* gcc-changelog/git_commit.py: Add author_tuple
only if not present in author_lines.
* gcc-changelog/test_email.py: New test.
* gcc-changelog/test_patches.txt: Add new patch.
This is new incarantion of patch to identify unmergeable tree at streaming out
time rather than streaming in and to avoid pickling them to sccs with with hash
codes.
Building cc1 plus this patch reduces:
[WPA] read 4452927 SCCs of average size 1.986030
[WPA] 8843646 tree bodies read in total
[WPA] tree SCC table: size 524287, 205158 elements, collision ratio: 0.505204
[WPA] tree SCC max chain length 43 (size 1)
[WPA] Compared 947551 SCCs, 780270 collisions (0.823460)
[WPA] Merged 944038 SCCs
[WPA] Merged 5253521 tree bodies
[WPA] Merged 590027 types
...
[WPA] Size of mmap'd section decls: 99229066 bytes
[WPA] Size of mmap'd section function_body: 18398837 bytes
[WPA] Size of mmap'd section refs: 733678 bytes
[WPA] Size of mmap'd section jmpfuncs: 2965981 bytes
[WPA] Size of mmap'd section pureconst: 170248 bytes
[WPA] Size of mmap'd section profile: 17985 bytes
[WPA] Size of mmap'd section symbol_nodes: 3392736 bytes
[WPA] Size of mmap'd section inline: 2693920 bytes
[WPA] Size of mmap'd section icf: 435557 bytes
[WPA] Size of mmap'd section offload_table: 0 bytes
[WPA] Size of mmap'd section lto: 4320 bytes
[WPA] Size of mmap'd section ipa_sra: 651660 bytes
... to ...
[WPA] read 3312246 unshared trees
[WPA] read 1144381 mergeable SCCs of average size 4.833785
[WPA] 8843938 tree bodies read in total
[WPA] tree SCC table: size 524287, 197767 elements, collision ratio: 0.506446
[WPA] tree SCC max chain length 43 (size 1)
[WPA] Compared 946614 SCCs, 775077 collisions (0.818789)
[WPA] Merged 943798 SCCs
[WPA] Merged 5253336 tree bodies
[WPA] Merged 590105 types
....
[WPA] Size of mmap'd section decls: 81262144 bytes
[WPA] Size of mmap'd section function_body: 14702611 bytes
[WPA] Size of mmap'd section ext_symtab: 0 bytes
[WPA] Size of mmap'd section refs: 733695 bytes
[WPA] Size of mmap'd section jmpfuncs: 2332150 bytes
[WPA] Size of mmap'd section pureconst: 170292 bytes
[WPA] Size of mmap'd section profile: 17986 bytes
[WPA] Size of mmap'd section symbol_nodes: 3393358 bytes
[WPA] Size of mmap'd section inline: 2567939 bytes
[WPA] Size of mmap'd section icf: 435633 bytes
[WPA] Size of mmap'd section lto: 4320 bytes
[WPA] Size of mmap'd section ipa_sra: 651824 bytes
so results in about 22% reduction in global decl stream and 24% reduction on
function bodies stream (which is read mostly by ICF)
Martin, the zstd compression breaks the compression statistics (it works when
GCC is configured for zlib)
At first ltrans I get:
[LTRANS] Size of mmap'd section decls: 3734248 bytes
[LTRANS] Size of mmap'd section function_body: 4895962 bytes
... to ...
[LTRANS] Size of mmap'd section decls: 3479850 bytes
[LTRANS] Size of mmap'd section function_body: 3722935 bytes
So 7% reduction of global stream and 31% reduction of function bodies.
Stream in seems to get about 3% faster and stream out about 5% but it is
close to noise factor of my experiment. I expect bigger speedups on
Firefox but I did not test it today since my Firefox setup broke again.
GCC is not very good example on the problem with anonymous namespace
types since we do not have so many of them.
Sice of object files in gcc directory is reduced by 11% (because hash
numbers do not compress well I guess).
