This patch adds a target selector for targets whose
preferred_vector_alignment is the alignment of one element. We'll never
peel in that case, and the step of a loop that operates on normal (as
opposed to packed) elements will always divide the preferred alignment.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* doc/sourcebuild.texi (vect_element_align_preferred): Document.
gcc/testsuite/
* lib/target-supports.exp
(check_effective_target_vect_element_align_preferred): New proc.
(check_effective_target_vect_peeling_profitable): Test it.
* gcc.dg/vect/no-section-anchors-vect-31.c: Don't expect peeling
if vect_element_align_preferred.
* gcc.dg/vect/no-section-anchors-vect-64.c: Likewise.
* gcc.dg/vect/pr65310.c: Likewise.
* gcc.dg/vect/vect-26.c: Likewise.
* gcc.dg/vect/vect-54.c: Likewise.
* gcc.dg/vect/vect-56.c: Likewise.
* gcc.dg/vect/vect-58.c: Likewise.
* gcc.dg/vect/vect-60.c: Likewise.
* gcc.dg/vect/vect-89-big-array.c: Likewise.
* gcc.dg/vect/vect-89.c: Likewise.
* gcc.dg/vect/vect-92.c: Likewise.
* gcc.dg/vect/vect-peel-1.c: Likewise.
* gcc.dg/vect/vect-outer-3a-big-array.c: Expect the step to
divide the alignment if vect_element_align_preferred.
* gcc.dg/vect/vect-outer-3a.c: Likewise.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254593
SLP load permutation fails if any individual permutation requires more
than two vector inputs. For 128-bit vectors, it's possible to permute
3 contiguous loads of 32-bit and 8-bit elements, but not 16-bit elements
or 64-bit elements. The results are reversed for 256-bit vectors,
and so on for wider vectors.
This patch adds a routine that tests whether a permute will require
three vectors for a given vector count and element size, then adds
vect_perm3_* target selectors for the cases that we currently use.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* doc/sourcebuild.texi (vect_perm_short, vect_perm_byte): Document
previously undocumented selectors.
(vect_perm3_byte, vect_perm3_short, vect_perm3_int): Document.
gcc/testsuite/
* lib/target-supports.exp (vect_perm_supported): New proc.
(check_effective_target_vect_perm3_int): Likewise.
(check_effective_target_vect_perm3_short): Likewise.
(check_effective_target_vect_perm3_byte): Likewise.
* gcc.dg/vect/slp-perm-1.c: Expect SLP load permutation to
succeed if vect_perm3_int.
* gcc.dg/vect/slp-perm-5.c: Likewise.
* gcc.dg/vect/slp-perm-6.c: Likewise.
* gcc.dg/vect/slp-perm-7.c: Likewise.
* gcc.dg/vect/slp-perm-8.c: Likewise vect_perm3_byte.
* gcc.dg/vect/slp-perm-9.c: Likewise vect_perm3_short.
Use vect_perm_short instead of vect_perm. Add a scan-tree-dump-not
test for vect_perm3_short targets.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254592
Some tests assumed that there would only be 2 vector sizes if
vect_multiple_sizes, whereas for SVE there are three (SVE, 128-bit
and 64-bit). This patch replaces scan-tree-dump-times with
scan-tree-dump for vect_multiple_sizes but keeps it for
!vect_multiple_sizes.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/testsuite/
* gcc.dg/vect/no-vfa-vect-101.c: Use scan-tree-dump rather than
scan-tree-dump-times for vect_multiple_sizes.
* gcc.dg/vect/no-vfa-vect-102.c: Likewise.
* gcc.dg/vect/no-vfa-vect-102a.c: Likewise.
* gcc.dg/vect/no-vfa-vect-37.c: Likewise.
* gcc.dg/vect/no-vfa-vect-79.c: Likewise.
* gcc.dg/vect/vect-104.c: Likewise.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254591
This patch adds a routine that lists the available vector sizes
for a target and uses it for some existing target conditions.
Later patches add more uses.
The cases are taken from multiple_sizes.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/testsuite/
* lib/target-supports.exp (available_vector_sizes): New proc.
(check_effective_target_vect_multiple_sizes): Use it.
(check_effective_target_vect64): Likewise.
(check_effective_target_vect_sizes_32B_16B): Likewise.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254590
Several vector tests are sensitive to the vector size. This patch adds
a VECTOR_BITS macro to tree-vect.h to select the expected vector size
and uses it to influence iteration counts and array sizes. The tests
keep the original values if the vector size is small enough.
For now VECTOR_BITS is always 128, but the SVE patches add other values.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/testsuite/
* gcc.dg/vect/tree-vect.h (VECTOR_BITS): Define.
* gcc.dg/vect/bb-slp-pr69907.c: Include tree-vect.h.
(N): New macro.
(foo): Use it instead of hard-coded 320.
* gcc.dg/vect/no-scevccp-outer-7.c (N): Redefine if the default
value is too small for VECTOR_BITS.
* gcc.dg/vect/no-scevccp-vect-iv-3.c (N): Likewise.
* gcc.dg/vect/no-section-anchors-vect-31.c (N): Likewise.
* gcc.dg/vect/no-section-anchors-vect-36.c (N): Likewise.
* gcc.dg/vect/slp-perm-9.c (N): Likewise.
* gcc.dg/vect/vect-32.c (N): Likewise.
* gcc.dg/vect/vect-75.c (N, OFF): Likewise.
* gcc.dg/vect/vect-77-alignchecks.c (N, OFF): Likewise.
* gcc.dg/vect/vect-78-alignchecks.c (N, OFF): Likewise.
* gcc.dg/vect/vect-89.c (N): Likewise.
* gcc.dg/vect/vect-96.c (N): Likewise.
* gcc.dg/vect/vect-multitypes-3.c (N): Likewise.
* gcc.dg/vect/vect-multitypes-6.c (N): Likewise.
* gcc.dg/vect/vect-over-widen-1.c (N): Likewise.
* gcc.dg/vect/vect-over-widen-4.c (N): Likewise.
* gcc.dg/vect/vect-reduc-pattern-1a.c (N): Likewise.
* gcc.dg/vect/vect-reduc-pattern-1b.c (N): Likewise.
* gcc.dg/vect/vect-reduc-pattern-2a.c (N): Likewise.
* gcc.dg/vect/no-section-anchors-vect-64.c (NINTS): New macro.
(N): Redefine in terms of NINTS.
(ia, ib, ic): Use NINTS instead of hard-coded constants in the
array bounds.
* gcc.dg/vect/no-section-anchors-vect-69.c (NINTS): New macro.
(N): Redefine in terms of NINTS.
(test1): Replace a and b fields with NINTS - 2 ints of padding.
(main1): Use NINTS instead of hard-coded constants.
* gcc.dg/vect/section-anchors-vect-69.c (NINTS): New macro.
(N): Redefine in terms of NINTS.
(test1): Replace a and b fields with NINTS - 2 ints of padding.
(test2): Remove incorrect comments about alignment.
(main1): Use NINTS instead of hard-coded constants.
* gcc.dg/vect/pr45752.c (N): Redefine if the default value is
too small for VECTOR_BITS.
