[llvm-branch-commits] [BOLT] Match functions with call graph (PR #98125)
https://github.com/shawbyoung created https://github.com/llvm/llvm-project/pull/98125 Test Plan: tbd ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [flang] [Flang][OpenMP] Add lowering support for DISTRIBUTE SIMD (PR #97819)
https://github.com/skatrak updated
https://github.com/llvm/llvm-project/pull/97819
>From a61b3069a400f43eaa4f71e443b8a85c90aea0f0 Mon Sep 17 00:00:00 2001
From: Sergio Afonso
Date: Fri, 5 Jul 2024 12:49:46 +0100
Subject: [PATCH] [Flang][OpenMP] Add lowering support for DISTRIBUTE SIMD
This patch adds support for lowering 'DISTRIBUTE SIMD' constructs to MLIR.
Translation of `omp.distribute` operations to LLVM IR is still not supported,
so its composition with `omp.simd` isn't either.
---
flang/lib/Lower/OpenMP/OpenMP.cpp | 38 +++-
flang/test/Lower/OpenMP/distribute-simd.f90 | 59 ++
flang/test/Lower/OpenMP/if-clause.f90 | 218 +++-
flang/test/Lower/OpenMP/loop-compound.f90 | 45 +++-
4 files changed, 351 insertions(+), 9 deletions(-)
create mode 100644 flang/test/Lower/OpenMP/distribute-simd.f90
diff --git a/flang/lib/Lower/OpenMP/OpenMP.cpp
b/flang/lib/Lower/OpenMP/OpenMP.cpp
index b1bb4c11f86dd..dffdb834d4e66 100644
--- a/flang/lib/Lower/OpenMP/OpenMP.cpp
+++ b/flang/lib/Lower/OpenMP/OpenMP.cpp
@@ -1976,7 +1976,43 @@ static void genCompositeDistributeSimd(
semantics::SemanticsContext &semaCtx, lower::pft::Evaluation &eval,
mlir::Location loc, const ConstructQueue &queue,
ConstructQueue::iterator item, DataSharingProcessor &dsp) {
- TODO(loc, "Composite DISTRIBUTE SIMD");
+ lower::StatementContext stmtCtx;
+
+ // Clause processing.
+ mlir::omp::DistributeClauseOps distributeClauseOps;
+ genDistributeClauses(converter, semaCtx, stmtCtx, item->clauses, loc,
+ distributeClauseOps);
+
+ mlir::omp::SimdClauseOps simdClauseOps;
+ genSimdClauses(converter, semaCtx, item->clauses, loc, simdClauseOps);
+
+ mlir::omp::LoopNestClauseOps loopNestClauseOps;
+ llvm::SmallVector iv;
+ genLoopNestClauses(converter, semaCtx, eval, item->clauses, loc,
+ loopNestClauseOps, iv);
+
+ // Operation creation.
+ // TODO: Populate entry block arguments with private variables.
+ auto distributeOp = genWrapperOp(
+ converter, loc, distributeClauseOps, /*blockArgTypes=*/{});
+
+ // TODO: Populate entry block arguments with reduction and private variables.
+ auto simdOp = genWrapperOp(converter, loc, simdClauseOps,
+/*blockArgTypes=*/{});
+
+ // Construct wrapper entry block list and associated symbols. It is important
+ // that the symbol order and the block argument order match, so that the
+ // symbol-value bindings created are correct.
+ // TODO: Add omp.distribute private and omp.simd private and reduction args.
+ auto wrapperArgs = llvm::to_vector(
+
llvm::concat(distributeOp.getRegion().getArguments(),
+simdOp.getRegion().getArguments()));
+
+ assert(wrapperArgs.empty() &&
+ "Block args for omp.simd and omp.distribute currently not expected");
+ genLoopNestOp(converter, symTable, semaCtx, eval, loc, queue, item,
+loopNestClauseOps, iv, /*wrapperSyms=*/{}, wrapperArgs,
+llvm::omp::Directive::OMPD_distribute_simd, dsp);
}
static void genCompositeDoSimd(lower::AbstractConverter &converter,
diff --git a/flang/test/Lower/OpenMP/distribute-simd.f90
b/flang/test/Lower/OpenMP/distribute-simd.f90
new file mode 100644
index 0..545037b7791b8
--- /dev/null
+++ b/flang/test/Lower/OpenMP/distribute-simd.f90
@@ -0,0 +1,59 @@
+! This test checks lowering of OpenMP DISTRIBUTE SIMD composite constructs.
+
+! RUN: bbc -fopenmp -emit-hlfir %s -o - | FileCheck %s
+! RUN: %flang_fc1 -fopenmp -emit-hlfir %s -o - | FileCheck %s
+
+! CHECK-LABEL: func.func @_QPdistribute_simd_aligned(
+subroutine distribute_simd_aligned(A)
+ use iso_c_binding
+ type(c_ptr) :: A
+
+ !$omp teams
+
+ ! CHECK: omp.distribute
+ ! CHECK-NOT: aligned({{.*}})
+ ! CHECK-SAME: {
+ ! CHECK-NEXT: omp.simd
+ ! CHECK-SAME: aligned({{.*}})
+ !$omp distribute simd aligned(A)
+ do index_ = 1, 10
+call c_test_call(A)
+ end do
+ !$omp end distribute simd
+
+ !$omp end teams
+end subroutine distribute_simd_aligned
+
+! CHECK-LABEL: func.func @_QPdistribute_simd_safelen(
+subroutine distribute_simd_safelen()
+ !$omp teams
+
+ ! CHECK: omp.distribute
+ ! CHECK-NOT: safelen({{.*}})
+ ! CHECK-SAME: {
+ ! CHECK-NEXT: omp.simd
+ ! CHECK-SAME: safelen({{.*}})
+ !$omp distribute simd safelen(4)
+ do index_ = 1, 10
+ end do
+ !$omp end distribute simd
+
+ !$omp end teams
+end subroutine distribute_simd_safelen
+
+! CHECK-LABEL: func.func @_QPdistribute_simd_simdlen(
+subroutine distribute_simd_simdlen()
+ !$omp teams
+
+ ! CHECK: omp.distribute
+ ! CHECK-NOT: simdlen({{.*}})
+ ! CHECK-SAME: {
+ ! CHECK-NEXT: omp.simd
+ ! CHECK-SAME: simdlen({{.*}})
+ !$omp distribute simd simdlen(4)
+ do index_ = 1, 10
+ end do
+ !$omp end distribute simd
+
+ !$omp end teams
+end subroutine distribute_simd_simdlen
diff --git a/flang/test/Lower/OpenMP/if-clause.f90
b/flang/tes
[llvm-branch-commits] [llvm] [BOLT] Added more details on heatmap docs. (PR #98162)
https://github.com/paschalis-mpeis created https://github.com/llvm/llvm-project/pull/98162 Suggesting a few more details for Heatmaps.md >From f209cca87cf7c53242a353a505e3bfe34688a1b2 Mon Sep 17 00:00:00 2001 From: Paschalis Mpeis Date: Tue, 9 Jul 2024 08:52:51 +0100 Subject: [PATCH] [BOLT] Added more details on heatmap docs. --- bolt/docs/Heatmaps.md | 53 +++ 1 file changed, 39 insertions(+), 14 deletions(-) diff --git a/bolt/docs/Heatmaps.md b/bolt/docs/Heatmaps.md index e1b59d49ad102..c1a01839d6682 100644 --- a/bolt/docs/Heatmaps.md +++ b/bolt/docs/Heatmaps.md @@ -1,9 +1,9 @@ # Code Heatmaps BOLT has gained the ability to print code heatmaps based on -sampling-based LBR profiles generated by `perf`. The output is produced -in colored ASCII to be displayed in a color-capable terminal. It looks -something like this: +sampling-based profiles generated by `perf`, either with `LBR` data or not. +The output is produced in colored ASCII to be displayed in a color-capable +terminal. It looks something like this:  @@ -32,17 +32,7 @@ $ llvm-bolt-heatmap -p perf.data ``` By default the heatmap will be dumped to *stdout*. You can change it -with `-o ` option. Each character/block in the heatmap -shows the execution data accumulated for corresponding 64 bytes of -code. You can change this granularity with a `-block-size` option. -E.g. set it to 4096 to see code usage grouped by 4K pages. -Other useful options are: - -```bash --line-size= - number of entries per line (default 256) --max-address= - maximum address considered valid for heatmap (default 4GB) --print-mappings= - print mappings in legend, between characters/blocks and text sections (default false) -``` +with `-o ` option. If you prefer to look at the data in a browser (or would like to share it that way), then you can use an HTML conversion tool. E.g.: @@ -50,3 +40,38 @@ it that way), then you can use an HTML conversion tool. E.g.: ```bash $ aha -b -f > .html ``` + +--- + +## Background on heatmaps: +A heatmap is effectively a histogram that is rendered into a grid for better +visualization. +In theory we can generate a heatmap using any binary and a perf profile. + +Each block/character in the heatmap shows the execution data accumulated for +corresponding 64 bytes of code. You can change this granularity with a +`-block-size` option. +E.g. set it to 4096 to see code usage grouped by 4K pages. + + +When a block is shown as a dot, it means that no samples were found for that address. +When it is shown as a letter, it indicates a captured sample on a particular text section of the binary. To show a mapping between letters and text sections in the legend, use `-print-mappings`. When a sampled address does not belong to any of the TextSegments, the characters 'o' or 'O' will be shown. + +The legend shows by default the ranges in the heatmap according to the number +of samples per block. +A color is assigned per range, except the first two ranges that distinguished by +lower and upper case letters. + +Each consecutive line in the heatmap advances by the same amount, +with the binary size covered by a line being dependent on the block size and the +line size. +An empty new line is inserted for bigger gaps between samples. + + +Some useful options are: + +``` +-line-size= - number of entries per line (default 256) +-max-address= - maximum address considered valid for heatmap (default 4GB) +-print-mappings - print mappings in the legend, between characters/blocks and text sections (default false) +``` ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Added more details on heatmap docs. (PR #98162)
llvmbot wrote: @llvm/pr-subscribers-bolt Author: Paschalis Mpeis (paschalis-mpeis) Changes Suggesting a few more details for Heatmaps.md --- Full diff: https://github.com/llvm/llvm-project/pull/98162.diff 1 Files Affected: - (modified) bolt/docs/Heatmaps.md (+39-14) ``diff diff --git a/bolt/docs/Heatmaps.md b/bolt/docs/Heatmaps.md index e1b59d49ad102..c1a01839d6682 100644 --- a/bolt/docs/Heatmaps.md +++ b/bolt/docs/Heatmaps.md @@ -1,9 +1,9 @@ # Code Heatmaps BOLT has gained the ability to print code heatmaps based on -sampling-based LBR profiles generated by `perf`. The output is produced -in colored ASCII to be displayed in a color-capable terminal. It looks -something like this: +sampling-based profiles generated by `perf`, either with `LBR` data or not. +The output is produced in colored ASCII to be displayed in a color-capable +terminal. It looks something like this:  @@ -32,17 +32,7 @@ $ llvm-bolt-heatmap -p perf.data ``` By default the heatmap will be dumped to *stdout*. You can change it -with `-o ` option. Each character/block in the heatmap -shows the execution data accumulated for corresponding 64 bytes of -code. You can change this granularity with a `-block-size` option. -E.g. set it to 4096 to see code usage grouped by 4K pages. -Other useful options are: - -```bash --line-size= - number of entries per line (default 256) --max-address= - maximum address considered valid for heatmap (default 4GB) --print-mappings= - print mappings in legend, between characters/blocks and text sections (default false) -``` +with `-o ` option. If you prefer to look at the data in a browser (or would like to share it that way), then you can use an HTML conversion tool. E.g.: @@ -50,3 +40,38 @@ it that way), then you can use an HTML conversion tool. E.g.: ```bash $ aha -b -f > .html ``` + +--- + +## Background on heatmaps: +A heatmap is effectively a histogram that is rendered into a grid for better +visualization. +In theory we can generate a heatmap using any binary and a perf profile. + +Each block/character in the heatmap shows the execution data accumulated for +corresponding 64 bytes of code. You can change this granularity with a +`-block-size` option. +E.g. set it to 4096 to see code usage grouped by 4K pages. + + +When a block is shown as a dot, it means that no samples were found for that address. +When it is shown as a letter, it indicates a captured sample on a particular text section of the binary. To show a mapping between letters and text sections in the legend, use `-print-mappings`. When a sampled address does not belong to any of the TextSegments, the characters 'o' or 'O' will be shown. + +The legend shows by default the ranges in the heatmap according to the number +of samples per block. +A color is assigned per range, except the first two ranges that distinguished by +lower and upper case letters. + +Each consecutive line in the heatmap advances by the same amount, +with the binary size covered by a line being dependent on the block size and the +line size. +An empty new line is inserted for bigger gaps between samples. + + +Some useful options are: + +``` +-line-size= - number of entries per line (default 256) +-max-address= - maximum address considered valid for heatmap (default 4GB) +-print-mappings - print mappings in the legend, between characters/blocks and text sections (default false) +``` `` https://github.com/llvm/llvm-project/pull/98162 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time constructors. (PR #95010)
https://github.com/ldionne milestoned https://github.com/llvm/llvm-project/pull/95010 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
https://github.com/ldionne edited https://github.com/llvm/llvm-project/pull/95026 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -0,0 +1,136 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+// UNSUPPORTED: no-filesystem, no-localization, no-tzdb
+
+// XFAIL: libcpp-has-no-experimental-tzdb
+// XFAIL: availability-tzdb-missing
+
+//
+
+// template
+// class zoned_time;
+//
+// zoned_time& operator=(const sys_time& st);
+
+#include
+#include
+#include
+#include
+
+namespace cr = std::chrono;
+
+int main(int, char**) {
+ {
+using duration = cr::nanoseconds;
+using time_point = cr::sys_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("UTC"));
+assert(zt.get_sys_time() == time_point{duration{42}});
+
+[[maybe_unused]] std::same_as decltype(auto) _ = zt =
time_point{duration{99}};
ldionne wrote:
Same comment here!
https://github.com/llvm/llvm-project/pull/95026
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -0,0 +1,135 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+// UNSUPPORTED: no-filesystem, no-localization, no-tzdb
+
+// XFAIL: libcpp-has-no-experimental-tzdb
+// XFAIL: availability-tzdb-missing
+
+//
+
+// template
+// class zoned_time;
+//
+// explicit operator local_time() const;
+
+#include
+#include
+
+#include "../test_offset_time_zone.h"
+
+namespace cr = std::chrono;
+
+static void test_const_member() {
+ {
+using duration = cr::nanoseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+static_assert(!std::is_convertible_v>);
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
ldionne wrote:
Same comment.
https://github.com/llvm/llvm-project/pull/95026
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
https://github.com/ldionne approved this pull request. LGTM with a few suggestions! https://github.com/llvm/llvm-project/pull/95026 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -0,0 +1,243 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+// UNSUPPORTED: no-filesystem, no-localization, no-tzdb
+
+// XFAIL: libcpp-has-no-experimental-tzdb
+// XFAIL: availability-tzdb-missing
+
+//
+
+// template
+// class zoned_time;
+//
+// zoned_time& operator=(const local_time& st);
+
+// TODO TZDB Investigate the issues in this test, this seems like
+// a design issue of the class.
