Dear Warner and Hackers,
--- "M. Warner Losh" <[EMAIL PROTECTED]> wrote:
> In message: <[EMAIL PROTECTED]>
> Maksim Yevmenkin <[EMAIL PROTECTED]> writes:
> : > :
> : > : Is there any way (on -current with NEWCARD) "devd" can
> : > : prevent "sio" driver from attaching to *ANY* pc-card
> : > : that has PCCARD_FUNCTION_SERIAL?
> : >
> : > Sure. Just have sio_pccard_match return -100. I've just committed
> : > the change to do this. No need to do anything else, I think.
> :
> : Nope :( It does not work. I applied patch to /sys/dev/sio/sio_pccard.c
> : and recompile my kernel with NEWCARD. It seems devd pays no attention
> : when i plug or unplug the 3COM card. I have attached dmesg output and
> : my devd.conf file. I was trying to get devd to kldload ng_bt3c module,
> : but it did not work. Am i missing something obvious here?
>
> Yes. You need to have ng_bt3c loaded before you insert the card.
> That's because of three reasons:
>
> 1) We don't detach a device when it 'won' the bidding on the device
> with a bid < 0 when a new driver is loaded.
> 2) There device is known, so devctl doesn't report anything to devd
> because it is known (it will report the sio attach).
> 3) devd ignores all unknown devices at the current time.
>
> I'm working on most of these issues, but not the 'rescan' issue.
Cool. It works if i kldload ng_bt3c before i insert the card.
However i could not get devd to execute proper "attach" commands
from the config file. It seems devd always wants to execute
/etc/devd-generic {start|stop} device
I took a quick look at the devd sources and could not find the
place where devd calls proper "attach" commands from the config
file. I saw few XXX comments in process_event() function and
almost convince myself that this is not done yet :) Anyway this
is not a big problem for me :)
thanks,
max
__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
