]>
git.ipfire.org Git - thirdparty/valgrind.git/log
Nicholas Nethercote [Thu, 6 Apr 2006 22:59:35 +0000 (22:59 +0000)]
Fix vg_perf to work with older Valgrind versions.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5833
Nicholas Nethercote [Thu, 6 Apr 2006 14:21:42 +0000 (14:21 +0000)]
Commit a patch from Olly Betts which avoids a possible problem with
COUNT_LEAKS on 64-bit machines.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5832
Nicholas Nethercote [Thu, 6 Apr 2006 14:04:48 +0000 (14:04 +0000)]
Add a FAQ.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5831
Nicholas Nethercote [Thu, 6 Apr 2006 13:33:48 +0000 (13:33 +0000)]
Use better variable names.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5830
Julian Seward [Thu, 6 Apr 2006 13:31:17 +0000 (13:31 +0000)]
Avoid compiler warnings on ppc64-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5829
Nicholas Nethercote [Thu, 6 Apr 2006 13:29:18 +0000 (13:29 +0000)]
update
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5828
Julian Seward [Wed, 5 Apr 2006 23:06:31 +0000 (23:06 +0000)]
Minor scheduler tidyings:
- rename os_thread_t to ThreadOSstate
- remove unused ThreadState.syscall_result_set field
- fix some comments
- add an assertion in VG_(scheduler_init)
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5827
Julian Seward [Tue, 4 Apr 2006 17:35:48 +0000 (17:35 +0000)]
Test program which uses templates to generate very long stabs (debug
info) strings. This crashes 3.1.1 when the test is compiled with
-gstabs.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5826
Julian Seward [Tue, 4 Apr 2006 16:57:15 +0000 (16:57 +0000)]
Comment-only change.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5825
Julian Seward [Tue, 4 Apr 2006 15:43:30 +0000 (15:43 +0000)]
Un-break make dist.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5824
Julian Seward [Tue, 4 Apr 2006 15:12:44 +0000 (15:12 +0000)]
Cleanup/restructure m_debuginfo, as described in
coregrind/m_debuginfo/README.txt.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5823
Julian Seward [Tue, 4 Apr 2006 13:28:50 +0000 (13:28 +0000)]
Round up known bugs in 3.1.1 and assign tentative priorities.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5822
Julian Seward [Tue, 4 Apr 2006 03:29:48 +0000 (03:29 +0000)]
Add a comment that really should have been in r5820.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5821
Julian Seward [Tue, 4 Apr 2006 03:08:49 +0000 (03:08 +0000)]
ppc32-linux: work around assemblers which can't do Altivec, by
emitting the relevant instruction directly. Fixes a build problem on
Debian 3.1 (ppc32).
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5820
Tom Hughes [Mon, 3 Apr 2006 16:38:33 +0000 (16:38 +0000)]
Enable sigsuspend on x86-linux and ppc32-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5819
Tom Hughes [Mon, 3 Apr 2006 16:37:30 +0000 (16:37 +0000)]
Don't use the presence of a filename to decide if a segment in the
initial /proc/self/maps to decide if the segment is an AnonV or FileV
segment as some systems don't report the filename. Use the device
and inode numbers instead. Fixes bug #124528.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5818
Nicholas Nethercote [Mon, 3 Apr 2006 14:25:23 +0000 (14:25 +0000)]
Add an optional post-test check. Based on a patch from Josef W.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5817
Josef Weidendorfer [Sun, 2 Apr 2006 22:23:27 +0000 (22:23 +0000)]
Callgrind: add 4 regression tests
The simwork tests check different cache simulator
options/modes. These tests should be extended to
check for the produced call graph.
The threads check tests the callgrind mode where
counts are aggregated separate per thread.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5816
Julian Seward [Sun, 2 Apr 2006 16:21:44 +0000 (16:21 +0000)]
Fix another place where function addresses passed to Vex needed to be
dereferenced properly. (Logically part of r5814).
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5815
Julian Seward [Sun, 2 Apr 2006 15:53:59 +0000 (15:53 +0000)]
Fix ppc64-linux breakage - get the function entry address, not the
descriptor address.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5814
Julian Seward [Sun, 2 Apr 2006 02:59:35 +0000 (02:59 +0000)]
Add 64-bit test output.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5813
Julian Seward [Sun, 2 Apr 2006 02:27:52 +0000 (02:27 +0000)]
Don't print an address on stdout as it doesn't get automagically
changed into 0x.........
