291 replies to this topic
#21
Posted 11 January 2010 - 04:12 PM
Still just getting self-assigned IP's going to move onto other things until I hear from you again (hnak & wayover especially.) Maybe Another User can give the kexts I provided a shot too...
I'm curious though, the ICH-X entries in the Info.plist, are those device id's for the ICH itself or for the ethernet?
I'm curious though, the ICH-X entries in the Info.plist, are those device id's for the ICH itself or for the ethernet?
#22
Posted 12 January 2010 - 03:01 AM
Well, 82575/82576 ( 0x10a7 ) is not supported by e1000e-1.1.2 of the original Linux project which I based on ( e1000.sourceforge.net ).
You need to build your own driver from Intel's igp-2.1.1 source, which is available from Intel site.
They are ethernet device IDs ( as far as I can read from the original source ).
I added a simple code fragment to enable/disable device in setPowerState ( in 1.0.2), however, as my system cannot sleep for some reasons, I don't think I can do anything further about the issue.
You need to build your own driver from Intel's igp-2.1.1 source, which is available from Intel site.
scalar, on Jan 11 2010, 04:12 PM, said:
I'm curious though, the ICH-X entries in the Info.plist, are those device id's for the ICH itself or for the ethernet?
wayover, on Jan 11 2010, 10:45 AM, said:
tested in 10.6, after wake up from sleep kp
#23
Posted 12 January 2010 - 03:23 PM
Quote
Linux* driver for Intel® 82575/6-based network connections using kernel version 2.4.21 through 2.6.x.
There are three Linux* Base Drivers for Intel® Gigabit Network Connections.
- e1000e-x.x.x.x.tar.gz supports Intel® 82571/2/3/4 or 82566/7-based PCI-E gigabit network adapters/connections.
- igb-x.x.x.tar.gz supports Intel® 82575/6-based gigabit network connections.
- e1000.x.x.x.tar.gz support all Intel® PCI and PCI-X gigabit network adapters/connections.
There are three Linux* Base Drivers for Intel® Gigabit Network Connections.
- e1000e-x.x.x.x.tar.gz supports Intel® 82571/2/3/4 or 82566/7-based PCI-E gigabit network adapters/connections.
- igb-x.x.x.tar.gz supports Intel® 82575/6-based gigabit network connections.
- e1000.x.x.x.tar.gz support all Intel® PCI and PCI-X gigabit network adapters/connections.
I'll look into building my own from igb-2.1.1.tar.gz, but I'm rather out of date on my coding skills so that might take a bit to setup. I'm wondering it's time to replace my Pro/100 with something more useful in the meantime since I suspect the kext I've used since 10.5.2 as being unstable in 10.5.8 for me now. Thought about DSDT to resolve this but most of the people I see right now doing that are focusing on 10.6.x.
#24
Posted 13 January 2010 - 12:23 AM
scalar, on Jan 12 2010, 04:23 PM, said:
I'm wondering it's time to replace my Pro/100 with something more useful in the meantime since I suspect the kext I've used since 10.5.2 as being unstable in 10.5.8 for me now.
#25
Posted 13 January 2010 - 07:01 AM
hnak, on Jan 12 2010, 05:01 AM, said:
I added a simple code fragment to enable/disable device in setPowerState ( in 1.0.2), however, as my system cannot sleep for some reasons, I don't think I can do anything further about the issue.
#26
Posted 13 January 2010 - 01:23 PM
#27
Posted 13 January 2010 - 04:32 PM
hnak, on Jan 12 2010, 04:23 PM, said:
If you plan to get a new NIC, Marvell 88E8053 works best within my experience, as it is natively supported by Apple. Realtek 8168/8169 should work with Apple or R1000.kext. Of course, you may try Intel 8257x though they are more expensive.
Also I'm guessing you did more than just package up the E1000e src into the 82566MM project structure, I haven't yet diff'd your E1000e src against the one from sourceforge to compare, but I did some reading about the differences between linux, bsd, darwin & OSX's kernel drivers and it doesn't seem like I can simply run a make (or build under XCode) and have it all 'just work'. Doing a kernel driver might be more than I take take on for a 'first OSX coding project'...?
Btw apologies for cluttering up this thread, but having played with OSX for 2 years on this box without 82575 support, I am ready to move forward in some way. Especially as I'm sans a 'real' Mac for 3-4 months now...
