48 replies to this topic

[Slice]

I want to present you updated RealtekR1000 driver. I added support for new chips according to linux's sources.
Now it supports RTL8168E,F/8111E,F.It works for me and works better then v2.0.6 from realtek.com.tw. Now I have no problems with sleep/wake. With native driver I sometimes lose LAN after wake.With this driver never.

I have a little problem. Network utility shows Link Speed = 0Can someone help me to resolve this Resolved!
Unplug cable
 Screen Shot 2013-03-11 at 21.05.46.png   45.32K   240 downloads
Plug cable
 Screen Shot 2013-03-11 at 21.05.25.png   47.58K   204 downloads

UPDATE. In this version I applied my own inventions so my driver differs from native.
 RealtekR1000SL.kext-3.0.2.zip   208.59K   207 downloads
 RealtekR1000SL_11.03.2013.zip   308.38K   152 downloads


25.04.2013
Corrected a mistake with warm boot after Windows. Thanks for Mieze, dmazar, rehabman.
Updated part for RTL 8105/RTL8136
This kext is compiled under SL, XCode 3.2.5. For use in ML I'll check little later. EDITED: I checked, it works in ML 10.8.3.
 RealtekR1000SL.kext-v304.zip   188.58K   130 downloads
 RealtekR1000SL_src_25.04.2013.zip   376.54K   24 downloads

Attached Files

•  RealtekR1000SL_09.03.2013.src.zip   272.79K   26 downloads
•  RealtekR1000SL__10.03.2013_.src.zip   290.84K   11 downloads

[buoo]

Thank you, very appreciated

[Maniac10]

Thanks a lot Slice, will test it later.

[Badeendje]

This kext is way better than the older ones, no more lag or loose of signal during normal use.
Also sleep seems to work fine.
Thank you, this really completes my stable build using an otherwise already well supported GA-Z68X-UD3H board.

[Rampage Dev]

Great work. Will have to fire up my testing rig and try it out.

[mitch_de]

THANKS!

I use (since 9 months) the  Lnx2Mac Realtek RTL81xx driver for my LAN (EP-35 Board) - worked fine (also has sleep / speed fixes) , but i always try some new things.

http://lnx2mac.blogs...osx-driver.html   (v0.0.90) Lnx2Mac's Realtek RTL81xx Driver... Due to the shortcomings of the current Realtek 81xx Hackintosh drivers (such as lack of or limited support of 8111E, 32/64-bit, sleep issues), I endeavored to port the Linux RTL81xx driver to Mac OS X.)

Problem with rtl8168 ( Hersteller-ID: 0x10ec,   Geräte-ID: 0x8168,   Subsystem-Hersteller-ID: 0x1458,   Subsystem-ID: 0xe000,   Versions-ID: 0x0001)
Your kext get loaded : org.chucko.driver.RealtekR1000 (3.0.0d1) <37 10 5 4 3 1>
but no Ethernet available.  OS X 10.8.3 dev D76, Gigabyte EP35

system.log (yours):
10.03.13 10:45:20,000 kernel[0]: RealtekR1000: init
10.03.13 10:45:20,000 kernel[0]: RealtekR1000: start
10.03.13 10:45:20,000 kernel[0]: RealtekR1000: R1000InitBoard @ PCI 0x4,00
10.03.13 10:45:20,000 kernel[0]: RTL8xxx@0xd000: Found chip reg=bb500700 // = val32 = ReadMMIO32(TxConfig);
10.03.13 10:45:20,000 kernel[0]: RTL8xxx@0xd000: Unknown device type, reg = 0xb8000000
// code section in voidRealtekR1000::R1000GetMacVersion(). mfg not set in case ...
bb500700   // u32 val32 = ReadMMIO32(TxConfig);
bb500700 & 0xFC800000 = 0xB8000000 // u32 reg = val32 & 0xFC800000; // was 0x7C800000
But there is no case for 0xB8000000 !
SOLUTION:
But if you would use // was 0x7C800000 vs 0xFC800000  for the & , then it would be in case as 0x38000000
0xbb500700 & 0x7C800000 = 0x38000000 (vs bb500700 & 0xFC800000 = 0xB8000000 ), and case 0x3800000, or change value the case to B8000000, it would work - side effects fpor others unknown of course.