(main): Continue to use canned results for the default value of N,
but compute the expected results from scratch for other values.
* gcc.dg/vect/slp-perm-1.c (N, main): As for pr45752.c.
* gcc.dg/vect/slp-perm-4.c (N, main): Likewise.
* gcc.dg/vect/slp-perm-5.c (N, main): Likewise.
* gcc.dg/vect/slp-perm-6.c (N, main): Likewise.
* gcc.dg/vect/slp-perm-7.c (N, main): Likewise.
* gcc.dg/vect/pr65518.c (NINTS, N, RESULT): New macros.
(giga): Use NINTS as the array bound.
(main): Use NINTS, N and RESULT.
* gcc.dg/vect/pr65947-5.c (N): Redefine if the default value is
too small for VECTOR_BITS.
(main): Fill in any remaining elements of A programmatically.
* gcc.dg/vect/pr81136.c: Include tree-vect.h.
(a): Use VECTOR_BITS to set the alignment of the target structure.
* gcc.dg/vect/slp-19c.c (N): Redefine if the default value is
too small for VECTOR_BITS.
(main1): Continue to use the canned input for the default value of N,
but compute the input from scratch for other values.
* gcc.dg/vect/slp-28.c (N): Redefine if the default value is
too small for VECTOR_BITS.
(in1, in2, in3): Remove initialization.
(check1, check2): Delete.
(main1): Initialize in1, in2 and in3 here. Check every element
of the vectors and compute the expected values directly instead
of using an array.
* gcc.dg/vect/slp-perm-8.c (N): Redefine if the default value is
too small for VECTOR_BITS.
(foo, main): Change type of "i" to int.
* gcc.dg/vect/vect-103.c (NINTS): New macro.
(N): Redefine in terms of N.
(c): Delete.
(main1): Use NINTS. Check the result from a and b directly.
* gcc.dg/vect/vect-67.c (NINTS): New macro.
(N): Redefine in terms of N.
(main1): Use NINTS for the inner array bounds.
* gcc.dg/vect/vect-70.c (NINTS, OUTERN): New macros.
(N): Redefine in terms of NINTS.
(s): Keep the outer dimensions as 4 even if N is larger than 24.
(tmp1): New variable.
(main1): Only define a local tmp1 if NINTS is relatively small.
Use OUTERN for the outer loops and NINTS for the inner loops.
* gcc.dg/vect/vect-91.c (OFF): New macro.
(a, main3): Use it.
* gcc.dg/vect/vect-92.c (NITER): New macro.
(main1, main2): Use it.
* gcc.dg/vect/vect-93.c (N): Rename to...
(N1): ...this.
(main): Update accordingly.
(N2): New macro.
(main1): Use N1 instead of 3001 and N2 insteaed of 10.
* gcc.dg/vect/vect-multitypes-1.c (NSHORTS, NINTS): New macros.
(N): Redefine in terms of NSHORTS.
(main1): Use NINTS - 1 instead of 3 and NSHORTS - 1 instead of 7.
(main): Likewise.
* gcc.dg/vect/vect-over-widen-3-big-array.c (N): Define to VECTOR_BITS.
(foo): Truncate the expected value to the type of *d.
* gcc.dg/vect/vect-peel-3.c (NINTS, EXTRA): New macros.
(ia, ib, ic, main): Use EXTRA.
(main): Use NINTS.
(RES_A, RES_B, REC_C): New macros.
(RES): Redefine as their sum.
* gcc.dg/vect/vect-reduc-or_1.c (N): New macro.
(in): Change number of elements to N.
(main): Update accordingly. Calculate the expected result.
* gcc.dg/vect/vect-reduc-or_2.c (N, in, main): As for
vect-reduc-or-1.c.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254589
The recent gen_vec_duplicate patches used CONST_VECTOR for all
constants, but the documentation says:
@findex const_vector
@item (const_vector:@var{m} [@var{x0} @var{x1} @dots{}])
Represents a vector constant. The square brackets stand for the vector
containing the constant elements. @var{x0}, @var{x1} and so on are
the @code{const_int}, @code{const_double} or @code{const_fixed} elements.
Both the AArch32 and AArch64 ports relied on the elements having
this form and would ICE if the element was something like a CONST
instead. This showed up as a failure in vect-126.c for both arm-eabi
and aarch64-elf (but not aarch64-linux-gnu, which is what the series
was tested on).
The two obvious options were to redefine CONST_VECTOR to accept all
constants or make gen_vec_duplicate honour the existing documentation.
It looks like other code also assumes that integer CONST_VECTORs contain
CONST_INTs, so the patch does the latter.
I deliberately didn't add an assert to gen_const_vec_duplicate
because it looks like the SPU port *does* expect to be able to create
CONST_VECTORs of symbolic constants.
Also, I think the list above should include const_wide_int for vectors
of TImode and wider.
The new routine takes a mode for consistency with the generators,
and because I think it does make sense to accept all constants for
variable-length:
(const (vec_duplicate ...))
rather than have some rtxes for which we instead use:
(vec_duplicate (const ...))
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
gcc/
* doc/rtl.texi (const_vector): Say that elements can be
const_wide_ints too.
* emit-rtl.h (valid_for_const_vec_duplicate_p): Declare.
* emit-rtl.c (valid_for_const_vec_duplicate_p): New function.
(gen_vec_duplicate): Use it instead of CONSTANT_P.
* optabs.c (expand_vector_broadcast): Likewise.
From-SVN: r254586
This patch improves the ivopts address cost calculation for modes
in which an index must be scaled rather than unscaled. Previously
we would only try the scaled form if the unscaled form was valid.
Many of the SVE tests rely on this when matching scaled indices.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-ssa-loop-ivopts.c (get_address_cost): Try using a
scaled index even if the unscaled address was invalid.
Don't increase the complexity of using a scale in that case.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254585
Originally subregs operated at the word level and subreg offsets
were measured in words. The offset units were later changed from
words to bytes (SUBREG_WORD became SUBREG_BYTE), but the fundamental
assumption that subregs should operate at the word level remained.
Whether (subreg:M1 (reg:M2 R2) N) is well-formed depended on the
way that M1 and M2 partitioned into words and whether the subword
part of N represented a lowpart. However, some questions depended
instead on the macro REGMODE_NATURAL_SIZE, which was introduced
as part of the patch that moved from SUBREG_WORD to SUBREG_BYTE.
It is used to decide whether setting (subreg:M1 (reg:M2 R2) N)
clobbers all of R2 or just part of it (df_read_modify_subreg).
Using words doesn't really make sense for modern vector
architectures. Vector registers are usually bigger than
a word and:
(a) setting the scalar lowpart of them usually clobbers the
rest of the register (contrary to the subreg rules,
where only the containing words should be clobbered).
(b) high words of vector registers are often not independently
addressable, even though that's what the subreg rules expect.
This patch therefore uses REGMODE_NATURAL_SIZE instead of
UNITS_PER_WORD to determine the size of the independently
addressable blocks in an inner register.
This is needed for SVE because the number of words in a vector
mode isn't known at compile time, so isn't a sensible basis
for calculating the number of registers.
