Petar Jovanovic [Sat, 13 Jul 2013 23:50:46 +0000 (23:50 +0000)]
mips32/mips64: Avoid breakpoints in branch delay slots
Reusing parts of Chris Dearman's change in GDB to avoid placing breakpoints
in a branch delay slot.
Fixes gdbserver_tests/mcbreak for MIPS32 and MIPS64.
Petar Jovanovic [Fri, 12 Jul 2013 15:32:27 +0000 (15:32 +0000)]
mips32: another VG_(am_get_advisory) needs non-single-page-size adjustment
Another mmap issue in which another VG_(am_get_advisory) needs adjustment
wrapper for cases when (VKI_SHMLBA > VKI_PAGE_SIZE) and argument is
VKI_MAP_SHARED.
Fix by DejanJ for Bug #320057.
Issue and the test case by Vasile Floroiu.
Add test cases pertaining to vex r2731, for the following instructions:
SSAX SXTAB16 SHASX SHSAX SHSUB16 SHSUB8
UASX USAX UQADD16 UQASX UQSAX UHASX UHSAX REVSH
Mark Wielaard [Fri, 28 Jun 2013 14:03:58 +0000 (14:03 +0000)]
Bug 289360 parse_type_DIE confused by DW_TAG_enumeration_type.
GCC allows incomplete enums as GNU extension.
http://gcc.gnu.org/onlinedocs/gcc/Incomplete-Enums.html
These are marked as DW_AT_declaration and won't have a size.
They can only be used in declaration or as pointer types.
You can't allocate variables or storage using such an enum type.
So don't require a size for such enum types.
Julian Seward [Thu, 27 Jun 2013 20:31:36 +0000 (20:31 +0000)]
Exit a bit more gracefully if a request to get part of an image
exceeds the allowable range. With this change, it should be
essentially impossible to crash V by feeding it invalid ELF or Dwarf.
Florian Krohm [Mon, 17 Jun 2013 19:04:24 +0000 (19:04 +0000)]
valgrind side changes to support the new IRops introduced
in VEX r2727 (conversion ops between binary floating point and
decimal floating point).
Patch by Maran Pakkirisamy (maranp@linux.vnet.ibm.com).
Part of fixing BZ 307113.
improve --help for --main-stacksize and supported ARM cpu
If the command line option --main-stacksize is not used,
the current ulimit value is used, with a min of 1MB
and a max of 16MB. Document this min/max default formula
in the --help.
Petar Jovanovic [Sun, 2 Jun 2013 18:08:04 +0000 (18:08 +0000)]
mips64: disable intercepting the spinlock function
This change disables intercepting pthread_spin_lock() for MIPS64, similar to
r13190 for MIPS32. Used in DRD and Helgrind as a workaround for the issue
#311690.
Petar Jovanovic [Sun, 2 Jun 2013 03:22:28 +0000 (03:22 +0000)]
update cond_ld_st test and expected outputs
Line numbers are off by one in some outputs since it differs on GCC version
how it will map source code to line number for the ternary operator written
in multiple lines.
This is small fix that replaces one occurrence of ternary operator with
if-then-else. The excepted outputs have been changed accordingly.
This fixes the test memcheck/tests/cond_ld_st for different platforms with
different GCC versions.
Petar Jovanovic [Sun, 2 Jun 2013 02:59:07 +0000 (02:59 +0000)]
add and update exp files for drd/tests/annotate_trace_memory
exp-32bit files needed minor update for the last store, and some mips32
platforms also need additional exp-mips32 due to different way of loading
and storing double values.
This fixes drd/tests/annotate_trace_memory on different 32-bit platforms.
Florian Krohm [Fri, 31 May 2013 15:44:06 +0000 (15:44 +0000)]
s390x: The CC_DEP1 field is accessed as a whole or bits [0:31]
in case the field holds a 32-bit floating point value.
Patch by Maran Pakkirisamy (maranp@linux.vnet.ibm.com).
Part of fixing BZ 307113.
fix 320211 Stack buffer overflow in ./coregrind/m_main.c with huge TMPDIR
* Addition of a function to compute size of buffer needed for VG_(mkstemp)
* Use it to dimension buffers for all VG_(mkstemp) calls.
Mark Wielaard [Wed, 22 May 2013 10:21:10 +0000 (10:21 +0000)]
Support Linux kernel AF_BLUETOOTH for bind().
Bug #320116. sockaddr_rc might contain some padding which might not be
initialized. Explicitly check the sockaddr_rc fields are set. That also
produces better diagnostics about which field is unitialized.
Petar Jovanovic [Sat, 11 May 2013 22:23:27 +0000 (22:23 +0000)]
mips: define allexec files as soft links
This change is a follow up to r13385. It removes physical copies of the
allexec c-files in none/tests/mips32/ and none/tests/mips64/ and defines
them as soft-links to a common file like other arches do.
Julian Seward [Sat, 11 May 2013 13:42:08 +0000 (13:42 +0000)]
complainIfUndefined: reinstate the 3rd argument (guard) so as to make
the definedness check and possible shadow temp set-to-defined be
optional. Use this to properly instrument IRLoadG and IRStoreG, so
that if the load/store does not happen, not only is the validity of
the address not checked, neither is the definedness.
This fixes a regression introduced by the COMEM branch on ARM, in
which conditional loads/stores with addresses which are undefined at
runtime and with guards which are false, would generate false errors.
Also extensively re-checked the check-generation machinery and updated
a bunch of comments.
fix 319235 --db-attach=yes is broken with Yama ptrace scoping enabled
On Ubuntu systems, ptrace_scoping could forbid a process to ptrace another.
This ptrace scoping was already handled for vgdb by using SET_PTRACER
(the valgrind process must be ptraced by vgdb when it is blocked
in a syscall).
set_ptracer is however also needed when the old mechanism --db-attach=yes
is used.
The following changes are done:
* make the set_ptracer logic callable outside gdbserver
* make set_ptracer less restrictive (i.e. allow all
processes of the user to ptrace). This removes a limitation for vgdb.
* call the set_ptracer in the child launched for --db-attach=yes
* cleaned up the ptrace scope restriction message and doc as vgdb
is now working properly by default, even with ptrace_scope enabled.
fix gdbsrv inferior calls when PT_GNU_STACK declares stack not executable
With rev 13368, Valgrind obeys PT_GNU_STACK making the stack not
executable. This makes inferior function call with GDB >= 7.5 failing,
as GDB places a breakpoint on the stack, which must be decoded
and translated by Valgrind to have the inferior function call properly done.
=> introduce a special case in the conditions to allow translation
when a segment is not executable but is readable and there is a
breakpoint at the address.
Minor changes to how V{LD}{1,2,3,4} are tested:
* print the output 64-bit results in most-to-least-significant
lane format
* change the memory values used in these tests to ones which
are reliably able to detect confusion of 8-bit lanes. The
previous values couldn't to that reliable. This change causes
the diff to be huge because the stdout.exp changes a lot.
Petar Jovanovic [Fri, 19 Apr 2013 15:23:44 +0000 (15:23 +0000)]
Calculate offsets in read_dwarf2_lineblock based on is64
Reading header length and values in external line info was incorrect at
some places as it used offsets based on dw64 that came from .debug_info.
Instead, offsets should be calculated based on is64 from .debug_line.
This issue surfaced in MIPS64 port, and it was discussed at:
Non-functional changes: flush stdout frequently, so as to make
it easier to sync with stderr output. Also, add some more helpful
data for testing Neon loads/stores.
Mark Wielaard [Wed, 17 Apr 2013 19:11:05 +0000 (19:11 +0000)]
Fix -Ttext-segment configure check.
Explicitly test together with -static -nodefaultlibs -nostartfiles to mimic
what the tools linking script does. At least on s390 the test might fail for
a non-static AC_LINK, while it does work when using those other flags too.
Mark Wielaard [Wed, 17 Apr 2013 13:48:29 +0000 (13:48 +0000)]
Simplify read_unitinfo_dwarf2. Only try to read the first DIE.
Bug #305513. We should only read the first DIE of a compilation unit.
Each compilation unit header is followed by a single DW_TAG_compile_unit
(or DW_TAG_partial_unit, but those aren't important here) and its children.
There is no reason to read any of the children at this point. If the first
DIE isn't a DW_TAG_compile_unit we are done, none of the child DIEs will
provide any useful information.
Mark Wielaard [Thu, 11 Apr 2013 17:55:39 +0000 (17:55 +0000)]
read_unitinfo_dwarf2 DW_FORM_ref_addr is address size in DWARF version 2.
Bug #305513 contained a patch for some extra robustness checks. But
the real cause of crashing in the read_unitinfo_dwarf2 DWARF reader
seemed to have been this issue where DWARF version 2 DWZ partial_units
were read and DW_FORM_ref_addr had an unexpected size. This combination
is rare. DWARF version 4 is the current default version of GCC.
Solve false negative for various malloc replaced functions arguments
The replacement functions are running on the simulated CPU.
The code on the simulated CPU does not necessarily use
all arguments. E.g. args can be ignored and/or only given
to a NON SIMD call.
The definedness of such 'unused' arguments will not be verified
by memcheck.
A call to 'trigger_memcheck_error_if_undefined' allows
memcheck to detect such errors for the otherwise unused args.
Apart of allowing memcheck to detect an error, the function
trigger_memcheck_error_if_undefined has no effect and
has a minimal cost for other tools replacing malloc functions.
(suggestion of the 'no operation check' from Julian).
tested on f12/x86, debian6/amd64, f18/ppc64
Note that some Darwin specific code has been modified
in coregrind/m_replace_malloc/vg_replace_malloc.c.
(Some of) this code has not been compiled (no access to a
Darwin system). The code changed is trivial, so there is
some chance it will compile and even maybe work.
Added a new test verifying that various malloc related
functions undefined args are triggering an error in memcheck.
Petar Jovanovic [Thu, 4 Apr 2013 10:55:09 +0000 (10:55 +0000)]
mips: add different value for SOCK_STREAM
MIPS uses different values for socket types.
This is protected by ARCH_HAS_SOCKET_TYPES in Linux kernel and we introduce
it here too. This is important for log-socket feature, and it resolves the
issue reported in https://bugs.kde.org/show_bug.cgi?id=313267#c21.
projects / thirdparty / valgrind.git / log
