maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/test/CodeGen/ppc64-align-struct.c

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

173 lines
6.6 KiB
C
Raw Normal View History

Remove the -faltivec alias option and replace it with -maltivec everywhere. The alias was only ever used on darwin and had some issues there, and isn't used in practice much. Also fixes a problem with -mno-altivec not turning off -maltivec. Also add a diagnostic for faltivec/fno-altivec that directs users to use maltivec options and include the altivec.h file explicitly. llvm-svn: 298449
2017-03-22 06:06:18 +08:00
// RUN: %clang_cc1 -target-feature +altivec -triple powerpc64-unknown-linux-gnu -emit-llvm -o - %s | FileCheck %s
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
#include <stdarg.h>
struct test1 { int x; int y; };
struct test2 { int x; int y; } __attribute__((aligned (16)));
struct test3 { int x; int y; } __attribute__((aligned (32)));
struct test4 { int x; int y; int z; };
[PowerPC] Optimize passing certain aggregates by value In addition to enabling ELFv2 homogeneous aggregate handling, LLVM support to pass array types directly also enables a performance enhancement. We can now pass (non-homogeneous) aggregates that fit fully in registers as direct integer arrays, using an element type to encode the alignment requirement (that would otherwise go to the "byval align" field). This is preferable since "byval" forces the back-end to write the aggregate out to the stack, even if it could be passed fully in registers. This is particularly annoying on ELFv2, if there is no parameter save area available, since we then need to allocate space on the callee's stack just to hold those aggregates. Note that to implement this optimization, this patch does not attempt to fully anticipate register allocation rules as (defined in the ABI and) implemented in the back-end. Instead, the patch is simply passing *any* aggregate passed by value using the array mechanism if its size is up to 64 bytes. This means that some of those will end up being passed in stack slots anyway, but the generated code shouldn't be any worse either. (*Large* aggregates remain passed using "byval" to enable optimized copying via memcpy etc.) llvm-svn: 213495
2014-07-21 08:56:36 +08:00
struct test5 { int x[17]; };
struct test6 { int x[17]; } __attribute__((aligned (16)));
struct test7 { int x[17]; } __attribute__((aligned (32)));
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test1(i32 signext %x, i64 %y.coerce)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
void test1 (int x, struct test1 y)
{
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test2(i32 signext %x, [1 x i128] %y.coerce)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
void test2 (int x, struct test2 y)
{
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test3(i32 signext %x, [2 x i128] %y.coerce)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
void test3 (int x, struct test3 y)
{
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test4(i32 signext %x, [2 x i64] %y.coerce)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
void test4 (int x, struct test4 y)
{
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test5(i32 signext %x, %struct.test5* byval(%struct.test5) align 8 %y)
[PowerPC] Optimize passing certain aggregates by value In addition to enabling ELFv2 homogeneous aggregate handling, LLVM support to pass array types directly also enables a performance enhancement. We can now pass (non-homogeneous) aggregates that fit fully in registers as direct integer arrays, using an element type to encode the alignment requirement (that would otherwise go to the "byval align" field). This is preferable since "byval" forces the back-end to write the aggregate out to the stack, even if it could be passed fully in registers. This is particularly annoying on ELFv2, if there is no parameter save area available, since we then need to allocate space on the callee's stack just to hold those aggregates. Note that to implement this optimization, this patch does not attempt to fully anticipate register allocation rules as (defined in the ABI and) implemented in the back-end. Instead, the patch is simply passing *any* aggregate passed by value using the array mechanism if its size is up to 64 bytes. This means that some of those will end up being passed in stack slots anyway, but the generated code shouldn't be any worse either. (*Large* aggregates remain passed using "byval" to enable optimized copying via memcpy etc.) llvm-svn: 213495
2014-07-21 08:56:36 +08:00
void test5 (int x, struct test5 y)
{
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test6(i32 signext %x, %struct.test6* byval(%struct.test6) align 16 %y)