void RealtekR1000::R1000GetMacVersion()
{
u32 val32 = ReadMMIO32(TxConfig);
u32 reg = val32 & 0xFC800000; // was 0x7C800000
u32 ICVerID = val32 & 0x00700000;
DLog("Found chip reg=%08x\n", val32);
switch (reg)
{
.......
case 0x38000000:
if(ICVerID == 0x00000000) {
mcfg = MCFG_8168B_2;
} else if(ICVerID == 0x00500000) {
mcfg = MCFG_8168B_3;
} else {
mcfg = MCFG_8168B_3;
}
break;
....
default:
mcfg = -1;
break;
}

if (mcfg < 0)
{
DLog("Unknown device type, reg = %#08x\n", reg);
}
else
{
...
}

10.03.13 10:45:20,000 kernel[0]: RTL8xxx@0xd000: R1000InitBoard: **** Ethernet chip unrecognized ****
10.03.13 10:45:20,000 kernel[0]: RTL8xxx@0xd000: start: Failed, R1000ProbeAndStartBoard returned false
10.03.13 10:45:20,000 kernel[0]: RTL8xxx@0xd000: start: returning '0'


system.log with working (v0.0.90) Lnx2Mac's Realtek RTL81xx Driver:
10.03.13 02:06:28,000 kernel[0]: [RealtekRTL81xx:init] RealtekRTL81xx.kext v0.0.90 ©2010-2011 by Lnx2Mac (lnx2mac@gmail.com)
10.03.13 02:06:28,000 kernel[0]: [RealtekRTL81xx:init] _logLevel is now 132 (0x0084)
10.03.13 02:06:28,000 kernel[0]: [RealtekRTL81xx:init] Using updated PHY config method
10.03.13 02:06:28,000 kernel[0]: [RealtekRTL81xx:init] init completed
10.03.13 02:06:28,000 kernel[0]: [RTL81xx@0xd000:rtl8168_init_board] NIC identified as RTL8168B/8111B mcfg=2

[Slice]

OK, thank you for the note. Will correct.
lnx2mac driver doesn't work for me, I don't remember exactly what is the problem, like my chip 8111F is not supported.

[Slice]

mitch_de, on 10 March 2013 - 08:41 AM, said:

But if you would use // was 0x7C800000 vs 0xFC800000  for the &amp; , then it would be in case as 0x38000000
0xbb500700 &amp; 0x7C800000 = 0x38000000 (vs bb500700 &amp; 0xFC800000 = 0xB8000000 ), and case 0x3800000, or change value the case to B8000000, it would work - side effects fpor others unknown of course.

The reason is an attempt to join with 8169

  case 0x98000000:
   mcfg = MCFG_8169SC_2;
   break;

I will add case 0xB8000000

[Warped]

Slice,

I start to love this driver as it is first RTL driver which offers me working WOL on my GA-P35-DS4 under ML 10.8.2!
I just started to play with it and currently I see there is problem with idle sleep.
With LNX2MAC driver sleep works perfectly - so it looks my system is OK.
Also - when system is woken via WOL, wakeup reason is undefined:

Mar 10 21:28:27 Piotr-Desktop kernel[0]: en0: disable
Mar 10 21:28:27 Piotr-Desktop kernel[0]: en0: setActivationLevel(0)
Mar 10 21:28:27 Piotr-Desktop kernel[0]: en0: setPowerState(0)
Mar 10 21:28:27 Piotr-Desktop kernel[0]: en0: R1000Suspend
Mar 10 21:28:27 Piotr-Desktop kernel[0]: Wake reason: ?
Mar 10 21:28:56 Piotr-Desktop kernel[0]: Previous Sleep Cause: 5
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: setPowerState(1)
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: R1000Resume
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: R1000ResetTask
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: enable
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: setActivationLevel(2)
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: Selected medium index 5
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: selectMedium, index=5
Mar 10 21:28:56 Piotr-Desktop kernel[0]: en0: R1000SetMedium(0x3, 0x2, 0x2)
Mar 10 21:28:56 Piotr-Desktop.local configd[19]: network changed: v4(en0-:192.168.1.1) DNS- Proxy- SMB 

Maybe this explains why WOL wake system with monitor off for 15-20sec and next system returns to sleep.
Setting ttyspreventsleep not helps, but using caffeination -s allows to keep system ON.

