]> git.ipfire.org Git - thirdparty/valgrind.git/log
thirdparty/valgrind.git
19 years agoAdd comments from email discussion re mprotect.
Julian Seward [Wed, 9 Nov 2005 14:51:59 +0000 (14:51 +0000)] 
Add comments from email discussion re mprotect.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5057

19 years agoThird pass at the ppc32 ptrace stuff. It compiles now, but gcc complains that
Julian Seward [Wed, 9 Nov 2005 14:43:03 +0000 (14:43 +0000)] 
Third pass at the ppc32 ptrace stuff.  It compiles now, but gcc complains that
the last two args to VG_(ptrace) are ints when they should be pointers.  So
I'm still not sure whether this is right or not.  Should I just put in a bunch
of casts to shut it up, or is it really expecting pointers and not ints?

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5056

19 years agoSave/Restore condition register, and VRSAVE register in core dispatch loop.
Cerion Armour-Brown [Wed, 9 Nov 2005 14:13:08 +0000 (14:13 +0000)] 
Save/Restore condition register, and VRSAVE register in core dispatch loop.

Cleaned up stack according to common abi constraints.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5055

19 years agoAdjust debugLog messages.
Julian Seward [Wed, 9 Nov 2005 14:09:14 +0000 (14:09 +0000)] 
Adjust debugLog messages.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5054

19 years agoInstall libvex.a and libcoregrind.a, needed for external tool builds.
Julian Seward [Wed, 9 Nov 2005 14:04:27 +0000 (14:04 +0000)] 
Install libvex.a and libcoregrind.a, needed for external tool builds.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5053

19 years agoDo ptrace_setregs in a ppc32 compatible way instead of a ppc64 way.
Tom Hughes [Wed, 9 Nov 2005 08:52:56 +0000 (08:52 +0000)] 
Do ptrace_setregs in a ppc32 compatible way instead of a ppc64 way.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5052

19 years agoAdd missing quote.
Tom Hughes [Wed, 9 Nov 2005 08:52:21 +0000 (08:52 +0000)] 
Add missing quote.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5051

19 years agoUpdate results - we now emulate the whole LDT.
Tom Hughes [Wed, 9 Nov 2005 08:48:38 +0000 (08:48 +0000)] 
Update results - we now emulate the whole LDT.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5050

19 years agoFactor out some common code in m_syswrap.
Nicholas Nethercote [Wed, 9 Nov 2005 04:49:28 +0000 (04:49 +0000)] 
Factor out some common code in m_syswrap.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5049

19 years agoupdate
Nicholas Nethercote [Wed, 9 Nov 2005 04:43:32 +0000 (04:43 +0000)] 
update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5048

19 years agostore & load callee-saved floating-point and vector registers in core dispatch loop.
Cerion Armour-Brown [Tue, 8 Nov 2005 22:03:07 +0000 (22:03 +0000)] 
store & load callee-saved floating-point and vector registers in core dispatch loop.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5047

19 years agoTry to fill in ptrace_setregs. Dunno if it's right.
Julian Seward [Tue, 8 Nov 2005 20:59:14 +0000 (20:59 +0000)] 
Try to fill in ptrace_setregs.  Dunno if it's right.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5046

19 years agoAugment a little.
Nicholas Nethercote [Tue, 8 Nov 2005 20:03:08 +0000 (20:03 +0000)] 
Augment a little.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5045

19 years agoupdate
Nicholas Nethercote [Tue, 8 Nov 2005 19:59:24 +0000 (19:59 +0000)] 
update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5044

19 years agoupdate
Nicholas Nethercote [Tue, 8 Nov 2005 19:38:58 +0000 (19:38 +0000)] 
update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5043

19 years agoReinstate timestamping, sort-of, having trashed VG_(ctime) as part of
Julian Seward [Tue, 8 Nov 2005 19:01:44 +0000 (19:01 +0000)] 
Reinstate timestamping, sort-of, having trashed VG_(ctime) as part of
the glibc-removal process.  Timestamps are now printed in terms of
elapsed wallclock time since startup, shown as days, hours, minutes,
seconds and milliseconds.  The arithmetic is done with 32-bit unsigned
ints, so people doing Valgrind runs that last longer than 49.71 days
are going to see some funny results :-)

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5042

19 years agoHandle the PROT_GROWSDOWN and PROT_GROWSUP mprotect flags properly
Tom Hughes [Tue, 8 Nov 2005 16:51:55 +0000 (16:51 +0000)] 
Handle the PROT_GROWSDOWN and PROT_GROWSUP mprotect flags properly
which stops glibc falling over when a program requires an executable
stack and glibc has been built to assume that PROT_GROWSDOWN will work.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5041

19 years agoDocument the --kernel-variant= option.
Julian Seward [Tue, 8 Nov 2005 15:52:36 +0000 (15:52 +0000)] 
Document the --kernel-variant= option.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5040

19 years agoGet rid of old flags that memcheck no longer has.
Julian Seward [Tue, 8 Nov 2005 14:10:24 +0000 (14:10 +0000)] 
Get rid of old flags that memcheck no longer has.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5039

19 years agoAdd tests for --partial-loads-ok= flag.
Julian Seward [Tue, 8 Nov 2005 13:20:24 +0000 (13:20 +0000)] 
Add tests for --partial-loads-ok= flag.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5038

19 years agoReworked to avoid depending on $(eval) in the make files as only fairly
Tom Hughes [Tue, 8 Nov 2005 12:28:35 +0000 (12:28 +0000)] 
Reworked to avoid depending on $(eval) in the make files as only fairly
new GNU make versions seem to have it.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5037

19 years agoSet CPPFLAGS correctly when building the core preload library.
Tom Hughes [Tue, 8 Nov 2005 12:27:42 +0000 (12:27 +0000)] 
Set CPPFLAGS correctly when building the core preload library.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5036

19 years agomemcheck: make --partial-loads-ok=yes work again, but now make it
Julian Seward [Tue, 8 Nov 2005 02:25:37 +0000 (02:25 +0000)] 
memcheck: make --partial-loads-ok=yes work again, but now make it
the non-default (it's a hack after all).

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5035

19 years agofwrite.stdout.exp seems to be not present and make dist doesn't like that.
Julian Seward [Tue, 8 Nov 2005 01:24:23 +0000 (01:24 +0000)] 
fwrite.stdout.exp seems to be not present and make dist doesn't like that.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5034

19 years agoSecond try at getting rid of the is_self() hack used to decide when to
Julian Seward [Tue, 8 Nov 2005 00:45:47 +0000 (00:45 +0000)] 
Second try at getting rid of the is_self() hack used to decide when to
load debug info from the V executable.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5033

19 years agoInitial description of multiple architecture support.
Tom Hughes [Mon, 7 Nov 2005 18:06:10 +0000 (18:06 +0000)] 
Initial description of multiple architecture support.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5032

19 years agoppc32 build fix.
Julian Seward [Mon, 7 Nov 2005 16:56:27 +0000 (16:56 +0000)] 
ppc32 build fix.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5031

19 years agoMy emacs is just too clever for its own good...
Tom Hughes [Mon, 7 Nov 2005 16:50:55 +0000 (16:50 +0000)] 
My emacs is just too clever for its own good...

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5030

19 years agoUse sysconf to get the page size as it is more portable.
Tom Hughes [Mon, 7 Nov 2005 16:46:55 +0000 (16:46 +0000)] 
Use sysconf to get the page size as it is more portable.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5029

19 years agoHandle sys_sigaction on ppc32-linux.
Julian Seward [Mon, 7 Nov 2005 15:48:31 +0000 (15:48 +0000)] 
Handle sys_sigaction on ppc32-linux.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5028

19 years agoDual architecture support - this commit is a major rework of the build
Tom Hughes [Mon, 7 Nov 2005 15:24:38 +0000 (15:24 +0000)] 
Dual architecture support - this commit is a major rework of the build
system that allows multiple copies of valgrind to be built so that we
can build both x86 and amd64 versions of the tools on amd64 machines.

The launcher is then modified to look at the program being run and
decide which tool to use to run it.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5027

19 years agoRun the right regression tests...
Nicholas Nethercote [Mon, 7 Nov 2005 14:43:12 +0000 (14:43 +0000)] 
Run the right regression tests...

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5026

19 years agoImprove factoring in m_coredump a little.
Nicholas Nethercote [Mon, 7 Nov 2005 04:52:41 +0000 (04:52 +0000)] 
Improve factoring in m_coredump a little.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5025

19 years agoAdd tests for 64 bit instructions.
Tom Hughes [Sun, 6 Nov 2005 12:29:50 +0000 (12:29 +0000)] 
Add tests for 64 bit instructions.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5024

19 years agoLiberalise x86 clone flag handling, so it's as per amd64/ppc32.
Julian Seward [Sat, 5 Nov 2005 15:13:23 +0000 (15:13 +0000)] 
Liberalise x86 clone flag handling, so it's as per amd64/ppc32.
However, leave the old code in place, disabled, for documentary
reasons.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5023

19 years agoFix stack unwinding on ppc.
Julian Seward [Sat, 5 Nov 2005 14:48:03 +0000 (14:48 +0000)] 
Fix stack unwinding on ppc.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5022

19 years agoCommit Tom's improved-stack-unwinding patch.
Julian Seward [Sat, 5 Nov 2005 14:22:03 +0000 (14:22 +0000)] 
Commit Tom's improved-stack-unwinding patch.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5021

19 years agoRead debug info from the V exe when it is installed as an in-place
Julian Seward [Sat, 5 Nov 2005 14:08:35 +0000 (14:08 +0000)] 
Read debug info from the V exe when it is installed as an in-place
build.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5020

19 years agoUpdate ignore lists.
Tom Hughes [Sat, 5 Nov 2005 12:16:03 +0000 (12:16 +0000)] 
Update ignore lists.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5019

19 years agoUpdate ignore lists.
Tom Hughes [Sat, 5 Nov 2005 12:13:41 +0000 (12:13 +0000)] 
Update ignore lists.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5018

19 years agoSync sse2 tests from x86 to amd64.
Tom Hughes [Sat, 5 Nov 2005 12:10:31 +0000 (12:10 +0000)] 
Sync sse2 tests from x86 to amd64.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5017

19 years agoAdd test cases for amd64 vector FP unordered compares.
Julian Seward [Sat, 5 Nov 2005 02:34:06 +0000 (02:34 +0000)] 
Add test cases for amd64 vector FP unordered compares.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5016

19 years agoOops, this should have been included in r5014.
Julian Seward [Sat, 5 Nov 2005 02:12:28 +0000 (02:12 +0000)] 
Oops, this should have been included in r5014.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5015

19 years agoAdd test cases for SSE unordered-compare instructions.
Julian Seward [Sat, 5 Nov 2005 01:57:02 +0000 (01:57 +0000)] 
Add test cases for SSE unordered-compare instructions.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5014

19 years agoHandle vector FP unordered compares.
Julian Seward [Sat, 5 Nov 2005 01:55:04 +0000 (01:55 +0000)] 
Handle vector FP unordered compares.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5013

19 years agoAdd a comment documenting the reason for subtracting one from the IP value
Tom Hughes [Sat, 5 Nov 2005 00:17:21 +0000 (00:17 +0000)] 
Add a comment documenting the reason for subtracting one from the IP value
after each pass of the loop.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5012

19 years agoEnable tests for shld/shrd instructions which are now supported on amd64.
Tom Hughes [Sat, 5 Nov 2005 00:10:36 +0000 (00:10 +0000)] 
Enable tests for shld/shrd instructions which are now supported on amd64.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5011

19 years agoPlacate compiler warning.
Julian Seward [Fri, 4 Nov 2005 20:19:09 +0000 (20:19 +0000)] 
Placate compiler warning.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5010

19 years agoTest shrd/shld.
Julian Seward [Fri, 4 Nov 2005 20:18:27 +0000 (20:18 +0000)] 
Test shrd/shld.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5009

19 years agoRe-enable a load of tests for instructions which VEX now implements.
Tom Hughes [Fri, 4 Nov 2005 18:07:10 +0000 (18:07 +0000)] 
Re-enable a load of tests for instructions which VEX now implements.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5008

19 years agoBackout additional space until we decide whether we want it.
Tom Hughes [Fri, 4 Nov 2005 17:18:18 +0000 (17:18 +0000)] 
Backout additional space until we decide whether we want it.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5007

19 years agoRemove "pid" to make docs align with reality.
Tom Hughes [Fri, 4 Nov 2005 17:13:24 +0000 (17:13 +0000)] 
Remove "pid" to make docs align with reality.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5006

19 years agoAdd an insn_basic test for amd64. This is based on the x86 one so
Tom Hughes [Fri, 4 Nov 2005 17:07:24 +0000 (17:07 +0000)] 
Add an insn_basic test for amd64. This is based on the x86 one so
doesn't test any 64 bit operations yet.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5005

19 years agoFix up references to log files names containing "pid" in them and
Tom Hughes [Fri, 4 Nov 2005 16:49:39 +0000 (16:49 +0000)] 
Fix up references to log files names containing "pid" in them and
a few other related items. Patch from James Begley <james@hafro.is>.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5004

19 years agoAllow rax/rbx/rcx/rdx (and their narrower counterparts) to be used
Tom Hughes [Fri, 4 Nov 2005 16:46:19 +0000 (16:46 +0000)] 
Allow rax/rbx/rcx/rdx (and their narrower counterparts) to be used
again but only after the high registers are exhausted or (and this is
the important point) if they are explicitly requested.

