->Ico.U8
@@ -2274,31 +2280,32 @@ static IRExpr* fold_Expr_WRK ( IRExpr** env, IRExpr* e )
                }
                break;
 
+            case Iop_And1:
             case Iop_And8:
             case Iop_And16:
             case Iop_And32:
             case Iop_And64:
-               /* And8/And16/And32/And64(x,1---1b) ==> x */
+               /* And1/And8/And16/And32/And64(x,1---1b) ==> x */
                if (isOnesU(e->Iex.Binop.arg2)) {
                   e2 = e->Iex.Binop.arg1;
                   break;
                }
-               /* And8/And16/And32/And64(1---1b,x) ==> x */
+               /* And1/And8/And16/And32/And64(1---1b,x) ==> x */
                if (isOnesU(e->Iex.Binop.arg1)) {
                   e2 = e->Iex.Binop.arg2;
                   break;
                }
-               /* And8/And16/And32/And64(x,0) ==> 0 */
+               /* And1/And8/And16/And32/And64(x,0) ==> 0 */
                if (isZeroU(e->Iex.Binop.arg2)) {
                   e2 = e->Iex.Binop.arg2;
                   break;
                }
-               /* And8/And16/And32/And64(0,x) ==> 0 */
+               /* And1/And8/And16/And32/And64(0,x) ==> 0 */
                if (isZeroU(e->Iex.Binop.arg1)) {
                   e2 = e->Iex.Binop.arg1;
                   break;
                }
-               /* And8/And16/And32/And64(t,t) ==> t, for some IRTemp t */
+               /* And1/And8/And16/And32/And64(t,t) ==> t, for some IRTemp t */
                if (sameIRExprs(env, e->Iex.Binop.arg1, e->Iex.Binop.arg2)) {
                   e2 = e->Iex.Binop.arg1;
                   break;

[Valgrind-developers] [valgrind] Use Ints for fds in PRE and POST sys_close_range

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=d984d9aabe5b89d6787c2604b72c09360e7b6fc0

commit d984d9aabe5b89d6787c2604b72c09360e7b6fc0
Author: Mark Wielaard <ma...@kl...>
Date:   Sat Dec 14 22:34:12 2024 +0000

    Use Ints for fds in PRE and POST sys_close_range
    
    The double_close_range test failed on riscv64-linux because the
    close_range wrapper is using unsigned int and the ARG regwords
    directly. Which causes the ARG2 == ~0U check to fail. Explicitly
    using Int for the fd arguments fixes this. I am not clear on why
    this was only an issue for the riscv port. It seems this patch is
    OK for other arches (tested on amd64 and i386).

Diff:
---
 coregrind/m_syswrap/syswrap-generic.c |  2 +-
 coregrind/m_syswrap/syswrap-linux.c   | 22 ++++++++++++----------
 2 files changed, 13 insertions(+), 11 deletions(-)

diff --git a/coregrind/m_syswrap/syswrap-generic.c b/coregrind/m_syswrap/syswrap-generic.c
index d6b10780dc..bfe4c6fe06 100644
--- a/coregrind/m_syswrap/syswrap-generic.c
+++ b/coregrind/m_syswrap/syswrap-generic.c
@@ -560,7 +560,7 @@ Int ML_(get_fd_count)(void)
 }
 
 /* Close_range caller might want to close very wide range of file descriptors,
-   up to 0U.  We want to avoid iterating through such a range in a normall
+   up to ~0U.  We want to avoid iterating through such a range in a normal
    close_range, just up to any open file descriptor.  Also, unlike
    record_fd_close_range, we assume the user might deliberately double closes
    any file descriptors in the range, so don't warn about double close here. */
diff --git a/coregrind/m_syswrap/syswrap-linux.c b/coregrind/m_syswrap/syswrap-linux.c
index e01d920e11..60e05a46c4 100644
--- a/coregrind/m_syswrap/syswrap-linux.c
+++ b/coregrind/m_syswrap/syswrap-linux.c
@@ -13625,8 +13625,9 @@ PRE(sys_execveat)
 PRE(sys_close_range)
 {
    SysRes res = VG_(mk_SysRes_Success)(0);
-   unsigned int beg, end;
-   unsigned int last = ARG2;
+   Int beg, end;
+   Int first = ARG1;
+   Int last = ARG2;
 
    FUSE_COMPATIBLE_MAY_BLOCK();
    PRINT("sys_close_range ( %" FMT_REGWORD "u, %" FMT_REGWORD "u, %"
@@ -13635,7 +13636,7 @@ PRE(sys_close_range)
                  unsigned int, first, unsigned int, last,
                  unsigned int, flags);
 
-   if (ARG1 > last) {
+   if (first > last) {
       SET_STATUS_Failure( VKI_EINVAL );
       return;
    }
@@ -13643,12 +13644,12 @@ PRE(sys_close_range)
    if (last >= VG_(fd_hard_limit))
       last = VG_(fd_hard_limit) - 1;
 
-   if (ARG1 > last) {
+   if (first > last) {
       SET_STATUS_Success ( 0 );
       return;
    }
 
-   beg = end = ARG1;
+   beg = end = first;
    do {
       if (end > last
 	  || (end == 2/*stderr*/ && VG_(debugLog_getLevel)() > 0)
@@ -13668,8 +13669,9 @@ PRE(sys_close_range)
 
 POST(sys_close_range)
 {
-   unsigned int fd;
-   unsigned int last = ARG2;
+   Int fd;
+   Int first = ARG1;
+   Int last = ARG2;
 
    if (!VG_(clo_track_fds)
        || (ARG3 & VKI_CLOSE_RANGE_CLOEXEC) != 0)
@@ -13680,10 +13682,10 @@ POST(sys_close_range)
 
    /* If the close_range range is too wide, we don't want to loop
       through the whole range.  */
-   if (ARG2 == ~0U)
-     ML_(record_fd_close_range)(tid, ARG1);
+   if (last == ~0U)
+     ML_(record_fd_close_range)(tid, first);
    else {
-     for (fd = ARG1; fd <= last; fd++)
+     for (fd = first; fd <= last; fd++)
        if ((fd != 2/*stderr*/ || VG_(debugLog_getLevel)() == 0)
            && fd != VG_(log_output_sink).fd
            && fd != VG_(xml_output_sink).fd)

Re: [Valgrind-developers] UPDATE: makefile patch to remove regtest artefacts

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

Any takers ?


