Re: [patch 00/13] devtmpfs patches

[Eric W. Biederman]

Kay Sievers <kay.sievers@xxxxxxxx> writes:

> On Mon, May 11, 2009 at 18:40, Eric W. Biederman <ebiederm@xxxxxxxxxxxx> wrote:
>> The goal for kernel compile options is that they do not affect the
>> kernels behavior. ÂThe behavior in devmtmpfs clearly does not match
>> that rule. ÂThe kernel acts very different with it compiled in
>> and with it not compiled in. So that section of the code deserves.
>
> It's the same as with any other option, like the ones that enable
> dynamic minors.

No it's not.

The practical goal is that a distribution can enable essentially every feature
in the kernel and not be forced to use one.

The most similar example I can think of is the dhcp client in the kernel.
Even when compiled in only if you enable it on the command line (aka ip=dhcp)
does it enable and attempt to network boot.

Mouting on /dev might be sane if it was enabled by a kernel command line option.
By default it is wrong.

Dynamic minors is right on that hairy edge.  I'm puzzled know why we
can't have the first N devices use the static assignment and the rest
of the devices use the dynamic minors.

Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/