+//
+// [time.zone.zonedtime.members]/3
+// Effects: After assignment, get_local_time() == lt.
+// This assignment has no effect on the return value of get_time_zone().
+//
+// The test cases describe the issues.
+
+#include
+#include
+#include
+#include
+
+#include "test_macros.h"
+
+namespace cr = std::chrono;
+
+// Tests unique conversions. To make sure the test is does not depend on
changes
+// in the database it uses a time zone with a fixed offset.
+static void test_unique() {
+ // common_type_t -> duration
+ {
+using duration = cr::nanoseconds;
+using sys_time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{"Etc/GMT+1", sys_time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{42}});
+assert(zt.get_local_time() == local_time_point{duration{42} -
cr::hours{1}});
+
+[[maybe_unused]] std::same_as decltype(auto) _ = zt =
local_time_point{duration{99}};
ldionne wrote:
```suggestion
std::same_as decltype(auto) result = zt =
local_time_point{duration{99}};
assert(&result == &zt);
```
Kinda pedantic but necessary!
https://github.com/llvm/llvm-project/pull/95026
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -0,0 +1,133 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+// UNSUPPORTED: no-filesystem, no-localization, no-tzdb
+
+// XFAIL: libcpp-has-no-experimental-tzdb
+// XFAIL: availability-tzdb-missing
+
+//
+
+// template
+// class zoned_time;
+//
+// local_time get_local_time() const;
+
+#include
+#include
+
+#include "../test_offset_time_zone.h"
+
+namespace cr = std::chrono;
+
+static void test_const_member() {
+ {
+using duration = cr::nanoseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
ldionne wrote:
This test seems out of place, since the function name suggests that we're only
testing const-qualification in it.
https://github.com/llvm/llvm-project/pull/95026
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -0,0 +1,243 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+// UNSUPPORTED: no-filesystem, no-localization, no-tzdb
+
+// XFAIL: libcpp-has-no-experimental-tzdb
+// XFAIL: availability-tzdb-missing
+
+//
+
+// template
+// class zoned_time;
+//
+// zoned_time& operator=(const local_time& st);
+
+// TODO TZDB Investigate the issues in this test, this seems like
+// a design issue of the class.
+//
+// [time.zone.zonedtime.members]/3
+// Effects: After assignment, get_local_time() == lt.
+// This assignment has no effect on the return value of get_time_zone().
+//
+// The test cases describe the issues.
+
+#include
+#include
+#include
+#include
+
+#include "test_macros.h"
+
+namespace cr = std::chrono;
+
+// Tests unique conversions. To make sure the test is does not depend on
changes
+// in the database it uses a time zone with a fixed offset.
+static void test_unique() {
+ // common_type_t -> duration
+ {
+using duration = cr::nanoseconds;
+using sys_time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{"Etc/GMT+1", sys_time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{42}});
+assert(zt.get_local_time() == local_time_point{duration{42} -
cr::hours{1}});
+
+[[maybe_unused]] std::same_as decltype(auto) _ = zt =
local_time_point{duration{99}};
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{99} + cr::hours{1}});
+assert(zt.get_local_time() == local_time_point{duration{99}});
+ }
+ {
+using duration = cr::microseconds;
+using sys_time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{"Etc/GMT+1", sys_time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{42}});
+assert(zt.get_local_time() == local_time_point{duration{42} -
cr::hours{1}});
+
+[[maybe_unused]] std::same_as decltype(auto) _ = zt =
local_time_point{duration{99}};
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{99} + cr::hours{1}});
+assert(zt.get_local_time() == local_time_point{duration{99}});
+ }
+ {
+using duration = cr::milliseconds;
+using sys_time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{"Etc/GMT+1", sys_time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{42}});
+assert(zt.get_local_time() == local_time_point{duration{42} -
cr::hours{1}});
+
+[[maybe_unused]] std::same_as decltype(auto) _ = zt =
local_time_point{duration{99}};
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{99} + cr::hours{1}});
+assert(zt.get_local_time() == local_time_point{duration{99}});
+ }
+ // common_type_t -> seconds
+ {
+using duration = cr::seconds;
+using sys_time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{"Etc/GMT+1", sys_time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{42}});
+assert(zt.get_local_time() == local_time_point{duration{42} -
cr::hours{1}});
+
+[[maybe_unused]] std::same_as decltype(auto) _ = zt =
local_time_point{duration{99}};
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == sys_time_point{duration{99} + cr::hours{1}});
+assert(zt.get_local_time() == local_time_point{duration{99}});
+ }
+ // common_type_t -> seconds
+ {
+using duration = cr::days;
+using sys_time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+using zoned_time = cr::zoned_time;
+zoned_time zt{"Etc/GMT+1", sys_time_point{duration{42}}};
+
+assert(zt.get_time_zone() == cr::locate_zone("Etc/GMT+1"));
+assert(zt.get_sys_time() == cr::sys_seconds{duration{42}});
+assert(zt.get_local_time() == cr::local_seconds{duration{42} -
cr::hours{1}});
+
+[[maybe_unused]] std::same_as decltyp
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -0,0 +1,135 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+// UNSUPPORTED: no-filesystem, no-localization, no-tzdb
+
+// XFAIL: libcpp-has-no-experimental-tzdb
+// XFAIL: availability-tzdb-missing
+
+//
+
+// template
+// class zoned_time;
+//
+// explicit operator local_time() const;
+
+#include
+#include
+
+#include "../test_offset_time_zone.h"
+
+namespace cr = std::chrono;
+
+static void test_const_member() {
+ {
+using duration = cr::nanoseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+static_assert(!std::is_convertible_v>);
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ {
+using duration = cr::nanoseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+static_assert(!std::is_convertible_v>);
+const cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+}
+
+static void test_duration_conversion() {
+ // common_type_t -> duration
+ {
+using duration = cr::nanoseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ {
+using duration = cr::microseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ {
+using duration = cr::milliseconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ // common_type_t -> seconds
+ {
+using duration = cr::seconds;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ // common_type_t -> seconds
+ {
+using duration = cr::days;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ {
+using duration = cr::weeks;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ {
+using duration = cr::months;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
+
+std::same_as decltype(auto) time =
static_cast(zt);
+assert(time == local_time_point{duration{42} - cr::hours{1}});
+ }
+ {
+using duration = cr::years;
+using time_point = cr::sys_time;
+using local_time_point = cr::local_time;
+cr::zoned_time zt{"Etc/GMT+1", time_point{duration{42}}};
ldionne wrote:
Note that you could make all of those `zt`'s `const` since the method is
`const`.
https://github.com/llvm/llvm-project/pull/95026
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Finishes zoned_time member functions. (PR #95026)
@@ -147,8 +148,29 @@ class zoned_time {
} && is_convertible_v, sys_time<_Duration>>)
: zoned_time{__traits::locate_zone(__name), __zt, __c} {}
+ _LIBCPP_HIDE_FROM_ABI zoned_time& operator=(const sys_time<_Duration>& __tp)
{
+__tp_ = __tp;
+return *this;
+ }
+
+ _LIBCPP_HIDE_FROM_ABI zoned_time& operator=(const local_time<_Duration>&
__tp) {
+// TODO TZDB This seems wrong.
+// Assigning a non-existant or abiguous time will throw and not satisfy
ldionne wrote:
```suggestion
// Assigning a non-existent or ambiguous time will throw and not satisfy
```
Can you file a LWG issue about this?
https://github.com/llvm/llvm-project/pull/95026
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Adds zoned_time deduction guides. (PR #95139)
https://github.com/ldionne milestoned https://github.com/llvm/llvm-project/pull/95139 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Adds zoned_time deduction guides. (PR #95139)
https://github.com/ldionne approved this pull request. https://github.com/llvm/llvm-project/pull/95139 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Implements zoned_time's operator==. (PR #95140)
https://github.com/ldionne milestoned https://github.com/llvm/llvm-project/pull/95140 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Implements zoned_time's operator==. (PR #95140)
@@ -0,0 +1,71 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+
+// REQUIRES: has-unix-headers
+// REQUIRES: libcpp-hardening-mode={{extensive|debug}}
+// XFAIL: libcpp-hardening-mode=debug && availability-verbose_abort-missing
ldionne wrote:
This looks like a leftover?
https://github.com/llvm/llvm-project/pull/95140
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Implements zoned_time's operator==. (PR #95140)
https://github.com/ldionne edited https://github.com/llvm/llvm-project/pull/95140 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Implements zoned_time's operator==. (PR #95140)
https://github.com/ldionne approved this pull request. https://github.com/llvm/llvm-project/pull/95140 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Implements zoned_seconds typedef. (PR #95141)
https://github.com/ldionne milestoned https://github.com/llvm/llvm-project/pull/95141 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libcxx] [libc++][TZDB] Implements zoned_seconds typedef. (PR #95141)
https://github.com/ldionne approved this pull request. https://github.com/llvm/llvm-project/pull/95141 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] b685845 - Revert "[lit] Implement builtin umask (#94621)"
Author: Jay Foad
Date: 2024-07-09T17:17:57+01:00
New Revision: b68584571de3370ba7655b62047908b28a81d56c
URL:
https://github.com/llvm/llvm-project/commit/b68584571de3370ba7655b62047908b28a81d56c
DIFF:
https://github.com/llvm/llvm-project/commit/b68584571de3370ba7655b62047908b28a81d56c.diff
LOG: Revert "[lit] Implement builtin umask (#94621)"
This reverts commit 9f6dd1f761a5121e9a69e5d1f67c2438c065b499.
Added:
Modified:
llvm/test/tools/llvm-dwarfutil/ELF/X86/mirror-permissions-unix.test
llvm/test/tools/llvm-objcopy/ELF/mirror-permissions-unix.test
llvm/test/tools/llvm-objcopy/ELF/respect-umask.test
llvm/utils/lit/lit/TestRunner.py
Removed:
llvm/utils/lit/tests/Inputs/shtest-umask/lit.cfg
llvm/utils/lit/tests/Inputs/shtest-umask/umask-bad-arg.txt
llvm/utils/lit/tests/Inputs/shtest-umask/umask-ok.txt
llvm/utils/lit/tests/Inputs/shtest-umask/umask-too-many-args.txt
llvm/utils/lit/tests/shtest-umask.py
diff --git
a/llvm/test/tools/llvm-dwarfutil/ELF/X86/mirror-permissions-unix.test
b/llvm/test/tools/llvm-dwarfutil/ELF/X86/mirror-permissions-unix.test
index fdcba4dcd666b..a95d1c0aafa21 100644
--- a/llvm/test/tools/llvm-dwarfutil/ELF/X86/mirror-permissions-unix.test
+++ b/llvm/test/tools/llvm-dwarfutil/ELF/X86/mirror-permissions-unix.test
@@ -3,6 +3,7 @@
## Setting the umask to 0 ensures deterministic permissions across
## test environments.
# UNSUPPORTED: system-windows
+# REQUIRES: shell
# RUN: touch %t
# RUN: chmod 0777 %t
diff --git a/llvm/test/tools/llvm-objcopy/ELF/mirror-permissions-unix.test
b/llvm/test/tools/llvm-objcopy/ELF/mirror-permissions-unix.test
index 66a481a2230d1..8f4993f4f3d29 100644
--- a/llvm/test/tools/llvm-objcopy/ELF/mirror-permissions-unix.test
+++ b/llvm/test/tools/llvm-objcopy/ELF/mirror-permissions-unix.test
@@ -6,6 +6,7 @@
## Setting the umask to 0 ensures deterministic permissions across
## test environments.
# UNSUPPORTED: system-windows
+# REQUIRES: shell
# RUN: touch %t
# RUN: chmod 0777 %t
diff --git a/llvm/test/tools/llvm-objcopy/ELF/respect-umask.test
b/llvm/test/tools/llvm-objcopy/ELF/respect-umask.test
index 02e9b93f5376f..376e33a217819 100644
--- a/llvm/test/tools/llvm-objcopy/ELF/respect-umask.test
+++ b/llvm/test/tools/llvm-objcopy/ELF/respect-umask.test
@@ -1,8 +1,10 @@
## This tests that the umask is respected when
## assigning permissions of output files.
-## Windows has no umask so this test makes no sense.
+## Windows has no umask so this test makes no sense, nor would
+## it work because there is no umask(1) in a Windows environment
# UNSUPPORTED: system-windows
+# REQUIRES: shell
# RUN: rm -f %t
# RUN: touch %t
diff --git a/llvm/utils/lit/lit/TestRunner.py
b/llvm/utils/lit/lit/TestRunner.py
index 865376e0e711e..da7fa86fd3917 100644
--- a/llvm/utils/lit/lit/TestRunner.py
+++ b/llvm/utils/lit/lit/TestRunner.py
@@ -87,10 +87,9 @@ class ShellEnvironment(object):
we maintain a dir stack for pushd/popd.
"""
-def __init__(self, cwd, env, umask=-1):
+def __init__(self, cwd, env):
self.cwd = cwd
self.env = dict(env)
-self.umask = umask
self.dirStack = []
def change_dir(self, newdir):
@@ -566,20 +565,6 @@ class SHFILEOPSTRUCTW(Structure):
return ShellCommandResult(cmd, "", stderr.getvalue(), exitCode, False)
-def executeBuiltinUmask(cmd, shenv):
-"""executeBuiltinUmask - Change the current umask."""
-if os.name != "posix":
-raise InternalShellError(cmd, "'umask' not supported on this system")
-if len(cmd.args) != 2:
-raise InternalShellError(cmd, "'umask' supports only one argument")
-try:
-# Update the umask in the parent environment.
-shenv.umask = int(cmd.args[1], 8)
-except ValueError as err:
-raise InternalShellError(cmd, "Error: 'umask': %s" % str(err))
-return ShellCommandResult(cmd, "", "", 0, False)
-
-
def executeBuiltinColon(cmd, cmd_shenv):
"""executeBuiltinColon - Discard arguments and exit with status 0."""
return ShellCommandResult(cmd, "", "", 0, False)
@@ -734,7 +719,6 @@ def _executeShCmd(cmd, shenv, results, timeoutHelper):
"popd": executeBuiltinPopd,
"pushd": executeBuiltinPushd,
"rm": executeBuiltinRm,
-"umask": executeBuiltinUmask,
":": executeBuiltinColon,
}
# To avoid deadlock, we use a single stderr stream for piped
@@ -756,7 +740,7 @@ def _executeShCmd(cmd, shenv, results, timeoutHelper):
# env FOO=1 llc < %s | env BAR=2 llvm-mc | FileCheck %s
# env FOO=1 %{another_env_plus_cmd} | FileCheck %s
if cmd_shenv is shenv:
-cmd_shenv = ShellEnvironment(shenv.cwd, shenv.env,
shenv.umask)
+cmd_shenv = ShellEnvironment(shenv.cwd, shenv.env)
args = updateEnv(cmd_shenv, args)
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
https://github.com/shawbyoung updated
https://github.com/llvm/llvm-project/pull/96596
>From 05d59574d6260b98a469921eb2fccf5398bfafb6 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Mon, 24 Jun 2024 23:00:59 -0700
Subject: [PATCH 01/16] Added call to matchWithCallsAsAnchors
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 3 +++
1 file changed, 3 insertions(+)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index aafffac3d4b1c..1a0e5d239d252 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -479,6 +479,9 @@ Error YAMLProfileReader::readProfile(BinaryContext &BC) {
if (!YamlBF.Used && BF && !ProfiledFunctions.count(BF))
matchProfileToFunction(YamlBF, *BF);
+ uint64_t MatchedWithCallsAsAnchors = 0;
+ matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
+
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
if (!YamlBF.Used && opts::Verbosity >= 1)
errs() << "BOLT-WARNING: profile ignored for function " << YamlBF.Name
>From 77ef0008f4f5987719555e6cc3e32da812ae0f31 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Mon, 24 Jun 2024 23:11:43 -0700
Subject: [PATCH 02/16] Changed CallHashToBF representation
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 1a0e5d239d252..91b01a99c7485 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -29,6 +29,10 @@ static llvm::cl::opt
cl::desc("ignore hash while reading function profile"),
cl::Hidden, cl::cat(BoltOptCategory));
+llvm::cl::opt MatchWithCallsAsAnchors("match-with-calls-as-anchors",
+ cl::desc("Matches with calls as anchors"),
+ cl::Hidden, cl::cat(BoltOptCategory));
+
llvm::cl::opt ProfileUseDFS("profile-use-dfs",
cl::desc("use DFS order for YAML profile"),
cl::Hidden, cl::cat(BoltOptCategory));
@@ -353,7 +357,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
llvm_unreachable("Unhandled HashFunction");
};
- std::unordered_map CallHashToBF;
+ std::unordered_map CallHashToBF;
for (BinaryFunction *BF : BC.getAllBinaryFunctions()) {
if (ProfiledFunctions.count(BF))
@@ -375,12 +379,12 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
for (const std::string &FunctionName : FunctionNames)
HashString.append(FunctionName);
}
-CallHashToBF.emplace(ComputeCallHash(HashString), BF);
+CallHashToBF[ComputeCallHash(HashString)] = BF;
}
std::unordered_map ProfiledFunctionIdToName;
- for (const yaml::bolt::BinaryFunctionProfile YamlBF : YamlBP.Functions)
+ for (const yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
ProfiledFunctionIdToName[YamlBF.Id] = YamlBF.Name;
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions) {
@@ -401,7 +405,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
auto It = CallHashToBF.find(Hash);
if (It == CallHashToBF.end())
continue;
-matchProfileToFunction(YamlBF, It->second);
+matchProfileToFunction(YamlBF, *It->second);
++MatchedWithCallsAsAnchors;
}
}
@@ -480,7 +484,8 @@ Error YAMLProfileReader::readProfile(BinaryContext &BC) {
matchProfileToFunction(YamlBF, *BF);
uint64_t MatchedWithCallsAsAnchors = 0;
- matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
+ if (opts::MatchWithCallsAsAnchors)
+matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
if (!YamlBF.Used && opts::Verbosity >= 1)
>From ea7cb68ab9e8e158412c2e752986968968a60d93 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 25 Jun 2024 09:28:39 -0700
Subject: [PATCH 03/16] Changed BF called FunctionNames to multiset
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 91b01a99c7485..3b3d73f7af023 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -365,7 +365,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
std::string HashString;
for (const auto &BB : BF->blocks()) {
- std::set FunctionNames;
+ std::multiset FunctionNames;
for (const MCInst &Instr : BB) {
// Skip non-call instructions.
if (!BC.MIB->isCall(Instr))
@@ -397,9 +397,8 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
std::string &FunctionName = ProfiledFunctionIdToName[CallSite.DestId];
FunctionNames.insert(FunctionName);
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
https://github.com/shawbyoung updated
https://github.com/llvm/llvm-project/pull/96596
>From 05d59574d6260b98a469921eb2fccf5398bfafb6 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Mon, 24 Jun 2024 23:00:59 -0700
Subject: [PATCH 01/16] Added call to matchWithCallsAsAnchors
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 3 +++
1 file changed, 3 insertions(+)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index aafffac3d4b1c..1a0e5d239d252 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -479,6 +479,9 @@ Error YAMLProfileReader::readProfile(BinaryContext &BC) {
if (!YamlBF.Used && BF && !ProfiledFunctions.count(BF))
matchProfileToFunction(YamlBF, *BF);
+ uint64_t MatchedWithCallsAsAnchors = 0;
+ matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
+
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
if (!YamlBF.Used && opts::Verbosity >= 1)
errs() << "BOLT-WARNING: profile ignored for function " << YamlBF.Name
>From 77ef0008f4f5987719555e6cc3e32da812ae0f31 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Mon, 24 Jun 2024 23:11:43 -0700
Subject: [PATCH 02/16] Changed CallHashToBF representation
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 1a0e5d239d252..91b01a99c7485 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -29,6 +29,10 @@ static llvm::cl::opt
cl::desc("ignore hash while reading function profile"),
cl::Hidden, cl::cat(BoltOptCategory));
+llvm::cl::opt MatchWithCallsAsAnchors("match-with-calls-as-anchors",
+ cl::desc("Matches with calls as anchors"),
+ cl::Hidden, cl::cat(BoltOptCategory));
+
llvm::cl::opt ProfileUseDFS("profile-use-dfs",
cl::desc("use DFS order for YAML profile"),
cl::Hidden, cl::cat(BoltOptCategory));
@@ -353,7 +357,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
llvm_unreachable("Unhandled HashFunction");
};
- std::unordered_map CallHashToBF;
+ std::unordered_map CallHashToBF;
for (BinaryFunction *BF : BC.getAllBinaryFunctions()) {
if (ProfiledFunctions.count(BF))
@@ -375,12 +379,12 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
for (const std::string &FunctionName : FunctionNames)
HashString.append(FunctionName);
}
-CallHashToBF.emplace(ComputeCallHash(HashString), BF);
+CallHashToBF[ComputeCallHash(HashString)] = BF;
}
std::unordered_map ProfiledFunctionIdToName;
- for (const yaml::bolt::BinaryFunctionProfile YamlBF : YamlBP.Functions)
+ for (const yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
ProfiledFunctionIdToName[YamlBF.Id] = YamlBF.Name;
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions) {
@@ -401,7 +405,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
auto It = CallHashToBF.find(Hash);
if (It == CallHashToBF.end())
continue;
-matchProfileToFunction(YamlBF, It->second);
+matchProfileToFunction(YamlBF, *It->second);
++MatchedWithCallsAsAnchors;
}
}
@@ -480,7 +484,8 @@ Error YAMLProfileReader::readProfile(BinaryContext &BC) {
matchProfileToFunction(YamlBF, *BF);
uint64_t MatchedWithCallsAsAnchors = 0;
- matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
+ if (opts::MatchWithCallsAsAnchors)
+matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
if (!YamlBF.Used && opts::Verbosity >= 1)
>From ea7cb68ab9e8e158412c2e752986968968a60d93 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 25 Jun 2024 09:28:39 -0700
Subject: [PATCH 03/16] Changed BF called FunctionNames to multiset
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 91b01a99c7485..3b3d73f7af023 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -365,7 +365,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
std::string HashString;
for (const auto &BB : BF->blocks()) {
- std::set FunctionNames;
+ std::multiset FunctionNames;
for (const MCInst &Instr : BB) {
// Skip non-call instructions.
if (!BC.MIB->isCall(Instr))
@@ -397,9 +397,8 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
std::string &FunctionName = ProfiledFunctionIdToName[CallSite.DestId];
FunctionNames.insert(FunctionName);
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
https://github.com/shawbyoung updated
https://github.com/llvm/llvm-project/pull/96596
>From 05d59574d6260b98a469921eb2fccf5398bfafb6 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Mon, 24 Jun 2024 23:00:59 -0700
Subject: [PATCH 01/17] Added call to matchWithCallsAsAnchors
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 3 +++
1 file changed, 3 insertions(+)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index aafffac3d4b1c..1a0e5d239d252 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -479,6 +479,9 @@ Error YAMLProfileReader::readProfile(BinaryContext &BC) {
if (!YamlBF.Used && BF && !ProfiledFunctions.count(BF))
matchProfileToFunction(YamlBF, *BF);
+ uint64_t MatchedWithCallsAsAnchors = 0;
+ matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
+
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
if (!YamlBF.Used && opts::Verbosity >= 1)
errs() << "BOLT-WARNING: profile ignored for function " << YamlBF.Name
>From 77ef0008f4f5987719555e6cc3e32da812ae0f31 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Mon, 24 Jun 2024 23:11:43 -0700
Subject: [PATCH 02/17] Changed CallHashToBF representation
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 1a0e5d239d252..91b01a99c7485 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -29,6 +29,10 @@ static llvm::cl::opt
cl::desc("ignore hash while reading function profile"),
cl::Hidden, cl::cat(BoltOptCategory));
+llvm::cl::opt MatchWithCallsAsAnchors("match-with-calls-as-anchors",
+ cl::desc("Matches with calls as anchors"),
+ cl::Hidden, cl::cat(BoltOptCategory));
+
llvm::cl::opt ProfileUseDFS("profile-use-dfs",
cl::desc("use DFS order for YAML profile"),
cl::Hidden, cl::cat(BoltOptCategory));
@@ -353,7 +357,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
llvm_unreachable("Unhandled HashFunction");
};
- std::unordered_map CallHashToBF;
+ std::unordered_map CallHashToBF;
for (BinaryFunction *BF : BC.getAllBinaryFunctions()) {
if (ProfiledFunctions.count(BF))
@@ -375,12 +379,12 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
for (const std::string &FunctionName : FunctionNames)
HashString.append(FunctionName);
}
-CallHashToBF.emplace(ComputeCallHash(HashString), BF);
+CallHashToBF[ComputeCallHash(HashString)] = BF;
}
std::unordered_map ProfiledFunctionIdToName;
- for (const yaml::bolt::BinaryFunctionProfile YamlBF : YamlBP.Functions)
+ for (const yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
ProfiledFunctionIdToName[YamlBF.Id] = YamlBF.Name;
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions) {
@@ -401,7 +405,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
auto It = CallHashToBF.find(Hash);
if (It == CallHashToBF.end())
continue;
-matchProfileToFunction(YamlBF, It->second);
+matchProfileToFunction(YamlBF, *It->second);
++MatchedWithCallsAsAnchors;
}
}
@@ -480,7 +484,8 @@ Error YAMLProfileReader::readProfile(BinaryContext &BC) {
matchProfileToFunction(YamlBF, *BF);
uint64_t MatchedWithCallsAsAnchors = 0;
- matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
+ if (opts::MatchWithCallsAsAnchors)
+matchWithCallsAsAnchors(BC, MatchedWithCallsAsAnchors);
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions)
if (!YamlBF.Used && opts::Verbosity >= 1)
>From ea7cb68ab9e8e158412c2e752986968968a60d93 Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 25 Jun 2024 09:28:39 -0700
Subject: [PATCH 03/17] Changed BF called FunctionNames to multiset
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 91b01a99c7485..3b3d73f7af023 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -365,7 +365,7 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
std::string HashString;
for (const auto &BB : BF->blocks()) {
- std::set FunctionNames;
+ std::multiset FunctionNames;
for (const MCInst &Instr : BB) {
// Skip non-call instructions.
