llvm/test/Transforms/DeadArgElim/keepalive.ll
David Majnemer 39a09d2b7c IR: Change inalloca's grammar a bit
The grammar for LLVM IR is not well specified in any document but seems
to obey the following rules:

 - Attributes which have parenthesized arguments are never preceded by
   commas.  This form of attribute is the only one which ever has
   optional arguments.  However, not all of these attributes support
   optional arguments: 'thread_local' supports an optional argument but
   'addrspace' does not.  Interestingly, 'addrspace' is documented as
   being a "qualifier".  What constitutes a qualifier?  I cannot find a
   definition.

 - Some attributes use a space between the keyword and the value.
   Examples of this form are 'align' and 'section'.  These are always
   preceded by a comma.

 - Otherwise, the attribute has no argument.  These attributes do not
   have a preceding comma.

Sometimes an attribute goes before the instruction, between the
instruction and it's type, or after it's type.  'atomicrmw' has
'volatile' between the instruction and the type while 'call' has 'tail'
preceding the instruction.

With all this in mind, it seems most consistent for 'inalloca' on an
'inalloca' instruction to occur before between the instruction and the
type.  Unlike the current formulation, there would be no preceding
comma.  The combination 'alloca inalloca' doesn't look particularly
appetizing, perhaps a better spelling of 'inalloca' is down the road.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@203376 91177308-0d34-0410-b5e6-96231b3b80d8
2014-03-09 06:41:58 +00:00

48 lines
1.5 KiB
LLVM

; RUN: opt < %s -deadargelim -S | FileCheck %s
%Ty = type <{ i32, i32 }>
; Check if the pass doesn't modify anything that doesn't need changing. We feed
; an unused argument to each function to lure it into changing _something_ about
; the function and then changing too much.
; This checks if the return value attributes are not removed
; CHECK: define internal zeroext i32 @test1() #0
define internal zeroext i32 @test1(i32 %DEADARG1) nounwind {
ret i32 1
}
; This checks if the struct doesn't get non-packed
; CHECK-LABEL: define internal <{ i32, i32 }> @test2(
define internal <{ i32, i32 }> @test2(i32 %DEADARG1) {
ret <{ i32, i32 }> <{ i32 1, i32 2 }>
}
; We use this external function to make sure the return values don't become dead
declare void @user(i32, <{ i32, i32 }>)
define void @caller() {
%B = call i32 @test1(i32 1)
%C = call <{ i32, i32 }> @test2(i32 2)
call void @user(i32 %B, <{ i32, i32 }> %C)
ret void
}
; We can't remove 'this' here, as that would put argmem in ecx instead of
; memory.
define internal x86_thiscallcc i32 @unused_this(i32* %this, i32* inalloca %argmem) {
%v = load i32* %argmem
ret i32 %v
}
; CHECK-LABEL: define internal x86_thiscallcc i32 @unused_this(i32* %this, i32* inalloca %argmem)
define i32 @caller2() {
%t = alloca i32
%m = alloca inalloca i32
store i32 42, i32* %m
%v = call x86_thiscallcc i32 @unused_this(i32* %t, i32* inalloca %m)
ret i32 %v
}
; CHECK: attributes #0 = { nounwind }