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5812
Julian Seward [Sun, 2 Apr 2006 01:53:01 +0000 (01:53 +0000)]
Another shadow memory test. This one does a huge number of loads and
stores of char/short/int/int64/double at random offsets and hence
alignments in an array. It does it in a way in which the computation
just computes the expected V bits, and hence can check whether these
seem correct.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5811
Julian Seward [Sun, 2 Apr 2006 00:58:45 +0000 (00:58 +0000)]
Get rid of unused #define.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5810
Julian Seward [Sun, 2 Apr 2006 00:58:01 +0000 (00:58 +0000)]
In check_all, allow two different acceptable byte values, so as to
cover the behaviour of all possible variants of float loads/stores.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5809
Nicholas Nethercote [Sat, 1 Apr 2006 23:06:29 +0000 (23:06 +0000)]
Partial fix for the sh-mem.c failure on PPC32. This should make it work
on PPC32 now but break it on the other platforms. Julian will commit a
change to ensure the 32-bit floats are copied through the FP regs on all
platforms to make the broken ones work again.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5808
Josef Weidendorfer [Fri, 31 Mar 2006 19:34:51 +0000 (19:34 +0000)]
Simple regression test for callgrind:
run a custom client request.
By doing this, I found out that callgrind.h still defined
client requests for VG 2 :-( Obviously, nobody was using
them. This is fixed, and other small things to make the
test run, too.
Notice for myself: regression tests are needed.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5807
Josef Weidendorfer [Fri, 31 Mar 2006 13:16:15 +0000 (13:16 +0000)]
Option --dump-every-bb now takes a 64bit value on user
request. Update/clarify documentation.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5806
Josef Weidendorfer [Fri, 31 Mar 2006 12:53:22 +0000 (12:53 +0000)]
Fix typos
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5805
Nicholas Nethercote [Fri, 31 Mar 2006 12:24:37 +0000 (12:24 +0000)]
update Memcheck description
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5804
Nicholas Nethercote [Fri, 31 Mar 2006 12:05:04 +0000 (12:05 +0000)]
Make the BSD-style license notices slightly more accurate.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5803
Nicholas Nethercote [Fri, 31 Mar 2006 11:57:59 +0000 (11:57 +0000)]
Terminology change: previously in Memcheck we had the four states:
noaccess, writable, readable, other
Now they are:
noaccess, undefined, defined, partdefined
As a result, the following names:
make_writable, make_readable,
check_writable, check_readable, check_defined
have become:
make_mem_undefined, make_mem_defined,
check_mem_is_addressable, check_mem_is_defined, check_value_is_defined
(and likewise for the upper-case versions for client request macros).
The old MAKE_* and CHECK_* macros still work for backwards compatibility.
This is much better, because the old names were subtly misleading. For
example:
- "readable" really meant "readable and writable".
- "writable" really meant "writable and maybe readable, depending on how
the read value is used".
- "check_writable" really meant "check writable or readable"
The new names avoid these problems.
The recently-added macro which was called MAKE_DEFINED is now
MAKE_MEM_DEFINED_IF_ADDRESSABLE.
I also corrected the spelling of "addressable" in numerous places in
memcheck.h.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5802
Nicholas Nethercote [Thu, 30 Mar 2006 10:26:44 +0000 (10:26 +0000)]
Avoid compile warnings for tinycc.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5801
Julian Seward [Wed, 29 Mar 2006 19:33:54 +0000 (19:33 +0000)]
Hmm. Why is this file still here? (It got folded into aspacemgr.c
a while back.)