[PowerPC] Optimize passing certain aggregates by value In addition to enabling ELFv2 homogeneous aggregate handling, LLVM support to pass array types directly also enables a performance enhancement. We can now pass (non-homogeneous) aggregates that fit fully in registers as direct integer arrays, using an element type to encode the alignment requirement (that would otherwise go to the "byval align" field). This is preferable since "byval" forces the back-end to write the aggregate out to the stack, even if it could be passed fully in registers. This is particularly annoying on ELFv2, if there is no parameter save area available, since we then need to allocate space on the callee's stack just to hold those aggregates. Note that to implement this optimization, this patch does not attempt to fully anticipate register allocation rules as (defined in the ABI and) implemented in the back-end. Instead, the patch is simply passing *any* aggregate passed by value using the array mechanism if its size is up to 64 bytes. This means that some of those will end up being passed in stack slots anyway, but the generated code shouldn't be any worse either. (*Large* aggregates remain passed using "byval" to enable optimized copying via memcpy etc.) llvm-svn: 213495
2014-07-21 08:56:36 +08:00
void test6 (int x, struct test6 y)
{
}
// This case requires run-time realignment of the incoming struct
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK-LABEL: define dso_local void @test7(i32 signext %x, %struct.test7* byval(%struct.test7) align 16 %0)
[PowerPC] Optimize passing certain aggregates by value In addition to enabling ELFv2 homogeneous aggregate handling, LLVM support to pass array types directly also enables a performance enhancement. We can now pass (non-homogeneous) aggregates that fit fully in registers as direct integer arrays, using an element type to encode the alignment requirement (that would otherwise go to the "byval align" field). This is preferable since "byval" forces the back-end to write the aggregate out to the stack, even if it could be passed fully in registers. This is particularly annoying on ELFv2, if there is no parameter save area available, since we then need to allocate space on the callee's stack just to hold those aggregates. Note that to implement this optimization, this patch does not attempt to fully anticipate register allocation rules as (defined in the ABI and) implemented in the back-end. Instead, the patch is simply passing *any* aggregate passed by value using the array mechanism if its size is up to 64 bytes. This means that some of those will end up being passed in stack slots anyway, but the generated code shouldn't be any worse either. (*Large* aggregates remain passed using "byval" to enable optimized copying via memcpy etc.) llvm-svn: 213495
2014-07-21 08:56:36 +08:00
// CHECK: %y = alloca %struct.test7, align 32
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64
void test7 (int x, struct test7 y)
{
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test1va(%struct.test1* noalias sret %[[AGG_RESULT:.*]], i32 signext %x, ...)
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 8
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: store i8* %[[NEXT]], i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[T0:%.*]] = bitcast i8* %[[CUR]] to %struct.test1*
Generalize NRVO to cover C structs. This commit generalizes NRVO to cover C structs (both trivial and non-trivial structs). rdar://problem/33599681 Differential Revision: https://reviews.llvm.org/D44968 llvm-svn: 328809
2018-03-30 01:56:24 +08:00
// CHECK: [[DEST:%.*]] = bitcast %struct.test1* %[[AGG_RESULT]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[SRC:%.*]] = bitcast %struct.test1* [[T0]] to i8*
Change memcpy/memove/memset to have dest and source alignment attributes. Summary: This change is step three in the series of changes to remove alignment argument from memcpy/memmove/memset in favour of alignment attributes. Steps: Step 1) Remove alignment parameter and create alignment parameter attributes for memcpy/memmove/memset. ( rL322965, rC322964, rL322963 ) Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing source and dest alignments. ( rL323597 ) Step 3) Update Clang to use the new IRBuilder API. Step 4) Update Polly to use the new IRBuilder API. Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API, and those that use use MemIntrinsicInst::[get|set]Alignment() to use getDestAlignment() and getSourceAlignment() instead. Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the MemIntrinsicInst::[get|set]Alignment() methods. Reference http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.html http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html Reviewers: rjmccall Subscribers: jyknight, nemanjai, nhaehnle, javed.absar, sbc100, aheejin, kbarton, fedor.sergeev, cfe-commits Differential Revision: https://reviews.llvm.org/D41677 llvm-svn: 323617
2018-01-29 01:27:45 +08:00
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[DEST]], i8* align 8 [[SRC]], i64 8, i1 false)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
struct test1 test1va (int x, ...)
