lkml.org 
[lkml]   [1998]   [Aug]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: DEVFSv50 and /dev/fb? (or /dev/fb/? ???)
David Luyer writes:
> > Richard you missed the whole point. Companies and clients read the FAQ
> > the run away screaming. All they see are this really "ugly" naming
> > scheme.
> >
> > --
> > Terry L. Ridder
> > Blue Danube Software (Blaue Donau Software)
> > "We do not write software, we compose it."
>
> Companies and clients use Solaris. I think that's all that needs to be said
> on that point.
>
> I haven't used devfs. However the two things I hope it would provide are...
> * dynamically-allocated PTYs beyond 256

Send me a patch.

> * fixed device names for SCSI devices independent of other SCSI devices
> in the system

Done ages ago.

> These two things would be really good, and if it does it in a less
> ugly way than Solaris devices (it's almost impossible not to...) I'd
> be happy. I'd also hope that it has the benefits I've listed below.
>
> I haven't looked at devfs. I don't know how good or bad it is.
> Personally, I think the solution should be a /proc/dev directory and
> people should be able to symlink /dev to /proc/dev if they so
> desire, or symlink only specific devices, or ignore it totally. If
> /dev is symlinked to /proc/dev, the proc filesystem needs to be able
> to support a /dev/log socket being created (and remaining until next
> reboot), and anything else userland does to /dev.

You can use devfs in this way too: just don't mount it at boot onto
/dev and configure it for full compatibility. In the early days while
I was developing devfs I mounted it onto /devfs.

> (just in case you haven't noticed how ugly Solaris devices are...)
> cos; ls -l /dev/sd0a
> lrwxrwxrwx 1 root root 12 Mar 1 1996 /dev/sd0a -> dsk/c0t3d0s0
> cos; ls -l /dev/dsk/c0t3d0s0
> lrwxrwxrwx 1 root root 51 Mar 1 1996 /dev/dsk/c0t3d0s0 -> ../../devices/sbus@1,f8000000/esp@0,800000/sd@3,0:a
> cos; ls -l /devices/sbus@1,f8000000/esp@0,800000/sd@3,0:a
> brw-r----- 1 root sys 32, 24 Mar 1 1996 /devices/sbus@1,f8000000/esp@0,800000/sd@3,0:a
>
> Now, if devfs is ugly compared to that, then yes, it doesn't really deserve
> to live.

% df
Filesystem 1024-blocks Used Available Capacity Mounted on
/dev/sd/c0b0t0u0p2 302379 217862 81444 73% /
/dev/ide/hd/c0b0t0u0p6
101925 10927 85859 11% /var
/dev/ide/hd/c0b0t0u0p5
101925 31998 64788 33% /tmp
/dev/sd/c0b0t1u0p3 102978 98182 3758 96% /home

% ls -lF /dev/sd
total 121
brw-r----- 1 root root 8, 0 Jan 1 1970 c0b0t0u0
brw-r----- 1 root root 8, 1 Jan 1 1970 c0b0t0u0p1
brw-r----- 1 root root 8, 2 Jan 1 1970 c0b0t0u0p2
brw-r----- 1 root root 8, 3 Jan 1 1970 c0b0t0u0p3
brw-r----- 1 root root 8, 4 Jan 1 1970 c0b0t0u0p4
brw-r----- 1 root root 8, 16 Jan 1 1970 c0b0t1u0
brw-r----- 1 root root 8, 17 Jan 1 1970 c0b0t1u0p1
brw-r----- 1 root root 8, 18 Jan 1 1970 c0b0t1u0p2
brw-r----- 1 root root 8, 19 Jan 1 1970 c0b0t1u0p3
brw-r----- 1 root root 8, 20 Jan 1 1970 c0b0t1u0p4

I trust that looks clean enough.

> However I believe there is room for a /proc/dev populated only with
> real devices, with four major advantages;
> * you can instantly see what is there
> * faster directory searches
> * makes things a lot easier for builders of distributions
> * makes it a lot easier (for the user) when adding a new device/driver

Devfs does that for you. You can mount it somewhere else if you like.

Regards,

Richard....

-
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.altern.org/andrebalsa/doc/lkml-faq.html

\
 
 \ /
  Last update: 2005-03-22 13:43    [W:0.189 / U:1.756 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site