[llvm-bugs] [Bug 26495] New: Inliner fails with negative threshholds
via llvm-bugs
llvm-bugs at lists.llvm.org
Fri Feb 5 09:42:45 PST 2016
https://llvm.org/bugs/show_bug.cgi?id=26495
Bug ID: 26495
Summary: Inliner fails with negative threshholds
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: Interprocedural Optimizations
Assignee: unassignedbugs at nondot.org
Reporter: hans at chromium.org
CC: david.majnemer at gmail.com, eraman at google.com,
llvm-bugs at lists.llvm.org
Classification: Unclassified
David reported on IRC the other week that we fail to inline g into f and f into
h here:
target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-pc-windows-msvc"
define void @h(i8* nocapture readnone %this) {
entry:
call void @f(i8* %this)
unreachable
}
define linkonce_odr void @f(i8* nocapture readnone %this) {
entry:
tail call void @g(i8* %this)
unreachable
}
define linkonce_odr void @g(i8* nocapture readnone %this) {
entry:
br label %for.cond
for.cond:
tail call void @foo(i32 1)
br label %for.cond
}
declare void @foo(i32)
$ bin/opt -inline /tmp/a.ll -S -debug
[..]
Analyzing call of g...
NumConstantArgs: 0
NumConstantOffsetPtrArgs: 1
NumAllocaArgs: 0
NumConstantPtrCmps: 0
NumConstantPtrDiffs: 0
NumInstructionsSimplified: 1
NumInstructions: 2
SROACostSavings: 0
SROACostSavingsLost: 0
ContainsNoDuplicateCall: 0
Cost: 30
Threshold: -337
NOT Inlining: cost=30, thres=-337, Call: tail call void @g(i8* %this)
Inliner visiting SCC: h: 1 call sites.
Analyzing call of f...
NumConstantArgs: 0
NumConstantOffsetPtrArgs: 1
NumAllocaArgs: 0
NumConstantPtrCmps: 0
NumConstantPtrDiffs: 0
NumInstructionsSimplified: 0
NumInstructions: 1
SROACostSavings: 0
SROACostSavingsLost: 0
ContainsNoDuplicateCall: 0
Cost: 30
Threshold: -337
NOT Inlining: cost=30, thres=-337, Call: call void @f(i8* %this)
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20160205/2cc70c16/attachment-0001.html>
More information about the llvm-bugs
mailing list