The only existing port to define REGMODE_NATURAL_SIZE is
64-bit SPARC, where FP registers are 32 bits. (This is the
opposite of the use case for SVE, since the natural division
is smaller than a word.) I compiled the testsuite before and
after the patch for sparc64-linux-gnu and the only test whose
assembly changed was g++.dg/debug/pr65678.C, where the order
of two independent stores was reversed and where a different
register was picked for one pseudo. The new code was
otherwise equivalent to the old code.
2017-11-09 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* doc/rtl.texi: Rewrite the subreg rules so that they partition
the inner register into REGMODE_NATURAL_SIZE bytes rather than
UNITS_PER_WORD bytes.
* emit-rtl.c (validate_subreg): Divide subregs into blocks
based on REGMODE_NATURAL_SIZE of the inner mode.
(gen_lowpart_common): Split the SCALAR_FLOAT_MODE_P and
!SCALAR_FLOAT_MODE_P cases. Use REGMODE_NATURAL_SIZE for the latter.
* expmed.c (lowpart_bit_field_p): Divide the value up into
chunks of REGMODE_NATURAL_SIZE rather than UNITS_PER_WORD.
* expr.c (store_constructor): Use REGMODE_NATURAL_SIZE to test
whether something is likely to occupy more than one register.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r254583
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* sem_ch12.adb (Analyze_Generic_Package_Declaration): Handle properly
the pragma Compile_Time_Error when it appears in a generic package
declaration and uses an expanded name to denote the current unit.
2017-11-09 Jerome Lambourg <lambourg@adacore.com>
* libgnarl/s-taprop__qnx.adb: Fix incorrect casing for pthread_self.
* tracebak.c: Add support for tracebacks in QNX.
2017-11-09 Eric Botcazou <ebotcazou@adacore.com>
* exp_aggr.adb (Aggr_Size_OK): Bump base limit from 50000 to 500000.
2017-11-09 Yannick Moy <moy@adacore.com>
* erroutc.adb, set_targ.adb: Remove pragma Annotate for CodePeer
justification.
2017-11-09 Joel Brobecker <brobecker@adacore.com>
* doc/gnat_ugn/platform_specific_information.rst: Document packages
needed on GNU/Linux by GNAT.
* gnat_ugn.texi: Regenerate.
2017-11-09 Hristian Kirtchev <kirtchev@adacore.com>
* contracts.adb (Analyze_Contracts): Remove the three parameter
version. This routine now only analyzes contracts and does not perform
any freezing actions.
(Analyze_Previous_Contracts): Removed.
(Freeze_Previous_Contracts): New routine.
* contracts.ads (Analyze_Previous_Contracts): Removed.
(Freeze_Previous_Contracts): New routine.
* sem_ch3.adb (Analyze_Declarations): Analyze the contract of an
enclosing package spec regardless of whether the list denotes the
visible or private declarations. Fix the removal of partial state
refinements when the context is a package spec.
* sem_ch6.adb (Analyze_Subprogram_Body_Helper): Freeze previous
contracts.
* sem_ch7.adb (Analyze_Package_Body_Helper): Freeze previous contracts.
* sem_ch9.adb (Analyze_Entry_Body): Freeze previous contracts.
(Analyze_Protected_Body): Freeze previous contracts.
(Analyze_Task_Body): Freeze previous contracts.
* sem_prag.adb: Comment reformatting.
2017-11-09 Bob Duff <duff@adacore.com>
* libgnarl/g-thread.ads, libgnarl/g-thread.adb: (Make_Independent):
Export this so users can use it without importing
System.Tasking.Utilities.
* libgnarl/s-tassta.adb (Vulnerable_Complete_Task): Relax assertion
that fails when Make_Independent is called on a user task.
* libgnarl/s-taskin.ads (Master_Of_Task): Avoid unusual
capitalization style ((style) bad casing of "Master_of_Task").
From-SVN: r254580
* gimple-ssa-store-merging.c (count_multiple_uses): New function.
(split_group): Add total_orig and total_new arguments, estimate the
number of statements related to the store group without store merging
and with store merging.
(imm_store_chain_info::output_merged_store): Adjust split_group
callers, punt if estimated number of statements with store merging
is not smaller than estimated number of statements without it.
Formatting fix.
(handled_load): Remove has_single_use checks.
(pass_store_merging::process_store): Likewise.
From-SVN: r254579
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* sem_ch12.adb (Analyze_Subprogram_Instantiation): Correct use of
uninitialized variable uncovered by Codepeer.
2017-11-09 Arnaud Charlet <charlet@adacore.com>
* namet.adb: Replace pragma Assume by pragma Assert to fix bootstrap.
2017-11-09 Javier Miranda <miranda@adacore.com>
* doc/gnat_rm/standard_and_implementation_defined_restrictions.rst:
(Static_Dispatch_Tables): Minor rewording.
* gnat_rm.texi: Regenerate.
2017-11-09 Justin Squirek <squirek@adacore.com>
* sem_ch8.adb (Analyze_Use_Package): Remove forced installation of
use_clauses within instances.
(Use_One_Package): Add condition to check for "hidden" open scopes to
avoid skipping over packages that have not been properly installed even
though they are visible.
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* sem_ch4.adb (Analyze_Selected_Component): Reject properly a call to a
private operation of a protected type, when the type has no visible
operations.
From-SVN: r254578
2017-11-09 Javier Miranda <miranda@adacore.com>
* rtsfind.ads (RE_Id, RE_Unit_Table): Add RE_HT_Link.
* exp_disp.adb (Make_DT): Initialize the HT_Link field of the TSD only
if available.
2017-11-09 Bob Duff <duff@adacore.com>
* exp_ch4.adb, exp_ch9.adb, exp_prag.adb, par-ch3.adb, sem_aggr.adb,
sem_ch12.adb, sem_ch13.adb, sem_ch4.adb, sem_disp.adb, sem_prag.adb,
sem_res.adb, sem_util.adb: Get rid of warnings about uninitialized
variables.
From-SVN: r254577
2017-11-09 Pascal Obry <obry@adacore.com>
* libgnarl/s-taprop__mingw.adb: On Windows, initialize the thead handle
only for foreign threads. We initialize the thread handle only if not
yet initialized. This happens in Enter_Task for foreign threads only.
But for native threads (Ada tasking) we do want to keep the real
handle (from Create_Task) to be able to free the corresponding
resources in Finalize_TCB (CloseHandle).
2017-11-09 Yannick Moy <moy@adacore.com>
* sem_attr.adb (Analyze_Attribute): Default initialize P_Type,
P_Base_Type.
(Error_Attr_P): Fix name in pragma No_Return.
(Unexpected_Argument): Add pragma No_Return.
(Placement_Error): Add pragma No_Return.
2017-11-09 Javier Miranda <miranda@adacore.com>
* exp_disp.adb (Elab_Flag_Needed): Elaboration flag not needed when the
dispatch table is statically built.
(Make_DT): Declare constant the Interface_Table object associated with
an statically built dispatch table. For this purpose the Offset_To_Top
value of each interface is computed using the dummy object.
* exp_ch3.adb (Build_Init_Procedure): Do not generate code initializing
the Offset_To_Top field of secondary dispatch tables when the dispatch
table is statically built.
(Initialize_Tag): Do not generate calls to Register_Interface_Offset
when the dispatch table is statically built.
* doc/gnat_rm/standard_and_implementation_defined_restrictions.rst:
Document the new GNAT restriction Static_Dispatch_Tables.
* gnat_rm.texi: Regenerate.
2017-11-09 Hristian Kirtchev <kirtchev@adacore.com>
* sem_aggr.adb (Resolve_Delta_Record_Aggregate): Reorder declarations
to avoid a dormant bug.
2017-11-09 Jerome Lambourg <lambourg@adacore.com>
* init.c: Define missing __gnat_alternate_stack for QNX. Set it to 0,
as such capability is not available on the OS.
* link.c: Make sure linker options for QNX are correct.
* libgnarl/s-osinte__qnx.ads: Add some missing bindings to pthread.
* libgnarl/s-taprop__qnx.adb: New, derived from s-taprop__posix.adb. This brings
in particular a workaround with locks priority ceiling where a higher
priority task is allowed to lock a lower ceiling priority lock. This
also fixes the scheduling of FIFO tasks when the priority of a task is
lowered.
* libgnat/system-qnx-aarch64.ads: Fix priority ranges.
2017-11-09 Yannick Moy <moy@adacore.com>
* erroutc.adb (Output_Error_Msgs): Justify CodePeer false positive
message.
* gnatbind.adb (Scan_Bind_Arg): Simplify test to remove always true
condition.
* namet.adb (Copy_One_Character): Add assumption for static analysis,
as knowledge that Hex(2) is in the range 0..255 is too complex for
CodePeer.
(Finalize): Add assumption for static analysis, as the fact that there
are symbols in the table depends on a global invariant at this point in
the program.
* set_targ.adb (Check_Spaces): Justify CodePeer false positive message.
* stylesw.adb (Save_Style_Check_Options): Rewrite to avoid test always
true.
From-SVN: r254573
2017-11-09 Javier Miranda <miranda@adacore.com>
* libgnat/s-rident.ads (Static_Dispatch_Tables): New restriction name.
* exp_disp.adb (Building_Static_DT): Check restriction.
(Building_Static_Secondary_DT): Check restriction.
(Make_DT): Initialize the HT_Link to No_Tag.
* opt.ads (Static_Dispatch_Tables): Rename flag...
(Building_Static_Dispatch_Tables): ... into this. This will avoid
conflict with the restriction name.
* gnat1drv.adb: Update.
* exp_aggr.adb (Is_Static_Dispatch_Table_Aggregate): Update.
* exp_ch3.adb (Expand_N_Object_Declaration): Update.
2017-11-09 Pascal Obry <obry@adacore.com>
* libgnarl/s-taprop__mingw.adb: Minor code clean-up. Better using a
named number.
From-SVN: r254572
gcc/ada/
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* exp_ch3.adb, gnat1drv.adb, namet.adb, namet.ads, sem_aggr.adb,
sem_ch2.adb, sem_ch4.adb: Minor reformatting.
* sem_res.adb (Resolve_Entity_Name): Suppress spurious error on read of
out parameter when in Ada_83 mode, the oarameter is of a composite
type, and it appears as the prefix of an attribute.
2017-11-09 Bob Duff <duff@adacore.com>
* sinfo.ads: Minor comment fix.
2017-11-09 Hristian Kirtchev <kirtchev@adacore.com>
* sem_prag.ads: Add pragmas Unmodified and Unreferenced to table
Pragma_Significant_In_SPARK.
gcc/testsuite/
2017-11-09 Hristian Kirtchev <kirtchev@adacore.com>
* gnat.dg/unreferenced.adb: New testcase.
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* gnat.dg/out_param.adb: New testcase.
From-SVN: r254571
2017-11-09 Yannick Moy <moy@adacore.com>
* binde.adb (Diagnose_Elaboration_Problem): Mark procedure No_Return.
* checks.adb (Apply_Scalar_Range_Check): Rescope variable OK closer to
use. Default initialize Hi, Lo.
(Selected_Range_Checks): Retype Num_Checks more precisely.
(Determine_Range, Determine_Range_R): Default initialize Hi_Right,
Lo_Right.
* contracts.adb (Process_Contract_Cases): Mark parameter Stmts as
Unmodified.
(Process_Postconditions): Mark parameter Stmts as Unmodified.
* exp_attr.adb (Expand_Loop_Entry_Attribute): Default initialize Blk.
* exp_ch4.adb (Expand_N_Allocator): Default initialize Typ.
(Expand_Concatenate): Default initialize High_Bound.
(Optimize_Length_Comparison): Default initialize Ent, Index.
* exp_ch5.adb (Expand_Predicated_Loop): Default initialize L_Hi and
L_Lo.
* exp_ch6.adb (Expand_N_Extended_Return_Statement): Default initialize
Return_Stmt.
* exp_ch9.adb (Expand_Entry_Barrier): Default initialize Func_Body and
remove pragma Warnings(Off).
* exp_imgv.adb (Expand_Image_Attribute): Default initialize Tent.
* exp_util.adb (Find_Interface_Tag): Default initialize AI_Tag.
* freeze.adb (Check_Component_Storage_Order): Default initialize
Comp_Byte_Aligned rather than silencing messages with pragma
Warnings(Off), which does not work for CodePeer initialization
messages, and given that here the possible read of an unitialized value
depends on a proper use of parameters by the caller.
* inline.adb (Expand_Inlined_Call): Default initialize Lab_Decl, Targ.
* sem_ch12.adb (Build_Operator_Wrapper): Default initialize Expr.
* sem_ch3.adb (Build_Derived_Array_Type): Default initialize
Implicit_Base.
* sem_ch4.adb (List_Operand_Interps): Default initialize Nam and remove
pragma Warnings(Off).
(Analyze_Case_Expression): Rescope checking block within branch where
Others_Present is set by the call to Check_Choices.
* sem_ch5.adb (Analyze_Assignment): Default initialize
Save_Full_Analysis.
* sem_ch6.adb (Analyze_Function_Return): Default initialize Obj_Decl,
and restructure code to defend against previous errors, so that, in
that case, control does not flow to the elsif condition which read an
uninitialized Obj_Decl.
* sem_ch9.adb (Analyze_Requeue): Default initialize Synch_Type.
(Check_Interfaces): Default initialize Full_T_Ifaces and Priv_T_Ifaces,
which seem to be left uninitialized and possibly read in some cases.
* sem_dim.adb (Analyze_Aspect_Dimension_System): Retype Position more
precisely. This requires to exchange the test for exiting in case of
too many positions and the increment to Position, inside the loop.
* sem_eval.adb (Eval_Concatenation): Default initialize Folded_Val,
which cannot be read uninitialized, but the reasons for that are quite
subtle.
* sem_intr.adb (Check_Intrinsic_Call): Default initialize Rtyp.
* sem_prag.adb (Collect_Subprogram_Inputs_Outputs): Default initialize
Spec_Id.
* sem_res.adb (Make_Call_Into_Operator): Default initialize Opnd_Type,
and test for presence of non-null Opnd_Type before testing its scope,
in a test which would read its value uninitialized, and is very rarely
exercized (it depends on the presence of an extension of System).
