glibc doesn't guarantee anything about setrlimit with a NULL limit argument.
It could just crash (if it needs to adjust the limit) or might silently
succeed (as newer glibc do). Just remove the extra check.
See also the "setrlimit change to prlimit change in behavior" thread:
https://sourceware.org/ml/libc-alpha/2017-10/threads.html#00830
Compiler may optimize out call to cbrt. Change test to prevent that.
Otherwise, the test does not exercise a desired codepath for cbrt, and it
prints precalculated value.
In 64 bit mode there's no way to just save eflags so we save the
whole of rflags but we were doing so to a 32 bit variable!
Replace that with proper rflags support that knows it is dealing
with the full 64 bit flags word in 64 bit mode.
Implements a new version of VEX register allocator which
keeps the main state per virtual registers, as opposed
to real registers in v2. This results in a simpler and
cleaner design and much simpler implementation.
It has been observed that the new allocator executes 20-30%
faster than the previous one but could produce slightly worse
spilling decisions. Overall performance improvement when running
the Valgrind performance regression test suite has been observed
in terms of a few percent.
The new register allocator (v3) is now the default one.
The old register allocator (v2) is still kept around and can be
activated with command line option '--vex-regalloc-version=2'.
Fixes BZ#381553.
This patch removes two compiler warnings from the test:
vfp.c: In function 'handler':
vfp.c:260:4: warning: implicit declaration of function 'exit'
[-Wimplicit-function-declaration]
exit(0);
^
vfp.c:260:4: warning: incompatible implicit declaration of built-in
function 'exit'
vfp.c: At top level:
vfp.c:258:13: warning: 'handler' defined but not used [-Wunused-function]
static void handler(int sig)
^
The #ifdef HAS_VSX guard is wrongly placed. It makes the standard
include headers not be used. Causing a build failure. Fix by moving
the #ifdef HAS_VSX after the standard includes.
https://bugs.kde.org/show_bug.cgi?id=381272
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16450
into three different files with regards to using
fs prefix, gs prefix or no fs/gs prefix.
Pertains to BZ#379525.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16421
Note: this just unbreaks the build on avx + non_adx capable systems
(such as gcc farm gcc20).
adx capable system should probably be better handled:
* ./tests/x86_amd64_features cannot check for adx flag
(so fb_test_amd64 is run if compiled and system is avx capable, which
might give problems if gcc/as can compile the test, but the cpu
cannot execute adx instructions)
* on an adx capable system, a native run of cpuid tells it is adx capable
but under valgrind, cpuid reports the valgrind synthetic cpu is not adx
capable.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16401
- Fix mffs instruction, FPSCR, C_FPCC field access is not
working correctly.
- Fix xscmpexpdp again still had a bug.
- Remove duplicate tests from test_isa_3_0.c
- Comment out he tests for the new ISA 3.0B instructions. There
seem to be some issues between the simulator and early HW that
need to be sorted out.
Bugzilla 379925, VEX commit 3375
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16396
Update inline assembly with .set noreorder.
This prevents assembler from reordering the instructions.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16394
Follow up to r16372.
The test is now built if the compiler understands AVX
instructions. The test is executed if the host speaks AVX
is well. Finally bogus printf format flags were fixed as
they caused the output to be inconsistent on Linux and Solaris.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16393
Fix for stxvx instruction.
Add LE support for stxv instruction
Fix for the xscmpexpdp instruction
Bugzilla: 379703
VEX commit: 3374
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16391
condition codes. This is originally by Fabrice Bellard (GPL2+'d), with MD5
support from Alexander Peslyak (public domain) and has been extended to cover
ADOX and ADCX as per bug #360415.
The program generates more than 800MB of output, which it MD5 sums, so the final
MD5 sum serves as the pass/fail check. It takes roughly a minute to run.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16372
- Invalid memory access fixed;
- Inappropriate immediate value on ROTR test changed;
- Stricter tests for LWL/LWR instructions.
Patch by Aleksandar Rikalo.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16363
Compiler produces position independent executables (PIE) by default
which gets in the way of some unit tests.
Fixes BZ#377066.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16349
mffscdrni, mffsce, mffscrn, mffscrni, mffsl. vmsumudm.
Forgot to add the new files.
Vex commit 3359 Has the source code changes for the instruction and OV32, CS32
support
Valgrind commit 16329 updated the existing files
This commit adds all the new files.
