maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/WebAssembly/userstack.ll

336 lines
12 KiB
LLVM
Raw Normal View History

[WebAssembly] Added default stack-only instruction mode for MC. Summary: Made it convert from register to stack based instructions, and removed the registers. Fixes to related code that was expecting register based instructions. Added the correct testing flag to all tests, depending on what the format they were expecting so far. Translated one test to stack format as example: reg-stackify-stack.ll tested: llvm-lit -v `find test -name WebAssembly` unittests/MC/* Reviewers: dschuff, sunfish Subscribers: sbc100, jgravelle-google, eraman, aheejin, llvm-commits, jfb Differential Revision: https://reviews.llvm.org/D51241 llvm-svn: 340750
2018-08-27 23:45:51 +08:00
; RUN: llc < %s -asm-verbose=false -disable-wasm-fallthrough-return-opt -wasm-keep-registers | FileCheck %s
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
[WebAssembly] Add -m:e to the target triple. This enables ELF-style name mangling, which primarily means using ".L" for private symbols. llvm-svn: 257020
2016-01-07 11:19:23 +08:00
target datalayout = "e-m:e-p:32:32-i64:64-n32:64-S128"
[WebAsembly] Update default triple in test files to wasm32-unknown-unkown. Summary: The final -wasm component has been the default for some time now. Subscribers: jfb, dschuff, jgravelle-google, eraman, aheejin, JDevlieghere, sunfish, llvm-commits Differential Revision: https://reviews.llvm.org/D46342 llvm-svn: 332007
2018-05-11 01:49:11 +08:00
target triple = "wasm32-unknown-unknown"
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
declare void @ext_func(i64* %ptr)
declare void @ext_func_i32(i32* %ptr)
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
; CHECK-LABEL: alloca32:
[WebAssembly] Print an extra local decl when the user stack pointer is used Differential Revision: http://reviews.llvm.org/D15546 llvm-svn: 255815
2015-12-17 04:43:06 +08:00
; Check that there is an extra local for the stack pointer.
[WebAssembly] Stackify function prologs and epilogs The instructions are the same, but fewer locals are used. Differential Revision: http://reviews.llvm.org/D17428 llvm-svn: 261452
2016-02-21 05:46:50 +08:00
; CHECK: .local i32{{$}}
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
define void @alloca32() noredzone {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: get_global $push[[L2:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Stackify function prologs and epilogs The instructions are the same, but fewer locals are used. Differential Revision: http://reviews.llvm.org/D17428 llvm-svn: 261452
2016-02-21 05:46:50 +08:00
; CHECK-NEXT: i32.const $push[[L3:.+]]=, 16
[WebAssembly] Disable the store-results optimization. The WebAssemly spec removing the return value from store instructions, so remove the associated optimization from LLVM. This patch leaves the store instruction operands in place for now, so stores now always write to "$drop"; these will be removed in a seperate patch. llvm-svn: 279100
2016-08-19 01:51:27 +08:00
; CHECK-NEXT: i32.sub $push[[L9:.+]]=, $pop[[L2]], $pop[[L3]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L8:.+]]=, [[SP:.+]], $pop[[L9]]{{$}}
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L8]]{{$}}
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
%retval = alloca i32
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L4:.+]]=, [[SP]]{{$}}
[WebAssembly] Stackify function prologs and epilogs The instructions are the same, but fewer locals are used. Differential Revision: http://reviews.llvm.org/D17428 llvm-svn: 261452
2016-02-21 05:46:50 +08:00
; CHECK: i32.const $push[[L0:.+]]=, 0
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: i32.store 12($pop[[L4]]), $pop[[L0]]
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
store i32 0, i32* %retval
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L6:.+]]=, [[SP]]{{$}}
[WebAssembly] Stackify code emitted by eliminateFrameIndex and SP writeback Summary: MRI::eliminateFrameIndex can emit several instructions to do address calculations; these can usually be stackified. Because instructions with FI operands can have subsequent operands which may be expression trees, find the top of the leftmost tree and insert the code before it, to keep the LIFO property. Also use stackified registers when writing back the SP value to memory in the epilog; it's unnecessary because SP will not be used after the epilog, and it results in better code. Differential Revision: http://reviews.llvm.org/D18234 llvm-svn: 263725
2016-03-18 01:00:29 +08:00
; CHECK-NEXT: i32.const $push[[L5:.+]]=, 16
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.add $push[[L7:.+]]=, $pop[[L6]], $pop[[L5]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L7]]
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
ret void
}
; CHECK-LABEL: alloca3264:
[WebAssembly] Stackify function prologs and epilogs The instructions are the same, but fewer locals are used. Differential Revision: http://reviews.llvm.org/D17428 llvm-svn: 261452
2016-02-21 05:46:50 +08:00
; CHECK: .local i32{{$}}
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
define void @alloca3264() {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L3:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: i32.const $push[[L4:.+]]=, 16
; CHECK-NEXT: i32.sub $push[[L6:.+]]=, $pop[[L3]], $pop[[L4]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L5:.+]]=, [[SP:.+]], $pop[[L6]]
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
%r1 = alloca i32
%r2 = alloca double
store i32 0, i32* %r1
Revert part of r308100 since the cause (r308025) was also reverted. The commit r308100 updated WebAssembly tests for r308025. In one case it merely made the test more resilient but in another case it made a substantive update. Because r308025 was reverted in r308271, these changes to the test also need to be reverted. They should be folded into the recommit of r308025 when it is ready. llvm-svn: 308273