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5800
Julian Seward [Wed, 29 Mar 2006 03:15:25 +0000 (03:15 +0000)]
Delete VG_(sigaltstack) and VG_(sigpending) as they are unused.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5799
Julian Seward [Tue, 28 Mar 2006 23:20:38 +0000 (23:20 +0000)]
Delete VG_(poll), as it is unused.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5798
Nicholas Nethercote [Tue, 28 Mar 2006 12:51:02 +0000 (12:51 +0000)]
Change to some better function names.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5797
Nicholas Nethercote [Tue, 28 Mar 2006 12:35:08 +0000 (12:35 +0000)]
Change some uses of bytes to bits for consistency.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5796
Julian Seward [Tue, 28 Mar 2006 01:02:38 +0000 (01:02 +0000)]
Only complain once for each missing wrapper.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5795
Nicholas Nethercote [Tue, 28 Mar 2006 00:05:16 +0000 (00:05 +0000)]
tweak
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5794
Nicholas Nethercote [Mon, 27 Mar 2006 12:23:44 +0000 (12:23 +0000)]
Partial release notes for 3.2.0, talking about the performance improvements
and removal of Addrcheck.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5793
Nicholas Nethercote [Mon, 27 Mar 2006 11:55:09 +0000 (11:55 +0000)]
update
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5792
Nicholas Nethercote [Mon, 27 Mar 2006 11:37:07 +0000 (11:37 +0000)]
Merge in the COMPVBITS branch to the trunk. This is a big change to
Memcheck, replacing the 9-bits-per-byte shadow memory representation to a
2-bits-per-byte representation (with possibly a little more on the side) by
taking advantage of the fact that extremely few memory bytes are partially
defined.
For the SPEC2k benchmarks with "test" inputs, this speeds up Memcheck by a
(geometric mean) factor of 1.20, and reduces the size of shadow memory by a
(geometric mean) factor of 4.26.
At the same time, Addrcheck is removed. It hadn't worked for quite some
time, and with these improvements in Memcheck its raisons-d'etre have
shrivelled so much that it's not worth the effort to keep around. Hooray!
Nb: this code hasn't been tested on PPC. If things go wrong, look first in
the fast stack-handling functions (eg. mc_new_mem_stack_160,
MC_(helperc_MAKE_STACK_UNINIT)).
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5791
Nicholas Nethercote [Mon, 27 Mar 2006 00:39:43 +0000 (00:39 +0000)]
Avoid unnecessary whitespace at end of printed lines.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5790
Tom Hughes [Fri, 24 Mar 2006 16:48:19 +0000 (16:48 +0000)]
Handle msg_flags correctly - for recvmsg this field is written by the
kernel not read.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5789
Julian Seward [Wed, 22 Mar 2006 11:18:50 +0000 (11:18 +0000)]
Update expected output following r5785; also add a missing 's'.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5788
Nicholas Nethercote [Tue, 21 Mar 2006 23:23:22 +0000 (23:23 +0000)]
Update
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5787
Nicholas Nethercote [Tue, 21 Mar 2006 23:15:43 +0000 (23:15 +0000)]
Make the unrecognised-instruction error message more informative.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5786
Tom Hughes [Tue, 21 Mar 2006 10:58:35 +0000 (10:58 +0000)]
Add wrappers for a load of new 2.6.16 system calls. Fixes bug #123248.
Also expands pathname arguments as strings in a lot more system call
trace messages and fixed the poll wrapper to not be x86 specific.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5785
Julian Seward [Tue, 21 Mar 2006 02:23:27 +0000 (02:23 +0000)]
Enable sys_fchdir on ppc32-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5784
Tom Hughes [Mon, 20 Mar 2006 15:36:45 +0000 (15:36 +0000)]
Sync system call lists with the 2.6.16 kernel.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5783
Tom Hughes [Mon, 20 Mar 2006 15:36:28 +0000 (15:36 +0000)]
Make hard limit enforcement work for RLIMIT_DATA and RLIMIT_STACK in
the setrlimit system call wrapper.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5782
Josef Weidendorfer [Mon, 20 Mar 2006 10:29:30 +0000 (10:29 +0000)]
Callgrind merge: documentation
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5781
Josef Weidendorfer [Mon, 20 Mar 2006 10:27:30 +0000 (10:27 +0000)]
Callgrind merge: code
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5780
Julian Seward [Mon, 20 Mar 2006 00:54:54 +0000 (00:54 +0000)]
Add a comment about the library's license.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5779
Julian Seward [Sun, 19 Mar 2006 18:19:11 +0000 (18:19 +0000)]
Yet another essay: document the MPI wrapper library.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5778
Julian Seward [Thu, 16 Mar 2006 11:33:29 +0000 (11:33 +0000)]
Update
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5776
Julian Seward [Thu, 16 Mar 2006 11:33:14 +0000 (11:33 +0000)]
Merge stuff for 3.1.1.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5775
Julian Seward [Thu, 16 Mar 2006 11:31:29 +0000 (11:31 +0000)]
upmerge r5768 (Allow a thread to spin longer when yielding before
switching to a different thread.)
