RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-23 21:13:02 +02:00
Code Issues Projects Releases Wiki Activity
f9a0655273
llvm-mirror/test/CodeGen/X86/cleanuppad-realign.ll

79 lines
2.6 KiB
LLVM
Raw Normal View History

Implemented stack symbol table ordering/packing optimization to improve data locality and code size from SP/FP offset encoding. Differential Revision: http://reviews.llvm.org/D15393 llvm-svn: 260917
2016-02-16 00:44:13 +01:00
; RUN: llc -mtriple=i686-pc-windows-msvc -stack-symbol-ordering=0 < %s | FileCheck --check-prefix=X86 %s
; RUN: llc -mtriple=x86_64-pc-windows-msvc -stack-symbol-ordering=0 < %s | FileCheck --check-prefix=X64 %s
[WinEH] Fix funclet prologues with stack realignment We already had a test for this for 32-bit SEH catchpads, but those don't actually create funclets. We had a bug that only appeared in funclet prologues, where we would establish EBP and ESI as our FP and BP, and then downstream prologue code would overwrite them. While I was at it, I fixed Win64+funclets+stackrealign. This issue doesn't come up as often there due to the ABI requring 16 byte stack alignment, but now we can rest easy that AVX and WinEH will work well together =P. llvm-svn: 252210
2015-11-05 22:09:49 +01:00
declare i32 @__CxxFrameHandler3(...)
declare void @Dtor(i64* %o)
declare void @f(i32)
define void @realigned_cleanup() personality i32 (...)* @__CxxFrameHandler3 {
entry:
; Overalign %o to cause stack realignment.
%o = alloca i64, align 32
invoke void @f(i32 1)
to label %invoke.cont unwind label %ehcleanup
invoke.cont: ; preds = %entry
call void @Dtor(i64* %o)
ret void
ehcleanup: ; preds = %entry
[IR] Reformulate LLVM's EH funclet IR While we have successfully implemented a funclet-oriented EH scheme on top of LLVM IR, our scheme has some notable deficiencies: - catchendpad and cleanupendpad are necessary in the current design but they are difficult to explain to others, even to seasoned LLVM experts. - catchendpad and cleanupendpad are optimization barriers. They cannot be split and force all potentially throwing call-sites to be invokes. This has a noticable effect on the quality of our code generation. - catchpad, while similar in some aspects to invoke, is fairly awkward. It is unsplittable, starts a funclet, and has control flow to other funclets. - The nesting relationship between funclets is currently a property of control flow edges. Because of this, we are forced to carefully analyze the flow graph to see if there might potentially exist illegal nesting among funclets. While we have logic to clone funclets when they are illegally nested, it would be nicer if we had a representation which forbade them upfront. Let's clean this up a bit by doing the following: - Instead, make catchpad more like cleanuppad and landingpad: no control flow, just a bunch of simple operands; catchpad would be splittable. - Introduce catchswitch, a control flow instruction designed to model the constraints of funclet oriented EH. - Make funclet scoping explicit by having funclet instructions consume the token produced by the funclet which contains them. - Remove catchendpad and cleanupendpad. Their presence can be inferred implicitly using coloring information. N.B. The state numbering code for the CLR has been updated but the veracity of it's output cannot be spoken for. An expert should take a look to make sure the results are reasonable. Reviewers: rnk, JosephTremoulet, andrew.w.kaylor Differential Revision: http://reviews.llvm.org/D15139 llvm-svn: 255422
2015-12-12 06:38:55 +01:00
%0 = cleanuppad within none []
[WinEH] Use operand bundles to describe call sites SimplifyCFG allows tail merging with code which terminates in unreachable which, in turn, makes it possible for an invoke to end up in a funclet which it was not originally part of. Using operand bundles on invokes allows us to determine whether or not an invoke was part of a funclet in the source program. Furthermore, it allows us to unambiguously answer questions about the legality of inlining into call sites which the personality may have trouble with. Differential Revision: http://reviews.llvm.org/D15517 llvm-svn: 255674
2015-12-15 22:27:27 +01:00
call void @Dtor(i64* %o) [ "funclet"(token %0) ]
[IR] Reformulate LLVM's EH funclet IR While we have successfully implemented a funclet-oriented EH scheme on top of LLVM IR, our scheme has some notable deficiencies: - catchendpad and cleanupendpad are necessary in the current design but they are difficult to explain to others, even to seasoned LLVM experts. - catchendpad and cleanupendpad are optimization barriers. They cannot be split and force all potentially throwing call-sites to be invokes. This has a noticable effect on the quality of our code generation. - catchpad, while similar in some aspects to invoke, is fairly awkward. It is unsplittable, starts a funclet, and has control flow to other funclets. - The nesting relationship between funclets is currently a property of control flow edges. Because of this, we are forced to carefully analyze the flow graph to see if there might potentially exist illegal nesting among funclets. While we have logic to clone funclets when they are illegally nested, it would be nicer if we had a representation which forbade them upfront. Let's clean this up a bit by doing the following: - Instead, make catchpad more like cleanuppad and landingpad: no control flow, just a bunch of simple operands; catchpad would be splittable. - Introduce catchswitch, a control flow instruction designed to model the constraints of funclet oriented EH. - Make funclet scoping explicit by having funclet instructions consume the token produced by the funclet which contains them. - Remove catchendpad and cleanupendpad. Their presence can be inferred implicitly using coloring information. N.B. The state numbering code for the CLR has been updated but the veracity of it's output cannot be spoken for. An expert should take a look to make sure the results are reasonable. Reviewers: rnk, JosephTremoulet, andrew.w.kaylor Differential Revision: http://reviews.llvm.org/D15139 llvm-svn: 255422
2015-12-12 06:38:55 +01:00
cleanupret from %0 unwind to caller
[WinEH] Fix funclet prologues with stack realignment We already had a test for this for 32-bit SEH catchpads, but those don't actually create funclets. We had a bug that only appeared in funclet prologues, where we would establish EBP and ESI as our FP and BP, and then downstream prologue code would overwrite them. While I was at it, I fixed Win64+funclets+stackrealign. This issue doesn't come up as often there due to the ABI requring 16 byte stack alignment, but now we can rest easy that AVX and WinEH will work well together =P. llvm-svn: 252210
2015-11-05 22:09:49 +01:00
}
; X86-LABEL: _realigned_cleanup: # @realigned_cleanup
; X86: pushl %ebp
; X86: movl %esp, %ebp
; X86: pushl %ebx
; X86: pushl %edi
; X86: pushl %esi
; X86: andl $-32, %esp
; X86: subl $96, %esp
; X86: movl %esp, %esi
; EBP will reload from this offset.
