nForce LAN driver

[Denicio]

eno,

could you please inform me what is the power save switch on info.plist?

it's set to true by default.

 

thanks.

 

edit: also, eno was right, Voodoo BETA 2c corrects lots of problems (no more cpus=1 for me).

 

oh... AND MAKE THiS GEM KEXT A STICKY.

[mitsos1os]

I tried the Voodo XNU Beta 2c kernel, and had exactly the same results. It even froze in less time with the voodoo kernel than the stagexnu. Eno, what did you mean that my system log has errors?

[Rob_Quads]

I've been using this driver fine in 10.5.2. I have upgraded up to 10.5.5 but am now having problems. Note I also went from default display to having hardware accelerated graphics via NVinject (Geforce 8800 GTS)

 

Currently when I boot the network card is detected and it assigns a network address to it (if I set it up for DHCP it seems to talk to the DHCP server fine and gets an IP address) but I am unable to ping anyone on the network or my router.

 

I have once booted up and the network was working but then a few minutes later it stopped working. Looking at the logs there was no apparent reason why it stopped. In the system.log the following info is logged by the driver

 

OfficeDesktop:Extensions convery$ ifconfig

lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 16384

inet6 fe80::1%lo0 prefixlen 64 scopeid 0x1

inet 127.0.0.1 netmask 0xff000000

inet6 ::1 prefixlen 128

gif0: flags=8010<POINTOPOINT,MULTICAST> mtu 1280

stf0: flags=0<> mtu 1280

en0: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500

inet 192.168.1.5 netmask 0xffffff00 broadcast 192.168.1.255

ether 00:1d:7d:d0:57:f5

media: autoselect (1000baseT <full-duplex,flow-control>) status: active

supported media: none autoselect 10baseT/UTP <half-duplex,flow-control> 10baseT/UTP <full-duplex> 10baseT/UTP <full-duplex,flow-control> 100baseTX <full-duplex> 100baseTX <full-duplex,flow-control> 100baseTX <half-duplex> 100baseTX <half-duplex,flow-control> 1000baseT <full-duplex> 1000baseT <full-duplex,flow-control>

 

Log for nForceLAN

 

Nov 6 19:00:53 OfficeDesktop kernel[0]: nForceLAN: Version 0.62.3n beta_osx initializing

Nov 6 19:00:54 OfficeDesktop kernel[0]: nForceLAN: setting optimization mode: cpu

Nov 6 19:00:54 OfficeDesktop kernel[0]: nForceLAN: found nVidia MCP73 Ethernet Controller with ID: 0x07dc10de Rev: 0xa2 (subsystem ID: 0xe0001458)

Nov 6 19:00:56 OfficeDesktop kernel[0]: nForceLAN: Ethernet address 00:1d:7d:d0:57:f5

Nov 6 19:00:56 OfficeDesktop kernel[0]: nForceLAN: Network link down.

Nov 6 19:00:56 OfficeDesktop kernel[0]: nForceLAN: Link speed now 1000Mbps(duplex = Full, flow control = Yes), code 0x10032.

 

 

Below you can see a sequence where it booted and had network access then it suddenly went. The timestamp is of when the pings stop

 

 

OfficeDesktop:~ convery$ ping 192.168.1.254

PING 192.168.1.254 (192.168.1.254): 56 data bytes

64 bytes from 192.168.1.254: icmp_seq=36 ttl=64 time=0.675 ms

64 bytes from 192.168.1.254: icmp_seq=37 ttl=64 time=0.710 ms

64 bytes from 192.168.1.254: icmp_seq=38 ttl=64 time=0.667 ms

64 bytes from 192.168.1.254: icmp_seq=39 ttl=64 time=0.702 ms

64 bytes from 192.168.1.254: icmp_seq=40 ttl=64 time=0.652 ms

 

OfficeDesktop:~ convery$ date

Thu Nov 6 19:31:28 GMT 2008

 

Log for that period...

 

Nov 6 19:31:20 OfficeDesktop com.apple.kdcmond[186]: kdcmond cannot retreive the computer's local Hostname , retrying ..

Nov 6 19:31:50: --- last message repeated 2 times ---

Nov 6 19:31:50 OfficeDesktop com.apple.kdcmond[186]: kdcmond cannot retreive the computer's local Hostname , retrying ..