if (!BC.MIB->isCall(Instr))
@@ -397,9 +397,8 @@ void YAMLProfileReader::matchWithCallsAsAnchors(
std::string &FunctionName = ProfiledFunctionIdToName[CallSite.DestId];
FunctionNames.insert(FunctionName);
[llvm-branch-commits] [llvm] [BOLT] Match functions with call graph (PR #98125)
https://github.com/shawbyoung updated
https://github.com/llvm/llvm-project/pull/98125
>From cf32a43e7c2b04079c6123fe13df4fb7226d771f Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 9 Jul 2024 10:04:25 -0700
Subject: [PATCH] Comments
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 69ea0899c5f2c..6753337c24ea7 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -501,7 +501,6 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
// Maps binary functions to adjacent functions in the FCG.
for (const BinaryFunction *CallerBF : BFs) {
-// Add all call targets to the hash map.
for (const BinaryBasicBlock &BB : CallerBF->blocks()) {
for (const MCInst &Inst : BB) {
if (!BC.MIB->isCall(Instr))
@@ -533,7 +532,8 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
}
}
- // Create mapping from neighbor hash to BFs.
+ // Using the constructed adjacent function mapping, creates mapping from
+ // neighbor hash to BFs.
std::unordered_map>
NeighborHashToBFs;
for (const BinaryFunction *BF : BFs) {
@@ -552,12 +552,12 @@ size_t
YAMLProfileReader::matchWithCallGraph(BinaryContext &BC) {
.push_back(BF);
}
- // TODO: change call anchor PR to have this representation - we need it here
+ // TODO: note, this will be introduced in the matching functions with calls
+ // as anchors pr
DenseMap
IdToYAMLBF;
- // TODO: change call anchor PR to have this representation - we need it here
- // Maps hashes to profiled functions.
+ // Maps YAML functions to adjacent functions in the profile FCG.
std::unordered_map
YamlBFToHashes(BFs.size());
@@ -590,7 +590,7 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
}
}
- // Matching YAMLBF with neighbor hashes.
+ // Matches YAMLBF to BFs with neighbor hashes.
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions) {
if (YamlBF.Used)
continue;
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
shawbyoung wrote: @maksfb Just addressed your comments, let me know if everything looks good. https://github.com/llvm/llvm-project/pull/96596 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [MLIR][LLVM] Use CyclicReplacerCache for recursive DIType import (PR #98203)
https://github.com/zyx-billy created
https://github.com/llvm/llvm-project/pull/98203
Use the new CyclicReplacerCache from
https://github.com/llvm/llvm-project/pull/98202 to support importing of
recursive DITypes in LLVM dialect's DebugImporter. This helps simplify the
implementation, allows for separate testing of the cache infra itself, and as a
result we even got more efficient translations.
---
Stacked PRs:
- https://github.com/llvm/llvm-project/pull/98202
>From 288234b7494eae0cd5f1e74db99edaf76290d1fd Mon Sep 17 00:00:00 2001
From: Billy Zhu
Date: Tue, 9 Jul 2024 10:28:21 -0700
Subject: [PATCH] use cyclic cache in importer and update tests
---
mlir/lib/Target/LLVMIR/DebugImporter.cpp | 142 ---
mlir/lib/Target/LLVMIR/DebugImporter.h | 101 ++---
mlir/test/Target/LLVMIR/Import/debug-info.ll | 34 +++--
3 files changed, 50 insertions(+), 227 deletions(-)
diff --git a/mlir/lib/Target/LLVMIR/DebugImporter.cpp
b/mlir/lib/Target/LLVMIR/DebugImporter.cpp
index 15737aa681c59..2bbaa485a07a7 100644
--- a/mlir/lib/Target/LLVMIR/DebugImporter.cpp
+++ b/mlir/lib/Target/LLVMIR/DebugImporter.cpp
@@ -28,7 +28,7 @@ using namespace mlir::LLVM::detail;
DebugImporter::DebugImporter(ModuleOp mlirModule,
bool dropDICompositeTypeElements)
-: recursionPruner(mlirModule.getContext()),
+: cache([&](llvm::DINode *node) { return createRecSelf(node); }),
context(mlirModule.getContext()), mlirModule(mlirModule),
dropDICompositeTypeElements(dropDICompositeTypeElements) {}
@@ -287,16 +287,9 @@ DINodeAttr DebugImporter::translate(llvm::DINode *node) {
return nullptr;
// Check for a cached instance.
- if (DINodeAttr attr = nodeToAttr.lookup(node))
-return attr;
-
- // Register with the recursive translator. If it can be handled without
- // recursing into it, return the result immediately.
- if (DINodeAttr attr = recursionPruner.pruneOrPushTranslationStack(node))
-return attr;
-
- auto guard = llvm::make_scope_exit(
- [&]() { recursionPruner.popTranslationStack(node); });
+ auto cacheEntry = cache.lookupOrInit(node);
+ if (std::optional result = cacheEntry.get())
+return *result;
// Convert the debug metadata if possible.
auto translateNode = [this](llvm::DINode *node) -> DINodeAttr {
@@ -335,20 +328,19 @@ DINodeAttr DebugImporter::translate(llvm::DINode *node) {
return nullptr;
};
if (DINodeAttr attr = translateNode(node)) {
-auto [result, isSelfContained] =
-recursionPruner.finalizeTranslation(node, attr);
-// Only cache fully self-contained nodes.
-if (isSelfContained)
- nodeToAttr.try_emplace(node, result);
-return result;
+// If this node is given a recursive ID, set the recId onto the result.
+if (cacheEntry.wasRepeated()) {
+ DistinctAttr recId = nodeToRecId.lookup(node);
+ auto recType = cast(attr);
+ attr = cast(recType.withRecId(recId));
+}
+cacheEntry.resolve(attr);
+return attr;
}
+ cacheEntry.resolve(nullptr);
return nullptr;
}
-//===--===//
-// RecursionPruner
-//===--===//
-
/// Get the `getRecSelf` constructor for the translated type of `node` if its
/// translated DITypeAttr supports recursion. Otherwise, returns nullptr.
static function_ref
@@ -361,104 +353,20 @@ getRecSelfConstructor(llvm::DINode *node) {
.Default(CtorType());
}
-DINodeAttr DebugImporter::RecursionPruner::pruneOrPushTranslationStack(
-llvm::DINode *node) {
- // If the node type is capable of being recursive, check if it's seen
- // before.
+std::optional DebugImporter::createRecSelf(llvm::DINode *node) {
auto recSelfCtor = getRecSelfConstructor(node);
- if (recSelfCtor) {
-// If a cyclic dependency is detected since the same node is being
-// traversed twice, emit a recursive self type, and mark the duplicate
-// node on the translationStack so it can emit a recursive decl type.
-auto [iter, inserted] = translationStack.try_emplace(node);
-if (!inserted) {
- // The original node may have already been assigned a recursive ID from
- // a different self-reference. Use that if possible.
- DIRecursiveTypeAttrInterface recSelf = iter->second.recSelf;
- if (!recSelf) {
-DistinctAttr recId = nodeToRecId.lookup(node);
-if (!recId) {
- recId = DistinctAttr::create(UnitAttr::get(context));
- nodeToRecId[node] = recId;
-}
-recSelf = recSelfCtor(recId);
-iter->second.recSelf = recSelf;
- }
- // Inject the self-ref into the previous layer.
- translationStack.back().second.unboundSelfRefs.insert(recSelf);
- return cast(recSelf);
-}
+ if (!recSelfCtor)
+return std::nullopt;
+
+ // The original node may have already been assigned a recursive ID from
+ //
[llvm-branch-commits] [mlir] [MLIR] Cyclic AttrType Replacer (PR #98206)
https://github.com/zyx-billy created
https://github.com/llvm/llvm-project/pull/98206
The current `AttrTypeReplacer` does not allow for custom handling of replacer
functions that may cause self-recursion. For example, the replacement of one
attr/type may depend on the replacement of another attr/type (by calling into
the replacer manually again), which in turn may depend on the replacement of
the original attr/type.
To enable this functionality, this PR broke out the original AttrTypeReplacer
into two parts:
- An uncached base version (`detail::AttrTypeReplacerBase`) that allows
registering replacer functions and has logic for invoking it on attr/types &
their sub-elements
- A cached version (`AttrTypeReplacer`) that provides the same caching as the
original one. This is still the one used everywhere and behavior is unchanged.
On top of the uncached base version, a `CyclicAttrTypeReplacer` is introduced
that provides caching & cycle-handling for replacer logic that is cyclic.
Cycle-breaking & caching is provided by the `CyclicReplacerCache` from
https://github.com/llvm/llvm-project/pull/98202.
Both concrete implementations of the uncached base version use CRTP to avoid
dynamic dispatch. The base class merely provides replacer registration &
invocation, and is not meant to be used, or otherwise extended elsewhere.
---
Stacked PRs:
- https://github.com/llvm/llvm-project/pull/98202
>From 0d38de6f61b4dcb6d0d3ba5ddc75a64819f64c0a Mon Sep 17 00:00:00 2001
From: Billy Zhu
Date: Tue, 9 Jul 2024 10:27:13 -0700
Subject: [PATCH] refactor attrtype replacers & add tests
---
mlir/include/mlir/IR/AttrTypeSubElements.h | 142 +++--
mlir/lib/IR/AttrTypeSubElements.cpp| 146 ++---
mlir/unittests/IR/AttrTypeReplacerTest.cpp | 231 +
mlir/unittests/IR/CMakeLists.txt | 1 +
4 files changed, 469 insertions(+), 51 deletions(-)
create mode 100644 mlir/unittests/IR/AttrTypeReplacerTest.cpp
diff --git a/mlir/include/mlir/IR/AttrTypeSubElements.h
b/mlir/include/mlir/IR/AttrTypeSubElements.h
index 3105040b87631..2e953d490f92a 100644
--- a/mlir/include/mlir/IR/AttrTypeSubElements.h
+++ b/mlir/include/mlir/IR/AttrTypeSubElements.h
@@ -16,6 +16,7 @@
#include "mlir/IR/MLIRContext.h"
#include "mlir/IR/Visitors.h"
+#include "mlir/Support/CyclicReplacerCache.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/DenseMap.h"
#include
@@ -116,9 +117,21 @@ class AttrTypeWalker {
/// AttrTypeReplacer
//===--===//
-/// This class provides a utility for replacing attributes/types, and their sub
-/// elements. Multiple replacement functions may be registered.
-class AttrTypeReplacer {
+namespace detail {
+
+/// This class provides a base utility for replacing attributes/types, and
their
+/// sub elements. Multiple replacement functions may be registered.
+///
+/// This base utility is uncached. Users can choose between two cached versions
+/// of this replacer:
+/// * For non-cyclic replacer logic, use `AttrTypeReplacer`.
+/// * For cyclic replacer logic, use `CyclicAttrTypeReplacer`.
+///
+/// Concrete implementations implement the following `replace` entry functions:
+/// * Attribute replace(Attribute attr);
+/// * Type replace(Type type);
+template
+class AttrTypeReplacerBase {
public:
//======//
// Application
@@ -139,12 +152,6 @@ class AttrTypeReplacer {
bool replaceLocs = false,
bool replaceTypes = false);
- /// Replace the given attribute/type, and recursively replace any sub
- /// elements. Returns either the new attribute/type, or nullptr in the case
of
- /// failure.
- Attribute replace(Attribute attr);
- Type replace(Type type);
-
//======//
// Registration
//======//
@@ -206,29 +213,122 @@ class AttrTypeReplacer {
});
}
-private:
- /// Internal implementation of the `replace` methods above.
- template
- T replaceImpl(T element, ReplaceFns &replaceFns);
-
- /// Replace the sub elements of the given interface.
- template
- T replaceSubElements(T interface);
+protected:
+ /// Invokes the registered replacement functions from most recently
registered
+ /// to least recently registered until a successful replacement is returned.
+ /// Unless skipping is requested, invokes `replace` on sub-elements of the
+ /// current attr/type.