Anyway - million thx for excellent work!

[Slice]

More correct version 3.0.2 is in topic.

[verleihnix]

Hi Slice,
Thanks for your efforts.
Tried to use your kext (3.0.1) for device 10ec:8136 in lenovo z560, but it won't work.


11.03.13 08:34:47,000 kernel[0]: RealtekR1000: init
11.03.13 08:34:47,000 kernel[0]: RealtekR1000: start
11.03.13 08:34:47,000 kernel[0]: RealtekR1000: R1000InitBoard @ PCI 0x7,00
11.03.13 08:34:47,000 kernel[0]: com_chucko_RealtekR1000: Ethernet address XX:XX:XX:XX:XX:XX
11.03.13 08:34:47,000 kernel[0]: RTL8xxx@0x2000: attaching as en0
11.03.13 08:34:47,000 kernel[0]: en0: start: returning '1'
11.03.13 08:34:47,000 kernel[0]: en0: enable
11.03.13 08:34:47,000 kernel[0]: en0: setActivationLevel(2)
11.03.13 08:34:47,000 kernel[0]: en0: Selected medium is NULL, forcing to autonegotiation
11.03.13 08:34:47,000 kernel[0]: en0: selectMedium, index=6
11.03.13 08:34:47,000 kernel[0]: en0: R1000SetMedium(0x2, 0x2, 0x1)
11.03.13 08:34:47,000 kernel[0]: en0: RTLN8101NicReset
11.03.13 08:34:47,000 kernel[0]: en0: RTL8100DSM
11.03.13 08:34:47,000 kernel[0]: en0: RTL8100DisableEEE called on an unsupported chip.

Is this device supported? ID was in plist.

Regards

[Warped]

Slice,
I don't know what things are improved in 3.0.2 but I decided give you just small update after going to 3.0.2:
-idle sleep is still not working
-WOL still gives "wake reason: ?" and turns-off PC after 15sec
-the same like previous version driver auto detects my GE network like 100Mb/s. Manual selection to 1GE works OK.

[tj5]

I've been testing this version on my MSI Z77MA-G45 mobo and so far it works flawlessly! Latency seems down, downloads about 10% faster - sleep/wake is problem free. LNX2MAC has issues on my mobo when waking sometimes the network is dead and the OS needs to be rebooted to get the network going again. Nice job, thanks!

[theconnactic]

Hi, Slice! Congratulations for your work!

Will your driver also work for the RTL8105E (10/100), which i think is a variation of the 8139? Or, better yet, can you add support for it in your driver (which i can personally confirm it works for the 8111).

Best regards!

[gothic860]

Works great on my GA-Z77N-Wifi thank you very much!

[Slice]

verleihnix, on 11 March 2013 - 08:16 PM, said:

Hi Slice,
Thanks for your efforts.
Tried to use your kext (3.0.1) for device 10ec:8136 in lenovo z560, but it won't work.


11.03.13 08:34:47,000 kernel[0]: RealtekR1000: init
11.03.13 08:34:47,000 kernel[0]: RealtekR1000: start
11.03.13 08:34:47,000 kernel[0]: RealtekR1000: R1000InitBoard @ PCI 0x7,00
11.03.13 08:34:47,000 kernel[0]: com_chucko_RealtekR1000: Ethernet address XX:XX:XX:XX:XX:XX
11.03.13 08:34:47,000 kernel[0]: RTL8xxx@0x2000: attaching as en0
11.03.13 08:34:47,000 kernel[0]: en0: start: returning '1'
11.03.13 08:34:47,000 kernel[0]: en0: enable
11.03.13 08:34:47,000 kernel[0]: en0: setActivationLevel(2)
11.03.13 08:34:47,000 kernel[0]: en0: Selected medium is NULL, forcing to autonegotiation
11.03.13 08:34:47,000 kernel[0]: en0: selectMedium, index=6
11.03.13 08:34:47,000 kernel[0]: en0: R1000SetMedium(0x2, 0x2, 0x1)
11.03.13 08:34:47,000 kernel[0]: en0: RTLN8101NicReset
11.03.13 08:34:47,000 kernel[0]: en0: RTL8100DSM
11.03.13 08:34:47,000 kernel[0]: en0: RTL8100DisableEEE called on an unsupported chip.