2017-07-18 16:20:50 +08:00
store double 0.0, double* %r2
Revert r308273 to reinstate part of r308100. That part was reverted because the underlying change necessitating it (r308025) was reverted in r308271. Nirav re-landed r308025 again in r308350, so re-landing this fix. llvm-svn: 308418
2017-07-19 12:15:30 +08:00
; CHECK-NEXT: i64.const $push[[L1:.+]]=, 0
; CHECK-NEXT: i64.store 0($pop[[L5]]), $pop[[L1]]
; CHECK-NEXT: get_local $push[[L2:.+]]=, [[SP]]{{$}}
; CHECK-NEXT: i32.const $push[[L0:.+]]=, 0
; CHECK-NEXT: i32.store 12($pop[[L2]]), $pop[[L0]]
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
; CHECK-NEXT: return
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
ret void
}
[WebAssembly] Print an extra local decl when the user stack pointer is used Differential Revision: http://reviews.llvm.org/D15546 llvm-svn: 255815
2015-12-17 04:43:06 +08:00
; CHECK-LABEL: allocarray:
[WebAssembly] Stackify code emitted by eliminateFrameIndex and SP writeback Summary: MRI::eliminateFrameIndex can emit several instructions to do address calculations; these can usually be stackified. Because instructions with FI operands can have subsequent operands which may be expression trees, find the top of the leftmost tree and insert the code before it, to keep the LIFO property. Also use stackified registers when writing back the SP value to memory in the epilog; it's unnecessary because SP will not be used after the epilog, and it results in better code. Differential Revision: http://reviews.llvm.org/D18234 llvm-svn: 263725
2016-03-18 01:00:29 +08:00
; CHECK: .local i32{{$}}
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
define void @allocarray() {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: get_global $push[[L4:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Disable the store-results optimization. The WebAssemly spec removing the return value from store instructions, so remove the associated optimization from LLVM. This patch leaves the store instruction operands in place for now, so stores now always write to "$drop"; these will be removed in a seperate patch. llvm-svn: 279100
2016-08-19 01:51:27 +08:00
; CHECK-NEXT: i32.const $push[[L5:.+]]=, 144{{$}}
; CHECK-NEXT: i32.sub $push[[L12:.+]]=, $pop[[L4]], $pop[[L5]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L11:.+]]=, 0, $pop[[L12]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L11]]
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
%r = alloca [33 x i32]
[WebAssembly] Don't stackify stores across instructions with side effects. llvm-svn: 258285
2016-01-20 12:21:16 +08:00
[WebAssembly] Model the stack evaluation order more precisely. We currently don't represent get_local and set_local explicitly; they are just implied by virtual register use and def. This avoids a lot of clutter, but it does complicate stackifying: get_locals read their operands at their position in the stack evaluation order, rather than at their parent instruction. This patch adds code to walk the stack to determine the precise ordering, when needed. llvm-svn: 269854
2016-05-18 06:24:18 +08:00
; CHECK: i32.const $push{{.+}}=, 24
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.add $push[[L3:.+]]=, $pop{{.+}}, $pop{{.+}}
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: i32.const $push[[L1:.+]]=, 1{{$}}
[WebAssembly] Remove the output operand from stores. Per spec changes, store instructions in WebAssembly no longer have a return value. Update the instruction descriptions. Differential Revision: https://reviews.llvm.org/D25122 llvm-svn: 283501
2016-10-07 06:08:28 +08:00
; CHECK-NEXT: i32.store 0($pop[[L3]]), $pop[[L1]]{{$}}
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: get_local $push[[L4:.+]]=, 0{{$}}
[WebAssembly] Disable the store-results optimization. The WebAssemly spec removing the return value from store instructions, so remove the associated optimization from LLVM. This patch leaves the store instruction operands in place for now, so stores now always write to "$drop"; these will be removed in a seperate patch. llvm-svn: 279100
2016-08-19 01:51:27 +08:00
; CHECK-NEXT: i32.const $push[[L10:.+]]=, 1{{$}}
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store 12($pop[[L4]]), $pop[[L10]]{{$}}
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
%p = getelementptr [33 x i32], [33 x i32]* %r, i32 0, i32 0
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
store i32 1, i32* %p
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
%p2 = getelementptr [33 x i32], [33 x i32]* %r, i32 0, i32 3
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
store i32 1, i32* %p2
[WebAssembly] Don't stackify stores across instructions with side effects. llvm-svn: 258285
2016-01-20 12:21:16 +08:00
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: get_local $push[[L2:.+]]=, [[SP]]{{$}}
[WebAssembly] Disable the store-results optimization. The WebAssemly spec removing the return value from store instructions, so remove the associated optimization from LLVM. This patch leaves the store instruction operands in place for now, so stores now always write to "$drop"; these will be removed in a seperate patch. llvm-svn: 279100
2016-08-19 01:51:27 +08:00
; CHECK-NEXT: i32.const $push[[L7:.+]]=, 144
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.add $push[[L8:.+]]=, $pop[[L2]], $pop[[L7]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L8]]
