cvs commit: src/share/man/man4 Makefile rr232x.4 src/sys/dev/rr232x LICENSE README amd64-elf.rr232x_lib.o.uu array.h him.h himfuncs.h hptintf.h i386-elf.rr232x_lib.o.uu ldm.h list.h os_bsd.c os_bsd.h osm.h osm_bsd.c rr232x_config.c ...

Maxim Sobolev sobomax at FreeBSD.ORG
Tue May 2 03:37:22 UTC 2006


IMHO it is misleading, since FreeBSD users get used to the fact
that the driver only announces itself if actual hardware is present
in the system. With number of supported devices in GENERIC growing
up dmesg will eventually turn into mess if each device driver will
announce itself unconditionally.

At the very least, please put those printf's under if (bootverbose).

-Maxim

On Mon, May 01, 2006 at 08:32:37PM -0600, Scott Long wrote:
> Sam Leffler wrote:
> >Scott Long wrote:
> >
> >>Sam Leffler wrote:
> >>
> >>>Scott Long wrote:
> >>>
> >>>>The architecture of the driver makes this a request hard to do.  I 
> >>>>don't like it, but there is precedence already with the ath driver.
> >>>
> >>>
> >>>
> >>>I'm guessing you're referring to ath bailing if the hal could not be 
> >>>attached or the card otherwise setup?  If so the card was actually 
> >>>recognized and failure to complete the attach is totally separate.
> >>>
> >>>ath doesn't print anything during probe.
> >>>
> >>>    Sam
> >>
> >>
> >>I thought that the ath hal printed a line early in boot with the
> >>version number.
> >
> >
> >The hal is a separate module.  It prints it's version string on module 
> >load.  I can put it under bootverbose if desired but it's way useful to 
> >tell people to send me:
> >
> >dmesg|grep ath
> >
> >and get the hal version and mac+phy revs for the hardware.
> >
> >    Sam
> 
> The same is basically true of rr232x, though there is no module
> separation of HAL vs OSM like with ath.
> 
> Scott
> 


More information about the cvs-src mailing list