| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The MOVC5 machine instruction has `memset' semantics if encoded with a
zero source length[1]:
"4. MOVC5 with a zero source length operand is the preferred way
to fill a block of memory with the fill character."
Use that instruction to implement the `setmemhi' instruction then. Use
the AP register in the register deferred mode for the source address to
yield the shortest possible encoding of the otherwise unused operand,
observing that the address is never dereferenced if the source length is
zero.
The use of this instruction yields steadily better performance, at least
with the Mariah VAX implementation, for a variable-length `memset' call
expanded inline as a single MOVC5 operation compared to an equivalent
libcall invocation:
Length: 1, time elapsed: 0.971789 (builtin), 2.847303 (libcall)
Length: 2, time elapsed: 0.907904 (builtin), 2.728259 (libcall)
Length: 3, time elapsed: 1.038311 (builtin), 2.917245 (libcall)
Length: 4, time elapsed: 0.775305 (builtin), 2.686088 (libcall)
Length: 7, time elapsed: 1.112331 (builtin), 2.992968 (libcall)
Length: 8, time elapsed: 0.856882 (builtin), 2.764885 (libcall)
Length: 15, time elapsed: 1.256086 (builtin), 3.096660 (libcall)
Length: 16, time elapsed: 1.001962 (builtin), 2.888131 (libcall)
Length: 31, time elapsed: 1.590456 (builtin), 3.774164 (libcall)
Length: 32, time elapsed: 1.288909 (builtin), 3.629622 (libcall)
Length: 63, time elapsed: 3.430285 (builtin), 5.269789 (libcall)
Length: 64, time elapsed: 3.265147 (builtin), 5.113156 (libcall)
Length: 127, time elapsed: 6.438772 (builtin), 8.268305 (libcall)
Length: 128, time elapsed: 6.268991 (builtin), 8.114557 (libcall)
Length: 255, time elapsed: 12.417338 (builtin), 14.259678 (libcall)
(times given in seconds per 1000000 `memset' invocations for the given
length made in a loop). It is clear from these figures that hardware
does data coalescence for consecutive bytes rather than naively copying
them one by one, as for lengths that are powers of 2 the figures are
consistently lower than ones for their respective next lower lengths.
The use of MOVC5 also requires at least 4 bytes less in terms of machine
code as it avoids encoding the address of `memset' needed for the CALLS
instruction used to make a libcall, as well as extra PUSHL instructions
needed to pass arguments to the call as those can be encoded directly as
the respective operands of the MOVC5 instruction.
It is perhaps worth noting too that for constant lengths we prefer to
emit up to 5 individual MOVx instructions rather than a single MOVC5
instruction to clear memory and for consistency we copy this behavior
here for filling memory with another value too, even though there may be
a performance advantage with a string copy in comparison to a piecemeal
copy, e.g.:
Length: 40, time elapsed: 2.183192 (string), 2.638878 (piecemeal)
But this is something for another change as it will have to be carefully
evaluated.
[1] DEC STD 032-0 "VAX Architecture Standard", Digital Equipment
Corporation, A-DS-EL-00032-00-0 Rev J, December 15, 1989, Section
3.10 "Character-String Instructions", p. 3-163
gcc/
* config/vax/vax.h (SET_RATIO): New macro.
* config/vax/vax.md (UNSPEC_SETMEM_FILL): New constant.
(setmemhi): New expander.
(setmemhi1): New insn and splitter.
(*setmemhi1): New insn.
gcc/testsuite/
* gcc.target/vax/setmem.c: New test.
|
|
This patch adds kill tracking to ipa-modref. This is representd by array
of accesses to memory locations that are known to be overwritten by the
function.
gcc/ChangeLog:
2021-11-14 Jan Hubicka <hubicka@ucw.cz>
* ipa-modref-tree.c (modref_access_node::update_for_kills): New
member function.
(modref_access_node::merge_for_kills): Likewise.
(modref_access_node::insert_kill): Likewise.
* ipa-modref-tree.h (modref_access_node::update_for_kills,
modref_access_node::merge_for_kills, modref_access_node::insert_kill):
Declare.
(modref_access_node::useful_for_kill): New member function.
* ipa-modref.c (modref_summary::useful_p): Release useless kills.
(lto_modref_summary): Add kills.
(modref_summary::dump): Dump kills.
(record_access): Add mdoref_access_node parameter.
(record_access_lto): Likewise.
(merge_call_side_effects): Merge kills.
(analyze_call): Add ALWAYS_EXECUTED param and pass it around.
(struct summary_ptrs): Add always_executed filed.
(analyze_load): Update.
(analyze_store): Update; record kills.
(analyze_stmt): Add always_executed; record kills in clobbers.
(analyze_function): Track always_executed.
(modref_summaries::duplicate): Duplicate kills.
(update_signature): Release kills.
* ipa-modref.h (struct modref_summary): Add kills.
* tree-ssa-alias.c (alias_stats): Add kill stats.
(dump_alias_stats): Dump kill stats.
(store_kills_ref_p): Break out from ...
(stmt_kills_ref_p): Use it; handle modref info based kills.
gcc/testsuite/ChangeLog:
2021-11-14 Jan Hubicka <hubicka@ucw.cz>
* gcc.dg/tree-ssa/modref-dse-3.c: New test.
|
|
gcc/testsuite/ChangeLog:
* gcc.dg/pr103229.c: Removed.
|
|
The code computing ranges in PHIs in the path solver reuses the
temporary ssa_global_cache by calling its clear method. Calling it on
an empty cache causes us to call memset with NULL.
