Skip site navigation (1)Skip section navigation (2) 
Date:      Wed, 2 Sep 1998 15:30:43 -0400 (EDT)
From:      Bill Paul <wpaul@skynet.ctr.columbia.edu>
To:        vance@usq.edu.au (Christopher JS Vance)
Cc:        stable@FreeBSD.ORG
Subject:   Re: Looking for feedback on xl (3c905/3c905B) driver
Message-ID:  <199809021930.PAA08981@skynet.ctr.columbia.edu>
In-Reply-To: <19980902042906.506.qmail@usq.edu.au> from "Christopher JS Vance" at Sep 2, 98 02:29:06 pm
next in thread | previous in thread | raw e-mail | index | archive | help 

Of all the gin joints in all the towns in all the world, Christopher JS 
Vance had to walk into mine and say:

> Bill Paul wrote:
> | Okay people, it's been some time since I put the xl0 driver in 
> | -current
> | and -stable. I need to know if people are having problems with the 
> | driver
> | or not. So far I've gotten startlingly little feedback. This either 
> | means
> | that it's working so well that nobody has had any trouble with it, or
> | you guys are all just slackers and are too lazy to speak up.
> 
> I get a kernel panic.  I didn't actually try the new kernel till
> several days after the discussion started...

*sigh*
 
> | - What kind of adapter you have. Note that a 3c905-TX is NOT THE SAME
> |   as a 3c905B-TX. (See the 'B'? It makes a difference.) Double and 
> | triple
> |   check the model. Don't tell me you have a 3c905B when you actually 
> | only
> |   have a 3c905. If you want to be really nice, do 'dmesg | grep xl' 
> | and
> |   show me the output.
> 
> My handwritten copy of what the 2.2.7-STABLE kernel (with -dv) gives
> on reboot just before the panic is:
> 
> xl0 <3Com 3c900 Etherlink XL 10BaseT Combo> rev 0 int a irq 11 on 
> pci0:9:0
> 	mapreg[10] type=1 addr=0000e000 size=0040.
> xl0: Ethernet address: 00:60:97:a4:18:de
> xl0: media options word: e138
> xl0: found 10BaseT
> xl0: found AUI
> xl0: found BNC
> xl0: unknown XCVR type: 14
> ifmedia_set: no match for 0x100026/0xffffffff
> panic: ifmedia_set

This is not normal: 14 is not a proper transceiver selection value
as far as I can tell. Why your EEPROM has this value programmed into
it I can't say, but it's not right. Grr. Looks like another oddball
condition I have to test for. Does it always report the same value
(14) or does it change from one reboot to the next?

I'd like you try a small patch for me. This will modify the xl_mediacheck()
routine to also test the XCVR value for sanity and choose a proper default
based on the card type.

This should at least get the system booted without a panic. I think
in your case it will default to 10baseT as the media; you'll need to
use ifconfig xl0 media <foo> to change it, where <foo> is one of
10baseT/UTP, 10base5/AUI or 10base2/BNC (assuming you have a COMBO
card, otherwise 10baseT will be right). Please let me know if this
works.

To apply this patch:

- Save this e-mail message to /tmp/xl.diff
- cd /sys/pci
- patch < /tmp/xl.diff
- make a new kernel and boot it

-Bill

-- 
=============================================================================
-Bill Paul            (212) 854-6020 | System Manager, Master of Unix-Fu
Work:         wpaul@ctr.columbia.edu | Center for Telecommunications Research
Home:  wpaul@skynet.ctr.columbia.edu | Columbia University, New York City
=============================================================================
 "It is not I who am crazy; it is I who am mad!" - Ren Hoek, "Space Madness"
=============================================================================

*** if_xl.c	1998/08/31 15:13:27	1.37
--- if_xl.c	1998/09/02 19:13:53
***************
*** 1152,1164 ****
  	 */
  	if (sc->xl_media & (XL_MEDIAOPT_MASK & ~XL_MEDIAOPT_VCO))
  		return;
  
! 	printf("xl%d: WARNING: no media options bits set in "
! 		"the media options register!!\n", sc->xl_unit);
! 	printf("xl%d: this could be a manufacturing defect in "
! 		"your adapter or system\n", sc->xl_unit);
! 	printf("xl%d: attempting to guess media type; you "
! 		"should probably consult your vendor\n", sc->xl_unit);
  
