Re: The guts of /dev/* ?

From: Jean-David Beyer (jdbeyer_at_exit109.com)
Date: 12/05/04 

Date: Sat, 04 Dec 2004 21:25:21 -0500

Bill Unruh wrote:
> Robert Heller <heller@deepsoft.com> writes:
>
> ] not@top-post,
> ] In a message on Sat, 04 Dec 2004 10:54:05 -0600, wrote :
>
> ]n> Where would I find text explaining the 'innards' of /dev/* ?
> ]n> Not at a 'salesClerk level': "try xyz [it worked for me]. ", but
> ]n> down at asm, interrupts ...etc. level.
>
> Read the source code. I highly doubt that this is what you want.
> The files in /dev/are links to the device drivers. The input and output
> depend on what the specific device driver will do.
> cat /dev/dsp reads from teh oss audio device, and pushes the stuff out on
> stdout.
> Exactly what is read depends on how the device driver happens to have been
> set by the ioctls for that device. (ioctls are commands sent to the device
> driver with routines which deliver the ioctl codes to the driver).
>
>
While you are not wrong, I would never call them _links_, since I use that
term mainly for the things the _ln_ command creates. I think of them as an
index to a two-dimensional table inside the kernel that the kernel uses to
call the proper driver. (I do not mean to imply that it is actually
implemented that way, though it might be: I never looked.) 

-- 
   .~.  Jean-David Beyer           Registered Linux User 85642.
   /V\                             Registered Machine   241939.
  /( )\ Shrewsbury, New Jersey     http://counter.li.org
  ^^-^^ 21:19:59 up 2 days, 1 min, 3 users, load average: 4.45, 4.29, 4.19

Relevant Pages

  • Re: i386: pata_cs5520 does not work
    ... Device driver platform lacks bus and class support for being resumed. ... Device driver pci0000:00 lacks bus and class support for being resumed. ...
    (Linux-Kernel)
  • Re: CE6.0 Exposing Driver Interfaces to an Application
    ... complicated mechanism is to create a simple stream driver and expose your ... for *kernel* ioctls, not device driver ioctls. ... components can call device driver ioctls though. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: 1394 Virtual Device Driver
    ... Try two machines connected via 1394, one with a virtual device ... > enumerates the bus successfully including the virtual device driver. ...
    (microsoft.public.development.device.drivers)
  • Re: The Kernal Is A Huge Security Whole In Windows
    ... The easiest way to exploit the security hole I am discussing is obviously ... driver writer is ridiculous, I'm sorry. ... responsible for making his device driver secure? ... > Performance Tab> View> Show Kernel Times. ...
    (microsoft.public.win2000.security)
  • Re: hard drive scrubbing utility
    ... we only teach the basics relative to PCI-style bus devices. ... The type of device driver you are writing makes a lot of difference also. ... There are some new books coming out from Microsoft Press on the new Kernel Mode Driver ...
    (microsoft.public.vc.mfc)