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5774
Julian Seward [Wed, 15 Mar 2006 17:27:56 +0000 (17:27 +0000)]
Recycle Dirk's glibc-2.3.90 suppressions, since at some point it will
really become glibc-2.4.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5769
Julian Seward [Wed, 15 Mar 2006 11:50:32 +0000 (11:50 +0000)]
A couple of initialisations to keep gcc-4.1.0 happy.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5765
Julian Seward [Tue, 14 Mar 2006 00:56:29 +0000 (00:56 +0000)]
Minor futzing (fontification, etc) of the function-wrappers documentation.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5764
Julian Seward [Mon, 13 Mar 2006 13:40:57 +0000 (13:40 +0000)]
First pass at documenting how to use the function-wrapping facility.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5763
Julian Seward [Sun, 12 Mar 2006 19:28:34 +0000 (19:28 +0000)]
Sigh; rename mpiwrap.c to libmpiwrap.c so that 'make dist' doesn't barf
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5762
Julian Seward [Sun, 12 Mar 2006 18:29:18 +0000 (18:29 +0000)]
When testing 'mpicc', try at least somewhat to test for the primary
target, as that is what it will be used to compile for.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5761
Julian Seward [Sun, 12 Mar 2006 17:27:44 +0000 (17:27 +0000)]
Unbreak mtoctf configure check.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5760
Julian Seward [Sun, 12 Mar 2006 16:57:07 +0000 (16:57 +0000)]
Mess with this even more so that fn_1 doesn't return zero.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5759
Julian Seward [Sun, 12 Mar 2006 16:48:14 +0000 (16:48 +0000)]
Add missing CALL_FN_* macros for ppc64-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5758
Julian Seward [Sun, 12 Mar 2006 16:47:10 +0000 (16:47 +0000)]
Work around lameness in older ppc assemblers.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5757
Julian Seward [Sun, 12 Mar 2006 16:44:05 +0000 (16:44 +0000)]
LAM/MPI compile fix
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5756
Julian Seward [Sun, 12 Mar 2006 14:15:23 +0000 (14:15 +0000)]
Updated expected output following r5753.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5755
Julian Seward [Sun, 12 Mar 2006 13:39:15 +0000 (13:39 +0000)]
Add CALL_FN_* missing cases for ppc32-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5754
Julian Seward [Sun, 12 Mar 2006 13:38:46 +0000 (13:38 +0000)]
TRASH_IREGS: add enough live variables to make ppc32 spill (not easy)
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5753
Julian Seward [Sun, 12 Mar 2006 13:37:19 +0000 (13:37 +0000)]
sizeofOneNamedTy(): handle long double correctly on non-x86/amd64 platforms.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5752
Julian Seward [Sun, 12 Mar 2006 13:36:06 +0000 (13:36 +0000)]
Add ppc32-linux details.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5751
Julian Seward [Sun, 12 Mar 2006 00:39:19 +0000 (00:39 +0000)]
Fix a redirection bug created by the overhaul of m_redir.c. Bug
caused hardwired Actives to be discarded at the first call to
VG_(redir_notify_delete_SegInfo). This causes crashes on amd64-linux
by throwing away the hardwired redirections
0xFFFFFFFFFF600000 -> VG_(amd64_linux_REDIR_FOR_vgettimeofday)
0xFFFFFFFFFF600400 -> VG_(amd64_linux_REDIR_FOR_vtime)
which I'm sure got mentioned at some point recently, but I cannot find
the email now.
Bug does not affect 3.1.X, since that line contains the old m_redir.c
implementation.
