Jump to content

Marvell Support Effort


mortis
 Share

223 posts in this topic

Recommended Posts

It takes a great deal of effort to write drivers for hardware who's manufacturer is uncooperative. Not just time, but effort. Even for experienced people, it's difficult.

 

 

Well Altaic:

 

I think I speak in everyone's name when I say I am very thankful fo your efforts. I know its hard to por drivers.

I'm not critizicing you or anyone else in special. It`s just that I feel that the aim of the community should be shifted right now towards drivers......

there are lots of unsupported hardware wih may be feasable to port and support.

As long as Apple doesn't set out their final version of OsX86, we are jus beta-testing their security systems!

on the other hand, open source drivers can remain for any version.

 

For most of us, running 10.4.1 or 10.4.3 (any of the 2 availabe versions) doesn't make any difference. It has only had "some" impact on ATI users. But it has been a major focus on these groups. Should we (the community) loose time and publish results on how to crack OsX install methosd??

Wouldnt it be better if only Maxxus ( who's the responsible for most of the advances) works in silence and only releases the "final" hack when OSX86 is officially out????

 

I think we've only taught Apple how to improve security measures while in the meantime, done nothing to expand the Os base for the rest.

Link to comment
Share on other sites

Whatever apple does to combat pirating or use on other hardware, it'll be cracked. Of that I'm certain, and I am sure they are too.

 

As to where efforts should be diverted, it seems there is a very small portion of the community capable of writing drivers. I would say, with the exception of Maxxus, everyone who is capable and willing is writing drivers. I mean, we're not seeing multiple TPM cracks or anything.

 

Also, it's somewhat improper to call what they're doing "security," since that implies that the installed OS is insecure to hackers, which it's not. I wish people would refrain from saying things like, "New OSX security broken again." It's misleading, and is bad for apple's reputation. They value their reputation greatly, and it's just begging the corporation to stomp on all of these projects. It sucks for everyone involved.

Link to comment
Share on other sites

  • 2 weeks later...
Also, it's somewhat improper to call what they're doing "security," since that implies that the installed OS is insecure to hackers, which it's not. I wish people would refrain from saying things like, "New OSX security broken again." It's misleading, and is bad for apple's reputation. They value their reputation greatly, and it's just begging the corporation to stomp on all of these projects. It sucks for everyone involved.

 

That's a very good point. I would hate to see people targeted just because Apple's reputation is in jeopardy. If people are going to get sued, it better be over their ingenious methods of getting around things like TPM chips. They really do pride themselves on their reputation, and they do have a right to (9 times out of 10).

 

I also agree that we are basically beta-testing their TPM system though. As long as we keep going with it, they'll keep improving it. As they showed us with 10.4.1 to 10.4.2 they can improve it even after the intial release. I doubt that the official release of Macintels will be the final word in TPM or whatever they end up using as a protection system.

Link to comment
Share on other sites

I honestly don't have a timeframe for you all. Neither manufacturer responded to my inquiries. With a doc explaining the various registers, this would be a *ton* easier. It seems like the more I work on this, the more I realize that there's much more to be done. It'll be quite some time. If anyone who has experience writing drivers would like to join me, pm me.

Edited by altaic
Link to comment
Share on other sites

The AppleYukon driver in 10.4.4 claims to be compatible with Marvell Yukon2 88E8053 PCI-Express chips, ID 0x11AB/0x4362. Some newer motherboards apparently use this chip for integrated LAN.

 

This chip is apparently newer and quite a bit different from the 3C2000 / 3C940 / Yukon / 88E8001, which have ID 0x11AB/0x4320. The older chip seems much more common than Yukon2 and is probably what most people reading this thread seek support for.

 

I tried installing only the AppleYukon plugin from the 10.4.4 OS into 8F1111. Is this approach fatally flawed? I haven't read around enough to understand, but kextload didn't seem to complain. Anyway, I did this on a machine with a Yukon1, and the driver loaded but didn't recognize the NIC. Tweaking Info.plist, this problem is solved -- the driver creates an enX device, correctly determines the MAC, and manages to send out a DHCP request. Unfortunately it seems to fail to receive DHCP replies and causes the system to crash shortly after being enabled.

 

Indeed SysKonnect maintains one of the Linux drivers for this series of chips, and apparently shoved support for both Yukon and Yukon2 into that same sk98lin driver. There is some criticism of this approach on the linux-kernel mailing list, suggesting that perhaps those chips are too different to be comfortably supported by the same device driver, and a driver written for one won't accidentially support the other.

