Richard Biener [Wed, 20 Feb 2013 11:39:39 +0000 (11:39 +0000)]
re PR tree-optimization/56396 (memory corruption in cc1)
2013-02-20 Richard Biener <rguenther@suse.de>
Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/56396
* tree-ssa-ccp.c (n_const_val): New static variable.
(get_value): Return NULL for SSA names we don't have a lattice
entry for.
(ccp_initialize): Initialize n_const_val.
* tree-ssa-copy.c (n_copy_of): New static variable.
(init_copy_prop): Initialize n_copy_of.
(get_value): Return NULL_TREE for SSA names we don't have a
lattice entry for.
* gcc.dg/pr56396.c: New testcase.
Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
From-SVN: r196170
Richard Biener [Wed, 20 Feb 2013 09:03:18 +0000 (09:03 +0000)]
tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits and ceil_log2.
2013-02-20 Richard Biener <rguenther@suse.de>
* tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
and ceil_log2.
(get_use_iv_cost): Terminate hashtable walk when coming across
an empty entry.
Benjamin Kosnik [Wed, 20 Feb 2013 01:52:36 +0000 (01:52 +0000)]
user.cfg.in: Set __cplusplus to 201103L.
2013-02-19 Benjamin Kosnik <bkoz@redhat.com>
* doc/doxygen/user.cfg.in: Set __cplusplus to 201103L. Change to
_GLIBCXX_INCLUDE_AS_CXX11. DIRECTORY_GRAPH, MARKDOWN_SUPPORT,
AUTOLINK_SUPPORT to NO. Update to doxygen 1.8.3.1.
* include/bits/stl_pair.h: Add to utilities group.
* include/std/tuple: Same.
* include/std/typeindex: Same.
Add support for the CLZ insn
* config/microblaze/microblaze.c: microblaze_has_clz = 0
Add version check for v8.10.a to enable microblaze_has_clz
* config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as
combined version and TARGET_PATTERN_COMPARE check
* config/microblaze/microblaze.md: New clzsi2 instruction
Jakub Jelinek [Tue, 19 Feb 2013 09:14:33 +0000 (10:14 +0100)]
re PR tree-optimization/56350 (ICE in vectorizable_reduction, at tree-vect-loop.c:4731)
PR tree-optimization/56350
* tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
if haven't found reduction or nested cycle operand, rather than
asserting we must find it.
re PR target/56214 (FAIL: gcc.c-torture/execute/20050121-1.c compilation, ICE)
PR target/56214
* config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
and HImode, require all displacements to be an integer multiple of the
mode size.
* config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
only allow QImode and HImode when reload is in progress and strict is
true. Likewise for symbolic addresses. Use base14_operand to check
displacements in REG+BASE addresses.
Richard Biener [Mon, 18 Feb 2013 13:03:15 +0000 (13:03 +0000)]
re PR tree-optimization/56350 (ICE in vectorizable_reduction, at tree-vect-loop.c:4731)
2013-02-18 Richard Biener <rguenther@suse.de>
PR middle-end/56350
* cfghooks.c (merge_blocks): If we merge a latch into another
block adjust references to it.
* cfgloop.c (flow_loops_find): Reset latch before recomputing it.
(verify_loop_structure): Verify that a recorded latch is in fact
a latch.
Add TARGET_SUPPORTS_PIC flag and check that the flag_pic = 2
* config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
* config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
* config/microblaze/microblaze.c (microblaze_option_override):
Bail out early for PIC modes when target does not support PIC.
* gcc.dg/20020312-2.c: Define MicroBlaze PIC register
microblaze.c (microblaze_asm_trampoline_template): Replace with a microblaze version.
* config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
Replace with a microblaze version.
(microblaze_trampoline_init): Adapt for microblaze.
* gcc/config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
microblaze.
Jakub Jelinek [Sat, 16 Feb 2013 09:32:56 +0000 (10:32 +0100)]
[asan] Fix for PR asan/56330
gcc/
* asan.c (get_mem_refs_of_builtin_call): White space and style
cleanup.
(instrument_mem_region_access): Do not forget to always put
instrumentation of the of 'base' and 'base + len' in a "if (len !=
0) statement, even for cases where either 'base' or 'base + len'
are not instrumented -- because they have been previously
instrumented. Simplify the logic by putting all the statements
instrument 'base + len' inside a sequence, and then insert that
sequence right before the current insertion point. Then, to
instrument 'base + len', just get an iterator on that statement.