The patch makes DFS walk to recognize trees that are not merged (anonymous
namespace, local function/variable decls, anonymous types etc). As discussed
on IRC this is now done during the SCC walk rather than during the hash
computation. When local tree is discovered we know that SCC components of everything that is on
the stack reffers to it and thus is also local. Moreover we mark trees into hash set in output block
so if we get a cross edge referring to local tree it gets marked too.
Patch also takes care of avoiding SCC wrappers around some trees. In particular
1) singleton unmergeable SCCs are now streamed inline in global decl stream
This includes INTEGER_CSTs and IDENTIFIER_NODEs that are shared by different
code than rest of tree merging.
2) We use LTO_trees instead of LTO_tree_scc to wrap unmergeable SCC components.
It is still necessary to mark them because of forward references. LTO_trees
has simple header with number of trees and then things are streamed same way
as for LTO_tree_scc. That is tree headers first followed by pickled references
so things may point to future.
Of course it is not necessary for LTO_tree_scc to be single component and
streamer out may group more components together, but I decided to not snowball
the patch even more
3) In local streams when lto_output_tree is called and the topmost SCC components
turns out to be singleton we stream the tree directly
instead of LTO_tree_scc, hash code, pickled tree, reference to just stremaed tree.
LTO_trees is used to wrap all trees needed to represent tree being streamed.
It would make sense again to use only one LTO_trees rather than one per SCC
but I think this can be done incrementally.
In general local trees are now recognized by new predicate local_tree_p
Bit subtle is handing of TRANLSATION_UNIT_DECL, INTEGER_CST and
IDENTIFIER_NODE.
TRANSLATION_UNIT_DECL a local tree but references to it does not make
other trees local (because we also understand local decls now).
So I check for it later after localness propagation is done.
INTEGER_CST and IDENTIFIER_NODEs are merged but not via the tree merging
machinery. So it makes sense to stream them as unmergeable trees but we
still need to compute their hashes so SCCs referring them do not get too
large collision chains. For this reason they are checked just prior
stream out.
lto-bootstrapped/regteted x86_64-linux, OK?
gcc/ChangeLog:
2020-05-19 Jan Hubicka <hubicka@ucw.cz>
* lto-streamer-in.c (lto_input_scc): Add SHARED_SCC parameter.
(lto_input_tree_1): Strenghten sanity check.
(lto_input_tree): Update call of lto_input_scc.
* lto-streamer-out.c: Include ipa-utils.h
(create_output_block): Initialize local_trees if merigng is going
to happen.
(destroy_output_block): Destroy local_trees.
(DFS): Add max_local_entry.
(local_tree_p): New function.
(DFS::DFS): Initialize and maintain it.
(DFS::DFS_write_tree): Decide on streaming format.
(lto_output_tree): Stream inline singleton SCCs
* lto-streamer.h (enum LTO_tags): Add LTO_trees.
(struct output_block): Add local_trees.
(lto_input_scc): Update prototype.
gcc/lto/ChangeLog:
2020-05-19 Jan Hubicka <hubicka@ucw.cz>
* lto-common.c (compare_tree_sccs_1): Sanity check that we never
read TRANSLATION_UNIT_DECL.
(process_dref): Break out from ...
(unify_scc): ... here.
(process_new_tree): Break out from ...
(lto_read_decls): ... here; handle streaming of singleton trees.
(print_lto_report_1): Update statistics.
The value 'ppr@' is set in the name of result symbol, the actual
name of the symbol is in the procedure name symbol pointed
to by the result symbol's namespace (ns). When reporting errors for
symbols that have the proc_pointer attribute check whether the
result attribute is set and set the name accordingly.
2020-05-20 Mark Eggleston <markeggleston@gcc.gnu.org>
gcc/fortran/
PR fortran/39695
* resolve.c (resolve_fl_procedure): Set name depending on
whether the result attribute is set. For PROCEDURE/RESULT
conflict use the name in sym->ns->proc_name->name.
* symbol.c (gfc_add_type): Add check for function and result
attributes use sym->ns->proc_name->name if both are set.
Where the symbol cannot have a type use the name in
sym->ns->proc_name->name.