Bug does not affect any platform other than amd64-linux since no
others have hardwired Actives at the start.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5750
Julian Seward [Sun, 12 Mar 2006 00:35:42 +0000 (00:35 +0000)]
A 64-bit fix
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5749
Julian Seward [Sat, 11 Mar 2006 16:19:14 +0000 (16:19 +0000)]
Complete CALL_FN_W_* wrappers for amd64-linux.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5748
Julian Seward [Sat, 11 Mar 2006 13:20:41 +0000 (13:20 +0000)]
Majorly beef up this test:
* test all wrapped-function arities from 0 to 12
* try hard to run both callers and callees out of integer registers,
so as to detect problems where the CALL_FN_* macros do not
properly save registers around the call
This will cause failure in building the regtests on all non-x86
platforms. Will fix shortly.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5747
Julian Seward [Sat, 11 Mar 2006 12:58:03 +0000 (12:58 +0000)]
Fix comment typo.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5746
Julian Seward [Sat, 11 Mar 2006 12:57:41 +0000 (12:57 +0000)]
Add missing x86-linux CALL_FN_W_11W macro.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5745
Nicholas Nethercote [Sat, 11 Mar 2006 06:48:20 +0000 (06:48 +0000)]
Add a comment.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5744
Julian Seward [Fri, 10 Mar 2006 22:17:57 +0000 (22:17 +0000)]
Quadrics MPI build fixes.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5743
Julian Seward [Fri, 10 Mar 2006 22:17:02 +0000 (22:17 +0000)]
Make sure we get the right headers, not whatever happens to be already
installed.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5742
Julian Seward [Fri, 10 Mar 2006 21:48:29 +0000 (21:48 +0000)]
Add a (rather ingenious, if I do say so myself) program to test the
type-walking stuff in mpiwrap.c.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5741
Julian Seward [Fri, 10 Mar 2006 21:41:01 +0000 (21:41 +0000)]
Even more MPI wrappers.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5740
Julian Seward [Fri, 10 Mar 2006 18:37:45 +0000 (18:37 +0000)]
Add a bunch more wrappers and generally tidy up a bit.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5739
Julian Seward [Fri, 10 Mar 2006 14:51:19 +0000 (14:51 +0000)]
* try to handle MPI_{LONG,DOUBLE,SHORT}_INT better (despite the best
efforts of the MPI spec to screw this up)
* Add locking to accesses to the "sReqs" shadow-requests table
used to keep track of uncompleted MPI_Irecv's. This might
even mean the wrappers are thread-safe now :-)
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5738
Julian Seward [Fri, 10 Mar 2006 13:43:49 +0000 (13:43 +0000)]
- Improve handling of MPI base types
- Use MAKE_DEFINED rather than MAKE_READABLE, as the latter's
behaviour - changing A bits as well as V bits - can hide
addressing problems later
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5737
Julian Seward [Fri, 10 Mar 2006 13:41:58 +0000 (13:41 +0000)]
Add a new kind of memory-painting primitive, which is: 'make_defined'.
For each byte in the range, if the byte is addressible, make it be
initialised, but if it isn't addressible, leave it alone. So it's
like a version of make_readable which doesn't alter addressibility.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5736
Julian Seward [Thu, 9 Mar 2006 19:08:20 +0000 (19:08 +0000)]
Give *all* MPI-1.1 and MPI-2 functions a wrapper; most of them get a
default wrapper which can either run transparently, run transparently
but give a warning, or abort the run -- depending on flags set in
$MPIWRAP_DEBUG. Mess with flag handling accordingly.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5735
Julian Seward [Thu, 9 Mar 2006 19:06:23 +0000 (19:06 +0000)]
Add a couple of missing cases. Function wrapping is really ugly :-(
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5734
Julian Seward [Thu, 9 Mar 2006 19:05:29 +0000 (19:05 +0000)]
Record list of MPI-2 entry points, laboriously extracted from Open MPI
1.0.X header files.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5733
Julian Seward [Thu, 9 Mar 2006 03:05:17 +0000 (03:05 +0000)]
Add a 'quiet' option to make it shut up except for showing errors.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5732
Julian Seward [Thu, 9 Mar 2006 02:49:56 +0000 (02:49 +0000)]
The mpicc being tested needs to be able to link stuff, not merely compile it.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5731
Julian Seward [Thu, 9 Mar 2006 02:41:41 +0000 (02:41 +0000)]
MPI checks: use AC_TRY_COMPILE correctly, so as to generate a test program
that icc can compile (just in case mpicc is a wrapper round icc).
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5730
Julian Seward [Wed, 8 Mar 2006 21:40:34 +0000 (21:40 +0000)]
More MPICH build fixes.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5729
Julian Seward [Wed, 8 Mar 2006 20:41:44 +0000 (20:41 +0000)]
Add showCombiner and move showTy to a better place.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5728
Julian Seward [Wed, 8 Mar 2006 15:27:37 +0000 (15:27 +0000)]
New wrapper, and an MPICH build fix.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5727
Julian Seward [Wed, 8 Mar 2006 15:26:10 +0000 (15:26 +0000)]
MPI build system refinements: just base everything off mpicc, and don't
try to figure out where mpi.h is.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5726