And do not forget to update the pointer to iterator the function
received as argument.
gcc/testsuite/
* c-c++-common/asan/no-redundant-instrumentation-4.c: New test file.
* c-c++-common/asan/no-redundant-instrumentation-5.c: Likewise.
* c-c++-common/asan/no-redundant-instrumentation-6.c: Likewise.
* c-c++-common/asan/no-redundant-instrumentation-7.c: Likewise.
* c-c++-common/asan/no-redundant-instrumentation-8.c: Likewise.
* c-c++-common/asan/pr56330.c: Likewise.
* c-c++-common/asan/no-redundant-instrumentation-1.c (test1):
Ensure the size argument of __builtin_memcpy is a constant.
Ian Lance Taylor [Fri, 15 Feb 2013 18:55:09 +0000 (18:55 +0000)]
runtime: Do not reserve huge amount of swap on 32 bit architectures.
The mmap() call which reserves the arena should have MAP_NORESERVE
flag as in typical cases this memory will never be (fully) needed.
This matters in environments which do not do Linux style memory
overcommit, such as OpenIndiana/OpenSolaris/Solaris.
The MAP_NORESERVE flag does not exist on all operating systems
(for example FreeBSD). Therefore we define it to zero value in
case it does not exist.
Alan Modra [Fri, 15 Feb 2013 13:53:40 +0000 (00:23 +1030)]
re PR target/55431 (Invalid auxv search in ppc linux-unwind code.)
PR target/55431
* config/rs6000/linux-unwind.h (ppc_linux_aux_vector): Delete.
(ppc_fallback_frame_state): Always set up save locations for fp
and altivec. Don't bother with non-callee-saved regs, r0-r13
except for r2 on ppc64, fr0-fr13, v0-v19, vscr.
Richard Biener [Thu, 14 Feb 2013 12:24:12 +0000 (12:24 +0000)]
re PR lto/50494 (gcc.dg/vect/vect-reduc-2char.c fails spuriously on ppc with -flto)
2013-02-14 Richard Biener <rguenther@suse.de>
PR lto/50494
* varasm.c (output_constant_def_1): Get the decl representing
the constant as argument.
(output_constant_def): Wrap output_constant_def_1.
(make_decl_rtl): Use output_constant_def_1 with the decl
representing the constant.
(build_constant_desc): Optionally re-use a decl already
representing the constant.
(tree_output_constant_def): Adjust.
Jakub Jelinek [Thu, 14 Feb 2013 08:40:13 +0000 (09:40 +0100)]
dejagnu-gtest.h: Add multiple inclusion guards.
* g++.dg/asan/dejagnu-gtest.h: Add multiple inclusion guards.
* asan_globals_test-wrapper.cc: New file.
* g++.dg/asan/asan_test.C: Use asan_globals_test-wrapper.cc
instead of asan_globals_test.cc as dg-additional-sources.
Include asan_mem_test.cc, asan_str_test.cc and asan_oob_test.cc.
* g++.dg/asan/asan_test_utils.h: Synced from upstream. Include
"sanitizer_test_utils.h" instead of
"sanitizer_common/tests/sanitizer_test_utils.h".
* g++.dg/asan/asan_str_test.cc: New file, synced from upstream.
* g++.dg/asan/asan_mem_test.cc: New file, synced from upstream.
* g++.dg/asan/asan_oob_test.cc: New file, synced from upstream.
* g++.dg/asan/asan_globals_test.cc: Synced from upstream.
* g++.dg/asan/asan_test.cc: Synced from upstream.
* g++.dg/asan/sanitizer_test_utils.h: New file, synced from upstream.
Marc Glisse [Wed, 13 Feb 2013 21:58:53 +0000 (22:58 +0100)]
re PR libstdc++/56111 ({float,double,long double} complex not accepted anymore)
2013-02-13 Marc Glisse <marc.glisse@inria.fr>
PR libstdc++/56111
* include/std/complex (complex): Undefine.
* include/c_compatibility/complex.h (complex): Only undefine if
<complex> has been included.
* testsuite/26_numerics/complex/56111.cc: New testcase.
* include/bits/hashtable_policy.h (_Hash_code_base): Restore
default constructor protected.
* include/bits/hashtable.h: static assert that _Hash_code_base has
a default constructor available through inheritance.
* asan.c (create_cond_insert_point): Add create_then_fallthru_edge
argument. If it is false, don't create edge from then_bb to
fallthru_bb.
(insert_if_then_before_iter): Pass true to it.
(build_check_stmt): Pass false to it.