Tested on x86-64 Linux.
gcc/ChangeLog:
PR tree-optimization/103229
* gimple-range-cache.cc (ssa_global_cache::clear): Do not pass
null value to memset.
gcc/testsuite/ChangeLog:
* gcc.dg/pr103229.c: New test.
|
|
gcc/testsuite/ChangeLog:
* c-c++-common/tsan/free_race.c: Remove unnecessary -ldl.
* c-c++-common/tsan/free_race2.c: Likewise.
|
|
Move code getting tree op from access_node and stmt to a common place. I also
commonized logic to build ao_ref. While I was on it I also replaced FOR_EACH_*
by range for since they reads better.
gcc/ChangeLog:
2021-11-14 Jan Hubicka <hubicka@ucw.cz>
* ipa-modref-tree.c (modref_access_node::get_call_arg): New member
function.
(modref_access_node::get_ao_ref): Likewise.
* ipa-modref-tree.h (modref_access_node::get_call_arg): Declare.
(modref_access_node::get_ao_ref): Declare.
* tree-ssa-alias.c (modref_may_conflict): Use new accessors.
* tree-ssa-dse.c (dse_optimize_call): Use new accessors.
gcc/testsuite/ChangeLog:
2021-11-14 Jan Hubicka <hubicka@ucw.cz>
* c-c++-common/asan/null-deref-1.c: Update template.
* c-c++-common/tsan/free_race.c: Update template.
* c-c++-common/tsan/free_race2.c: Update template.
* gcc.dg/ipa/ipa-sra-4.c: Update template.
|
|
|
|
gcc/ChangeLog:
PR lto/103211
* dbgcnt.def (ipa_attr): New counters.
* ipa-pure-const.c: Include dbgcnt.c
(ipa_make_function_const): Use debug counter.
(ipa_make_function_pure): Likewise.
(propagate_pure_const): Fix bug in my previous change.
|
|
Move access dumping to member function and cleanup formating.
gcc/ChangeLog:
2021-11-13 Jan Hubicka <hubicka@ucw.cz>
* ipa-modref-tree.c (modref_access_node::range_info_useful_p):
Offline from ipa-modref-tree.h.
(modref_access_node::dump): Move from ipa-modref.c; make member
function.
* ipa-modref-tree.h (modref_access_node::range_info_useful_p.
modref_access_node::dump): Declare.
* ipa-modref.c (dump_access): Remove.
(dump_records): Update.
(dump_lto_records): Update.
(record_access): Update.
(record_access_lto): Update.
|
|
gcc/ChangeLog:
2021-11-13 Jan Hubicka <hubicka@ucw.cz>
* ipa-modref.c (modref_summary::modref_summary): Clear new flags.
(modref_summary::dump): Dump try_dse.
(modref_summary::finalize): Add FUN attribute; compute try-dse.
(analyze_function): Update.
(read_section): Update.
(update_signature): Update.
(pass_ipa_modref::execute): Update.
* ipa-modref.h (struct modref_summary):
* tree-ssa-alias.c (ao_ref_init_from_ptr_and_range): Export.
* tree-ssa-alias.h (ao_ref_init_from_ptr_and_range): Declare.
* tree-ssa-dse.c (dse_optimize_call): New function.
(dse_optimize_stmt): Use it.
gcc/testsuite/ChangeLog:
2021-11-13 Jan Hubicka <hubicka@ucw.cz>
* g++.dg/cpp1z/inh-ctor23.C: Fix template
* g++.dg/ipa/ipa-icf-4.C: Fix template
* gcc.dg/tree-ssa/modref-dse-1.c: New test.
* gcc.dg/tree-ssa/modref-dse-2.c: New test.
|
|
gcc/ChangeLog:
2021-11-13 Jan Hubicka <hubicka@ucw.cz>
* ipa-modref-tree.c: Move #if CHECKING_P to proper place.
|
|
move member functions of modref_access_node from ipa-modref-tree.h to
ipa-modref-tree.c since they become long and not fitting for inlines anyway. I
also cleaned up the interface by making static insert method (which handles
inserting accesses into a vector and optimizing them) which makes it possible
to hide most of the interface handling interval merging private.
Honza
gcc/ChangeLog:
* ipa-modref-tree.h
(struct modref_access_node): Move longer member functions to
ipa-modref-tree.c
(modref_ref_node::try_merge_with): Turn into modreef_acces_node member
function.
* ipa-modref-tree.c (modref_access_node::contains): Move here
from ipa-modref-tree.h.
(modref_access_node::update): Likewise.
(modref_access_node::merge): Likewise.
(modref_access_node::closer_pair_p): Likewise.
(modref_access_node::forced_merge): Likewise.
(modref_access_node::update2): Likewise.
(modref_access_node::combined_offsets): Likewise.
(modref_access_node::try_merge_with): Likewise.
(modref_access_node::insert): Likewise.
|
|
gcc/ChangeLog:
* ipa-modref.c (modref_summary::global_memory_read_p): Remove.
(modref_summary::global_memory_written_p): Remove.
(modref_summary::dump): Dump new flags.
(modref_summary::finalize): New member function.
(analyze_function): Call it.
(read_section): Call it.
(update_signature): Call it.
(pass_ipa_modref::execute): Call it.
* ipa-modref.h (struct modref_summary): Remove
global_memory_read_p and global_memory_written_p.
Add global_memory_read, global_memory_written.
* tree-ssa-structalias.c (determine_global_memory_access):
Update.
|
|
gcc/ChangeLog:
* ipa-fnsummary.c (compute_fn_summary): Use type_attribut_allowed_p
* ipa-param-manipulation.c
(ipa_param_adjustments::type_attribute_allowed_p):
New member function.
(drop_type_attribute_if_params_changed_p): New function.
(build_adjusted_function_type): Use it.
* ipa-param-manipulation.h: Add type_attribute_allowed_p.
|
|
The initial commit of the analyzer in GCC 10 had a single warning,
-Wanalyzer-tainted-array-index
and required manually enabling the taint checker with
-fanalyzer-checker=taint (due to scaling issues).
