Linux MIPS Architecture development
 help / color / mirror / Atom feed
From: Manish Lachwani <mlachwani@mvista.com>
To: Ralf Baechle <ralf@linux-mips.org>
Cc: TheNop <TheNop@gmx.net>, linux-mips@linux-mips.org
Subject: Re: Titan ethernet driver broken
Date: Sat, 20 Nov 2004 23:26:03 -0800	[thread overview]
Message-ID: <41A0430B.6080201@mvista.com> (raw)
In-Reply-To: <20041120095445.GA12870@linux-mips.org>

Ralf Baechle wrote:
> On Thu, Nov 18, 2004 at 11:43:51PM +0100, TheNop wrote:
> 
> 
>>I use the chip version 1.1.
>>Now I have the problem, that I can not use the newest code,  until 1.2 
>>version of the chip is available.
>>Is it possible to make the code usable for all chip version by choosing 
>>the version at the kernel configuration?
> 
> 
> Titan 1.2 is available since quite a while - the dust on my board is
> proof ;-)  Since Titan 1.0 and 1.0 were shipped in very low numbers to
> early customers only and will never be available in volume production the
> support for them was removed.  As I recall there were at least these
> problems with Titan 1.0 and 1.1 in Linux:
> 
>   - Linux uses the prefetch prepare for store operation.
>   - Coherency mode 5 which is mandatory for good performance and any kind
>     of sanity on SMP is now being used.
>   - The problem with the third ethernet port which Manish just had
>     described.
> 
> You can dig through XCVS, WebCVS or the linux-cvs archive to find where
> I broke backward compatibility.
> 
>   Ralf
> 
Hello !

Ralf, thanks for the good description. Anyways, just to make it a little 
more clear. Port #2 (third ethernet port) was not working on Titan 1.0 
and 1.1. This is because there is no interrupt line on which interrupts 
for port #2 could be routed. And Titan MACs cannot share interrupts.

One other problem with 1.0 and 1.1 was that for incoming packets, the IP 
header is not aligned. As a result, an extra copy was implemented in the 
driver to align this IP header. This problem exists on all ports (0 and 1).

With Titan 1.2, there is no need for the extra copy and port #2 can be 
enabled since there is an interrupt line which it can use.

And of course, 1.0 and 1.1 did not support the five state MOESI 
protocol. They supported MEI only.

Thanks
Manish Lachwani

  reply	other threads:[~2004-11-21  7:26 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-11-18 20:19 Titan ethernet driver broken TheNop
2004-11-18 20:23 ` Manish Lachwani
2004-11-18 21:41   ` TheNop
2004-11-18 21:42     ` Manish Lachwani
2004-11-18 21:54       ` TheNop
2004-11-18 22:17         ` TheNop
2004-11-18 22:23           ` Manish Lachwani
2004-11-18 22:43             ` TheNop
2004-11-18 23:43               ` Manish Lachwani
2004-11-20  9:54               ` Ralf Baechle
2004-11-21  7:26                 ` Manish Lachwani [this message]
2004-11-19 17:32           ` Manish Lachwani

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=41A0430B.6080201@mvista.com \
    --to=mlachwani@mvista.com \
    --cc=TheNop@gmx.net \
    --cc=linux-mips@linux-mips.org \
    --cc=ralf@linux-mips.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox