Re: [RFC/PATCH] revoke/frevoke system calls V2

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

Ar Mer, 2006-08-09 am 10:41 +0200, ysgrifennodd Edgar Toernig:
> If I read the code correctly, the behaviour for hung up ttys is completely
> different: read returns EOF, write returns EIO, select/poll/epoll return
> ready, close works.  As rather boring but totally sane behaviour for an fd.
> 
> But after revoke you get EBADF for any operation, even select or close.

Thats a detail of the proposed implementation that isn't hard to fix.

> And IMHO that's insane that a regular user may close fds in someone else's
> processes (or munmap some of its memory).  I already see people trying
> to exploit bugs in system services:

I can do this already today. In fact the index.html one can be used to
crash certain products now depending on their configuration. Just do

	while { true } do
		cp some.html index.html
		> index.html
	done

with a shell that truncates on > and you'll be able to bus error them if
they mmap and are not well written.

These are not actually changes in behaviour. At any point I can shrink a
file I own and you get read -> 0, mmap access -> bus error. Write
behaviour is new but thats no different to filling the disk up or other
real write errors.