+ Attribute replaceBase(Attribute attr);
+ Type replaceBase(Type type);
+private:
/// The set of replacement functions that map sub elements.
std::vector> attrReplacementFns;
std::vector> typeReplacementFns;
+};
+
+} // namespace detail
+
+/// This is an attribute/type replacer that is naively cached. It is best
[llvm-branch-commits] [mlir] [MLIR][LLVM] Use CyclicReplacerCache for recursive DIType import (PR #98203)
https://github.com/zyx-billy edited https://github.com/llvm/llvm-project/pull/98203 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [MLIR] Cyclic AttrType Replacer (PR #98206)
https://github.com/zyx-billy edited https://github.com/llvm/llvm-project/pull/98206 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [MLIR] Cyclic AttrType Replacer (PR #98206)
https://github.com/zyx-billy updated
https://github.com/llvm/llvm-project/pull/98206
>From 8d1dd886c4a80507c8a97dda15e91acbfa7c3619 Mon Sep 17 00:00:00 2001
From: Billy Zhu
Date: Tue, 9 Jul 2024 10:27:13 -0700
Subject: [PATCH] refactor attrtype replacers & add tests
---
mlir/include/mlir/IR/AttrTypeSubElements.h | 138 ++--
mlir/lib/IR/AttrTypeSubElements.cpp| 146 ++---
mlir/unittests/IR/AttrTypeReplacerTest.cpp | 231 +
mlir/unittests/IR/CMakeLists.txt | 1 +
4 files changed, 467 insertions(+), 49 deletions(-)
create mode 100644 mlir/unittests/IR/AttrTypeReplacerTest.cpp
diff --git a/mlir/include/mlir/IR/AttrTypeSubElements.h
b/mlir/include/mlir/IR/AttrTypeSubElements.h
index 3105040b87631..234767deea00a 100644
--- a/mlir/include/mlir/IR/AttrTypeSubElements.h
+++ b/mlir/include/mlir/IR/AttrTypeSubElements.h
@@ -16,6 +16,7 @@
#include "mlir/IR/MLIRContext.h"
#include "mlir/IR/Visitors.h"
+#include "mlir/Support/CyclicReplacerCache.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/DenseMap.h"
#include
@@ -116,9 +117,21 @@ class AttrTypeWalker {
/// AttrTypeReplacer
//===--===//
-/// This class provides a utility for replacing attributes/types, and their sub
-/// elements. Multiple replacement functions may be registered.
-class AttrTypeReplacer {
+namespace detail {
+
+/// This class provides a base utility for replacing attributes/types, and
their
+/// sub elements. Multiple replacement functions may be registered.
+///
+/// This base utility is uncached. Users can choose between two cached versions
+/// of this replacer:
+/// * For non-cyclic replacer logic, use `AttrTypeReplacer`.
+/// * For cyclic replacer logic, use `CyclicAttrTypeReplacer`.
+///
+/// Concrete implementations implement the following `replace` entry functions:
+/// * Attribute replace(Attribute attr);
+/// * Type replace(Type type);
+template
+class AttrTypeReplacerBase {
public:
//======//
// Application
@@ -139,12 +152,6 @@ class AttrTypeReplacer {
bool replaceLocs = false,
bool replaceTypes = false);
- /// Replace the given attribute/type, and recursively replace any sub
- /// elements. Returns either the new attribute/type, or nullptr in the case
of
- /// failure.
- Attribute replace(Attribute attr);
- Type replace(Type type);
-
//======//
// Registration
//======//
@@ -206,21 +213,114 @@ class AttrTypeReplacer {
});
}
-private:
- /// Internal implementation of the `replace` methods above.
- template
- T replaceImpl(T element, ReplaceFns &replaceFns);
-
- /// Replace the sub elements of the given interface.
- template
- T replaceSubElements(T interface);
+protected:
+ /// Invokes the registered replacement functions from most recently
registered
+ /// to least recently registered until a successful replacement is returned.
+ /// Unless skipping is requested, invokes `replace` on sub-elements of the
+ /// current attr/type.
+ Attribute replaceBase(Attribute attr);
+ Type replaceBase(Type type);
+private:
/// The set of replacement functions that map sub elements.
std::vector> attrReplacementFns;
std::vector> typeReplacementFns;
+};
+
+} // namespace detail
+
+/// This is an attribute/type replacer that is naively cached. It is best used
+/// when the replacer logic is guaranteed to not contain cycles. Otherwise, any
+/// re-occurrence of an in-progress element will be skipped.
+class AttrTypeReplacer : public detail::AttrTypeReplacerBase
{
+public:
+ Attribute replace(Attribute attr);
+ Type replace(Type type);
+
+private:
+ /// Shared concrete implementation of the public `replace` functions. Invokes
+ /// `replaceBase` with caching.
+ template
+ T cachedReplaceImpl(T element);
+
+ // Stores the opaque pointer of an attribute or type.
+ DenseMap cache;
+};
+
+/// This is an attribute/type replacer that supports custom handling of cycles
+/// in the replacer logic. In addition to registering replacer functions, it
+/// allows registering cycle-breaking functions in the same style.
+class CyclicAttrTypeReplacer
+: public detail::AttrTypeReplacerBase {
+public:
+ CyclicAttrTypeReplacer();
- /// The set of cached mappings for attributes/types.
- DenseMap attrTypeMap;
+
//======//
+ // Application
+
//======//
+
+ Attribute replace(Attribute attr);
+ Type replace(Type type);
+
+
//======//
+ // Registration
+
//===-
[llvm-branch-commits] [mlir] [MLIR][LLVM] Use CyclicReplacerCache for recursive DIType import (PR #98203)
https://github.com/zyx-billy updated
https://github.com/llvm/llvm-project/pull/98203
>From a0b59b246a1b8860fad16f1f74537c38b5abf2cf Mon Sep 17 00:00:00 2001
From: Billy Zhu
Date: Tue, 9 Jul 2024 10:28:21 -0700
Subject: [PATCH] use cyclic cache in importer and update tests
---
mlir/lib/Target/LLVMIR/DebugImporter.cpp | 143 ---
mlir/lib/Target/LLVMIR/DebugImporter.h | 101 ++---
mlir/test/Target/LLVMIR/Import/debug-info.ll | 34 +++--
3 files changed, 51 insertions(+), 227 deletions(-)
diff --git a/mlir/lib/Target/LLVMIR/DebugImporter.cpp
b/mlir/lib/Target/LLVMIR/DebugImporter.cpp
index 15737aa681c59..f104b72209c39 100644
--- a/mlir/lib/Target/LLVMIR/DebugImporter.cpp
+++ b/mlir/lib/Target/LLVMIR/DebugImporter.cpp
@@ -28,7 +28,7 @@ using namespace mlir::LLVM::detail;
DebugImporter::DebugImporter(ModuleOp mlirModule,
bool dropDICompositeTypeElements)
-: recursionPruner(mlirModule.getContext()),
+: cache([&](llvm::DINode *node) { return createRecSelf(node); }),
context(mlirModule.getContext()), mlirModule(mlirModule),
dropDICompositeTypeElements(dropDICompositeTypeElements) {}
@@ -287,16 +287,9 @@ DINodeAttr DebugImporter::translate(llvm::DINode *node) {
return nullptr;
// Check for a cached instance.
- if (DINodeAttr attr = nodeToAttr.lookup(node))
-return attr;
-
- // Register with the recursive translator. If it can be handled without
- // recursing into it, return the result immediately.
- if (DINodeAttr attr = recursionPruner.pruneOrPushTranslationStack(node))
-return attr;
-
- auto guard = llvm::make_scope_exit(
- [&]() { recursionPruner.popTranslationStack(node); });
+ auto cacheEntry = cache.lookupOrInit(node);
+ if (std::optional result = cacheEntry.get())
+return *result;
// Convert the debug metadata if possible.
auto translateNode = [this](llvm::DINode *node) -> DINodeAttr {
@@ -335,20 +328,20 @@ DINodeAttr DebugImporter::translate(llvm::DINode *node) {
return nullptr;
};
if (DINodeAttr attr = translateNode(node)) {
-auto [result, isSelfContained] =
-recursionPruner.finalizeTranslation(node, attr);
-// Only cache fully self-contained nodes.
-if (isSelfContained)
- nodeToAttr.try_emplace(node, result);
-return result;
+// If this node was repeated, lookup its recursive ID and assign it to the
+// base result.
+if (cacheEntry.wasRepeated()) {
+ DistinctAttr recId = nodeToRecId.lookup(node);
+ auto recType = cast(attr);
+ attr = cast(recType.withRecId(recId));
+}
+cacheEntry.resolve(attr);
+return attr;
}
+ cacheEntry.resolve(nullptr);
return nullptr;
}
-//===--===//
-// RecursionPruner
-//===--===//
-
/// Get the `getRecSelf` constructor for the translated type of `node` if its
/// translated DITypeAttr supports recursion. Otherwise, returns nullptr.
static function_ref
@@ -361,104 +354,20 @@ getRecSelfConstructor(llvm::DINode *node) {
.Default(CtorType());
}
-DINodeAttr DebugImporter::RecursionPruner::pruneOrPushTranslationStack(
-llvm::DINode *node) {
- // If the node type is capable of being recursive, check if it's seen
- // before.
+std::optional DebugImporter::createRecSelf(llvm::DINode *node) {
auto recSelfCtor = getRecSelfConstructor(node);
- if (recSelfCtor) {
-// If a cyclic dependency is detected since the same node is being
-// traversed twice, emit a recursive self type, and mark the duplicate
-// node on the translationStack so it can emit a recursive decl type.
-auto [iter, inserted] = translationStack.try_emplace(node);
-if (!inserted) {
- // The original node may have already been assigned a recursive ID from
- // a different self-reference. Use that if possible.
- DIRecursiveTypeAttrInterface recSelf = iter->second.recSelf;
- if (!recSelf) {
-DistinctAttr recId = nodeToRecId.lookup(node);
-if (!recId) {
- recId = DistinctAttr::create(UnitAttr::get(context));
- nodeToRecId[node] = recId;
-}
-recSelf = recSelfCtor(recId);
-iter->second.recSelf = recSelf;
- }
- // Inject the self-ref into the previous layer.
- translationStack.back().second.unboundSelfRefs.insert(recSelf);
- return cast(recSelf);
-}
+ if (!recSelfCtor)
+return std::nullopt;
+
+ // The original node may have already been assigned a recursive ID from
+ // a different self-reference. Use that if possible.
+ DistinctAttr recId = nodeToRecId.lookup(node);
+ if (!recId) {
+recId = DistinctAttr::create(UnitAttr::get(context));
+nodeToRecId[node] = recId;
}
-
- return lookup(node);
-}
-
-std::pair
-DebugImporter::RecursionPruner::finalizeTranslation(llvm::DINode *node,
-
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
@@ -40,6 +40,8 @@ class YAMLProfileReader : public ProfileReaderBase {
/// Check if the file contains YAML.
static bool isYAML(StringRef Filename);
+ using FunctionMap = DenseMap;
aaupov wrote:
```suggestion
using ProfileLookupMap = DenseMap;
```
https://github.com/llvm/llvm-project/pull/96596
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
@@ -181,20 +182,19 @@ std::string hashBlockCalls(BinaryContext &BC, const
BinaryBasicBlock &BB) {
/// The same as the $hashBlockCalls function, but for profiled functions.
std::string
-hashBlockCalls(const DenseMap &IdToFunctionName,
+hashBlockCalls(const DenseMap
+ &IdToYamlFunction,
const yaml::bolt::BinaryBasicBlockProfile &YamlBB) {
- std::multiset FunctionNames;
+ std::vector FunctionNames;
for (const yaml::bolt::CallSiteInfo &CallSiteInfo : YamlBB.CallSites) {
-auto It = IdToFunctionName.find(CallSiteInfo.DestId);
-if (It == IdToFunctionName.end())
+auto It = IdToYamlFunction.find(CallSiteInfo.DestId);
+if (It == IdToYamlFunction.end())
continue;
-StringRef Name = It->second;
-const size_t Pos = Name.find("(*");
-if (Pos != StringRef::npos)
- Name = Name.substr(0, Pos);
-FunctionNames.insert(std::string(Name));
+StringRef Name =
+NameResolver::removeSuffix(It->second->Name, StringRef("(*"));
aaupov wrote:
I think the intent of Maksim's comment was to hide an implementation detail
(using `(*N)` suffix) in relevant component (NameResolver).
https://github.com/llvm/llvm-project/pull/96596
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Match functions with call graph (PR #98125)
https://github.com/shawbyoung updated
https://github.com/llvm/llvm-project/pull/98125
>From cf32a43e7c2b04079c6123fe13df4fb7226d771f Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 9 Jul 2024 10:04:25 -0700
Subject: [PATCH 1/2] Comments
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 69ea0899c5f2c..6753337c24ea7 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -501,7 +501,6 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
// Maps binary functions to adjacent functions in the FCG.
for (const BinaryFunction *CallerBF : BFs) {
-// Add all call targets to the hash map.
for (const BinaryBasicBlock &BB : CallerBF->blocks()) {
for (const MCInst &Inst : BB) {
if (!BC.MIB->isCall(Instr))
@@ -533,7 +532,8 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
}
}
- // Create mapping from neighbor hash to BFs.
+ // Using the constructed adjacent function mapping, creates mapping from
+ // neighbor hash to BFs.
std::unordered_map>
NeighborHashToBFs;
for (const BinaryFunction *BF : BFs) {
@@ -552,12 +552,12 @@ size_t
YAMLProfileReader::matchWithCallGraph(BinaryContext &BC) {
.push_back(BF);
}
- // TODO: change call anchor PR to have this representation - we need it here
+ // TODO: note, this will be introduced in the matching functions with calls
+ // as anchors pr
DenseMap
IdToYAMLBF;
- // TODO: change call anchor PR to have this representation - we need it here
- // Maps hashes to profiled functions.
+ // Maps YAML functions to adjacent functions in the profile FCG.
std::unordered_map
YamlBFToHashes(BFs.size());
@@ -590,7 +590,7 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
}
}
- // Matching YAMLBF with neighbor hashes.
