Ivan Maidanski [Fri, 21 Apr 2017 21:28:35 +0000 (00:28 +0300)]
Eliminate 'possible loss of data' MS VC warning in disclaim_and_reclaim
* reclaim.c [ENABLE_DISCLAIM] (GC_disclaim_and_reclaim): Change type
of bit_no local variable from int to word (matching that in other
GC_reclaim_* functions).
Ivan Maidanski [Thu, 20 Apr 2017 08:50:03 +0000 (11:50 +0300)]
Eliminate 'possible loss of data' compiler warnings in cord (MS VC)
(Cherry-pick commit 1060ec9 from 'release-7_6' branch.)
* cord/cordbscs.c (CORD_cat_char_star, CORD_cat, CORD_from_fn_inner):
Add explicit cast in assignment of result->depth, result->len.
* cord/cordbscs.c (CORD_substr_checked): Change type of "j", "lim"
local variables from int to size_t.
* cord/cordbscs.c (CORD_init_min_len): Add explicit cast (to int) in
assignment of CORD_max_len.
* cord/cordxtra.c (CORD_ncmp): Add explicit cast (to long) when "avail"
local variable is updated.
* cord/cordxtra.c (CORD_chr, CORD_rchr, CORD_from_file_eager): Add
explicit cast to char (from int) for "c" local variable in assignment
of d.target and in CORD_ec_append call.
* cord/cordxtra.c (refill_cache): Add explicit cast to long (from
size_t) for line_start in fseek call.
* cord/tests/de.c (line_pos, fix_pos, do_command): Add explicit cast to
int in assignment of *c, line, col.
* cord/tests/de.c (do_command): Reset line, col and file_pos variables
(of different types) in separate statements.
* cord/tests/de_win.c (WinMain): Add explicit cast of msg.wParam to int
in return statement.
Ivan Maidanski [Thu, 13 Apr 2017 16:19:00 +0000 (19:19 +0300)]
Eliminate 'incompatible function pointer' warning in mark_some (MinGW/x86)
* include/private/gc_priv.h [__MINGW32__ && __i386__ && GC_EXTERN]
(__MINGW_EXCPT_DEFINE_PSDK): Define (before any system header include);
this macro definition is needed only when extra/gc.c is compiled.
* mark.c [__MINGW32__ && __i386__] (__MINGW_EXCPT_DEFINE_PSDK): Define
(before include gc_pmark.h).
Ivan Maidanski [Fri, 7 Apr 2017 16:14:24 +0000 (19:14 +0300)]
Update shared libraries version info to differentiate against v7.2
There some API changes in libgc against v7.2, so the "current" number
of the version info is incremented, thus the version info is now 3:0:1.
There are no API changes in libcord and libgccpp against v7.2
but some implementation changes exist, 2:*:* version info is reserved
for v7.2, so the version info for libcord and libgccpp is now 3:0:2.
* Makefile.am (LIBGC_VER_INFO): Change version info to 3:0:1.
* Makefile.am (LIBGCCPP_VER_INFO): Change version info to 3:0:2.
* cord/cord.am (LIBCORD_VER_INFO): Likewise.
Ivan Maidanski [Fri, 17 Mar 2017 08:05:56 +0000 (11:05 +0300)]
Allow HAVE_DL_ITERATE_PHDR to be defined by client (musl)
Do not redefine HAVE_DL_ITERATE_PHDR macro if already defined.
* dyn_load.c [OPENBSD>=200519] (HAVE_DL_ITERATE_PHDR): Do not define
if already defined.
* dyn_load.c [!USE_PROC_FOR_LIBRARIES && (__GLIBC__>=2
|| PLATFORM_ANDROID)] (HAVE_DL_ITERATE_PHDR): Likewise.
* dyn_load.c [!USE_PROC_FOR_LIBRARIES && HAVE_DL_ITERATE_PHDR
&& !DL_ITERATE_PHDR_STRONG] (dl_iterate_phdr): Define as weak symbol
(even if HAVE_DL_ITERATE_PHDR is defined outside this file).
Ivan Maidanski [Thu, 16 Mar 2017 07:22:29 +0000 (10:22 +0300)]
Eliminate 'ISO C forbids object to function pointer conversion' warning
(Cherry-pick commit c69ae9c from 'release-7_6' branch.)
* dyn_load.c [SOLARISDL && !USE_PROC_FOR_LIBRARIES
&& SUNOS53_SHARED_LIB] (GC_FirstDLOpenedLinkMap): Cast result of dlsym()
to word type before casting it to a function pointer.
* pthread_support.c [GC_USE_DLOPEN_WRAP] (GC_init_real_syms): Likewise.
Ivan Maidanski [Thu, 30 Mar 2017 19:06:09 +0000 (22:06 +0300)]
Fix mixed include of GC public header and gc_priv.h in disclaim bench/test
Include gc_priv.h should be avoided in tests (if possible) or, at
least, this include should not precede include of the public GC
headers (because, e.g., gc_priv.h define GC_BUILD which influences
declspec of the GC public symbols in case of a shared library).
* tests/disclaim_bench.c: Move include gc_priv.h to be after
gc_disclaim.h; add comment.
* tests/disclaim_bench.c [HAVE_CONFIG_H]: Include "config.h" (before
include gc_disclaim.h).
Ivan Maidanski [Tue, 14 Mar 2017 22:58:41 +0000 (01:58 +0300)]
Fix GC_SIZE_MAX definition (Linux/musl-gcc)
(Cherry-pick commit f40fab8 from 'release-7_6' branch.)
* include/private/gc_priv.h [SIZE_MAX] (GC_SIZE_MAX): Cast
SIZE_MAX to size_t explicitly (to workaround incorrect definition
of SIZE_MAX - e.g. it is defined as ~0ULL in musl-tools which is not
correct for a 32-bit target).
Ivan Maidanski [Thu, 30 Mar 2017 21:11:04 +0000 (00:11 +0300)]
Fix gc.h compliance to strict ANSI (pthreads)
(Cherry-pick commits 1554297, 0970110 from 'master' branch.)
If -std=cXX option is passed to gcc/clang then sigset_t is undefined,
thus GC_pthread_sigmask declaration should be omitted. However, this
symbol should be available during GC source code audit by cppcheck.
* include/gc_pthread_redirects.h [!GC_NO_PTHREAD_SIGMASK]
(GC_pthread_sigmask): Declare only if GC_PTHREAD_SIGMASK_NEEDED or
_POSIX_C_SOURCE >= 199506 or _XOPEN_SOURCE >= 500, or _BSD_SOURCE
or _GNU_SOURCE.
* tests/test.c [CPPCHECK && !_GNU_SOURCE && GC_PTHREADS] (_GNU_SOURCE):
Likewise.
Ivan Maidanski [Thu, 23 Mar 2017 20:12:07 +0000 (23:12 +0300)]
Fix finalize.c compilation in 'strict ANSI' mode
This and some other GC .c files include gc_pmark.h directly (not
gc_priv.h), so the compiler complains that sigset_t (used in
gc_pthread_redirects.h) is undefined if 'strict ANSI' mode is on and
_GNU_SOURCE is not defined before including system headers.
* include/private/gc_pmark.h [(__linux__ || __GLIBC__ || __GNU__)
&& GC_PTHREADS && !GC_NO_PTHREAD_SIGMASK] (_GNU_SOURCE): Define (before
including other headers).
Ivan Maidanski [Thu, 23 Mar 2017 19:44:11 +0000 (22:44 +0300)]
Fix static assertion violation in LONG_MULT for 64-bit targets
* include/private/gc_pmark.h [MARK_BIT_PER_OBJ] (PUSH_CONTENTS_HDR):
Cast displ to unsigned32 when passed to LONG_MULT (it is safe because
displ is less than HBLKSIZE).
* include/private/gc_pmark.h [MARK_BIT_PER_OBJ] (PUSH_CONTENTS_HDR):
Declare "obj_displ" local variable at the beginning of the block (not
after GC_STATIC_ASSERT).
Ivan Maidanski [Wed, 1 Mar 2017 08:10:00 +0000 (11:10 +0300)]
Move libraries version info to the beginning of Makefile.am
(code refactoring)
* Makefile.am (LIBGC_VER_INFO, LIBGCCPP_VER_INFO): New variable; add
comment about the version update rule.
* cord/cord.am (LIBCORD_VER_INFO): Likewise.
* Makefile.am (libgc_la_LDFLAGS, libgccpp_la_LDFLAGS): Use
LIBGC[CPP]_VER_INFO instead the direct hard-coding of version info.
* cord/cord.am (libcord_la_LDFLAGS): Likewise.
Ivan Maidanski [Tue, 31 Jan 2017 23:30:18 +0000 (02:30 +0300)]
Fix enable_parallel_mark condition in CMake script
* CMakeLists.txt [enable_parallel_mark] (ADD_DEFINITIONS): Add
-DPARALLEL_MARK; remove ADD_DEFINITIONS("-DPARALLEL_MARK") in other
places; remove ${} for enable_parallel_mark in IF.
Ivan Maidanski [Fri, 9 Dec 2016 23:07:25 +0000 (02:07 +0300)]
Workaround 'index out of bounds' UBSan false warning in push_marked
* mark.c (GC_push_marked): Cast (h+1)->hb_body to word before
subtraction of sz value (and cast the result back to ptr_t).
* mark.c [ENABLE_DISCLAIM] (GC_push_unconditionally): Likewise.
* tools/if_mach.c (main): Cast the 2nd argument of execvp to void* (to
avoid compiler warning as the argument of execvp() prototype could be
of "const char* const*" or "char**" type).
* tools/if_not_there.c (main): Likewise.
Ivan Maidanski [Tue, 15 Nov 2016 07:48:31 +0000 (10:48 +0300)]
Suppress 'tainted string passed to vulnerable operation' false defects
* include/private/gc_priv.h (TRUSTED_STRING): New tagging macro; add
comment.
* misc.c (GC_init): Process the result of GETENV("GC_LOG_FILE") by
TRUSTED_STRING.
* tools/if_mach.c (main): Process argv[3] by TRUSTED_STRING (before
passing the string to execvp).
* tools/if_not_there.c (main): Declare "fname" local variable;
process argv[1] and argv[2] by TRUSTED_STRING (before passing the
strings to fopen/opendir and execvp, respectively).
Ivan Maidanski [Tue, 20 Dec 2016 07:21:07 +0000 (10:21 +0300)]
Fix (adjust) GC_scratch_alloc actual argument type
* dyn_load.c [IRIX5 || USE_PROC_FOR_LIBRARIES && !LINUX]
(GC_register_dynamic_libraries): Do not cast GC_scratch_alloc argument
to word type (it should be of size_t).
* headers.c (alloc_hdr, GC_init_headers, get_index): Likewise.
* os_dep.c [PROC_VDB] (GC_proc_buf_size): Change type from word to size_t.
* os_dep.c [PROC_VDB] (GC_read_dirty): Change type of new_size local
variable (which is passed to GC_scratch_alloc) from word to size_t.
Ivan Maidanski [Mon, 12 Dec 2016 23:34:28 +0000 (02:34 +0300)]
Fix GC_noop6 definition to avoid its calls to be optimized away
(fix commits e3ec4a5, 1586a21)
* include/private/gc_priv.h (GC_ATTR_NOINLINE): New macro (effective
for GCC/Clang and MS VC currently).
* mark.c (GC_noop6): Use GC_ATTR_NOINLINE attribute; call
AO_compiler_barrier if PARALLEL_MARK (or GC_PTHREADS and not
GC_WIN32_THREADS), otherwise call GC_noop1(0); add
comment.
Ivan Maidanski [Thu, 22 Dec 2016 21:39:10 +0000 (00:39 +0300)]
Fix bm_huge initialization for 64-bit targets (gctest)
* tests/test.c (bm_huge): Make array size depend on word size
(5 elements for 64-bit targets); initialize elements (except the last
one) to -1 instead of 0xffffffff (so that all bits are ones even on
a 64-bit target); initialize last element to ((word)-1)>>8 instead of
0x00ffffff value.
* tests/test.c (typed_test): Fail if bm_huge descriptor contains wrong
bit values at certain positions (as it had before this fix).
projects / gc / log