[lkml]   [1999]   [Jun]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Migrating to larger numbers
In article <>,
H. Peter Anvin <> wrote:

>devfs is based on a completely bogus idea: that a device driver is a
>kernel thing.


A device driver advertising the devices it supports doesn't require
it to be a kernel thing. The current interface may be restricted
to kernel-space device drivers, but that can be corrected by
publishing the appropriate interfaces and doing kernel-space glue
so that user-mode device drivers (or whatever) can sensibly advertise
the devices they're servicing.

>*THIS* is the supreme fallacy, and a major pitfall for
>Linux today. devfs in fact helps perpetualize this problem, in large
>part by moving policy into the kernel that has no business being there.

Which is worse; the current secret policy where the device drivers
have interface information that they don't bother to tell you
(unless you dig through /proc/devices) or a policy where a
namespace is exported? In either case, you can link your own
naming convention to the actual devices (conventionally via
mknod, which is symbolic links for devices, or with real symbolic
links for devfs), but in the second case the device drivers are
telling you this information instead of you hoping you've got
the correct majors and minors.

david parsons \bi/ A great fan of having the kernel tell all.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:52    [W:0.075 / U:0.740 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site