[LLVMbugs] [Bug 2104] New: Solaris test failure: Analysis/BasicAA/ 2007-08-01-NoAliasAndGEP.ll
bugzilla-daemon at cs.uiuc.edu
bugzilla-daemon at cs.uiuc.edu
Wed Feb 27 02:02:58 PST 2008
http://llvm.org/bugs/show_bug.cgi?id=2104
Summary: Solaris test failure: Analysis/BasicAA/2007-08-01-
NoAliasAndGEP.ll
Product: libraries
Version: trunk
Platform: Sun
OS/Version: Solaris
Status: NEW
Severity: minor
Priority: P2
Component: Global Analyses
AssignedTo: unassignedbugs at nondot.org
ReportedBy: ggreif at gmail.com
CC: llvmbugs at cs.uiuc.edu
Here is the failure:
Running .../llvm/test/Analysis/BasicAA/dg.exp ...
FAIL: .../llvm/test/Analysis/BasicAA/2007-08-01-NoAliasAndGEP.ll
Failed with exit(1) at line 3
while running: llvm-as
.../llvm/test/Analysis/BasicAA/2007-08-01-NoAliasAndGEP.ll -o - | opt -basicaa
-aa-eval -print-all-alias-modref-info -disable-output |& grep {MayAlias:.*i32\*
%Jpointer, i32\* %Ipointer}
child process exited abnormally
Looks like the analysis is correct, only the displaying is mixed up:
ggreif at my [!345] cd Debug/bin/
ggreif at my [!346] llvm-as
.../llvm/test/Analysis/BasicAA/2007-08-01-NoAliasAndGEP.ll -o - | opt -basicaa
-aa-eval -print-all-alias-modref-info -disable-output
Function: foo: 6 pointers, 0 call sites
NoAlias: i32* %Jpointer, i32* %qi
MayAlias: i32* %qj, i32* %qi
NoAlias: i32* %qj, i32* %Jpointer
NoAlias: i32* %Ipointer, i32* %qi
MayAlias: i32* %Ipointer, i32* %Jpointer <<<<< HERE WE GO
NoAlias: i32* %Ipointer, i32* %qj
MayAlias: i32* %q, i32* %qi
NoAlias: i32* %q, i32* %Jpointer
MayAlias: i32* %q, i32* %qj
NoAlias: i32* %q, i32* %Ipointer
NoAlias: i32* %p, i32* %qi
MayAlias: i32* %p, i32* %Jpointer
NoAlias: i32* %p, i32* %qj
MayAlias: i32* %p, i32* %Ipointer
NoAlias: i32* %p, i32* %q
===== Alias Analysis Evaluator Report =====
15 Total Alias Queries Performed
9 no alias responses (60.0%)
6 may alias responses (40.0%)
0 must alias responses (0.0%)
Alias Analysis Evaluator Pointer Alias Summary: 60%/40%/0%
Alias Analysis Mod/Ref Evaluator Summary: no mod/ref!
My theory is that on solaris the map/set is built up in a different way,
indeterministically.
For output we might need to sort the results, or fix "operator <" on the key
type to start with.
I am looking into this.
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
More information about the llvm-bugs
mailing list