#28
Posted 13 January 2010 - 09:08 PM
scalar, on Jan 13 2010, 04:32 PM, said:
What about a PCI card with native support?
scalar, on Jan 13 2010, 04:32 PM, said:
Also I'm guessing you did more than just package up the E1000e src into the 82566MM project structure, I haven't yet diff'd your E1000e src against the one from sourceforge to compare, but I did some reading about the differences between linux, bsd, darwin & OSX's kernel drivers and it doesn't seem like I can simply run a make (or build under XCode) and have it all 'just work'. Doing a kernel driver might be more than I take take on for a 'first OSX coding project'...?
So I use most of Linux C files with no modification other than netdev.c, which implements Linux driver interface and must be ported to IONetworking framework ( AppleIntelE1000e.cpp ), though I tried to copy the original code as much as possible. I took the basic structure of the framework interface from 82566MM project.
I wrote kcompat.h so that Linux C files can be compiled ( almost ) without modidication. I hope you can utilize kcompat.h to compile igb-xxx files and use AppleIntelE1000e.cpp as the starting point.
#29
Posted 14 January 2010 - 01:50 AM
hnak, on Jan 13 2010, 01:08 PM, said:
I think that Realtek 8169 is easier to get than other 100-base NICs ( and sold at very affordable price ) even if you do not need Gigabit.
hnak, on Jan 13 2010, 01:08 PM, said:
The original idea was to keep up with Linux driver update easily as possible.
So I use most of Linux C files with no modification other than netdev.c, which implements Linux driver interface and must be ported to IONetworking framework ( AppleIntelE1000e.cpp ), though I tried to copy the original code as much as possible. I took the basic structure of the framework interface from 82566MM project.
I wrote kcompat.h so that Linux C files can be compiled ( almost ) without modidication. I hope you can utilize kcompat.h to compile igb-xxx files and use AppleIntelE1000e.cpp as the staring point.
So I use most of Linux C files with no modification other than netdev.c, which implements Linux driver interface and must be ported to IONetworking framework ( AppleIntelE1000e.cpp ), though I tried to copy the original code as much as possible. I took the basic structure of the framework interface from 82566MM project.
I wrote kcompat.h so that Linux C files can be compiled ( almost ) without modidication. I hope you can utilize kcompat.h to compile igb-xxx files and use AppleIntelE1000e.cpp as the staring point.
#30
Posted 26 January 2010 - 12:15 AM
i successed to sleep with AppleIntelE1000e.kext 1.0.2a on DFI P55-T36 (Intel 82578DC), but kp is happed after wake up on s3 sleep.

my pc configurations
if anynone have some problem of sleep, let's check to them, and try!
DSDT Database for P55 Motherboards.
Advanced DSDT Fixes: Enable Sleep om P55 Motherboards

my pc configurations
if anynone have some problem of sleep, let's check to them, and try!
DSDT Database for P55 Motherboards.
Advanced DSDT Fixes: Enable Sleep om P55 Motherboards
#31
Posted 06 February 2010 - 08:10 AM
It took me a while to set up sleep-capable PC.
I uploaded 1.0.2b, which does not cause KP at least in my environment.
It is a crude implementation, which does disable/stop interface on suspend ( as if it was kextunload-ed ) and start/enable on resume.
I uploaded 1.0.2b, which does not cause KP at least in my environment.
It is a crude implementation, which does disable/stop interface on suspend ( as if it was kextunload-ed ) and start/enable on resume.
#32
Posted 06 February 2010 - 11:20 PM
hnak, on Feb 6 2010, 05:10 PM, said:
It took me a while to set up sleep-capable PC.
I uploaded 1.0.2b, which does not cause KP at least in my environment.
It is a crude implementation, which does disable/stop interface on suspend ( as if it was kextunload-ed ) and start/enable on resume.
I uploaded 1.0.2b, which does not cause KP at least in my environment.
It is a crude implementation, which does disable/stop interface on suspend ( as if it was kextunload-ed ) and start/enable on resume.
Thank you! but 1.0.2b 's link is normal ? i can not access to it
#33
Posted 07 February 2010 - 12:00 AM
#34
Posted 07 February 2010 - 08:19 PM
Thank you , fix source!
i test it , improve kernel panic. but after s3 sleep wake up, ethernet is linkdown (not active).
dmesg is no output linkup messsage. no ip address.
AppleIntelE1000e.kext (1.0.2b)
example nvenet's message on my Mac mini(early 2009). there is Mac mini in same network environment.
nvenet (1.0.69)
i think Link Speed's 0 Mbit/s is releated. i check router setting , but doses'nt improve.
before s3 sleep

after s3 sleep
i test it , improve kernel panic. but after s3 sleep wake up, ethernet is linkdown (not active).
dmesg is no output linkup messsage. no ip address.
