maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/Attributor/callbacks.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

454 lines
33 KiB
LLVM
Raw Normal View History

[IR] Value::getPointerAlignment(): handle pointer constants Summary: New `@test13` in `Attributor/align.ll` is the main motivation - `null` pointer really does not limit our alignment knowledge, in fact it is fully aligned since it has no bits set. Here we don't special-case `null` pointer because it is somewhat controversial to add one more place where we enforce that `null` pointer is zero, but instead we do the more general thing of trying to perform constant-fold of pointer constant to an integer, and perform alignment inferrment on that. Reviewers: jdoerfert, gchatelet, courbet, sstefan1 Reviewed By: jdoerfert Subscribers: hiraditya, arphaman, jfb, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D73131
2020-01-22 06:04:47 +08:00
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --scrub-attributes
[Attributor][FIX] Record dependences for assumed dead abstract attributes In a recent patch we introduced a problem with abstract attributes that were assumed dead at some point. Since `Attributor::updateAA` was introduced in 95e0d28b71e42c9b7cd77c96f728311981a021f6, we did not remember the dependence on the liveness AA when an abstract attribute was assumed dead and therefore not updated. Explicit reproducer added in liveness.ll. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 509242 (345483/s) temporary memory allocations: 98666 (66937/s) peak heap memory consumption: 18.60MB peak RSS (including heaptrack overhead): 103.29MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 529332 (355494/s) temporary memory allocations: 102107 (68574/s) peak heap memory consumption: 19.40MB peak RSS (including heaptrack overhead): 102.79MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 20090 (1339333/s) temporary memory allocations: 3441 (229400/s) peak heap memory consumption: 801.45KB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ```
2020-05-08 05:50:50 +08:00
; RUN: opt -attributor -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=4 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_NPM,NOT_CGSCC_OPM,NOT_TUNIT_NPM,IS__TUNIT____,IS________OPM,IS__TUNIT_OPM
[Attributor][NFC] Remove obsolete option from tests Since D76871 it is sufficient to run `opt -atributor` or `-attributor-cgscc`.
2020-04-18 10:09:16 +08:00
; RUN: opt -aa-pipeline=basic-aa -passes=attributor -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=3 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_OPM,NOT_CGSCC_NPM,NOT_TUNIT_OPM,IS__TUNIT____,IS________NPM,IS__TUNIT_NPM
; RUN: opt -attributor-cgscc -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_TUNIT_NPM,NOT_TUNIT_OPM,NOT_CGSCC_NPM,IS__CGSCC____,IS________OPM,IS__CGSCC_OPM
; RUN: opt -aa-pipeline=basic-aa -passes=attributor-cgscc -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_TUNIT_NPM,NOT_TUNIT_OPM,NOT_CGSCC_OPM,IS__CGSCC____,IS________NPM,IS__CGSCC_NPM
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
[Attributor] Use abstract call sites for call site callback Summary: When we iterate over uses of functions and expect them to be call sites, we now use abstract call sites to allow callback calls. Reviewers: sstefan1, uenoku Subscribers: hiraditya, bollu, hfinkel, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D67871 llvm-svn: 373985
2019-10-08 07:14:58 +08:00
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
; Test 0
;
; Make sure we propagate information from the caller to the callback callee but
; only for arguments that are mapped through the callback metadata. Here, the
; first two arguments of the call and the callback callee do not correspond to
; each other but argument 3-5 of the transitive call site in the caller match
; arguments 2-4 of the callback callee. Here we should see information and value
; transfer in both directions.