{
struct test1 y;
va_list ap;
va_start(ap, x);
y = va_arg (ap, struct test1);
va_end(ap);
return y;
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test2va(%struct.test2* noalias sret %[[AGG_RESULT:.*]], i32 signext %x, ...)
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: %[[TMP0:[^ ]+]] = ptrtoint i8* %[[CUR]] to i64
// CHECK: %[[TMP1:[^ ]+]] = add i64 %[[TMP0]], 15
// CHECK: %[[TMP2:[^ ]+]] = and i64 %[[TMP1]], -16
// CHECK: %[[ALIGN:[^ ]+]] = inttoptr i64 %[[TMP2]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[ALIGN]], i64 16
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: store i8* %[[NEXT]], i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[T0:%.*]] = bitcast i8* %[[ALIGN]] to %struct.test2*
Generalize NRVO to cover C structs. This commit generalizes NRVO to cover C structs (both trivial and non-trivial structs). rdar://problem/33599681 Differential Revision: https://reviews.llvm.org/D44968 llvm-svn: 328809
2018-03-30 01:56:24 +08:00
// CHECK: [[DEST:%.*]] = bitcast %struct.test2* %[[AGG_RESULT]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[SRC:%.*]] = bitcast %struct.test2* [[T0]] to i8*
Change memcpy/memove/memset to have dest and source alignment attributes (Step 1). Summary: Upstream LLVM is changing the the prototypes of the @llvm.memcpy/memmove/memset intrinsics. This change updates the Clang tests for this change. The @llvm.memcpy/memmove/memset intrinsics currently have an explicit argument which is required to be a constant integer. It represents the alignment of the dest (and source), and so must be the minimum of the actual alignment of the two. This change removes the alignment argument in favour of placing the alignment attribute on the source and destination pointers of the memory intrinsic call. For example, code which used to read: call void @llvm.memcpy.p0i8.p0i8.i32(i8* %dest, i8* %src, i32 100, i32 4, i1 false) will now read call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %dest, i8* align 4 %src, i32 100, i1 false) At this time the source and destination alignments must be the same (Step 1). Step 2 of the change, to be landed shortly, will relax that contraint and allow the source and destination to have different alignments. llvm-svn: 322964
2018-01-20 01:12:54 +08:00
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[DEST]], i8* align 16 [[SRC]], i64 16, i1 false)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
struct test2 test2va (int x, ...)
{
struct test2 y;
va_list ap;
va_start(ap, x);
y = va_arg (ap, struct test2);
va_end(ap);
return y;
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test3va(%struct.test3* noalias sret %[[AGG_RESULT:.*]], i32 signext %x, ...)