This patch extends the taint detection to add four new taint-based
warnings:
-Wanalyzer-tainted-allocation-size
for e.g. attacker-controlled malloc/alloca
-Wanalyzer-tainted-divisor
for detecting where an attacker can inject a divide-by-zero
-Wanalyzer-tainted-offset
for attacker-controlled pointer offsets
-Wanalyzer-tainted-size
for e.g. attacker-controlled memset
and rewords all the warnings to talk about "attacker-controlled" values
rather than "tainted" values.
Unfortunately I haven't yet addressed the scaling issues, so all of
these still require -fanalyzer-checker=taint (in addition to -fanalyzer).
gcc/analyzer/ChangeLog:
* analyzer.opt (Wanalyzer-tainted-allocation-size): New.
(Wanalyzer-tainted-divisor): New.
(Wanalyzer-tainted-offset): New.
(Wanalyzer-tainted-size): New.
* engine.cc (impl_region_model_context::get_taint_map): New.
* exploded-graph.h (impl_region_model_context::get_taint_map):
New decl.
* program-state.cc (sm_state_map::get_state): Call
alt_get_inherited_state.
(sm_state_map::impl_set_state): Modify states within
compound svalues.
(program_state::impl_call_analyzer_dump_state): Undo casts.
(selftest::test_program_state_1): Update for new context param of
create_region_for_heap_alloc.
(selftest::test_program_state_merging): Likewise.
* region-model-impl-calls.cc (region_model::impl_call_alloca):
Likewise.
(region_model::impl_call_calloc): Likewise.
(region_model::impl_call_malloc): Likewise.
(region_model::impl_call_operator_new): Likewise.
(region_model::impl_call_realloc): Likewise.
* region-model.cc (region_model::check_region_access): Call
check_region_for_taint.
(region_model::get_representative_path_var_1): Handle binops.
(region_model::create_region_for_heap_alloc): Add "ctxt" param and
pass it to set_dynamic_extents.
(region_model::create_region_for_alloca): Likewise.
(region_model::set_dynamic_extents): Add "ctxt" param and use it
to call check_dynamic_size_for_taint.
(selftest::test_state_merging): Update for new context param of
create_region_for_heap_alloc.
(selftest::test_malloc_constraints): Likewise.
(selftest::test_malloc): Likewise.
(selftest::test_alloca): Likewise for create_region_for_alloca.
* region-model.h (region_model::create_region_for_heap_alloc): Add
"ctxt" param.
(region_model::create_region_for_alloca): Likewise.
(region_model::set_dynamic_extents): Likewise.
(region_model::check_dynamic_size_for_taint): New decl.
(region_model::check_region_for_taint): New decl.
(region_model_context::get_taint_map): New vfunc.
(noop_region_model_context::get_taint_map): New.
* sm-taint.cc: Remove include of "diagnostic-event-id.h"; add
includes of "gimple-iterator.h", "tristate.h", "selftest.h",
"ordered-hash-map.h", "cgraph.h", "cfg.h", "digraph.h",
"analyzer/supergraph.h", "analyzer/call-string.h",
"analyzer/program-point.h", "analyzer/store.h",
"analyzer/region-model.h", and "analyzer/program-state.h".
(enum bounds): Move to top of file.
(class taint_diagnostic): New.
(class tainted_array_index): Convert to subclass of taint_diagnostic.
(tainted_array_index::emit): Add CWE-129. Reword warning to use
"attacker-controlled" rather than "tainted".
(tainted_array_index::describe_state_change): Move to
taint_diagnostic::describe_state_change.
(tainted_array_index::describe_final_event): Reword to use
"attacker-controlled" rather than "tainted".
(class tainted_offset): New.
(class tainted_size): New.
(class tainted_divisor): New.
(class tainted_allocation_size): New.
(taint_state_machine::alt_get_inherited_state): New.
(taint_state_machine::on_stmt): In assignment handling, remove
ARRAY_REF handling in favor of check_region_for_taint. Add
detection of tainted divisors.
(taint_state_machine::get_taint): New.
(taint_state_machine::combine_states): New.
(region_model::check_region_for_taint): New.
(region_model::check_dynamic_size_for_taint): New.
* sm.h (state_machine::alt_get_inherited_state): New.
gcc/ChangeLog:
* doc/invoke.texi (Static Analyzer Options): Add
-Wno-analyzer-tainted-allocation-size,
-Wno-analyzer-tainted-divisor, -Wno-analyzer-tainted-offset, and
-Wno-analyzer-tainted-size to list. Add
-Wanalyzer-tainted-allocation-size, -Wanalyzer-tainted-divisor,
-Wanalyzer-tainted-offset, and -Wanalyzer-tainted-size to list
of options effectively enabled by -fanalyzer.
(-Wanalyzer-tainted-allocation-size): New.
(-Wanalyzer-tainted-array-index): Tweak wording; add link to CWE.
(-Wanalyzer-tainted-divisor): New.
(-Wanalyzer-tainted-offset): New.
(-Wanalyzer-tainted-size): New.
gcc/testsuite/ChangeLog:
* gcc.dg/analyzer/pr93382.c: Tweak expected wording.
* gcc.dg/analyzer/taint-alloc-1.c: New test.
* gcc.dg/analyzer/taint-alloc-2.c: New test.
* gcc.dg/analyzer/taint-divisor-1.c: New test.
* gcc.dg/analyzer/taint-1.c: Rename to...
* gcc.dg/analyzer/taint-read-index-1.c: ...this. Tweak expected
wording. Mark some events as xfail.
* gcc.dg/analyzer/taint-read-offset-1.c: New test.