define void @t0_caller(i32* %a) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-LABEL: define {{[^@]+}}@t0_caller
; IS__TUNIT_OPM-SAME: (i32* align 256 [[A:%.*]])
; IS__TUNIT_OPM-NEXT: entry:
; IS__TUNIT_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t0_callback_broker(i32* noalias nocapture align 536870912 null, i32* nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t0_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 undef, i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t0_caller
; IS__TUNIT_NPM-SAME: (i32* align 256 [[A:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t0_callback_broker(i32* noalias nocapture align 536870912 null, i32* nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t0_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 undef, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@t0_caller
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-SAME: (i32* align 256 [[A:%.*]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: entry:
; IS__CGSCC_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t0_callback_broker(i32* noalias nocapture align 536870912 null, i32* nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t0_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 99, i32** nonnull align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t0_caller
; IS__CGSCC_NPM-SAME: (i32* align 256 [[A:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t0_callback_broker(i32* noalias nocapture align 536870912 null, i32* nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t0_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 99, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Use abstract call sites for call site callback Summary: When we iterate over uses of functions and expect them to be call sites, we now use abstract call sites to allow callback calls. Reviewers: sstefan1, uenoku Subscribers: hiraditya, bollu, hfinkel, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D67871 llvm-svn: 373985
2019-10-08 07:14:58 +08:00
;
entry:
%b = alloca i32, align 32
%c = alloca i32*, align 64
%ptr = alloca i32, align 128
%0 = bitcast i32* %b to i8*
store i32 42, i32* %b, align 4
store i32* %b, i32** %c, align 8
call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t0_callback_broker(i32* null, i32* %ptr, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t0_callback_callee to void (i32*, i32*, ...)*), i32* %a, i64 99, i32** %c)
ret void
}
; Note that the first two arguments are provided by the callback_broker according to the callback in !1 below!
; The others are annotated with alignment information, amongst others, or even replaced by the constants passed to the call.
define internal void @t0_callback_callee(i32* %is_not_null, i32* %ptr, i32* %a, i64 %b, i32** %c) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
;
; IS________OPM-LABEL: define {{[^@]+}}@t0_callback_callee
; IS________OPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS________OPM-NEXT: entry:
; IS________OPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS________OPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS________OPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
; IS________OPM-NEXT: tail call void @t0_check(i32* align 256 [[A]], i64 99, i32* [[TMP0]])
; IS________OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t0_callback_callee
; IS__TUNIT_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* align 256 [[A:%.*]], i64 [[B:%.*]], i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__TUNIT_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
; IS__TUNIT_NPM-NEXT: tail call void @t0_check(i32* align 256 [[A]], i64 99, i32* [[TMP0]])
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t0_callback_callee
; IS__CGSCC_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 8 dereferenceable(8) [[C:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__CGSCC_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 8
; IS__CGSCC_NPM-NEXT: tail call void @t0_check(i32* align 256 [[A]], i64 99, i32* [[TMP0]])
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Use abstract call sites for call site callback Summary: When we iterate over uses of functions and expect them to be call sites, we now use abstract call sites to allow callback calls. Reviewers: sstefan1, uenoku Subscribers: hiraditya, bollu, hfinkel, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D67871 llvm-svn: 373985
2019-10-08 07:14:58 +08:00
;
entry:
%ptr_val = load i32, i32* %ptr, align 8
store i32 %ptr_val, i32* %is_not_null
%0 = load i32*, i32** %c, align 8
tail call void @t0_check(i32* %a, i64 %b, i32* %0)
ret void
}
declare void @t0_check(i32* align 256, i64, i32*)
declare !callback !0 void @t0_callback_broker(i32*, i32*, void (i32*, i32*, ...)*, ...)
[Attributor] Pointer privatization attribute (argument promotion) A pointer is privatizeable if it can be replaced by a new, private one. Privatizing pointer reduces the use count, interaction between unrelated code parts. This is a first step towards replacing argument promotion. While we can already handle recursion (unlike argument promotion!) we are restricted to stack allocations for now because we do not analyze the uses in the callee. Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D68852
2019-10-31 06:20:20 +08:00
; Test 1
;
; Similar to test 0 but with some additional annotations (noalias/nocapute) to make sure
; we deduce and propagate noalias and others properly.
