Subject: None
To: None <jonathan@dsg.stanford.edu>
From: Steven M. Bellovin <smb@cs.columbia.edu>
List: tech-kern
Date: 01/04/2006 20:57:16
In message <E1EuJYs-0008Qp-00@smeg.dsg.stanford.edu>, jonathan@dsg.stanford.edu
 writes:
>In message <20060105003422.22C5B3C01E3@berkshire.machshav.com>,
>"Steven M. Bellovin" writes:
>
>>In message <43BC5508.3040709@anastigmatix.net>, Chapman Flack writes:
>>>matthew green wrote:
>>>> EPERM?
>>>
>>>Not the way I read it ... EPERM is specifically about whether
>>>the open would be allowed by the checks on your identity and
>>>the mode bits on the node.  But you can have all the permissions
>>>in the world and when you try to open /dev/cardpunch for reading,
>>>something's still gotta give. It's an inherent device capability
>>>issue, not a permission issue.
>>>
>>What do you get if you open a device on a read-only file system for 
>>writing?   [...]
>
>Err, EROFS, surely? Was this a trick question or a rhetorical question?
>
A question that came from not reading errno(2) first....

		--Steven M. Bellovin, http://www.cs.columbia.edu/~smb