[oi-dev] atge driver issue

[Carsten Grzemba]

Am 01.07.21 21:57 schrieb Gary Mills  <gary_mills at fastmail.fm>: 
> 
> On Thu, Jul 01, 2021 at 06:27:55PM +0200, Carsten Grzemba via oi-dev wrote:
> > 
> > I use the atge driver on my laptop with AR8151V2 chip. There I have the
> > problem if the cable disconnects 10 sec later I get a L1C error and as
> > a result it seems that the upper layer assumes that the link is up with
> > 10MB half duplex.
> [...]
> > But it seems there is very less similarity of the illumos driver with
> > the BSD or Linux driver.
> > 
> > So I am wondering which source was the blue print for the Illumos atge
> > driver?
> 
> I started with the Freebsd alc driver, and wound up with a new version
> of the illumos atge driver. This was in 2010. Much has changed since
> then. At the time, I tested the AR8132 on my ASUS laptop. Masayuki
> Murayama tested the AR8131 on his laptop. This was my first driver.
> I had a great deal of help from other people. There was one report in
> 2012 of a driver failure with AR8151 v2.0 hardware.
> 
> 
> -- 
> -Gary Mills- -refurb- -Winnipeg, Manitoba, Canada-
> 

 
It seems the fix is simple: 

 
--- a/usr/src/uts/common/io/atge/atge_main.c
+++ b/usr/src/uts/common/io/atge/atge_main.c
@@ -256,7 +256,7 @@ static mii_ops_t atge_l1c_mii_ops = {
 atge_l1c_mii_read,
 atge_l1c_mii_write,
 atge_mii_notify,
- NULL
+ atge_l1c_mii_reset
 };
 
the function  atge_l1c_mii_reset exist already but is not registered in the mii_ops table.  
With this the "atge0: L1C chip detected a fatal error, interrupt status: 2200" did not occur again on cable disconnect.  
The interface is working if the cable is connected again.
  
 
-- 
Carsten
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openindiana.org/pipermail/oi-dev/attachments/20210702/ca1458d0/attachment.html>