+ // Matches YAMLBF to BFs with neighbor hashes.
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions) {
if (YamlBF.Used)
continue;
>From ee9049fc4bd3d4203c19c9c0982a78ab3b47666f Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 9 Jul 2024 13:52:05 -0700
Subject: [PATCH 2/2] Moved blended hash definition
Created using spr 1.3.4
---
bolt/include/bolt/Profile/YAMLProfileReader.h | 69 ++-
bolt/lib/Profile/StaleProfileMatching.cpp | 65 ---
bolt/lib/Profile/YAMLProfileReader.cpp| 110 --
3 files changed, 119 insertions(+), 125 deletions(-)
diff --git a/bolt/include/bolt/Profile/YAMLProfileReader.h
b/bolt/include/bolt/Profile/YAMLProfileReader.h
index 36e8f8739eee1..e8a34ecad9a08 100644
--- a/bolt/include/bolt/Profile/YAMLProfileReader.h
+++ b/bolt/include/bolt/Profile/YAMLProfileReader.h
@@ -16,6 +16,73 @@
namespace llvm {
namespace bolt {
+/// An object wrapping several components of a basic block hash. The combined
+/// (blended) hash is represented and stored as one uint64_t, while individual
+/// components are of smaller size (e.g., uint16_t or uint8_t).
+struct BlendedBlockHash {
+private:
+ using ValueOffset = Bitfield::Element;
+ using ValueOpcode = Bitfield::Element;
+ using ValueInstr = Bitfield::Element;
+ using ValuePred = Bitfield::Element;
+ using ValueSucc = Bitfield::Element;
+
+public:
+ explicit BlendedBlockHash() {}
+
+ explicit BlendedBlockHash(uint64_t Hash) {
+Offset = Bitfield::get(Hash);
+OpcodeHash = Bitfield::get(Hash);
+InstrHash = Bitfield::get(Hash);
+PredHash = Bitfield::get(Hash);
+SuccHash = Bitfield::get(Hash);
+ }
+
+ /// Combine the blended hash into uint64_t.
+ uint64_t combine() const {
+uint64_t Hash = 0;
+Bitfield::set(Hash, Offset);
+Bitfield::set(Hash, OpcodeHash);
+Bitfield::set(Hash, InstrHash);
+Bitfield::set(Hash, PredHash);
+Bitfield::set(Hash, SuccHash);
+return Hash;
+ }
+
+ /// Compute a distance between two given blended hashes. The smaller the
+ /// distance, the more similar two blocks are. For identical basic blocks,
+ /// the distance is zero.
+ uint64_t distance(const BlendedBlockHash &BBH) const {
+assert(OpcodeHash == BBH.OpcodeHash &&
+ "incorrect blended hash distance computation");
+uint64_t Dist = 0;
+// Account for NeighborHash
+Dist += SuccHash == BBH.SuccHash ? 0 : 1;
+Dist += PredHash == BBH.PredHash ? 0 : 1;
+Dist <<= 16;
+// Account for InstrHash
+Dist += InstrHash == BBH.InstrHash ? 0 : 1;
+Dist <<= 16;
+// Account for Offset
+Dist += (Offset >= BBH.Offset ? Offset - BBH.Offset : BBH.Offset - Offset);
+return Dist;
+ }
+
+ /// The offset of the basic block from the function start.
+ uint16_t Offset{0};
+ /// (Loose) Hash of the basic block instructions, excluding operands.
+ uint16_t OpcodeHash{0};
+ /// (Strong)
[llvm-branch-commits] [mlir] [MLIR][LLVM] Use CyclicReplacerCache for recursive DIType import (PR #98203)
https://github.com/zyx-billy ready_for_review https://github.com/llvm/llvm-project/pull/98203 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [MLIR] Cyclic AttrType Replacer (PR #98206)
https://github.com/zyx-billy ready_for_review https://github.com/llvm/llvm-project/pull/98206 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [MLIR][LLVM] Use CyclicReplacerCache for recursive DIType import (PR #98203)
llvmbot wrote:
@llvm/pr-subscribers-mlir-llvm
@llvm/pr-subscribers-mlir
Author: Billy Zhu (zyx-billy)
Changes
Use the new CyclicReplacerCache from
https://github.com/llvm/llvm-project/pull/98202 to support importing of
recursive DITypes in LLVM dialect's DebugImporter. This helps simplify the
implementation, allows for separate testing of the cache infra itself, and as a
result we even got more efficient translations.
---
Stacked PRs:
- https://github.com/llvm/llvm-project/pull/98202
- ➡️ https://github.com/llvm/llvm-project/pull/98203
---
Full diff: https://github.com/llvm/llvm-project/pull/98203.diff
3 Files Affected:
- (modified) mlir/lib/Target/LLVMIR/DebugImporter.cpp (+26-117)
- (modified) mlir/lib/Target/LLVMIR/DebugImporter.h (+9-92)
- (modified) mlir/test/Target/LLVMIR/Import/debug-info.ll (+16-18)
``diff
diff --git a/mlir/lib/Target/LLVMIR/DebugImporter.cpp
b/mlir/lib/Target/LLVMIR/DebugImporter.cpp
index 15737aa681c59..f104b72209c39 100644
--- a/mlir/lib/Target/LLVMIR/DebugImporter.cpp
+++ b/mlir/lib/Target/LLVMIR/DebugImporter.cpp
@@ -28,7 +28,7 @@ using namespace mlir::LLVM::detail;
DebugImporter::DebugImporter(ModuleOp mlirModule,
bool dropDICompositeTypeElements)
-: recursionPruner(mlirModule.getContext()),
+: cache([&](llvm::DINode *node) { return createRecSelf(node); }),
context(mlirModule.getContext()), mlirModule(mlirModule),
dropDICompositeTypeElements(dropDICompositeTypeElements) {}
@@ -287,16 +287,9 @@ DINodeAttr DebugImporter::translate(llvm::DINode *node) {
return nullptr;
// Check for a cached instance.
- if (DINodeAttr attr = nodeToAttr.lookup(node))
-return attr;
-
- // Register with the recursive translator. If it can be handled without
- // recursing into it, return the result immediately.
- if (DINodeAttr attr = recursionPruner.pruneOrPushTranslationStack(node))
-return attr;
-
- auto guard = llvm::make_scope_exit(
- [&]() { recursionPruner.popTranslationStack(node); });
+ auto cacheEntry = cache.lookupOrInit(node);
+ if (std::optional result = cacheEntry.get())
+return *result;
// Convert the debug metadata if possible.
auto translateNode = [this](llvm::DINode *node) -> DINodeAttr {
@@ -335,20 +328,20 @@ DINodeAttr DebugImporter::translate(llvm::DINode *node) {
return nullptr;
};
if (DINodeAttr attr = translateNode(node)) {
-auto [result, isSelfContained] =
-recursionPruner.finalizeTranslation(node, attr);
-// Only cache fully self-contained nodes.
-if (isSelfContained)
- nodeToAttr.try_emplace(node, result);
-return result;
+// If this node was repeated, lookup its recursive ID and assign it to the
+// base result.
+if (cacheEntry.wasRepeated()) {
+ DistinctAttr recId = nodeToRecId.lookup(node);
+ auto recType = cast(attr);
+ attr = cast(recType.withRecId(recId));
+}
+cacheEntry.resolve(attr);
+return attr;
}
+ cacheEntry.resolve(nullptr);
return nullptr;
}
-//===--===//
-// RecursionPruner
-//===--===//
-
/// Get the `getRecSelf` constructor for the translated type of `node` if its
/// translated DITypeAttr supports recursion. Otherwise, returns nullptr.
static function_ref
@@ -361,104 +354,20 @@ getRecSelfConstructor(llvm::DINode *node) {
.Default(CtorType());
}
-DINodeAttr DebugImporter::RecursionPruner::pruneOrPushTranslationStack(
-llvm::DINode *node) {
- // If the node type is capable of being recursive, check if it's seen
- // before.
+std::optional DebugImporter::createRecSelf(llvm::DINode *node) {
auto recSelfCtor = getRecSelfConstructor(node);
- if (recSelfCtor) {
-// If a cyclic dependency is detected since the same node is being
-// traversed twice, emit a recursive self type, and mark the duplicate
-// node on the translationStack so it can emit a recursive decl type.
-auto [iter, inserted] = translationStack.try_emplace(node);
-if (!inserted) {
- // The original node may have already been assigned a recursive ID from
- // a different self-reference. Use that if possible.
- DIRecursiveTypeAttrInterface recSelf = iter->second.recSelf;
- if (!recSelf) {
-DistinctAttr recId = nodeToRecId.lookup(node);
-if (!recId) {
- recId = DistinctAttr::create(UnitAttr::get(context));
- nodeToRecId[node] = recId;
-}
-recSelf = recSelfCtor(recId);
-iter->second.recSelf = recSelf;
- }
- // Inject the self-ref into the previous layer.
- translationStack.back().second.unboundSelfRefs.insert(recSelf);
- return cast(recSelf);
-}
+ if (!recSelfCtor)
+return std::nullopt;
+
+ // The original node may have already been assigned a recursive ID from
+ // a different self-reference. Use that
[llvm-branch-commits] [mlir] [MLIR] Cyclic AttrType Replacer (PR #98206)
llvmbot wrote:
@llvm/pr-subscribers-mlir
@llvm/pr-subscribers-mlir-core
Author: Billy Zhu (zyx-billy)
Changes
The current `AttrTypeReplacer` does not allow for custom handling of replacer
functions that may cause self-recursion. For example, the replacement of one
attr/type may depend on the replacement of another attr/type (by calling into
the replacer manually again), which in turn may depend on the replacement of
the original attr/type.
To enable this functionality, this PR broke out the original AttrTypeReplacer
into two parts:
- An uncached base version (`detail::AttrTypeReplacerBase`) that allows
registering replacer functions and has logic for invoking it on attr/types
& their sub-elements
- A cached version (`AttrTypeReplacer`) that provides the same caching as the
original one. This is still the one used everywhere and behavior is unchanged.
On top of the uncached base version, a `CyclicAttrTypeReplacer` is introduced
that provides caching & cycle-handling for replacer logic that is cyclic.
Cycle-breaking & caching is provided by the `CyclicReplacerCache` from
https://github.com/llvm/llvm-project/pull/98202.
Both concrete implementations of the uncached base version use CRTP to avoid
dynamic dispatch. The base class merely provides replacer registration &
invocation, and is not meant to be used, or otherwise extended elsewhere.
---
Stacked PRs:
- https://github.com/llvm/llvm-project/pull/98202
- ➡️ https://github.com/llvm/llvm-project/pull/98206
---
Patch is 24.36 KiB, truncated to 20.00 KiB below, full version:
https://github.com/llvm/llvm-project/pull/98206.diff
4 Files Affected:
- (modified) mlir/include/mlir/IR/AttrTypeSubElements.h (+119-19)
- (modified) mlir/lib/IR/AttrTypeSubElements.cpp (+116-30)
- (added) mlir/unittests/IR/AttrTypeReplacerTest.cpp (+231)
- (modified) mlir/unittests/IR/CMakeLists.txt (+1)
``diff
diff --git a/mlir/include/mlir/IR/AttrTypeSubElements.h
b/mlir/include/mlir/IR/AttrTypeSubElements.h
index 3105040b87631..234767deea00a 100644
--- a/mlir/include/mlir/IR/AttrTypeSubElements.h
+++ b/mlir/include/mlir/IR/AttrTypeSubElements.h
@@ -16,6 +16,7 @@
#include "mlir/IR/MLIRContext.h"
#include "mlir/IR/Visitors.h"
+#include "mlir/Support/CyclicReplacerCache.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/DenseMap.h"
#include
@@ -116,9 +117,21 @@ class AttrTypeWalker {
/// AttrTypeReplacer
//===--===//
-/// This class provides a utility for replacing attributes/types, and their sub
-/// elements. Multiple replacement functions may be registered.
-class AttrTypeReplacer {
+namespace detail {
+
+/// This class provides a base utility for replacing attributes/types, and
their
+/// sub elements. Multiple replacement functions may be registered.
+///
+/// This base utility is uncached. Users can choose between two cached versions
+/// of this replacer:
+/// * For non-cyclic replacer logic, use `AttrTypeReplacer`.
+/// * For cyclic replacer logic, use `CyclicAttrTypeReplacer`.
+///
+/// Concrete implementations implement the following `replace` entry functions:
+/// * Attribute replace(Attribute attr);
+/// * Type replace(Type type);
+template
+class AttrTypeReplacerBase {
public:
//======//
// Application
@@ -139,12 +152,6 @@ class AttrTypeReplacer {
bool replaceLocs = false,
bool replaceTypes = false);
- /// Replace the given attribute/type, and recursively replace any sub
- /// elements. Returns either the new attribute/type, or nullptr in the case
of
- /// failure.
- Attribute replace(Attribute attr);
- Type replace(Type type);
-
//======//
// Registration
//======//
@@ -206,21 +213,114 @@ class AttrTypeReplacer {
});
}
-private:
- /// Internal implementation of the `replace` methods above.
- template
- T replaceImpl(T element, ReplaceFns &replaceFns);
-
- /// Replace the sub elements of the given interface.
- template
- T replaceSubElements(T interface);
+protected:
+ /// Invokes the registered replacement functions from most recently
registered
+ /// to least recently registered until a successful replacement is returned.
+ /// Unless skipping is requested, invokes `replace` on sub-elements of the
+ /// current attr/type.
