[Fwd: bogusness in the 21143]

[Bob]

Bob wrote:

> "bogusness in the 21143...I'm guessing the 98715A was
> the merger ("alignment") of the PNIC-II evolution with
> the 98715.", said Donald Becker.
>
> The Macronix 98715A must be what we should prefer now?
> How are the boards with that chip designated, as far as
> names and numbers?
>
> -Bob
>
>
> Donald Becker wrote:
>
>
>> On Sat, 1 Apr 2000, Bill Paul wrote:
>>
>> > >> I knew the chip was supposed to be compatible with Macronix's
>> > >> products. So, I e-mailed Macronix and asked them for info on
>> which
>> > >> of their chips was the same as Lite-On chip. They have told me
>> that
>> > >> the 98715 is an exact match.
>> >
>> > Not 100% exact, but pretty close.
>>
>> > > Donald Becker wrote.
>> > > While Macronix and Lite-On have a close relationship, the
>> PNIC-II is *not*
>> > > the same as the 98715. Beyond both being loosely Tulip clones,
>> they are not
>> > > even close. A quick look at the driver sections will confirm
>> this.
>> >
>> > I'm sorry, but this is wrong. The LC82C115 chip is very, *very*
>> similar
>> > to the 98715A.
>>
>> It similar mostly in the sense that they are both trying to be
>> work-alikes
>> to the 21143.  Their differences from the 21143 were both driven by
>> the
>> bogusness in the 21143.
>>
>> > The differences are that it supports magic packet/Wake on
>> > LAN and has a 128-bit multicast hash table instead of a 512-bit
>> table
>> > like most tulip clones.
>>
>> It points out that the chips are not the same.  Macronix now owns
>> the
>> Lite-On design, but the 98715 and the PNIC-II started from different
>> bases.
>> At least they both were trying to copy the good Digital
>> architecture, rather
>> than force an old 8390-like design into going faster.
>>
>> > It has the same 21143-alike NWAY support as
>> > the 98715A, the same register layout (except for the new ones
>> needed
>> > for the magic packet/WOL stuff) and the same rx filter programming
>> scheme
>> > (except for the smaller hash table). The PNIC II is basically the
>> 98715A
>> > with a few extra features (and one misfeature).
>>
>> The c115 document that I have uses the MXIC logo, but in much the
>> same way
>> that "Digital" was replaced by "Intel" in the 21143, without
>> changing the
>> contents.  It doesn't say 98715 anywhere in the text.  I'm guessing
>> the
>> 98715A was the merger ("alignment") of the PNIC-II evolution with
>> the
>> 98715.
>>
>> > > The PNIC-II datasheet is (or was at one time), proprietary. I
>> have no
>> > > rights to redistribute it.
>> >
>> > I got my copy from LinkSys when the asked me to do support for
>> their
>> > new card, and they never said anything about limiting
>> redistribution.
>> ...
>> > http://www.freebsd.org/~wpaul/Macronix/PNIC_II.PDF
>> > http://www.freebsd.org/~wpaul/PNIC/pnic.ps.gz
>> >
>> > The latter is the datasheet I have for the original PNIC. It's not
>>
>> > terribly spectacular. These datasheets have been sitting at these
>> > URLs for a while.
>>
>> The original PNIC datasheet evolution was amusing.  It started out
>> as a few
>> quick notes, and moved to increasing detail.  All the while keeping
>> the
>> Jan. 95 date.  (Uhmm, cred gap.)  You always needed a 21140 databook
>>
>> around to understand what they had intended to write.
>>
>> Donald Becker
>> Scyld Computing Corporation, becker@scyld.com
>>
>> -------------------------------------------------------------------
>> To unsubscribe send a message body containing "unsubscribe"
>> to linux-tulip-request@beowulf.org
>

-------------------------------------------------------------------
To unsubscribe send a message body containing "unsubscribe"
to linux-tulip-request@beowulf.org