]> git.ipfire.org Git - thirdparty/valgrind.git/log
thirdparty/valgrind.git
21 years agocomment wibbles
Nicholas Nethercote [Mon, 25 Oct 2004 20:43:14 +0000 (20:43 +0000)] 
comment wibbles

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2841

21 years agocomment wibbles
Nicholas Nethercote [Mon, 25 Oct 2004 20:21:47 +0000 (20:21 +0000)] 
comment wibbles

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2840

21 years agoAvoid strange warnings about dereferencing type-punned pointers that occurs
Nicholas Nethercote [Mon, 25 Oct 2004 19:46:07 +0000 (19:46 +0000)] 
Avoid strange warnings about dereferencing type-punned pointers that occurs
with GCC 3.3+.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2839

21 years agotweak
Nicholas Nethercote [Mon, 25 Oct 2004 19:33:26 +0000 (19:33 +0000)] 
tweak

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2838

21 years agoArch-abstraction:
Nicholas Nethercote [Mon, 25 Oct 2004 19:20:14 +0000 (19:20 +0000)] 
Arch-abstraction:
- factor out varying ASM flags

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2837

21 years agoRevert last commit because stupid older versions (eg. 1.6.3) of automake can't
Nicholas Nethercote [Mon, 25 Oct 2004 19:13:41 +0000 (19:13 +0000)] 
Revert last commit because stupid older versions (eg. 1.6.3) of automake can't
handle it.  Pah.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2836

21 years agoFix up stage2's dependencies.
Nicholas Nethercote [Mon, 25 Oct 2004 17:59:43 +0000 (17:59 +0000)] 
Fix up stage2's dependencies.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2835

21 years agowhoops
Nicholas Nethercote [Mon, 25 Oct 2004 17:28:26 +0000 (17:28 +0000)] 
whoops

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2834

21 years agoIf PIE (position-independent executables) are supported, build valgrind's
Nicholas Nethercote [Mon, 25 Oct 2004 17:18:24 +0000 (17:18 +0000)] 
If PIE (position-independent executables) are supported, build valgrind's
stage2 as one.  This means that we're not hard-wiring stage2 in at 0xb0000000,
which means our memory layout is a bit more flexible, yay.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2833

21 years agoActually call VGA_(has_tls) rather than just testing the function
Tom Hughes [Mon, 25 Oct 2004 16:59:46 +0000 (16:59 +0000)] 
Actually call VGA_(has_tls) rather than just testing the function
address - this replaces Nick's replacing of the has_tls call with
an inline version.

BUG: 91844

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2832

21 years agoRename two params named 'syscall' as 'syscallname' to avoid clashes with a new
Nicholas Nethercote [Mon, 25 Oct 2004 15:43:21 +0000 (15:43 +0000)] 
Rename two params named 'syscall' as 'syscallname' to avoid clashes with a new
name in recent headers.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2831

21 years agoUse renamed and improved vex-supplied guest-state-handling fns.
Julian Seward [Mon, 25 Oct 2004 15:34:19 +0000 (15:34 +0000)] 
Use renamed and improved vex-supplied guest-state-handling fns.

git-svn-id: svn://svn.valgrind.org/vex/trunk@425

21 years agoImprove naming scheme for library-visible fns (infrastructural improvement)
Julian Seward [Mon, 25 Oct 2004 15:33:26 +0000 (15:33 +0000)] 
Improve naming scheme for library-visible fns (infrastructural improvement)

git-svn-id: svn://svn.valgrind.org/vex/trunk@424

21 years agoDon't bother with -fomit-frame-pointer for the demangler.
Nicholas Nethercote [Mon, 25 Oct 2004 15:32:09 +0000 (15:32 +0000)] 
Don't bother with -fomit-frame-pointer for the demangler.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2830

21 years agoArch-abstraction:
Nicholas Nethercote [Mon, 25 Oct 2004 15:21:00 +0000 (15:21 +0000)] 
Arch-abstraction:
- factor out different CFLAGS in Makefiles

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2829

21 years agoGeneric fixes to iropt to enable it to be as sloppy with exceptions as
Julian Seward [Mon, 25 Oct 2004 14:50:21 +0000 (14:50 +0000)] 
Generic fixes to iropt to enable it to be as sloppy with exceptions as
we can get away with.