* sem_spark.ads: Update comment to fix name of main analysis procedure.
* sem_warn.adb (Warn_On_Known_Condition): Default initialize
Test_Result.
* set_targ.adb (FailN): Mark procedure with No_Return.
* stylesw.adb (Save_Style_Check_Options): Delete useless code to
initialize all array Options to white space, as there is already code
doing the same for the remaining positions in Options at the end of the
procedure.
2017-11-09 Eric Botcazou <ebotcazou@adacore.com>
* exp_ch11.adb (Possible_Local_Raise): Do not issue the warning for
generic instantiations either.
From-SVN: r254570
2017-11-09 Piotr Trojanek <trojanek@adacore.com>
* sem_prag.adb (Analyze_Part_Of): Reword error message.
(Get_SPARK_Mode_Type): Do not raise Program_Error in case pragma
SPARK_Mode appears with an illegal mode, treat this as a non-existent
mode.
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* sem_ch4.adb (Analyze_Call): Reject a call to a function that returns
a limited view of a type T declared in unit U1, when the function is
declared in another unit U2 and the call appears in a procedure within
another unit.
2017-11-09 Justin Squirek <squirek@adacore.com>
* sem_ch8.adb (Analyze_Use_Package): Force installation of use_clauses
when processing generic instances.
2017-11-09 Bob Duff <duff@adacore.com>
* namet.ads, namet.adb (Valid_Name_Id): New subtype that excludes
Error_Name and No_Name. Use this (versus Name_Id) to indicate which
objects can have those special values. Valid_Name_Id could usefully be
used all over the compiler front end, but that's too much trouble for
now. If we did that, we might want to rename:
Name_Id --> Optional_Name_Id, Valid_Name_Id --> Name_Id.
For parameters of type Valid_Name_Id, remove some redundant tests,
including the ones found by CodePeer. Use Is_Valid_Name instead of
membership test when appropriate.
(Error_Name_Or_No_Name): Delete this; it's no longer needed.
* sem_ch2.adb (Analyze_Identifier): Use "not Is_Valid_Name" instead of
"in Error_Name_Or_No_Name".
(Check_Parameterless_Call): Use "not Is_Valid_Name" instead of "in
Error_Name_Or_No_Name".
From-SVN: r254569
gcc/ada/
2017-11-09 Arnaud Charlet <charlet@adacore.com>
* gnat1drv.adb (Adjust_Global_Switches): Suppress warnings in codepeer
mode here unless -gnateC is specified.
* switch-c.adb (Scan_Front_End_Switches): Do not suppress warnings with
-gnatC here.
2017-11-09 Piotr Trojanek <trojanek@adacore.com>
* lib-writ.adb (Write_ALI): Remove processing of the frontend xrefs as
part of the ALI writing; they are now processed directly from memory
when requested by the backend.
* lib-xref.ads (Collect_SPARK_Xrefs): Remove.
(Iterate_SPARK_Xrefs): New routine for iterating over frontend xrefs.
* lib-xref-spark_specific.adb (Traverse_Compilation_Unit): Remove.
(Add_SPARK_File): Remove.
(Add_SPARK_Xref): Refactored from removed code; filters xref entries
that are trivially uninteresting to the SPARK backend.
* spark_xrefs.ads: Remove code that is no longer needed.
* spark_xrefs.adb (dspark): Adapt to use Iterate_SPARK_Xrefs.
2017-11-09 Hristian Kirtchev <kirtchev@adacore.com>
* sem_elab.adb: Update the documentation on adding a new elaboration
schenario. Add new hash table Recorded_Top_Level_Scenarios.
(Is_Check_Emitting_Scenario): Removed.
(Is_Recorded_Top_Level_Scenario): New routine.
(Kill_Elaboration_Scenario): Reimplemented.
(Record_Elaboration_Scenario): Mark the scenario as recorded.
(Set_Is_Recorded_Top_Level_Scenario): New routine.
(Update_Elaboration_Scenario): Reimplemented.
* sinfo.adb (Is_Recorded_Scenario): Removed.
(Set_Is_Recorded_Scenario): Removed.
* sinfo.ads: Remove attribute Is_Recorded_Scenario along with
occurrences in nodes.
(Is_Recorded_Scenario): Removed along with pragma Inline.
(Set_Is_Recorded_Scenario): Removed along with pragma Inline.
2017-11-09 Piotr Trojanek <trojanek@adacore.com>
* sem_prag.adb (Analyze_Part_Of): Change "designate" to "denote" in
error message.
2017-11-09 Justin Squirek <squirek@adacore.com>
* sem_res.adb (Resolve_Allocator): Add warning messages corresponding
to the allocation of an anonymous access-to-controlled object.
gcc/testsuite/
2017-11-09 Hristian Kirtchev <kirtchev@adacore.com>
* gnat.dg/elab3.adb, gnat.dg/elab3.ads, gnat.dg/elab3_pkg.adb,
gnat.dg/elab3_pkg.ads: New testcase.
2017-11-09 Pierre-Marie de Rodat <derodat@adacore.com>
* gnat.dg/controlled2.adb, gnat.dg/controlled4.adb, gnat.dg/finalized.adb:
Disable the new warning from GNAT.
From-SVN: r254568
2017-11-09 Jerome Lambourg <lambourg@adacore.com>
* sigtramp-qnx.c: Fix obvious typo.
2017-11-09 Doug Rupp <rupp@adacore.com>
* libgnarl/s-taprop__linux.adb (Monotonic_Clock): Minor reformatting.
2017-11-09 Ed Schonberg <schonberg@adacore.com>
* sem_res.adb (Resolve): If expression is an entity whose type has
implicit dereference, generate reference to it, because no reference is
generated for an overloaded entity during analysis, given that its
identity may not be known.
2017-11-09 Javier Miranda <miranda@adacore.com>
* exp_disp.adb (Expand_Interface_Thunk): Replace substraction of
offset-to-top field by addition.
(Make_Secondary_DT): Initialize the offset-to-top field with a negative
offset.
* exp_ch3.adb (Build_Offset_To_Top_Function): Build functions that
return a negative offset-to-top value.
(Initialize_Tag): Invoke runtime services Set_Dynamic_Offset_To_Top and
Set_Static_Offset_To_Top passing a negative offet-to-top value;
initialize also the offset-to-top field with a negative offset.
* libgnat/a-tags.adb (Base_Address): Displace the pointer by means of
an addition since the offset-to-top field is now a negative value.
(Displace): Displace the pointer to the object means of a substraction
since it is now a negative value.
(Set_Dynamic_Offset_to_top): Displace the pointer to the object by
means of a substraction since it is now a negative value.
2017-11-09 Eric Botcazou <ebotcazou@adacore.com>
* gnat1drv.adb (Gnat1drv): Call Errout.Finalize (Last_Call => True)
before Errout.Output_Messages also in the case of compilation errors.
2017-11-09 Javier Miranda <miranda@adacore.com>
* doc/gnat_ugn/the_gnat_compilation_model.rst (Interfacing with C++ at
the Class Level): Fix error interfacing with C strings.
* gnat_ugn.texi: Regenerate.
