Mark Wielaard [Sun, 16 Jun 2024 22:27:12 +0000 (00:27 +0200)]
Close both internal pipe fds after VG_(fork) in parent and child
An VG_fork() creates a pipe between parent and child to syncronize the
two processes. The parent wants to register the child pid before the
child can run. This is done in register_sigchld_ignore.
Make sure both the parent and the child close both the read and write
file descriptors so none leak.
Mark Wielaard [Sun, 16 Jun 2024 19:23:08 +0000 (21:23 +0200)]
Don't leave fds created with --log-file, --xml-file or --log-socket open
prepare_sink_fd and prepare_sink_socket will create a new file
descriptor for the output sink. finalize_sink_fd then copies the fd
to the safe range, so it doesn't conflict with any application fds.
If we created the original fd ourselves, it was a VgLogTo_File or
VgLogTo_Socket, not VgLogTo_Fd, finalize_sink_fd should close it.
Also close socket when connecting fails in VG_(connect_via_socket).
Add a testcase for --log-file and --xml-file which prints output to
/dev/stderr
Add a new filter_xml. Note the use of --child-silent-after-fork=yes
usage in two vgtests. Maybe this should be the default for --xml=yes?
Otherwise xml output will be "corrupted" by output from a fork.
Paul Floyd [Mon, 10 Jun 2024 05:14:40 +0000 (07:14 +0200)]
FreeBSD: fixes for version 14.1
There were several leftovers from when I split FREEBSD_14
into 14_0 and 14_1 versions.
sys_break doesn't exist on arm64
There's a really annoying conditional jump error in a static copy
of strlen in ld-elf.so.1. We can't redirect the strlen, so I've
added a suppression. But it messes up test cases that use -s
to count errors.
Finally, FreeBSD 14.1 has removed a few old FreeBSD 7 syscalls.
David Benjamin [Thu, 16 May 2024 14:12:59 +0000 (10:12 -0400)]
Extract common arm64 SIMD helpers into a single header
This was copy-pasted between two files and, with the number of
extensions in aarch64, will likely need to be in many more. As the
header file defines a bunch of static, mutable state, some functions
needed to be moved to a separate .c file, to avoid weird behaviors from
C's textual inclusion model.
This also required refreshing fp_and_simd's expected output. The
fp_and_simd and fp_and_simd_v82 copies of randV128 produced slightly
different output because fp_and_simd_v82 also checked for valid f16s.
Deduplicating the code means we now apply that across the board.
NB: The fp_and_simd expected output was synthesized from what valgrind
thought the correct output was, *not* running the executable directly.
Valgrind does not seem to actually match a real Arm machine. This
divergence already existed before this commit. The divergence is in the
fmla, fcvtxn, and fcvtxn2 instructions. Looking at the corresponding
code in guest_arm64_toIR.c, I see various comments discussion how they
don't quite round correctly, so I'm guessing this is a known bug. For
now, as before this commit, I've generated the test expectations based
on the bug.
Andreas Arnez [Wed, 15 May 2024 12:32:42 +0000 (14:32 +0200)]
s390x: Support the deflate-conversion facility (DFLTCC)
So far the DFLTCC (deflate conversion call) instruction is not supported
by Valgrind. Similar to PRNO and NNPA, it is a "complex" instruction
whose memory effects cannot be adequately expressed with a dirty helper.
Add support for the DFLTCC instruction using the new "extension" mechanism
and reflect this accordingly in the supported facilities and HWCAPs.
Andreas Arnez [Wed, 15 May 2024 12:32:42 +0000 (14:32 +0200)]
Avoid use of guest_IP_AT_SYSCALL in handle_extension()
The guest state field guest_IP_AT_SYSCALL is referenced in
handle_extension(), even though it may not be defined by all
architectures. Avoid its use altogether.
Andreas Arnez [Wed, 15 May 2024 12:32:42 +0000 (14:32 +0200)]
Fix uninitialized `err' in handle_extension()
In handle_extension(), in the case of a second return from SCHEDSETJMP the
variable `err' would be used uninitialized. Fix this by avoiding any
access to `err' in this case.
Mark Wielaard [Mon, 13 May 2024 10:30:13 +0000 (12:30 +0200)]
README_DEVELOPERS: Replace b vgPlain_do_exec with b vgPlain_do_exec_inner
When building with --enable-lto vgPlain_do_exec is optimized out.
So replace the breakpoint example with vgPlain_do_exec_inner and
add a note that this is just an example and internal symbol names
might change or get optimized out.
Paul Floyd [Wed, 1 May 2024 13:17:50 +0000 (15:17 +0200)]
FreeBSD ioctl: CAMIOCOMMAND with subcommmand XPT_DEV_MATCH
CAM is the FreeBSD disk I/O subsystem (Common Access Method).
I had picked up an implementation of this from bitbucket done
by a FreeBSD dev. It was done a long time ago, around 2013 on
Valgrind 3.8. Mercurial bitbucket is now kickbucket. There
were 2 big problems.
1. Bit rot. The CAM ioctl is versioned and the version that I
copies was 0x15 whilst the current version is 0x2a, 6 versions
later. That meant that the ioctl wrapper wasn't even getting
called. In addition, as the versioning implies, the data structures
were also out of date.
2. Might also be bit rot but the data structures were copied with
QUEUE_MACRO_DEBUG_TRACE enabled. For that to work the whole
FreeBSD system needs to be built with it turned on. With that
macro turned on in Valgrind the data structures were offset
compared to the kernel, and nothing worked correctly.
There are still a huge number of CAM ioctls unimplemented.
"camcontrol devlist" now works wihtout generating false positives.
Paul Floyd [Tue, 30 Apr 2024 05:58:38 +0000 (07:58 +0200)]
DRD regtest: try to make drd/tests/pth_mutex_signal less flaky
Use a loop of 1ms usleeps rather than a single 1s usleep.
This is usually more releable in ensuring that there's a context switch and that signals
get handled in the order that we want.
Paul Floyd [Mon, 29 Apr 2024 19:11:53 +0000 (21:11 +0200)]
Darwin regtest: don't visit dSYM directories with tests/vg_regtest
Due to the extra directory depth of dSYM directories the following line
my $dashes = "-" x (50 - length $full_dir);
generates a warning on Darwin because the length of $full_dir is more than 50
("Negative repeat count does nothing"). This change makes test_one_dir return
early if the directory contains "dSYM".
Paul Floyd [Sun, 28 Apr 2024 16:19:55 +0000 (18:19 +0200)]
fd leaks: add an assert for the error extra data
Coverity claims that it could cause a NULL dereference. I don't
think it's possible. Adding an assert will at least make things
clearer if Coverity ever turns out to be right.
One reported by Fabien Launay on valgrind-developers. The other
two found with grep. There's a fourth one, but it's in a patch
file so I've left it for now.
Andreas Arnez [Thu, 25 Apr 2024 12:38:45 +0000 (14:38 +0200)]
s390x: Add support for the `NNPA' instruction
Finalize the NNPA facility support by handling the NNPA instruction
itself, using the new "extension" approach. Also reflect NNPA facility
support in HWCAP and in the STFLE bits.
Paul Floyd [Wed, 24 Apr 2024 19:34:03 +0000 (21:34 +0200)]
FreeBSD syscall: add wrapper for kcmp
For FreeBSD 15.0-CURRENT. Will need to add for 14.1 when it is out.
Mostly the same as the Linux kcmp syscall (small differences in the
supported types).
Mark Wielaard [Wed, 24 Apr 2024 18:50:43 +0000 (20:50 +0200)]
Add prereqs for tests using python 3.9+
Adds a script tests/python_test.sh that checks whether there is a python3
binary and that it supports python version 3.9 or higher. Use this script
in the various cachegrind/tests annotate vgtests as prereq.
Paul Floyd [Wed, 24 Apr 2024 18:14:40 +0000 (20:14 +0200)]
FreeBSD syswrap: wrong length for __sysctlbyname(name)
Copied and pasted from syscall where name is a pointer to array of ints
so the size has a '*sizeof(int)'. -byname name is a char* so should not
have the *4 factor.