[WebAssembly] Implement prolog/epilog insertion and FrameIndex elimination Summary: Use the SP32 physical register as the base for FrameIndex lowering. Update it and the __stack_pointer global var in the prolog and epilog. Extend the mapping of virtual registers to wasm locals to include the physical registers. Rather than modify the target-independent PrologEpilogInserter (which asserts that there are no virtual registers left) include a slightly-modified copy for Wasm that does not have this assertion and only clears the virtual registers if scavenging was needed (which of course it isn't for wasm). Differential Revision: http://reviews.llvm.org/D15344 llvm-svn: 255392
2015-12-12 07:49:46 +08:00
ret void
}
[WebAssembly] Omit no-op adds for non-mem uses of FrameIndex Differential Revision: http://reviews.llvm.org/D16554 llvm-svn: 258872
2016-01-27 06:47:43 +08:00
; CHECK-LABEL: non_mem_use
[WebAssembly] Fix uses of FrameIndex as store values Previously the code assumed all uses of FI on loads and stores were as addresses. This checks whether the use is the address or a value and handles the latter case as it does for non-memory instructions. llvm-svn: 259306
2016-01-31 05:43:08 +08:00
define void @non_mem_use(i8** %addr) {
[WebAssembly] Disable the store-results optimization. The WebAssemly spec removing the return value from store instructions, so remove the associated optimization from LLVM. This patch leaves the store instruction operands in place for now, so stores now always write to "$drop"; these will be removed in a seperate patch. llvm-svn: 279100
2016-08-19 01:51:27 +08:00
; CHECK: i32.const $push[[L2:.+]]=, 48
; CHECK-NEXT: i32.sub $push[[L12:.+]]=, {{.+}}, $pop[[L2]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L11:.+]]=, [[SP:.+]], $pop[[L12]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L11]]
[WebAssembly] Fix uses of FrameIndex as store values Previously the code assumed all uses of FI on loads and stores were as addresses. This checks whether the use is the address or a value and handles the latter case as it does for non-memory instructions. llvm-svn: 259306
2016-01-31 05:43:08 +08:00
%buf = alloca [27 x i8], align 16
[WebAssembly] Omit no-op adds for non-mem uses of FrameIndex Differential Revision: http://reviews.llvm.org/D16554 llvm-svn: 258872
2016-01-27 06:47:43 +08:00
%r = alloca i64
%r2 = alloca i64
; %r is at SP+8
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L3:.+]]=, [[SP]]
[WebAssembly] Stackify code emitted by eliminateFrameIndex and SP writeback Summary: MRI::eliminateFrameIndex can emit several instructions to do address calculations; these can usually be stackified. Because instructions with FI operands can have subsequent operands which may be expression trees, find the top of the leftmost tree and insert the code before it, to keep the LIFO property. Also use stackified registers when writing back the SP value to memory in the epilog; it's unnecessary because SP will not be used after the epilog, and it results in better code. Differential Revision: http://reviews.llvm.org/D18234 llvm-svn: 263725
2016-03-18 01:00:29 +08:00
; CHECK: i32.const $push[[OFF:.+]]=, 8
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.add $push[[ARG1:.+]]=, $pop[[L3]], $pop[[OFF]]
[WebAssembly] Stackify code emitted by eliminateFrameIndex and SP writeback Summary: MRI::eliminateFrameIndex can emit several instructions to do address calculations; these can usually be stackified. Because instructions with FI operands can have subsequent operands which may be expression trees, find the top of the leftmost tree and insert the code before it, to keep the LIFO property. Also use stackified registers when writing back the SP value to memory in the epilog; it's unnecessary because SP will not be used after the epilog, and it results in better code. Differential Revision: http://reviews.llvm.org/D18234 llvm-svn: 263725
2016-03-18 01:00:29 +08:00
; CHECK-NEXT: call ext_func@FUNCTION, $pop[[ARG1]]
[WebAssembly] Omit no-op adds for non-mem uses of FrameIndex Differential Revision: http://reviews.llvm.org/D16554 llvm-svn: 258872
2016-01-27 06:47:43 +08:00
call void @ext_func(i64* %r)
; %r2 is at SP+0, no add needed
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L4:.+]]=, [[SP]]
; CHECK-NEXT: call ext_func@FUNCTION, $pop[[L4]]
[WebAssembly] Omit no-op adds for non-mem uses of FrameIndex Differential Revision: http://reviews.llvm.org/D16554 llvm-svn: 258872
2016-01-27 06:47:43 +08:00
call void @ext_func(i64* %r2)
[WebAssembly] Fix uses of FrameIndex as store values Previously the code assumed all uses of FI on loads and stores were as addresses. This checks whether the use is the address or a value and handles the latter case as it does for non-memory instructions. llvm-svn: 259306
2016-01-31 05:43:08 +08:00
; Use as a value, but in a store
; %buf is at SP+16
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L5:.+]]=, [[SP]]
[WebAssembly] Stackify code emitted by eliminateFrameIndex and SP writeback Summary: MRI::eliminateFrameIndex can emit several instructions to do address calculations; these can usually be stackified. Because instructions with FI operands can have subsequent operands which may be expression trees, find the top of the leftmost tree and insert the code before it, to keep the LIFO property. Also use stackified registers when writing back the SP value to memory in the epilog; it's unnecessary because SP will not be used after the epilog, and it results in better code. Differential Revision: http://reviews.llvm.org/D18234 llvm-svn: 263725
2016-03-18 01:00:29 +08:00
; CHECK: i32.const $push[[OFF:.+]]=, 16
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.add $push[[VAL:.+]]=, $pop[[L5]], $pop[[OFF]]
; CHECK-NEXT: i32.store 0($pop{{.+}}), $pop[[VAL]]
[WebAssembly] Fix uses of FrameIndex as store values Previously the code assumed all uses of FI on loads and stores were as addresses. This checks whether the use is the address or a value and handles the latter case as it does for non-memory instructions. llvm-svn: 259306
2016-01-31 05:43:08 +08:00