(transform_statements): Flush hash table only on extended basic
block boundaries, rather than at the beginning of every bb.
Don't flush hash table on nonfreeing_call_p calls.
* tree-flow.h (nonfreeing_call_p): New prototype.
* tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
Jakub Jelinek [Wed, 13 Feb 2013 16:09:27 +0000 (17:09 +0100)]
re PR c++/56302 (sys/sdt.h asms stopped working at -O0)
PR c++/56302
* semantics.c (finish_asm_stmt): If input constraints allow
neither register nor memory, try maybe_constant_value to get
a constant if possible.
* g++.dg/torture/pr56302.C: New test.
* g++.dg/cpp0x/constexpr-56302.C: New test.
* c-c++-common/pr56302.c: New test.
Richard Biener [Wed, 13 Feb 2013 11:42:04 +0000 (11:42 +0000)]
loop-init.c (loop_optimizer_init): Clear loop state when re-initializing preserved loops.
2013-02-13 Richard Biener <rguenther@suse.de>
* loop-init.c (loop_optimizer_init): Clear loop state when
re-initializing preserved loops.
* loop-unswitch.c (unswitch_single_loop): Return whether
we unswitched the loop. Do not verify loop state here.
(unswitch_loops): When we unswitched a loop discover new
loops.
Dodji Seketeli [Wed, 13 Feb 2013 10:32:26 +0000 (10:32 +0000)]
[asan] Avoid instrumenting duplicated memory access in the same basic block
Like what Address Sanitizer does in LLVM, this patch avoids instrumented
duplicated memory accesses in the same basic blocks.
The approach taken is very conservative, to keep the pass simple, for
a start.
A memory access is considered to be a pair made of an expression tree
representing the beginning of the memory region that is accessed and
a the size of the access, in byte. For now that size is either 1, 2,
4, 8 or 16 bytes.
The patch builds a hash table of the memory accesses that have been
instrumented in the current basic block. Then it walks the gimple
statements of the current basic block. For each statement, it tests
if the memory regions it references have already been instrumented.
If not, the statement is instrumented and each memory references that
are actually instrumented are added to the hash table. When a memory
region is accessed (usually through builtin functions like memset),
then what gets added to the hash table is actually two memory
accesses: one for the beginning of the region, and the other for the
its end.
When the patch crosses a function call that is not a built-in function
that we ought to instrument, the hash table is cleared, because that
function call can possibly e.g free some memory that was instrumented.
Likewise, when a new basic block is visited, the hash table is
cleared. I guess we could be smarter than just unconditionally
clearing the hash table in this later case, but this is what asan@llvm
does, and for now, I thought starting in a conservative manner might
have some value.
The hash table is destroyed at the end of the pass.
Bootstrapped and tested against trunk on x86-64-unknown-linux-gnu.
gcc/
* Makefile.in (asan.o): Add new dependency on hash-table.h
* asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
(asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
(has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
(free_mem_ref_resources, has_mem_ref_been_instrumented)
(has_stmt_been_instrumented_p, update_mem_ref_hash_table)
(get_mem_ref_of_assignment): New functions.
(get_mem_refs_of_builtin_call): Extract from
instrument_builtin_call and tweak a little bit to make it fit with
the new signature.
(instrument_builtin_call): Use the new
get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
of is_gimple_builtin_call.
(instrument_derefs, instrument_mem_region_access): Insert the
instrumented memory reference into the hash table.
(maybe_instrument_assignment): Renamed instrument_assignment into
this, and change it to advance the iterator when instrumentation
actually happened and return true in that case. This makes it
homogeneous with maybe_instrument_assignment, and thus give a
chance to callers to be more 'regular'.
(transform_statements): Clear the memory reference hash table
whenever we enter a new BB, when we cross a function call, or when
we are done transforming statements. Use
maybe_instrument_assignment instead of instrumentation. No more
need to special case maybe_instrument_assignment and advance the
iterator after calling it; it's now handled just like
maybe_instrument_call. Update comment.
Vladimir Makarov [Tue, 12 Feb 2013 17:44:47 +0000 (17:44 +0000)]
re PR inline-asm/56148 (inline asm matching constraint with different mode)
2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
PR inline-asm/56148
* lra-constraints.c (process_alt_operands): Match early clobber
operand with itself. Check conflicts with earlyclobner only if
the operand is not reloaded. Prefer to reload conflicting operand
if earlyclobber and matching operands are the same.
2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
PR inline-asm/56148
* gcc.target/i386/pr56148.c: New test.
projects / thirdparty / gcc.git / log