define void @t1_caller(i32* noalias %a) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-LABEL: define {{[^@]+}}@t1_caller
; IS__TUNIT_OPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__TUNIT_OPM-NEXT: entry:
; IS__TUNIT_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t1_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t1_callback_callee to void (i32*, i32*, ...)*), i32* nocapture align 256 [[A]], i64 undef, i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t1_caller
; IS__TUNIT_NPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t1_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t1_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 undef, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@t1_caller
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-SAME: (i32* noalias align 256 [[A:%.*]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: entry:
; IS__CGSCC_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t1_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t1_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 99, i32** nonnull align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t1_caller
; IS__CGSCC_NPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t1_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t1_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 99, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Pointer privatization attribute (argument promotion) A pointer is privatizeable if it can be replaced by a new, private one. Privatizing pointer reduces the use count, interaction between unrelated code parts. This is a first step towards replacing argument promotion. While we can already handle recursion (unlike argument promotion!) we are restricted to stack allocations for now because we do not analyze the uses in the callee. Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D68852
2019-10-31 06:20:20 +08:00
;
entry:
%b = alloca i32, align 32
%c = alloca i32*, align 64
%ptr = alloca i32, align 128
%0 = bitcast i32* %b to i8*
store i32 42, i32* %b, align 4
store i32* %b, i32** %c, align 8
call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t1_callback_broker(i32* null, i32* %ptr, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t1_callback_callee to void (i32*, i32*, ...)*), i32* %a, i64 99, i32** %c)
ret void
}
; Note that the first two arguments are provided by the callback_broker according to the callback in !1 below!
; The others are annotated with alignment information, amongst others, or even replaced by the constants passed to the call.
define internal void @t1_callback_callee(i32* %is_not_null, i32* %ptr, i32* %a, i64 %b, i32** %c) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
;
; IS________OPM-LABEL: define {{[^@]+}}@t1_callback_callee
; IS________OPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS________OPM-NEXT: entry:
; IS________OPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS________OPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS________OPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS________OPM-NEXT: tail call void @t1_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS________OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t1_callback_callee
; IS__TUNIT_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* noalias nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__TUNIT_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: tail call void @t1_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t1_callback_callee
; IS__CGSCC_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 8 dereferenceable(8) [[C:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__CGSCC_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 8
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: tail call void @t1_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Pointer privatization attribute (argument promotion) A pointer is privatizeable if it can be replaced by a new, private one. Privatizing pointer reduces the use count, interaction between unrelated code parts. This is a first step towards replacing argument promotion. While we can already handle recursion (unlike argument promotion!) we are restricted to stack allocations for now because we do not analyze the uses in the callee. Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D68852
2019-10-31 06:20:20 +08:00
;
entry:
%ptr_val = load i32, i32* %ptr, align 8
store i32 %ptr_val, i32* %is_not_null
%0 = load i32*, i32** %c, align 8
tail call void @t1_check(i32* %a, i64 %b, i32* %0)
ret void
}
declare void @t1_check(i32* nocapture align 256, i64, i32* nocapture) nosync
declare !callback !0 void @t1_callback_broker(i32* nocapture , i32* nocapture , void (i32*, i32*, ...)* nocapture, ...)
; Test 2
;
; Similar to test 1 but checking that the noalias is only placed if potential synchronization through @t2_check is preserved.
define void @t2_caller(i32* noalias %a) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-LABEL: define {{[^@]+}}@t2_caller
; IS__TUNIT_OPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__TUNIT_OPM-NEXT: entry:
; IS__TUNIT_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t2_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t2_callback_callee to void (i32*, i32*, ...)*), i32* nocapture align 256 [[A]], i64 undef, i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t2_caller
; IS__TUNIT_NPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t2_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t2_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 undef, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@t2_caller
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-SAME: (i32* noalias align 256 [[A:%.*]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: entry:
; IS__CGSCC_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t2_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t2_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 99, i32** nonnull align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t2_caller
; IS__CGSCC_NPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t2_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t2_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 99, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Pointer privatization attribute (argument promotion) A pointer is privatizeable if it can be replaced by a new, private one. Privatizing pointer reduces the use count, interaction between unrelated code parts. This is a first step towards replacing argument promotion. While we can already handle recursion (unlike argument promotion!) we are restricted to stack allocations for now because we do not analyze the uses in the callee. Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D68852
2019-10-31 06:20:20 +08:00
;
entry:
%b = alloca i32, align 32
%c = alloca i32*, align 64
%ptr = alloca i32, align 128
%0 = bitcast i32* %b to i8*
store i32 42, i32* %b, align 4
store i32* %b, i32** %c, align 8
call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t2_callback_broker(i32* null, i32* %ptr, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t2_callback_callee to void (i32*, i32*, ...)*), i32* %a, i64 99, i32** %c)
ret void
}
; Note that the first two arguments are provided by the callback_broker according to the callback in !1 below!