From
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=278566
Andreas Arnez [Tue, 23 Apr 2024 14:27:11 +0000 (16:27 +0200)]
s390x: Improve operand names in trackers for PRNO
The operand names in the trackers called in the implementation of the PRNO
extension do not indicate the extension being handled. Change this by
adding the name PRNO to all these names.
Mark Wielaard [Tue, 23 Apr 2024 12:14:33 +0000 (14:14 +0200)]
Filter away "main" differences in filter_fdleak
Stack traces showing where fds were created show some differences in
the "main" function, different line numbers, or (in binary) or (below
main). Since the precise location of the original call in the main
function isn't the goal of these tests just filer them all out and
replace them with a simple "main".
Paul Floyd [Sat, 20 Apr 2024 05:54:05 +0000 (07:54 +0200)]
FreeBSD suppression: reachable for setlocale
Normally I have LANG=en_GB.UTF-8 and LC_ALL=C but if I switch to
C.UTF-8 for everything (which is the default for a fresh install)
then setlocale comes into the picture with a reachable allocation
on exit.
Mark Wielaard [Fri, 19 Apr 2024 20:46:11 +0000 (22:46 +0200)]
core errors suppression support
Add two new functions core_get_extra_suppression_info and
core_get_error_name as alternatives for tool suppression callbacks.
These functions are used in gen_suppression.
Instead of a tool name, a core error component name is "CoreError".
Two new suppression kinds FdBadCloseSupp and FdNotClosedSupp
were added. Corresponding to the FdBadClose and FdNotClosed
error kinds.
core_error_matches_suppression matches these suppression kinds
with error kinds.
core_get_extra_suppression_info and core_print_extra_suppression_use
are noops for core errors.
is_suppressible_error, supp_matches_error, load_one_suppressions_file
and show_used_suppressions have been adjusted to work with core
error kinds.
A new function VG_(found_or_suppressed_errs) helps to not output
an empty error summary if only core errors are requested, but no
errors were detected.
VG_(clo_track_fds) has been moved from pub_core_options.h to
pub_tool_options.h. And VG_(needs_core_errors) now takes a Bool
that can be set to false in the tool post_clo_init handler. This
is used in the none tool to request core errors, but disable all
reporting if no option has been given that enables such errors.
This make sure only the none/tests/fdleak_ipv4.stderr.exp needs
adjustment. For all other tests the output is exactly as before.
Improve file descriptor xml output, add fd and path elements
Add needs_xml_output to none tool so it can also output xml when
--track-fds enabled.
Use xml protocolversion 5 when clo_track_fds is enabled
Split OpenFD pathname and description. Add description when a file
descriptor is closed so it can be used in a future error. On error
print <fd> element and (if known) a <path> element.
Mark Wielaard [Wed, 20 Mar 2024 05:44:23 +0000 (01:44 -0400)]
Add core errors and use them to implement file descriptor tracker
All the tool error callbacks now have a core error equivalent.
core errors are negative (while tool errors are positive).
There are two new ones for tracking issues with file descriptors.
FdBadClose (-2) and FdNotClosed (-3).
Add following core error functions with delegates to file descriptor
specific functions (implemented in syswrap-generic):
- core_eq_Error (fd_eq_Error)
Compares core errors to detect duplicates
- core_before_pp_Error (fd_before_pp_Error)
Currently prints nothing for known core errors and
exists with FATAL for unknown core errors
- core_pp_Error (fd_pp_Error)
For FdBadClose prints the backtraces for the file descriptor was
opened, where it was originally closed and where it was closed again.
For FdNotClosed prints where the file descriptor was opened.
- core_update_extra (fd_update_extra)
Returns the size of the BadCloseExtra or FdNotClosedExtra struct
which data needs to be saved (the fd number, pathname/description
and previous backtraces).
We now accept the error (ExeContext) where to be NULL.
This is necessary for reporting not closed file descriptors when
the descriptor is inherited from the parent (so wasn't actually
created and doesn't have a 'where' in the current process code).
All the testcases still pass since the (stderr) output is the
same. But now they count as "real" errors. And so --error-exitcode
does now also work for file descriptor errors or leaks.
https://bugs.kde.org/show_bug.cgi?id=362680
Co-authored-by: Alexandra Hájková <ahajkova@redhat.com>
projects / thirdparty / valgrind.git / log