On 07.12.24 22:23, Florian Krohm wrote:
> Here is an updated patch which also removes .post.out files.
> Also, turns out, that .diff files can have a suffix as derived from the .exp 
> file that was used in the comparison. Those get removed now as well.
> 
> diff --git a/Makefile.am b/Makefile.am
> index b3e5be5f4..bace4a499 100644
> --- a/Makefile.am
> +++ b/Makefile.am
> @@ -139,6 +139,16 @@ clean-local:
>       rm -rf $(inplacedir) include/vgversion.h
>       find . -name '*.vgtest.log' -print0 | xargs -r -0 rm
>       find . -name '*.vgtest.trs' -print0 | xargs -r -0 rm
> +    find . -name '*.stdout.out' -print0 | xargs -r -0 rm
> +    find . -name '*.stderr.out' -print0 | xargs -r -0 rm
> +    find . -name '*.stdoutB.out' -print0 | xargs -r -0 rm
> +    find . -name '*.stderrB.out' -print0 | xargs -r -0 rm
> +    find . -name '*.post.out' -print0 | xargs -r -0 rm
> +    find . -name '*.stdout.diff*' -print0 | xargs -r -0 rm
> +    find . -name '*.stderr.diff*' -print0 | xargs -r -0 rm
> +    find . -name '*.stdoutB.diff*' -print0 | xargs -r -0 rm
> +    find . -name '*.stderrB.diff*' -print0 | xargs -r -0 rm
> +    find . -name '*.post.diff*' -print0 | xargs -r -0 rm
>       find . -name test-suite.log -print0 | xargs -r -0 rm
>       rm -f test-suite-overall.log
> 
> 
> 
> On 02.12.24 22:00, Florian Krohm wrote:
>> The patch below removes artefacts from failing regression tests.
>> Specifically, "make distclean" and "make maintainer-clean" removes them. 
>> Previously they were not removed which was a surprise.
>> post_regtest_checks does not appear to create any files that ought to be 
>> cleaned up.
>> Thanks to Paul for pointing me to the file to change.
>>
>>    Florian
>>
>>
>> diff --git a/Makefile.am b/Makefile.am
>> index b3e5be5f4..b584cb34b 100644
>> --- a/Makefile.am
>> +++ b/Makefile.am
>> @@ -139,6 +139,14 @@ clean-local:
>>          rm -rf $(inplacedir) include/vgversion.h
>>          find . -name '*.vgtest.log' -print0 | xargs -r -0 rm
>>          find . -name '*.vgtest.trs' -print0 | xargs -r -0 rm
>> +       find . -name '*.stdout.out' -print0 | xargs -r -0 rm
>> +       find . -name '*.stderr.out' -print0 | xargs -r -0 rm
>> +       find . -name '*.stdoutB.out' -print0 | xargs -r -0 rm
>> +       find . -name '*.stderrB.out' -print0 | xargs -r -0 rm
>> +       find . -name '*.stdout.diff' -print0 | xargs -r -0 rm
>> +       find . -name '*.stderr.diff' -print0 | xargs -r -0 rm
>> +       find . -name '*.stdoutB.diff' -print0 | xargs -r -0 rm
>> +       find . -name '*.stderrB.diff' -print0 | xargs -r -0 rm
>>          find . -name test-suite.log -print0 | xargs -r -0 rm
>>          rm -f test-suite-overall.log
>>
> 
> 
> 
> _______________________________________________
> Valgrind-developers mailing list
> Val...@li...
> https://lists.sourceforge.net/lists/listinfo/valgrind-developers

[Valgrind-developers] zstd.h missing in make dist tarball

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

I was just testing a larger patch making sure the new stuff makes it into the 
tarball.. Turns out zstd.h is missing.

diff --git a/coregrind/Makefile.am b/coregrind/Makefile.am
index 4eeb6e3d2..63bbc5ef5 100644
--- a/coregrind/Makefile.am
+++ b/coregrind/Makefile.am
@@ -265,6 +265,7 @@ noinst_HEADERS = \
         m_debuginfo/lzoconf.h           \
         m_debuginfo/lzodefs.h           \
         m_debuginfo/minilzo.h           \
+       m_debuginfo/zstd.h              \
         m_demangle/ansidecl.h   \
         m_demangle/cp-demangle.h \
         m_demangle/dyn-string.h \

[Valgrind-developers] Sourceware infrastructure updates for Q4 2024

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

Sourceware infrastructure community updates for Q4 2024

Sourceware has provided the infrastructure for core toolchain and
developer tools projects for more than 25 years.
https://sourceware.org/sourceware-25-roadmap.html

The last couple of years it has transformed from a purely volunteer
into a professional organization with an eight person strong Project
Leadership Committee, monthly open office hours, multiple hardware
services partners, expanded services, the Software Freedom Conservancy
as fiscal sponsor and a more diverse funding model that allows us to
enter into contracts with paid contractors or staff when appropriate.

Every quarter we provide a summary of news about Sourceware, the core
toolchain and developer tools infrastructure, from the last 3 months.

