On Mon, 6 Nov 2000, Al B. Snell wrote:

> On Tue, 7 Nov 2000 itojun@iijlab.net wrote:
> 
> > 	just curious, how can you identify the existence of those LEDs?
> > 	are there any special I/O sequence to identify them?
> > 	if there's no way to identify those, it would be hard to probe/attach
> > 	drivers... (or we end up hardcoding it for ISP1100)
> 
> I'm looking into that. Intel helpfully supply the server with the
> technical specs on CD-ROM, down to the port access and connector pinout
> level - nice!
> 
> I hope there'll be something we can do to detect this baby; I need to
> analyse the PCI devices and stuff to see if it comes up in any kind of PnP
> registry - otherwise, it'll have to be an uncommentable line in the config
> file, I'm afraid :-(
> 
> # /dev/panel drivers
> 
> #inisppa0	at isa? port 0x...	# Intel ISP1100 front panel LEDs
> 
> #lptpa0		at lpt0			# Hardwired switches and LEDs on a
> 					# centronics port
> 
> #usbpa0		at uhid X		# Using a USB HID as a front panel
> 
> ...or something like that.

look for LED in src/sys/arch/sparc/sparc/auxreg.[ch] for another one...

To take this to extremes it would be nice to have a list of blink sources
(network activity, routing table lookups, harddisk access, page ins/outs
etc), that could be attached to an led. (useing the rndctl sources would be
a good start).

-- 
http://pointless.net/     : Another Great waste of bandwidth!
http://advogato.com/      : nothing to do with me, but worth a look.
"Ever since then I've wanted to live in a shipping container" - Lee.
2ECA0975     ::     45 52 65 E9 33 C7 72 4B  96 F4 44 2D 79 27 BF 08