Also, don't use r8 as it confuses things as we can't tell an explicit
request from a generic request for a byte register.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5003

19 years agoFix the amd64 version of gen_insn_test.pl to strip any b/w/d suffix
Tom Hughes [Fri, 4 Nov 2005 15:36:05 +0000 (15:36 +0000)] 
Fix the amd64 version of gen_insn_test.pl to strip any b/w/d suffix
from register names when generating the clobber list as gcc won't
recognise things like r8d but will recognise r8.

This allows us to use the high number integer registers for the tests
which is something Julian asked for ages ago.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5002

19 years agoAlso test jecxz.
Julian Seward [Fri, 4 Nov 2005 14:38:48 +0000 (14:38 +0000)] 
Also test jecxz.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5001

19 years agoAdd missing tag.
Tom Hughes [Fri, 4 Nov 2005 14:25:09 +0000 (14:25 +0000)] 
Add missing tag.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5000

19 years agoMake the tool writing documentation align a bit more closely
Tom Hughes [Fri, 4 Nov 2005 14:11:05 +0000 (14:11 +0000)] 
Make the tool writing documentation align a bit more closely
with reality.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4999

19 years agoRemove remaining references to the old tool.h header.
Tom Hughes [Fri, 4 Nov 2005 12:27:58 +0000 (12:27 +0000)] 
Remove remaining references to the old tool.h header.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4998

19 years agoTool writing documentation updates from Yao Qi <qiyaoltc@cn.ibm.com>.
Tom Hughes [Fri, 4 Nov 2005 12:17:20 +0000 (12:17 +0000)] 
Tool writing documentation updates from Yao Qi <qiyaoltc@cn.ibm.com>.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4997

19 years agoWhen unwinding the stack on x86/amd64 subtract one from the value of
Tom Hughes [Fri, 4 Nov 2005 11:31:33 +0000 (11:31 +0000)] 
When unwinding the stack on x86/amd64 subtract one from the value of
ip before starting a new pass of the loop.

The reason for this is that (except for the first pass of the loop) the
value of ip is actually a return address, which is therefore after the
instruction that was executing at the time. This means that if there is
a boundary in the CFI information at that point we can wind up using the
wrong CFI data to do the next unwind if we do it based on the return
address.

This most commonly happens with a tail call where we wind up using the
data for the next function to do the unwind and getting hopelessly lost.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4996

19 years agoRecord a useful nugget from Julian.
Nicholas Nethercote [Thu, 3 Nov 2005 16:24:53 +0000 (16:24 +0000)] 
Record a useful nugget from Julian.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4995

19 years agoTrack API change in vex r1429.
Julian Seward [Thu, 3 Nov 2005 13:46:30 +0000 (13:46 +0000)] 
Track API change in vex r1429.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4994

19 years agoDon't try to honour the ELF interpreter's load address if it is zero.
Julian Seward [Wed, 2 Nov 2005 19:41:21 +0000 (19:41 +0000)] 
Don't try to honour the ELF interpreter's load address if it is zero.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4993

19 years agoSplit faultstatus into the platform independent tests and those
Tom Hughes [Wed, 2 Nov 2005 16:15:55 +0000 (16:15 +0000)] 
Split faultstatus into the platform independent tests and those
which are x86 specific. The first three x86 specific ones should
work on amd64 as well so I have added those as amd64 tests.

Note that the x86/amd64 tests will still fail as VEX doesn't
always trigger the right sort of signal for faulting instructions
at the moment.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4992

19 years agoRemove the as_mmap and as_shm tests which don't make much sense since
Tom Hughes [Wed, 2 Nov 2005 15:47:43 +0000 (15:47 +0000)] 
Remove the as_mmap and as_shm tests which don't make much sense since
the address space manager was rewritten.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4991

19 years agoThe memcheck/tests/sigprocmask test is designed to test that we handle
Tom Hughes [Wed, 2 Nov 2005 15:46:07 +0000 (15:46 +0000)] 
The memcheck/tests/sigprocmask test is designed to test that we handle
the old style sigprocmask system call correctly without corrupting
memory when we copy out the new (larger) signal mask into the user
provided old (smaller) signal mask.