2020-05-20 Mark Eggleston <markeggleston@gcc.gnu.org>
gcc/testsuite/
PR fortran/39695
* gfortran.dg/pr39695_1.f90: New test.
* gfortran.dg/pr39695_2.f90: New test.
* gfortran.dg/pr39695_3.f90: New test.
* gfortran.dg/pr39695_4.f90: New test.
_cpp_find_file has 3 bool arguments, at most one of which is ever set.
Ripe for replacing with a 4-state enum. Also, this is C++, so
'typedef struct Foo Foo' is unnecessary.
* internal.h (typedef _cpp_file): Delete, unnecessary in C++.
(enum _cpp_find_file_kind): New.
(_cpp_find_file): Use it, not 3 bools.
* files.c (_cpp_find_file): Use _cpp_find_file_kind enum, not
bools.
(cpp_make_system_header): Break overly long line.
(_cpp_stack_include, _cpp_fake_include)
(_cpp_do_file_change, _cpp_compare_file_date, _cpp_has_header): Adjust.
* init.c (cpp_read_main): Adjust _cpp_find_file call.
After enabling sanitization of the specialization tables, we are
triggering one of the hash table sanity checks in the below testcase.
The reason is that when looking up the specialization j<int> in the
type_specializations table, the sanity check finds that the existing
entry j<n<t>::m> compares equal to j<int> but hashes differently.
The discrepancy is due to structural_comptypes looking through
TYPENAME_TYPEs (via resolve_typename_type), something which
iterative_hash_template_arg doesn't do. So the TYPENAME_TYPE n<t>::m is
considered equal to int, but the hashes of these two template arguments
are different.
It seems wrong for the result of a specialization table lookup to depend
on the current scope, so this patch makes structural_comptypes avoid
calling resolve_typename_type when comparing_specializations.
In order for the below testcase to deterministically trigger the
sanitization error without this patch, we also need to fix the location
of the call to hash_table::verify within hash_table::find_with_hash.
gcc/ChangeLog:
PR c++/95223
* hash-table.h (hash_table::find_with_hash): Move up the call to
hash_table::verify.
gcc/cp/ChangeLog:
PR c++/95223
* typeck.c (structural_comptypes): Don't perform
context-dependent resolution of TYPENAME_TYPEs when
comparing_specializations.
gcc/testsuite/ChangeLog:
PR c++/95223
* g++.dg/template/typename23.C: New test.
This improves code generation with SSE2 for the testcase by
making sure to only generate a single IV when the group size
is a multiple of the vector size. It also adjusts the testcase
which was passing before.
2020-05-20 Richard Biener <rguenther@suse.de>
PR tree-optimization/95219
* tree-vect-loop.c (vectorizable_induction): Reduce
group_size before computing the number of required IVs.
* gcc.dg/vect/costmodel/x86_64/costmodel-pr30843.c: Adjust.
I mistook the opportunity to also "fix" the [VEC_]COND_EXPR case
for PR95171 but I was wrong in that it doesn't need the fix and
in the actual fix as well. The following just reverts that part.
2020-05-20 Richard Biener <rguenther@suse.de>
PR middle-end/95231
* tree-inline.c (remap_gimple_stmt): Revert adjusting
COND_EXPR and VEC_COND_EXPR for a -fnon-call-exception boundary.
* g++.dg/other/pr95231.C: New testcase.
Update VPCLMULQDQ check to support processors with AVX version of
VPCLMULQDQ.
PR target/91695
* config/i386/cpuinfo.c (get_available_features): Fix VPCLMULQDQ
check.
Few MVE intrinsics like vldrbq_s32, vldrhq_s32 etc., the assembler instructions
generated by current compiler are wrong.
eg: vldrbq_s32 generates an assembly instructions `vldrb.s32 q0,[ip]`.
But as per Arm-arm second argument in above instructions must also be a low
register (<= r7). This patch fixes this issue by creating a new predicate
"mve_memory_operand" and constraint "Ux" which allows low registers as arguments
to the generated instructions depending on the mode of the argument. A new constraint
"Ul" is created to handle loading to PC-relative addressing modes for vector
store/load intrinsiscs.