Valgrind bugzilla 378931
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16330
mffscdrni, mffsce, mffscrn, mffscrni, mffsl. vmsumudm.
Additionally, the OV32 and CA32 bits were introduced in ISA 3.0 but
Valgrind add support for setting these bits for ISA 3.0. The OV32 and CA32
bits must now be set on a number of pre ISA 3.0 instructions. So now the
instructions produce different results in the XER register. Thus we need pre
and post ISA 3.0 expect files. Command line options were added to thee
pre ISA test cases so instructions that didn't change could be run with one
set of command line args. The instructions that have different XER results
are run using a different set of command line args. The tests were split into
two, one for instructions that didn't change on for instructions that do
change under ISA 3.0. We then create ISA3.0 expect files only for the tests
that run differently. By doing this we minimized the size of the expect files
needed.
Vex commit 3359 Has the source code changes for the instruction and OV32, CS32
support
This commit is all the test case changes, adding the new test case files.
Valgrind bugzilla 378931
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16329
This implements a fallback LL/SC implementation as described in bug 344524.
Valgrind side changes:
* Command line plumbing for --sim-hints=fallback-llsc
* memcheck: handle new arm64 guest state in memcheck/mc_machine.c
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16309
Update the test cvm_atomic_thread so it can be executed on BE boards too.
Reuse the stdout.exp file.
Based on patch from Tamara Vlahovic.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16298
Patch from Matthias Schwarzott (zzam@gentoo.org). The patch removes
a volatile memory read which was only there to stop compilers warning
about |format| being unused.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16290
Leave the old exp file that covers cases in which __addtf3 and __subtf3
did not take into account rounding modes. New exp file is the same file
that already exists in mips32 folder, so we just create a symbolic link
to it.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16201
Fix 373192 Calling posix_spawn in glibc 2.24 completely broken
Functionally, this patch just does the following 2 changes to the
fork clone handling:
* It does not mask anymore CLONE_VFORK :
The only effect of this flag is to suspend the parent, waiting for
the child to either exit or execve.
If some applications depends on this synchronisation, better keep it,
as it will not harm to suspend the parent valgrind waiting for the
child valgrind to exit or execve.
* In case the guest calls the clone syscall providing a non zero client stack,
set the child guest SP after the syscall, before executing guest instructions.
Not setting the guest stack ptr was the source of the problem reported
in the bugs.
This also adds a test case none/tests/linux/clonev.
Before this patch, test gives a SEGV, which is fixed by the patch.
The patch is however a lot bigger : this fix was touching some (mostly
identical/duplicated) code in all the linux platforms.
So, the clone/fork code has been factorised as much as possible.
This removes about 1700 lines of code.
This has been tested on:
* amd64
* x86
* ppc64 be and le
* ppc32
* arm64
This has been compiled on but *not really tested* on:
* mips64 (not too clear how to properly build and run valgrind on gcc22)
It has *not* been compiled and *not* tested on:
* arm
* mips32
* tilegx
* darwin (normally, no impact)
* solaris (normally, no impact)
The changes are relatively mechanical, so it is not impossible that
it will compile and work out of the box on these platforms.
Otherwise, questions welcome.
A few points of interest:
* Some platforms did have a typedef void vki_modify_ldt_t,
and some platforms had no definition for this type at all.
To make it easier to factorise, for such platforms, the following has
been used:
typedef char vki_modify_ldt_t;
When the sizeof vki_modify_ldt_t is > 1, then the arg syscall is checked.
This is somewhat a hack, but was simplifying the factorisation.
* for mips32/mips64 and tilegx, there is a strange unconditional assignment
of 0 to a register (guest_r2 on mips, guest_r0 on tilegx).
Unclear what this is, in particular because this is assigned whatever
the result of the syscall (success or not).
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16186
Add BE variant of exp file for cvm_atomic test.
This fixes none/tests/mips64/cvm_atomic for Cavium MIPS64 BE systems.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16183
Allow execution of none/tests/mips32/round_fpu64 test if it is built with
-mfpxx. It gets us more coverage eventually.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16181
When signal SIGINT was used, it got propagated to the parent shell
and therefore the regression test suite was terminated.
With SIGTERM only the parent exits, as is supposed.
Verified that this changed test case still reproduces the original problem
on Linux and Solaris, when r16140 is temporarily undone.
Follow up to r16140 and r16143 and BZ#372600.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16147
