[Webkit-unassigned] [Bug 142625] New: Crash in JSC::Interpreter::execute
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu Mar 12 08:57:17 PDT 2015
https://bugs.webkit.org/show_bug.cgi?id=142625
Bug ID: 142625
Summary: Crash in JSC::Interpreter::execute
Classification: Unclassified
Product: WebKit
Version: 528+ (Nightly build)
Hardware: PC
OS: Linux
Status: NEW
Severity: Normal
Priority: P2
Component: JavaScriptCore
Assignee: webkit-unassigned at lists.webkit.org
Reporter: rhodovan.u-szeged at partner.samsung.com
CC: fpizlo at apple.com, ggaren at apple.com, msaboff at apple.com,
oliver at apple.com
Blocks: 116980
Created attachment 248519
--> https://bugs.webkit.org/attachment.cgi?id=248519&action=review
Test case
If you load this with TotT JSC (on Ubuntu 14.04, x86_64):
do
eval("function fuzz() {}");
while(true);
then you will get a crash with the following backtrace:
Program received signal SIGSEGV, Segmentation fault.
0x00007ffff73de2d9 in WTFCrash () at ../../Source/WTF/wtf/Assertions.cpp:321
321 *(int *)(uintptr_t)0xbbadbeef = 0;
(gdb) bt
#0 0x00007ffff73de2d9 in WTFCrash () at ../../Source/WTF/wtf/Assertions.cpp:321
#1 0x00000000004280fa in WTF::CrashOnOverflow::overflowed () at ../../Source/WTF/wtf/CheckedArithmetic.h:78
#2 0x00007ffff6ce7899 in WTF::Vector<JSC::WriteBarrier<JSC::FunctionExecutable>, 0ul, WTF::CrashOnOverflow>::at (this=0x7fffef7f06d0, i=0)
at ../../Source/WTF/wtf/Vector.h:659
#3 0x00007ffff6cdf303 in WTF::Vector<JSC::WriteBarrier<JSC::FunctionExecutable>, 0ul, WTF::CrashOnOverflow>::operator[] (this=0x7fffef7f06d0, i=0)
at ../../Source/WTF/wtf/Vector.h:679
#4 0x00007ffff6dd1ba8 in JSC::CodeBlock::functionDecl (this=0x7fffef7f04d0, index=0) at ../../Source/JavaScriptCore/bytecode/CodeBlock.h:657
#5 0x00007ffff709b0ee in JSC::Interpreter::execute (this=0x7fffefff6000, eval=0x7fffee16fb70, callFrame=0x7fffffffcac0, thisValue=...,
scope=0x7fffee0af970) at ../../Source/JavaScriptCore/interpreter/Interpreter.cpp:1119
#6 0x00007ffff7096ca1 in JSC::eval (callFrame=0x7fffffffcac0) at ../../Source/JavaScriptCore/interpreter/Interpreter.cpp:134
#7 0x00007ffff70ce0a9 in JSC::operationCallEval (exec=0x7fffffffcb10, execCallee=0x7fffffffcac0)
at ../../Source/JavaScriptCore/jit/JITOperations.cpp:638
#8 0x00007fffadfffe06 in ?? ()
#9 0x00007fffffffcb10 in ?? ()
#10 0x00007ffff738d751 in llint_entry () from /home/reni/data/REPOS/webkit/WebKitBuild/Debug/lib/libjavascriptcoregtk-4.0.so.18
#11 0x00007ffff7387966 in vmEntryToJavaScript () from /home/reni/data/REPOS/webkit/WebKitBuild/Debug/lib/libjavascriptcoregtk-4.0.so.18
#12 0x00007ffff70be1fb in JSC::JITCode::execute (this=0x7fffefff7900, vm=0x7fffee010000, protoCallFrame=0x7fffffffcd30)
at ../../Source/JavaScriptCore/jit/JITCode.cpp:77
#13 0x00007ffff70997dc in JSC::Interpreter::execute (this=0x7fffefff6000, program=0x7fffee16fc70, callFrame=0x7fffee0af9b0, thisObj=0x7fffee0cfaf0)
at ../../Source/JavaScriptCore/interpreter/Interpreter.cpp:855
#14 0x00007ffff72299ca in JSC::evaluate (exec=0x7fffee0af9b0, source=..., thisValue=..., returnedException=0x7fffffffd6b0)
at ../../Source/JavaScriptCore/runtime/Completion.cpp:81
#15 0x000000000042648f in runWithScripts (globalObject=0x7fffee0af970, scripts=..., dump=false) at ../../Source/JavaScriptCore/jsc.cpp:1264
#16 0x00000000004272c4 in jscmain (argc=2, argv=0x7fffffffd928) at ../../Source/JavaScriptCore/jsc.cpp:1481
#17 0x000000000042627a in main (argc=2, argv=0x7fffffffd928) at ../../Source/JavaScriptCore/jsc.cpp:1222
>From Source/JavaScriptCore/interpreter/Interpreter.cpp:
1118 for (int i = 0; i < numFunctions; ++i) {
1119 FunctionExecutable* function = codeBlock->functionDecl(i);
...
...
1122 }
The crash happens in line 1119. After a short debugging it seems that at the moment of the crash the index of the loop is 0 and the value of |numFunctions| is 1. However, since |codeBlock| doesn't contain any functionDeclarations at this point, we crash. One more note that could be important: the crash happens around the 4096th eval execution.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150312/4c49167a/attachment-0001.html>
More information about the webkit-unassigned
mailing list