All the corresponding MVE intrinsic generating wrong code-gen as vldrbq_s32
are modified in this patch.
gcc/ChangeLog:
2020-05-20 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
Andre Vieira <andre.simoesdiasvieira@arm.com>
PR target/94959
* config/arm/arm-protos.h (arm_mode_base_reg_class): Function
declaration.
(mve_vector_mem_operand): Likewise.
* config/arm/arm.c (thumb2_legitimate_address_p): For MVE target check
the load from memory to a core register is legitimate for give mode.
(mve_vector_mem_operand): Define function.
(arm_print_operand): Modify comment.
(arm_mode_base_reg_class): Define.
* config/arm/arm.h (MODE_BASE_REG_CLASS): Modify to add check for
TARGET_HAVE_MVE and expand to arm_mode_base_reg_class on TRUE.
* config/arm/constraints.md (Ux): Likewise.
(Ul): Likewise.
* config/arm/mve.md (mve_mov): Replace constraint Us with Ux and also
add support for missing Vector Store Register and Vector Load Register.
Add a new alternative to support load from memory to PC (or label) in
vector store/load.
(mve_vstrbq_<supf><mode>): Modify constraint Us to Ux.
(mve_vldrbq_<supf><mode>): Modify constriant Us to Ux, predicate to
mve_memory_operand and also modify the MVE instructions to emit.
(mve_vldrbq_z_<supf><mode>): Modify constraint Us to Ux.
(mve_vldrhq_fv8hf): Modify constriant Us to Ux, predicate to
mve_memory_operand and also modify the MVE instructions to emit.
(mve_vldrhq_<supf><mode>): Modify constriant Us to Ux, predicate to
mve_memory_operand and also modify the MVE instructions to emit.
(mve_vldrhq_z_fv8hf): Likewise.
(mve_vldrhq_z_<supf><mode>): Likewise.
(mve_vldrwq_fv4sf): Likewise.
(mve_vldrwq_<supf>v4si): Likewise.
(mve_vldrwq_z_fv4sf): Likewise.
(mve_vldrwq_z_<supf>v4si): Likewise.
(mve_vld1q_f<mode>): Modify constriant Us to Ux.
(mve_vld1q_<supf><mode>): Likewise.
(mve_vstrhq_fv8hf): Modify constriant Us to Ux, predicate to
mve_memory_operand.
(mve_vstrhq_p_fv8hf): Modify constriant Us to Ux, predicate to
mve_memory_operand and also modify the MVE instructions to emit.
(mve_vstrhq_p_<supf><mode>): Likewise.
(mve_vstrhq_<supf><mode>): Modify constriant Us to Ux, predicate to
mve_memory_operand.
(mve_vstrwq_fv4sf): Modify constriant Us to Ux.
(mve_vstrwq_p_fv4sf): Modify constriant Us to Ux and also modify the MVE
instructions to emit.
(mve_vstrwq_p_<supf>v4si): Likewise.
(mve_vstrwq_<supf>v4si): Likewise.Modify constriant Us to Ux.
* config/arm/predicates.md (mve_memory_operand): Define.
gcc/testsuite/ChangeLog:
2020-05-20 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
PR target/94959
* gcc.target/arm/mve/intrinsics/mve_vector_float2.c: Modify.
* gcc.target/arm/mve/intrinsics/mve_vldr.c: New test.
* gcc.target/arm/mve/intrinsics/mve_vldr_z.c: Likewise.
* gcc.target/arm/mve/intrinsics/mve_vstr.c: Likewise.
* gcc.target/arm/mve/intrinsics/mve_vstr_p.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_f16.c: Modify.
* gcc.target/arm/mve/intrinsics/vld1q_f32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_s16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_s8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_u16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_u8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_f16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_f32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_s16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_s8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_u16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vld1q_z_u8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrbq_s8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrbq_u8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrbq_z_s8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrbq_z_u8.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrdq_gather_base_wb_s64.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrdq_gather_base_wb_u64.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrdq_gather_base_wb_z_s64.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrdq_gather_base_wb_z_u64.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_f16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_s16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_u16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_z_f16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_z_s16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_z_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_z_u16.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrhq_z_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_f32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_gather_base_wb_f32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_gather_base_wb_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_gather_base_wb_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_gather_base_wb_z_f32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_gather_base_wb_z_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_gather_base_wb_z_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_z_f32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_z_s32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vldrwq_z_u32.c: Likewise.
* gcc.target/arm/mve/intrinsics/vuninitializedq_float.c: Likewise.
* gcc.target/arm/mve/intrinsics/vuninitializedq_float1.c: Likewise.
* gcc.target/arm/mve/intrinsics/vuninitializedq_int.c: Likewise.
* gcc.target/arm/mve/intrinsics/vuninitializedq_int1.c: Likewise.
* gcc-changelog/git_commit.py: Refactor to make flake8 happy.
* gcc-changelog/test_email.py: Add new test.
* gcc-changelog/test_patches.txt: Add new patch.
git-check-commit.py does not know about "Reviewed-by",
"Reviewed-on", and "Signed-off-by" lines and hence it
expects those lines which follow the ChangeLog entries
to be indented by a tab.
This commit makes the script skip those lines. No further
processing is attempted because the review information
is not part of the ChangeLogs.
contrib/
2020-05-20 Frederik Harwath <frederik@codesourcery.com>
* gcc-changelog/git_commit.py: Skip over lines starting
with "Reviewed-by: ", "Reviewed-on: ", or "Signed-off-by: "
* gcc-changelog/git_commit.py: Support cherry pick
prefix.
* gcc-changelog/test_email.py: Test it.
* gcc-changelog/test_patches.txt: Add new patch.
* gcc-git-customization.sh: Add gcc-backport.
This fixes an integer overflow warning that ultimatively happens because
of TREE_OVERFLOW propagating through transforms and the existing guard
against this,
375 if (TREE_OVERFLOW_P (ret)
376 && !TREE_OVERFLOW_P (op0)
377 && !TREE_OVERFLOW_P (op1))
378 overflow_warning (EXPR_LOC_OR_LOC (expr, input_location,
being insufficient. Rather than trying to use sth like walk_tree to
exhaustively walk operands (with the possibility of introducing
quadraticness when folding larger expressions recursively) the
following amends the above with an ad-hoc test for a binary op0
with a possibly constant op1.
2020-05-30 Richard Biener <rguenther@suse.de>
PR c/95141
gcc/c
* c-fold.c (c_fully_fold_internal): Enhance guard on
overflow_warning.
gcc/testsuite
* gcc.dg/pr95141.c: New testcase.
- The alignment for local variable was adjust during estimate_stack_frame_size,
however it seems wrong spot to adjust that, expand phase will adjust that
but it little too late to some gimple optimization, which rely on certain
target hooks need to check alignment, forwprop is an example for
that, result of simplify_builtin_call rely on the alignment on some
target like ARM or RISC-V.
- Exclude static local var and hard register var in the process of
alignment adjustment.
- This patch fix gfortran.dg/pr45636.f90 for arm and riscv.
- Regression test on riscv32/riscv64 and x86_64-linux-gnu, no new fail
introduced.
gcc/ChangeLog
PR target/90811
* Makefile.in (OBJS): Add adjust-alignment.o.
* adjust-alignment.c (pass_data_adjust_alignment): New.
(pass_adjust_alignment): New.
(pass_adjust_alignment::execute): New.
(make_pass_adjust_alignment): New.
* tree-pass.h (make_pass_adjust_alignment): New.
* passes.def: Add pass_adjust_alignment.
It's not difficult for multiple threads to drain the entropy available
to the RDSEED instruction, at which point we throw an exception. This
change will try to use RDRAND after RDSEED fails repeatedly, and only
throw if RDRAND also fails repeatedly. This doesn't guarantee a random
value can always be read, but reduces the likelihood of failure when
using the RDSEED instruction.
PR libstdc++/94087
* src/c++11/random.cc (__x86_rdseed): Allow fallback function to be
passed in.
(__x86_rdseed_rdrand): New function that uses rdseed with rdrand
fallback.
(random_device::_M_init): Use __x86_rdseed_rdrand when both
instructions are available.
* testsuite/26_numerics/random/random_device/94087.cc: New test.
Add FEATURE_AVX512VP2INTERSECT to libgcc so that enum processor_features
in libgcc matches enum processor_features in i386-builtins.c. Update
GFNI check to support processors with SSE and AVX versions of GFNI.
PR target/95212
PR target/95220
* config/i386/cpuinfo.c (get_available_features): Fix
FEATURE_GFNI check. Also check FEATURE_AVX512VP2INTERSECT.
* config/i386/cpuinfo.h (processor_features): Add
FEATURE_AVX512VP2INTERSECT.
I discovered that the alias instantiation machinery would setup
template_info, and then sometime later overwrite that with equivalent
info. This broke modules, because the template info, once set, is
logically immutable. Let's just not do that.
* pt.c (lookup_template_class_1): Do not reinit template_info of an
alias here.
This fixes a bunch of poorly formatted decls, marks some getters as
PURE, deletes some C-relevant bool hackery, and finally uses a
passed-in location rather than deducing a closely-related but not
necessarily the same location.
* include/cpplib.h (cpp_get_otions, cpp_get_callbacks)
(cpp_get_deps): Mark as PURE.
* include/line-map.h (get_combined_adhoc_loc)
(get_location_from_adhoc_loc, get_pure_location): Reformat decls.
* internal.h (struct lexer_state): Clarify comment.
* system.h: Remove now-unneeded bool hackery.
* files.c (_cpp_find_file): Store LOC not highest_location.
This fixes PR94591. The problem was the function aarch64_evpc_rev_local()
matching vector permutations that were not reversals. In particular, prior to
this patch, this function matched the identity permutation which led to
generating bogus REV64 insns which were rejected by the assembler.
gcc/
PR target/94591
* config/aarch64/aarch64.c (aarch64_evpc_rev_local): Don't match
identity permutation.
gcc/testsuite/
PR target/94591
* gcc.c-torture/execute/pr94591.c: New test.
pr95149 is a false positive static analysis checker. But it
encouranged me to fix raw string lexing, which does contain a
complicated macro and pointers to local variables. The
reimplementation does away with that macro. Part of the complication
is we need to undo some of the fresh line processing -- trigraph notes
and escaped line continuations. But the undone characters need to go
through the raw string processing, as they can legitimately be part of
the prefix marker. however, in this reformulation we only process one
line marker at a time[*], so there's a limited number of undone
characters. We can arrange the buffering to make sure we don't split
such an append sequence, and then simply take the characters from the
append buffer.
The prefix scanner had a switch statement, which I discovered was not
optimized as well as an if of a bunch of explicit comparisons (pr
95208 filed).
Finally I adjusted the failure mode. When we get a bad prefix, we lex
up until the next '"', thus often swallowing the whole raw string.
Previously we'd bail and then the lexer would usually generate stupid
tokens, particularly when meeting the ending '"'.
libcpp/
* lex.c (struct lit_accum): New.
(bufring_append): Replace by lit_accum::append.
(lex_raw_string): Reimplement, using fragments of the old version.
(lex_string): Adjust lex_raw_string call.
gcc/testsuite/
* c-c++-common/raw-string-14.c: Adjust errors.
* c-c++-common/raw-string-16.c: Likewise.
* c-c++-common/raw-string-5.c: Likewise.
This adjusts the testcase for the introduced vector promotion/demotion
support.
2020-05-19 Richard Biener <rguenther@suse.de>
* gcc.target/i386/pr92645-4.c: Adjust expected pattern.
This was another latent case of us losing an EOF token, but succeeding
anyway. Since my patch to make us pay more attention to EOFs it came
to light. We also need to keep the EOF if we fall off the end of the
main file. Forced includes look like regular nested includes at this
point.
PR preprocessor/95182
libcpp/
* macro.c (collect_args): Preserve EOFif we fell out of the main
file.
(cpp_get_token_1): Reformat a couple of short lines.