- Donations and Software Freedom Conservancy match fundraiser
- Sourceware Cyber Security FAQ
- New incoming email dmarc p=reject policy
- Sourceware forge experiment
- Signed-commit census report
- Sourceware Organization, Contact and Open Office hours

= Donations and Software Freedom Conservancy match fundraiser

  We like to thank all the individuals who donated this year and
  helped us create a hardware backup fund in case one of our hardware
  sponsors has to drop out. https://sourceware.org/donate.html

  But various hardware and service partners have actually offered more
  resources this year. So we haven't needed to spend much money this
  year. And could still expand the buildbot CI, snapshot server and
  run the Forge experiment. https://sourceware.org/mission.html#sponsors

  Currently the best way to support Sourceware is to join the Software
  Freedom Conservancy match fundraiser:
  https://sfconservancy.org/news/2024/nov/26/2024-fundraiser/

  By becoming a Software Freedom Conservancy Sustainer the Conservancy
  can keep supporting Sourceware and other Software Freedom
  communities https://sfconservancy.org/sustainer/

= Sourceware Cyber Security FAQ

  After lots of discussions at some of our Open Office hours, at the
  Cauldron, with other Software Freedom organizations and some of our
  hardware and services providers we created an Sourceware Cyber
  Security FAQ "explainer" about topics like the "US Improving the
  Nation's Cybersecurity Executive Order 14028", "EU Cyber Resilience
  Act (EU CRA)" and "Secure Software Development Framework (NIST SP
  800-218)" or who is required to use Zero Trust (NIST SP 800-207)
  cloud-computing environments.

  https://sourceware.org/cyber-security-faq.html

  Somewhat related the Software Freedom Conservancy published a blog
  post about the recent bans of Russian contributors in the Linux
  project and whether Free Software projects need to worry about
  U.S. Sanctions.

  https://sfconservancy.org/blog/2024/dec/12/linux-banned-russian-contributors-do-i-need-to/

= New incoming email dmarc p=reject policy

  We are now rejecting any incoming emails that fail dmarc for domains
  that have a p=reject policy. Because if we forward such emails
  through mailinglists (without From rewriting) they will cause
  bounces for some subscribers (eventually causing those subscribers
  to be unsubscribed).

  https://sourceware.org/bugzilla/show_bug.cgi?id=32255  
  https://inbox.sourceware.org/e13...@kl.../

= Sourceware forge experiment

  https://forge.sourceware.org

  As an experiment Sourceware is now running an Forgejo v9 instance.

  The full setup can be found in this git repository:
  https://sourceware.org/cgit/forge/

  For now it is only open for existing maintainers/developers for
  experimenting with a pull-request model.

  There is also a new mailinglist for discussion about the setup and
  the best way to create a pull-request workflow:
  https://sourceware.org/mailman/listinfo/forge

  And a wiki with more background information:
  https://gcc.gnu.org/wiki/ForgeExperiment

= Signed-commit census report

  Analyzing branch HEAD since 2024-01-01
  annobin        85 commits  16 signed  18%   1 committers  1 signers 100%
  binutils-gdb 3433 commits  74 signed   2%  99 committers  3 signers   3%
  builder       141 commits  19 signed  13%   3 committers  1 signers  33%
  bunsen        163 commits 102 signed  62%   4 committers  3 signers  75%
  bzip2           1 commits   0 signed   0%   1 committers  0 signers   0%
  cgen            1 commits   0 signed   0%   1 committers  0 signers   0%
  cygwin-calm    68 commits  68 signed 100%   1 committers  1 signers 100%
  cygwin-setup   50 commits  49 signed  98%   3 committers  2 signers  66%
  debugedit      22 commits   1 signed   4%   2 committers  1 signers  50%
  dwz             2 commits   0 signed   0%   1 committers  0 signers   0%
  elfutils      161 commits  21 signed  13%   4 committers  2 signers  50%
  forge           7 commits   0 signed   0%   1 committers  0 signers   0%
  gcc          9654 commits 721 signed   7% 216 committers 14 signers   6%
  gitsigur        9 commits   9 signed 100%   1 committers  1 signers 100%
  glibc        1147 commits  82 signed   7%  39 committers  3 signers   7%
  insight        54 commits   0 signed   0%   1 committers  0 signers   0%
  libabigail    185 commits   1 signed   0%   2 committers  1 signers  50%
  lvm2          833 commits  42 signed   5%   8 committers  1 signers  12%
  newlib-cygwin 376 commits  16 signed   4%  12 committers  1 signers   8%
  systemtap     155 commits  87 signed  56%   5 committers  3 signers  60%
  valgrind      448 commits   0 signed   0%   8 committers  0 signers   0%