AppleIntelE1000e.kext (1.0.2b)
bash-3.2# dmesg AppleIntelE1000e(Info): Link is Up 100 Mbps Full Duplex, Flow Control: None IOHIDSystem::relativePointerEventGated: VBL too high (20004354), capping to 20000000 0 [Time 1265572832] [Message System Sleep (S3 Sleep) Wake reason = System Wake Previous Sleep Cause: 0 USB (EHCI):Port 1 on bus 0x1a has remote wakeup from some device USBF: 864.420 AppleUSBEHCI[0xb9f7000]::ResumeUSBBus - controller took (9) turns to get going
example nvenet's message on my Mac mini(early 2009). there is Mac mini in same network environment.
nvenet (1.0.69)
bash-3.2# dmesg NVEthernet: Ethernet address 00:26:b0:e9:d6:76 NVEthernet::setLinkStatus - Valid but not Active Ethernet [nvenet]: Link up on en0, 100-Megabit, Full-duplex, No flow-control, Debug [796d,0000,0de1,0005,41e1,0000] NVEthernet::setLinkStatus - link Valid and Active (S3 Sleep) Wake reason = OHC1 System Wake Previous Sleep Cause: 5 USB (OHCI):Port 7 on bus 0x4 has remote wakeup from some device NVEthernet::setLinkStatus - Valid but not Active Ethernet [nvenet]: Link up on en0, 100-Megabit, Full-duplex, No flow-control, Debug [796d,0000,0de1,0005,41e1,0000] NVEthernet::setLinkStatus - link Valid and Active
i think Link Speed's 0 Mbit/s is releated. i check router setting , but doses'nt improve.
before s3 sleep

after s3 sleep
#35
Posted 26 February 2010 - 05:37 PM
test 1.0.2b and sometimes there is a falling with kernel panic and the mouse cursor stops, in 10.5.8 and 10.6 too. How it occurs on my system: I have created more than three network locations and when there is a switching between them there is a panic. At first I change locations in network preferences then put Apply button = KP. Any idea? please help!
#36
Posted 26 February 2010 - 08:39 PM
Thanks.. at least my Intel d945gtp board see the on board card now.. What a struggle to get it to work under SL..
Yeah.. it works after removing the other AppleIntel82xxx.kext's out of IONetwokingFamily..
Yeah.. it works after removing the other AppleIntel82xxx.kext's out of IONetwokingFamily..
#37
Posted 28 February 2010 - 04:56 AM
wayover, on Feb 26 2010, 05:37 PM, said:
test 1.0.2b and sometimes there is a falling with kernel panic and the mouse cursor stops, in 10.5.8 and 10.6 too. How it occurs on my system: I have created more than three network locations and when there is a switching between them there is a panic. At first I change locations in network preferences then put Apply button = KP. Any idea? please help!
Switching network configuration may cause disable/enable sequence similar to sleep.
#38
Posted 28 February 2010 - 01:47 PM
#39
Posted 01 March 2010 - 03:41 AM
Thanks, 1.0.3 work after sleep. good job!!
#40
Posted 18 March 2010 - 09:25 PM
hnak, on Jan 4 2010, 07:34 PM, said:
AppleIntelE1000e.kext 32/64 for 10.6.
This driver is based on Intel Wired Ethernet for Linux ( e1000.sourceforge.net ) and 82566mm OSX driver ( 82566mm-osx-driver.googlecode.com ). As it uses ( almost as is ) Linux driver's source code, the license is GPL.
The base Linux driver is e100e-1.1.2 and can work with Intel PCI Express ethernet chips like 82578 used with P55 chipset.
( I tested 82573 and 82578 only as I do not have other NICs. )
List of the IDs supported by original driver:
0x105E,0x105F,0x1060,0x10D9,0x10DA,0x10A4,0x10D5,0x10A5,0x10BC,0x107D,
0x107E,0x107F,0x10B9,0x108B,0x108C,0x109A,0x10D3,0x10F6,0x150C,0x1096,
0x1098,0x10BA,0x10BB,0x1501,0x1049,0x104A,0x104B,0x104C,0x10C4,0x10C5,
0x104D,0x10BF,0x10F5,0x10CB,0x10BD,0x10E5,0x294C,0x10C0,0x10C3,0x10C2,
0x10CC,0x10CD,0x10CE,0x10DE,0x10DF,0x10EA,0x10EB,0x10EF,0x10F0
0x104B and 0x10F6 are supported by Apple's Intel82574L.kext ( so removed from this driver's Info.plist ).
2010/2/28: Sleep support.
1.0.3
This driver is based on Intel Wired Ethernet for Linux ( e1000.sourceforge.net ) and 82566mm OSX driver ( 82566mm-osx-driver.googlecode.com ). As it uses ( almost as is ) Linux driver's source code, the license is GPL.
The base Linux driver is e100e-1.1.2 and can work with Intel PCI Express ethernet chips like 82578 used with P55 chipset.
( I tested 82573 and 82578 only as I do not have other NICs. )
List of the IDs supported by original driver:
0x105E,0x105F,0x1060,0x10D9,0x10DA,0x10A4,0x10D5,0x10A5,0x10BC,0x107D,
0x107E,0x107F,0x10B9,0x108B,0x108C,0x109A,0x10D3,0x10F6,0x150C,0x1096,
0x1098,0x10BA,0x10BB,0x1501,0x1049,0x104A,0x104B,0x104C,0x10C4,0x10C5,
0x104D,0x10BF,0x10F5,0x10CB,0x10BD,0x10E5,0x294C,0x10C0,0x10C3,0x10C2,
0x10CC,0x10CD,0x10CE,0x10DE,0x10DF,0x10EA,0x10EB,0x10EF,0x10F0
0x104B and 0x10F6 are supported by Apple's Intel82574L.kext ( so removed from this driver's Info.plist ).
2010/2/28: Sleep support.
1.0.3
Thank you for this driver, my Dell StudioXPS i7 with built-in NIC 82567LV-2 work perfect with Sleep in Snow 10.6.2
hnak, on Jan 4 2010, 07:34 PM, said:
AppleIntelE1000e.kext 32/64 for 10.6.
This driver is based on Intel Wired Ethernet for Linux ( e1000.sourceforge.net ) and 82566mm OSX driver ( 82566mm-osx-driver.googlecode.com ). As it uses ( almost as is ) Linux driver's source code, the license is GPL.
The base Linux driver is e100e-1.1.2 and can work with Intel PCI Express ethernet chips like 82578 used with P55 chipset.
( I tested 82573 and 82578 only as I do not have other NICs. )
List of the IDs supported by original driver:
0x105E,0x105F,0x1060,0x10D9,0x10DA,0x10A4,0x10D5,0x10A5,0x10BC,0x107D,
0x107E,0x107F,0x10B9,0x108B,0x108C,0x109A,0x10D3,0x10F6,0x150C,0x1096,
0x1098,0x10BA,0x10BB,0x1501,0x1049,0x104A,0x104B,0x104C,0x10C4,0x10C5,
0x104D,0x10BF,0x10F5,0x10CB,0x10BD,0x10E5,0x294C,0x10C0,0x10C3,0x10C2,
0x10CC,0x10CD,0x10CE,0x10DE,0x10DF,0x10EA,0x10EB,0x10EF,0x10F0
0x104B and 0x10F6 are supported by Apple's Intel82574L.kext ( so removed from this driver's Info.plist ).
2010/2/28: Sleep support.
1.0.3
This driver is based on Intel Wired Ethernet for Linux ( e1000.sourceforge.net ) and 82566mm OSX driver ( 82566mm-osx-driver.googlecode.com ). As it uses ( almost as is ) Linux driver's source code, the license is GPL.
The base Linux driver is e100e-1.1.2 and can work with Intel PCI Express ethernet chips like 82578 used with P55 chipset.
( I tested 82573 and 82578 only as I do not have other NICs. )
List of the IDs supported by original driver:
0x105E,0x105F,0x1060,0x10D9,0x10DA,0x10A4,0x10D5,0x10A5,0x10BC,0x107D,
0x107E,0x107F,0x10B9,0x108B,0x108C,0x109A,0x10D3,0x10F6,0x150C,0x1096,
0x1098,0x10BA,0x10BB,0x1501,0x1049,0x104A,0x104B,0x104C,0x10C4,0x10C5,
0x104D,0x10BF,0x10F5,0x10CB,0x10BD,0x10E5,0x294C,0x10C0,0x10C3,0x10C2,
0x10CC,0x10CD,0x10CE,0x10DE,0x10DF,0x10EA,0x10EB,0x10EF,0x10F0
0x104B and 0x10F6 are supported by Apple's Intel82574L.kext ( so removed from this driver's Info.plist ).
2010/2/28: Sleep support.
1.0.3
2 user(s) are reading this topic
0 members, 2 guests, 0 anonymous users



Sign In
Create Account









