rs6000: Add "cannot_copy" attribute, use it (PR67788, PR67789)
After the shrink-wrapping patches the prologue will often be pushed
"deeper" into the function, which in turn means the software trace cache
pass will more often want to duplicate the basic block containing the
prologue. This caused failures for 32-bit SVR4 with -msecure-plt PIC.
This configuration uses the load_toc_v4_PIC_1 instruction, which creates
assembler labels without using the normal machinery for that. If now
the compiler decides to duplicate the insn, it will emit the same label
twice. Boom.
It isn't so easy to fix this to use labels the compiler knows about (let
alone test that properly). Instead, this patch wires up a "cannot_copy"
attribute to be used by TARGET_CANNOT_COPY_P, and sets that attribute on
these insns we do not want copied.
Aditya Kumar [Thu, 1 Oct 2015 22:04:20 +0000 (22:04 +0000)]
use sese_l throughout scop-detection
Use sese_l throughout SCoP detection and create vec<scop_p> at the very end when
all SCoPs have been identified. 'struct sese_l' is very lightweight (two
pointers) compared to 'struct scop'.
No functional change intended. Passes regtest and bootstrap.
gcc/ChangeLog:
2015-10-01 Aditya Kumar <hiraditya@msn.com>
* graphite-scop-detection.c (struct sese_l): New conversion constructor
so that this type can be pushed into a vec.
(class scop_builder): use sese_l to collect scops.
(get_scops): New getter function.
(remove_intersecting_scops): Use sese_l instead of scops_p.
(intersects): Same.
(add_scop): Same.
(subsumes): Same.
(remove_subscops): Same.
(build_scops): Add scops to vec<scops_p> once all the scops have been
detected.
Aditya Kumar [Thu, 1 Oct 2015 22:04:14 +0000 (22:04 +0000)]
Rename gimple_bb to gimple_poly_bb
Renaming gimple_bb to gimple_poly_bb because there is a function gimple_bb
by the same name in gimple.h. No functional change intended.
Passes regtest and bootstrap.
gcc/ChangeLog:
2015-10-01 Aditya Kumar <hiraditya@msn.com>
* graphite-isl-ast-to-gimple.c (class translate_isl_ast_to_gimple):
Renamed type from gimple_bb_p to gimple_poly_bb_p.
(translate_isl_ast_node_user): Same.
* graphite-poly.c (new_poly_bb): Same.
* graphite-poly.h (gbb_from_bb): Same.
* sese.h: Same.
* graphite-sese-to-poly.c (new_gimple_bb): gimple_bb_p -> gimple_poly_bb_p
(build_scop_scattering): Same.
(find_params_in_bb): Same.
(add_conditions_to_domain): Same.
(sese_dom_walker::before_dom_children): Same.
(analyze_drs_in_stmts): Same.
(new_pbb_from_pbb): Same.
(free_data_refs_aux): New pointer to type base_alias_pair.
* graphite-sese-to-poly.h: Same.
* sese.c (if_region_set_false_region): Fixed Indentation.
(move_sese_in_condition): Same.
Sebastian Pop [Thu, 1 Oct 2015 22:04:07 +0000 (22:04 +0000)]
correctly handle non affine data references
2015-10-01 Sebastian Pop <s.pop@samsung.com>
Aditya Kumar <aditya.k7@samsung.com>
PR tree-optimization/66980
* graphite-scop-detection.c (stmt_has_simple_data_refs_p): Return false
when data reference analysis has failed.
re PR c++/54430 ([C++11] For-Loop: Scope of iterating variable begins too early)
PR c++/54430
/cp
2015-10-01 Ville Voutilainen <ville.voutilainen@gmail.com>
PR c++/54430
* name-lookup.c (push_binding): Make non-static.
* name-lookup.h (push_binding): Declare it.
* parser.c (cp_parser_range_for): Use it, get the range
declaration away from the scope until the range expression
has been parsed, then restore the declaration.
/testsuite
2015-10-01 Ville Voutilainen <ville.voutilainen@gmail.com>
Sebastian Pop [Thu, 1 Oct 2015 15:17:58 +0000 (15:17 +0000)]
call scev analysis in scop-detection as in sese-to-poly
Before our rewrite of the scop detection, we used to not have a valid SESE
region under hand, and so we used to do more ad-hoc analysis of data references
by trying to prove that at all levels of a loop nest the data references would
be still valid.
Now that we have a valid SESE region, we can call the scev analysis in the same
way on the same computed loop nest in the scop-detection as in the sese-to-poly.
Next step will be to cache the data references analyzed in the scop detection
and not compute the same info in sese-to-poly.
The patch fixes block-1.f90 that used to ICE on x86_64-linux when compiled with
-m32. Patch passed bootstrap with BOOT_CFLAGS="-g -O2 -fgraphite-identity
-floop-nest-optimize" and check on x86_64-linux using ISL-0.15.
2015-09-28 Sebastian Pop <s.pop@samsung.com>
Aditya Kumar <aditya.k7@samsung.com>
PR tree-optimization/67754
* graphite-scop-detection.c (stmt_has_simple_data_refs_p): Call
scev analysis on the same loop nest as analyze_drs_in_stmts.
* graphite-sese-to-poly.c (outermost_loop_in_sese_1): Moved and renamed...
(try_generate_gimple_bb): Call outermost_loop_in_sese.
(analyze_drs_in_stmts): Same.
* sese.c (outermost_loop_in_sese): ...here.
Sebastian Pop [Thu, 1 Oct 2015 15:17:51 +0000 (15:17 +0000)]
add recursion on the inner loops
We now check that all data references in the current loop and inner loops
contained within loop are valid in an outer region before declaring that the
outer loop is a valid scop.
2015-09-30 Sebastian Pop <s.pop@samsung.com>
Aditya Kumar <aditya.k7@samsung.com>
PR tree-optimization/67754
* graphite-scop-detection.c (loop_body_is_valid_scop): Add missing
recursion on the inner loops.
Jonathan Wakely [Thu, 1 Oct 2015 11:23:39 +0000 (12:23 +0100)]
Save-and-restore errno more carefully in libstdc++
* doc/xml/manual/diagnostics.xml: Document use of errno.
* doc/html/*: Regenerate.
* config/locale/generic/c_locale.cc (_Save_errno): New helper.
(__convert_to_v): Use _Save_errno.
* include/ext/string_conversions.h (__stoa): Only restore errno when
it isn't set to non-zero.
bb-reorder: Add -freorder-blocks-algorithm= and wire it up
This adds an -freorder-blocks-algorithm=[simple|stc] flag, with "simple"
as default. For -O2 and up (except -Os) it is switched to "stc" instead.
Targets that never want STC can override this. This changes -freorder-blocks
to be on at -O1 and up (was -O2 and up).
In effect, the changes are for -O1 (which now gets "simple" instead of
nothing), -Os (which now gets "simple" instead of "stc", since STC results
in much bigger code), and for targets that wish to never use STC (not in
this patch though).
* bb-reorder.c (reorder_basic_blocks): Use the algorithm selected
with flag_reorder_blocks_algorithm.
* common.opt (freorder-blocks-algorithm=): New flag.
(reorder_blocks_algorithm): New enum.
* flag-types.h (reorder_blocks_algorithm): New enum.
* opts.c (default_options_table): Use -freorder-blocks at -O1 and up,
and -freorder-blocks-algorithm=stc at -O2 and up (not at -Os).
* bb-reorder.c: Add intro comment.
(reorder_basic_blocks_software_trace_cache): Print a header to
the dump file.
(edge_order): New function.
(reorder_basic_blocks_simple): New function.
(reorder_basic_blocks): Choose between the STC and the simple
algorithms (always choose the former).
Tom de Vries [Thu, 1 Oct 2015 05:42:50 +0000 (05:42 +0000)]
Make dumping of function attributes resemble source syntax
2015-10-01 Tom de Vries <tom@codesourcery.com>
* tree-cfg.c (dump_function_to_file): Dump function attributes using
__attribute__(()) string. Move dumping of function attributes to before
function name.
re PR go/66870 (split stack issues on ppc64le and ppc64)
PR target/66870
* config/rs6000/sysv4.h (TARGET_CAN_SPLIT_STACK_64BIT): Define.
