]> git.ipfire.org Git - thirdparty/valgrind.git/log
thirdparty/valgrind.git
19 years agoSuppress any source filename in a "(below main)" line in a stack trace.
Tom Hughes [Thu, 13 Apr 2006 12:44:31 +0000 (12:44 +0000)] 
Suppress any source filename in a "(below main)" line in a stack trace.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5847

19 years agoClarify the Helgrind-doesn't-work message.
Nicholas Nethercote [Wed, 12 Apr 2006 22:53:23 +0000 (22:53 +0000)] 
Clarify the Helgrind-doesn't-work message.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5846

19 years agoRecent GCCs (3.4+ at least) optimize static unused functions out, so
Julian Seward [Wed, 12 Apr 2006 17:15:35 +0000 (17:15 +0000)] 
Recent GCCs (3.4+ at least) optimize static unused functions out, so
making VALGRIND_PRINTF and VALGRIND_PRINTF_BACKTRACE static and
attribute unused proved to be much better than always compiling it as
exported weak function.  (Jakub Jelinek)

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5845

19 years agoAdd a suppression for yet another glibc string function: __strcpy_chk.
Julian Seward [Wed, 12 Apr 2006 13:41:51 +0000 (13:41 +0000)] 
Add a suppression for yet another glibc string function: __strcpy_chk.
We really ought intercept/replace this, and that would be easy, except
__strcpy_chk uses __chk_fail and I haven't figured out what that
is/does.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5844

19 years agoOn x86, don't use x87 registers for 8-byte FP loads/stores; instead
Julian Seward [Wed, 12 Apr 2006 11:42:54 +0000 (11:42 +0000)] 
On x86, don't use x87 registers for 8-byte FP loads/stores; instead
use an mmx register (which is the same thing in disguise) since mmx
loads/stores are guaranteed to be the identity.  This should fix
failures of this test on x86-linux.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5843

19 years agoAdd the O_LARGEFILE constant so people can use it when writing tools that
Nicholas Nethercote [Wed, 12 Apr 2006 11:10:46 +0000 (11:10 +0000)] 
Add the O_LARGEFILE constant so people can use it when writing tools that
generate large trace files (> 2GB).

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5842

19 years agoProperly handle external debuginfo (.gnu_debuglink) for stripped and
Julian Seward [Wed, 12 Apr 2006 10:08:05 +0000 (10:08 +0000)] 
Properly handle external debuginfo (.gnu_debuglink) for stripped and
prelinked libraries (patch from L. David Baron).

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5841

19 years agoRedid the --trace-mem=yes option of Lackey properly. Updated some related
Nicholas Nethercote [Sun, 9 Apr 2006 01:23:29 +0000 (01:23 +0000)] 
Redid the --trace-mem=yes option of Lackey properly.  Updated some related
stuff along with it, such as the NEWS file.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5840

19 years agoRe-enable amd64 test for pmaddwd.
Julian Seward [Sat, 8 Apr 2006 16:52:42 +0000 (16:52 +0000)] 
Re-enable amd64 test for pmaddwd.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5839

19 years agoFold in a patch which appeared in FC5's default valgrind build, which
Julian Seward [Sat, 8 Apr 2006 16:22:53 +0000 (16:22 +0000)] 
Fold in a patch which appeared in FC5's default valgrind build, which
causes V to ignore more DWARF3 CFA expressions on amd64 and so gets
rid of complaints from the CFA reader.  Why didn't Red Hat push this
patch upstream?  I don't know.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5838

19 years agoInclude missing files in tarball (== unbreak from-tarball builds)
Julian Seward [Sat, 8 Apr 2006 13:17:25 +0000 (13:17 +0000)] 
Include missing files in tarball (== unbreak from-tarball builds)

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5837

19 years agoUpdate
Julian Seward [Sat, 8 Apr 2006 13:16:36 +0000 (13:16 +0000)] 
Update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5836

19 years agoOops: when adding translations to the auxiliary transtab, don't forget to
Julian Seward [Fri, 7 Apr 2006 12:47:05 +0000 (12:47 +0000)] 
Oops: when adding translations to the auxiliary transtab, don't forget to
ensure D-I cache coherence.  Fixes SIGILLs in fn wrapping failures on low end
PowerPCs.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5835

19 years agoTweaked Lackey. Main change is that the default instrumentation is now only
Nicholas Nethercote [Fri, 7 Apr 2006 11:52:55 +0000 (11:52 +0000)] 
Tweaked Lackey.  Main change is that the default instrumentation is now only
added if you specify --basic-counts=yes (which is the default).  So
all of the instrumentation is now controlled by a command-line option (one
of --basic-counts, --detailed-counts or --trace-mem) and so if you turn them
all off it behaves like Nulgrind.  This makes it clearer what's going on and
easier for newbies to modify.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5834

19 years agoFix vg_perf to work with older Valgrind versions.
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

19 years agoCommit a patch from Olly Betts which avoids a possible problem with
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

19 years agoAdd a FAQ.
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

19 years agoUse better variable names.
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

19 years agoAvoid compiler warnings on ppc64-linux.
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

19 years agoupdate
Nicholas Nethercote [Thu, 6 Apr 2006 13:29:18 +0000 (13:29 +0000)] 
update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5828

19 years agoMinor scheduler tidyings:
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

19 years agoTest program which uses templates to generate very long stabs (debug
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

19 years agoComment-only change.
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

19 years agoUn-break make dist.
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

19 years agoCleanup/restructure m_debuginfo, as described in
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

