Conversation

With 64 bit address spaces, thread stacks should have GBs reserved by default, and programming with alloca should be encouraged. A lot of string and container work could be done at high performance with alloca and no frees if you don't let them escape.

545

Replying to

Gotta love the man page. Don't get carried away. "The alloca() function returns a pointer to the beginning of the allocated space. If the allocation causes stack overflow, program behavior is undefined"

Old Man Yells At Cloud Provider

Replying to

and

Surely touching anything outside the stack limit would result in a page fault, no?

Replying to

and

It depends on the platform including the OS and compiler. You need -fstack-clash-protection with Clang and GCC on Linux to guarantee stack overflows are detected. Otherwise, you can easily skip over the guard page. It's not inherently caught from going past the end of the stack.

Touisteur EmporteUneVache

Replying to

Never quite understood the difference between that and -fstack-check?

Replying to

GCC -fstack-check doesn't provide protection in the general case and is actively dangerous on architectures without new-style checking. Clang doesn't actually implement it. -fstack-clash-protection is safe and provides full protection on architectures with proper support for it.

Touisteur EmporteUneVache

Replying to

Oh OK thanks. I've had great success catching stack overflows w/ -fstack-check and was wondering how they differed. Will read on this, thanks.

Replying to

It's covered in the GCC man page. Definitely stop using -fstack-check and use -fstack-clash-protection instead.

7:57 AM · May 2, 2021Twitter Web App

Touisteur EmporteUneVache

Replying to

Mmmh developers.redhat.com/blog/2019/04/3 helps with one detail, I write lots of Ada, rarely mixed. Mmmh. Still the 1-3 pages might be why I had some trouble with some stack overflow behaviour... Thanks! Will read the gcc man page too... TIL

developers.redhat.com

Stack Clash mitigation in GCC: Why -fstack-check is not the answer | Red Hat Developer

In our previous article about Stack Clash, we covered the basics of the Stack Clash vulnerability. To summarize, an attacker first uses various means to bring the heap and stack close together. A...

Replying to

It's safe on an architecture with a proper implementation if you have pure Ada processes without any C code (no libc usage, etc.) or assembly code that's not doing the probes, etc. It's unsafe in the presence of code without the checks. It can actually be worse than not doing it.

Show replies