Nov 6 19:32:35: --- last message repeated 1 time ---

Nov 6 19:32:35 OfficeDesktop com.apple.kdcmond[186]: kdcmond cannot retreive the computer's local Hostname , retrying ..

[Denicio]

I've been using this driver fine in 10.5.2. I have upgraded up to 10.5.5 but am now having problems. Note I also went from default display to having hardware accelerated graphics via NVinject (Geforce 8800 GTS)

 

Currently when I boot the network card is detected and it assigns a network address to it (if I set it up for DHCP it seems to talk to the DHCP server fine and gets an IP address) but I am unable to ping anyone on the network or my router.

 

I have once booted up and the network was working but then a few minutes later it stopped working. Looking at the logs there was no apparent reason why it stopped. In the system.log the following info is logged by the driver

 

OfficeDesktop:Extensions convery$ ifconfig

lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 16384

inet6 fe80::1%lo0 prefixlen 64 scopeid 0x1

inet 127.0.0.1 netmask 0xff000000

inet6 ::1 prefixlen 128

gif0: flags=8010<POINTOPOINT,MULTICAST> mtu 1280

stf0: flags=0<> mtu 1280

en0: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500

inet 192.168.1.5 netmask 0xffffff00 broadcast 192.168.1.255

ether 00:1d:7d:d0:57:f5

media: autoselect (1000baseT <full-duplex,flow-control>) status: active

supported media: none autoselect 10baseT/UTP <half-duplex,flow-control> 10baseT/UTP <full-duplex> 10baseT/UTP <full-duplex,flow-control> 100baseTX <full-duplex> 100baseTX <full-duplex,flow-control> 100baseTX <half-duplex> 100baseTX <half-duplex,flow-control> 1000baseT <full-duplex> 1000baseT <full-duplex,flow-control>

 

Log for nForceLAN

 

Nov 6 19:00:53 OfficeDesktop kernel[0]: nForceLAN: Version 0.62.3n beta_osx initializing

Nov 6 19:00:54 OfficeDesktop kernel[0]: nForceLAN: setting optimization mode: cpu

Nov 6 19:00:54 OfficeDesktop kernel[0]: nForceLAN: found nVidia MCP73 Ethernet Controller with ID: 0x07dc10de Rev: 0xa2 (subsystem ID: 0xe0001458)

Nov 6 19:00:56 OfficeDesktop kernel[0]: nForceLAN: Ethernet address 00:1d:7d:d0:57:f5

Nov 6 19:00:56 OfficeDesktop kernel[0]: nForceLAN: Network link down.

Nov 6 19:00:56 OfficeDesktop kernel[0]: nForceLAN: Link speed now 1000Mbps(duplex = Full, flow control = Yes), code 0x10032.

Below you can see a sequence where it booted and had network access then it suddenly went. The timestamp is of when the pings stop

OfficeDesktop:~ convery$ ping 192.168.1.254

PING 192.168.1.254 (192.168.1.254): 56 data bytes

64 bytes from 192.168.1.254: icmp_seq=36 ttl=64 time=0.675 ms

64 bytes from 192.168.1.254: icmp_seq=37 ttl=64 time=0.710 ms

64 bytes from 192.168.1.254: icmp_seq=38 ttl=64 time=0.667 ms

64 bytes from 192.168.1.254: icmp_seq=39 ttl=64 time=0.702 ms

64 bytes from 192.168.1.254: icmp_seq=40 ttl=64 time=0.652 ms

 

OfficeDesktop:~ convery$ date

Thu Nov 6 19:31:28 GMT 2008

 

Log for that period...

 

Nov 6 19:31:20 OfficeDesktop com.apple.kdcmond[186]: kdcmond cannot retreive the computer's local Hostname , retrying ..

Nov 6 19:31:50: --- last message repeated 2 times ---

Nov 6 19:31:50 OfficeDesktop com.apple.kdcmond[186]: kdcmond cannot retreive the computer's local Hostname , retrying ..

Nov 6 19:32:35: --- last message repeated 1 time ---

Nov 6 19:32:35 OfficeDesktop com.apple.kdcmond[186]: kdcmond cannot retreive the computer's local Hostname , retrying ..

 

Hey, i see we have an almost identical mobo, and of course the same PHY.

Tell me, after your connection fails, does it say that the iP is Self-Assigned?

