Re: For the love of God, is it even possible to make the Atheros ath.patch & updated HAL actually work?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ryan R wrote:
> *Sigh*
> 
> Hi again everyone..
> 
> I'm a novice BSD user who recently purchased an Atheros wireless card.  The
> card is an "Engenius EMP 8602", 6th Generation Atheros AR5006 a/b/g chipset.
> ( http://www.netgate.com/product_info.php?products_id=279 for product
> specs).
> 
> I have been racking my brain for DAYS trying to get Sam Leffler's (
> http://people.freebsd.org/~sam ) ath.patch and updated HAL binary to work in
> *ANY* version of FreeBSD.  First, I was told to do a MINIMAL install, then
> cvsup to RELENG_6 and attempt to apply the ath.patch.  The patch fails at
> least 7 hunks in various places in various files, 90% of the failed hunks
> are in if_ath.c.  I was then told to
> 
> # cd /usr/src/sys/contrib/dev
> # mv ath ath.original
> # tar -xzvf ath_hal_20051212.tgz
> # mv ath_hal_20051212 ath
> 
> to update the /usr/src/sys/contrib/dev directory.  Well lo and behold, the
> bloody thing wont compile.. It spits out all kinds of errors from if_ath.c
> about 'undeclared functions' and 'needing more parameters to function'
> 
> So I scrapped that installation and figured I'd give 7.0-CURRENT a try, in
> hopes that maybe the patch was FINALLY applied to the kernel for me, but
> nope..  It seems only tiny bits and pieces from the ath.patch have made it
> into the actual kernel source.  The stock kernel source does not have enough
> of the patch in it to make my wireless card work, and trying to apply the
> patch to the 7.0 kernel is a complete failure as well.
> 
> So my question is; if the patch doesn't apply cleanly to either RELENG_6 nor
> HEAD, then what the heck WILL it apply cleanly to?  This FreeBSD
> installation is sitting here doing absolutely nothing because it's a
> complete waste on my Laptop since there is NO wireless support for it at
> all.  Yet it works just fine in Linux..
> 
> 
> 
> Does anybody have even the slightest clue when this code may actually make
> it into the kernel sources?  Or better yet, does the patch cleanly apply and
> compile for ANYBODY?  I posted this on bsdforums.org and people there can't
> get it to compile either.. I e-mailed the author of the patch but I guess
> he's way too busy and didn't get a chance to respond to me
> 
> I'm at my wits end here so any help would be appreciated, I really don't
> want to have to just depend on Linux or Windows XP for this system when I've
> fallen in love with everything OTHER than this major problem in FreeBSD :(

Ryan,

If I recall the ath patch was pretty small. You could try applying the bits
manually.

I'm using DragonFly .. I think it took me about an hour to manually merge in the
changes. And it works for me now.

Having said that, the new hal isn't entirely well behaved on FreeBSD (or DragonFly).
Under load you'll see hw error interrupts and the occasional watchdog timeout.

Not-under-load, it works just fine.

Sam's aware of the problems, but has said that tracking down the root cause is
non-trivial because the vap changes to the madwifi-ng code obscure the probable
differences between it and the FreeBSD driver.

I was thinking it was init or dma timing related. Sam thinks it might be calibration.
I haven't been able to track it down (no time, and even if I had time it would still
be a considerable challenge), and Sam, I believe, is short of time himself these
days. I shouldn't speak for him actually. Sam ?

Andrew.

> 
> Thanks guys; anxiously awaiting any help
> -Ryan
> _______________________________________________
> freebsd-current_at_freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (FreeBSD)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFD4SzO8It2CaCdeMwRAnAmAJ9lq6DFgBpfSLpQ7maob8ex39gFBwCgmSzh
I1Gvsq5yE7wAGapl/oYrNek=
=BKPd
-----END PGP SIGNATURE-----