* gcc.dg/analyzer/taint-size-1.c: New test.
* gcc.dg/analyzer/taint-write-index-1.c: New test.
* gcc.dg/analyzer/taint-write-offset-1.c: New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
gcc/ChangeLog:
* attr-fnspec.h (attr_fnspec::arg_eaf_flags): Break out from ...
* gimple.c (gimple_call_arg_flags): ... here.
* ipa-modref.c (analyze_parms): Record flags known from fnspec.
(modref_merge_call_site_flags): Use arg_eaf_flags.
|
|
PHIs must be resolved simulatenously, otherwise we may not pick up the
ranges incoming to the block.
For example. If we put p3_7 in the cache before all PHIs have been
computed, we will pick up the wrong p3_7 value for p2_17:
# p3_7 = PHI <1(2), 0(5)>
# p2_17 = PHI <1(2), p3_7(5)>
This patch delays updating the cache until all PHIs have been
analyzed.
gcc/ChangeLog:
PR tree-optimization/103222
* gimple-range-path.cc (path_range_query::compute_ranges_in_phis):
New.
(path_range_query::compute_ranges_in_block): Call
compute_ranges_in_phis.
* gimple-range-path.h (path_range_query::compute_ranges_in_phis):
New.
gcc/testsuite/ChangeLog:
* gcc.dg/pr103222.c: New test.
|
|
Enable some ipa-sra on fortran by allowing signature changes on functions
with "fn spec" attribute when ipa-modref is enabled. This is possible since ipa-modref
knows how to preserve things we trace in fnspec and fnspec generated by fortran forntend
are quite simple and can be analysed automatically now. To be sure I will also add
code that merge fnspec to parameters.
This unfortunately hits bug in ipa-param-manipulation when we remove parameter
that specifies size of variable length parameter. For this reason I added a hack
that prevent signature changes on such functions and will handle it incrementally.
I tried creating C testcase but it is blocked by another problem that we punt ipa-sra
on access attribute. This is optimization regression we ought to fix so I filled
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103223.
As a followup I will add code classifying the type attributes (we have just few) and
get stats on access attribute.
gcc/ChangeLog:
* ipa-fnsummary.c (compute_fn_summary): Do not give up on signature
changes on "fn spec" attribute; give up on varadic types.
* ipa-param-manipulation.c: Include attribs.h.
(build_adjusted_function_type): New parameter ARG_MODIFIED; if it is
true remove "fn spec" attribute.
(ipa_param_adjustments::build_new_function_type): Update.
(ipa_param_body_adjustments::modify_formal_parameters): update.
* ipa-sra.c: Include attribs.h.
(ipa_sra_preliminary_function_checks): Do not check for TYPE_ATTRIBUTES.
|
|
There's no need for two constructors, when we can do it all with one
that defaults to the common behavior:
path_range_query (bool resolve = true, gimple_ranger *ranger = NULL);
Tested on x86-64 Linux.
gcc/ChangeLog:
* gimple-range-path.cc (path_range_query::path_range_query): Merge
ctors.
(path_range_query::import_p): Move from header file.
(path_range_query::~path_range_query): Adjust for combined ctors.
* gimple-range-path.h: Merge ctors.
(path_range_query::import_p): Move to .cc file.
|
|
ipa-modref gets confused by EAF flags of memcpy becuase parameter 1 is
escaping but used only directly. In modref we do not track values saved to
memory and thus we clear all other flags on each store. This needs to also
happen when called function escapes parameter.
gcc/ChangeLog:
PR tree-optimization/103182
* ipa-modref.c (callee_to_caller_flags): Fix merging of flags.
(modref_eaf_analysis::analyze_ssa_name): Fix merging of flags.
|
|
|
|
Recently we changed the PROFILE_HOOK _mcount call to pass in the link
register as an argument. This actually does not work when the _mcount
call uses a PLT because the GOT register setup code ends up getting
inserted before the PROFILE_HOOK and clobbers the link register
argument.
These glibc tests are failing:
gmon/tst-gmon-pie-gprof
gmon/tst-gmon-static-gprof
This patch fixes this by saving the instruction that stores the Link
Register to the _mcount argument and then inserts the GOT register setup
instructions after that.
For example:
main.c:
extern int e;
int f2(int a) {
return a + e;
}
int f1(int a) {
return f2 (a + a);
}
int main(int argc, char ** argv) {
return f1 (argc);
}
Compiled:
or1k-smh-linux-gnu-gcc -Wall -c -O2 -fPIC -pg -S main.c
Before Fix:
main:
l.addi r1, r1, -16
l.sw 8(r1), r2
l.sw 0(r1), r16
l.addi r2, r1, 16 # Keeping FP, but not needed
l.sw 4(r1), r18
l.sw 12(r1), r9
l.jal 8 # GOT Setup clobbers r9 (Link Register)
l.movhi r16, gotpchi(_GLOBAL_OFFSET_TABLE_-4)
l.ori r16, r16, gotpclo(_GLOBAL_OFFSET_TABLE_+0)
l.add r16, r16, r9
l.or r18, r3, r3
l.or r3, r9, r9 # This is not the original LR
l.jal plt(_mcount)
l.nop
l.jal plt(f1)
l.or r3, r18, r18
l.lwz r9, 12(r1)
l.lwz r16, 0(r1)
l.lwz r18, 4(r1)
l.lwz r2, 8(r1)
l.jr r9
l.addi r1, r1, 16
After the fix:
main:
l.addi r1, r1, -12
l.sw 0(r1), r16
l.sw 4(r1), r18
l.sw 8(r1), r9
l.or r18, r3, r3
l.or r3, r9, r9 # We now have r9 (LR) set early
l.jal 8 # Clobbers r9 (Link Register)
l.movhi r16, gotpchi(_GLOBAL_OFFSET_TABLE_-4)
l.ori r16, r16, gotpclo(_GLOBAL_OFFSET_TABLE_+0)
l.add r16, r16, r9
l.jal plt(_mcount)
l.nop
l.jal plt(f1)
l.or r3, r18, r18
l.lwz r9, 8(r1)
l.lwz r16, 0(r1)
l.lwz r18, 4(r1)
l.jr r9
l.addi r1, r1, 12
Fixes: 308531d148a ("or1k: Add return address argument to _mcount call")
gcc/ChangeLog:
* config/or1k/or1k-protos.h (or1k_profile_hook): New function.
* config/or1k/or1k.h (PROFILE_HOOK): Change macro to reference
new function or1k_profile_hook.
* config/or1k/or1k.c (struct machine_function): Add new field
set_mcount_arg_insn.
(or1k_profile_hook): New function.
(or1k_init_pic_reg): Update to inject pic rtx after _mcount arg
when profiling.
(or1k_frame_pointer_required): Frame pointer no longer needed
when profiling.
|
|
I introduced bug into find_func_aliases_for_call in handling pure functions.
Instead of reading global memory pure functions are believed to write global
memory. This results in misoptimization of the testcase at -O1.
The change to pta-callused.c updates the template for new behaviour of the
constraint generation. We copy nonlocal memory to calluse which is correct but
also not strictly necessary because later we take care to add nonlocal_p flag
manually.
gcc/ChangeLog:
PR tree-optimization/103209
* tree-ssa-structalias.c (find_func_aliases_for_call): Fix
use of handle_rhs_call
gcc/testsuite/ChangeLog:
PR tree-optimization/103209
* gcc.dg/tree-ssa/pta-callused.c: Update template.
* gcc.c-torture/execute/pr103209.c: New test.
|
|
PHIs must be resolved first while solving ranges in a block,
regardless of where they appear in the import bitmap. We went through
a similar exercise for the relational code, but missed these.
Tested on x86-64 & ppc64le Linux.
gcc/ChangeLog:
PR tree-optimization/103202
* gimple-range-path.cc
(path_range_query::compute_ranges_in_block): Solve PHI imports first.
|
|
gcc/ChangeLog:
* ipa-pure-const.c (propagate_pure_const): Remove redundant check;
fix call of ipa_make_function_const and ipa_make_function_pure.
|
|
gcc/analyzer/ChangeLog:
* engine.cc (exploded_node::on_stmt_pre): Return when handling
"__analyzer_dump_state".
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
Previous patches made some of the complex parts of the issue rate
code redundant.
gcc/
* config/aarch64/aarch64.c (aarch64_vector_op::n_advsimd_ops): Delete.
(aarch64_vector_op::m_seen_loads): Likewise.
(aarch64_vector_costs::aarch64_vector_costs): Don't push to
m_advsimd_ops.
(aarch64_vector_op::count_ops): Remove vectype and factor parameters.
Remove code that tries to predict different vec_flags from the
current loop's.
(aarch64_vector_costs::add_stmt_cost): Update accordingly.
Remove m_advsimd_ops handling.
|
|
Previously we tried to account for the different issue rates of
the various vector modes by guessing what the Advanced SIMD version
of an SVE loop would look like and what its issue rate was likely to be.
We'd then increase the cost of the SVE loop if the Advanced SIMD loop
might issue more quickly.
This patch moves that logic to better_main_loop_than_p, so that we
can compare loops side-by-side rather than having to guess. This also
means we can apply the issue rate heuristics to *any* vector loop
comparison, rather than just weighting SVE vs. Advanced SIMD.
The actual heuristics are otherwise unchanged. We're just
applying them in a different place.
gcc/
* config/aarch64/aarch64.c (aarch64_vector_costs::m_saw_sve_only_op)
(aarch64_sve_only_stmt_p): Delete.
(aarch64_vector_costs::prefer_unrolled_loop): New function,
extracted from adjust_body_cost.
(aarch64_vector_costs::better_main_loop_than_p): New function,
using heuristics extracted from adjust_body_cost and
adjust_body_cost_sve.
(aarch64_vector_costs::adjust_body_cost_sve): Remove
advsimd_cycles_per_iter and could_use_advsimd parameters.
Update after changes above.
(aarch64_vector_costs::adjust_body_cost): Update after changes above.
|
|
-mtune=neoverse-512tvb sets the likely SVE vector length to 128 bits,
but it also takes into account Neoverse V1, which is a 256-bit target.
This patch adds this VF (VL) factor to aarch64_vec_op_count.
gcc/
* config/aarch64/aarch64.c (aarch64_vec_op_count::m_vf_factor):
New member variable.
(aarch64_vec_op_count::aarch64_vec_op_count): Add a parameter for it.
(aarch64_vec_op_count::vf_factor): New function.
(aarch64_vector_costs::aarch64_vector_costs): When costing for
neoverse-512tvb, pass a vf_factor of 2 for the Neoverse V1 version
of an SVE loop.
(aarch64_vector_costs::adjust_body_cost): Read the vf factor
instead of hard-coding 2.
|
|
This patch just moves the main cycle estimation routines
into aarch64_vec_op_count.
gcc/
* config/aarch64/aarch64.c
(aarch64_vec_op_count::rename_cycles_per_iter): New function.
(aarch64_vec_op_count::min_nonpred_cycles_per_iter): Likewise.
(aarch64_vec_op_count::min_pred_cycles_per_iter): Likewise.
(aarch64_vec_op_count::min_cycles_per_iter): Likewise.
(aarch64_vec_op_count::dump): Move earlier in file. Dump the
above properties too.
(aarch64_estimate_min_cycles_per_iter): Delete.
