Date: Tue, 10 Jul 2007 20:57:51 -0600 (MDT) From: "M. Warner Losh" <imp@bsdimp.com> To: ambrisko@ambrisko.com Cc: rpaulo@fnop.net, gurney_j@resnet.uoregon.edu, andre@freebsd.org, perforce@freebsd.org, cnst@freebsd.org, syrinx@freebsd.org, freebsd-arch@freebsd.org Subject: Re: Porting OpenBSD's sysctl hw.sensors framework to FreeBSD Message-ID: <20070710.205751.-1962670861.imp@bsdimp.com> In-Reply-To: <200707101833.l6AIX0xl049962@ambrisko.com> References: <20070708081511.GX1221@funkthat.com> <200707101833.l6AIX0xl049962@ambrisko.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message: <200707101833.l6AIX0xl049962@ambrisko.com>
Doug Ambrisko <ambrisko@ambrisko.com> writes:
: John-Mark Gurney writes:
: | Constantine A. Murenin wrote this message on Sat, Jul 07, 2007 at 21:43 -0400:
: | > Hardware sensors tree is going to be pretty deep down. Under sysctl(8)
: | > the variable names will look like this:
: | >
: | > hw.sensors.lm0.temp0
: | >
: | > whereas in reality, the tree has five levels:
: | >
: | > hw.sensors.lm0.temp.0
: |
: | I'm curious, why do we want/need these in the kernel as opposed to a
: | userland library/utility to provide this info?
:
: I agree. There are so many different flavours of HW monitoring chips
: and several tools that can read them live in ports. Lots of them are
: slightly different, intefaces can be i2c or direct I/O. We are already
: somewhat battling with the various ways IPMI controllers can be attached
: to the system. Now in the case of IPMI there is a good win in providing
: a device driver interface for the HW and user land tools to get info.
: out of them. I don't see a win with this in the various HW monitoring
: chips.
My big concern is the 'automatic' probing. It doesn't mix well with
i2c eeprom chips :-(
Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070710.205751.-1962670861.imp>