6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-29 10:32:47 +00:00
Code Issues Projects Releases Wiki Activity
faf98904b7
llvm-6502/test/Transforms/Inline/inline_invoke.ll

353 lines
11 KiB
LLVM
Raw Normal View History

Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
; RUN: opt < %s -inline -S | FileCheck %s
; Test that the inliner correctly handles inlining into invoke sites
; by appending selectors and forwarding _Unwind_Resume directly to the
; enclosing landing pad.
First, do no harm -- even if we can't find a selector for an enclosing landing pad, forward llvm.eh.resume calls to it instead of turning them invalidly into invokes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132382 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-01 02:17:11 +00:00
;; Test 0 - basic functionality.
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
%struct.A = type { i8 }
@_ZTIi = external constant i8*
declare void @_ZN1AC1Ev(%struct.A*)
declare void @_ZN1AD1Ev(%struct.A*)
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
declare void @use(i32) nounwind
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
declare void @opaque()
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
declare i32 @llvm.eh.typeid.for(i8*) nounwind
declare i32 @__gxx_personality_v0(...)
declare i8* @__cxa_begin_catch(i8*)
declare void @__cxa_end_catch()
declare void @_ZSt9terminatev()
define internal void @test0_in() alwaysinline uwtable ssp {
entry:
%a = alloca %struct.A, align 1
%b = alloca %struct.A, align 1
call void @_ZN1AC1Ev(%struct.A* %a)
invoke void @_ZN1AC1Ev(%struct.A* %b)
to label %invoke.cont unwind label %lpad
invoke.cont:
invoke void @_ZN1AD1Ev(%struct.A* %b)
to label %invoke.cont1 unwind label %lpad
invoke.cont1:
call void @_ZN1AD1Ev(%struct.A* %a)
ret void
lpad:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
cleanup
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
invoke void @_ZN1AD1Ev(%struct.A* %a)
to label %invoke.cont2 unwind label %terminate.lpad
invoke.cont2:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
resume { i8*, i32 } %exn
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
terminate.lpad:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn1 = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
catch i8* null
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
call void @_ZSt9terminatev() noreturn nounwind
unreachable
}
define void @test0_out() uwtable ssp {
entry:
invoke void @test0_in()
to label %ret unwind label %lpad
ret:
ret void
lpad: ; preds = %entry
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
catch i8* bitcast (i8** @_ZTIi to i8*)
%eh.exc = extractvalue { i8*, i32 } %exn, 0
%eh.selector = extractvalue { i8*, i32 } %exn, 1
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
%0 = call i32 @llvm.eh.typeid.for(i8* bitcast (i8** @_ZTIi to i8*)) nounwind
%1 = icmp eq i32 %eh.selector, %0
br i1 %1, label %catch, label %eh.resume
catch:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%ignored = call i8* @__cxa_begin_catch(i8* %eh.exc) nounwind
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
call void @__cxa_end_catch() nounwind
br label %ret
eh.resume:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
resume { i8*, i32 } %exn
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
}
Implement and document the llvm.eh.resume intrinsic, which is transformed by the inliner into a branch to the enclosing landing pad (when inlined through an invoke). If not so optimized, it is lowered DWARF EH preparation into a call to _Unwind_Resume (or _Unwind_SjLj_Resume as appropriate). Its chief advantage is that it takes both the exception value and the selector value as arguments, meaning that there is zero effort in recovering these; however, the frontend is required to pass these down, which is not actually particularly difficult. Also document the behavior of landing pads a bit better, and make it clearer that it's okay that personality functions don't always land at landing pads. This is just a fact of life. Don't write optimizations that rely on pushing things over an unwind edge. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132253 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-28 07:45:59 +00:00
; CHECK: define void @test0_out()
; CHECK: [[A:%.*]] = alloca %struct.A,
; CHECK: [[B:%.*]] = alloca %struct.A,
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[A]])
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[B]])
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[B]])
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[A]])
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK: landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: cleanup
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Update test. There may be multiple catches, but those will be cleaned up. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177758 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-22 20:36:39 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Fix the inliner to maintain the current de facto invoke semantics: - the selector for the landing pad must provide all available information about the handlers, filters, and cleanups within that landing pad - calls to _Unwind_Resume must be converted to branches to the enclosing lpad so as to avoid re-entering the unwinder when the lpad claimed it was going to handle the exception in some way This is quite specific to libUnwind-based unwinding. In an effort to not interfere too badly with other unwinders, and with existing hacks in frontends, this only triggers on _Unwind_Resume (not _Unwind_Resume_or_Rethrow) and does nothing with selectors if it cannot find a selector call for either lpad. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132200 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-27 18:34:38 +00:00