= Sourceware Organization, Contact and Open Office hours

  We can be reached through irc, email and bugzilla
  https://sourceware.org/mission.html#organization

  There is also a fediverse account for for announcements, notices
  about downtime and temporary issues with our network.
  https://fosstodon.org/@sourceware

  Every second Friday of the month is the Sourceware Overseers Open
  Office hour in #overseers on irc.libera.chat from 16:00 till 17:00
  UTC. Please feel free to drop by with any Sourceware services and
  hosting questions.

  If you aren't already and want to keep up to date on Sourceware
  infrastructure services then please also subscribe to the overseers
  mailinglist. https://sourceware.org/mailman/listinfo/overseers

Sourceware PLC,

 Frank Ch. Eigler, Christopher Faylor, Ian Kelling, Ian Lance Taylor,
 Tom Tromey, Jon Turney, Mark J. Wielaard, Elena Zannoni

Re: [Valgrind-developers] [PATCH v5] Fixes PLAT_amd64_win64 support with GCC, CLANG, MSVC

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

Hi,

On Wed, Dec 18, 2024 at 04:45:17AM +0800, 罗勇刚(Yonggang Luo) wrote:
> sorry for the ping

Sorry, but could you split this into "one change per patch" like Bart
asked.  And could you answer some of the questions I asked.
Specifically about how you tested your changes.

Thanks,

Mark

Re: [Valgrind-developers] [PATCH v5] Fixes PLAT_amd64_win64 support with GCC, CLANG, MSVC

[罗勇刚(Yonggang L. <luo...@gm...>]