; X86: movl %ebp, 28(%esi)
; The last EH reg field is the state number, so dtor adjust is this +4.
; X86: movl $-1, 72(%esi)
; X86-LABEL: "?dtor$2@?0?realigned_cleanup@4HA":
; X86: pushl %ebp
; X86: leal -76(%ebp), %esi
; X86: movl 28(%esi), %ebp
; We used to have a bug where we clobbered ESI after the prologue.
; X86-NOT: movl {{.*}}, %esi
; X86: popl %ebp
; X86: retl # CLEANUPRET
; X64-LABEL: realigned_cleanup: # @realigned_cleanup
; X64: pushq %rbp
; X64: .seh_pushreg 5
; X64: pushq %rbx
; X64: .seh_pushreg 3
[WinEH] Make UnwindHelp a fixed stack object allocated after XMM CSRs Now the offset of UnwindHelp in our EH tables and the offset that we store to in the prologue agree. llvm-svn: 253059
2015-11-13 20:06:01 +01:00
; X64: subq $104, %rsp
; X64: .seh_stackalloc 104
; X64: leaq 96(%rsp), %rbp
; X64: .seh_setframe 5, 96
[WinEH] Fix funclet prologues with stack realignment We already had a test for this for 32-bit SEH catchpads, but those don't actually create funclets. We had a bug that only appeared in funclet prologues, where we would establish EBP and ESI as our FP and BP, and then downstream prologue code would overwrite them. While I was at it, I fixed Win64+funclets+stackrealign. This issue doesn't come up as often there due to the ABI requring 16 byte stack alignment, but now we can rest easy that AVX and WinEH will work well together =P. llvm-svn: 252210
2015-11-05 22:09:49 +01:00
; X64: .seh_endprologue
; X64: andq $-32, %rsp
; X64: movq %rsp, %rbx
; RBP will reload from this offset.
[WinEH] Make UnwindHelp a fixed stack object allocated after XMM CSRs Now the offset of UnwindHelp in our EH tables and the offset that we store to in the prologue agree. llvm-svn: 253059
2015-11-13 20:06:01 +01:00
; X64: movq %rbp, 56(%rbx)
; X64: movq $-2, (%rbp)
[WinEH] Fix funclet prologues with stack realignment We already had a test for this for 32-bit SEH catchpads, but those don't actually create funclets. We had a bug that only appeared in funclet prologues, where we would establish EBP and ESI as our FP and BP, and then downstream prologue code would overwrite them. While I was at it, I fixed Win64+funclets+stackrealign. This issue doesn't come up as often there due to the ABI requring 16 byte stack alignment, but now we can rest easy that AVX and WinEH will work well together =P. llvm-svn: 252210
2015-11-05 22:09:49 +01:00
; X64-LABEL: "?dtor$2@?0?realigned_cleanup@4HA":
; X64: movq %rdx, 16(%rsp)
; X64: pushq %rbp
; X64: .seh_pushreg 5
; X64: pushq %rbx
; X64: .seh_pushreg 3
; X64: subq $40, %rsp
; X64: .seh_stackalloc 40
[WinEH] Make UnwindHelp a fixed stack object allocated after XMM CSRs Now the offset of UnwindHelp in our EH tables and the offset that we store to in the prologue agree. llvm-svn: 253059
2015-11-13 20:06:01 +01:00
; X64: leaq 96(%rdx), %rbp
[WinEH] Fix funclet prologues with stack realignment We already had a test for this for 32-bit SEH catchpads, but those don't actually create funclets. We had a bug that only appeared in funclet prologues, where we would establish EBP and ESI as our FP and BP, and then downstream prologue code would overwrite them. While I was at it, I fixed Win64+funclets+stackrealign. This issue doesn't come up as often there due to the ABI requring 16 byte stack alignment, but now we can rest easy that AVX and WinEH will work well together =P. llvm-svn: 252210
2015-11-05 22:09:49 +01:00
; X64: .seh_endprologue
; X64: andq $-32, %rdx
; X64: movq %rdx, %rbx
; X64-NOT: mov{{.*}}, %rbx
; X64: popq %rbp
; X64: retq # CLEANUPRET
Reference in New Issue Copy Permalink