git-svn-id: svn://svn.valgrind.org/vex/trunk@423

21 years agoArch-abstraction:
Nicholas Nethercote [Mon, 25 Oct 2004 14:05:56 +0000 (14:05 +0000)] 
Arch-abstraction:
- factor out KICKSTART_BASE

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2828

21 years agoTrack vex controllability changes, and other hacks-de-jour.
Julian Seward [Mon, 25 Oct 2004 13:06:17 +0000 (13:06 +0000)] 
Track vex controllability changes, and other hacks-de-jour.

git-svn-id: svn://svn.valgrind.org/vex/trunk@422

21 years agoTrack Vex controllability changes.
Julian Seward [Mon, 25 Oct 2004 13:04:59 +0000 (13:04 +0000)] 
Track Vex controllability changes.

git-svn-id: svn://svn.valgrind.org/vex/trunk@421

21 years agoAdd run-time controllability to iropt.
Julian Seward [Mon, 25 Oct 2004 13:01:45 +0000 (13:01 +0000)] 
Add run-time controllability to iropt.

git-svn-id: svn://svn.valgrind.org/vex/trunk@420

21 years agoFix for 91844...
Nicholas Nethercote [Mon, 25 Oct 2004 11:25:22 +0000 (11:25 +0000)] 
Fix for 91844...

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2827

21 years ago* re-enable some obscure sbb case
Julian Seward [Mon, 25 Oct 2004 09:52:18 +0000 (09:52 +0000)] 
* re-enable some obscure sbb case
* add printer for cpuid

git-svn-id: svn://svn.valgrind.org/vex/trunk@419

21 years agoRedundant-get removal: invalidate the entire env when passing
Julian Seward [Mon, 25 Oct 2004 09:48:47 +0000 (09:48 +0000)] 
Redundant-get removal: invalidate the entire env when passing
a dirty helper which might write any part of the guest state.
TODO: check redundant GetI removal.

git-svn-id: svn://svn.valgrind.org/vex/trunk@418

21 years agoEven more stuff.
Julian Seward [Mon, 25 Oct 2004 02:58:30 +0000 (02:58 +0000)] 
Even more stuff.

git-svn-id: svn://svn.valgrind.org/vex/trunk@417

21 years agoAdd various FP insns apparently needed by mozilla. Not convinced
Julian Seward [Mon, 25 Oct 2004 02:27:01 +0000 (02:27 +0000)] 
Add various FP insns apparently needed by mozilla.  Not convinced
they are all correct.

git-svn-id: svn://svn.valgrind.org/vex/trunk@416

21 years agoMerge together the two live-range-calculating loops, so we don't call
Julian Seward [Mon, 25 Oct 2004 01:21:08 +0000 (01:21 +0000)] 
Merge together the two live-range-calculating loops, so we don't call
getRegUsage twice for each insn.  This significantly reduces the
number of calls thereto, but doesn't halve them for some reason
(removes about 1/3 of the calls).

git-svn-id: svn://svn.valgrind.org/vex/trunk@415

21 years agoMess with logic for debug printing a bit.
Julian Seward [Mon, 25 Oct 2004 00:45:30 +0000 (00:45 +0000)] 
Mess with logic for debug printing a bit.

git-svn-id: svn://svn.valgrind.org/vex/trunk@414

21 years agoTry to further clear up confusion re notation for live range in
Julian Seward [Mon, 25 Oct 2004 00:43:36 +0000 (00:43 +0000)] 
Try to further clear up confusion re notation for live range in
reg-alloc.  I think it is much improved.  This commit also undoes
the bogus fix rev 119.

git-svn-id: svn://svn.valgrind.org/vex/trunk@413

21 years agoCorrect a longstanding bug in how branch target addresses were
Julian Seward [Mon, 25 Oct 2004 00:42:16 +0000 (00:42 +0000)] 
Correct a longstanding bug in how branch target addresses were
computed.

git-svn-id: svn://svn.valgrind.org/vex/trunk@412

21 years agoDuh; fix use of IROPT_LEVEL.
Julian Seward [Mon, 25 Oct 2004 00:41:11 +0000 (00:41 +0000)] 
Duh; fix use of IROPT_LEVEL.

git-svn-id: svn://svn.valgrind.org/vex/trunk@411

21 years agoFix yet more problems to do with confusion over live range
Julian Seward [Mon, 25 Oct 2004 00:04:52 +0000 (00:04 +0000)] 
Fix yet more problems to do with confusion over live range
definitions.  This commit gets rid of rev 180, which was wrong, and
rev 317, which was a kludge to try and cover 180's wrongness.