This patch adds very basic allocator support (omp_{init,destroy}_allocator,
omp_{alloc,free}, omp_[sg]et_default_allocator).
The plan is to use memkind (likely dlopened) for high bandwidth memory, but
that part isn't implemented yet, probably mlock for pinned memory and see
what other options there are for other kinds of memory.
For offloading targets, we need to decide if we want to support the
dynamic allocators (and on which targets), or if e.g. all we do is at compile
time replace omp_alloc/omp_free calls with constexpr predefined allocators
with something special.
And allocate directive and allocator/uses_allocators clauses are future work
too.
2020-05-19 Jakub Jelinek <jakub@redhat.com>
* allocator.c: New file.
BPF considers that every call to a function allocates a fresh set of
registers that are available to the callee, of which the first five
may have bee initialized with the function arguments. This is
implemented by both interpreter and JIT in the Linux kernel.
This is enforced by the kernel BPF verifier, which will reject any
code in which non-initialized registers are accessed before being
written. Consequently, the spill instructions generated in function
prologue were causing the verifier to reject our compiled programs.
This patch makes GCC to not save/restore callee-saved registers in
function prologue/epilogue, unless xBPF mode is enabled.
2020-05-19 Jose E. Marchesi <jose.marchesi@oracle.com>
gcc/
* config/bpf/bpf.c (bpf_compute_frame_layout): Include space for
callee saved registers only in xBPF.
(bpf_expand_prologue): Save callee saved registers only in xBPF.
(bpf_expand_epilogue): Likewise for restoring.
* doc/invoke.texi (eBPF Options): Document this is activated by
-mxbpf.
gcc/testsuite/
* gcc.target/bpf/xbpf-callee-saved-regs-1.c: New test.
* gcc.target/bpf/xbpf-callee-saved-regs-2.c: Likewise.
This patch adds support for a new option -mxbpf. This tells GCC to
generate code for an expanded version of BPF that relaxes some of the
restrictions imposed by BPF.
2020-05-19 Jose E. Marchesi <jose.marchesi@oracle.com>
gcc/
* config/bpf/bpf.opt (mxbpf): New option.
* doc/invoke.texi (Option Summary): Add -mxbpf.
(eBPF Options): Document -mxbbpf.
contrib/ChangeLog:
2020-05-15 Martin Liska <mliska@suse.cz>
* gcc-git-customization.sh: Add
alias.gcc-mklog new hook.
* mklog.py: New file.
* test_mklog.py: New file.
This patch adds very basic allocator support (omp_{init,destroy}_allocator,
omp_{alloc,free}, omp_[sg]et_default_allocator).
The plan is to use memkind (likely dlopened) for high bandwidth memory, but
that part isn't implemented yet, probably mlock for pinned memory and see
what other options there are for other kinds of memory.
For offloading targets, we need to decide if we want to support the
dynamic allocators (and on which targets), or if e.g. all we do is at compile
time replace omp_alloc/omp_free calls with constexpr predefined allocators
with something special.
And allocate directive and allocator/uses_allocators clauses are future work
too.
2020-05-19 Jakub Jelinek <jakub@redhat.com>
* omp.h.in (omp_uintptr_t): New typedef.
(__GOMP_UINTPTR_T_ENUM): Define.
(omp_memspace_handle_t, omp_allocator_handle_t, omp_alloctrait_key_t,
omp_alloctrait_value_t, omp_alloctrait_t): New typedefs.
(__GOMP_DEFAULT_NULL_ALLOCATOR): Define.
(omp_init_allocator, omp_destroy_allocator, omp_set_default_allocator,
omp_get_default_allocator, omp_alloc, omp_free): Declare.
* libgomp.h (struct gomp_team_state): Add def_allocator field.
(gomp_def_allocator): Declare.
* libgomp.map (OMP_5.0.1): Export omp_set_default_allocator,
omp_get_default_allocator, omp_init_allocator, omp_destroy_allocator,
omp_alloc and omp_free.
* team.c (gomp_team_start): Copy over ts.def_allocator.
* env.c (gomp_def_allocator): New variable.