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: %[[TMP0:[^ ]+]] = ptrtoint i8* %[[CUR]] to i64
// CHECK: %[[TMP1:[^ ]+]] = add i64 %[[TMP0]], 15
// CHECK: %[[TMP2:[^ ]+]] = and i64 %[[TMP1]], -16
// CHECK: %[[ALIGN:[^ ]+]] = inttoptr i64 %[[TMP2]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[ALIGN]], i64 32
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: store i8* %[[NEXT]], i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[T0:%.*]] = bitcast i8* %[[ALIGN]] to %struct.test3*
Generalize NRVO to cover C structs. This commit generalizes NRVO to cover C structs (both trivial and non-trivial structs). rdar://problem/33599681 Differential Revision: https://reviews.llvm.org/D44968 llvm-svn: 328809
2018-03-30 01:56:24 +08:00
// CHECK: [[DEST:%.*]] = bitcast %struct.test3* %[[AGG_RESULT]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[SRC:%.*]] = bitcast %struct.test3* [[T0]] to i8*
Change memcpy/memove/memset to have dest and source alignment attributes. Summary: This change is step three in the series of changes to remove alignment argument from memcpy/memmove/memset in favour of alignment attributes. Steps: Step 1) Remove alignment parameter and create alignment parameter attributes for memcpy/memmove/memset. ( rL322965, rC322964, rL322963 ) Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing source and dest alignments. ( rL323597 ) Step 3) Update Clang to use the new IRBuilder API. Step 4) Update Polly to use the new IRBuilder API. Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API, and those that use use MemIntrinsicInst::[get|set]Alignment() to use getDestAlignment() and getSourceAlignment() instead. Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the MemIntrinsicInst::[get|set]Alignment() methods. Reference http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.html http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html Reviewers: rjmccall Subscribers: jyknight, nemanjai, nhaehnle, javed.absar, sbc100, aheejin, kbarton, fedor.sergeev, cfe-commits Differential Revision: https://reviews.llvm.org/D41677 llvm-svn: 323617
2018-01-29 01:27:45 +08:00
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 32 [[DEST]], i8* align 16 [[SRC]], i64 32, i1 false)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
struct test3 test3va (int x, ...)
{
struct test3 y;
va_list ap;
va_start(ap, x);
y = va_arg (ap, struct test3);
va_end(ap);
return y;
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @test4va(%struct.test4* noalias sret %[[AGG_RESULT:.*]], i32 signext %x, ...)
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 16
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: store i8* %[[NEXT]], i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[T0:%.*]] = bitcast i8* %[[CUR]] to %struct.test4*
Generalize NRVO to cover C structs. This commit generalizes NRVO to cover C structs (both trivial and non-trivial structs). rdar://problem/33599681 Differential Revision: https://reviews.llvm.org/D44968 llvm-svn: 328809
2018-03-30 01:56:24 +08:00
// CHECK: [[DEST:%.*]] = bitcast %struct.test4* %[[AGG_RESULT]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[SRC:%.*]] = bitcast %struct.test4* [[T0]] to i8*
Change memcpy/memove/memset to have dest and source alignment attributes. Summary: This change is step three in the series of changes to remove alignment argument from memcpy/memmove/memset in favour of alignment attributes. Steps: Step 1) Remove alignment parameter and create alignment parameter attributes for memcpy/memmove/memset. ( rL322965, rC322964, rL322963 ) Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing source and dest alignments. ( rL323597 ) Step 3) Update Clang to use the new IRBuilder API. Step 4) Update Polly to use the new IRBuilder API. Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API, and those that use use MemIntrinsicInst::[get|set]Alignment() to use getDestAlignment() and getSourceAlignment() instead. Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the MemIntrinsicInst::[get|set]Alignment() methods. Reference http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.html http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html Reviewers: rjmccall Subscribers: jyknight, nemanjai, nhaehnle, javed.absar, sbc100, aheejin, kbarton, fedor.sergeev, cfe-commits Differential Revision: https://reviews.llvm.org/D41677 llvm-svn: 323617
2018-01-29 01:27:45 +08:00
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[DEST]], i8* align 8 [[SRC]], i64 12, i1 false)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
struct test4 test4va (int x, ...)
{
struct test4 y;
va_list ap;
va_start(ap, x);
y = va_arg (ap, struct test4);
va_end(ap);
return y;
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @testva_longdouble(%struct.test_longdouble* noalias sret %[[AGG_RESULT:.*]], i32 signext %x, ...)