 

Also, try the Voodoo Kernel BETA 2 and see what happens.

Check if you have 2 cores enabled, if so boot with cpus=1 and do some testing.

 

(i stopped using cpus=1 with the Voodoo Kernel).

[Denicio]

Anyone knows what is this?

 

Nov 7 11:17:32 dennis-kozlowskis-macpro31 kernel[0]: nForceLAN: NIC ring full, stalling.

 

At this point i lose my connection.

Still shows connected but there are no incoming/outgoing packets.

[macguybern]

My RTL8139 card bit the dust this past weekend so I have been stuck without a working ethernet card. I had been using the RTL8139 because the built-in network adapter is a nForce 590i MCP55. The MCP55 ethernet adapter has up until now not worked at all under osx (at least with my gigabyte motherboard).

 

The attached driver is a combination of planetbeing's nForce4 0.3c driver and the forcedeth 0.61 driver from linux 2.6.27. In theory the attached driver should work with any of the network adapters supported by the forcedeth 0.61 driver (from the nForce1 to the MCP79) but it has so far only been tested with my MCP55. This driver can be considered a replacement for both forcedeth.kext and nForceEthernet.kext. Please let me know how well this works (or doesn't work for that matter).

 

This driver is being released under the GPL.

 

The original nForce4 driver is posted here: nForce4 LAN driver

 

Known working physical devices:

Realtek 8211, 8211B, 8211C

Realtek/Cicada/Vitesse 8201

Marvel 88E1116, 88E3016

 

Supported PCI ids:

0x01c310de, 0x006610de, 0x00d610de, 0x008610de, 0x008c10de, 0x00e610de, 0x00df10de,

0x005610de, 0x005710de 0x003710de, 0x003810de, 0x026810de, 0x026910de, 0x037210de,

0x037310de, 0x03e510de, 0x03e610de, 0x03ee10de, 0x03ef10de, 0x045010de, 0x045110de,

0x045210de, 0x045310de, 0x054c10de, 0x054d10de, 0x054e10de, 0x054f10de, 0x07dc10de,

0x07dd10de, 0x07de10de, 0x07df10de, 0x076010de, 0x076110de, 0x076210de, 0x076310de,

0x0ab010de, 0x0ab110de, 0x0ab210de, 0x0ab310de

 

Updates:

- removed rx/tx checksum overrides. Should work with more cards out of the box now.

- [0.61.1alpha_osx]added support for sleep/wake functions.

- [0.61.2alpha_osx]added OptimizationMode option (0 = throughput, 1 = cpu). CPU optimization is the default since it appears to works better with some devices (my MCP55). The option can be set in the Info.plist.

- [0.61.3alpha_osx]renamed to nForceLAN.kext and some small bug-fixes.

- [0.61.4alpha_osx]fixed link timer bug.

- [0.61.5alpha_osx]fixed multicast bug.

- [0.62.0n alpha_osx]merged some of the code from the nVidia 0.62 linux driver into nForceLAN. Also added proper sleep (D3 mode).

- [0.62.1n alpha_osx]Macbook support (the MCP79 on the macbook doesn't work with the 8211C initialization code).

- [0.62.2n alpha_osx]Fixed media bug.

- [0.62.3n beta_osx]Fixed media with 1000BaseT and added autodetection of flow control. Driver appears stable so moved into Beta status.

 

enjoy!

-Nate

 

Kernel extension:

0.61.5alpha_osx

0.62.0n alpha_osx

0.62.1n alpha_osx

0.62.2n alpha_osx

0.62.3n beta_osx

 

Source:

0.62.3n beta_osx (source)

 

 

Hi to all fellow hackintosh geeks out there, thanks to this lively forum I was able to get my system fully operational.

Many thanks to ENO..tnx man.

To those guys with an AMDTurion64x2 specifically the HPdv6550 series im happy to announce that the above post here is indeed a perfect solution for the nvidia nforce networking controller.Im running kalyway 10.5.2 AMD/Intel EFI,all devices runs smoothly,all apps are functional too.

Im posting this comment as my way of gratifying all of you guys...after i stumbled upon this thread...my ethernet just pop!auto configured using DHCP settings.Man have i got sleepless nights trying every kext posted on every thread here in this forum..till i bump into one of Eno's miracle kext.."nForceLAN_0.62.3n.zip"which worked flawlessly on my onboard ethernet.I was about to give it a go buyin an Apple USB Ethernet, which seems to be the most viable solution to gettin hooked on the net, still the said device im not quite sure will work for me.