(parse_wait_policy): Adjust function comment.
(parse_allocator): New function.
(handle_omp_display_env): Print OMP_ALLOCATOR.
(initialize_env): Call parse_allocator.
* Makefile.am (libgomp_la_SOURCES): Add allocator.c.
* allocator.c: New file.
* icv.c (omp_set_default_allocator, omp_get_default_allocator): New
functions.
* testsuite/libgomp.c-c++-common/alloc-1.c: New test.
* testsuite/libgomp.c-c++-common/alloc-2.c: New test.
* testsuite/libgomp.c-c++-common/alloc-3.c: New test.
* Makefile.in: Regenerated.
- Implied rule are introduced into latest RISC-V ISA spec.
- Only implemented D implied F-extension. Zicsr and Zifence are not
implement yet, so the rule not included in this patch.
- Pass preprocessed arch string to arch.
- Verified with binutils 2.30 and 2.34.
gcc/ChangeLog
* common/config/riscv/riscv-common.c (riscv_implied_info_t): New.
(riscv_implied_info): New.
(riscv_subset_list): Add handle_implied_ext.
(riscv_subset_list::to_string): New parameter version_p to
control output format.
(riscv_subset_list::handle_implied_ext): New.
(riscv_subset_list::parse_std_ext): Call handle_implied_ext.
(riscv_arch_str): New parameter version_p to control output format.
(riscv_expand_arch): New.
* config/riscv/riscv-protos.h (riscv_arch_str): New parameter,
version_p.
* config/riscv/riscv.h (riscv_expand_arch): New,
(EXTRA_SPEC_FUNCTIONS): Define.
(ASM_SPEC): Transform -march= via riscv_expand_arch.
gcc/testsuite/ChangeLog
* gcc.target/riscv/arch-6.c: New.
* gcc.target/riscv/attribute-11.c: New.
* gcc.target/riscv/attribute-12.c: New.
- The arch string rule has changed in latest spec, it introduced new
multi-letter extension prefix with 'h' and 'z', and drop `sx`. also
adjust parsing order for 's' and 'x'.
gcc/ChangeLog
* riscv-common.c (parse_sv_or_non_std_ext): Rename to
parse_multiletter_ext.
(parse_multiletter_ext): Add parsing `h` and `z`, drop `sx`,
adjust parsing order for 's' and 'x'.
gcc/testsuite/ChangeLog
* gcc.target/riscv/arch-3.c: Adjust option.
* gcc.target/riscv/arch-5.c: New.
* gcc.target/riscv/attribute-9.c: Adjust option and test
condition.
2020-05-19 Richard Biener <rguenther@suse.de>
* tree-vectorizer.h (_slp_tree::vectype): Add field.
(SLP_TREE_VECTYPE): New.
* tree-vect-slp.c (vect_create_new_slp_node): Initialize
SLP_TREE_VECTYPE.
(vect_create_new_slp_node): Likewise.
(vect_prologue_cost_for_slp): Move here from tree-vect-stmts.c
and simplify.
(vect_slp_analyze_node_operations): Walk nodes children for
invariant costing.
(vect_get_constant_vectors): Use local scope op variable.
* tree-vect-stmts.c (vect_prologue_cost_for_slp_op): Remove here.
(vect_model_simple_cost): Adjust.
(vect_model_store_cost): Likewise.
(vectorizable_store): Likewise.
It looks like hash table sanitization is now safe to enable for the
decl_specializations and type_specializations tables, probably ever
since PR94454 was fixed.
gcc/cp/ChangeLog:
PR c++/87847
* pt.c (init_template_processing): Enable sanitization for
decl_specializations and type_specializations.
Martin Liska
Jan Hubicka
Mark Eggleston
Nathan Sidwell
Patrick Palka
Srinath Parvathaneni
Richard Biener
H.J. Lu
Frederik Harwath
Kito Cheng
GCC Administrator
Jonathan Wakely
Alex Coplan
Martin Sebor
Jozef Lawrynowicz
Jakub Jelinek
Jose E. Marchesi
Uros Bizjak