(adjust_body_cost): Use aarch64_vec_op_count::min_cycles_per_iter
instead of aarch64_estimate_min_cycles_per_iter. Rely on the dump
routine to print CPI estimates.
(adjust_body_cost_sve): Likewise. Use the other functions above
instead of doing the work inline.
|
|
-mtune=neoverse-512tvb uses two issue rates, one for Neoverse V1
and one with more generic parameters. We use both rates when
making a choice between scalar, Advanced SIMD and SVE code.
Previously we calculated the Neoverse V1 issue rates from the
more generic issue rates, but by removing m_scalar_ops and
(later) m_advsimd_ops, it becomes easier to track multiple
issue rates directly.
This patch therefore converts m_ops and (temporarily) m_advsimd_ops
into arrays.
gcc/
* config/aarch64/aarch64.c (aarch64_vec_op_count): Allow default
initialization.
(aarch64_vec_op_count::base_issue_info): Remove handling of null
issue_infos.
(aarch64_vec_op_count::simd_issue_info): Likewise.
(aarch64_vec_op_count::sve_issue_info): Likewise.
(aarch64_vector_costs::m_ops): Turn into a vector.
(aarch64_vector_costs::m_advsimd_ops): Likewise.
(aarch64_vector_costs::aarch64_vector_costs): Add entries to
the vectors based on aarch64_tune_params.
(aarch64_vector_costs::analyze_loop_vinfo): Update the pred_ops
of all entries in m_ops.
(aarch64_vector_costs::add_stmt_cost): Call count_ops for all
entries in m_ops.
(aarch64_estimate_min_cycles_per_iter): Remove issue_info
parameter and get the information from the ops instead.
(aarch64_vector_costs::adjust_body_cost_sve): Take a
aarch64_vec_issue_info instead of a aarch64_vec_op_count.
(aarch64_vector_costs::adjust_body_cost): Update call accordingly.
Exit earlier if m_ops is empty for either cost structure.
|
|
Now that vector finish_costs is passed the associated scalar costs,
we can record the scalar issue information while computing the scalar
costs, rather than trying to estimate it while computing the vector
costs.
This simplifies things a little, but the main motivation is to improve
accuracy.
gcc/
* config/aarch64/aarch64.c (aarch64_vector_costs::m_scalar_ops)
(aarch64_vector_costs::m_sve_ops): Replace with...
(aarch64_vector_costs::m_ops): ...this.
(aarch64_vector_costs::analyze_loop_vinfo): Update accordingly.
(aarch64_vector_costs::adjust_body_cost_sve): Likewise.
(aarch64_vector_costs::aarch64_vector_costs): Likewise.
Initialize m_vec_flags here rather than in add_stmt_cost.
(aarch64_vector_costs::count_ops): Test for scalar reductions too.
Allow vectype to be null.
(aarch64_vector_costs::add_stmt_cost): Call count_ops for scalar
code too. Don't require vectype to be nonnull.
(aarch64_vector_costs::adjust_body_cost): Take the loop_vec_info
and scalar costs as parameters. Use the scalar costs to determine
the cycles per iteration of the scalar loop, then multiply it
by the estimated VF.
(aarch64_vector_costs::finish_cost): Update call accordingly.
|
|
This patch gets the floatness of a memory access from the data
reference rather than the vectype. This makes it more suitable
for use in scalar costing code.
gcc/
* config/aarch64/aarch64.c (aarch64_dr_type): New function.
(aarch64_vector_costs::count_ops): Use it rather than the
vectype to determine floatness.
|
|
This patch gets the scalar mode of a reduction operation from the
gimple stmt rather than the vectype. This makes it more suitable
for use in scalar costs.
gcc/
* config/aarch64/aarch64.c (aarch64_sve_in_loop_reduction_latency):
Remove vectype parameter and get floatness from the type of the
stmt lhs instead.
(arch64_in_loop_reduction_latency): Likewise.
(aarch64_detect_vector_stmt_subtype): Update caller.
(aarch64_vector_costs::count_ops): Likewise.
|
|
Later patches make aarch64 use the new vector hooks. We then
only need to track one set of ops for each aarch64_vector_costs
structure. This in turn means that it's more convenient to merge
aarch64_sve_op_count and aarch64_vec_op_count.
The patch also adds issue info and vec flags to aarch64_vec_op_count,
so that the structure is more self-descriptive. This simplifies some
things later.
gcc/
* config/aarch64/aarch64.c (aarch64_sve_op_count): Fold into...
(aarch64_vec_op_count): ...this. Add a constructor.
(aarch64_vec_op_count::vec_flags): New function.
(aarch64_vec_op_count::base_issue_info): Likewise.
(aarch64_vec_op_count::simd_issue_info): Likewise.
(aarch64_vec_op_count::sve_issue_info): Likewise.
(aarch64_vec_op_count::m_issue_info): New member variable.
(aarch64_vec_op_count::m_vec_flags): Likewise.
(aarch64_vector_costs): Add a constructor.
(aarch64_vector_costs::m_sve_ops): Change type to aarch64_vec_op_count.
(aarch64_vector_costs::aarch64_vector_costs): New function.
Initialize m_scalar_ops, m_advsimd_ops and m_sve_ops.
(aarch64_vector_costs::count_ops): Remove vec_flags and
issue_info parameters, using the new aarch64_vec_op_count
functions instead.
(aarch64_vector_costs::add_stmt_cost): Update call accordingly.
(aarch64_sve_op_count::dump): Fold into...