; CHECK-NEXT: invoke void @_ZN1AD1Ev(%struct.A* [[A]])
; CHECK-NEXT: to label %[[LBL:[^\s]+]] unwind
; CHECK: [[LBL]]:
; CHECK-NEXT: br label %[[LPAD:[^\s]+]]
Implement and document the llvm.eh.resume intrinsic, which is transformed by the inliner into a branch to the enclosing landing pad (when inlined through an invoke). If not so optimized, it is lowered DWARF EH preparation into a call to _Unwind_Resume (or _Unwind_SjLj_Resume as appropriate). Its chief advantage is that it takes both the exception value and the selector value as arguments, meaning that there is zero effort in recovering these; however, the frontend is required to pass these down, which is not actually particularly difficult. Also document the behavior of landing pads a bit better, and make it clearer that it's okay that personality functions don't always land at landing pads. This is just a fact of life. Don't write optimizations that rely on pushing things over an unwind edge. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132253 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-28 07:45:59 +00:00
; CHECK: ret void
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK: landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Implement and document the llvm.eh.resume intrinsic, which is transformed by the inliner into a branch to the enclosing landing pad (when inlined through an invoke). If not so optimized, it is lowered DWARF EH preparation into a call to _Unwind_Resume (or _Unwind_SjLj_Resume as appropriate). Its chief advantage is that it takes both the exception value and the selector value as arguments, meaning that there is zero effort in recovering these; however, the frontend is required to pass these down, which is not actually particularly difficult. Also document the behavior of landing pads a bit better, and make it clearer that it's okay that personality functions don't always land at landing pads. This is just a fact of life. Don't write optimizations that rely on pushing things over an unwind edge. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132253 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-28 07:45:59 +00:00
; CHECK-NEXT: br label %[[LPAD]]
; CHECK: [[LPAD]]:
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: phi { i8*, i32 } [
; CHECK-NEXT: extractvalue { i8*, i32 }
; CHECK-NEXT: extractvalue { i8*, i32 }
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
; CHECK-NEXT: call i32 @llvm.eh.typeid.for(
First, do no harm -- even if we can't find a selector for an enclosing landing pad, forward llvm.eh.resume calls to it instead of turning them invalidly into invokes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132382 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-01 02:17:11 +00:00
;; Test 1 - Correctly handle phis in outer landing pads.
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
define void @test1_out() uwtable ssp {
entry:
invoke void @test0_in()
to label %cont unwind label %lpad
cont:
invoke void @test0_in()
to label %ret unwind label %lpad
ret:
ret void
lpad:
%x = phi i32 [ 0, %entry ], [ 1, %cont ]
%y = phi i32 [ 1, %entry ], [ 4, %cont ]
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
catch i8* bitcast (i8** @_ZTIi to i8*)
%eh.exc = extractvalue { i8*, i32 } %exn, 0
%eh.selector = extractvalue { i8*, i32 } %exn, 1
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
%0 = call i32 @llvm.eh.typeid.for(i8* bitcast (i8** @_ZTIi to i8*)) nounwind
%1 = icmp eq i32 %eh.selector, %0
br i1 %1, label %catch, label %eh.resume
catch:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%ignored = call i8* @__cxa_begin_catch(i8* %eh.exc) nounwind
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
call void @use(i32 %x)
call void @use(i32 %y)
call void @__cxa_end_catch() nounwind
br label %ret
eh.resume:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
resume { i8*, i32 } %exn
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
}
; CHECK: define void @test1_out()
; CHECK: [[A2:%.*]] = alloca %struct.A,
; CHECK: [[B2:%.*]] = alloca %struct.A,
; CHECK: [[A1:%.*]] = alloca %struct.A,
; CHECK: [[B1:%.*]] = alloca %struct.A,
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[A1]])
; CHECK-NEXT: unwind label %[[LPAD:[^\s]+]]
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[B1]])
; CHECK-NEXT: unwind label %[[LPAD1:[^\s]+]]
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[B1]])
; CHECK-NEXT: unwind label %[[LPAD1]]
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[A1]])
; CHECK-NEXT: unwind label %[[LPAD]]
; Inner landing pad from first inlining.
; CHECK: [[LPAD1]]:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK-NEXT: [[LPADVAL1:%.*]] = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: cleanup
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Update test. There may be multiple catches, but those will be cleaned up. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177758 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-22 20:36:39 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
; CHECK-NEXT: invoke void @_ZN1AD1Ev(%struct.A* [[A1]])
; CHECK-NEXT: to label %[[RESUME1:[^\s]+]] unwind
; CHECK: [[RESUME1]]:
; CHECK-NEXT: br label %[[LPAD_JOIN1:[^\s]+]]
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[A2]])
; CHECK-NEXT: unwind label %[[LPAD]]
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[B2]])
; CHECK-NEXT: unwind label %[[LPAD2:[^\s]+]]
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[B2]])
; CHECK-NEXT: unwind label %[[LPAD2]]
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[A2]])
; CHECK-NEXT: unwind label %[[LPAD]]
; Inner landing pad from second inlining.
