Re: Linprocfs observation.

2000-03-28 Thread Dag-Erling Smorgrav
David Malone <[EMAIL PROTECTED]> writes: > I haven't checked carefully, but I expect that the linprocfs code > has the same problem as the FreeBSD procfs code, in that it can > expose suid executables which would not usually be run 'cos they > are in inaccessible directories. That is indeed corre

Re: Linprocfs observation.

2000-03-27 Thread Warner Losh
In message <[EMAIL PROTECTED]> David Malone writes: : The "file" file was removed from FreeBSD's /proc code (in 4.0 and : 5.0) because of this, but it is probably important for Linux : emulation so it can't really be removed from the linprocfs code. : I guess this probably warrants at least a note

Re: Linprocfs observation.

2000-03-27 Thread Warner Losh
In message <[EMAIL PROTECTED]> David Malone writes: : > File was removed because it was a huge, gaping security hole. It was : > effectively hard link to the file in question and circumvented some of : > the usual security protections that the file would otherwise be : > protected by. : : I know

Re: Linprocfs observation.

2000-03-27 Thread David Malone
> File was removed because it was a huge, gaping security hole. It was > effectively hard link to the file in question and circumvented some of > the usual security protections that the file would otherwise be > protected by. I know - AFAIK I was the one who reported it ;-) > : Linux itself is

Linprocfs observation.

2000-03-27 Thread David Malone
I notice that we've just gained a linprocfs which aims to provide a Linux style procfs for the Linux binary compatibility stuff. It looks quite neat, and provides lots of the odd files those linux programs go looking for. However... I haven't checked carefully, but I expect that the linprocfs cod