Link to comment
Share on other sites

brokenlink: I have the AppleYukon driver working perfectly on my 8F1111 system as we speak. I simply replaced the IONetworkingFamily and IOPCIFamily kexts on my system, and deleted all the unnecessary plug-ins. Because my board is the D915GEV (88E8050), I had to change the PCI match ID to 0x436111AB.

Link to comment
Share on other sites

It half works on the nforce3 marvel yukon, the card is detected and works but ppoe is screwed . It hangs solid the whole os if i try to connect to the inet with my ethernet ppoe modem

 

the device id for the nforce3 marvel yukon is 0x432011AB

Link to comment
Share on other sites

brokenlink: I have the AppleYukon driver working perfectly on my 8F1111 system as we speak. I simply replaced the IONetworkingFamily and IOPCIFamily kexts on my system, and deleted all the unnecessary plug-ins. Because my board is the D915GEV (88E8050), I had to change the PCI match ID to 0x436111AB.

 

This is good news.

 

It might be time to remove the two incompatible hardware list entries for the 88E8050 and 88E8053 chips.

 

The 88E8001 situation seems much more bleak. :D Still working on that driver altaic?

Link to comment
Share on other sites

This is the full ionetworking kext + the marvel driver from 10.4.4

 

<Snipped - Violation>

 

copy it in the extension folder overwriting the old or first delete the old and then copy the new.

 

Repair permissions delete the kextcache

 

sudo -s

kextcache -k /System/Library/Extensions/

 

reboot and pray, trying to do a ppoe connection hangs my system but the lan works so use it at your own risk

Link to comment
Share on other sites

Which is your chip version Myzar?

 

88E8001/8003/8010 the onboard nforce3 250 marvel

 

it's matched with the id 0x432011AB.

 

This pisses me off , the lan works but i can't connect with my ethernet alcatel speed touch home modem

 

it tries to connect then the connection times out and the whole os freezes , i'm non sure if it's the card or ppoe is fuxored on 10.4.3

Link to comment
Share on other sites

Are there reports of ppoe having problems with other cards? This may not be specific to the Yukons.

 

I haven't abandoned my code. It seems like the apple driver does what we need it to, albeit a bit unstably. Ideally, 10.4.4's TPM will be thwarted and we'll see if they're caused by mixing 10.4.4 and 10.4.3. My driver is a lower priority now that I know there is something that looks like it'll work for us. It happens to work for me right now, in fact.

 

As far as the whole "Don't Steal Mac OS X" thing goes, if you buy OS X and use it on non-apple hardware, the message has nothing to do with you. Apple just wants to stem the flow of OS X piracy created by the beta-testing. It was a clever way to get the message out without appearing to be heavy-handed.

Link to comment
Share on other sites

tried on 88E8001/8003/8010, with DHCP ethernet connection, gave me freezing OS...so it's not the problem of PPPoE i guess.

when I boot in -v, some weird messages showed up.

it was something like '00000,00000 starting BMU, bug hit.'

I dunno what this mean, but OSX seems trying BMU thing continuously but keep failing, and it may eventually kills OSX.

 

any ideas?

Link to comment
Share on other sites

hmm... maybe this driver needs to be initialized via EFI like the graphics one, or maybe it needs the new kernel somehow.

 

Looks like the kext loads ok, same as the new ATI ones.

 

Altaic, you say yours works fully, what chip are you on?

Edited by autoy
Link to comment
Share on other sites

hmm... maybe this driver needs to be initialized via EFI like the graphics one, or maybe it needs the new kernel somehow.

 

Looks like the kext loads ok, same as the new ATI ones.

 

Altaic, you say yours works fully, what chip are you on?

 

it's more like the driver is for Yukon-88E8053 only and has a bug with the old marvell and we are screwed again

 

in the plist it says

 

Apple-Marvell Yukon-2 Gigabit Ethernet Driver

 

and we are forcing it on the old card. oh well sounds like it's time to buy a cheap realtek card unless someone can disassemble the driver and figure out what goes wrong with the old card

Link to comment
Share on other sites

Up and running un a Toshiba M45sp351

Marvell is 88E8036

Just extracted the AppleYukon.kext plugin and placed it on my ionetworking family.kext plugijn folder, edited the usual plist.info to add the model, and chowned the whole thing.....

 

so far so good... gets recognized as a 88e8053

 

Os is 10.4.3, ( the 099 one not the 11a g or whatever)

Edited by mortis
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
 Share

×
×
  • Create New...