; CHECK: [[LPAD2]]:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK-NEXT: [[LPADVAL2:%.*]] = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: cleanup
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Update test. There may be multiple catches, but those will be cleaned up. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177758 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-22 20:36:39 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
; CHECK-NEXT: invoke void @_ZN1AD1Ev(%struct.A* [[A2]])
; CHECK-NEXT: to label %[[RESUME2:[^\s]+]] unwind
; CHECK: [[RESUME2]]:
; CHECK-NEXT: br label %[[LPAD_JOIN2:[^\s]+]]
; CHECK: ret void
; CHECK: [[LPAD]]:
; CHECK-NEXT: [[X:%.*]] = phi i32 [ 0, %entry ], [ 0, {{%.*}} ], [ 1, %cont ], [ 1, {{%.*}} ]
; CHECK-NEXT: [[Y:%.*]] = phi i32 [ 1, %entry ], [ 1, {{%.*}} ], [ 4, %cont ], [ 4, {{%.*}} ]
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK-NEXT: [[LPADVAL:%.*]] = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
; CHECK-NEXT: br label %[[LPAD_JOIN2]]
; CHECK: [[LPAD_JOIN2]]:
; CHECK-NEXT: [[XJ2:%.*]] = phi i32 [ [[X]], %[[LPAD]] ], [ 1, %[[RESUME2]] ]
; CHECK-NEXT: [[YJ2:%.*]] = phi i32 [ [[Y]], %[[LPAD]] ], [ 4, %[[RESUME2]] ]
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: [[EXNJ2:%.*]] = phi { i8*, i32 } [ [[LPADVAL]], %[[LPAD]] ], [ [[LPADVAL2]], %[[RESUME2]] ]
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
; CHECK-NEXT: br label %[[LPAD_JOIN1]]
; CHECK: [[LPAD_JOIN1]]:
; CHECK-NEXT: [[XJ1:%.*]] = phi i32 [ [[XJ2]], %[[LPAD_JOIN2]] ], [ 0, %[[RESUME1]] ]
; CHECK-NEXT: [[YJ1:%.*]] = phi i32 [ [[YJ2]], %[[LPAD_JOIN2]] ], [ 1, %[[RESUME1]] ]
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: [[EXNJ1:%.*]] = phi { i8*, i32 } [ [[EXNJ2]], %[[LPAD_JOIN2]] ], [ [[LPADVAL1]], %[[RESUME1]] ]
; CHECK-NEXT: extractvalue { i8*, i32 } [[EXNJ1]], 0
; CHECK-NEXT: [[SELJ1:%.*]] = extractvalue { i8*, i32 } [[EXNJ1]], 1
Add the test case for phis in the outer landing pad during the inliner's forwarding of eh.resume that I promised yesterday. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132307 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-30 01:08:04 +00:00
; CHECK-NEXT: [[T:%.*]] = call i32 @llvm.eh.typeid.for(
; CHECK-NEXT: icmp eq i32 [[SELJ1]], [[T]]
; CHECK: call void @use(i32 [[XJ1]])
; CHECK: call void @use(i32 [[YJ1]])
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK: resume { i8*, i32 }
First, do no harm -- even if we can't find a selector for an enclosing landing pad, forward llvm.eh.resume calls to it instead of turning them invalidly into invokes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132382 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-01 02:17:11 +00:00
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
;; Test 2 - Don't make invalid IR for inlines into landing pads without eh.exception calls
First, do no harm -- even if we can't find a selector for an enclosing landing pad, forward llvm.eh.resume calls to it instead of turning them invalidly into invokes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132382 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-01 02:17:11 +00:00
define void @test2_out() uwtable ssp {
entry:
invoke void @test0_in()
to label %ret unwind label %lpad
ret:
ret void
lpad:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
cleanup
First, do no harm -- even if we can't find a selector for an enclosing landing pad, forward llvm.eh.resume calls to it instead of turning them invalidly into invokes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132382 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-01 02:17:11 +00:00
call void @_ZSt9terminatev()
unreachable
}
; CHECK: define void @test2_out()
; CHECK: [[A:%.*]] = alloca %struct.A,
; CHECK: [[B:%.*]] = alloca %struct.A,
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[A]])
; CHECK-NEXT: unwind label %[[LPAD:[^\s]+]]
; CHECK: invoke void @_ZN1AC1Ev(%struct.A* [[B]])
; CHECK-NEXT: unwind label %[[LPAD2:[^\s]+]]
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[B]])
; CHECK-NEXT: unwind label %[[LPAD2]]
; CHECK: invoke void @_ZN1AD1Ev(%struct.A* [[A]])
; CHECK-NEXT: unwind label %[[LPAD]]
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
;; Test 3 - Deal correctly with split unwind edges.
define void @test3_out() uwtable ssp {
entry:
invoke void @test0_in()
to label %ret unwind label %lpad
ret:
ret void
lpad:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
catch i8* bitcast (i8** @_ZTIi to i8*)
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
br label %lpad.cont
lpad.cont:
call void @_ZSt9terminatev()
unreachable
}
; CHECK: define void @test3_out()
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK: landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: cleanup
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Update test. There may be multiple catches, but those will be cleaned up. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177758 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-22 20:36:39 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: invoke void @_ZN1AD1Ev(
; CHECK-NEXT: to label %[[L:[^\s]+]] unwind
; CHECK: [[L]]:
; CHECK-NEXT: br label %[[JOIN:[^\s]+]]
; CHECK: [[JOIN]]:
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: phi { i8*, i32 }
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: br label %lpad.cont
; CHECK: lpad.cont:
; CHECK-NEXT: call void @_ZSt9terminatev()
;; Test 4 - Split unwind edges with a dominance problem
define void @test4_out() uwtable ssp {
entry:
invoke void @test0_in()
to label %cont unwind label %lpad.crit
cont:
invoke void @opaque()
to label %ret unwind label %lpad
ret:
ret void
lpad.crit:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
catch i8* bitcast (i8** @_ZTIi to i8*)
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
call void @opaque() nounwind
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
br label %terminate
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
lpad:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
%exn2 = landingpad {i8*, i32} personality i32 (...)* @__gxx_personality_v0
catch i8* bitcast (i8** @_ZTIi to i8*)
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
br label %terminate
terminate:
%phi = phi i32 [ 0, %lpad.crit ], [ 1, %lpad ]
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
call void @use(i32 %phi)
call void @_ZSt9terminatev()
unreachable
}
; CHECK: define void @test4_out()
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK: landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: cleanup
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
Update test. There may be multiple catches, but those will be cleaned up. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177758 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-22 20:36:39 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: invoke void @_ZN1AD1Ev(
; CHECK-NEXT: to label %[[L:[^\s]+]] unwind
; CHECK: [[L]]:
; CHECK-NEXT: br label %[[JOIN:[^\s]+]]
; CHECK: invoke void @opaque()
; CHECK-NEXT: unwind label %lpad
; CHECK: lpad.crit:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK-NEXT: landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: br label %[[JOIN]]
; CHECK: [[JOIN]]:
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: phi { i8*, i32 }
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175877 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-22 09:09:42 +00:00
; CHECK-NEXT: call void @opaque() [[NUW:#[0-9]+]]
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: br label %[[FIX:[^\s]+]]
; CHECK: lpad:
Update to new EH scheme. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-02 01:25:11 +00:00
; CHECK-NEXT: landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: catch i8* bitcast (i8** @_ZTIi to i8*)
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: br label %[[FIX]]
; CHECK: [[FIX]]:
Auto upgrade the old EH scheme to use the new one. This is on a trial basis. If things to disasterously over night, this can be reverted. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@138702 91177308-0d34-0410-b5e6-96231b3b80d8
2011-08-27 06:11:03 +00:00
; CHECK-NEXT: [[T1:%.*]] = phi i32 [ 0, %[[JOIN]] ], [ 1, %lpad ]
SplitCriticalEdge can sometimes split the edge from an invoke to a landing pad, separating the exception and selector calls from the new lpad. Teaching it not to do that, or to properly adjust the CFG afterwards, is out of scope because it would require the other edges to the landing pad to be split as well (effectively). Instead, just recover from the most likely cases during inlining. The best long-term solution is to change the exception representation and commit to either requiring or not requiring the more complex edge-splitting logic; this is just a shorter-term hack. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132799 91177308-0d34-0410-b5e6-96231b3b80d8
2011-06-09 20:06:24 +00:00
; CHECK-NEXT: call void @use(i32 [[T1]])
; CHECK-NEXT: call void @_ZSt9terminatev()
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175877 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-22 09:09:42 +00:00
; CHECK: attributes [[NUW]] = { nounwind }
; CHECK: attributes #1 = { nounwind readnone }
; CHECK: attributes #2 = { ssp uwtable }
; CHECK: attributes #3 = { noreturn nounwind }
Reference in New Issue Copy Permalink