sorry for the ping

On Sat, Dec 14, 2024 at 2:32 AM Yonggang Luo <luo...@gm...> wrote:
>
> Always include uintptr_t for win32, as we use uintptr_t.
>
> sizeof(unsigned long int) is 4 on mingw 64bit, use uintptr_t int instead,
that's the same with MSVC 64bit
>
> __clang__ for MSVC support for GCC inline assembly
>
> _MSC_VER for PLAT_amd64_win64 does not support inline assembly, use stub
instead,
> may be fixed use machine code.
>
> The compiling for MSVC,CLANG,GCC are checked
>
> Signed-off-by: Yonggang Luo <luo...@gm...>
> ---
>  include/valgrind.h.in | 121 ++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 116 insertions(+), 5 deletions(-)
>
> diff --git a/include/valgrind.h.in b/include/valgrind.h.in
> index bc18f40f7..e4867c333 100644
> --- a/include/valgrind.h.in
> +++ b/include/valgrind.h.in
> @@ -94,6 +94,10 @@
>
>  #include <stdarg.h>
>
> +#ifdef _WIN32
> +#include <stdint.h>
> +#endif
> +
>  /* Nb: this file might be included in a file compiled with -ansi.  So
>     we can't use C++ style "//" comments nor the "asm" keyword (instead
>     use "__asm__"). */
> @@ -403,12 +407,11 @@ valgrind_do_client_request_expr(uintptr_t
_zzq_default, uintptr_t _zzq_request,
>
>  #endif /* PLAT_x86_win32 */
>
> -/* ----------------- amd64-{linux,darwin,solaris} --------------- */
> +/* ----------------- amd64-{linux,darwin,solaris,freebsd}
--------------- */
>
>  #if defined(PLAT_amd64_linux)  ||  defined(PLAT_amd64_darwin) \
>      ||  defined(PLAT_amd64_solaris) \
> -    ||  defined(PLAT_amd64_freebsd) \
> -    ||  (defined(PLAT_amd64_win64) && defined(__GNUC__))
> +    ||  defined(PLAT_amd64_freebsd)
>
>  typedef
>     struct {
> @@ -468,14 +471,122 @@ typedef
>                      );                                           \
>   } while (0)
>
> -#endif /* PLAT_amd64_linux || PLAT_amd64_darwin || PLAT_amd64_solaris */
> +#endif /* defined(PLAT_amd64_linux)  ||  defined(PLAT_amd64_darwin) \
> +    ||  defined(PLAT_amd64_solaris) \
> +    ||  defined(PLAT_amd64_freebsd) */
>
>  /* ------------------------- amd64-Win64 ------------------------- */
>
> -#if defined(PLAT_amd64_win64) && !defined(__GNUC__)
> +#if defined(PLAT_amd64_win64)
> +
> +typedef
> +   struct {
> +      uintptr_t nraddr; /* where's the code? */
> +   }
> +   OrigFn;
> +
> +#if defined(__GNUC__) || defined(__clang__)
> +
> +#define __SPECIAL_INSTRUCTION_PREAMBLE                            \
> +                     "rolq $3,  %%rdi ; rolq $13, %%rdi\n\t"      \
> +                     "rolq $61, %%rdi ; rolq $51, %%rdi\n\t"
> +
> +#define VALGRIND_DO_CLIENT_REQUEST_EXPR(                          \
> +        _zzq_default, _zzq_request,                               \
> +        _zzq_arg1, _zzq_arg2, _zzq_arg3, _zzq_arg4, _zzq_arg5)    \
> +    __extension__                                                 \
> +    ({ volatile uintptr_t _zzq_args[6];                           \
> +    volatile uintptr_t _zzq_result;                               \
> +    _zzq_args[0] = (uintptr_t)(_zzq_request);                     \
> +    _zzq_args[1] = (uintptr_t)(_zzq_arg1);                        \
> +    _zzq_args[2] = (uintptr_t)(_zzq_arg2);                        \
> +    _zzq_args[3] = (uintptr_t)(_zzq_arg3);                        \
> +    _zzq_args[4] = (uintptr_t)(_zzq_arg4);                        \
> +    _zzq_args[5] = (uintptr_t)(_zzq_arg5);                        \
> +    __asm__ volatile(__SPECIAL_INSTRUCTION_PREAMBLE               \
> +                     /* %RDX = client_request ( %RAX ) */         \
> +                     "xchgq %%rbx,%%rbx"                          \
> +                     : "=d" (_zzq_result)                         \
> +                     : "a" (&_zzq_args[0]), "0" (_zzq_default)    \
> +                     : "cc", "memory"                             \
> +                    );                                            \
> +    _zzq_result;                                                  \
> +    })
> +
> +#define VALGRIND_GET_NR_CONTEXT(_zzq_rlval)                       \
> +  { volatile OrigFn* _zzq_orig = &(_zzq_rlval);                   \
> +    volatile uintptr_t __addr;                                    \
> +    __asm__ volatile(__SPECIAL_INSTRUCTION_PREAMBLE               \
> +                     /* %RAX = guest_NRADDR */                    \
> +                     "xchgq %%rcx,%%rcx"                          \