* configure.ac: Define HAVE_GOLD_ALTERNATE_SPLIT_STACK on Power
based on gold linker version.
* gcc.c: Add -fuse-ld=gold to STACK_SPLIT_SPEC if
HAVE_GOLD_ALTERNATE_SPLIT_STACK defined.
* configure, config.in: Regenerate.
go:
* gospec.c (lang_specific_driver): Set appropriate split stack
options for 64 bit compiles based on TARGET_CAN_SPLIT_STACK_64BIT.
ppc32 starts using the stack for integer arg passing when we run out
of integer arg passing registers. Similarly, we start using the stack
for floating point args when we run out of floating point registers.
The decision on where an integer arg goes does not depend on number of
floating point args, nor does the decision on where a floating point
arg goes depend on number of integer args. Alignment of stack args
also simply depends on number of stack args.
This patch untangles the horrible mess we had, with intarg_count being
wrongly used to count both integer args and stack words.
* src/powerpc/ffi_sysv.c (ffi_prep_cif_sysv_core): Count fprs,
gprs, and stack words separately.
(ffi_prep_args_SYSV): Similarly.
James Bowman [Tue, 29 Sep 2015 23:34:34 +0000 (23:34 +0000)]
Limit the MEMSET (setmemsi pattern) and MEMCPY (movmemsi pattern) instructions to 0-511 bytes.
Limit the MEMSET (setmemsi pattern) and MEMCPY (movmemsi pattern)
instructions to 0-511 bytes. There is a hardware limitation on large
MEMSET, MEMCPY operations that the library versions of memset() and
memcpy() deal with.
2015-09-29 James Bowman <james.bowman@ftdichip.com>
* config/ft32/predicates.md (ft32_imm_operand): New predicate.
* config/ft32/ft32.md (movmemsi, setmemsi): Use ft32_imm_operand
predicate, disallow register for operand 2.
Ian Lance Taylor [Tue, 29 Sep 2015 22:27:53 +0000 (22:27 +0000)]
compiler: Accept untyped integral values as string/array indices.
When determining the type of an index for a string/array indexing
expression, the gofrontend would disallow floating-point and complex
values even if they were integral and throw an internal error. This
patch changes gofrontend to use an integral type context when
determining the types of a string/array index.
No functional changes intended. Passes make check and bootstrap.
gcc/ChangeLog:
2015-09-29 Aditya Kumar <hiraditya@msn.com>
* graphite-dependences.c (scop_get_dependences): Moved in down
in order to be visible to its caller.
* graphite-poly.h: Removed compute_deps, and extend_schedule.
Sebastian Pop [Tue, 29 Sep 2015 16:33:45 +0000 (16:33 +0000)]
remove dead code of commutative_reductions
This code is not used anymore after we removed the previous loop optimizer (not
based on the ISL scheduler.) We will add back the detection of commutative
reductions after we improve the code generation of scalar dependences (by not
going out of SSA for scalar dependences just to expose them to the data
dependence graph.)
Patch passed bootstrap and check on x86_64-linux with ISL-0.15.
2015-09-29 Sebastian Pop <s.pop@samsung.com>
Aditya Kumar <aditya.k7@samsung.com>
libgomp-plugin-intelmic.cpp (OFFLOAD_ACTIVE_WAIT_ENV): New define.
liboffloadmic/
* plugin/libgomp-plugin-intelmic.cpp (OFFLOAD_ACTIVE_WAIT_ENV): New
define.
(init): Set OFFLOAD_ACTIVE_WAIT env var to 0, if it is not set.
* runtime/emulator/coi_common.h (PIPE_HOST_PATH): Replace with ...
(PIPE_HOST2TGT_NAME): ... this.
(PIPE_TARGET_PATH): Replace with ...
(PIPE_TGT2HOST_NAME): ... this.
(MALLOCN): New define.
(READN): Likewise.
(WRITEN): Likewise.
(enum cmd_t): Replace CMD_RUN_FUNCTION with CMD_PIPELINE_RUN_FUNCTION.