The correct solution is to compute live ranges correctly in the first
place.  In particular, if an insn #i reads a reg and that's the end of
the live range, we must say that insn's dead-before value is i+1 and
not i.

git-svn-id: svn://svn.valgrind.org/vex/trunk@410

21 years agox86 guest: implement more convincing version of CPUID, using a dirty
Julian Seward [Sun, 24 Oct 2004 22:31:25 +0000 (22:31 +0000)] 
x86 guest: implement more convincing version of CPUID, using a dirty
helper to modify guest state.  Fix the oversight that dirty helpers
don't currently have access to guest state by passing a pointer to it
as the first argument to all dirty helpers.

git-svn-id: svn://svn.valgrind.org/vex/trunk@409

21 years agoMake this more controllable.
Julian Seward [Sun, 24 Oct 2004 22:29:19 +0000 (22:29 +0000)] 
Make this more controllable.

git-svn-id: svn://svn.valgrind.org/vex/trunk@408

21 years agoIncrease size of translation buffer.
Julian Seward [Sun, 24 Oct 2004 22:28:57 +0000 (22:28 +0000)] 
Increase size of translation buffer.

git-svn-id: svn://svn.valgrind.org/vex/trunk@407

21 years agox86 guest: support rcr{bwl}. Includes some cleaning up of macro
Julian Seward [Sun, 24 Oct 2004 19:20:43 +0000 (19:20 +0000)] 
x86 guest: support rcr{bwl}.  Includes some cleaning up of macro
definitions in gdefs.h and ghelpers.c.

git-svn-id: svn://svn.valgrind.org/vex/trunk@406

21 years agoiselIntExp64: handle calls to clean helpers which return a 64-bit
Julian Seward [Sun, 24 Oct 2004 19:19:37 +0000 (19:19 +0000)] 
iselIntExp64: handle calls to clean helpers which return a 64-bit
result.

git-svn-id: svn://svn.valgrind.org/vex/trunk@405

21 years agoRe-enable tests for rcr{bwl}.
Julian Seward [Sun, 24 Oct 2004 19:18:44 +0000 (19:18 +0000)] 
Re-enable tests for rcr{bwl}.

git-svn-id: svn://svn.valgrind.org/vex/trunk@404

21 years agoComment changes only.
Julian Seward [Sun, 24 Oct 2004 13:11:39 +0000 (13:11 +0000)] 
Comment changes only.

git-svn-id: svn://svn.valgrind.org/vex/trunk@403

21 years agoFast-case some more stuff.
Julian Seward [Sat, 23 Oct 2004 01:13:57 +0000 (01:13 +0000)] 
Fast-case some more stuff.

git-svn-id: svn://svn.valgrind.org/vex/trunk@402

21 years agoOops -- properly handle fast case of Carry after Subtract.
Julian Seward [Sat, 23 Oct 2004 00:23:07 +0000 (00:23 +0000)] 
Oops -- properly handle fast case of Carry after Subtract.

git-svn-id: svn://svn.valgrind.org/vex/trunk@401

21 years agoAdd handy way to disable iropt completely.
Julian Seward [Sat, 23 Oct 2004 00:22:32 +0000 (00:22 +0000)] 
Add handy way to disable iropt completely.

git-svn-id: svn://svn.valgrind.org/vex/trunk@400

21 years agox86 guest: handle more bswap insns
Julian Seward [Sat, 23 Oct 2004 00:21:50 +0000 (00:21 +0000)] 
x86 guest: handle more bswap insns

git-svn-id: svn://svn.valgrind.org/vex/trunk@399

21 years agoFix "make check" for when src_dir != build_dir
Nicholas Nethercote [Thu, 21 Oct 2004 14:07:58 +0000 (14:07 +0000)] 
Fix "make check" for when src_dir != build_dir