> +                     : "=a" (__addr)                              \
> +                     :                                            \
> +                     : "cc", "memory"                             \
> +                    );                                            \
> +    _zzq_orig->nraddr = __addr;                                   \
> +  }
> +
> +#define VALGRIND_CALL_NOREDIR_RAX                                 \
> +                     __SPECIAL_INSTRUCTION_PREAMBLE               \
> +                     /* call-noredir *%RAX */                     \
> +                     "xchgq %%rdx,%%rdx\n\t"
> +
> +#define VALGRIND_VEX_INJECT_IR()                                  \
> + do {                                                             \
> +    __asm__ volatile(__SPECIAL_INSTRUCTION_PREAMBLE               \
> +                     "xchgq %%rdi,%%rdi\n\t"                      \
> +                     : : : "cc", "memory"                         \
> +                    );                                            \
> + } while (0)
> +
> +#elif defined(_MSC_VER)
> +
> +#define VALGRIND_DO_CLIENT_REQUEST_EXPR(                          \
> +        _zzq_default, _zzq_request,                               \
> +        _zzq_arg1, _zzq_arg2, _zzq_arg3, _zzq_arg4, _zzq_arg5)    \
> +    valgrind_do_client_request_expr((uintptr_t)(_zzq_default),    \
> +        (uintptr_t)(_zzq_request), (uintptr_t)(_zzq_arg1),        \
> +        (uintptr_t)(_zzq_arg2), (uintptr_t)(_zzq_arg3),           \
> +        (uintptr_t)(_zzq_arg4), (uintptr_t)(_zzq_arg5))
> +
> +static __inline uintptr_t
> +valgrind_do_client_request_expr(uintptr_t _zzq_default, uintptr_t
_zzq_request,
> +                                uintptr_t _zzq_arg1, uintptr_t _zzq_arg2,
> +                                uintptr_t _zzq_arg3, uintptr_t _zzq_arg4,
> +                                uintptr_t _zzq_arg5)
> +{
> +    volatile uintptr_t _zzq_args[6];
> +    volatile uintptr_t _zzq_result = 0;
> +    _zzq_args[0] = (uintptr_t)(_zzq_request);
> +    _zzq_args[1] = (uintptr_t)(_zzq_arg1);
> +    _zzq_args[2] = (uintptr_t)(_zzq_arg2);
> +    _zzq_args[3] = (uintptr_t)(_zzq_arg3);
> +    _zzq_args[4] = (uintptr_t)(_zzq_arg4);
> +    _zzq_args[5] = (uintptr_t)(_zzq_arg5);
> +    /* TODO: use machine code to support for MSVC 64 bit*/
> +    return _zzq_result;
> +}
> +
> +#define VALGRIND_GET_NR_CONTEXT(_zzq_rlval)                       \
> +  { volatile OrigFn* _zzq_orig = &(_zzq_rlval);                   \
> +    volatile uintptr_t __addr;                                    \
> +    /* TODO: use machine code to support for MSVC 64 bit*/        \
> +    _zzq_orig->nraddr = __addr;                                   \
> +  }
> +
> +#define VALGRIND_CALL_NOREDIR_RAX ERROR
> +
> +#define VALGRIND_VEX_INJECT_IR()                                  \
> + do {                                                             \
> +   /* TODO: use machine code to support for MSVC 64 bit*/         \
> +} while (0)
> +
> +#else /* !defined(__GNUC__) && !defined(__clang__) && !defined(_MSC_VER)
*/
>
>  #error Unsupported compiler.
>
> +#endif /* defined(__GNUC__) || defined(__clang__) */
> +
>  #endif /* PLAT_amd64_win64 */
>
>  /* ------------------------ ppc32-linux ------------------------ */
> --
> 2.47.1.windows.1
>


--
         此致
礼
罗勇刚
Yours
    sincerely,
Yonggang Luo

[Valgrind-developers] [valgrind] Add 469782 to NEWS

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

https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=04edd792bd707bace8ccbf947b48b0c70b7dc5d0

commit 04edd792bd707bace8ccbf947b48b0c70b7dc5d0
Author: Paul Floyd <pj...@wa...>
Date:   Tue Dec 17 20:14:05 2024 +0100

    Add 469782 to NEWS

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

diff --git a/NEWS b/NEWS
index eb12b115cc..d78d9e9a5d 100644
--- a/NEWS
+++ b/NEWS
@@ -29,6 +29,7 @@ are not entered into bugzilla tend to get forgotten about or ignored.
 494337  All threaded applications cause still holding lock errors
 495488  Add FreeBSD getrlimitusage syscall wrapper
 496571  False positive for null key passed to bpf_map_get_next_key syscall.
+469782  Valgrind does not support zstd-compressed debug sections
 497130  Recognize new DWARF5 DW_LANG constants
 497455  Update drd/scripts/download-and-build-gcc