It therefore makes no sense to run it on amd64 or any other platform
which only has the newer rt_sigprocmask system call, and indeed it
wasn't working because we weren't passing the extra argument which
that call expects.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4990

19 years agoFix format string in signal mask printing.
Tom Hughes [Wed, 2 Nov 2005 15:32:19 +0000 (15:32 +0000)] 
Fix format string in signal mask printing.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4989

19 years agoDisplay signal masks correctly when --trace-signals=yes is used
Tom Hughes [Wed, 2 Nov 2005 15:31:21 +0000 (15:31 +0000)] 
Display signal masks correctly when --trace-signals=yes is used
on 64 bit platforms.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4988

19 years agoUpdate EXTRA_DIST to include new results files.
Tom Hughes [Wed, 2 Nov 2005 15:17:43 +0000 (15:17 +0000)] 
Update EXTRA_DIST to include new results files.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4987

19 years agoAdd alternate results for memcheck/tests/strchr on amd64.
Tom Hughes [Wed, 2 Nov 2005 15:03:12 +0000 (15:03 +0000)] 
Add alternate results for memcheck/tests/strchr on amd64.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4986

19 years agoThe sloppyRXcheck logic in the sync checker was not correct - it was
Tom Hughes [Wed, 2 Nov 2005 14:42:39 +0000 (14:42 +0000)] 
The sloppyRXcheck logic in the sync checker was not correct - it was
simply treating R and X as equivalent but the real problem is that
mappings can appear to have X permission entirely indepenent of anything
else with recent x86 kernels.

If a mapping is inside the (deliberately constrained) code segment then
it will appear to have X permission regardless of whether R or X was asked
for when it was mapped, so what we really need to do is allow the kernel
to add X to any mapping but not to take it away if we were expecting it.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4985

19 years agoRemoved memcheck/tests/weirdioctl as it is badly broken and seems
Tom Hughes [Wed, 2 Nov 2005 13:58:39 +0000 (13:58 +0000)] 
Removed memcheck/tests/weirdioctl as it is badly broken and seems
to serve no useful purpose.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4984

19 years agoDon't give the heap execute permission - the linux kernel doesn't
Tom Hughes [Wed, 2 Nov 2005 13:02:40 +0000 (13:02 +0000)] 
Don't give the heap execute permission - the linux kernel doesn't
normally give execute permission to memory allocated from the heap
with sbrk.

This also required fixing the smc1 test for amd64 to use mmap to
allocate memory so that it can have execute permission.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4983

19 years agoUpdate memcheck/tests/xml1 results.
Tom Hughes [Wed, 2 Nov 2005 13:00:41 +0000 (13:00 +0000)] 
Update memcheck/tests/xml1 results.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4982

19 years agoUpdate memcheck/tests/xml1 results.
Tom Hughes [Wed, 2 Nov 2005 13:00:03 +0000 (13:00 +0000)] 
Update memcheck/tests/xml1 results.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4981

19 years agoMake the request for space for the interpreter into a hint so that we
Tom Hughes [Tue, 1 Nov 2005 19:46:07 +0000 (19:46 +0000)] 
Make the request for space for the interpreter into a hint so that we
can still proceed if the preferred load address is not available.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4980

19 years agoReinstate and extend lackey's functionality (a major overhaul), from
Julian Seward [Tue, 1 Nov 2005 17:52:34 +0000 (17:52 +0000)] 
Reinstate and extend lackey's functionality (a major overhaul), from
Jeroen N. Witmond.  Thanks for that.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4979

19 years agoFixed test to not assume that malloc(10) will return 10 nul bytes!
Tom Hughes [Tue, 1 Nov 2005 17:21:03 +0000 (17:21 +0000)] 
Fixed test to not assume that malloc(10) will return 10 nul bytes!

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4978

19 years agoupdate
Nicholas Nethercote [Tue, 1 Nov 2005 17:15:50 +0000 (17:15 +0000)] 
update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4977

19 years agoWhitespace only change (test commit to see if auto-mail-on-commit is
Julian Seward [Tue, 1 Nov 2005 12:32:24 +0000 (12:32 +0000)] 
Whitespace only change (test commit to see if auto-mail-on-commit is
still working).