Add CMD_PIPELINE_CREATE, CMD_PIPELINE_DESTROY.
* runtime/emulator/coi_device.cpp (engine_dir): New static variable.
(pipeline_thread_routine): New static function.
(COIProcessWaitForShutdown): Use global engine_dir instead of mic_dir.
Rename pipe_host and pipe_target to pipe_host2tgt and pipe_tgt2host.
If cmd is CMD_PIPELINE_CREATE, create a new thread for the pipeline.
Remove cmd == CMD_RUN_FUNCTION case.
* runtime/emulator/coi_device.h (COIERRORN): New define.
* runtime/emulator/coi_host.cpp: Include set, map, queue.
Replace typedefs with enums and structs.
(struct Function): Remove name, add num_buffers, bufs_size,
bufs_data_target, misc_data_len, misc_data, return_value_len,
return_value, completion_event.
(struct Callback): New.
(struct Process): Remove pipeline. Add pipe_host2tgt and pipe_tgt2host.
(struct Pipeline): Remove pipe_host and pipe_target. Add thread,
destroy, is_destroyed, pipe_host2tgt_path, pipe_tgt2host_path,
pipe_host2tgt, pipe_tgt2host, queue, process.
(max_pipeline_num): New static variable.
(pipelines): Likewise.
(max_event_num): Likewise.
(non_signalled_events): Likewise.
(errored_events): Likewise.
(callbacks): Likewise.
(cleanup): Do not check tmp_dirs before free.
(start_critical_section): New static function.
(finish_critical_section): Likewise.
(pipeline_is_destroyed): Likewise.
(maybe_invoke_callback): Likewise.
(signal_event): Likewise.
(get_event_result): Likewise.
(COIBufferCopy): Rename arguments according to headers. Add asserts.
Use process' main pipes, instead of pipeline's pipes. Signal completion
event.
(COIBufferCreate): Rename arguments according to headers. Add asserts.
Use process' main pipes, instead of pipeline's pipes.
(COIBufferCreateFromMemory): Rename arguments according to headers.
Add asserts.
(COIBufferDestroy): Rename arguments according to headers. Add asserts.
Use process' main pipes, instead of pipeline's pipes.
(COIBufferGetSinkAddress): Rename arguments according to headers.
Add asserts.
(COIBufferMap): Rename arguments according to headers. Add asserts.
Signal completion event.
(COIBufferRead): Likewise.
(COIBufferSetState): Likewise.
(COIBufferUnmap): Likewise.
(COIBufferWrite): Likewise.
(COIEngineGetCount): Add assert.
(COIEngineGetHandle): Rename arguments according to headers.
Add assert.
(COIEventWait): Rename arguments according to headers. Add asserts.
Implement waiting for events with zero or infinite timeout.
(COIEventRegisterCallback): New function.
(pipeline_thread_routine): New static function.
(COIPipelineCreate): Create a new thread for the pipeline.
(COIPipelineDestroy): Exit pipeline thread.
(COIPipelineRunFunction): Add the function into pipeline's queue,
instead running it here. Wait for it's completion in case of
synchronous execution.
(COIProcessCreateFromMemory): Rename arguments according to headers.
Add asserts. Create process' main pipes, instead of pipeline's pipes.
(COIProcessDestroy): Rename arguments according to headers.
Add asserts. Destroy all undestroyed pipelines.
(COIProcessGetFunctionHandles): Rename arguments according to headers.
Add asserts. Use process' main pipes, instead of pipeline's pipes.
Remove useless function names.
(COIProcessLoadLibraryFromMemory): Add asserts. Use process' main
pipes, instead of pipeline's pipes.
(COIProcessUnloadLibrary): Likewise.
(COIEngineGetInfo): Add assert.
* runtime/emulator/coi_host.h (COIERRORN): New define.
Richard Biener [Tue, 29 Sep 2015 13:04:18 +0000 (13:04 +0000)]
re PR fortran/67170 (PRE can't hoist out a readonly argument)
2015-09-29 Richard Biener <rguenther@suse.de>
PR tree-optimization/67170
* tree-ssa-alias.h (get_continuation_for_phi): Adjust
the translate function pointer parameter to get the
bool whether to disambiguate only by reference.
