Messages in this thread | | | Date | Fri, 18 Jun 1999 03:17:29 -0700 | From | "H. Peter Anvin" <> | Subject | Re: UUIDs (and devfs and major/minor numbers) |
| |
Alan Cox wrote: > > This was Richard's argument. It is true you can do devfs in user space. It > is also true (as Richard pointed out) that the user space solution takes > about as much unswappable memory, and requires a pile of new kernel files > and also as I realised when I had a hack on this it needs select() and the > like on /proc files >
I disagree with the claim that it requires a pile of new kernel files. It basically means you have started in the wrong end of the problem. There is a need for Linux do develop a decent device driver management; unfortunately devfs is a hack to patch it rather than tacking the admittedly hard problem.
> > You can do it in userspace, but once you are handling swap its on the wrong > side of ugly to do so. > > What gets me is there are people saying khttpd is a good idea, when its > reporting lower performance than a user space httpd and who are anti-devfs >
I don't necessarily think it's a good idea, but unlike devfs it isn't an all-pervasive thing and wouldn't be running on most -- if any -- systems. Consequently I don't particularly care one way or the other; I can imagine a very small set of machines for which it could make sense.
Similarly, I don't care if Richard Gooch runs devfs on his machine. However, for devfs to make sense, it has to be everywhere, and I do object to *that*.
-hpa
-- "The user's computer downloads the ActiveX code and simulates a 'Blue Screen' crash, a generally benign event most users are familiar with and that would not necessarily arouse suspicions." -- Security exploit description on http://www.zks.net/p3/how.asp
- 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/
| |