Re: sleep mode bit causes Pro/100 V 82559 to transmit timeout at 10mbps

Donald Becker <[email protected]> Fri, 19 Jul 2002 06:33:24 -0400 (EDT)
Newsgroups gmane.linux.drivers.eepro100.user
Message-ID <[email protected]>
On Thu, 18 Jul 2002, Kallol Biswas wrote:

> Date: Thu, 18 Jul 2002 15:57:14 -0700
> From: Kallol Biswas <[email protected]>
> To: [email protected], [email protected]
> Subject: [eepro100-bug] sleep mode bit causes Pro/100 V 82559 to transmit
>     timeout at 10mbps
> 
> We have an embedded board  that uses 82801BA chips.
> (http://www.intel.com/design/chipsets/specupdt/29824217.pdf)
> The Pro/100 V embedded (0x2449) controller would experience
> transmit time out at 10Mbps. Turning off  the sleep mode bit (word 10
> bit 1 of eeprom)
> fixed the problem.

The various eepro100 driver releases should soon all have code
similar to
	if (eeprom[10] & 0x0002)
		printk(KERN_INFO "\n" KERN_INFO "  ** The configuration "
			   "EEPROM enables Sleep Mode.\n" KERN_INFO "\n"
			   "  ** This will cause PCI bus errors!\n"
			   KERN_INFO "  ** Update the configuration EEPROM "
			   "with the eepro100-diag program.\n"  );

and the July 2002 release of  eepro100-diag  now has a long usage
message that should make it possible to avoid reading the manual page.

From the CVS log:
----------------------------
revision 2.8
date: 2002/07/15 17:14:06;  author: becker;  state: Exp;  lines: +111 -30
eepro100-diag.c:v2.09 7/15/2002

Added a extended usage message, and tweaked the long option names.

Added flash support and changed MII support to use weak linker symbols.
  The default warning message is shown when the library is not avaiable.
----------------------------


-- 
Donald Becker				[email protected]
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993