audit2 for -current

Poul-Henning Kamp phk at phk.freebsd.dk
Thu Mar 25 17:32:05 GMT 2004


In message <Pine.NEB.3.96L.1040325110554.53816D-100000 at fledge.watson.org>, Robe
rt Watson writes:

>I've CC'd
>Poul-Henning since he has his hands in the dev_t support in FreeBSD.

Back when I roto-tiled dev_t, I didn't want to change the inkernel type
because it meant hitting all the device drivers hard.  Today I am more
open to the idea, but still doubt it would carry concensus on the list.

We can either disband dev_t in the kernel for "struct cdev", or we
could introduce a ${something}dev_t typedef, I don't really have
preferences either way, but the cost of doing either of them is
device driver portability with other BSD kernels.

I'm not sure which of the two evils will be easier to support.

Anything passing a udev_t or for that matter dev_t from kernel to
userland on FreeBSD 5.x and later is in a state of sin because mostly
the major/minors are dynamic and therefore only marginally useful
as identifiers.

I'm not sure I know anything about the API you talk about, and certainly
not enough to know if it does something marginal or even fully stupid
with major/minor numbers.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

To Unsubscribe: send mail to majordomo at trustedbsd.org
with "unsubscribe trustedbsd-audit" in the body of the message



More information about the trustedbsd-audit mailing list