The comments about broken boards due with the hardware not wiring up byte enables. The problem in that case is a board limitation, not a 91c1111.

After you pointed out the mac needs to be unsigned, I went though and changed "all" mac references. Seems like there were a couple more than your patch had. The kernel drivers I have for this already have your fix, and treat mac's correctly through out.

Regards,

Richard W.

-----Original Message----- From: Stephan Linz [mailto:linz@mazet.de] Sent: Tuesday, November 04, 2003 12:45 PM To: Woodruff, Richard; 'wd@denx.de' Cc: u-boot-users@lists.sourceforge.net Subject: Re: [U-Boot-Users] [PATCH-1/2] LAN91C111

Hi,

I'm just back to NIOS.

Am Montag, 3. November 2003 19:04 schrieb Woodruff, Richard:

...

As to the 8bit to 16 access change, at both the boot and

kernel I use

...

16 bit accesses. I'm ok using 8bit access on my board, but

based on

...

comments in code and the ifdef's in the shipping kernel code, switching that access size will break a few boards. As word access works, I wouldn't switch that bit.

Oops, very interesting informations, because the SMSC's manuals for LAN91C110 and LAN91C1111 sayes about SMC91111_INT_REG: "... This register can be read and write as a word or two individual bytes ...". So I'm wondering about a few boards will be broken.

Nevertheless, let's leave it at 16bit access. The NIOS board is running in both cases, so there is no really important reason to change this bit.

...

As far as testing goes, if it works on my board I'm usually

happy. My

...

opinion would be if you are submitting new code, it doesn't really matter if it works for anyone else as long as it doesn't break the compile. If its existing code, then some more care is obviously necessary.

Yes of course, the 16bit to 8 change isn't really important. I've thought it could be a good minor code generalization. It wasn't my intention to make broken code.

So, here my second try. In attachment yo will find the patch again without this critical change.

...

From a group submission aspect, Wolfgang does a good job of

filtering

...

and moderating, and likely applies similar logic for

submission. He

...

is probably is a bit more stringent in certain areas as the

code would

...

become more difficult to maintain with out some control.

Oh yes, he does a good job for us.

Regards, Stephan

...

...

From: wd@denx.de [mailto:wd@denx.de] Sent: Monday, November 03, 2003 9:40 AM To: Stephan Linz Cc: Woodruff, Richard; u-boot-users@lists.sourceforge.net Subject: Re: [U-Boot-Users] [PATCH-1/2] LAN91C111

Dear Stephan,

in message 0311031553480Q.02205@pcj86 you wrote:

...

I'm curious about it. When your test phase / cross check

--8<--snipp-->8--