Date: Tue, 30 Apr 2002 16:38:14 -0500 (CDT) From: John Utz <john@utzweb.net> To: freebsd-mobile@freebsd.org Subject: is there a technical reason for apm or pnp to NOT be klds? Message-ID: <Pine.LNX.4.44.0204301608460.30583-100000@jupiter.linuxengine.net>
next in thread | raw e-mail | index | archive | help
hi;
i might resend this to -arch, but i'll start here :-)
i dont think i'll bother with -hardware because that seems to be a black
hole...
The Story So Far:
i've been looking at the tanimura newmidi code most intently lately.
strange as this may seem, that lead me to:
1. look at pnpinfo,
2. discover that it appears to be *broken*; it always sez that it cant
find any pnp devices even tho the compiled in 'options PNPBIOS' code
finds plenty. this code finds and boots the ESS ISA AudioEngine
OEM chip without fail.
3. decide that fixing pnpinfo
a. is virtuous :-)
b. moves forward my midi agenda by hopefully allowing the removal of
some device detection code that is in the newmidi stuff.
4. conclude that the pnp code /usr/src/sys/isa/{pnp,pnpparse}.c should be
reimplemented as a KLD.
5. decide to find another good example to work from
6. pick apm as a good example because it behaves in a way that i'd like to
see pnpinfo behave...
7. discover that apm isnt a KLD either. :-(
8. wonder why apm isnt a KLD...
9. decide to ask list.
so...'sup with that? is the only reason that apm isnt a kld is because
nobody's done it? or does it have tendrils into too many things
(apm_ioctrl() is a significant chunk of code ) for this to be possble or
practical?
likewise, is there anything in the way of making pnp a KLD? i am betting
not, because there isnt a device for it in LINT or /dev, the only way it
shows up is if you put options PNPBIOS or something in your config file
straying further into vapourous work that i havent done yet... it seems to
me that the program name 'pnpinfo' should be deprecated and the program
should be called pnpconf ( like pciconf ) and follow pciconf's semantics
as much as possible. pnpinfo would simply become a wrapper around 'pnpconf
-vl' or something.
and, taken to absurdity, would there be a perf hit if *everything* execpt
malloc() and free() (showing my WinNT/WinCE professional experience
here :-) ) where implemented as a KLD?
tnx for listening! feedback appreciated...
johnu
--
John L. Utz III
john@utzweb.net
Idiocy is the Impulse Function in the Convolution of Life
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-mobile" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.44.0204301608460.30583-100000>