; The others are annotated with alignment information, amongst others, or even replaced by the constants passed to the call.
;
; FIXME: We should derive noalias for %a and add a "fake use" of %a in all potentially synchronizing calls.
define internal void @t2_callback_callee(i32* %is_not_null, i32* %ptr, i32* %a, i64 %b, i32** %c) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
;
; IS________OPM-LABEL: define {{[^@]+}}@t2_callback_callee
; IS________OPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS________OPM-NEXT: entry:
; IS________OPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS________OPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS________OPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS________OPM-NEXT: tail call void @t2_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS________OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t2_callback_callee
; IS__TUNIT_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__TUNIT_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: tail call void @t2_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t2_callback_callee
; IS__CGSCC_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 8 dereferenceable(8) [[C:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__CGSCC_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 8
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: tail call void @t2_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Pointer privatization attribute (argument promotion) A pointer is privatizeable if it can be replaced by a new, private one. Privatizing pointer reduces the use count, interaction between unrelated code parts. This is a first step towards replacing argument promotion. While we can already handle recursion (unlike argument promotion!) we are restricted to stack allocations for now because we do not analyze the uses in the callee. Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D68852
2019-10-31 06:20:20 +08:00
;
entry:
%ptr_val = load i32, i32* %ptr, align 8
store i32 %ptr_val, i32* %is_not_null
%0 = load i32*, i32** %c, align 8
tail call void @t2_check(i32* %a, i64 %b, i32* %0)
ret void
}
declare void @t2_check(i32* nocapture align 256, i64, i32* nocapture)
declare !callback !0 void @t2_callback_broker(i32* nocapture , i32* nocapture , void (i32*, i32*, ...)* nocapture, ...)
[Attributor] Allow multiple uses of a casted function pointer If a function pointer is casted into a different type the resulting expression can be a constant. If so, it can be used multiple times which cannot be handled by the AbstractCallSite constructor alone. Instead, we follow the cast expression uses now explicitly during the call site traversal.
2020-01-26 10:24:38 +08:00
; Test 3
;
; Basically test 2 with the casted callback callee used twice.
define void @t3_caller(i32* noalias %a) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-LABEL: define {{[^@]+}}@t3_caller
; IS__TUNIT_OPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__TUNIT_OPM-NEXT: entry:
; IS__TUNIT_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* nocapture align 256 [[A]], i64 undef, i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
; IS__TUNIT_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* nocapture align 256 [[A]], i64 undef, i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t3_caller
; IS__TUNIT_NPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__TUNIT_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__TUNIT_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__TUNIT_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__TUNIT_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 undef, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
; IS__TUNIT_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture nonnull bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 undef, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_OPM-LABEL: define {{[^@]+}}@t3_caller
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-SAME: (i32* noalias align 256 [[A:%.*]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: entry:
; IS__CGSCC_OPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_OPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_OPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_OPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_OPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_OPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor] Inititialize "value attributes" w/ must-be-executed-context info Attributes that only depend on the value (=bit pattern) can be initialized from uses in the must-be-executed-context (MBEC). We did use `AAComposeTwoGenericDeduction` and `AAFromMustBeExecutedContext` before to do this for some positions of these attributes but not for all. This was fairly complicated and also problematic as we did run it in every `updateImpl` call even though we only use known information. The new implementation removes `AAComposeTwoGenericDeduction`* and `AAFromMustBeExecutedContext` in favor of a simple interface `AddInformation::fromMBEContext(...)` which we call from the `initialize` methods of the "value attribute" `Impl` classes, e.g. `AANonNullImpl:initialize`. There can be two types of test changes: 1) Artifacts were we miss some information that was known before a global fixpoint was reached and therefore available in an update but not at the beginning. 2) Deduction for values we did not derive via the MBEC before or which were not found as the `AAFromMustBeExecutedContext::updateImpl` was never invoked. * An improved version of AAComposeTwoGenericDeduction can be found in D78718. Once we find a new use case that implementation will be able to handle "generic" AAs better. --- Single run of the Attributor module and then CGSCC pass (oldPM) for SPASS/clause.c (~10k LLVM-IR loc): Before: ``` calls to allocation functions: 468428 (328952/s) temporary memory allocations: 77480 (54410/s) peak heap memory consumption: 32.71MB peak RSS (including heaptrack overhead): 122.46MB total memory leaked: 269.10KB ``` After: ``` calls to allocation functions: 554720 (351310/s) temporary memory allocations: 101650 (64376/s) peak heap memory consumption: 28.46MB peak RSS (including heaptrack overhead): 116.75MB total memory leaked: 269.10KB ``` Difference: ``` calls to allocation functions: 86292 (556722/s) temporary memory allocations: 24170 (155935/s) peak heap memory consumption: -4.25MB peak RSS (including heaptrack overhead): 0B total memory leaked: 0B ``` Reviewed By: uenoku Differential Revision: https://reviews.llvm.org/D78719
2020-04-23 06:49:58 +08:00
; IS__CGSCC_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 99, i32** nonnull align 64 dereferenceable(8) [[C]])
; IS__CGSCC_OPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* align 256 [[A]], i64 99, i32** nonnull align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_OPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t3_caller
; IS__CGSCC_NPM-SAME: (i32* noalias nocapture align 256 [[A:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[B:%.*]] = alloca i32, align 32
; IS__CGSCC_NPM-NEXT: [[C:%.*]] = alloca i32*, align 64
; IS__CGSCC_NPM-NEXT: [[PTR:%.*]] = alloca i32, align 128
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = bitcast i32* [[B]] to i8*
; IS__CGSCC_NPM-NEXT: store i32 42, i32* [[B]], align 32
; IS__CGSCC_NPM-NEXT: store i32* [[B]], i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 99, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
; IS__CGSCC_NPM-NEXT: call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* noalias nocapture align 536870912 null, i32* noalias nocapture nonnull align 128 dereferenceable(4) [[PTR]], void (i32*, i32*, ...)* nocapture bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* noalias nocapture align 256 [[A]], i64 99, i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Allow multiple uses of a casted function pointer If a function pointer is casted into a different type the resulting expression can be a constant. If so, it can be used multiple times which cannot be handled by the AbstractCallSite constructor alone. Instead, we follow the cast expression uses now explicitly during the call site traversal.
2020-01-26 10:24:38 +08:00
;
entry:
%b = alloca i32, align 32
%c = alloca i32*, align 64
%ptr = alloca i32, align 128
%0 = bitcast i32* %b to i8*
store i32 42, i32* %b, align 4
store i32* %b, i32** %c, align 8
call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* null, i32* %ptr, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* %a, i64 99, i32** %c)
call void (i32*, i32*, void (i32*, i32*, ...)*, ...) @t3_callback_broker(i32* null, i32* %ptr, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i32**)* @t3_callback_callee to void (i32*, i32*, ...)*), i32* %a, i64 99, i32** %c)
ret void
}
; Note that the first two arguments are provided by the callback_broker according to the callback in !1 below!
; The others are annotated with alignment information, amongst others, or even replaced by the constants passed to the call.
;
; FIXME: We should derive noalias for %a and add a "fake use" of %a in all potentially synchronizing calls.