Im postin this now on my hackintosh laptop trio boot with Vista Ultimate and Mandriva Powerpack2009, everything works fine..nvidia8400GS graphics card runs crispy on it, even my airport Broadcom43xx wifi and bluetooth and all..

[eno]

Anyone knows what is this?

 

Nov 7 11:17:32 dennis-kozlowskis-macpro31 kernel[0]: nForceLAN: NIC ring full, stalling.

 

At this point i lose my connection.

Still shows connected but there are no incoming/outgoing packets.

 

That means that the transmission ring buffer is full. It either means the ring is too small or that the NIC has stopped sending packets for some reason. You can try increasing the size of the TX ring in the Info.plist. The maximum size is 16384.

 

I tried the Voodo XNU Beta 2c kernel, and had exactly the same results. It even froze in less time with the voodoo kernel than the stagexnu. Eno, what did you mean that my system log has errors?

 

There are a bunch of finder opengl errors and a launchd error that seems to occur right before a reboot.

[awesomerobot]

This worked like a charm... for about 30 mins, and then it keeps defaulting back to a static IP and not working! arrr! so close! can anyone help? I'm on a P6N Diamond mobo

[dafart]

Hi Eno,

 

 

I don't know if it's meant to be or it was just a mistake on your part but in your 1st post where you have the list of all the supported PCI ids, mine is in it:

0x003710de

 

Since, your driver (0.62.3n beta) was causing me random freeze issues I checked the info.plist and noticed that 0x003710de was not in there.

So I removed all the PCI ids and just put 0x037310de&0xfffeffff.

 

I also set optimization mode to 0 and now the driver works like a charm, for now. I've been downloading a lot stuff and no problems so far.

Hopefully it stays that way.

 

 

Will report back if I have any issues.

 

Good work on the driver

 

PS: so maybe you can add that PCI id in the list?

This very same solution appears to be working on my system as well (for specs see my signature). Just changing the device/vendor ID did not seem to work, I also had to change the optimization mode to 0.

After that no more freezing for me.

[rerven]

this is what i want, thanks so much....

[Envying]

The current version of driver is working like a charm on my Happle, this post's really worth a sticky!

[djk314]

Ok 0.62.3n beta_osx is now working without cpus=1, and without freezing on my ds4 Gigabyte n650i. Hope that it won't freeze again...

[KarateMiszcz]

Tried all versions of this driver, no luck. I have to use cpus=1 flag to make the system stable (by stable i mean working networking since i dont have any mouse/keyboard/other issues, just network stops working). Actually tried all drivers available and all have the same issue. My mobo is MSI P6N Diamond with 2 gigabit nvidia cards onboard (tried even to disable one) and device id is on the list. Oh, and system logs are very clear like everything is working fine. I think I'll end up with some PCI Intel NIC :/

[Denicio]

This worked like a charm... for about 30 mins, and then it keeps defaulting back to a static IP and not working! arrr! so close! can anyone help? I'm on a P6N Diamond mobo

 

Try setting your MTU to 1350, i had the same problem before.

[apc]

The best driver for my Gigabyte GA-M55S-S3 I ever tried. No "network link is down" problem, no need in cpus=1. 99,9% applications work excellent, but I have some strange network issues:

 

1. Latest Skype version. Even echo123 test call lasts for 3-4 seconds, not more. No need to say about video calls with friends, etc. Everything is like a charm with the same hardware & ISP under WinXP.

 

2. Apple Mail or latest Thunderbird work very unstable with IMAP4 mail servers. The same Thunderbird works excellent on MacBook with the same mail servers.

 

Since the most connections (even PPTP VPN) works fine, I think that bug is not so common, but I can easily reproduce it in my configuration. How can I help to resolve this issue?

[lord_muad_dib]

skype issue is not due the nic driver, it's a timing problem, you probably got an amd or an intel notebook with crappy bios, use voodoo kernel

[Envying]

skype issue is not due the nic driver, it's a timing problem, you probably got an amd or an intel notebook with crappy bios, use voodoo kernel

 

LMD, please sticky this post, cause this driver provides more stable traffic flow than of the drivers in other two stick "nforcelan driver" posts.