2017-11-09 Jerome Lambourg <lambourg@adacore.com>
* system-qnx-aarch64.ads: Fix the priority constants.
* s-osinte__qnx.ads: Fix constants for handling the locking protocols
and scheduling.
* s-osinte__qnx.adb: New file , prevents the use of priority 0 that
corresponds to an idle priority on QNX.
2017-11-09 Piotr Trojanek <trojanek@adacore.com>
* sem_prag.adb, sem_util.adb, sem_elab.adb: Fix minor typos in
comments.
From-SVN: r254566
2017-11-09 Piotr Trojanek <trojanek@adacore.com>
* lib-xref-spark_specific.adb (Add_SPARK_Xrefs): Ignore loop parameters
in expression funtions that are expanded into variables.
2017-11-09 Piotr Trojanek <trojanek@adacore.com>
* sem_util.adb: Minor whitespace cleanup.
2017-11-09 Jerome Lambourg <lambourg@adacore.com>
* libgnarl/s-taprop__qnx.adb: Refine aarch64-qnx. Use the POSIX
s-taprop version rather than a custom one.
* sigtramp-qnx.c (aarch64-qnx): Implement the signal trampoline.
From-SVN: r254563
2017-11-09 Martin Liska <mliska@suse.cz>
PR gcov-profile/48463
* coverage.c (coverage_begin_function): Output also end locus
of a function and information whether the function is
artificial.
* gcov-dump.c (tag_function): Parse and print the information.
* gcov.c (INCLUDE_MAP): Add include.
(INCLUDE_SET): Likewise.
(struct line_info): Move earlier in the source file because
of vector<line_info> in function_info structure.
(line_info::line_info): Likewise.
(line_info::has_block): Likewise.
(struct source_info): Add new member index.
(source_info::get_functions_at_location): New function.
(function_info::group_line_p): New function.
(output_intermediate_line): New function.
(output_intermediate_file): Use the mentioned function.
(struct function_start): New.
(struct function_start_pair_hash): Likewise.
(process_file): Add code that identifies group functions.
Assign lines either to global or function scope.
(generate_results): Skip artificial functions.
(find_source): Assign index for each source file.
(read_graph_file): Read new flag artificial and end_line.
(add_line_counts): Assign it either to global of function scope.
(accumulate_line_counts): Isolate core of the function to
accumulate_line_info and call it for both function and global
scope lines.
(accumulate_line_info): New function.
(output_line_beginning): Fix GNU coding style.
(print_source_line): New function.
(output_line_details): Likewise.
(output_function_details): Likewise.
(output_lines): Iterate both source (global) scope and function
scope.
(struct function_line_start_cmp): New class.
* doc/gcov.texi: Reflect changes in documentation.
From-SVN: r254562
PR debug/82837
* dwarf2out.c (const_ok_for_output_1): Reject NEG in addition to NOT.
(mem_loc_descriptor): Handle (const (neg (...))) as (neg (const (...)))
and similarly for not instead of neg.
* gcc.dg/debug/dwarf2/pr82837.c: New test.
From-SVN: r254561
This patch adds a -mforce-indirect-call option to force all calls
or tail calls on x86_64 between functions to indirect. This is similar to the
large code model, but doesn't affect jumps inside functions, so has much
less run time overhead.
This is useful with Intel Processor Trace (PT). PT has precise timing
for indirect calls/jumps, but not for direct ones. So if we can force
them to indirect it allows to time every function relatively accurately
(minus the overhead of the indirect branch)
Without this short functions often don't see a timing update and cannot
be measured.
The timing requires at least Skylake or Goldmont based CPUs.
I made it an option. Originally I tried to make it a new code model,
but since it can be combined with other code models (medium, pic, kernel
etc.) this turned out to be too many combinations.
For example with gcc. This first column is a ns time stamp for the
functions.
$ perf record -e intel_pt/noretcomp=1,cyc=1,cyc_thresh=1/u ./cc1 -O3 hello.c
$ perf script --itrace=cr -F callindent,time,sym,addr --ns | sed -n 180000,182000p | less
...
1184596.432756920: build_int_cst => 79c9de c_common_nodes_and_builtins
1184596.432756921: tree_cons => ee2080 tree_cons
1184596.432756938: ggc_internal_alloc => 80f3e0 ggc_internal_alloc
1184596.432756951: memset@plt => 598af0 memset@plt
1184596.432756967: __memset_avx2_unaligned_erms => 80f605 ggc_internal_alloc
1184596.432756969: ggc_internal_alloc => ee20a2 tree_cons
1184596.432756973: tree_cons => 79c9f4 c_common_nodes_and_builtins
1184596.432756974: build_int_cst => ef9a40 build_int_cst
1184596.432756996: wide_int_to_tree => ef93a0 wide_int_to_tree
1184596.432757000: wi::force_to_size => f48f70 wi::force_to_size
1184596.432757005: canonize => ef94de wide_int_to_tree
1184596.432757021: get_int_cst_ext_nunits => ee1960 get_int_cst_ext_nunits
1184596.432757026: get_int_cst_ext_nunits => ef94fe wide_int_to_tree
1184596.432757042: tree_int_cst_elt_check => 83e310 tree_int_cst_elt_check
1184596.432757044: tree_int_cst_elt_check => ef9761 wide_int_to_tree
1184596.432757046: wide_int_to_tree => ef9a9b build_int_cst
gcc/:
2017-11-08 Andi Kleen <ak@linux.intel.com>
* config/i386/i386.opt: Add -mforce-indirect-call.
* config/i386/predicates.md: Check for flag_force_indirect_call.
* doc/invoke.texi: Document -mforce-indirect-call
gcc/testsuite/:
2017-11-08 Andi Kleen <ak@linux.intel.com>
* gcc.target/i386/force-indirect-call-1.c: New test.
* gcc.target/i386/force-indirect-call-2.c: New test.
* gcc.target/i386/force-indirect-call-3.c: New test.
From-SVN: r254560
2017-11-08 Steven G. Kargl <kargl@kgcc.gnu.org>
PR Fortran/82841
* simplify.c(gfc_simplify_transfer): Do not dereference a NULL pointer.
Unwrap a short line.
2017-11-08 Steven G. Kargl <kargl@kgcc.gnu.org>
PR Fortran/82841
* gfortran.dg/transfer_simplify_11.f90: new test.
From-SVN: r254555
- This build error was indroduced by "RISC-V: Implement movmemsi"
and "RISC-V: Support -mpreferred-stack-boundary flag"
gcc/ChangeLog
2017-11-08 Kito Cheng <kito.cheng@gmail.com>
* config/riscv/riscv-protos.h (riscv_slow_unaligned_access_p):
New extern.
(MOVE_RATIO): Use riscv_slow_unaligned_access_p.
config/riscv/riscv.c (predict.h): New include.
(riscv_slow_unaligned_access_p): No longer static.
(riscv_block_move_straight): Add require.
config/riscv/riscv-protos.h (riscv_hard_regno_nregs): Delete.
From-SVN: r254554
2017-11-08 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/82884
* arith.c (gfc_hollerith2character): Clear pad.