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2826

21 years agoReinstate conservative version of redundant-Put elimination which
Julian Seward [Thu, 21 Oct 2004 10:22:10 +0000 (10:22 +0000)] 
Reinstate conservative version of redundant-Put elimination which
flushes all guest state prior to any memory reference, hence giving
precise (memory-) exceptions.

git-svn-id: svn://svn.valgrind.org/vex/trunk@398

21 years agoFix some startup messages.
Julian Seward [Thu, 21 Oct 2004 09:38:48 +0000 (09:38 +0000)] 
Fix some startup messages.

git-svn-id: svn://svn.valgrind.org/vex/trunk@397

21 years agoReinstate various floating-point tests which are now supported.
Julian Seward [Thu, 21 Oct 2004 09:20:46 +0000 (09:20 +0000)] 
Reinstate various floating-point tests which are now supported.

git-svn-id: svn://svn.valgrind.org/vex/trunk@396

21 years agoFast-track another case in calculate_eflags_c (needs verification).
Julian Seward [Wed, 20 Oct 2004 23:42:56 +0000 (23:42 +0000)] 
Fast-track another case in calculate_eflags_c (needs verification).

git-svn-id: svn://svn.valgrind.org/vex/trunk@395

21 years agoRe-enable most IR optimisations. It's only really redundant-Put
Julian Seward [Wed, 20 Oct 2004 23:14:42 +0000 (23:14 +0000)] 
Re-enable most IR optimisations.  It's only really redundant-Put
elimination which is causing a problem right now and so is disabled;
tree-building is also potentially dodgy.

git-svn-id: svn://svn.valgrind.org/vex/trunk@394

21 years agoClean up BB counting stuff a bit.
Julian Seward [Wed, 20 Oct 2004 19:38:50 +0000 (19:38 +0000)] 
Clean up BB counting stuff a bit.

git-svn-id: svn://svn.valgrind.org/vex/trunk@393

21 years agoBuild vg_replace_malloc.c into a library, because it makes the Makefiles a bit
Nicholas Nethercote [Wed, 20 Oct 2004 18:40:40 +0000 (18:40 +0000)] 
Build vg_replace_malloc.c into a library, because it makes the Makefiles a bit
neater.  Also remove some dodgy CFLAGS+= lines.

I had to change the expected output of pth_once.c, because the change has
altered the order of the (non-deterministic) output.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2825

21 years agoGet rid of -fomit-frame-pointer; it's mostly just a pain in the parts
Julian Seward [Wed, 20 Oct 2004 18:39:58 +0000 (18:39 +0000)] 
Get rid of -fomit-frame-pointer; it's mostly just a pain in the parts
that point southwards when one is walking northward.

git-svn-id: svn://svn.valgrind.org/vex/trunk@392

21 years agoOops! Further update spill slot offsets following add of
Julian Seward [Wed, 20 Oct 2004 18:38:07 +0000 (18:38 +0000)] 
Oops! Further update spill slot offsets following add of
x86 guest segment registers.  TODO: SANELY.

git-svn-id: svn://svn.valgrind.org/vex/trunk@391

21 years agoRearrange stuff a bit more logically.
Julian Seward [Wed, 20 Oct 2004 18:36:54 +0000 (18:36 +0000)] 
Rearrange stuff a bit more logically.

git-svn-id: svn://svn.valgrind.org/vex/trunk@390

21 years agoIn the signal handler, make debug printing a bit more verbose.
Julian Seward [Wed, 20 Oct 2004 18:36:23 +0000 (18:36 +0000)] 
In the signal handler, make debug printing a bit more verbose.

git-svn-id: svn://svn.valgrind.org/vex/trunk@389

21 years agoGive the skiplist mechanism its own hacky allocator, so as to get
Julian Seward [Wed, 20 Oct 2004 18:35:33 +0000 (18:35 +0000)] 
Give the skiplist mechanism its own hacky allocator, so as to get
rid of the following call cycle:

    VG_(mmap)
 -> VG_(map_fd_segment)
 -> VG_(map_file_segment)
 -> VG_(SkipNode_Alloc)
 -> VG_(arena_malloc)
 -> newSuperblock
 -> VG_(get_memory_from_mmap)
 -> VG_(mmap)

git-svn-id: svn://svn.valgrind.org/vex/trunk@388

21 years agoApparently editing tool.h directly was not a clever thing to do :-)
Julian Seward [Wed, 20 Oct 2004 18:33:03 +0000 (18:33 +0000)] 
Apparently editing tool.h directly was not a clever thing to do :-)

git-svn-id: svn://svn.valgrind.org/vex/trunk@387

21 years agoFix typo, courtesy of Aleksander Salwa.
Nicholas Nethercote [Wed, 20 Oct 2004 14:22:58 +0000 (14:22 +0000)] 
Fix typo, courtesy of Aleksander Salwa.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2823

21 years agoMake tests that cause coredumps remove them. Don't need the core entries in
Nicholas Nethercote [Wed, 20 Oct 2004 13:33:07 +0000 (13:33 +0000)] 
Make tests that cause coredumps remove them.  Don't need the core entries in
.cvsignore any more as a result.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2822

21 years agoArch-abstraction:
Nicholas Nethercote [Wed, 20 Oct 2004 13:21:54 +0000 (13:21 +0000)] 
Arch-abstraction:
- conditionally compile x86-specific line within yield.c.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2821

21 years agowhoops
Nicholas Nethercote [Wed, 20 Oct 2004 13:02:21 +0000 (13:02 +0000)] 
whoops

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2820

21 years agoUse the vex-supplied simulation of segment registers.
Julian Seward [Wed, 20 Oct 2004 12:37:45 +0000 (12:37 +0000)] 
Use the vex-supplied simulation of segment registers.

git-svn-id: svn://svn.valgrind.org/vex/trunk@386

21 years agoArch-abstraction:
Nicholas Nethercote [Wed, 20 Oct 2004 12:36:47 +0000 (12:36 +0000)] 
Arch-abstraction:
- Factor out remaining x86-specific regtests in none/ into an x86/ subdir.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2819

21 years agoStart to teach Vex (guest x86) about segment registers.
Julian Seward [Wed, 20 Oct 2004 12:36:12 +0000 (12:36 +0000)] 
Start to teach Vex (guest x86) about segment registers.

git-svn-id: svn://svn.valgrind.org/vex/trunk@385

21 years agoWhen nuking threads we set their status to VgTs_Empty. This patch
Tom Hughes [Wed, 20 Oct 2004 12:25:59 +0000 (12:25 +0000)] 
When nuking threads we set their status to VgTs_Empty. This patch
ensures that any associated mutex and conditional variable pointers
are also cleared so that the thread state appears sane.

BUG: 91599

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2818

21 years agoArch-abstraction:
Nicholas Nethercote [Wed, 20 Oct 2004 11:43:01 +0000 (11:43 +0000)] 
Arch-abstraction:
- move remaining x86-specific Memcheck tests into an x86/ subdir.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2817

21 years agoArch-abstraction:
Nicholas Nethercote [Wed, 20 Oct 2004 11:05:34 +0000 (11:05 +0000)] 
Arch-abstraction:
- Abstractify one remaining mention of the stack pointer.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2816

21 years agoArch-abstraction:
Nicholas Nethercote [Wed, 20 Oct 2004 10:58:09 +0000 (10:58 +0000)] 
Arch-abstraction:
- Make Helgrind totally arch-agnostic.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2815

21 years agoArch-abstraction:
Nicholas Nethercote [Wed, 20 Oct 2004 10:37:00 +0000 (10:37 +0000)] 
Arch-abstraction:
- Make Lackey talk about "machine instructions" rather than "x86 instructions".

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2814

21 years agocat_n_send: get rid of bogus dependency on snprintf(). This should
Julian Seward [Wed, 20 Oct 2004 09:56:55 +0000 (09:56 +0000)] 
cat_n_send: get rid of bogus dependency on snprintf().  This should
possibly go in the CVS repo.

git-svn-id: svn://svn.valgrind.org/vex/trunk@384

21 years agoDisable all optimisation until such time as the 2.3.X route works
Julian Seward [Wed, 20 Oct 2004 09:38:58 +0000 (09:38 +0000)] 
Disable all optimisation until such time as the 2.3.X route works
better.  In particular, redundant-Put removal is removing assignments
to x86 guest %esp, which messes up the segfault-based stack extender.

git-svn-id: svn://svn.valgrind.org/vex/trunk@383

21 years agoRemove debug printing.
Julian Seward [Wed, 20 Oct 2004 09:37:05 +0000 (09:37 +0000)] 
Remove debug printing.

git-svn-id: svn://svn.valgrind.org/vex/trunk@382

21 years agoIf Vex calls the failure-exit function, bomb out in such a way as
Julian Seward [Wed, 20 Oct 2004 09:36:13 +0000 (09:36 +0000)] 
If Vex calls the failure-exit function, bomb out in such a way as
to show stack traces of all current threads.

git-svn-id: svn://svn.valgrind.org/vex/trunk@381

21 years agoShow Vex storage stats at end of run.
Julian Seward [Wed, 20 Oct 2004 09:35:35 +0000 (09:35 +0000)] 
Show Vex storage stats at end of run.

git-svn-id: svn://svn.valgrind.org/vex/trunk@380

21 years agoProperly initialise baseBlock (I overlooked that fact that suitable
Julian Seward [Wed, 20 Oct 2004 01:17:19 +0000 (01:17 +0000)] 
Properly initialise baseBlock (I overlooked that fact that suitable
starting values need to be put into it)

git-svn-id: svn://svn.valgrind.org/vex/trunk@379

21 years agoMake the number of spill slots agree with what Vex's
Julian Seward [Wed, 20 Oct 2004 01:16:01 +0000 (01:16 +0000)] 
Make the number of spill slots agree with what Vex's
priv/host-x86/hdefs.c (genSpill, genRestore) think it should be.

git-svn-id: svn://svn.valgrind.org/vex/trunk@378

21 years agoGet rid of translation chaining stuff from dispatcher.
Julian Seward [Wed, 20 Oct 2004 01:13:44 +0000 (01:13 +0000)] 
Get rid of translation chaining stuff from dispatcher.

git-svn-id: svn://svn.valgrind.org/vex/trunk@377

21 years agoTemporary kludge to locate spill slots on cvs head. TODO: better
Julian Seward [Wed, 20 Oct 2004 01:13:02 +0000 (01:13 +0000)] 
Temporary kludge to locate spill slots on cvs head.  TODO: better

git-svn-id: svn://svn.valgrind.org/vex/trunk@376

21 years agofind_helper: ensure function addresses are unsignedly widened to Addr64
Julian Seward [Wed, 20 Oct 2004 01:12:08 +0000 (01:12 +0000)] 
find_helper: ensure function addresses are unsignedly widened to Addr64

git-svn-id: svn://svn.valgrind.org/vex/trunk@375

21 years agoMucho hackery to get it to compile whilst attached to vex.
Julian Seward [Tue, 19 Oct 2004 23:43:04 +0000 (23:43 +0000)] 
Mucho hackery to get it to compile whilst attached to vex.

git-svn-id: svn://svn.valgrind.org/vex/trunk@374

21 years agoImplement vex_initialise_x87().
Julian Seward [Tue, 19 Oct 2004 23:42:00 +0000 (23:42 +0000)] 
Implement vex_initialise_x87().

git-svn-id: svn://svn.valgrind.org/vex/trunk@373

21 years agoas_shm wasn't being included in 'make dist'.
Nicholas Nethercote [Tue, 19 Oct 2004 19:39:59 +0000 (19:39 +0000)] 
as_shm wasn't being included in 'make dist'.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2813

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 19:39:39 +0000 (19:39 +0000)] 
Arch-abstraction:
- final fix for insn_* test moving;  they now work correctly

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2812

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 19:12:21 +0000 (19:12 +0000)] 
Arch-abstraction:
- Forgot to move the insn_*.def files when moving the insn_* tests.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2811

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 19:07:32 +0000 (19:07 +0000)] 
Arch-abstraction:
- things I forgot to do when moving the insn_* tests...

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2810

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 18:54:11 +0000 (18:54 +0000)] 
Arch-abstraction:
- Moved all the insn_* tests into x86/ subdirectories.  What are the chances of
  me getting this right on the first attempt?

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2809

21 years agoreadability fiddle
Nicholas Nethercote [Tue, 19 Oct 2004 17:54:02 +0000 (17:54 +0000)] 
readability fiddle

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2808

21 years agoForgot to cvs remove this when I moved it to x86/.
Nicholas Nethercote [Tue, 19 Oct 2004 17:49:39 +0000 (17:49 +0000)] 
Forgot to cvs remove this when I moved it to x86/.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2807

21 years agoUpdate
Nicholas Nethercote [Tue, 19 Oct 2004 17:30:47 +0000 (17:30 +0000)] 
Update

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2806

21 years agoWhoops, forgot this file.
Nicholas Nethercote [Tue, 19 Oct 2004 17:30:25 +0000 (17:30 +0000)] 
Whoops, forgot this file.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2805

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 17:00:59 +0000 (17:00 +0000)] 
Arch-abstraction:
- Move x86-specific regtests for Cachegrind into an x86/ subdir.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2804

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 16:56:41 +0000 (16:56 +0000)] 
Arch-abstraction:
- Change regtest script to print a message if a test is skipped due to the cpu
  being of the wrong type.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2803

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 16:29:30 +0000 (16:29 +0000)] 
Arch-abstraction:
- Rewrote tests/cputest.c so that it can apply to different kinds of
  processors.  The idea being that any arch-specific tests have a cpu_test:
  label in their .vgtest file, so they'll only get executed if the right
  machine is being used.
- Rewrote a bunch of .vgtest files accordingly.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2802

21 years agoTest commit.
Julian Seward [Tue, 19 Oct 2004 15:55:45 +0000 (15:55 +0000)] 
Test commit.

git-svn-id: svn://svn.valgrind.org/vex/trunk@372

21 years agoAdd a function to initialise the x87 FPU guest state.
Julian Seward [Tue, 19 Oct 2004 15:47:32 +0000 (15:47 +0000)] 
Add a function to initialise the x87 FPU guest state.

git-svn-id: svn://svn.valgrind.org/vex/trunk@371

21 years agoUse "ppc" rather than "powerpc" -- it's less typing.
Nicholas Nethercote [Tue, 19 Oct 2004 14:24:42 +0000 (14:24 +0000)] 
Use "ppc" rather than "powerpc" -- it's less typing.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2801

21 years agoWhoops, syscall.S is platform-specific, and so must go in x86-linux/ rather
Nicholas Nethercote [Tue, 19 Oct 2004 14:23:46 +0000 (14:23 +0000)] 
Whoops, syscall.S is platform-specific, and so must go in x86-linux/ rather
than x86/.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2800

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 13:48:06 +0000 (13:48 +0000)] 
Arch-abstraction:
- Moved all assembly files in coregrind/ into coregrind/x86/.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2799

21 years agoArch-abstraction:
Nicholas Nethercote [Tue, 19 Oct 2004 13:18:00 +0000 (13:18 +0000)] 
Arch-abstraction:
- Factored out the remaining arch-specific code from vg_libpthread.c.
- Also fixed up the build process for x86/libpthread.c, which was done
  wrongly in the previous commit.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2798

21 years agoFix some typos spotted by Nick.
Julian Seward [Tue, 19 Oct 2004 13:17:33 +0000 (13:17 +0000)] 
Fix some typos spotted by Nick.

git-svn-id: svn://svn.valgrind.org/vex/trunk@370

21 years agoAdd a complete copy of the cvs head as of 19 Oct 04, for experimenting
Julian Seward [Tue, 19 Oct 2004 13:11:35 +0000 (13:11 +0000)] 
Add a complete copy of the cvs head as of 19 Oct 04, for experimenting
with integration purposes.

git-svn-id: svn://svn.valgrind.org/vex/trunk@369