This required some tricks with casting to maintain Memcheck's silly (ie.
negative) arg checking. The allocator was also changed accordingly. It
should now be able to allocate more than 4GB blocks on 64-bit platforms.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2906
- Added include/x86-linux/ and include/linux/ subdirectories, with Makefile.am
files.
- Overhauled the definitions of kernel types. include/vg_kerneliface.h is now
three files, include/linux/vki.h, include/x86-linux/vki_arch.h, and
include/x86-linux/vki_arch_posixtypes.h. These files separate the
common/Linux and x86/Linux parts cleanly. All code is copied verbatim from
the relevant kernel headers, except that VKI_/vki_ prefixes are added as
necessary to distinguish them from glibc types. (This is done consistently,
unlike previously when some types did not have the prefixes.)
All code is clearly marked to show which particular header file it came from,
and the Linux version used. (I used 2.6.8.1, the most recent stable release,
for all of them.)
A few of the types changed; this is because they changed between the older
versions of Linux and the current 2.6.8.1. I checked that all these changes
were ok with respect to backwards compatibility for our purposes.
- vg_unsafe.h has been removed; we are no longer including any kernel headers,
as we have our own copies for everything. This is because installed kernel
headers are not reliable, and often cause compilation problems. (bug
#92420 is a recent example)
- Removed some no-longer-needed header-presence tests from configure.in.
- Some code in the rest of Valgrind was changed to account for some slight
changes in the names of our VKI_/vki_ kernel constants and types.
- Updated README_MISSING_SYSCALL_OR_IOCTL accordingly.
- Fixed off-by-one error with VKI_GDT_ENTRY_TLS_MAX (merged from stable branch)
The end result is that the kernel types situation should be much clearer, and
similar files can be created relatively easily for other architectures as
necessary.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2884
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
- 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
- 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
of using an assembly hack to find the stack pointer at startup, we find it from
argv. It's much simpler, avoids linking games, is platform independent, and
works on PPC.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2782
Valgrind itself (the files ume.c, ume_entry.c and jmp_with_stack.c). Thus,
we are using Memcheck to check these files in a unit test setting.
I hope to do unit self-testing for many more parts of Valgrind, eventually all
the bits that can be pulled out into any kind of sensible stand-alone form.
Doing so achieves two things:
a) it introduces unit testing into our framework (a serious shortcoming at the
moment)
b) it lets us use Valgrind (esp. Memcheck) on itself, to some extent
This should help reliability. This first unit self-test isn't very exhaustive,
but it's a start.
Note that this involves something like bootstrapping, in that we are checking
parts of a Valgrind build with itself. I don't think this will be a problem,
since we (at least, I do) tend to only run the regtests when we think the
Valgrind build is ok.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2760
signal handler present -- previously, Valgrind would abort unnecessarily on
this case.
Added a regression test for it.
MERGE TO STABLE
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2743
actually does what the comment says it does. Patch from Nathan Kurz.
CCMAIL: 90778-done@bugs.kde.org
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2728
- Added include/x86/: contains tool_arch.h, Makefile.am, .cvsignore.
- Added coregrind/x86/state.c. Contains some arch-specific code for dealing
with x86 registers -- eg. setting up the baseBlock, loading/saving the whole
register state. It is compiled into coregrind/x86/libarch.a and linked via
${VG_ARCH} with the core.
Relatedly, also added coregrind/x86/{core_arch.h,core_arch_asm.h}.
- Correspondingly abstracted the register state out of ThreadState. This
affected every place that touches registers, and there are a lot of them.
(Eventually all the register touching should be abstracted out in an
arch-neutral way, but not yet; one step at a time.)
- Added some declarations about register loading/saving functions to core.h;
all architectures will have to provide these functions.
- Rejigged the build system so that the arch-specific stuff is all done via
${VG_ARCH}, rather than naming e.g. x86/ directly. Appropriate -I arguments
are used so that all the headers are found, etc.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2656
Two were trivial. The one for strncmp is slightly trickier; you have to be
careful with the signedness of chars when comparing them... but the code
already casts s1 and s2 to (unsigned char*) before comparing them, so it's not
a problem.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2618
dist". This avoids the need to put it in "EXTRA_DIST", and saves a few lines
in all the docs Makefile.am files.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2614
__attribute((regparm(n))) with REGPARM(n) everywhere. REGPARM() is defined in
vg_skin.h, but will eventually be defined separately for each arch.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2601
lots of the details changed. Made the following generalisations:
- Recast everything to be entirely terms of bytes, instead of a mixture
of (32-bit) words and bytes. This is a bit easier to understand, and
made the following generalisations possible...
- Almost 64-bit clean; no longer assuming 32-bit words/pointers. Only
(I think) non-64-bit clean part is that VG_(malloc)() et al take an
Int as the size arg, and size_t is 64-bits on 64-bit machines.
- Made the alignment of blocks returned by malloc() et al completely
controlled by a single value, VG_MIN_MALLOC_SZB. (Previously there
were various magic numbers and assumptions about block alignment
scattered throughout.) I tested this, all the regression tests pass
with VG_MIN_MALLOC_SZB of 4, 8, 16, 32, 64. One thing required for
this was to make redzones elastic; the asked-for redzone size is now
the minimum size; it will use bigger ones if necessary to get the
required alignment.
Some other specific changes:
- Made use of types a bit more; ie. actually using the type 'Block',
rather than just having everything as arrays of words, so that should
be a bit safer.
- Removed the a->rz_check field, which was redundant wrt. a->clientmem.
- Fixed up the decision about which list to use so the 4 lists which
weren't ever being used now are -- the problem was that this hasn't
been properly updated when alignment changed from 4 to 8 bytes.
- Added a regression test for memalign() and posix_memalign().
memalign() was aborting if passed a bad alignment argument.
- Added some high-level comments in various places, explaining how the
damn thing works.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@2579