(walk_non_aliased_vuses): Likewise.
* tree-ssa-alias.c (maybe_skip_until): Adjust.
(get_continuation_for_phi_1): Likewise.
(get_continuation_for_phi): Likewise.
(walk_non_aliased_vuses): Likewise.
* tree-ssa-sccvn.c (const_parms): New bitmap.
(vn_reference_lookup_3): Adjust for interface change.
Disambiguate parameters pointing to readonly memory.
(free_scc_vn): Free const_parms.
(run_scc_vn): Initialize const_parms from a fn spec attribute.
Kaz Kojima [Tue, 29 Sep 2015 05:36:01 +0000 (05:36 +0000)]
re PR target/67716 ([5] [SH]: Miscompiles libraw: Assembler: unaligned opcodes detected in executable segment)
PR target/67716
* [SH] Implement targetm.override_options_after_change hook
so to avoid resetting loop, jump and function alignment values with
function-wise optimization flags.
Double-quoted words in Tcl have substitutions performed on them, including
backslash substitutions. That isn't terribly nice for regular expressions,
so use braced words instead.
Joseph Myers [Mon, 28 Sep 2015 21:50:22 +0000 (22:50 +0100)]
Update soft-fp from glibc.
This patch updates the soft-fp code in libgcc from glibc. There are
no changes here of significance to the use of soft-fp in GCC (and so
no testsuite additions); it's simply an update to bring in the latest
soft-fp version (which will also hopefully go into Linux 4.4 to
replace the 15-year-old copy currently in Linux).
Bootstrapped with no regressions on x86_64-pc-linux-gnu.
Paul Thomas [Mon, 28 Sep 2015 21:18:38 +0000 (21:18 +0000)]
[multiple changes]
2015-09-28 Paul Thomas <pault@gcc.gnu.org>
PR fortran/40054
PR fortran/63921
* decl.c (get_proc_name): Return if statement function is
found.
* expr.c (gfc_check_vardef_context): Add error return for
derived type expression lacking the derived type itself.
* match.c (gfc_match_ptr_fcn_assign): New function.
* match.h : Add prototype for gfc_match_ptr_fcn_assign.
* parse.c : Add static flag 'in_specification_block'.
(decode_statement): If in specification block match a statement
function, then, if no error arising from statement function
matching, try to match pointer function assignment.
(parse_interface): Set 'in_specification_block' on exiting from
parse_spec.
(parse_spec): Set and then reset 'in_specification_block'.
(gfc_parse_file): Set 'in_specification_block'.
* resolve.c (get_temp_from_expr): Extend to include functions
and array constructors as rvalues..
(resolve_ptr_fcn_assign): New function.
(gfc_resolve_code): Call it on finding a pointer function as an
lvalue. If valid or on error, go back to start of resolve_code.
* symbol.c (gfc_add_procedure): Add a sentence to the error to
flag up the ambiguity between a statement function and pointer
function assignment at the end of the specification block.
2015-09-28 Paul Thomas <pault@gcc.gnu.org>
PR fortran/40054
PR fortran/63921
* gfortran.dg/fmt_tab_1.f90: Change from run to compile and set
standard as legacy.
* gfortran.dg/fmt_tab_2.f90: Add extra tab error.
* gfortran.dg/function_types_3.f90: Change error message to
"Type inaccessible...."
* gfortran.dg/ptr_func_assign_1.f08: New test.
* gfortran.dg/ptr_func_assign_2.f08: New test.
2015-09-25 Mikael Morin <mikael.morin@sfr.fr>
PR fortran/40054
PR fortran/63921
* gfortran.dg/ptr_func_assign_3.f08: New test.
* gfortran.dg/ptr_func_assign_4.f08: New test.
libgomp/
* libgomp.h (acc_dispatch_t): Replace separate geometry args with
array.
* libgomp.map (GOACC_parallel_keyed): New.
* oacc-parallel.c (goacc_wait): Take pointer to va_list. Adjust
all callers.
(GOACC_parallel_keyed): New interface. Lose geometry arguments
and take keyed varargs list. Adjust call to exec_func.
(GOACC_parallel): Force host fallback.
* libgomp_g.h (GOACC_parallel): Remove.
(GOACC_parallel_keyed): Declare.
* plugin/plugin-nvptx.c (struct targ_fn_launch): New struct.
(stuct targ_gn_descriptor): Replace name field with launch field.
(nvptx_exec): Lose separate geometry args, take array. Process
dynamic dimensions and adjust.
(struct nvptx_tdata): Replace fn_names field with fn_descs.
(GOMP_OFFLOAD_load_image): Adjust for change in function table
data.
(GOMP_OFFLOAD_openacc_parallel): Adjust for change in dimension
passing.
* oacc-host.c (host_openacc_exec): Adjust for change in dimension
passing.
gcc/
* config/nvptx/nvptx.c: Include omp-low.h and gomp-constants.h.
(nvptx_record_offload_symbol): Record function execution geometry.
* config/nvptx/mkoffload.c (process): Include launch geometry in
function data.
* omp-low.c (oacc_launch_pack): New.
(replace_oacc_fn_attrib): New.
(set_oacc_fn_attrib): New.
(get_oacc_fn_attrib): New.
(expand_omp_target): Create keyed varargs for GOACC_parallel call
generation.
* omp-low.h (get_oacc_fn_attrib): Declare.
* builtin-types.def (DEF_FUNCTION_TyPE_VAR_6): New.
(DEF_FUNCTION_TYPE_VAR_11): Delete.
* tree.h (OMP_CLAUSE_EXPR): New.
* omp-builtins.def (BUILT_IN_GOACC_PARALLEL): Change target fn name.
Jeff Law [Mon, 28 Sep 2015 19:27:29 +0000 (13:27 -0600)]
Re: [PATCH] Fix undefined behaviour in arc port
Re: [PATCH] Fix undefined behaviour in arc port
* config/arc/arc.c (arc_legitimize_reload_address): Fix stupid
thinko in last change.
* config/arc/constraints.md (C2a): Fix typos in last change.
Sebastian Pop [Mon, 28 Sep 2015 17:30:09 +0000 (17:30 +0000)]
Redesign Graphite scop detection
Redesign Graphite scop detection for faster compiler time and detecting more SCoPs.
Existing algorithm for SCoP detection in graphite was based on dominator tree
where a tree (CFG) traversal was required for analyzing an SESE. The tree
traversal is linear in the number of basic blocks and SCoP detection is
(probably) linear in number of instructions. That algorithm utilized a generic
infrastructure of SESE which does not directly represent loops. With regards to
graphite framework, we are only interested in subtrees with loops. The new
algorithm is geared towards tree traversal on loop structure. The algorithm is
linear in number of loops which is faster than the previous algorithm.
Briefly, we start the traversal at a loop-nest and analyze it recursively for
validity. Once a valid loop is found we find a valid adjacent loop. If an
adjacent loop is found and is valid, we merge both loop nests otherwise we form
a SCoP from the previous loop nest, and resume the algorithm from the adjacent
loop nest. The data structure to represent an SESE is an ordered pair of edges
(entry, exit). The new algoritm can extend a SCoP in both the directions. With
this approach, the number of instructions to be analyzed for validity reduces to
a minimal set. We start by analyzing those statements which are inside a loop,
because validity of those statements is necessary for the validity of loop. The
statements outside the loop nest can be just excluded from the SESE if they are
not valid.
This patch depends on: https://gcc.gnu.org/ml/gcc-patches/2015-09/msg02024.html
Passes (c,c++,fortran) regtest and bootstrap.
gcc/ChangeLog:
2015-09-27 Aditya Kumar <hiraditya@msn.com>
Sebastian Pop <s.pop@samsung.com>
* graphite-optimize-isl.c (optimize_isl):
* graphite-scop-detection.c (struct sese_l): New type.
(get_entry_bb): API for getting entry bb of SESE.
(get_exit_bb): API for getting exit bb of SESE.