%gep = getelementptr inbounds [27 x i8], [27 x i8]* %buf, i32 0, i32 0
store i8* %gep, i8** %addr
[WebAssembly] Omit no-op adds for non-mem uses of FrameIndex Differential Revision: http://reviews.llvm.org/D16554 llvm-svn: 258872
2016-01-27 06:47:43 +08:00
ret void
}
[WebAssembly] Don't stackify stores across instructions with side effects. llvm-svn: 258285
2016-01-20 12:21:16 +08:00
; CHECK-LABEL: allocarray_inbounds:
[WebAssembly] Stackify function prologs and epilogs The instructions are the same, but fewer locals are used. Differential Revision: http://reviews.llvm.org/D17428 llvm-svn: 261452
2016-02-21 05:46:50 +08:00
; CHECK: .local i32{{$}}
[WebAssembly] Support combining GEP and FrameIndex offsets in memory operand offset field Previously we only supported putting the FI into memory operand offset fields if there was nothing there already. Now combine them. Differential Revision: http://reviews.llvm.org/D15941 llvm-svn: 257084
2016-01-08 02:55:52 +08:00
define void @allocarray_inbounds() {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L3:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Disable the store-results optimization. The WebAssemly spec removing the return value from store instructions, so remove the associated optimization from LLVM. This patch leaves the store instruction operands in place for now, so stores now always write to "$drop"; these will be removed in a seperate patch. llvm-svn: 279100
2016-08-19 01:51:27 +08:00
; CHECK-NEXT: i32.const $push[[L4:.+]]=, 32{{$}}
; CHECK-NEXT: i32.sub $push[[L11:.+]]=, $pop[[L3]], $pop[[L4]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L10:.+]]=, [[SP:.+]], $pop[[L11]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L10]]{{$}}
[WebAssembly] Support combining GEP and FrameIndex offsets in memory operand offset field Previously we only supported putting the FI into memory operand offset fields if there was nothing there already. Now combine them. Differential Revision: http://reviews.llvm.org/D15941 llvm-svn: 257084
2016-01-08 02:55:52 +08:00
%r = alloca [5 x i32]
; CHECK: i32.const $push[[L3:.+]]=, 1
[WebAssembly] Remove the output operand from stores. Per spec changes, store instructions in WebAssembly no longer have a return value. Update the instruction descriptions. Differential Revision: https://reviews.llvm.org/D25122 llvm-svn: 283501
2016-10-07 06:08:28 +08:00
; CHECK-DAG: i32.store 24(${{.+}}), $pop[[L3]]
[WebAssembly] Support combining GEP and FrameIndex offsets in memory operand offset field Previously we only supported putting the FI into memory operand offset fields if there was nothing there already. Now combine them. Differential Revision: http://reviews.llvm.org/D15941 llvm-svn: 257084
2016-01-08 02:55:52 +08:00
%p = getelementptr inbounds [5 x i32], [5 x i32]* %r, i32 0, i32 0
store i32 1, i32* %p
; This store should have both the GEP and the FI folded into it.
[WebAssembly] Remove the output operand from stores. Per spec changes, store instructions in WebAssembly no longer have a return value. Update the instruction descriptions. Differential Revision: https://reviews.llvm.org/D25122 llvm-svn: 283501
2016-10-07 06:08:28 +08:00
; CHECK-DAG: i32.store 12(${{.+}}), $pop
[WebAssembly] Implement unaligned loads and stores. Differential Revision: http://reviews.llvm.org/D16534 llvm-svn: 258779
2016-01-26 11:39:31 +08:00
%p2 = getelementptr inbounds [5 x i32], [5 x i32]* %r, i32 0, i32 3
[WebAssembly] Support combining GEP and FrameIndex offsets in memory operand offset field Previously we only supported putting the FI into memory operand offset fields if there was nothing there already. Now combine them. Differential Revision: http://reviews.llvm.org/D15941 llvm-svn: 257084
2016-01-08 02:55:52 +08:00
store i32 1, i32* %p2
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
call void @ext_func(i64* null);
[WebAssembly] Put __stack_pointer in the offset field of loads and stores. Instead of this: i32.const $push10=, __stack_pointer i32.load $push11=, 0($pop10) Emit this: i32.const $push10=, 0 i32.load $push11=, __stack_pointer($pop10) It's not currently clear which is better, though there's a chance the second form may be better at overall compression. We can revisit this when we have more data; for now it makes sense to make PEI consistent with isel. Differential Revision: http://reviews.llvm.org/D20411 llvm-svn: 270635
2016-05-25 07:47:41 +08:00
; CHECK: call ext_func
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: i32.const $push[[L5:.+]]=, 32{{$}}
[WebAssembly] Model the stack evaluation order more precisely. We currently don't represent get_local and set_local explicitly; they are just implied by virtual register use and def. This avoids a lot of clutter, but it does complicate stackifying: get_locals read their operands at their position in the stack evaluation order, rather than at their parent instruction. This patch adds code to walk the stack to determine the precise ordering, when needed. llvm-svn: 269854
2016-05-18 06:24:18 +08:00
; CHECK-NEXT: i32.add $push[[L7:.+]]=, ${{.+}}, $pop[[L5]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L7]]
[WebAssembly] Support combining GEP and FrameIndex offsets in memory operand offset field Previously we only supported putting the FI into memory operand offset fields if there was nothing there already. Now combine them. Differential Revision: http://reviews.llvm.org/D15941 llvm-svn: 257084
2016-01-08 02:55:52 +08:00
ret void
}
[WebAssembly] Don't stackify stores across instructions with side effects. llvm-svn: 258285
2016-01-20 12:21:16 +08:00
; CHECK-LABEL: dynamic_alloca:
[WebAssembly] Implement eliminateCallFramePseudo Summary: Implement eliminateCallFramePsuedo to handle ADJCALLSTACKUP/DOWN pseudo-instructions. Add a test calling a vararg function which causes non-0 adjustments. This revealed an issue with RegisterCoalescer wherein it eliminates a COPY from SP32 to a vreg but failes to update the live ranges of EXPR_STACK, causing a machineinstr verifier failure (so this test is commented out). Also add a dynamic alloca test, which causes a callseq_end dag node with a 0 (instead of undef) second argument to be generated. We currently fail to select that, so adjust the ADJCALLSTACKUP tablegen code to handle it. Differential Revision: http://reviews.llvm.org/D15587 llvm-svn: 255844
2015-12-17 07:21:30 +08:00
define void @dynamic_alloca(i32 %alloc) {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L13:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: tee_local $push[[L12:.+]]=, [[SP:.+]], $pop[[L13]]{{$}}
[WebAssembly] Fix writeback of stack pointer with dynamic alloca Previously the stack pointer was only written back to memory in the prolog. But this is wrong for dynamic allocas, for which target-independent codegen handles SP updates after the prolog (and possibly even in another BB). Instead update the SP global in ADJCALLSTACKDOWN which is generated after the SP update sequence. This will have further refinements when we add red zone support. llvm-svn: 261579
2016-02-23 05:57:17 +08:00
; Target independent codegen bumps the stack pointer.
; CHECK: i32.sub
; Check that SP is written back to memory after decrement
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: set_global __stack_pointer@GLOBAL,
[WebAssembly] Support frame pointer Add support for frame pointer use in prolog/epilog. Supports dynamic allocas but not yet over-aligned locals. Target-independend CG generates SP updates, but we still need to write back the SP value to memory when necessary. llvm-svn: 259220
2016-01-30 02:37:49 +08:00
%r = alloca i32, i32 %alloc
; Target-independent codegen also calculates the store addr
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
; CHECK: call ext_func_i32@FUNCTION
call void @ext_func_i32(i32* %r)
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: set_global __stack_pointer@GLOBAL, $pop{{.+}}
[WebAssembly] Support frame pointer Add support for frame pointer use in prolog/epilog. Supports dynamic allocas but not yet over-aligned locals. Target-independend CG generates SP updates, but we still need to write back the SP value to memory when necessary. llvm-svn: 259220
2016-01-30 02:37:49 +08:00
ret void
}
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
; CHECK-LABEL: dynamic_alloca_redzone:
define void @dynamic_alloca_redzone(i32 %alloc) {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L13:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: tee_local $push[[L12:.+]]=, [[SP:.+]], $pop[[L13]]{{$}}
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
; Target independent codegen bumps the stack pointer
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK: i32.sub
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
%r = alloca i32, i32 %alloc
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L8:.+]]=, {{.+}}, $pop
; CHECK: get_local $push[[L7:.+]]=, 0{{$}}
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: i32.const $push[[L6:.+]]=, 0{{$}}
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store 0($pop[[L7]]), $pop[[L6]]{{$}}
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
store i32 0, i32* %r
; CHECK-NEXT: return
ret void
}
[WebAssembly] Support frame pointer Add support for frame pointer use in prolog/epilog. Supports dynamic allocas but not yet over-aligned locals. Target-independend CG generates SP updates, but we still need to write back the SP value to memory when necessary. llvm-svn: 259220
2016-01-30 02:37:49 +08:00
; CHECK-LABEL: dynamic_static_alloca:
[WebAssembly] Implement red zone for user stack Implements a mostly-conventional redzone for the userspace stack. Because we have unsigned load/store offsets we continue to use a local SP subtracted from the incoming SP but do not write it back to memory. Differential Revision: http://reviews.llvm.org/D17525 llvm-svn: 261662
2016-02-24 02:13:07 +08:00
define void @dynamic_static_alloca(i32 %alloc) noredzone {
[WebAssembly] Fix writeback of stack pointer with dynamic alloca Previously the stack pointer was only written back to memory in the prolog. But this is wrong for dynamic allocas, for which target-independent codegen handles SP updates after the prolog (and possibly even in another BB). Instead update the SP global in ADJCALLSTACKDOWN which is generated after the SP update sequence. This will have further refinements when we add red zone support. llvm-svn: 261579
2016-02-23 05:57:17 +08:00
; Decrement SP in the prolog by the static amount and writeback to memory.
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L11:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK-NEXT: i32.const $push[[L12:.+]]=, 16
; CHECK-NEXT: i32.sub $push[[L23:.+]]=, $pop[[L11]], $pop[[L12]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L22:.+]]=, [[SP:.+]], $pop[[L23]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L22]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; Alloc and write to a static alloca
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L21:.+]]=, [[SP:.+]]
; CHECK-NEXT: tee_local $push[[pushedFP:.+]]=, [[FP:.+]], $pop[[L21]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK-NEXT: i32.const $push[[L0:.+]]=, 101
[WebAssembly] Remove the output operand from stores. Per spec changes, store instructions in WebAssembly no longer have a return value. Update the instruction descriptions. Differential Revision: https://reviews.llvm.org/D25122 llvm-svn: 283501
2016-10-07 06:08:28 +08:00
; CHECK-NEXT: i32.store [[static_offset:.+]]($pop[[pushedFP]]), $pop[[L0]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
%static = alloca i32
store volatile i32 101, i32* %static
; Decrement SP in the body by the dynamic amount.
; CHECK: i32.sub
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: tee_local $push[[L16:.+]]=, [[dynamic_local:.+]], $pop{{.+}}
; CHECK: tee_local $push[[L15:.+]]=, [[other:.+]], $pop[[L16]]{{$}}
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: set_global __stack_pointer@GLOBAL, $pop[[L15]]{{$}}
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
%dynamic = alloca i32, i32 %alloc
; Ensure we don't modify the frame pointer after assigning it.
; CHECK-NOT: $[[FP]]=
; Ensure the static address doesn't change after modifying the stack pointer.
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L17:.+]]=, [[FP]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK: i32.const $push[[L7:.+]]=, 102
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store [[static_offset]]($pop[[L17]]), $pop[[L7]]
; CHECK-NEXT: get_local $push[[L9:.+]]=, [[dynamic_local]]{{$}}
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK-NEXT: i32.const $push[[L8:.+]]=, 103
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store 0($pop[[L9]]), $pop[[L8]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
store volatile i32 102, i32* %static
store volatile i32 103, i32* %dynamic
[WebAssembly] Fix writeback of stack pointer with dynamic alloca Previously the stack pointer was only written back to memory in the prolog. But this is wrong for dynamic allocas, for which target-independent codegen handles SP updates after the prolog (and possibly even in another BB). Instead update the SP global in ADJCALLSTACKDOWN which is generated after the SP update sequence. This will have further refinements when we add red zone support. llvm-svn: 261579
2016-02-23 05:57:17 +08:00
; Decrement SP in the body by the dynamic amount.
; CHECK: i32.sub
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: tee_local $push{{.+}}=, [[dynamic2_local:.+]], $pop{{.+}}
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
%dynamic.2 = alloca i32, i32 %alloc
; CHECK-NOT: $[[FP]]=
; Ensure neither the static nor dynamic address changes after the second
; modification of the stack pointer.
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L22:.+]]=, [[FP]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK: i32.const $push[[L9:.+]]=, 104
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store [[static_offset]]($pop[[L22]]), $pop[[L9]]
; CHECK-NEXT: get_local $push[[L23:.+]]=, [[dynamic_local]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK-NEXT: i32.const $push[[L10:.+]]=, 105
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store 0($pop[[L23]]), $pop[[L10]]
; CHECK-NEXT: get_local $push[[L23:.+]]=, [[dynamic2_local]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
; CHECK-NEXT: i32.const $push[[L11:.+]]=, 106
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: i32.store 0($pop[[L23]]), $pop[[L11]]
[WebAssembly] Use the frame pointer instead of the stack pointer When we have dynamic allocas we have a frame pointer, and when we're lowering frame indexes we should make sure we use it. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24889 llvm-svn: 282442
2016-09-27 05:18:03 +08:00
store volatile i32 104, i32* %static
store volatile i32 105, i32* %dynamic
store volatile i32 106, i32* %dynamic.2
[WebAssembly] Fix writeback of stack pointer with dynamic alloca Previously the stack pointer was only written back to memory in the prolog. But this is wrong for dynamic allocas, for which target-independent codegen handles SP updates after the prolog (and possibly even in another BB). Instead update the SP global in ADJCALLSTACKDOWN which is generated after the SP update sequence. This will have further refinements when we add red zone support. llvm-svn: 261579
2016-02-23 05:57:17 +08:00
; Writeback to memory.
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L24:.+]]=, [[FP]]{{$}}
; CHECK: i32.const $push[[L18:.+]]=, 16
; CHECK-NEXT: i32.add $push[[L19:.+]]=, $pop[[L24]], $pop[[L18]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L19]]
[WebAssembly] Implement eliminateCallFramePseudo Summary: Implement eliminateCallFramePsuedo to handle ADJCALLSTACKUP/DOWN pseudo-instructions. Add a test calling a vararg function which causes non-0 adjustments. This revealed an issue with RegisterCoalescer wherein it eliminates a COPY from SP32 to a vreg but failes to update the live ranges of EXPR_STACK, causing a machineinstr verifier failure (so this test is commented out). Also add a dynamic alloca test, which causes a callseq_end dag node with a 0 (instead of undef) second argument to be generated. We currently fail to select that, so adjust the ADJCALLSTACKUP tablegen code to handle it. Differential Revision: http://reviews.llvm.org/D15587 llvm-svn: 255844
2015-12-17 07:21:30 +08:00
ret void
}
[WebAssembly] Fix uses of FrameIndex as store values Previously the code assumed all uses of FI on loads and stores were as addresses. This checks whether the use is the address or a value and handles the latter case as it does for non-memory instructions. llvm-svn: 259306
2016-01-31 05:43:08 +08:00
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
declare i8* @llvm.stacksave()
declare void @llvm.stackrestore(i8*)
; CHECK-LABEL: llvm_stack_builtins:
define void @llvm_stack_builtins(i32 %alloc) noredzone {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L11:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L10:.+]]=, {{.+}}, $pop[[L11]]
; CHECK-NEXT: set_local [[STACK:.+]], $pop[[L10]]
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
%stack = call i8* @llvm.stacksave()
; Ensure we don't reassign the stacksave local
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NOT: set_local [[STACK]],
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
%dynamic = alloca i32, i32 %alloc
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L12:.+]]=, [[STACK]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L12]]
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
call void @llvm.stackrestore(i8* %stack)
ret void
}
; Not actually using the alloca'd variables exposed an issue with register
; stackification, where copying the stack pointer into the frame pointer was
; moved after the stack pointer was updated for the dynamic alloca.
; CHECK-LABEL: dynamic_alloca_nouse:
define void @dynamic_alloca_nouse(i32 %alloc) noredzone {
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L11:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L10:.+]]=, {{.+}}, $pop[[L11]]
; CHECK-NEXT: set_local [[FP:.+]], $pop[[L10]]
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
%dynamic = alloca i32, i32 %alloc
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NOT: set_local [[FP]],
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L12:.+]]=, [[FP]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L12]]
[WebAssembly] Make register stackification more conservative Register stackification currently checks VNInfo for changes. Make that more accurate by testing each intervening instruction for any other defs to the same virtual register. Patch by Jacob Gravelle Differential Revision: https://reviews.llvm.org/D24942 llvm-svn: 282886
2016-10-01 02:02:54 +08:00
ret void
}
[WebAssembly] Insert COPY_LOCAL between CopyToReg and FrameIndex DAG nodes CopyToReg nodes don't support FrameIndex operands. Other targets select the FI to some LEA-like instruction, but since we don't have that, we need to insert some kind of instruction that can take an FI operand and produces a value usable by CopyToReg (i.e. in a vreg). So insert a dummy copy_local between Op and its FI operand. This results in a redundant copy which we should optimize away later (maybe in the post-FI-lowering peephole pass). Differential Revision: http://reviews.llvm.org/D17213 llvm-svn: 260987
2016-02-17 02:18:36 +08:00
; The use of the alloca in a phi causes a CopyToReg DAG node to be generated,
; which has to have special handling because CopyToReg can't have a FI operand
; CHECK-LABEL: copytoreg_fi:
define void @copytoreg_fi(i1 %cond, i32* %b) {
entry:
[WebAssembly] Stackify function prologs and epilogs The instructions are the same, but fewer locals are used. Differential Revision: http://reviews.llvm.org/D17428 llvm-svn: 261452
2016-02-21 05:46:50 +08:00
; CHECK: i32.const $push[[L1:.+]]=, 16
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: i32.sub $push[[L3:.+]]=, {{.+}}, $pop[[L1]]
[WebAssembly] Insert COPY_LOCAL between CopyToReg and FrameIndex DAG nodes CopyToReg nodes don't support FrameIndex operands. Other targets select the FI to some LEA-like instruction, but since we don't have that, we need to insert some kind of instruction that can take an FI operand and produces a value usable by CopyToReg (i.e. in a vreg). So insert a dummy copy_local between Op and its FI operand. This results in a redundant copy which we should optimize away later (maybe in the post-FI-lowering peephole pass). Differential Revision: http://reviews.llvm.org/D17213 llvm-svn: 260987
2016-02-17 02:18:36 +08:00
%addr = alloca i32
[WebAssembly] Stackify code emitted by eliminateFrameIndex and SP writeback Summary: MRI::eliminateFrameIndex can emit several instructions to do address calculations; these can usually be stackified. Because instructions with FI operands can have subsequent operands which may be expression trees, find the top of the leftmost tree and insert the code before it, to keep the LIFO property. Also use stackified registers when writing back the SP value to memory in the epilog; it's unnecessary because SP will not be used after the epilog, and it results in better code. Differential Revision: http://reviews.llvm.org/D18234 llvm-svn: 263725
2016-03-18 01:00:29 +08:00
; CHECK: i32.const $push[[OFF:.+]]=, 12
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: i32.add $push[[ADDR:.+]]=, $pop[[L3]], $pop[[OFF]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: set_local [[COPY:.+]], $pop[[ADDR]]
[WebAssembly] Insert COPY_LOCAL between CopyToReg and FrameIndex DAG nodes CopyToReg nodes don't support FrameIndex operands. Other targets select the FI to some LEA-like instruction, but since we don't have that, we need to insert some kind of instruction that can take an FI operand and produces a value usable by CopyToReg (i.e. in a vreg). So insert a dummy copy_local between Op and its FI operand. This results in a redundant copy which we should optimize away later (maybe in the post-FI-lowering peephole pass). Differential Revision: http://reviews.llvm.org/D17213 llvm-svn: 260987
2016-02-17 02:18:36 +08:00
br label %body
body:
%a = phi i32* [%addr, %entry], [%b, %body]
store i32 1, i32* %a
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK: get_local $push[[L12:.+]]=, [[COPY]]
; CHECK: i32.store 0($pop[[L12]]),
[WebAssembly] Insert COPY_LOCAL between CopyToReg and FrameIndex DAG nodes CopyToReg nodes don't support FrameIndex operands. Other targets select the FI to some LEA-like instruction, but since we don't have that, we need to insert some kind of instruction that can take an FI operand and produces a value usable by CopyToReg (i.e. in a vreg). So insert a dummy copy_local between Op and its FI operand. This results in a redundant copy which we should optimize away later (maybe in the post-FI-lowering peephole pass). Differential Revision: http://reviews.llvm.org/D17213 llvm-svn: 260987
2016-02-17 02:18:36 +08:00
br i1 %cond, label %body, label %exit
exit:
ret void
}
[WebAssembly] Implement __builtin_frame_address. Differential Revision: http://reviews.llvm.org/D17307 llvm-svn: 261032
2016-02-17 07:48:04 +08:00
declare void @use_i8_star(i8*)
declare i8* @llvm.frameaddress(i32)
; Test __builtin_frame_address(0).
; CHECK-LABEL: frameaddress_0:
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global $push[[L3:.+]]=, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: tee_local $push[[L2:.+]]=, [[FP:.+]], $pop[[L3]]{{$}}
[WebAssembly] Move register stackification and coloring to a late phase. Move the register stackification and coloring passes to run very late, after PEI, tail duplication, and most other passes. This means that all code emitted and expanded by those passes is now exposed to these passes. This also eliminates the need for prologue/epilogue code to be manually stackified, which significantly simplifies the code. This does require running LiveIntervals a second time. It's useful to think of these late passes not as late optimization passes, but as a domain-specific compression algorithm based on knowledge of liveness information. It's used to compress the code after all conventional optimizations are complete, which is why it uses LiveIntervals at a phase when actual optimization passes don't typically need it. Differential Revision: http://reviews.llvm.org/D20075 llvm-svn: 269012
2016-05-10 12:24:02 +08:00
; CHECK-NEXT: call use_i8_star@FUNCTION, $pop[[L2]]
[WebAssembly] Convert the remaining unit tests to the new wasm-object-file target. To facilitate this, add a new hidden command-line option to disable the explicit-locals pass. That causes llc to emit invalid code that doesn't have all locals converted to get_local/set_local, however it simplifies testwriting in many cases. llvm-svn: 296540
2017-03-01 07:37:04 +08:00
; CHECK-NEXT: get_local $push[[L5:.+]]=, [[FP]]
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK-NEXT: set_global __stack_pointer@GLOBAL, $pop[[L5]]
[WebAssembly] Implement __builtin_frame_address. Differential Revision: http://reviews.llvm.org/D17307 llvm-svn: 261032
2016-02-17 07:48:04 +08:00
define void @frameaddress_0() {
%t = call i8* @llvm.frameaddress(i32 0)
call void @use_i8_star(i8* %t)
ret void
}
; Test __builtin_frame_address(1).
; CHECK-LABEL: frameaddress_1:
[WebAssembly] replaced .param/.result by .functype Summary: This makes it easier/cleaner to generate a single signature from this directive. Also: - Adds the symbol name, such that we don't depend on the location of this directive anymore. - Actually constructs the signature in the assembler, and make the assembler own it. - Refactor the use of MVT vs ValType in the streamer and assembler to require less conversions overall. - Changed 700 or so tests to use it. Reviewers: sbc100, dschuff Subscribers: jgravelle-google, eraman, aheejin, sunfish, jfb, llvm-commits Differential Revision: https://reviews.llvm.org/D54652 llvm-svn: 347228
2018-11-20 01:10:36 +08:00
; CHECK: i32.const $push0=, 0{{$}}
[WebAssembly] Implement __builtin_frame_address. Differential Revision: http://reviews.llvm.org/D17307 llvm-svn: 261032
2016-02-17 07:48:04 +08:00
; CHECK-NEXT: call use_i8_star@FUNCTION, $pop0{{$}}
; CHECK-NEXT: return{{$}}
define void @frameaddress_1() {
%t = call i8* @llvm.frameaddress(i32 1)
call void @use_i8_star(i8* %t)
ret void
}
[WebAssembly] Handle CopyToReg nodes with flag results in LowerCopyToReg. llvm-svn: 261457
2016-02-21 07:09:44 +08:00
; Test a stack address passed to an inline asm.
; CHECK-LABEL: inline_asm:
[WebAssembly] Cleanup of the way globals and global flags are handled Differential Revision: https://reviews.llvm.org/D44030 llvm-svn: 338894
2018-08-03 22:33:37 +08:00
; CHECK: get_global {{.+}}, __stack_pointer@GLOBAL{{$}}
[WebAssembly] Handle CopyToReg nodes with flag results in LowerCopyToReg. llvm-svn: 261457
2016-02-21 07:09:44 +08:00
; CHECK: #APP
; CHECK-NEXT: # %{{[0-9]+}}{{$}}
; CHECK-NEXT: #NO_APP
define void @inline_asm() {
%tmp = alloca i8
call void asm sideeffect "# %0", "r"(i8* %tmp)
ret void
}
[WebAssembly] Added a .globaltype directive to .s output. Summary: Assembly output can use globals like __stack_pointer implicitly, but has no way of indicating the type of such a global, which makes it hard for tools processing it (such as the MC Assembler) to reconstruct this information. The improved assembler directives parsing (in progress in https://reviews.llvm.org/D53842) will make use of this information. Also deleted code for the .import_global directive which was unused. New test case in userstack.ll Reviewers: dschuff, sbc100 Subscribers: jgravelle-google, aheejin, sunfish, llvm-commits Differential Revision: https://reviews.llvm.org/D54012 llvm-svn: 345917
2018-11-02 08:45:00 +08:00
; CHECK: .globaltype __stack_pointer, i32{{$}}
[WebAssembly] Fix uses of FrameIndex as store values Previously the code assumed all uses of FI on loads and stores were as addresses. This checks whether the use is the address or a value and handles the latter case as it does for non-memory instructions. llvm-svn: 259306
2016-01-31 05:43:08 +08:00
; TODO: test over-aligned alloca