News for hackers. Stuff that matters.
Re:Syswall: a firewall for syscalls by equalunique on Sunday February 17, 2019 @12:57AM (#19182893)
Sounds a little bit like OpenBSD's pledge - was that a source of inspiration?
Re:Syswall: a firewall for syscalls by polaris64 on Sunday February 17, 2019 @04:24AM (#19183542)
Yes, partially, although I wanted to create a more interactive system for end-users to reason about software. I wouldn't recommend it (certainly not yet at least) for system security, tools like seccomp and pledge will do a better job there.
Re:Syswall: a firewall for syscalls by lsofzz on Monday February 18, 2019 @03:22AM (#19189278)
Nice work. Just read this :-)
Re:Syswall: a firewall for syscalls by jquast on Saturday February 16, 2019 @08:10PM (#19181987)
sounds like https://en.wikipedia.org/wiki/Systrace#Features

more ... http://www.citi.umich.edu/u/provos/systrace/

it didn't work out then, but best wishes to the new generation

Re:Syswall: a firewall for syscalls by yazr on Sunday February 17, 2019 @04:37AM (#19183594)
What do you mean by it didn't work? Not adopted or are there implementation issues?

Also - for the OP. Don't be discouraged there's so many features to add, management, attribution, machine learning etc

Re:Syswall: a firewall for syscalls by zxombie on Sunday February 17, 2019 @07:03AM (#19184046)
There were some security issues in systrace. /http://www.watson.org/~robert/2007woot/
Re:Syswall: a firewall for syscalls by polaris64 on Sunday February 17, 2019 @04:27AM (#19183555)
You're right; I hadn't come across systrace before actually, but it's very similar to what syswall is trying to achieve. I'll be sure to take a look!
Re:Syswall: a firewall for syscalls by lsofzz on Monday February 18, 2019 @03:21AM (#19189272)
How is this different from seccomp?
Re:Syswall: a firewall for syscalls by polaris64 on Monday February 18, 2019 @12:24PM (#19192293)
seccomp is a robust way of restricting a process's syscalls so that it can only do what you allow it to.

syswall is more of an interactive tool (similar to systrace as mentioned in another comment). The goal is not to replace seccomp (it's certainly not meant to provide complete security), but rather to allow users to reason about what a process is actually doing. For example, allowing users to see if a new version does something different from the previous, perhaps meaning that malicious code was added unexpectedly.