Is this device supported? ID was in plist.

Regards

theconnactic, on 12 March 2013 - 04:32 PM, said:

Hi, Slice! Congratulations for your work!

Will your driver also work for the RTL8105E (10/100), which i think is a variation of the 8139? Or, better yet, can you add support for it in your driver (which i can personally confirm it works for the 8111).

Best regards!

Sorry guys, I made no new support for 8105/8136 series cards although it is possible.
Just some work needed. If someone do this....
The new sources include a folder with recent linux driver for those.
 RealtekR1000SL_12.03.2013-src-with8101.zip   368K   22 downloads

[beta992]

Thanks Slice!

Going to try with my P8H67-M Pro MB, seems that I have a Realtek 8111E.

I never had any problems with their driver (Realtek), but it is always nice to have a 'bug free' kext.

[Maniac10]

Slice, with the latest version it gets an ip even though the cable is not even connected. The one before that worked great. It's a RTL8111E.

3.0.2d1:

12/03/13 19:58:52,000 kernel[0]: en0: setActivationLevel(2)
12/03/13 19:58:52,000 kernel[0]: en0: Selected medium index 6
12/03/13 19:58:52,000 kernel[0]: en0: selectMedium, index=6
12/03/13 19:58:52,000 kernel[0]: en0: R1000SetMedium(0x2, 0x2, 0x1)
12/03/13 19:58:52,000 kernel[0]: en0: setLinkStatus but no PHYStatus

Older (2.0.0d1?):

12/03/13 20:00:55,000 kernel[0]: en0: setActivationLevel(2)
12/03/13 20:00:55,000 kernel[0]: en0: Selected medium is NULL, forcing to autonegotiation
12/03/13 20:00:55,000 kernel[0]: en0: selectMedium, index=6
12/03/13 20:00:55,000 kernel[0]: en0: R1000SetMedium(0x2, 0x2, 0x1)

[theconnactic]

Thank you, Slice! I'm going to see what can i do!

Best regards!

EDIT: Slice, are you sure the new Linux drivers for the 8136/8105 are included with your sources? In Which folder are they? I only found 8101, 8102, 8103 and 8139: are these the drivers?

[Slice]

theconnactic, on 13 March 2013 - 02:01 AM, said:

Thank you, Slice! I'm going to see what can i do!

Best regards!

EDIT: Slice, are you sure the new Linux drivers for the 8136/8105 are included with your sources? In Which folder are they? I only found 8101, 8102, 8103 and 8139: are these the drivers?

As I see

static const struct {
const char *name;
u8 mcfg;
u32 RxConfigMask; /* Clears the bits supported by this chip */
} rtl_chip_info[] = {
_R("RTL8101E", CFG_METHOD_1, 0xff7e1880),
_R("RTL8101E", CFG_METHOD_2, 0xff7e1880),
_R("RTL8101E", CFG_METHOD_3, 0xff7e1880),
_R("RTL8102E", CFG_METHOD_4, 0xff7e1880),
_R("RTL8102E", CFG_METHOD_5, 0xff7e1880),
_R("RTL8103E", CFG_METHOD_6, 0xff7e1880),
_R("RTL8103E", CFG_METHOD_7, 0xff7e1880),
_R("RTL8103E", CFG_METHOD_8, 0xff7e1880),
_R("RTL8401E", CFG_METHOD_9, 0xff7e1880),
_R("RTL8105E", CFG_METHOD_10, 0xff7e1880),
_R("RTL8105E", CFG_METHOD_11, 0xff7e1880),
_R("RTL8105E", CFG_METHOD_12, 0xff7e1880),
_R("RTL8105E", CFG_METHOD_13, 0xff7e1880),
_R("RTL8402", CFG_METHOD_14, 0xff7e1880),
_R("RTL8106E", CFG_METHOD_15, 0xff7e1880),
_R("RTL8106E", CFG_METHOD_16, 0xff7e1880)
};