+ Attribute replaceBase(Attribute attr);
+ Type replaceBase(Type type);
+private:
/// The set of replacement functions that map sub elements.
std::vector> attrReplacementFns;
std::vector> typeReplacementFns;
+};
+
+} // namespace detail
+
+/// This is an attribute/type replacer that is naively cached. It is best used
+/// when the replacer logic is guaranteed to not contain cycles. O
[llvm-branch-commits] Improve tests, change API (PR #98235)
https://github.com/pcc created https://github.com/llvm/llvm-project/pull/98235 None ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] Improve tests, change API (PR #98235)
llvmbot wrote:
@llvm/pr-subscribers-llvm-binary-utilities
Author: None (pcc)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/98235.diff
3 Files Affected:
- (modified) llvm/include/llvm/Object/ArchiveWriter.h (+6-2)
- (modified) llvm/lib/Object/ArchiveWriter.cpp (+22-14)
- (modified) llvm/test/Object/archive-malformed-object.test (+14-1)
``diff
diff --git a/llvm/include/llvm/Object/ArchiveWriter.h
b/llvm/include/llvm/Object/ArchiveWriter.h
index a19f8fcc79d74..f965a12c45267 100644
--- a/llvm/include/llvm/Object/ArchiveWriter.h
+++ b/llvm/include/llvm/Object/ArchiveWriter.h
@@ -48,18 +48,22 @@ enum class SymtabWritingMode {
BigArchive64 // Only write the 64-bit symbol table.
};
+void warnToStderr(Error Err);
+
Error writeArchive(StringRef ArcName, ArrayRef NewMembers,
SymtabWritingMode WriteSymtab, object::Archive::Kind Kind,
bool Deterministic, bool Thin,
std::unique_ptr OldArchiveBuf = nullptr,
- std::optional IsEC = std::nullopt);
+ std::optional IsEC = std::nullopt,
+ function_ref Warn = warnToStderr);
// writeArchiveToBuffer is similar to writeArchive but returns the Archive in a
// buffer instead of writing it out to a file.
Expected>
writeArchiveToBuffer(ArrayRef NewMembers,
SymtabWritingMode WriteSymtab, object::Archive::Kind Kind,
- bool Deterministic, bool Thin);
+ bool Deterministic, bool Thin,
+ function_ref Warn = warnToStderr);
}
#endif
diff --git a/llvm/lib/Object/ArchiveWriter.cpp
b/llvm/lib/Object/ArchiveWriter.cpp
index c6d443ff9d15a..1c75268322257 100644
--- a/llvm/lib/Object/ArchiveWriter.cpp
+++ b/llvm/lib/Object/ArchiveWriter.cpp
@@ -483,7 +483,7 @@ static uint64_t computeHeadersSize(object::Archive::Kind
Kind,
static Expected>
getSymbolicFile(MemoryBufferRef Buf, LLVMContext &Context,
-object::Archive::Kind Kind) {
+object::Archive::Kind Kind, function_ref Warn) {
const file_magic Type = identify_magic(Buf.getBuffer());
// Don't attempt to read non-symbolic file types.
if (!object::SymbolicFile::isSymbolicFile(Type, &Context))
@@ -512,9 +512,7 @@ getSymbolicFile(MemoryBufferRef Buf, LLVMContext &Context,
case object::Archive::K_BSD:
case object::Archive::K_GNU:
case object::Archive::K_GNU64:
-llvm::logAllUnhandledErrors(ObjOrErr.takeError(), llvm::errs(),
-"warning: " + Buf.getBufferIdentifier() +
-": ");
+Warn(ObjOrErr.takeError());
return nullptr;
case object::Archive::K_AIXBIG:
case object::Archive::K_COFF:
@@ -782,7 +780,7 @@ computeMemberData(raw_ostream &StringTable, raw_ostream
&SymNames,
object::Archive::Kind Kind, bool Thin, bool Deterministic,
SymtabWritingMode NeedSymbols, SymMap *SymMap,
LLVMContext &Context, ArrayRef NewMembers,
- std::optional IsEC) {
+ std::optional IsEC, function_ref Warn) {
static char PaddingData[8] = {'\n', '\n', '\n', '\n', '\n', '\n', '\n',
'\n'};
uint64_t MemHeadPadSize = 0;
uint64_t Pos =
@@ -850,8 +848,10 @@ computeMemberData(raw_ostream &StringTable, raw_ostream
&SymNames,
if (NeedSymbols != SymtabWritingMode::NoSymtab || isAIXBigArchive(Kind)) {
for (const NewArchiveMember &M : NewMembers) {
- Expected> SymFileOrErr =
- getSymbolicFile(M.Buf->getMemBufferRef(), Context, Kind);
+ Expected> SymFileOrErr = getSymbolicFile(
+ M.Buf->getMemBufferRef(), Context, Kind, [&](Error Err) {
+Warn(createFileError(M.MemberName, std::move(Err)));
+ });
if (!SymFileOrErr)
return createFileError(M.MemberName, SymFileOrErr.takeError());
SymFiles.push_back(std::move(*SymFileOrErr));
@@ -1031,7 +1031,8 @@ Expected
computeArchiveRelativePath(StringRef From, StringRef To) {
static Error
writeArchiveToStream(raw_ostream &Out, ArrayRef NewMembers,
SymtabWritingMode WriteSymtab, object::Archive::Kind Kind,
- bool Deterministic, bool Thin, std::optional IsEC) {
+ bool Deterministic, bool Thin, std::optional IsEC,
+ function_ref Warn) {
assert((!Thin || !isBSDLike(Kind)) && "Only the gnu format has a thin mode");
SmallString<0> SymNamesBuf;
@@ -1053,7 +1054,7 @@ writeArchiveToStream(raw_ostream &Out,
ArrayRef NewMembers,
Expected> DataOrErr = computeMemberData(
StringTable, SymNames, Kind, Thin, Deterministic, WriteSymtab,
- isCOFFArchive(Kind) ? &SymMap : nullptr, Context, NewMembers, IsEC);
+ isCOFFArchive(Kind) ? &SymMap : nullptr, Context, NewMembers, IsEC,
Warn);
if (Error E = DataOrErr.takeError())
return E;
std::vector &Da
[llvm-branch-commits] Improve tests, change API (PR #98235)
https://github.com/pcc closed https://github.com/llvm/llvm-project/pull/98235 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Match functions with call graph (PR #98125)
https://github.com/shawbyoung updated
https://github.com/llvm/llvm-project/pull/98125
>From cf32a43e7c2b04079c6123fe13df4fb7226d771f Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 9 Jul 2024 10:04:25 -0700
Subject: [PATCH 1/3] Comments
Created using spr 1.3.4
---
bolt/lib/Profile/YAMLProfileReader.cpp | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/bolt/lib/Profile/YAMLProfileReader.cpp
b/bolt/lib/Profile/YAMLProfileReader.cpp
index 69ea0899c5f2c..6753337c24ea7 100644
--- a/bolt/lib/Profile/YAMLProfileReader.cpp
+++ b/bolt/lib/Profile/YAMLProfileReader.cpp
@@ -501,7 +501,6 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
// Maps binary functions to adjacent functions in the FCG.
for (const BinaryFunction *CallerBF : BFs) {
-// Add all call targets to the hash map.
for (const BinaryBasicBlock &BB : CallerBF->blocks()) {
for (const MCInst &Inst : BB) {
if (!BC.MIB->isCall(Instr))
@@ -533,7 +532,8 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
}
}
- // Create mapping from neighbor hash to BFs.
+ // Using the constructed adjacent function mapping, creates mapping from
+ // neighbor hash to BFs.
std::unordered_map>
NeighborHashToBFs;
for (const BinaryFunction *BF : BFs) {
@@ -552,12 +552,12 @@ size_t
YAMLProfileReader::matchWithCallGraph(BinaryContext &BC) {
.push_back(BF);
}
- // TODO: change call anchor PR to have this representation - we need it here
+ // TODO: note, this will be introduced in the matching functions with calls
+ // as anchors pr
DenseMap
IdToYAMLBF;
- // TODO: change call anchor PR to have this representation - we need it here
- // Maps hashes to profiled functions.
+ // Maps YAML functions to adjacent functions in the profile FCG.
std::unordered_map
YamlBFToHashes(BFs.size());
@@ -590,7 +590,7 @@ size_t YAMLProfileReader::matchWithCallGraph(BinaryContext
&BC) {
}
}
- // Matching YAMLBF with neighbor hashes.
+ // Matches YAMLBF to BFs with neighbor hashes.
for (yaml::bolt::BinaryFunctionProfile &YamlBF : YamlBP.Functions) {
if (YamlBF.Used)
continue;
>From ee9049fc4bd3d4203c19c9c0982a78ab3b47666f Mon Sep 17 00:00:00 2001
From: shawbyoung
Date: Tue, 9 Jul 2024 13:52:05 -0700
Subject: [PATCH 2/3] Moved blended hash definition
Created using spr 1.3.4
---
bolt/include/bolt/Profile/YAMLProfileReader.h | 69 ++-
bolt/lib/Profile/StaleProfileMatching.cpp | 65 ---
bolt/lib/Profile/YAMLProfileReader.cpp| 110 --
3 files changed, 119 insertions(+), 125 deletions(-)
diff --git a/bolt/include/bolt/Profile/YAMLProfileReader.h
b/bolt/include/bolt/Profile/YAMLProfileReader.h
index 36e8f8739eee1..e8a34ecad9a08 100644
--- a/bolt/include/bolt/Profile/YAMLProfileReader.h
+++ b/bolt/include/bolt/Profile/YAMLProfileReader.h
@@ -16,6 +16,73 @@
namespace llvm {
namespace bolt {
+/// An object wrapping several components of a basic block hash. The combined
+/// (blended) hash is represented and stored as one uint64_t, while individual
+/// components are of smaller size (e.g., uint16_t or uint8_t).
+struct BlendedBlockHash {
+private:
+ using ValueOffset = Bitfield::Element;
+ using ValueOpcode = Bitfield::Element;
+ using ValueInstr = Bitfield::Element;
+ using ValuePred = Bitfield::Element;
+ using ValueSucc = Bitfield::Element;
+
+public:
+ explicit BlendedBlockHash() {}
+
+ explicit BlendedBlockHash(uint64_t Hash) {
+Offset = Bitfield::get(Hash);
+OpcodeHash = Bitfield::get(Hash);
+InstrHash = Bitfield::get(Hash);
+PredHash = Bitfield::get(Hash);
+SuccHash = Bitfield::get(Hash);
+ }
+
+ /// Combine the blended hash into uint64_t.
+ uint64_t combine() const {
+uint64_t Hash = 0;
+Bitfield::set(Hash, Offset);
+Bitfield::set(Hash, OpcodeHash);
+Bitfield::set(Hash, InstrHash);
+Bitfield::set(Hash, PredHash);
+Bitfield::set(Hash, SuccHash);
+return Hash;
+ }
+
+ /// Compute a distance between two given blended hashes. The smaller the
+ /// distance, the more similar two blocks are. For identical basic blocks,
+ /// the distance is zero.
+ uint64_t distance(const BlendedBlockHash &BBH) const {
+assert(OpcodeHash == BBH.OpcodeHash &&
+ "incorrect blended hash distance computation");
+uint64_t Dist = 0;
+// Account for NeighborHash
+Dist += SuccHash == BBH.SuccHash ? 0 : 1;
+Dist += PredHash == BBH.PredHash ? 0 : 1;
+Dist <<= 16;
+// Account for InstrHash
+Dist += InstrHash == BBH.InstrHash ? 0 : 1;
+Dist <<= 16;
+// Account for Offset
+Dist += (Offset >= BBH.Offset ? Offset - BBH.Offset : BBH.Offset - Offset);
+return Dist;
+ }
+
+ /// The offset of the basic block from the function start.
+ uint16_t Offset{0};
+ /// (Loose) Hash of the basic block instructions, excluding operands.
+ uint16_t OpcodeHash{0};
+ /// (Strong)
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
https://github.com/dcci approved this pull request. LGTM https://github.com/llvm/llvm-project/pull/96596 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Match blocks with calls as anchors (PR #96596)
dcci wrote: (assuming Amir is happy) https://github.com/llvm/llvm-project/pull/96596 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [libc] b7fa6ce - Revert "[libc] Move off_t and stdio macros to proxy hdrs (#98215)"
Author: Michael Jones
Date: 2024-07-09T16:22:13-07:00
New Revision: b7fa6cee242086ae030a74cc3894c22ecc79b6c4
URL:
https://github.com/llvm/llvm-project/commit/b7fa6cee242086ae030a74cc3894c22ecc79b6c4
DIFF:
https://github.com/llvm/llvm-project/commit/b7fa6cee242086ae030a74cc3894c22ecc79b6c4.diff
LOG: Revert "[libc] Move off_t and stdio macros to proxy hdrs (#98215)"
This reverts commit 240ec5a37517f55a64cd0c69fd4fa1613ab30f97.
