On Fri, 21 Jul 2000, Andre Hedrick wrote:
> I am talking about attempting to invoke unknown vender specific commands
> They do comply with the SPEC but are not part of the SPEC.
> Since I do not have the priviledge of knowing these facts, but know they
> exist. You can not allow a rouge driver attempt to invoke these commands.
>
> This is how disks are/can be damaged.
>
> Controllers are limited to what the driver tells it to do.
>
> This is why it is a kernel level issue, I am taking responsiblity for not
> fixing this earlier but I just figured it out have being wacked in the
> head with it three months ago.
The problem, as I understand it, is that the patch isn't a fix so
much as it's a hack. Perhaps I've gotten the wrong impression, but I've
not seen many reasoned explanations about what the proposed patch does.
I say it sounds like a hack because it disables a valid part of
the SPEC because of concern that it could possibly be used
incorrectly/improperly. One suggestion that I've heard so far is that
CAPs be used and having things by default give up those CAPs on boot. I
suspect anyone as root could re-aquire the CAPs, but then, they could do
the same thing in other ways once they have root anyway. This would
prevent accidental destruction via this method though (I believe). Of
course, you could accidentally perform the same operation in other ways
as well given the moon is in the right phase.
Another thought, I would think, would be to place a CAP such that
instead of the entire method being protected just the specific calls
against specific drives require a CAP. Of course, this would require
actually having imformation it sounds like we don't have currently...
Now, if a drive can be written to in such a way that this
information is corrupted, one would *tend* to think that if the *correct*
information is written back to that same area you might be able to
recover the disk. I have no clue as to if this is actually the case or
not or if there is a way this could be done, but it seemed like a
reasonable question to bring up: Can a drive be fixed as it was broken?
Stephen
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jul 23 2000 - 21:00:17 EST