Re: [PATCHv2 1/2] cld: fix CLD_INODE_NAME_MAX woes

[Jeff Garzik <jeff@garzik.org>]

On 02/03/2010 08:45 AM, Colin McCabe wrote:
> When we create a static buffer for an inode name, and treat it like a
> null-terminated string, it needs to be of length CLD_INODE_NAME_MAX + 1 so
> that it can hold the NULL-terminator.
>
> In cldc_del and cldc_open, we should check that the user-submitted inode name
> is less than or equal to CLD_INODE_NAME_MAX. Formerly we were just checking
> that it wasn't too big to fit in the packet.
>
> When copying the inode name out of struct cld_dirent_cur, use snprintf rather
> than strcpy to ensure that we never overflow the buffer. This isn't strictly
> necessary if all other checks are working perfectly, but it seems prudent.
>
> Signed-off-by: Colin McCabe<cmccabe@alumni.cmu.edu>

applied, after s/snprintf/strncpy/

In general, too, you should never pass a variable string into snprintf, 
as that may make a program vulnerable to printf format string attacks 
(user supplies "%s" as a username, for example).

A few other changes made to your XDR work:

* "\n" removed from log messages, as that is appended as needed by log 
implementation

* user_key() restored.  that is our authentication hook, and it must be 
called, even though it merely returns the username passed to it at present.

* msg type renamed back to msg op

	Jeff