Messages in this thread | | | Date | Mon, 16 Feb 1998 11:40:12 -0800 (PST) | From | Linus Torvalds <> | Subject | Re: How to read-protect a vm_area? |
| |
On Mon, 16 Feb 1998, Gabriel Paubert wrote: > > Does it mean that a write(2) with a buffer from an > mprotect( , , PROT_NONE)'ed area will succeed ?
Yes. I think a read() will modify the pages too (although that would be easy to change).
> I can't see right now how it could fail (on Intel with kernel v2.1.xx).
Do people really care? I've never seen this as really being an issue, but PROT_NONE can certainly be implemented other ways too. Any other way is likely to be rather more involved, so I really would prefer not to do anything about it.
Note that this is an issue only for something that _has_ been accessible and then has turned into PROT_NONE by using mprotect() - because if the pages aren't there at all then the page fault handler will refuse to page them in even when the access is done from kernel space.
Note that passing a bad pointer to a system call falls under "undefined behaviour" anyway (although most systems try to return EFAULT), so the answer as far as I'm concerned is "don't do that, then". Note that the system itself (or security) is not compromised by any of this.
Linus
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |