Re: 3com Driver and the 3XP Processor

Re: 3com Driver and the 3XP Processor

[Brent D. Norris]

Thanks to Kip for clarifying why Linux doesn't use this feature, but now I
wonder why you beleive this?


--hahn@coffee.psychology.mcmaster.ca-- wrote:

> > hype it does DES without using the CPU, does linux take advantage of that?
>
> no, as far as I've heard.  and I wouldn't expect it to either.
> further, it's highly questionable whether the feature makes sense...

If they release the API for it why would you not expect it to use it, and
why would getting a complex computation and encryption off of the CPU to a
more specialized hardware not make sense?

Brent Norris

Executive Advisor -- WKU-Linux

3com Driver and the 3XP Processor

[Brent D. Norris]

I just had one of the "3com Etherlink 10/100 PCI NIC with 3XP processor"
float accross my desk, I was wondering how much the linux kernel uses the
3xp processor for its encryption offloading and such.  According to the
hype it does DES without using the CPU, does linux take advantage of that?

Brent Norris

Executive Advisor -- WKU-Linux

System Administrator -- WKU-Center for Biodiversity
                        Best Mechanical

W: 270-745-8864
H: 270-563-9226

"The problem with the Linux learning curve is that it is _so_ steep once
 at the top you can't see the people at the bottom"  --Doug Hagan

Re: 3com Driver and the 3XP Processor

[Kip Macy]

It can't because 3com hasn't implemented in the driver and they won't
publish the interface.
		-Kip

On Mon, 11 Jun 2001, Brent D. Norris wrote:

> I just had one of the "3com Etherlink 10/100 PCI NIC with 3XP processor"
> float accross my desk, I was wondering how much the linux kernel uses the
> 3xp processor for its encryption offloading and such.  According to the
> hype it does DES without using the CPU, does linux take advantage of that?
> 
> Brent Norris
> 
> Executive Advisor -- WKU-Linux
> 
> System Administrator -- WKU-Center for Biodiversity
>                         Best Mechanical
> 
> W: 270-745-8864
> H: 270-563-9226
> 
> "The problem with the Linux learning curve is that it is _so_ steep once
>  at the top you can't see the people at the bottom"  --Doug Hagan
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: 3com Driver and the 3XP Processor

[Brent D. Norris]

I thought 3com was pretty friendly to the Linux Community, was that a
misconception?

> It can't because 3com hasn't implemented in the driver and they won't
> publish the interface.
> 		-Kip
>

Brent

Executive Advisor -- WKU-Linux

Re: 3com Driver and the 3XP Processor

[Kip Macy]

I think that they are relatively friendly. However, if they publish the
interface to their card another company could come along with a card
with the same functionality and take advantage of pre-existing drivers and
undercut their price, thus taking away their margins. At least that is the
rationale I have been given and this has occurred on at least one
occasion to Adaptec. 

My opinion is that if you have to obscure your interface to protect your
margins because you are making a commodity component then you are in the
wrong business. Nonetheless they can correctly point out that they are
still making a lot more money than I am :-).

		-Kip

On Mon, 11 Jun 2001, Brent D. Norris wrote:

> I thought 3com was pretty friendly to the Linux Community, was that a
> misconception?
> 
> > It can't because 3com hasn't implemented in the driver and they won't
> > publish the interface.
> > 		-Kip
> >
> 
> Brent
> 
> Executive Advisor -- WKU-Linux
> 
> 

Re: 3com Driver and the 3XP Processor

[Alan Cox]

> My opinion is that if you have to obscure your interface to protect your
> margins because you are making a commodity component then you are in the
> wrong business. Nonetheless they can correctly point out that they are

Cryto hardware is commodity. In fact its questionable it has any value below
1Gbit/second anyway because the cheapest low speed crypto coprocessors are
made by AMD and intel. They fit into the second socket on your dual cpu
motherboard and as well as being mass market are conveinently reprogrammable
and able to run your applications when not doing crypto. The cheapest raid
accelerator is the same story. It costs a lot of money to build custom
hardwae, an x86 is the wrong solution but its sufficiently large a hamemr
that it works better than the elegant approach for most cases.

> still making a lot more money than I am :-).

That I doubt looking at recent financial reports

But I'd be inclined to ask 3com 

   What other large company helped design the interface and owns some of the IP
   Does that other company have an interest in the OS business.

I don't know for sure but I suspect that would give a most interesting answer

Re: 3com Driver and the 3XP Processor

[Pavel Machek]

Hi!

> I just had one of the "3com Etherlink 10/100 PCI NIC with 3XP processor"
> float accross my desk, I was wondering how much the linux kernel uses the
> 3xp processor for its encryption offloading and such.  According to the
> hype it does DES without using the CPU, does linux take advantage of that?

Doing DES is uninteresting these days...

That feature is useless --- everything but IPsec does encryption at
application layer where NIC can not help.
								Pavel
-- 
Philips Velo 1: 1"x4"x8", 300gram, 60, 12MB, 40bogomips, linux, mutt,
details at http://atrey.karlin.mff.cuni.cz/~pavel/velo/index.html.

Re: 3com Driver and the 3XP Processor

[James Sutherland]

On Tue, 12 Jun 2001, Pavel Machek wrote:

> Hi!
>
> > I just had one of the "3com Etherlink 10/100 PCI NIC with 3XP processor"
> > float accross my desk, I was wondering how much the linux kernel uses the
> > 3xp processor for its encryption offloading and such.  According to the
> > hype it does DES without using the CPU, does linux take advantage of that?
>
> Doing DES is uninteresting these days...
>
> That feature is useless --- everything but IPsec does encryption at
> application layer where NIC can not help.

Now, if the NIC were to integrate with OpenSSL and offload some of THAT
donkey work... Just offloading DES isn't terribly useful, as Pavel says:
apart from anything else, DES is a bit elderly now - SSH using 3DES or
Blowfish etc... How dedicated is this card? Could it be used to offload
other work?


James.

Re: 3com Driver and the 3XP Processor

[Brent D. Norris]

> Now, if the NIC were to integrate with OpenSSL and offload some of THAT
> donkey work... Just offloading DES isn't terribly useful, as Pavel says:
> apart from anything else, DES is a bit elderly now - SSH using 3DES or
> Blowfish etc... How dedicated is this card? Could it be used to offload
> other work?

Sorry my bad it is 3DES that they have on it, but I don't know how
in-grained it is in it.  Like I sad it just floated across my desk a few
days ago and it sounded like a cool bit of hardware.

Brent Norris

Executive Advisor -- WKU-Linux

System Administrator -- WKU-Center for Biodiversity
                        Best Mechanical

W: 270-745-8864
H: 270-563-9226

"The problem with the Linux learning curve is that it is _so_ steep once
 at the top you can't see the people at the bottom"  --Doug Hagan

Re: 3com Driver and the 3XP Processor

[Martin Moerman]

On Thu, 14 Jun 2001, Brent D. Norris wrote:

> > Now, if the NIC were to integrate with OpenSSL and offload some of THAT
> > donkey work... Just offloading DES isn't terribly useful, as Pavel says:
> > apart from anything else, DES is a bit elderly now - SSH using 3DES or
> > Blowfish etc... How dedicated is this card? Could it be used to offload
> > other work?
> 
> Sorry my bad it is 3DES that they have on it, but I don't know how
> in-grained it is in it.  Like I sad it just floated across my desk a few
> days ago and it sounded like a cool bit of hardware.


The card is offloading TCP/IP checksums, TCP/IP packet fragmentation, and
does IPSEC through the ARM9 proc.

I like the card. but no real real linux drivers yet. only basic network
card drivers for linux.

/Martin
martin_moerman@eur.3com.com



> 
> Brent Norris
> 
> Executive Advisor -- WKU-Linux
> 
> System Administrator -- WKU-Center for Biodiversity
>                         Best Mechanical
> 
> W: 270-745-8864
> H: 270-563-9226
> 
> "The problem with the Linux learning curve is that it is _so_ steep once
>  at the top you can't see the people at the bottom"  --Doug Hagan
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: 3com Driver and the 3XP Processor

[nick]

So what is the truth to the rumors 3com was throwing around about the
"linux driver with ipsec support"?
	Nick

On Thu, 14 Jun 2001, Martin Moerman wrote:

> 
> 
> On Thu, 14 Jun 2001, Brent D. Norris wrote:
> 
> > > Now, if the NIC were to integrate with OpenSSL and offload some of THAT
> > > donkey work... Just offloading DES isn't terribly useful, as Pavel says:
> > > apart from anything else, DES is a bit elderly now - SSH using 3DES or
> > > Blowfish etc... How dedicated is this card? Could it be used to offload
> > > other work?
> > 
> > Sorry my bad it is 3DES that they have on it, but I don't know how
> > in-grained it is in it.  Like I sad it just floated across my desk a few
> > days ago and it sounded like a cool bit of hardware.
> 
> 
> The card is offloading TCP/IP checksums, TCP/IP packet fragmentation, and
> does IPSEC through the ARM9 proc.
> 
> I like the card. but no real real linux drivers yet. only basic network
> card drivers for linux.
> 
> /Martin
> martin_moerman@eur.3com.com
> 
> 
> 
> > 
> > Brent Norris
> > 
> > Executive Advisor -- WKU-Linux
> > 
> > System Administrator -- WKU-Center for Biodiversity
> >                         Best Mechanical
> > 
> > W: 270-745-8864
> > H: 270-563-9226
> > 
> > "The problem with the Linux learning curve is that it is _so_ steep once
> >  at the top you can't see the people at the bottom"  --Doug Hagan
> > 
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> > 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: 3com Driver and the 3XP Processor

[Kip Macy]

IPsec support will be binary only.

	-Kip
On Thu, 14 Jun 2001 nick@snowman.net wrote:

> So what is the truth to the rumors 3com was throwing around about the
> "linux driver with ipsec support"?
> 	Nick
> 
> On Thu, 14 Jun 2001, Martin Moerman wrote:
> 
> > 
> > 
> > On Thu, 14 Jun 2001, Brent D. Norris wrote:
> > 
> > > > Now, if the NIC were to integrate with OpenSSL and offload some of THAT
> > > > donkey work... Just offloading DES isn't terribly useful, as Pavel says:
> > > > apart from anything else, DES is a bit elderly now - SSH using 3DES or
> > > > Blowfish etc... How dedicated is this card? Could it be used to offload
> > > > other work?
> > > 
> > > Sorry my bad it is 3DES that they have on it, but I don't know how
> > > in-grained it is in it.  Like I sad it just floated across my desk a few
> > > days ago and it sounded like a cool bit of hardware.
> > 
> > 
> > The card is offloading TCP/IP checksums, TCP/IP packet fragmentation, and
> > does IPSEC through the ARM9 proc.
> > 
> > I like the card. but no real real linux drivers yet. only basic network
> > card drivers for linux.
> > 
> > /Martin
> > martin_moerman@eur.3com.com
> > 
> > 
> > 
> > > 
> > > Brent Norris
> > > 
> > > Executive Advisor -- WKU-Linux
> > > 
> > > System Administrator -- WKU-Center for Biodiversity
> > >                         Best Mechanical
> > > 
> > > W: 270-745-8864
> > > H: 270-563-9226
> > > 
> > > "The problem with the Linux learning curve is that it is _so_ steep once
> > >  at the top you can't see the people at the bottom"  --Doug Hagan
> > > 
> > > -
> > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > the body of a message to majordomo@vger.kernel.org
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > Please read the FAQ at  http://www.tux.org/lkml/
> > > 
> > 
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> > 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: 3com Driver and the 3XP Processor

[nick]

Erm, that is going to be a problem.  Crypto benifits more from open source
than any other market segment, and binary only drivers for linux are not
the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
them with someone else's product?
	Nick

On Thu, 14 Jun 2001, Kip Macy wrote:

> IPsec support will be binary only.
> 
> 	-Kip
> On Thu, 14 Jun 2001 nick@snowman.net wrote:
> 
> > So what is the truth to the rumors 3com was throwing around about the
> > "linux driver with ipsec support"?
> > 	Nick
> > 
> > On Thu, 14 Jun 2001, Martin Moerman wrote:
> > 
> > > 
> > > 
> > > On Thu, 14 Jun 2001, Brent D. Norris wrote:
> > > 
> > > > > Now, if the NIC were to integrate with OpenSSL and offload some of THAT
> > > > > donkey work... Just offloading DES isn't terribly useful, as Pavel says:
> > > > > apart from anything else, DES is a bit elderly now - SSH using 3DES or
> > > > > Blowfish etc... How dedicated is this card? Could it be used to offload
> > > > > other work?
> > > > 
> > > > Sorry my bad it is 3DES that they have on it, but I don't know how
> > > > in-grained it is in it.  Like I sad it just floated across my desk a few
> > > > days ago and it sounded like a cool bit of hardware.
> > > 
> > > 
> > > The card is offloading TCP/IP checksums, TCP/IP packet fragmentation, and
> > > does IPSEC through the ARM9 proc.
> > > 
> > > I like the card. but no real real linux drivers yet. only basic network
> > > card drivers for linux.
> > > 
> > > /Martin
> > > martin_moerman@eur.3com.com
> > > 
> > > 
> > > 
> > > > 
> > > > Brent Norris
> > > > 
> > > > Executive Advisor -- WKU-Linux
> > > > 
> > > > System Administrator -- WKU-Center for Biodiversity
> > > >                         Best Mechanical
> > > > 
> > > > W: 270-745-8864
> > > > H: 270-563-9226
> > > > 
> > > > "The problem with the Linux learning curve is that it is _so_ steep once
> > > >  at the top you can't see the people at the bottom"  --Doug Hagan
> > > > 
> > > > -
> > > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > > the body of a message to majordomo@vger.kernel.org
> > > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > > Please read the FAQ at  http://www.tux.org/lkml/
> > > > 
> > > 
> > > -
> > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > the body of a message to majordomo@vger.kernel.org
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > Please read the FAQ at  http://www.tux.org/lkml/
> > > 
> > 
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> > 
> 

Re: 3com Driver and the 3XP Processor

[Kip Macy]

So it would seem. Here is the polite message I received in response my
inquiry regarding the crypto interface to the card:
> 
> 
> Thank you for your inquiry.  We do not offer the
> technical spec;s for the IPSec
> features of this NIC, due to the intellectual
> property-heavy nature of this
> product.  We will, however, be releasing a Linux
> driver which supports this
> feature in the very near future under the beta
> section of our support site.  The
> base driver for the 990 will be open source, the
> advanced driver (offloads,
> etc.) will not be open source.  Hope that somewhat
> helps you.

On Thu, 14 Jun 2001 nick@snowman.net wrote:

> Erm, that is going to be a problem.  Crypto benifits more from open source
> than any other market segment, and binary only drivers for linux are not
> the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
> them with someone else's product?
> 	Nick
> 
> On Thu, 14 Jun 2001, Kip Macy wrote:
> 
> > IPsec support will be binary only.
> > 

As I mentioned previously IP heavy is a euphemism for commodity.

Re: 3com Driver and the 3XP Processor

[Dr. Kelsey Hudson]

On Thu, 14 Jun 2001, Kip Macy wrote:

> As I mentioned previously IP heavy is a euphemism for commodity.

...and 3Com is notoriuos for putting out commodity, cheesy hardware.

 Kelsey Hudson                                           khudson@ctica.com
 Software Engineer
 Compendium Technologies, Inc                               (619) 725-0771
---------------------------------------------------------------------------

Re: 3com Driver and the 3XP Processor

[Dr. Kelsey Hudson]

On Thu, 14 Jun 2001 nick@snowman.net wrote:

> Erm, that is going to be a problem.  Crypto benifits more from open source
> than any other market segment, and binary only drivers for linux are not
> the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
> them with someone else's product?

I would... 3Com hardware (as far as i'm concerned) is garbage

 Kelsey Hudson                                           khudson@ctica.com
 Software Engineer
 Compendium Technologies, Inc                               (619) 725-0771
---------------------------------------------------------------------------

Re: 3com Driver and the 3XP Processor

[nick]

I've installed several thousand 3com cards of various ages and
types.  I've had less than 20 bad cards.
	Nick

On Thu, 14 Jun 2001, Dr. Kelsey Hudson wrote:

> On Thu, 14 Jun 2001 nick@snowman.net wrote:
> 
> > Erm, that is going to be a problem.  Crypto benifits more from open source
> > than any other market segment, and binary only drivers for linux are not
> > the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
> > them with someone else's product?
> 
> I would... 3Com hardware (as far as i'm concerned) is garbage
> 
>  Kelsey Hudson                                           khudson@ctica.com
>  Software Engineer
>  Compendium Technologies, Inc                               (619) 725-0771
> ---------------------------------------------------------------------------
> 

Re: 3com Driver and the 3XP Processor

[Alan Cox]

> I've installed several thousand 3com cards of various ages and
> types.  I've had less than 20 bad cards.
> 	Nick

Seconded. 3Com stuff is overpriced but reliable. They have also (prior to this
event) been very good at working with the Linux community, including digging out
docs for old MCA hardware they no longer even sell

Alan

Re: 3com Driver and the 3XP Processor

[Venkatesh Ramamurthy]

> Seconded. 3Com stuff is overpriced but reliable. They have also (prior to
this
> event) been very good at working with the Linux community, including
digging out
> docs for old MCA hardware they no longer even sell

Gee... I dont know whether they are good in working with Linux community. I
asked them documentation for thier PCI based ADSL modem... so that i can
write driver for it, which they decently refused to give it.... i tried hard
for months....

Re: 3com Driver and the 3XP Processor

[Pavel Machek]

Hi!

> > Seconded. 3Com stuff is overpriced but reliable. They have also (prior to
> this
> > event) been very good at working with the Linux community, including
> digging out
> > docs for old MCA hardware they no longer even sell
> 
> Gee... I dont know whether they are good in working with Linux community. I
> asked them documentation for thier PCI based ADSL modem... so that i can
> write driver for it, which they decently refused to give it.... i tried hard
> for months....

And do they really have that docs? I asked them for docs on 3com
homeconnect. Guess what, they probably can't give it to me. 3com
homeconnect == vicam.
								Pavel
-- 
I'm pavel@ucw.cz. "In my country we have almost anarchy and I don't care."
Panos Katsaloulis describing me w.r.t. patents at discuss@linmodems.org

Re: 3com Driver and the 3XP Processor

[Martin Moerman]

Well nick, that is your choice.

/Martin
martin_moerman@eur.3com.com

On Thu, 14 Jun 2001 nick@snowman.net wrote:

> Erm, that is going to be a problem.  Crypto benifits more from open source
> than any other market segment, and binary only drivers for linux are not
> the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
> them with someone else's product?
> 	Nick
> 
> On Thu, 14 Jun 2001, Kip Macy wrote:
> 
> > IPsec support will be binary only.
> > 
> > 	-Kip
> > On Thu, 14 Jun 2001 nick@snowman.net wrote:
> > 
> > > So what is the truth to the rumors 3com was throwing around about the
> > > "linux driver with ipsec support"?
> > > 	Nick
> > > 
> > > On Thu, 14 Jun 2001, Martin Moerman wrote:
> > > 
> > > > 
> > > > 
> > > > On Thu, 14 Jun 2001, Brent D. Norris wrote:
> > > > 
> > > > > > Now, if the NIC were to integrate with OpenSSL and offload some of THAT
> > > > > > donkey work... Just offloading DES isn't terribly useful, as Pavel says:
> > > > > > apart from anything else, DES is a bit elderly now - SSH using 3DES or
> > > > > > Blowfish etc... How dedicated is this card? Could it be used to offload
> > > > > > other work?
> > > > > 
> > > > > Sorry my bad it is 3DES that they have on it, but I don't know how
> > > > > in-grained it is in it.  Like I sad it just floated across my desk a few
> > > > > days ago and it sounded like a cool bit of hardware.
> > > > 
> > > > 
> > > > The card is offloading TCP/IP checksums, TCP/IP packet fragmentation, and
> > > > does IPSEC through the ARM9 proc.
> > > > 
> > > > I like the card. but no real real linux drivers yet. only basic network
> > > > card drivers for linux.
> > > > 
> > > > /Martin
> > > > martin_moerman@eur.3com.com
> > > > 
> > > > 
> > > > 
> > > > > 
> > > > > Brent Norris
> > > > > 
> > > > > Executive Advisor -- WKU-Linux
> > > > > 
> > > > > System Administrator -- WKU-Center for Biodiversity
> > > > >                         Best Mechanical
> > > > > 
> > > > > W: 270-745-8864
> > > > > H: 270-563-9226
> > > > > 
> > > > > "The problem with the Linux learning curve is that it is _so_ steep once
> > > > >  at the top you can't see the people at the bottom"  --Doug Hagan
> > > > > 
> > > > > -
> > > > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > > > the body of a message to majordomo@vger.kernel.org
> > > > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > > > Please read the FAQ at  http://www.tux.org/lkml/
> > > > > 
> > > > 
> > > > -
> > > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > > the body of a message to majordomo@vger.kernel.org
> > > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > > Please read the FAQ at  http://www.tux.org/lkml/
> > > > 
> > > 
> > > -
> > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > > the body of a message to majordomo@vger.kernel.org
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > Please read the FAQ at  http://www.tux.org/lkml/
> > > 
> > 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: 3com Driver and the 3XP Processor

[David S. Miller]

nick@snowman.net writes:
 > Erm, that is going to be a problem.  Crypto benifits more from open source
 > than any other market segment, and binary only drivers for linux are not
 > the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
 > them with someone else's product?

Many of us on the networking developer team believe that making the
programming interface to the cpus on the Tigon3 is the biggest mistake
3com could ever make.

What made the Acenic so ubiquitous and interesting was that you could
program the firmware on the board to do whatever you like.  They even
provided an entire firmware developer kit so you could hack on it.

So many useful projects came from this capability.

I feel dirty working on the Tigon3 driver for 2.4.x because of this.

Later,
David S. Miller
davem@redhat.com

Re: 3com Driver and the 3XP Processor

[nick]

So are there any intresting changes one can make to the acenic?  I've got
one, mostly unused right now.  I've been told it is mostly a pair or R5ks
hooked back to back.  Would anyone have a recommendation for a replacement
to the 3cr990?
	Nick

On Thu, 14 Jun 2001, David S. Miller wrote:

> 
> nick@snowman.net writes:
>  > Erm, that is going to be a problem.  Crypto benifits more from open source
>  > than any other market segment, and binary only drivers for linux are not
>  > the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
>  > them with someone else's product?
> 
> Many of us on the networking developer team believe that making the
> programming interface to the cpus on the Tigon3 is the biggest mistake
> 3com could ever make.
> 
> What made the Acenic so ubiquitous and interesting was that you could
> program the firmware on the board to do whatever you like.  They even
> provided an entire firmware developer kit so you could hack on it.
> 
> So many useful projects came from this capability.
> 
> I feel dirty working on the Tigon3 driver for 2.4.x because of this.
> 
> Later,
> David S. Miller
> davem@redhat.com
> 

Re: 3com Driver and the 3XP Processor

[Kip Macy]

The acenic is definitely a kick-ass card. One's natural
inclination is to assume that an interface is obscured because 
it is second rate.
		-Kip

On Thu, 14 Jun 2001, David S. Miller wrote:

> 
> nick@snowman.net writes:
>  > Erm, that is going to be a problem.  Crypto benifits more from open source
>  > than any other market segment, and binary only drivers for linux are not
>  > the way to go.  I guess I need to get rid of my 5-10 3cr990s and replace
>  > them with someone else's product?
> 
> Many of us on the networking developer team believe that making the
> programming interface to the cpus on the Tigon3 is the biggest mistake
> 3com could ever make.
> 
> What made the Acenic so ubiquitous and interesting was that you could
> program the firmware on the board to do whatever you like.  They even
> provided an entire firmware developer kit so you could hack on it.
> 
> So many useful projects came from this capability.
> 
> I feel dirty working on the Tigon3 driver for 2.4.x because of this.
> 
> Later,
> David S. Miller
> davem@redhat.com
> 

Re: 3com Driver and the 3XP Processor

[David S. Miller]

nick@snowman.net writes:
 > So are there any intresting changes one can make to the acenic?

Like I said, there is an entire firmware developer kit, so the only
limit is your imagination and coding skills :-)

Later,
David S. Miller
davem@redhat.com

Re: 3com Driver and the 3XP Processor

[Pete Wyckoff]

davem@redhat.com said:
> nick@snowman.net writes:
>  > So are there any intresting changes one can make to the acenic?
> 
> Like I said, there is an entire firmware developer kit, so the only
> limit is your imagination and coding skills :-)

I wrote a new firmware from scratch to offload most of a message passing
implementation (MPI, actually) into the Alteon NIC, including timeout
and retransmit, message matching, header building, etc.  It's almost
ready for release.  We're currently working on using both processors
of the Tigon in parallel.

There is other work which has modified the Alteon firmware to do, for
example, segmentation and reassembly (Underwood et al.), and further
work in progress on other topics.

Programmable NICs are fun.  I'd like to get one of those Tigon3 cards
to see if any of the register layout on the Tigon2 is still there,
hoping the interface is similar at least.

		-- Pete

Re: 3com Driver and the 3XP Processor

[David S. Miller]

Pete Wyckoff writes:
 > We're currently working on using both processors
 > of the Tigon in parallel.

It is my understanding that on the Tigon2, the second processor is
only for working around hw bugs in the DMA controller of the board and
cannot be used for other tasks.

WRT. tigon3, it was mentioned on this list that it is a pair of arm9
cpus, one for rx and one for tx.

Later,
David S. Miller
davem@redhat.com

Re: 3com Driver and the 3XP Processor

[Pekka Pietikainen]

On Fri, Jun 15, 2001 at 08:37:15AM -0700, David S. Miller wrote:
> 
> Pete Wyckoff writes:
>  > We're currently working on using both processors
>  > of the Tigon in parallel.
> 
> It is my understanding that on the Tigon2, the second processor is
> only for working around hw bugs in the DMA controller of the board and
> cannot be used for other tasks.
> 
> WRT. tigon3, it was mentioned on this list that it is a pair of arm9
> cpus, one for rx and one for tx.
> 
Might be worth asking broadcom instead of 3com for the specs, 
as they seem to be selling it as a chip (BCM5700/5701), whereas 3com sells a
board (3c996).

-- 
Pekka Pietikainen

Re: 3com Driver and the 3XP Processor

[Martin Moerman]

On Fri, 15 Jun 2001, Pekka Pietikainen wrote:

> On Fri, Jun 15, 2001 at 08:37:15AM -0700, David S. Miller wrote:
> > 
> > Pete Wyckoff writes:
> >  > We're currently working on using both processors
> >  > of the Tigon in parallel.
> > 
> > It is my understanding that on the Tigon2, the second processor is
> > only for working around hw bugs in the DMA controller of the board and
> > cannot be used for other tasks.
> > 
> > WRT. tigon3, it was mentioned on this list that it is a pair of arm9
> > cpus, one for rx and one for tx.
> > 
> Might be worth asking broadcom instead of 3com for the specs, 
> as they seem to be selling it as a chip (BCM5700/5701), whereas 3com sells a
> board (3c996).
> 

Guys,

To make it easier, Tell me exactly what you need in documentation and I
will try to get it for you. 

Martin Moerman
martin_moerman@eur.3com.com




> -- 
> Pekka Pietikainen
> 
> 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: 3com Driver and the 3XP Processor

[Jamie Lokier]

David S. Miller wrote:
> Pete Wyckoff writes:
>  > We're currently working on using both processors
>  > of the Tigon in parallel.
> 
> It is my understanding that on the Tigon2, the second processor is
> only for working around hw bugs in the DMA controller of the board and
> cannot be used for other tasks.

It certainly can be used for other tasks.  At CERN we have programmed
both processors on the Tigon2 to act as a traffic generator &
measurement tool to test switches.  We use lots of cards in parallel to
drive all ports on a switch.  It's given us much more useful results
than the expensive S*******s tester.

We even went so far as to build a Tigon2 development kit for Linux, and
got about 5% better performance just from using a better version of GCC and
linker tricks.

The Tigon2 cards are wonderful because of our ability to program it
however we like, and I really hope we can build similar interesting devices
from a crop of someone's Tigon3 cards.

-- Jamie

Re: 3com Driver and the 3XP Processor

[Jes Sorensen]

>>>>> "David" == David S Miller <davem@redhat.com> writes:

David> Pete Wyckoff writes:
>> We're currently working on using both processors of the Tigon in
>> parallel.

David> It is my understanding that on the Tigon2, the second processor
David> is only for working around hw bugs in the DMA controller of the
David> board and cannot be used for other tasks.

Actually it was intended to be used for other stuff but they ended up
having to use it for workarounds.

Jes

Re: 3com Driver and the 3XP Processor

[David S. Miller]

Kip Macy writes:
 > The acenic is definitely a kick-ass card. One's natural
 > inclination is to assume that an interface is obscured because 
 > it is second rate.

No, that's not my take.  My personal view is that 3com thinks that
allowing anyone to program the card like that was the biggest mistake
Alteon made with the Acenic.

Heh, "heavy intellectual property", it's a friggin' cpu that can
access the cards onboard memory and registers.  Sounds like just
another Acenic with a new PCI front-end to me.  If they are referring
to their IPSEC microcode when they say this, we don't care about that
just show us how to program the chip and we'll write our own :-)

Later,
David S. Miller
davem@redhat.com

Re: 3com Driver and the 3XP Processor

[Kip Macy]

On Thu, 14 Jun 2001, David S. Miller wrote:

> 
> No, that's not my take.  My personal view is that 3com thinks that
> allowing anyone to program the card like that was the biggest mistake
> Alteon made with the Acenic.
> 
Why do you think that they think that? Alteon certainly avoids exerting
itself in supporting the open firmware kit, so cost is not a good
justification. What do they lose?
	
			-Kip

Re: 3com Driver and the 3XP Processor

[David S. Miller]

Kip Macy writes:
 > Why do you think that they think that?

They are going to provide a proprietary ipsec module object for
Linux.  That's pissing off a bunch of folks right?

Well, what are you able to do about it without docs on how to program
the tigon3 on-board cpus?  Right?

The name of the game is control.

Later,
David S. Miller
davem@redhat.com

Re: 3com Driver and the 3XP Processor

[Jes Sorensen]

>>>>> "David" == David S Miller <davem@redhat.com> writes:

David> Kip Macy writes:
>> The acenic is definitely a kick-ass card. One's natural inclination
>> is to assume that an interface is obscured because it is second
>> rate.

David> No, that's not my take.  My personal view is that 3com thinks
David> that allowing anyone to program the card like that was the
David> biggest mistake Alteon made with the Acenic.

Of course one then goes to wonder why the AceNIC suddenly became the
most sold Gigabit Ethernet card on the market ... oh my, I love 3Com's
way of thinking ;-(

Jes (yes I know it's late responding to this)

Re: 3com Driver and the 3XP Processor

[Alan Cox]

> I just had one of the "3com Etherlink 10/100 PCI NIC with 3XP processor"
> float accross my desk, I was wondering how much the linux kernel uses the
> 3xp processor for its encryption offloading and such.  According to the
> hype it does DES without using the CPU, does linux take advantage of that?

No. Sell the card to a windows user buy a cheap taiwanese mass market ethernet
and spend the rest on the faster CPU. I bet that is more cost effective for
DES performance..