(class debug_printer): New type. Simple printer in debug mode.
(trivially_empty_bb_p): New. Return true when BB is empty or
contains only debug instructions.
(graphite_can_represent_expr): Call scalar_evoution_in_region
instead of analyze_scalar_evolution. Pass in scop instead of only
the scop entry.
(stmt_has_simple_data_refs_p): Pass in scop instead of only the
scop entry.
(stmt_simple_for_scop_p): Same.
(harmful_stmt_in_bb): Same.
(graphite_can_represent_loop): Deleted.
(struct scopdet_info): Deleted.
(scopdet_basic_block_info): Deleted.
(build_scops_1): Deleted.
(bb_in_sd_region): Deleted.
(find_single_entry_edge): Deleted.
(find_single_exit_edge): Deleted.
(create_single_entry_edge): Deleted.
(sd_region_without_exit): Deleted.
(create_single_exit_edge): Deleted.
(unmark_exit_edges): Deleted.
(mark_exit_edges): Deleted.
(create_sese_edges): Deleted.
(build_graphite_scops): Deleted.
(canonicalize_loop_closed_ssa): Recompute all dominators at the
end.
(build_scops): Use the new scop_builder to build scops.
(dot_all_scops_1): Use the new pretty printer. Print loop father
as well.
(loop_body_is_valid_scop): New. Return true if loop body is a
valid scop.
(class scop_builder): New. Builds SCoPs for polyhedral
optimizatios.
(scop_builder): New. Constructor.
(static sese_l invalid_sese): sese_l with invalid edges.
(get_sese): Get an sese (from a loop) if possible, invalid_sese
otherwise.
(get_nearest_dom_with_single_entry): Get nearest dominator of a
basic_block with single entry. Return NULL if we get to the
beginning of a function.
(get_nearest_pdom_with_single_exit): Get nearest post-dominator of
a basic_block with single exit. Return NULL if we get to the
beginning of a function.
(print_sese): Pretty-print SESE.
(merge_sese): Merge two SESEs if possible and return the new SESE.
(build_scop_depth): Start building the SCoP within a loop nest.
(build_scop_breadth): Start building the SCoP at a single loop
depth. Merge adjacent SESEs if valid.
(can_represent_loop_1): Returns true if Graphite can represent
loop inside SCoP. Helper for can_represent_loop.
(can_represent_loop): Returns true if Graphite can represent LOOP
and all its nested loops in SCoP.
(loop_is_valid_scop): Returns true if LOOP and all its nests
constitute a valid SCoP.
(region_has_one_loop): Returns true of a region has only one loop.
(add_scop): Add SCoP to the list of valid scops. Removes an
already existing scop if it intersects with or subsumed by this
one.
(harmful_stmt_in_region): Returns true if SCoP has any statment
which cannot be represented by Graphite.
(subsumes): Returns true of SCoP S1 subsumes SCoP S2.
(remove_subscops): Remove any SCoP from the list of already found
SCoPs, if subsumed by S1.
(intersects): Return true if region bounded by SCoPs S1 and S2
intersect.
(remove_intersecting_scops): Remove any SCoP which intersects with
S1.
* graphite.c (print_graphite_scop_statistics):
(print_graphite_statistics): Print SCoP info while debugging.
(graphite_initialize): Early exit in case number of loops in a
function is less than PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION or
basic blocks are more than PARAM_GRAPHITE_MAX_BBS_PER_FUNCTION.
(graphite_finalize):
* params.def: Add PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION.
* sese.h (sese_loop_depth): Remove unnecessary gcc_assert.
(recompute_all_dominators): Recalculate POST_DOMINATORS.
* tree-cfg.c (print_loops): Print the function name while printing
loops.
re PR tree-optimization/67700 ([graphite] miscompile due to wrong codegen)
fix PR67700
The patch makes the detection of scop parameters in parameter_index_in_region a
bit more conservative by discarding scalar variables defined in function of data
references defined in the scop.
2015-09-25 Aditya Kumar <aditya.k7@samsung.com>
Sebastian Pop <s.pop@samsung.com>
projects / thirdparty / gcc.git / log
