Re: Call for axe(4) testers.

On Sat, Oct 04, 2008 at 08:14:33PM +0400, Boris Samorodov wrote:
 > Pyun YongHyeon <pyunyh_at_gmail.com> writes:
 > > On Fri, Oct 03, 2008 at 10:31:07PM +0400, Boris Samorodov wrote:
 > >  > Norikatsu Shigemura <nork_at_FreeBSD.org> writes:
 > >  > > On Fri, 03 Oct 2008 21:50:52 +0400
 > >  > > Boris Samorodov <bsam_at_ipt.ru> wrote:
 > >  > 
 > >  > >> The adapter is Trendnet TU2-ET100. It works fine without the patch.
 > >  > >> Let me know if you need some more info/testing.
 > >  > > 	Please show us 'devinfo -rv | fgrep -A2 axe'.
 > >  > >
 > >  > > 	In my case(Planex GU-1000T), I got following results:
 > >  > > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 > >  > >               axe0 pnpinfo vendor=0x0b95 product=0x1780 devclass=0xff devsubclass=0xff release=0x0001 sernum="" intclass=0xff intsubclass=0xff at port=4 interface=0
 > >  > >                 miibus0
 > >  > >                   ciphy0 pnpinfo oui=0x3f1 model=0xb rev=0x1 at phyno=24
 > >  > > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 > >  > 
 > >  > Here it is:
 > >  > -----
 > >  >               axe0 pnpinfo vendor=0x0b95 product=0x7720 devclass=0xff devsubclass=0xff release=0x0001 sernum="" intclass=0xff intsubclass=0xff at port=1 interface=0
 > >  >                 miibus0
 > >  >                   ukphy0 pnpinfo oui=0xec6 model=0x1 rev=0x1 at phyno=16
 > >
 > > Thanks for your testing!
 > >
 > > Would you repreat the command above without applying patch and show
 > > me the output?
 > 
 > They are identical.
 > 

OK.

 > > And try attached patch at the following URL.
 > > http://people.freebsd.org/~yongari/axe.88178.patch2
 > 
 > After applying the patch nothing changed except boot messages:
 > -----
 > diff -u dmesg.axe dmesg.axe.patch2
 > --- dmesg.axe	2008-10-03 20:58:06.000000000 +0400
 > +++ dmesg.axe.patch2	2008-10-04 19:51:27.000000000 +0400
 > _at__at_ -1,5 +1,6 _at__at_
 >  axe0: <vendor 0x0b95 product 0x7720, class 255/255, rev 2.00/0.01, addr 2> on uhub4
 >  axe0: AX88772, bufsz 4096, boundary 512
 > +axe0: PHYADDR 0xe0:0x10
          ^^^^^^^^^^^^^^^^^

This is expected one. Primary PHY address is at 16, secondary PHY
does not exist on this controller.

 >  miibus0: <MII bus> on axe0
 >  axe0: WARNING: using obsoleted if_watchdog interface
 >  axe0: WARNING: using obsoleted IFF_NEEDSGIANT flag
 > -----
 > 
 > > This patch just adds a couple of printfs to see what's going on in
 > > link state change handling. Normally you should see "LINK UP 10Mbps"
 > > or "LINK UP 100Mbps" message on console. If you just see "LINK DOWN"
 > > that indicates a problem that should be resolved.
 > 
 > But those words make me think you want me to plug/unplug the adapter
 > (which I didn't do previously -- I've been booting the system with
 > the plugged adapter). And I was very surprised when after plugging
 > the adapter back I saw the line "LINK UP 100Mbps"! And I've got
 > a working adapter!
 > 
 > Well, with your patches my adapter don't work only if it is *plugged*
 > while booting.
 > 

Ok, that indicates a problem in PHY driver but I'm not sure why
ukphy(4) couldn't report correct link state. The only guess I can
think of at this time is fake the extended capability bit of BMSR
register. Datasheet says that extended status register is reserved
one even though BMSR cleary indicates the existence of the
register. Please backout previous patch and try again the patch
in the following URL.

http://people.freebsd.org/~yongari/axe.88178.patch3

-- 
Regards,
Pyun YongHyeon