(aarch64_vec_op_count::dump): ..here.
|
|
For tests like:
int res[2];
void
f1 (int x, int y)
{
res[0] = res[1] = x + y;
}
we generated:
add w0, w0, w1
adrp x1, .LANCHOR0
add x2, x1, :lo12:.LANCHOR0
str w0, [x1, #:lo12:.LANCHOR0]
str w0, [x2, 4]
ret
Using [x1, #:lo12:.LANCHOR0] for the first store prevented the
two stores being recognised as a pair. However, the MEM_EXPR
and MEM_OFFSET information tell us that the MEMs really are
consecutive. The peehole2 context then guarantees that the
first address is equivalent to [x2, 0].
While there: the reg_mentioned_p tests for loads were probably correct,
but seemed a bit indirect. We're matching two consecutive loads,
so the thing we need to test is that the second MEM in the original
sequence doesn't depend on the result of the first load in the
original sequence.
gcc/
* config/aarch64/aarch64.c: Include tree-dfa.h.
(aarch64_check_consecutive_mems): New function that takes MEM_EXPR
and MEM_OFFSET into account.
(aarch64_swap_ldrstr_operands): Use it.
(aarch64_operands_ok_for_ldpstp): Likewise. Check that the
address of the second memory doesn't depend on the result of
the first load.
gcc/testsuite/
* gcc.target/aarch64/stp_1.c: New test.
|
|
gcc/fortran/ChangeLog:
* parse.c (decode_omp_directive): Fix permitting 'nowait' for some
combined directives, add missing 'omp end ... loop'.
(gfc_ascii_statement): Fix ST_OMP_END_TEAMS_LOOP result.
* openmp.c (resolve_omp_clauses): Add missing combined loop constructs
case values to the 'if(directive-name: ...)' check.
* trans-openmp.c (gfc_split_omp_clauses): Put nowait on target if
first leaf construct accepting it.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/unexpected-end.f90: Update dg-error.
* gfortran.dg/gomp/clauses-1.f90: New test.
* gfortran.dg/gomp/nowait-2.f90: New test.
* gfortran.dg/gomp/nowait-3.f90: New test.
|
|
gcc/ChangeLog:
* ipa-pure-const.c (ipa_make_function_pure): Fix exit condition.
|
|
PR tree-optimization/103175
* ipa-modref.c (modref_lattice::merge): Add sanity check.
(callee_to_caller_flags): Make flags adjustment sane.
(modref_eaf_analysis::analyze_ssa_name): Likewise.
|
|
This patch converts one more open coded construction of a
DEBUG_EXPR_DECL to a call of build_debug_expr_decl that I missed in my
previous patch befause it happens to be in the Fortran front-end.
gcc/fortran/ChangeLog:
2021-11-11 Martin Jambor <mjambor@suse.cz>
* trans-types.c (gfc_get_array_descr_info): Use build_debug_expr_decl
instead of building DEBUG_EXPR_DECL manually.
|
|
PR testsuite/103051
gcc/testsuite/ChangeLog:
* gcc.dg/vect/tsvc/vect-tsvc-s112.c: Skip test for old Power
CPUs.
|
|
gcc/jit/ChangeLog:
PR jit/103199
* docs/examples/tut04-toyvm/toyvm.c (toyvm_function_compile):
Increase size of buffer.
* docs/examples/tut04-toyvm/toyvm.cc
(compilation_state::create_function): Likewise.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
PR ipa/103200
* ipa-modref.c (analyze_function, modref_propagate_in_scc): Do
not mark pure/const function if there are side-effects.
|
|
This patch implements relaxing the requirements when a map with the implicit
attribute encounters an overlapping existing map. As the OpenMP 5.0 spec
describes on page 320, lines 18-27 (and 5.1 spec, page 352, lines 13-22):
"If a single contiguous part of the original storage of a list item with an
implicit data-mapping attribute has corresponding storage in the device data
environment prior to a task encountering the construct that is associated with
the map clause, only that part of the original storage will have corresponding
storage in the device data environment as a result of the map clause."
2021-11-12 Chung-Lin Tang <cltang@codesourcery.com>
include/ChangeLog:
* gomp-constants.h (GOMP_MAP_FLAG_SPECIAL_3): Define special bit macro.
(GOMP_MAP_IMPLICIT): New special map kind bits value.
(GOMP_MAP_FLAG_SPECIAL_BITS): Define helper mask for whole set of
special map kind bits.
(GOMP_MAP_IMPLICIT_P): New predicate macro for implicit map kinds.
gcc/ChangeLog:
* tree.h (OMP_CLAUSE_MAP_RUNTIME_IMPLICIT_P): New access macro for
'implicit' bit, using 'base.deprecated_flag' field of tree_node.
* tree-pretty-print.c (dump_omp_clause): Add support for printing
implicit attribute in tree dumping.
* gimplify.c (gimplify_adjust_omp_clauses_1):
Set OMP_CLAUSE_MAP_RUNTIME_IMPLICIT_P to 1 if map clause is implicitly
created.
(gimplify_adjust_omp_clauses): Adjust place of adding implicitly created
clauses, from simple append, to starting of list, after non-map clauses.
* omp-low.c (lower_omp_target): Add GOMP_MAP_IMPLICIT bits into kind
values passed to libgomp for implicit maps.
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/target-implicit-map-1.c: New test.
* c-c++-common/goacc/combined-reduction.c: Adjust scan test pattern.
* c-c++-common/goacc/firstprivate-mappings-1.c: Likewise.
* c-c++-common/goacc/mdc-1.c: Likewise.
* g++.dg/goacc/firstprivate-mappings-1.C: Likewise.
libgomp/ChangeLog:
* target.c (gomp_map_vars_existing): Add 'bool implicit' parameter, add
implicit map handling to allow a "superset" existing map as valid case.
(get_kind): Adjust to filter out GOMP_MAP_IMPLICIT bits in return value.
(get_implicit): New function to extract implicit status.
(gomp_map_fields_existing): Adjust arguments in calls to
gomp_map_vars_existing, and add uses of get_implicit.
(gomp_map_vars_internal): Likewise.
* testsuite/libgomp.c-c++-common/target-implicit-map-1.c: New test.
|
|
The KIND argument of the INDEX intrinsic is a compile time constant
that is used at compile time only to resolve to a kind-specific library
function. That argument is otherwise completely ignored at runtime, and there is
no code generated for it as the library procedure has no kind argument.
This confuses the scalarizer which expects to see every argument
of elemental functions used when calling a procedure.
This change removes the argument from the scalarization lists
at the beginning of the scalarization process, so that the argument
is completely ignored.
This also reverts the existing workaround
(commit d09847357b965a2c2cda063827ce362d4c9c86f2 except for its testcase).
PR fortran/97896
gcc/fortran/ChangeLog:
* intrinsic.c (add_sym_4ind): Remove.
(add_functions): Use add_sym4 instead of add_sym4ind.
Don’t special case the index intrinsic.
* iresolve.c (gfc_resolve_index_func): Use the individual arguments
directly instead of the full argument list.
* intrinsic.h (gfc_resolve_index_func): Update the declaration
accordingly.
* trans-decl.c (gfc_get_extern_function_decl): Don’t modify the
list of arguments in the case of the index intrinsic.
* trans-array.h (gfc_get_intrinsic_for_expr,
gfc_get_proc_ifc_for_expr): New.
* trans-array.c (gfc_get_intrinsic_for_expr,
arg_evaluated_for_scalarization): New.
(gfc_walk_elemental_function_args): Add intrinsic procedure
as argument. Count arguments. Check arg_evaluated_for_scalarization.
* trans-intrinsic.c (gfc_walk_intrinsic_function): Update call.
* trans-stmt.c (get_intrinsic_for_code): New.
(gfc_trans_call): Update call.
gcc/testsuite/ChangeLog:
* gfortran.dg/index_5.f90: New.
|
|
The following patch implements what I've been talking about earlier,
honor that for explicit num_teams clause we create at least the
lower-bound (if not specified, upper-bound) teams in the league.
For host fallback, it still means we only have one thread doing all the
teams, sequentially one after another.
For PTX and GCN, I think the new teams-2.c test and maybe teams-4.c too
will or might fail.
For these offloads, I think it is ok to remove symbols no longer used
from libgomp.a.
If num_teams_lower is bigger than the provided num_blocks or num_workgroups,
we should arrange for gomp_num_teams_var to be num_teams_lower - 1,
stop using the %ctaid.x or __builtin_gcn_dim_pos (0) for omp_get_team_num ()
and instead use for it some .shared var that GOMP_teams4 initializes to
%ctaid.x or __builtin_gcn_dim_pos (0) when first and for !first
increment that by num_blocks or num_workgroups each time and only
return false when we are above num_teams_lower.
Any help with actually implementing this for the 2 architectures highly
appreciated.
2021-11-12 Jakub Jelinek <jakub@redhat.com>
gcc/
* omp-builtins.def (BUILT_IN_GOMP_TEAMS): Remove.
(BUILT_IN_GOMP_TEAMS4): New.
* builtin-types.def (BT_FN_VOID_UINT_UINT): Remove.
(BT_FN_BOOL_UINT_UINT_UINT_BOOL): New.
* omp-low.c (lower_omp_teams): Use GOMP_teams4 instead of
GOMP_teams, pass to it also num_teams lower-bound expression
or a dup of upper-bound if it is missing and a flag whether
it is the first call or not.
gcc/fortran/
* types.def (BT_FN_VOID_UINT_UINT): Remove.
(BT_FN_BOOL_UINT_UINT_UINT_BOOL): New.
libgomp/
* libgomp_g.h (GOMP_teams4): Declare.
* libgomp.map (GOMP_5.1): Export GOMP_teams4.
* target.c (GOMP_teams4): New function.
* config/nvptx/target.c (GOMP_teams): Remove.
(GOMP_teams4): New function.
* config/gcn/target.c (GOMP_teams): Remove.
(GOMP_teams4): New function.
* testsuite/libgomp.c/teams-4.c (main): Expect exactly 2
teams instead of <= 2.
* testsuite/libgomp.c-c++-common/teams-2.c: New test.
|
|
PR tree-optimization/102497
gcc/ChangeLog:
* gimple-predicate-analysis.cc (add_pred): Remove unused
function:
|
|
The following fixes a missed valueization when simplifying
a MEM[&...] combination during valueization.
2021-11-12 Richard Biener <rguenther@suse.de>
PR tree-optimization/103204
* tree-ssa-sccvn.c (valueize_refs_1): Re-valueize the
top operand after folding in an address.
* gcc.dg/torture/pr103204.c: New testcase.
|
|
|
|
All users of path_range_query are currently allocating a gimple_ranger
only to pass it to the query object. It's tidier to just do it from
path_range_query if no ranger was passed.
Tested on x86-64 Linux.
gcc/ChangeLog:
* gimple-range-path.cc (path_range_query::path_range_query): New
ctor without a ranger.
(path_range_query::~path_range_query): Free ranger if necessary.
(path_range_query::range_on_path_entry): Adjust m_ranger for pointer.
(path_range_query::ssa_range_in_phi): Same.
(path_range_query::compute_ranges_in_block): Same.
(path_range_query::compute_imports): Same.
(path_range_query::compute_ranges): Same.
(path_range_query::range_of_stmt): Same.
(path_range_query::compute_outgoing_relations): Same.
* gimple-range-path.h (class path_range_query): New ctor.
* tree-ssa-loop-ch.c (ch_base::copy_headers): Remove gimple_ranger
as path_range_query allocates one.
* tree-ssa-threadbackward.c (class back_threader): Remove m_ranger.
(back_threader::~back_threader): Same.
|