2017-11-08 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/82884
* gfortran.dg/hollerith_character_array_constructor.f90: New test.
From-SVN: r254553
On top of the previous vec_merge simplifications [1] we can add this pattern to perform
a store of a vec_concat of two 64-bit values in distinct registers as an STP.
This avoids constructing such a vector explicitly in a register and storing it as
a Q register.
This way for the code in the testcase we can generate:
construct_lane_1:
ldp d1, d0, [x0]
fmov d3, 1.0e+0
fmov d2, 2.0e+0
fadd d4, d1, d3
fadd d5, d0, d2
stp d4, d5, [x1, 32]
ret
construct_lane_2:
ldp x2, x0, [x0]
add x3, x2, 1
add x4, x0, 2
stp x3, x4, [x1, 32]
ret
instead of the current:
construct_lane_1:
ldp d0, d1, [x0]
fmov d3, 1.0e+0
fmov d2, 2.0e+0
fadd d0, d0, d3
fadd d1, d1, d2
dup v0.2d, v0.d[0]
ins v0.d[1], v1.d[0]
str q0, [x1, 32]
ret
construct_lane_2:
ldp x2, x3, [x0]
add x0, x2, 1
add x2, x3, 2
dup v0.2d, x0
ins v0.d[1], x2
str q0, [x1, 32]
ret
Bootstrapped and tested on aarch64-none-linux-gnu.
[1] https://gcc.gnu.org/ml/gcc-patches/2017-06/msg00272.htmlhttps://gcc.gnu.org/ml/gcc-patches/2017-06/msg00273.htmlhttps://gcc.gnu.org/ml/gcc-patches/2017-06/msg00274.html
* config/aarch64/aarch64-simd.md (store_pair_lanes<mode>):
New pattern.
* config/aarch64/constraints.md (Uml): New constraint.
* config/aarch64/predicates.md (aarch64_mem_pair_lanes_operand): New
predicate.
* gcc.target/aarch64/store_v2vec_lanes.c: New test.
From-SVN: r254551
Another vec_merge simplification that's missing from simplify-rtx.c is transforming
a vec_merge of two vec_duplicates. For example:
(set (reg:V2DF 80)
(vec_merge:V2DF (vec_duplicate:V2DF (reg:DF 84))
(vec_duplicate:V2DF (reg:DF 81))
(const_int 2)))
Can be transformed into the simpler:
(set (reg:V2DF 80)
(vec_concat:V2DF (reg:DF 81)
(reg:DF 84)))
I believe this should always be beneficial.
I'm still looking into finding a small testcase demonstrating this, but on aarch64 SPEC
I've seen this eliminate some really bizzare codegen where GCC was generating nonsense like:
ldr q18, [sp, 448]
ins v18.d[0], v23.d[0]
ins v18.d[1], v22.d[0]
With q18 being pushed and popped off the stack in the prologue and epilogue of the function!
These are large files from SPEC that I haven't been able to analyse yet as to why GCC even attempts
to do that, but with this patch it doesn't try to load a register and overwrite all its lanes.
This patch shaves off about 5k of code size from zeusmp on aarch64 at -O3, so I believe it's a good
thing to do.
* simplify-rtx.c (simplify_ternary_operation): Simplify vec_merge
of two vec_duplicates into a vec_concat.
From-SVN: r254550
Another vec_merge simplification that's missing is transforming:
(vec_merge (vec_duplicate x) (vec_concat (y) (z)) (const_int N))
into
(vec_concat x z) if N == 1 (0b01) or
(vec_concat y x) if N == 2 (0b10)
For the testcase in this patch on aarch64 this allows us to try matching during combine the pattern:
(set (reg:V2DI 78 [ x ])
(vec_concat:V2DI
(mem:DI (reg/v/f:DI 76 [ y ]) [1 *y_4(D)+0 S8 A64])
(mem:DI (plus:DI (reg/v/f:DI 76 [ y ])
(const_int 8 [0x8])) [1 MEM[(long long int *)y_4(D) + 8B]+0 S8 A64])))
rather than the more complex:
(set (reg:V2DI 78 [ x ])
(vec_merge:V2DI (vec_duplicate:V2DI (mem:DI (plus:DI (reg/v/f:DI 76 [ y ])
(const_int 8 [0x8])) [1 MEM[(long long int *)y_4(D) + 8B]+0 S8 A64]))
(vec_duplicate:V2DI (mem:DI (reg/v/f:DI 76 [ y ]) [1 *y_4(D)+0 S8 A64]))
(const_int 2 [0x2])))
We don't actually have an aarch64 pattern for the simplified version above, but it's a simple enough
form to add, so this patch adds such a pattern that performs a concatenated load of two 64-bit vectors
in adjacent memory locations as a single Q-register LDR. The new aarch64 pattern is needed to demonstrate
the effectiveness of the simplify-rtx change, so I've kept them together as one patch.
Now for the testcase in the patch we can generate:
construct_lanedi:
ldr q0, [x0]
ret
construct_lanedf:
ldr q0, [x0]
ret
instead of:
construct_lanedi:
ld1r {v0.2d}, [x0]
ldr x0, [x0, 8]
ins v0.d[1], x0
ret
construct_lanedf:
ld1r {v0.2d}, [x0]
ldr d1, [x0, 8]
ins v0.d[1], v1.d[0]
ret
The new memory constraint Utq is needed because we need to allow only the Q-register addressing modes but
the MEM expressions in the RTL pattern have 64-bit vector modes, and if we don't constrain them they will
allow the D-register addressing modes during register allocation/address mode selection, which will produce
invalid assembly.
Bootstrapped and tested on aarch64-none-linux-gnu.
* simplify-rtx.c (simplify_ternary_operation, VEC_MERGE):
Simplify vec_merge of vec_duplicate and vec_concat.
* config/aarch64/constraints.md (Utq): New constraint.
* config/aarch64/aarch64-simd.md (load_pair_lanes<mode>): New
define_insn.
* gcc.target/aarch64/load_v2vec_lanes_1.c: New test.
From-SVN: r254549
I'm trying to improve some of the RTL-level handling of vector lane operations on aarch64 and that
involves dealing with a lot of vec_merge operations. One simplification that I noticed missing
from simplify-rtx are combinations of vec_merge with vec_duplicate.
In this particular case:
(vec_merge (vec_duplicate (X)) (const_vector [A, B]) (const_int N))
which can be replaced with
(vec_concat (X) (B)) if N == 1 (0b01) or
(vec_concat (A) (X)) if N == 2 (0b10).
For the aarch64 testcase in this patch this simplifications allows us to try to combine:
(set (reg:V2DI 77 [ x ])
(vec_concat:V2DI (mem:DI (reg:DI 0 x0 [ y ]) [1 *y_3(D)+0 S8 A64])
(const_int 0 [0])))
instead of the more complex:
(set (reg:V2DI 77 [ x ])
(vec_merge:V2DI (vec_duplicate:V2DI (mem:DI (reg:DI 0 x0 [ y ]) [1 *y_3(D)+0 S8 A64]))
(const_vector:V2DI [
(const_int 0 [0])
(const_int 0 [0])
])
(const_int 1 [0x1])))
For the simplified form above we already have an aarch64 pattern: *aarch64_combinez<mode> which
is missing a DI/DFmode version due to an oversight, so this patch extends that pattern as well to
use the VDC mode iterator that includes DI and DFmode (as well as V2HF which VD_BHSI was missing).
