3adc6e0308
As pointed out in https://discourse.llvm.org/t/undeterministic-thin-index-file/69985, the block count added to distributed ThinLTO index files breaks incremental builds on ThinLTO - if any linked file has a different number of BBs, then the accumulated sum placed in the index files will change, causing all ThinLTO backend compiles to be redone. The block count is only used for scaling of partial sample profiles, and was added in D80403 for D79831. This patch simply removes this field from the index files of non partial sample profile compiles, which is NFC on the output of the compiler. We subsequently need to see if this can be removed for partial sample profiles without signficant performance loss, or redesigned in a way that does not destroy caching. Differential Revision: https://reviews.llvm.org/D148746
82 lines
2.3 KiB
LLVM
82 lines
2.3 KiB
LLVM
; RUN: opt -passes=name-anon-globals -module-summary < %s | llvm-bcanalyzer -dump | FileCheck %s -check-prefix=BC
|
|
; Check for summary block/records.
|
|
|
|
; BC: <SOURCE_FILENAME
|
|
; "h"
|
|
; BC-NEXT: <GLOBALVAR {{.*}} op0=0 op1=1
|
|
; "foo"
|
|
; BC-NEXT: <FUNCTION op0=1 op1=3
|
|
; "bar"
|
|
; BC-NEXT: <FUNCTION op0=4 op1=3
|
|
; "anon.[32 chars].0"
|
|
; BC-NEXT: <FUNCTION op0=7 op1=39
|
|
; "variadic"
|
|
; BC-NEXT: <FUNCTION op0=46 op1=8
|
|
; "llvm.va_start"
|
|
; BC-NEXT: <FUNCTION op0=54 op1=13
|
|
; "f"
|
|
; BC-NEXT: <ALIAS op0=67 op1=1
|
|
; BC: <GLOBALVAL_SUMMARY_BLOCK
|
|
; BC-NEXT: <VERSION
|
|
; BC-NEXT: <FLAGS
|
|
; BC-NEXT: <PERMODULE {{.*}} op0=1 op1=0
|
|
; BC-NEXT: <PERMODULE {{.*}} op0=2 op1=0
|
|
; BC-NEXT: <PERMODULE {{.*}} op0=3 op1=7
|
|
; BC-NEXT: <PERMODULE {{.*}} op0=4 op1=0 op2=4 op3=0
|
|
; BC-NEXT: <ALIAS {{.*}} op0=6 op1=0 op2=3
|
|
; BC-NEXT: </GLOBALVAL_SUMMARY_BLOCK
|
|
; BC: <STRTAB_BLOCK
|
|
; BC-NEXT: blob data = 'hfoobaranon.{{................................}}.0variadicllvm.va_startf{{.*}}'
|
|
|
|
|
|
; RUN: opt -passes=name-anon-globals -module-summary < %s | llvm-dis | FileCheck %s
|
|
; RUN: opt -passes=name-anon-globals -module-summary -S < %s | FileCheck %s
|
|
; RUN: opt -passes=name-anon-globals -module-summary -S < %s | llvm-as | llvm-dis | FileCheck %s
|
|
; Check that this round-trips correctly.
|
|
|
|
; ModuleID = '<stdin>'
|
|
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
|
|
target triple = "x86_64-unknown-linux-gnu"
|
|
|
|
; CHECK: define i32 @foo()
|
|
|
|
; Function Attrs: nounwind uwtable
|
|
define i32 @foo() #0 {
|
|
entry:
|
|
ret i32 1
|
|
}
|
|
|
|
; CHECK: define i32 @bar(i32 %x)
|
|
|
|
; Function Attrs: nounwind uwtable
|
|
define i32 @bar(i32 %x) #0 {
|
|
entry:
|
|
ret i32 %x
|
|
}
|
|
|
|
; FIXME: Anonymous function and alias not currently in summary until
|
|
; follow on fixes to rename anonymous globals and emit alias summary
|
|
; entries are committed.
|
|
; Check an anonymous function as well, since in that case only the alias
|
|
; ends up in the value symbol table and having a summary.
|
|
@f = alias void (), void ()* @0 ; <ptr> [#uses=0]
|
|
@h = external global void ()* ; <ptr> [#uses=0]
|
|
|
|
define internal void @0() nounwind {
|
|
entry:
|
|
store void()* @0, void()** @h
|
|
br label %return
|
|
|
|
return: ; preds = %entry
|
|
ret void
|
|
}
|
|
|
|
define i32 @variadic(...) {
|
|
%ap = alloca i8*, align 8
|
|
%ap.0 = bitcast i8** %ap to i8*
|
|
call void @llvm.va_start(i8* %ap.0)
|
|
ret i32 42
|
|
}
|
|
|
|
declare void @llvm.va_start(i8*) nounwind
|