Added:
Modified:
libc/config/gpu/api.td
libc/config/linux/api.td
libc/hdr/CMakeLists.txt
libc/hdr/types/CMakeLists.txt
libc/include/llvm-libc-macros/stdio-macros.h
libc/newhdrgen/yaml/stdio.yaml
libc/src/__support/File/CMakeLists.txt
libc/src/__support/File/file.cpp
libc/src/__support/File/file.h
libc/src/__support/File/linux/CMakeLists.txt
libc/src/__support/File/linux/file.cpp
libc/src/__support/File/linux/file.h
libc/src/__support/File/linux/lseekImpl.h
libc/src/__support/OSUtil/linux/CMakeLists.txt
libc/src/__support/OSUtil/linux/fcntl.cpp
libc/src/stdio/CMakeLists.txt
libc/src/stdio/fopencookie.cpp
libc/src/stdio/setbuf.cpp
utils/bazel/llvm-project-overlay/libc/BUILD.bazel
Removed:
libc/hdr/stdio_macros.h
libc/hdr/types/off_t.h
diff --git a/libc/config/gpu/api.td b/libc/config/gpu/api.td
index 21ddbb95b70c9..523ad49ffa3fd 100644
--- a/libc/config/gpu/api.td
+++ b/libc/config/gpu/api.td
@@ -59,6 +59,11 @@ def FenvAPI: PublicAPI<"fenv.h"> {
}
def StdIOAPI : PublicAPI<"stdio.h"> {
+ let Macros = [
+SimpleMacroDef<"_IOFBF", "0">,
+SimpleMacroDef<"_IOLBF", "1">,
+SimpleMacroDef<"_IONBF", "2">,
+ ];
let Types = [
"FILE",
"off_t",
diff --git a/libc/config/linux/api.td b/libc/config/linux/api.td
index 60e9b70f0d8a4..eb0090c80b0da 100644
--- a/libc/config/linux/api.td
+++ b/libc/config/linux/api.td
@@ -76,6 +76,9 @@ def StdIOAPI : PublicAPI<"stdio.h"> {
SimpleMacroDef<"stderr", "stderr">,
SimpleMacroDef<"stdin", "stdin">,
SimpleMacroDef<"stdout", "stdout">,
+SimpleMacroDef<"_IOFBF", "0">,
+SimpleMacroDef<"_IOLBF", "1">,
+SimpleMacroDef<"_IONBF", "2">,
];
let Types = [
"FILE",
diff --git a/libc/hdr/CMakeLists.txt b/libc/hdr/CMakeLists.txt
index da640e8c0f70c..1303280c2c5ef 100644
--- a/libc/hdr/CMakeLists.txt
+++ b/libc/hdr/CMakeLists.txt
@@ -69,16 +69,6 @@ add_proxy_header_library(
libc.include.signal
)
-add_proxy_header_library(
- stdio_macros
- HDRS
-stdio_macros.h
- FULL_BUILD_DEPENDS
-libc.include.stdio
-libc.include.llvm-libc-macros.stdio_macros
-libc.include.llvm-libc-macros.file_seek_macros
-)
-
add_proxy_header_library(
sys_epoll_macros
HDRS
diff --git a/libc/hdr/stdio_macros.h b/libc/hdr/stdio_macros.h
deleted file mode 100644
index a212846dd8f41..0
--- a/libc/hdr/stdio_macros.h
+++ /dev/null
@@ -1,23 +0,0 @@
-//===-- Definition of macros from stdio.h ===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===--===//
-
-#ifndef LLVM_LIBC_HDR_STDIO_MACROS_H
-#define LLVM_LIBC_HDR_STDIO_MACROS_H
-
-#ifdef LIBC_FULL_BUILD
-
-#include "include/llvm-libc-macros/file-seek-macros.h"
-#include "include/llvm-libc-macros/stdio-macros.h"
-
-#else // Overlay mode
-
-#include
-
-#endif // LLVM_LIBC_FULL_BUILD
-
-#endif // LLVM_LIBC_HDR_STDIO_MACROS_H
diff --git a/libc/hdr/types/CMakeLists.txt b/libc/hdr/types/CMakeLists.txt
index f66f2cf5dda9f..1cab1d3b812cf 100644
--- a/libc/hdr/types/CMakeLists.txt
+++ b/libc/hdr/types/CMakeLists.txt
@@ -135,12 +135,3 @@ add_proxy_header_library(
libc.include.llvm-libc-types.struct_sigaction
libc.include.signal
)
-
-add_proxy_header_library(
- off_t
- HDRS
-off_t.h
- FULL_BUILD_DEPENDS
-libc.include.llvm-libc-types.off_t
-libc.include.stdio
-)
diff --git a/libc/hdr/types/off_t.h b/libc/hdr/types/off_t.h
deleted file mode 100644
index abc3aa659365f..0
--- a/libc/hdr/types/off_t.h
+++ /dev/null
@@ -1,22 +0,0 @@
-//===-- Proxy for off_t
---===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===--===//
-
-#ifndef LLVM_LIBC_HDR_TYPES_OFF_T_H
-#define LLVM_LIBC_HDR_TYPES_OFF_T_H
-
-#ifdef LIBC_FULL_BUILD
-
-#include "include/llvm-libc-types/off_t.h"
-
-#else // Overlay mode
-
-#include
-
-#endif // LLVM_LIBC_FULL_BUILD
-
-#endif // LLVM_LIBC_HDR_TYPES_OFF_T_H
diff --git a/libc
[llvm-branch-commits] [compiler-rt] Cleanup use of COMPILER_RT_INCLUDE_TESTS (PR #98246)
https://github.com/vitalybuka created https://github.com/llvm/llvm-project/pull/98246 1. Move checks into parent test/CMakeLists.txt 2. COMPILER_RT_INCLUDE_TESTS disable both lit and gtests. Before it was very inconsistent between sanitizers. ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [compiler-rt] Cleanup use of COMPILER_RT_INCLUDE_TESTS (PR #98246)
llvmbot wrote:
@llvm/pr-subscribers-compiler-rt-sanitizer
Author: Vitaly Buka (vitalybuka)
Changes
1. Move checks into parent test/CMakeLists.txt
2. COMPILER_RT_INCLUDE_TESTS disable both lit and
gtests. Before it was very inconsistent between
sanitizers.
---
Full diff: https://github.com/llvm/llvm-project/pull/98246.diff
14 Files Affected:
- (modified) compiler-rt/cmake/base-config-ix.cmake (+2-2)
- (modified) compiler-rt/test/CMakeLists.txt (+1-1)
- (modified) compiler-rt/test/asan/CMakeLists.txt (+27-29)
- (modified) compiler-rt/test/ctx_profile/CMakeLists.txt (+9-11)
- (modified) compiler-rt/test/fuzzer/CMakeLists.txt (+7-11)
- (modified) compiler-rt/test/gwp_asan/CMakeLists.txt (+2-2)
- (modified) compiler-rt/test/interception/CMakeLists.txt (+1-1)
- (modified) compiler-rt/test/memprof/CMakeLists.txt (+9-11)
- (modified) compiler-rt/test/msan/CMakeLists.txt (+1-3)
- (modified) compiler-rt/test/nsan/CMakeLists.txt (+1-3)
- (modified) compiler-rt/test/rtsan/CMakeLists.txt (+10-12)
- (modified) compiler-rt/test/sanitizer_common/CMakeLists.txt (+7-9)
- (modified) compiler-rt/test/scudo/standalone/CMakeLists.txt (+16-18)
- (modified) compiler-rt/test/tsan/CMakeLists.txt (+10-12)
``diff
diff --git a/compiler-rt/cmake/base-config-ix.cmake
b/compiler-rt/cmake/base-config-ix.cmake
index 87137b3d111b6..5b52a30a94aaf 100644
--- a/compiler-rt/cmake/base-config-ix.cmake
+++ b/compiler-rt/cmake/base-config-ix.cmake
@@ -43,7 +43,7 @@ if (LLVM_TREE_AVAILABLE)
get_clang_resource_dir(COMPILER_RT_OUTPUT_DIR PREFIX
${LLVM_LIBRARY_OUTPUT_INTDIR}/..)
set(COMPILER_RT_EXEC_OUTPUT_DIR ${LLVM_RUNTIME_OUTPUT_INTDIR})
get_clang_resource_dir(COMPILER_RT_INSTALL_PATH)
- option(COMPILER_RT_INCLUDE_TESTS "Generate and build compiler-rt unit tests."
+ option(COMPILER_RT_INCLUDE_TESTS "Generate and build compiler-rt tests."
${LLVM_INCLUDE_TESTS})
option(COMPILER_RT_ENABLE_WERROR "Fail and stop if warning is triggered"
${LLVM_ENABLE_WERROR})
@@ -70,7 +70,7 @@ else()
"Path where built compiler-rt executables should be stored.")
set(COMPILER_RT_INSTALL_PATH "" CACHE PATH
"Prefix for directories where built compiler-rt artifacts should be
installed.")
- option(COMPILER_RT_INCLUDE_TESTS "Generate and build compiler-rt unit
tests." OFF)
+ option(COMPILER_RT_INCLUDE_TESTS "Generate and build compiler-rt tests." OFF)
option(COMPILER_RT_ENABLE_WERROR "Fail and stop if warning is triggered" OFF)
# Use a host compiler to compile/link tests.
set(COMPILER_RT_TEST_COMPILER ${CMAKE_C_COMPILER} CACHE PATH "Compiler to
use for testing")
diff --git a/compiler-rt/test/CMakeLists.txt b/compiler-rt/test/CMakeLists.txt
index 5982260c2d854..3d52bb552ccda 100644
--- a/compiler-rt/test/CMakeLists.txt
+++ b/compiler-rt/test/CMakeLists.txt
@@ -48,7 +48,7 @@ umbrella_lit_testsuite_begin(check-compiler-rt)
function(compiler_rt_test_runtime runtime)
string(TOUPPER ${runtime} runtime_uppercase)
- if(COMPILER_RT_HAS_${runtime_uppercase})
+ if(COMPILER_RT_HAS_${runtime_uppercase} AND COMPILER_RT_INCLUDE_TESTS)
if (${runtime} STREQUAL scudo_standalone)
add_subdirectory(scudo/standalone)
else()
diff --git a/compiler-rt/test/asan/CMakeLists.txt
b/compiler-rt/test/asan/CMakeLists.txt
index 2d683e61d6954..fb9e81bbe8082 100644
--- a/compiler-rt/test/asan/CMakeLists.txt
+++ b/compiler-rt/test/asan/CMakeLists.txt
@@ -130,39 +130,37 @@ if(APPLE)
endif()
# Add unit tests.
-if(COMPILER_RT_INCLUDE_TESTS)
- foreach(arch ${ASAN_TEST_ARCH})
-string(TOUPPER ${arch} ARCH_UPPER_CASE)
-set(CONFIG_NAME ${ARCH_UPPER_CASE}${OS_NAME}Config)
-set(CONFIG_NAME_DYNAMIC ${ARCH_UPPER_CASE}${OS_NAME}DynamicConfig)
+foreach(arch ${ASAN_TEST_ARCH})
+ string(TOUPPER ${arch} ARCH_UPPER_CASE)
+ set(CONFIG_NAME ${ARCH_UPPER_CASE}${OS_NAME}Config)
+ set(CONFIG_NAME_DYNAMIC ${ARCH_UPPER_CASE}${OS_NAME}DynamicConfig)
-if(NOT MINGW)
- # MinGW environments don't provide a statically linked CRT, so only the
- # dynamic asan test configuration can be expected to work.
- set(ASAN_TEST_DYNAMIC False)
- configure_lit_site_cfg(
-${CMAKE_CURRENT_SOURCE_DIR}/Unit/lit.site.cfg.py.in
-${CMAKE_CURRENT_BINARY_DIR}/Unit/${CONFIG_NAME}/lit.site.cfg.py)
+ if(NOT MINGW)
+# MinGW environments don't provide a statically linked CRT, so only the
+# dynamic asan test configuration can be expected to work.
+set(ASAN_TEST_DYNAMIC False)
+configure_lit_site_cfg(
+ ${CMAKE_CURRENT_SOURCE_DIR}/Unit/lit.site.cfg.py.in
+ ${CMAKE_CURRENT_BINARY_DIR}/Unit/${CONFIG_NAME}/lit.site.cfg.py)
+ endif()
+ if(COMPILER_RT_ASAN_HAS_STATIC_RUNTIME)
+set(ASAN_TEST_DYNAMIC True)
+configure_lit_site_cfg(
+ ${CMAKE_CURRENT_SOURCE_DIR}/Unit/lit.site.cfg.py.in
+ ${CMAKE_CURRENT_BINARY_DIR}/Unit/${CONFIG_NAME_DYNAMIC}/lit.site.cfg.py)
+ endif()
+ # FIXME: support unit test in the android test runner
+ if (NOT ANDR
[llvm-branch-commits] [compiler-rt] Cleanup use of COMPILER_RT_INCLUDE_TESTS (PR #98246)
https://github.com/MaskRay approved this pull request.
I am not familiar with how the tests cmake work, but this patch indeed improves
consistency.
For example, `foreach(arch ${MEMPROF_TEST_ARCH})` is at the toplevel while its
unittest `foreach` is guarded by `COMPILER_RT_INCLUDE_TESTS`. Removing
`COMPILER_RT_INCLUDE_TESTS` is great.
https://github.com/llvm/llvm-project/pull/98246
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Added more details on heatmap docs. (PR #98162)
https://github.com/paschalis-mpeis ready_for_review https://github.com/llvm/llvm-project/pull/98162 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [EarlyIfCvt] Take branch probablities into consideration (PR #97808)
@@ -913,6 +913,10 @@ class TargetInstrInfo : public MCInstrInfo {
return false;
}
+ /// Return true if the target will always try to convert predictable branches
+ /// to selects.
+ virtual bool shouldConvertPredictableBranches() const { return true; }
+
mgudim wrote:
The default should be `false`
https://github.com/llvm/llvm-project/pull/97808
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [EarlyIfCvt] Take branch probablities into consideration (PR #97808)
mgudim wrote: > [EarlyIfCvt] Take branch probablities into consideration It looks like this MR is only adding a target hook, so this title doesn't make sense to me https://github.com/llvm/llvm-project/pull/97808 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