19 years agoRound up known bugs in 3.1.1 and assign tentative priorities.
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

19 years agoAdd a comment that really should have been in r5820.
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

19 years agoppc32-linux: work around assemblers which can't do Altivec, by
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

19 years agoEnable sigsuspend on x86-linux and ppc32-linux.
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

19 years agoDon't use the presence of a filename to decide if a segment in the
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

19 years agoAdd an optional post-test check. Based on a patch from Josef W.
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

19 years agoCallgrind: add 4 regression tests
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

19 years agoFix another place where function addresses passed to Vex needed to be
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

19 years agoFix ppc64-linux breakage - get the function entry address, not the
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

19 years agoAdd 64-bit test output.
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

19 years agoDon't print an address on stdout as it doesn't get automagically
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

19 years agoAnother shadow memory test. This one does a huge number of loads and
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

19 years agoGet rid of unused #define.
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

19 years agoIn check_all, allow two different acceptable byte values, so as to
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

19 years agoPartial fix for the sh-mem.c failure on PPC32. This should make it work
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

19 years agoSimple regression test for callgrind:
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

19 years agoOption --dump-every-bb now takes a 64bit value on user
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

19 years agoFix typos
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

19 years agoupdate Memcheck description
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

19 years agoMake the BSD-style license notices slightly more accurate.
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

19 years agoTerminology change: previously in Memcheck we had the four states:
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

19 years agoAvoid compile warnings for tinycc.
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

19 years agoHmm. Why is this file still here? (It got folded into aspacemgr.c
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

19 years agoDelete VG_(sigaltstack) and VG_(sigpending) as they are unused.
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

19 years agoDelete VG_(poll), as it is unused.
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

19 years agoChange to some better function names.
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

19 years agoChange some uses of bytes to bits for consistency.
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

19 years agoOnly complain once for each missing wrapper.
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

19 years agotweak
Nicholas Nethercote [Tue, 28 Mar 2006 00:05:16 +0000 (00:05 +0000)] 
tweak

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5794

19 years agoPartial release notes for 3.2.0, talking about the performance improvements
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

19 years agoupdate
Nicholas Nethercote [Mon, 27 Mar 2006 11:55:09 +0000 (11:55 +0000)] 
update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5792

19 years agoMerge in the COMPVBITS branch to the trunk. This is a big change to
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

19 years agoAvoid unnecessary whitespace at end of printed lines.
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

19 years agoHandle msg_flags correctly - for recvmsg this field is written by the
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

19 years agoUpdate expected output following r5785; also add a missing 's'.
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

19 years agoUpdate
Nicholas Nethercote [Tue, 21 Mar 2006 23:23:22 +0000 (23:23 +0000)] 
Update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5787

19 years agoMake the unrecognised-instruction error message more informative.
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

19 years agoAdd wrappers for a load of new 2.6.16 system calls. Fixes bug #123248.
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

19 years agoEnable sys_fchdir on ppc32-linux.
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

19 years agoSync system call lists with the 2.6.16 kernel.
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

19 years agoMake hard limit enforcement work for RLIMIT_DATA and RLIMIT_STACK in
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

19 years agoCallgrind merge: documentation
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

19 years agoCallgrind merge: code
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

19 years agoAdd a comment about the library's license.
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

19 years agoYet another essay: document the MPI wrapper library.
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

19 years agoUpdate
Julian Seward [Thu, 16 Mar 2006 11:33:29 +0000 (11:33 +0000)] 
Update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5776

19 years agoMerge stuff for 3.1.1.
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

19 years agoupmerge r5768 (Allow a thread to spin longer when yielding before
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

19 years agoRecycle Dirk's glibc-2.3.90 suppressions, since at some point it will
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

19 years agoA couple of initialisations to keep gcc-4.1.0 happy.
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

19 years agoMinor futzing (fontification, etc) of the function-wrappers documentation.
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

19 years agoFirst pass at documenting how to use the function-wrapping facility.
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

19 years agoSigh; rename mpiwrap.c to libmpiwrap.c so that 'make dist' doesn't barf
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

19 years agoWhen testing 'mpicc', try at least somewhat to test for the primary
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

19 years agoUnbreak mtoctf configure check.
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

19 years agoMess with this even more so that fn_1 doesn't return zero.
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

19 years agoAdd missing CALL_FN_* macros for ppc64-linux.
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

19 years agoWork around lameness in older ppc assemblers.
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

19 years agoLAM/MPI compile fix
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

19 years agoUpdated expected output following r5753.
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

19 years agoAdd CALL_FN_* missing cases for ppc32-linux.
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

19 years agoTRASH_IREGS: add enough live variables to make ppc32 spill (not easy)
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

19 years agosizeofOneNamedTy(): handle long double correctly on non-x86/amd64 platforms.
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

19 years agoAdd ppc32-linux details.
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

19 years agoFix a redirection bug created by the overhaul of m_redir.c. Bug
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

19 years agoA 64-bit fix
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

19 years agoComplete CALL_FN_W_* wrappers for amd64-linux.
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

19 years agoMajorly beef up this test:
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

19 years agoFix comment typo.
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

19 years agoAdd missing x86-linux CALL_FN_W_11W macro.
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

19 years agoAdd a comment.
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

19 years agoQuadrics MPI build fixes.
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

19 years agoMake sure we get the right headers, not whatever happens to be already
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

19 years agoAdd a (rather ingenious, if I do say so myself) program to test the
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

19 years agoEven more MPI wrappers.
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