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[CUR]], i64 16
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: store i8* %[[NEXT]], i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[T0:%.*]] = bitcast i8* %[[CUR]] to %struct.test_longdouble*
Generalize NRVO to cover C structs. This commit generalizes NRVO to cover C structs (both trivial and non-trivial structs). rdar://problem/33599681 Differential Revision: https://reviews.llvm.org/D44968 llvm-svn: 328809
2018-03-30 01:56:24 +08:00
// CHECK: [[DEST:%.*]] = bitcast %struct.test_longdouble* %[[AGG_RESULT]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[SRC:%.*]] = bitcast %struct.test_longdouble* [[T0]] to i8*
Change memcpy/memove/memset to have dest and source alignment attributes. Summary: This change is step three in the series of changes to remove alignment argument from memcpy/memmove/memset in favour of alignment attributes. Steps: Step 1) Remove alignment parameter and create alignment parameter attributes for memcpy/memmove/memset. ( rL322965, rC322964, rL322963 ) Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing source and dest alignments. ( rL323597 ) Step 3) Update Clang to use the new IRBuilder API. Step 4) Update Polly to use the new IRBuilder API. Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API, and those that use use MemIntrinsicInst::[get|set]Alignment() to use getDestAlignment() and getSourceAlignment() instead. Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the MemIntrinsicInst::[get|set]Alignment() methods. Reference http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.html http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html Reviewers: rjmccall Subscribers: jyknight, nemanjai, nhaehnle, javed.absar, sbc100, aheejin, kbarton, fedor.sergeev, cfe-commits Differential Revision: https://reviews.llvm.org/D41677 llvm-svn: 323617
2018-01-29 01:27:45 +08:00
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[DEST]], i8* align 8 [[SRC]], i64 16, i1 false)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
struct test_longdouble { long double x; };
struct test_longdouble testva_longdouble (int x, ...)
{
struct test_longdouble y;
va_list ap;
va_start(ap, x);
y = va_arg (ap, struct test_longdouble);
va_end(ap);
return y;
}
[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition Summary: Clang -fpic defaults to -fno-semantic-interposition (GCC -fpic defaults to -fsemantic-interposition). Users need to specify -fsemantic-interposition to get semantic interposition behavior. Semantic interposition is currently a best-effort feature. There may still be some cases where it is not handled well. Reviewers: peter.smith, rnk, serge-sans-paille, sfertile, jfb, jdoerfert Subscribers: dschuff, jyknight, dylanmckay, nemanjai, jvesely, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, arphaman, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D73865
2020-02-03 04:23:47 +08:00
// CHECK: define dso_local void @testva_vector(%struct.test_vector* noalias sret %[[AGG_RESULT:.*]], i32 signext %x, ...)
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: %[[CUR:[^ ]+]] = load i8*, i8** %ap
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: %[[TMP0:[^ ]+]] = ptrtoint i8* %[[CUR]] to i64
// CHECK: %[[TMP1:[^ ]+]] = add i64 %[[TMP0]], 15
// CHECK: %[[TMP2:[^ ]+]] = and i64 %[[TMP1]], -16
// CHECK: %[[ALIGN:[^ ]+]] = inttoptr i64 %[[TMP2]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: %[[NEXT:[^ ]+]] = getelementptr inbounds i8, i8* %[[ALIGN]], i64 16
Fix (and reenable) ppc64-align-struct.c test for non-assert builds. llvm-svn: 212757
2014-07-11 03:19:03 +08:00
// CHECK: store i8* %[[NEXT]], i8** %ap
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[T0:%.*]] = bitcast i8* %[[ALIGN]] to %struct.test_vector*
Generalize NRVO to cover C structs. This commit generalizes NRVO to cover C structs (both trivial and non-trivial structs). rdar://problem/33599681 Differential Revision: https://reviews.llvm.org/D44968 llvm-svn: 328809
2018-03-30 01:56:24 +08:00
// CHECK: [[DEST:%.*]] = bitcast %struct.test_vector* %[[AGG_RESULT]] to i8*