The aarch64 hunk is needed to see the benefit of the simplify-rtx.c hunk, so I didn't split them
into separate patches.
Before this for the testcase we'd generate:
construct_lanedi:
movi v0.4s, 0
ldr x0, [x0]
ins v0.d[0], x0
ret
construct_lanedf:
movi v0.2d, 0
ldr d1, [x0]
ins v0.d[0], v1.d[0]
ret
but now we can generate:
construct_lanedi:
ldr d0, [x0]
ret
construct_lanedf:
ldr d0, [x0]
ret
Bootstrapped and tested on aarch64-none-linux-gnu.
* simplify-rtx.c (simplify_ternary_operation, VEC_MERGE):
Simplify vec_merge of vec_duplicate and const_vector.
* config/aarch64/predicates.md (aarch64_simd_or_scalar_imm_zero):
New predicate.
* config/aarch64/aarch64-simd.md (*aarch64_combinez<mode>): Use VDC
mode iterator. Update predicate on operand 1 to
handle non-const_vec constants. Delete constraints.
(*aarch64_combinez_be<mode>): Likewise for operand 2.
* gcc.target/aarch64/construct_lane_zero_1.c: New test.
From-SVN: r254548
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* lib-xref.ads, lib-xref-spark_specific.adb
(Traverse_Compilation_Unit): Move declaration to package body.
2017-11-08 Hristian Kirtchev <kirtchev@adacore.com>
* exp_spark.adb (Expand_SPARK_N_Object_Renaming_Declaration): Obtain
the type of the renaming from its defining entity, rather then the
subtype mark as there may not be a subtype mark.
2017-11-08 Jerome Lambourg <lambourg@adacore.com>
* adaint.c, s-oscons-tmplt.c, init.c, libgnat/system-qnx-aarch64.ads,
libgnarl/a-intnam__qnx.ads, libgnarl/s-intman__qnx.adb,
libgnarl/s-osinte__qnx.ads, libgnarl/s-qnx.ads,
libgnarl/s-taprop__qnx.adb, s-oscons-tmplt.c, sigtramp-qnx.c,
terminals.c: Initial port of GNAT for aarch64-qnx
2017-11-08 Elisa Barboni <barboni@adacore.com>
* exp_util.adb (Find_DIC_Type): Move...
* sem_util.ads, sem_util.adb (Find_DIC_Type): ... here.
2017-11-08 Justin Squirek <squirek@adacore.com>
* sem_res.adb (Resolve_Allocator): Add info messages corresponding to
the owner and corresponding coextension.
2017-11-08 Ed Schonberg <schonberg@adacore.com>
* sem_aggr.adb (Resolve_Delta_Aggregate): Divide into the
following separate procedures.
(Resolve_Delta_Array_Aggregate): Previous code form
Resolve_Delta_Aggregate.
(Resolve_Delta_Record_Aggregate): Extend previous code to cover latest
ARG decisions on the legality rules for delta aggregates for records:
in the case of a variant record, components from different variants
cannot be specified in the delta aggregate, and this must be checked
statically.
From-SVN: r254547
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* spark_xrefs.ads (SPARK_Scope_Record): Remove Spec_File_Num and
Spec_Scope_Num components.
* spark_xrefs.adb (dspark): Skip pretty-printing to removed components.
* lib-xref-spark_specific.adb (Add_SPARK_Scope): Skip initialization of
removed components.
(Collect_SPARK_Xrefs): Skip setting proper values of removed
components.
2017-11-08 Gary Dismukes <dismukes@adacore.com>
* exp_ch4.adb (Expand_N_Type_Conversion): Add test that the selector
name is a discriminant in check for unconditional accessibility
violation within instances.
From-SVN: r254545
gcc/ada/
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* lib-xref-spark_specific.adb (Add_SPARK_Xrefs): Remove special-case
for constants (with variable input).
(Is_Constant_Object_Without_Variable_Input): Remove.
2017-11-08 Hristian Kirtchev <kirtchev@adacore.com>
* exp_ch9.adb, sem_disp.adb, sem_util.adb: Minor reformatting.
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* spark_xrefs.ads (Rtype): Remove special-casing of constants for SPARK
cross-references.
(dspark): Remove hardcoded table bound.
2017-11-08 Ed Schonberg <schonberg@adacore.com>
* sem_ch4.adb (Analyze_Aggregate): For Ada2020 delta aggregates, use
the type of the base of the construct to determine the type (or
candidate interpretations) of the delta aggregate. This allows the
construct to appear in a context that expects a private extension.
* sem_res.adb (Resolve): Handle properly a delta aggregate with an
overloaded base.
gcc/testsuite/
2017-11-08 Ed Schonberg <schonberg@adacore.com>
* gnat.dg/delta_aggr.adb: New testcase.
From-SVN: r254544
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* spark_xrefs.ads (SPARK_Xref_Record): Replace file and scope indices
with Entity_Id of the reference.
* spark_xrefs.adb (dspark): Adapt pretty-printing routine.
* lib-xref-spark_specific.adb (Add_SPARK_Xrefs): Store Entity_Id of the
reference, not the file and scope indices.
2017-11-08 Arnaud Charlet <charlet@adacore.com>
* errout.ads (Current_Node): New.
* errout.adb (Error_Msg): Use Current_Node.
* par-ch6.adb, par-ch7.adb, par-ch9.adb, par-util.adb: Set Current_Node
when relevant.
* style.adb: Call Error_Msg_N when possible.
From-SVN: r254543
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* spark_xrefs.ads (SPARK_Scope_Record): Rename Scope_Id component to
Entity.
* lib-xref-spark_specific.adb, spark_xrefs.adb: Propagate renaming of
the Scope_Id record component.
From-SVN: r254542
2017-11-08 Piotr Trojanek <trojanek@adacore.com>
* spark_xrefs.ads (SPARK_Xref_Record): Referenced object is now
represented by Entity_Id.
(SPARK_Scope_Record): Referenced scope (e.g. subprogram) is now
represented by Entity_Id; this information is not repeated as
Scope_Entity.
(Heap): Moved from lib-xref-spark_specific.adb, to reside next to
Name_Of_Heap_Variable.
* spark_xrefs.adb (dspark): Adapt debug routine to above changes in
data types.
* lib-xref-spark_specific.adb: Adapt routines for populating SPARK
scope and xrefs tables to above changes in data types.
2017-11-08 Justin Squirek <squirek@adacore.com>
* sem_ch8.adb (Mark_Use_Clauses): Add condition to always mark the
primitives of generic actuals.
(Mark_Use_Type): Add recursive call to properly mark class-wide type's
base type clauses as per ARM 8.4 (8.2/3).
2017-11-08 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Analyze_Generic_Subprobram_Body): Validate
categorization dependency of the body, as is done for non-generic
units.
(New_Overloaded_Entity, Visible_Part_Type): Remove linear search
through declarations (Simple optimization, no behavior change).
From-SVN: r254539