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4976

19 years agoIf pdfxmltex bombs, as it so often does, at least cause the build to
Julian Seward [Tue, 1 Nov 2005 03:08:41 +0000 (03:08 +0000)] 
If pdfxmltex bombs, as it so often does, at least cause the build to
come to a halt rather than just hanging.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4975

19 years agoUpdate.
Julian Seward [Tue, 1 Nov 2005 03:07:38 +0000 (03:07 +0000)] 
Update.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4974

19 years agoMake ULong_width(0) return 1 instead of 1431655765. This stops
Julian Seward [Tue, 1 Nov 2005 02:20:19 +0000 (02:20 +0000)] 
Make ULong_width(0) return 1 instead of 1431655765.  This stops
cachegrind segfaulting if it can't open its output file.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4973

19 years agoBuild fixes following recent coredump hackery.
Julian Seward [Tue, 1 Nov 2005 00:42:07 +0000 (00:42 +0000)] 
Build fixes following recent coredump hackery.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4972

19 years agoFix a couple of gcc warnings following recent coredumping changes
Julian Seward [Tue, 1 Nov 2005 00:03:40 +0000 (00:03 +0000)] 
Fix a couple of gcc warnings following recent coredumping changes
(r4970).

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4971

19 years agoGet core dumping working again - the architecture specific code that
Tom Hughes [Mon, 31 Oct 2005 17:05:21 +0000 (17:05 +0000)] 
Get core dumping working again - the architecture specific code that
was in the sigframe module has been moved into the coredump module
where it belongs and things fixed up to compiler again.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4970

19 years agoThe version string from --version was being printed to stderr. This
Nicholas Nethercote [Wed, 26 Oct 2005 16:17:46 +0000 (16:17 +0000)] 
The version string from --version was being printed to stderr.  This
commit fixes it to print to stdout.  I added a regression test for this.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4969

19 years agoTrawl through bugzilla/users-list and add various bugs.
Julian Seward [Wed, 26 Oct 2005 12:21:16 +0000 (12:21 +0000)] 
Trawl through bugzilla/users-list and add various bugs.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4968

20 years agoDon't assume the first statement is an IRMark, since it could instead
Julian Seward [Sun, 23 Oct 2005 12:06:55 +0000 (12:06 +0000)] 
Don't assume the first statement is an IRMark, since it could instead
be part of a self-check.  Instead, copy verbatim any IR preamble
preceding the first IMark.  This stops cachegrind asserting on
self-checking translations.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4967

20 years agoUpdate bug status information.
Tom Hughes [Sun, 23 Oct 2005 09:30:35 +0000 (09:30 +0000)] 
Update bug status information.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4966

20 years agoEnable Altivec by default if it is available.
Julian Seward [Sat, 22 Oct 2005 12:52:25 +0000 (12:52 +0000)] 
Enable Altivec by default if it is available.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4965

20 years agoHandle a few more primops result from running Altivec code.
Julian Seward [Sat, 22 Oct 2005 12:50:43 +0000 (12:50 +0000)] 
Handle a few more primops result from running Altivec code.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4964

20 years agoPass args to match format string.
Julian Seward [Fri, 21 Oct 2005 12:05:05 +0000 (12:05 +0000)] 
Pass args to match format string.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4963

20 years agoHandle SHT_NOBITS sections when checking for section overruns. This
Julian Seward [Fri, 21 Oct 2005 02:37:40 +0000 (02:37 +0000)] 
Handle SHT_NOBITS sections when checking for section overruns.  This
fixes a problem where debug info was sometimes not being read on
ppc32.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4962

20 years agoDocument bug #114757.
Tom Hughes [Thu, 20 Oct 2005 18:42:26 +0000 (18:42 +0000)] 
Document bug #114757.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4961

20 years agoDon't assert if the DWARF line info reader is given so little data that
Tom Hughes [Thu, 20 Oct 2005 18:38:08 +0000 (18:38 +0000)] 
Don't assert if the DWARF line info reader is given so little data that
it can't even read the length of the block - just report an error as we
do if there isn't enough data for the rest of the block. Fix bug #114757.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4960

20 years agoDocument bug #114756.
Tom Hughes [Thu, 20 Oct 2005 17:03:53 +0000 (17:03 +0000)] 
Document bug #114756.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4959

20 years agoAdd support for the mbind system call on x86 and amd64. Fixes bug #114756.
Tom Hughes [Thu, 20 Oct 2005 17:00:23 +0000 (17:00 +0000)] 
Add support for the mbind system call on x86 and amd64. Fixes bug #114756.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@4958