  	/*
  	 * Read the device ID from the EEPROM.
--- 1152,1178 ----
  	 */
  	if (sc->xl_media & (XL_MEDIAOPT_MASK & ~XL_MEDIAOPT_VCO))
  		return;
+ 	else {
+ 		printf("xl%d: WARNING: no media options bits set in "
+ 			"the media options register!!\n", sc->xl_unit);
+ 		printf("xl%d: this could be a manufacturing defect in "
+ 			"your adapter or system\n", sc->xl_unit);
+ 		printf("xl%d: attempting to guess media type; you "
+ 			"should probably consult your vendor\n", sc->xl_unit);
+ 	}
  
! 	/*
! 	 * Check the XCVR value. If it's not in the normal range
! 	 * of values, we need to fake it up here.
! 	 */
! 	if (sc->xl_xcvr <= XL_XCVR_AUTO)
! 		return;
! 	else {
! 		printf("xl%d: warning: bogus xcvr value in EEPROM (%x)\n",
! 					sc->xl_unit, sc->xl_xcvr);
! 		printf("xl%d: choosing new default based on card type\n",
! 					sc->xl_unit);
! 	}
  
  	/*
  	 * Read the device ID from the EEPROM.
***************
*** 1171,1194 ****
--- 1185,1213 ----
  	case TC_DEVICEID_BOOMERANG_10BT:	/* 3c900-TP */
  	case TC_DEVICEID_CYCLONE_10BT:		/* 3c905B-TP */
  		sc->xl_media = XL_MEDIAOPT_BT;
+ 		sc->xl_xcvr = XL_XCVR_10BT;
  		printf("xl%d: guessing 10BaseT transceiver\n", sc->xl_unit);
  		break;
  	case TC_DEVICEID_BOOMERANG_10BT_COMBO:	/* 3c900-COMBO */
  	case TC_DEVICEID_CYCLONE_10BT_COMBO:	/* 3c905B-COMBO */
  		sc->xl_media = XL_MEDIAOPT_BT|XL_MEDIAOPT_BNC|XL_MEDIAOPT_AUI;
+ 		sc->xl_xcvr = XL_XCVR_10BT;
  		printf("xl%d: guessing COMBO (AUI/BNC/TP)\n", sc->xl_unit);
  		break;
  	case TC_DEVICEID_BOOMERANG_10_100BT:	/* 3c905-TX */
  		sc->xl_media = XL_MEDIAOPT_MII;
+ 		sc->xl_xcvr = XL_XCVR_MII;
  		printf("xl%d: guessing MII\n", sc->xl_unit);
  		break;
  	case TC_DEVICEID_BOOMERANG_100BT4:	/* 3c905-T4 */
  	case TC_DEVICEID_CYCLONE_10_100BT4:	/* 3c905B-T4 */
  		sc->xl_media = XL_MEDIAOPT_BT4;
+ 		sc->xl_xcvr = XL_XCVR_MII;
  		printf("xl%d: guessing 100BaseT4/MII\n", sc->xl_unit);
  		break;
  	case TC_DEVICEID_CYCLONE_10_100BT:	/* 3c905B-TX */
  		sc->xl_media = XL_MEDIAOPT_BTX;
+ 		sc->xl_xcvr = XL_XCVR_AUTO;
  		printf("xl%d: guessing 10/100 internal\n", sc->xl_unit);
  		break;
  	default:
***************
*** 1388,1398 ****
  		printf("xl%d: media options word: %x\n", sc->xl_unit,
  							 sc->xl_media);
  
- 	xl_mediacheck(sc);
- 
  	xl_read_eeprom(sc, (char *)&sc->xl_xcvr, XL_EE_ICFG_0, 2, 0);
  	sc->xl_xcvr &= XL_ICFG_CONNECTOR_MASK;
  	sc->xl_xcvr >>= XL_ICFG_CONNECTOR_BITS;
  	if (sc->xl_media & XL_MEDIAOPT_MII || sc->xl_media & XL_MEDIAOPT_BTX
  			|| sc->xl_media & XL_MEDIAOPT_BT4) {
  		/*
--- 1407,1418 ----
  		printf("xl%d: media options word: %x\n", sc->xl_unit,
  							 sc->xl_media);
  
  	xl_read_eeprom(sc, (char *)&sc->xl_xcvr, XL_EE_ICFG_0, 2, 0);
  	sc->xl_xcvr &= XL_ICFG_CONNECTOR_MASK;
  	sc->xl_xcvr >>= XL_ICFG_CONNECTOR_BITS;
+ 
+ 	xl_mediacheck(sc);
+ 
  	if (sc->xl_media & XL_MEDIAOPT_MII || sc->xl_media & XL_MEDIAOPT_BTX
  			|| sc->xl_media & XL_MEDIAOPT_BT4) {
  		/*

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message

Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199809021930.PAA08981>