[apc]

skype issue is not due the nic driver, it's a timing problem, you probably got an amd or an intel notebook with crappy bios, use voodoo kernel

 

Guys, I can confirm that described bugs are caused definitely by this nic driver. I just installed RTL8139 LAN adapter and described Skype and Thunderbird bugs vanished immediately. No more changes, just replaced LAN adapter and stopped using nForceLAN.kext

[Guest cavallo]

Guys, I can confirm that described bugs are caused definitely by this nic driver. I just installed RTL8139 LAN adapter and described Skype and Thunderbird bugs vanished immediately. No more changes, just replaced LAN adapter and stopped using nForceLAN.kext

 

I have done it few monthes ago.

 

skype issue is not due the nic driver, it's a timing problem, you probably got an amd or an intel notebook with crappy bios, use voodoo kernel

 

I have tried a lot of nforce (forcedeth) kext with a lot of kernels, no way.

I am obbliged to enable lan on board in bios to get final cut and cs3 working.

I still have problems to get taost and popcorn working, also updated, without perfoming with rosetta.

[xhess]

Works perfect with Leo4all4.1 AMD only Edition.

 

I use this one:

http://forum.insanelymac.com/index.php?act...st&id=37188

for my Network. The driver inside of edition dosnt work.

 

System:

 

GA-M57SLI-S4 MoBo

AMD X2 5200+

2 GB Ram 800

external USB Drive 120 GB

GFX 8800 GT 512 MB

 

Big THX!

 

X.

[eno]

Bump. Sticky.

 

Guys, I can confirm that described bugs are caused definitely by this nic driver. I just installed RTL8139 LAN adapter and described Skype and Thunderbird bugs vanished immediately. No more changes, just replaced LAN adapter and stopped using nForceLAN.kext

 

As far as I have seen the problems you describe are not likely caused by only this driver. The driver works perfectly when none of the other components of the system are screwed up (shutdown fix is a known unstable component). For example I use my nForceLAN driver on my macbook pro (MCP79) and AMD hackintosh (MCP55) without a single issue. The later machine is a server. BTW, I have been using Skype regularly for the past week without any problems.

 

EDIT: acp, have you tried changing the optimization mode. I forgot to mention that there seems to be problems with nForce chipsets and optimization mode 1 (cpu) (linux driver defaults to 0). Please let me know if that alleviates your problems.

 

Please remember that many of the "fixes" we have seen in the past are really buggy. In fact, the kernel included with Leo4All causes kernel panics with the simple call to replaceOrCopyPacket (a standard networking controller call) but it works perfectly with the Voodoo XNU kernel as well as the vanilla kernel.

 

BTW, to anyone seeing slower than expected speeds with 1000BaseT. I think I figured out why the speeds remain slow. Expect 0.62.4n to enhance LAN transfer speeds on 1000BaseT.

 

Try setting your MTU to 1350, i had the same problem before.

 

Very interesting. I wonder if there is a problem with the standard IONetworkingInterface in some kernels. The next version of the driver will override some of the functionality of the IONetworkingInterface in order to provide gigabit speeds. It might address some issues that still exist with some modified kernels.

[eno]

bump

[icedtrip]

Can we go ahead and get this thing stickied? This is a big advance for onboard LAN and nVidia boards.

 

Please Sticky!!

[mastegr]

Hi everybody!

 

Im finding a problem with my ASUS M2N-VM HDMI and the nForceLAN.kext's. It loads OK, it works fine in normal conditions, BUT when I send several files to another machine on the network, the LAN crashes.

 

If I receive info, It performs great! I've made rsync's of 4Gb, and receiving info works perfect, my problem is sending... it simply crashes after a few files. Anyone has a similar problem?

 

Cheers!

[icedtrip]

Hi everybody!

 

Im finding a problem with my ASUS M2N-VM HDMI and the nForceLAN.kext's. It loads OK, it works fine in normal conditions, BUT when I send several files to another machine on the network, the LAN crashes.

 

If I receive info, It performs great! I've made rsync's of 4Gb, and receiving info works perfect, my problem is sending... it simply crashes after a few files. Anyone has a similar problem?

 

Cheers!

 

Mine has the exact same issue. Hate to say it, but I am glad I'm not alone. See my sig for my hardware.

 

Eno....any ideas on this one?