define internal void @t3_callback_callee(i32* %is_not_null, i32* %ptr, i32* %a, i64 %b, i32** %c) {
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
;
; IS________OPM-LABEL: define {{[^@]+}}@t3_callback_callee
; IS________OPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS________OPM-NEXT: entry:
; IS________OPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS________OPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS________OPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS________OPM-NEXT: tail call void @t3_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS________OPM-NEXT: ret void
;
; IS__TUNIT_NPM-LABEL: define {{[^@]+}}@t3_callback_callee
; IS__TUNIT_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** noalias nocapture nonnull readonly align 64 dereferenceable(8) [[C:%.*]])
; IS__TUNIT_NPM-NEXT: entry:
; IS__TUNIT_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__TUNIT_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__TUNIT_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 64
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__TUNIT_NPM-NEXT: tail call void @t3_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__TUNIT_NPM-NEXT: ret void
;
; IS__CGSCC_NPM-LABEL: define {{[^@]+}}@t3_callback_callee
; IS__CGSCC_NPM-SAME: (i32* nocapture nonnull writeonly align 4 dereferenceable(4) [[IS_NOT_NULL:%.*]], i32* nocapture nonnull readonly align 8 dereferenceable(4) [[PTR:%.*]], i32* nocapture align 256 [[A:%.*]], i64 [[B:%.*]], i32** nocapture nonnull readonly align 8 dereferenceable(8) [[C:%.*]])
; IS__CGSCC_NPM-NEXT: entry:
; IS__CGSCC_NPM-NEXT: [[PTR_VAL:%.*]] = load i32, i32* [[PTR]], align 8
; IS__CGSCC_NPM-NEXT: store i32 [[PTR_VAL]], i32* [[IS_NOT_NULL]], align 4
; IS__CGSCC_NPM-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C]], align 8
[Attributor][NFC] Proactively ask for `nocapure` on call site arguments This minimizes test noise later on and is in line with other attributes we derive proactively.
2020-05-04 08:08:30 +08:00
; IS__CGSCC_NPM-NEXT: tail call void @t3_check(i32* nocapture align 256 [[A]], i64 99, i32* nocapture [[TMP0]])
[Attributor] Unify testing (=updates,prefixes,run configurations,...) When the Attributor was created the test update scripts were not well suited to deal with the challenges of IR attribute checking. This partially improved. Since then we also added three additional configurations that need testing; in total we now have the following four: { TUNIT, CGSCC } x { old pass manager (OPM), new pass manager (NPM) } Finally, the number of developers and tests grew rapidly (partially due to the addition of ArgumentPromotion and IPConstantProp tests), which resulted in tests only being run in some configurations, different prefixes being used, and different "styles" of checks being used. Due to the above reasons I believed we needed to take another look at the test update scripts. While we started to use them, via UTC_ARGS: --enable/disable, the other problems remained. To improve the testing situation for *all* configurations, to simplify future updates to the test, and to help identify subtle effects of future changes, we now use the test update scripts for (almost) all Attributor tests. An exhaustive prefix list minimizes the number of check lines and makes it easy to identify and compare configurations. Tests have been adjusted in the process but we tried to keep their intend unchanged. Reviewed By: sstefan1 Differential Revision: https://reviews.llvm.org/D76588
2020-04-09 08:04:57 +08:00
; IS__CGSCC_NPM-NEXT: ret void
[Attributor] Allow multiple uses of a casted function pointer If a function pointer is casted into a different type the resulting expression can be a constant. If so, it can be used multiple times which cannot be handled by the AbstractCallSite constructor alone. Instead, we follow the cast expression uses now explicitly during the call site traversal.
2020-01-26 10:24:38 +08:00
;
entry:
%ptr_val = load i32, i32* %ptr, align 8
store i32 %ptr_val, i32* %is_not_null
%0 = load i32*, i32** %c, align 8
tail call void @t3_check(i32* %a, i64 %b, i32* %0)
ret void
}
declare void @t3_check(i32* nocapture align 256, i64, i32* nocapture)
declare !callback !0 void @t3_callback_broker(i32* nocapture , i32* nocapture , void (i32*, i32*, ...)* nocapture, ...)
[Attributor] Use abstract call sites for call site callback Summary: When we iterate over uses of functions and expect them to be call sites, we now use abstract call sites to allow callback calls. Reviewers: sstefan1, uenoku Subscribers: hiraditya, bollu, hfinkel, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D67871 llvm-svn: 373985
2019-10-08 07:14:58 +08:00
!0 = !{!1}
!1 = !{i64 2, i64 -1, i64 -1, i1 true}