Compute and preserve alignment more faithfully in IR-generation. Introduce an Address type to bundle a pointer value with an alignment. Introduce APIs on CGBuilderTy to work with Address values. Change core APIs on CGF/CGM to traffic in Address where appropriate. Require alignments to be non-zero. Update a ton of code to compute and propagate alignment information. As part of this, I've promoted CGBuiltin's EmitPointerWithAlignment helper function to CGF and made use of it in a number of places in the expression emitter. The end result is that we should now be significantly more correct when performing operations on objects that are locally known to be under-aligned. Since alignment is not reliably tracked in the type system, there are inherent limits to this, but at least we are no longer confused by standard operations like derived-to-base conversions and array-to-pointer decay. I've also fixed a large number of bugs where we were applying the complete-object alignment to a pointer instead of the non-virtual alignment, although most of these were hidden by the very conservative approach we took with member alignment. Also, because IRGen now reliably asserts on zero alignments, we should no longer be subject to an absurd but frustrating recurring bug where an incomplete type would report a zero alignment and then we'd naively do a alignmentAtOffset on it and emit code using an alignment equal to the largest power-of-two factor of the offset. We should also now be emitting much more aggressive alignment attributes in the presence of over-alignment. In particular, field access now uses alignmentAtOffset instead of min. Several times in this patch, I had to change the existing code-generation pattern in order to more effectively use the Address APIs. For the most part, this seems to be a strict improvement, like doing pointer arithmetic with GEPs instead of ptrtoint. That said, I've tried very hard to not change semantics, but it is likely that I've failed in a few places, for which I apologize. ABIArgInfo now always carries the assumed alignment of indirect and indirect byval arguments. In order to cut down on what was already a dauntingly large patch, I changed the code to never set align attributes in the IR on non-byval indirect arguments. That is, we still generate code which assumes that indirect arguments have the given alignment, but we don't express this information to the backend except where it's semantically required (i.e. on byvals). This is likely a minor regression for those targets that did provide this information, but it'll be trivial to add it back in a later patch. I partially punted on applying this work to CGBuiltin. Please do not add more uses of the CreateDefaultAligned{Load,Store} APIs; they will be going away eventually. llvm-svn: 246985
2015-09-08 16:05:57 +08:00
// CHECK: [[SRC:%.*]] = bitcast %struct.test_vector* [[T0]] to i8*
Change memcpy/memove/memset to have dest and source alignment attributes (Step 1). Summary: Upstream LLVM is changing the the prototypes of the @llvm.memcpy/memmove/memset intrinsics. This change updates the Clang tests for this change. The @llvm.memcpy/memmove/memset intrinsics currently have an explicit argument which is required to be a constant integer. It represents the alignment of the dest (and source), and so must be the minimum of the actual alignment of the two. This change removes the alignment argument in favour of placing the alignment attribute on the source and destination pointers of the memory intrinsic call. For example, code which used to read: call void @llvm.memcpy.p0i8.p0i8.i32(i8* %dest, i8* %src, i32 100, i32 4, i1 false) will now read call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %dest, i8* align 4 %src, i32 100, i1 false) At this time the source and destination alignments must be the same (Step 1). Step 2 of the change, to be landed shortly, will relax that contraint and allow the source and destination to have different alignments. llvm-svn: 322964
2018-01-20 01:12:54 +08:00
// CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 16 [[DEST]], i8* align 16 [[SRC]], i64 16, i1 false)
[PowerPC] ABI support for aligned by-value aggregates This patch adds support for respecting the ABI and type alignment of aggregates passed by value. Currently, all aggregates are aligned at 8 bytes in the parameter save area. This is incorrect for two reasons: - Aggregates that need alignment of 16 bytes or more should be aligned at 16 bytes in the parameter save area. This is implemented by using an appropriate "byval align" attribute in the IR. - Aggregates that need alignment beyond 16 bytes need to be dynamically realigned by the caller. This is implemented by setting the Realign flag of the ABIArgInfo::getIndirect call. In addition, when expanding a va_arg call accessing a type that is aligned at 16 bytes in the argument save area (either one of the aggregate types as above, or a vector type which is already aligned at 16 bytes), code needs to align the va_list pointer accordingly. Reviewed by Hal Finkel. llvm-svn: 212743
2014-07-11 01:20:07 +08:00
struct test_vector { vector int x; };
struct test_vector testva_vector (int x, ...)
{
struct test_vector y;
va_list ap;
va_start(ap, x);
y = va_arg (ap, struct test_vector);
va_end(ap);
return y;
}