Petar Jovanovic [Sun, 5 Aug 2012 02:22:52 +0000 (02:22 +0000)]
Fixing incorrect case in putSyscallArgsIntoGuestState for __NR_syscall.
Previously unused arg8 is now used to hold NR_syscall value when that call is
executed. This is important, so putSyscallArgsIntoGuestState can correctly
fill up the guest state. This is MIPS-only change.
Julian Seward [Sat, 4 Aug 2012 19:23:54 +0000 (19:23 +0000)]
Try to fix up ARM stack alignment stuff following r12811. This
fixes two problems: first, r11 (aka fp) can't be used in assembly
for whatever reason. Secondly, the "bic sp,sp,#7" is not allowed
in Thumb mode, so work around that too.
Remove 284540 from NEWS fixed section.
284540 was not about performance but about the presentation
of results.
Revision 12824 (optimising the suppr matching) should not have
marked 284540 as fixed.
fix 284540 (optimise suppression matching)
Before this patch, matching an error stack trace with many suppression
patterns was implying to repeating the translation of the IPs of the
stack trace to the function name or object name for each suppr pattern.
This patch introduces a "lazy input completer" in the generic match
so that an IP is (in the worst case) translated once to its function
name and once to its object name.
It is a "lazy" completer in the sense that only the needed IP to fun or obj
name are done.
On a artificial test case, has given a factor 3 in performance.
On another big (real) application, gave a factor 2 to 3.
(there was less matching to do, but probably more debug info to search).
match-overrun.supp completed to have non matching suppr first to
better exercise the lazy completer.
Florian Krohm [Fri, 3 Aug 2012 20:28:21 +0000 (20:28 +0000)]
Fix testcase. Iterating over an array of invalid insns to test them all
does not work, as conversion will stop at the first invalid character.
So... need to loop over them.
Julian Seward [Thu, 2 Aug 2012 22:08:53 +0000 (22:08 +0000)]
Back off from 32 alignment of various guest state bits. We don't
actually need it, and gcc on MacOS simply fails (for whatever reason)
to actually honour requests for 32 alignment, and fall back to the
previous setting (16 alignment).
Julian Seward [Thu, 2 Aug 2012 10:40:31 +0000 (10:40 +0000)]
Back out r12466, which stopped non-MAP_FIXED mmaps by the client on
Darwin from returning address zero (however insane that is). r12466
appears to cause other applications to break (TextEdit, for one).
Tom Hughes [Thu, 2 Aug 2012 09:23:45 +0000 (09:23 +0000)]
Ensure CALL_FN_xx macros align the stack properly
The CALL_FN_xx macros in valgrind.h perform function calls by
signalling to valgrind using the client request system. Because
they are making function calls which are invisible to the compiler
they need to make sure that any stack alignment constraints
imposed by the ABI are enforced when making the call.
This commit enforces 16 byte alignment for x86, amd64, ppc32 and
ppc64 platforms, and 8 byte alignment for arm platforms.
It does not touch s390x where the ABI requires 8 byte alignment to
be maintained at all times, not just when making a function call.
It also does not touch mips32 as I'm not currently aware what if
any alignment constraints exist there.
Fixes BZ#304054 and observed alignment faults on amd64 when running
the regtests using a valgrind compiled with gcc 4.7 releases.
Avoid asserting when a segment is mapped both rw and rx.
If a segment is mapped with permission rwx, then map->rx
and map->rw will be true.
But due to the if (map->rx) {
...
} else if (map->rw) {
...
the (map->rw) part will not be executed.
If this mapping is the one which "gives" the nonempty rw map,
then this mapping will not be seen, and the following
vg_assert(has_nonempty_rw);
will fail.
This assert can be reproduced by doing
setarch i686 -X
./vg-in-place --tool=none none/tests/map_unmap
Note: the setarch i686 -X above has as effect to make all read
mapping also executable. So, a rw mapping becomes rwx and then
triggers the above asserts.
The setarch i686 -X also introduces a discrepancy between
the kernel mappings (rwx) and the valgrind aspacemgr view
(which believes it is a rw mapping).
This discrepancy causes a crash if giving --sanity-level=3.
A possible fix is to have valgrind calling the personality system call
and detecting if the READ_IMPLIES_EXEC bit (the -X arg to setarch)
was set, and then modify aspacemgr so that all read mapped segments
are automatically mapped x also.
This commit is the minimal fix allowing to run executables
launched with this READ_IMPLIES_EXEC.
* Option --vex-iropt-precise-memory-exns has been removed.
It is replaced by --vex-iropt-register-updates which accepts
3 values : 'unwindregs-at-mem-access' (replacing
--vex-iropt-precise-memory-exns=no), 'allregs-at-mem-access'
(replacing --vex-iropt-precise-memory-exns=yes)
and a new value 'allregs-at-each-insn'.
'allregs-at-each-insn' allows the Valgrind gdbserver to always
show up to date values to GDB.
* For tools replacing the malloc library (e.g. Memcheck, Helgrind, ...),
the option --redzone-size=<number> allows to control the padding
blocks (redzones) added before and after each client allocated block.
Smaller redzones decrease the memory needed by Valgrind. Bigger
redzones increase the chance to detect blocks overrun or underrun.
Petar Jovanovic [Fri, 27 Jul 2012 22:35:35 +0000 (22:35 +0000)]
MIPS specific expected output for memcheck/tests/sigkill.
MIPS has different signal values, so it has to have its own expected output for
the tests that deal with signal values.
This fixes (false) failure in memcheck/tests/sigkill.
Petar Jovanovic [Fri, 27 Jul 2012 16:25:32 +0000 (16:25 +0000)]
MIPS specific exp-out for none/tests/async-sigs.
MIPS has different signal values, so it has to have its own expected output for
the tests that deal with signal values.
This fixes (false) failure in none/tests/async-sigs.
Due to s390's dealings with invalid insns we need to do two things
- advance the guest_IA to the next insn after raising the signal
- adjusting the address in a complaint to point to the failing insn
(after guest_IA has been advanced)
Update testcases .exp files.
Better document the limitation of gdbsrv for registers and flags values
During investigations of 303963, Josef found that flags are not always
up to date and that --vex-guest-max-insns=1 ensures flags values
are (more?) correct.
=> enhance the paragraph in the gdbserver limitations to reference
this option and give an idea of the performance impact of the other
options helping to increase the precision of registers and flags.
Have perf/vg_perf handling EXTRA_REGTEST_OPTS
Similarly to tests/vg_regtest, allow to run all perf tests with extra options.
(note: it was preferred to use the same env var name).
Petar Jovanovic [Mon, 23 Jul 2012 00:11:10 +0000 (00:11 +0000)]
Avoid checking for bitset initialization if futex_wait_bitset is meant to fail.
Glibc deliberately passes random value for the sixth parameter when calling
FUTEX_WAIT_BITSET | FUTEX_CLOCK_REALTIME. This is a regular case of using the
Futex API, so V should not complain that "Syscall param futex(val3) contains
uninitialised byte(s)", if the futex does not have a specified value initially.
For more info, see function pthread_initialize_minimal_internal at:
glibc/nptl/nptl-init.c.
Add 'aesdec' to the set of insns checked for assembler level SSE4.2
support (Rich Coe) and tidy up a couple of other bits of assembly by
giving them trashed-register lists.
ppc64-linux _start: don't assume ._start_in_C_linux -- the function
descriptor for _start_in_C_linux -- actually exists. Instead call the
entry point directly.
If the compiler being used to build is clang, remember to add
-new_linker to the stuff given to the linker, else link failures
result on x86. Based on a suggestion by Christoph Erhardt. Fixes
#295427.
Add testcases for CU21. Update opcode table.
WRT the memcheck test: the good news is we get all the complaints
we want. The bad news is:
- the line numbers from within do_cu21 are off
- there are 2 complaints when attempting to convert from
an uninitialised input buffer instead of one. One of them has a
completely bogus line number.
Possibly due to bad debug info? Let's see what happens on machines
with more recent GCCs.
Observe guards on dirty helpers in memcheck.
This means, that any guest state and/or memory accesses of the helper (and
complaints about those) only occur if the guard expression is true at runtime.
Definedness of parameters that the helper might have is *always* checked,
as parameters are evaluated regardless of the guard expression.
New functions: expr2vbits_guarded_Load and gen_guarded_load_b.
Fix 303624 segmentation fault on Android 4.1 (e.g. on android emulator or Galaxy Nexus OMAP)
Valgrind was crashing systematically on Android 4.1.
This crash is caused by AT_IGNORE-ing AT_BASE.
This AT_IGNORE was needed to have breakpoints in shared libs
be handled properly (not very clear what is the problem
in the interaction between Valgrind GDBSERVER, AT_BASE and GDB).
Waiting to better understand all this, as a temporary bypass,
this patch ensures we do not ignore the AT_BASE on android.
The possible consequence is that breakpoints might be inserted
by the Valgrind gdbserver at wrong addresses in shared lib.
(any feedback on that is welcome).
Valgrind was build and then "proved" to work on Android emulator 4.0
and emulator 4.1, by using memcheck on one executable.
patch that improves the speed of the leak search by up to 40% (on amd64)
Scanning 1GB of random values made of 200_000 malloc-ed block is
(on amd64) changing from (about) 17 seconds to (about) 10 seconds.
On x86, it goes from 153 seconds to 129 seconds.
(this huge difference between x86 and amd64 leak search time
for this random test is because a random value has about one chance
on 4 to be in the addressable memory on x86 and so the dichotomic
search in the list of malloc-ed blocks is called for a lot more
values than on amd64).
Basically, there are 3 optimisations:
1. call MC_(is_within_valid_secondary) only at the beginning of a
secondary map (and not for each Word).
2. call SETJMP only at the beginning of a page (and not for each Word)
3. Validate an aligned word using get_vabits8 rather than get_vabits2.
Each of the above optimisation more or less improves by 2 seconds.
(to go from 17 seconds to 10 seconds).
Small cleanup: use VG_TRACK (when possible) to call tool tracking functions
(spotted by Julian)
Note: there is a second occurence of call to track_post_mem_write in the
same file; but this second occurence is better done with an "if".
Petar Jovanovic [Mon, 16 Jul 2012 14:35:26 +0000 (14:35 +0000)]
Add some missing syscall numbers for MIPS.
The header file has been extended with the missing syscall numbers for MIPS,
and sys_prlimit64 has been enabled. This will make none/tests/rlimit64_nofile
pass.
projects / thirdparty / valgrind.git / log
