Jump to content
InsanelyMac Forum
hnak

AppleIntelE1000e.kext for 10.8/10.7/10.6/10.5

Recommended Posts

yaaaaaaa

i bought a cat6

i replaced the 5e existent

 

it run!! automatic!!

 

yaaa

 

I can make pictures of cables.. I'm sure of what I did!!

so strange it's not the same as you

Share this post


Link to post
Share on other sites
Advertisement

Hello guys.

After 10.8.5 update my network gone...

My MB is Asus Sabertooth P67 (lan - Intel® 82579)

during loading it even doesn't print anything related to network, as I compared logs lines from system.log

 

I updated kext to the latest version and still nothing. how and where can I check what's wrong and find out the source of the problem?

 

- I fixed permissions

- I don't use DSDT

 

thanx

 

UPD. something strange, it works times through time

Share this post


Link to post
Share on other sites

on this build i am unable to get link aggregation bond to work, in advanced i get red dots and 'invalid link'.  on the same system if i boot to windows, lagg is working just fine.  honestly i have never gotten lagg to work on osx hackintosh.

 

does anyone have ideas on how to get LAGG working with these drivers?

 

thanks

 

dave

Share this post


Link to post
Share on other sites

on this build i am unable to get link aggregation bond to work, in advanced i get red dots and 'invalid link'.  on the same system if i boot to windows, lagg is working just fine.  honestly i have never gotten lagg to work on osx hackintosh.

 

does anyone have ideas on how to get LAGG working with these drivers?

 

thanks

 

dave

You have LACP-supporting (IEEE 802.3ad compliant) switches ?

Share this post


Link to post
Share on other sites

There's something still wrong with 2.5.4. It seems to work fine when you surf the web (both IPv4 and IPv6), but network printing to an AirPort printer literally takes ages.

Older versions work perfectly though.

Share this post


Link to post
Share on other sites

I am on a Intel Haswell board (Intel DH87MC Motherboard).

 

I recently added a machine to my network, which is on a 50 ft Cat 5e through two gigabit switches. There are two additional connectors along the way to go through a wall. There is probably 60-65 feet of total cable.

 

When I am moving large amounts of data to that machine (10-60 gigs), after 10-15 gigs or so, the driver or the networking hardware seem to kick out. The transfer hangs, the network is no longer accessible, but as far as the OSX is concerned, the network is still connected. The machine itself, which is a server, also becomes inaccessible on my network. There is no OS crash. I traced the issue to the machine with this driver.

 

The only way to revive the  network seems to be restarting the machine. Then all is fine again for weeks at a time, until large amounts of data need to be copied to that 'remote' machine again. Any ideas?

 

EDIT: I can confirm that this issue does not occur with version 2.4.14, but with 2.5.4c.

Share this post


Link to post
Share on other sites

I get kernel panics with 2.5.4c when used in a 10.8.5 installation disc. Just before the language select screen pops up it panics referencing AppleIntelE1000e and IONetworkingFamily. This didn't happen with previous kext or System versions so I just downgraded the copy on the install disc to 2.4.14 and all is well.

 

Haven't noticed any issues with 2.5.4c on a running 10.8.5 system yet, but then my traffic isn't very heavy.

Share this post


Link to post
Share on other sites

I am on a Intel Haswell board (Intel DH87MC Motherboard).

 

I recently added a machine to my network, which is on a 50 ft Cat 5e through two gigabit switches. There are two additional connectors along the way to go through a wall. There is probably 60-65 feet of total cable.

 

When I am moving large amounts of data to that machine (10-60 gigs), after 10-15 gigs or so, the driver or the networking hardware seem to kick out. The transfer hangs, the network is no longer accessible, but as far as the OSX is concerned, the network is still connected. The machine itself, which is a server, also becomes inaccessible on my network. There is no OS crash. I traced the issue to the machine with this driver.

 

The only way to revive the  network seems to be restarting the machine. Then all is fine again for weeks at a time, until large amounts of data need to be copied to that 'remote' machine again. Any ideas?

 

EDIT: I can confirm that this issue does not occur with version 2.4.14, but with 2.5.4c.

I am having the exact same problem with a Gigabyte X79S-UP5. As long as there is no major load on the NIC, the network behaves fine. But today I was trying to move 60 GB of data over the network and the NIC crashed. Cannot ping or anything and there is no network access from the computer.

Share this post


Link to post
Share on other sites

I was having issues with the 2.5.4c driver on 10.8.5 with a Gigabyte z77x-up5th which I believe has the 82579 chip.  I was getting really bad Skype send packet loss, Google hangouts were disconnecting and DNS with long lookup names would fail.  When I ran "nslookup some.long.dns.name.that.should.fail.but.not.timeout.cnn.com 8.8.8.8" I would get no response from the server.  In capturing the packets at my firewall I saw that random bits of the name in the DNS query would be corrupted.  I tried forcing both sides to 100Mb instead of 1Gbps but that didn't help.  I reverted back to r66 in SVN and that is working great.

 

Oh and in case anyone else runs into this, the name lookups was also causing sporadic failures in iTunes Radio not being able to load the next song because they were querying something like streamingaudio.itunes.apple.com.c.footprint.net which was apparently long enough to cause packet problems as well.

 

Is there anything I could do to help to debug the issue further?

Share this post


Link to post
Share on other sites

 

Is there anything I could do to help to debug the issue further?

 

Interesting,

I was having all sorts of problems on my system too (also up5 th) and after pulling my hair out during repeated router (apple airport extreme), modem, computer and switch reboots. i found that if i changed from manual network settings to DHCP things worked perfectly again. now i had been on 2.5.4c since its release and only started having problems after the 10.8.5 update. I also can reserve my mac address for a specific IP in the airport extreme so that manual settings are not required but i found this very perplexing since my manual settings contain the exact same info as the DHCP delivered settings.

 

g\

Share this post


Link to post
Share on other sites

I found UDP csum does not work in 10.8.5.

I uploaded 2.5.4d, in which the function is disabled.

I will give it a test and report back... Many thanks...

 

EDIT: Tried 2.5.4d. Resulted at a hangup at the 10 gig mark of a 65 gig copy as it did previously with 2.5.4c. Reverted back to 2.4.14. Retested to confirm no such issue exists with that version.

Share this post


Link to post
Share on other sites

Does anyone have a download link to 2.4.14 , I've searched the whole internet for 2 hours and I cant seem to find it   :(

Share this post


Link to post
Share on other sites

I have kernel panic with version 2.5.4c I downloaded the commit r65 and have compiled by myself meanwhile

Share this post


Link to post
Share on other sites

I have kernel panic with version 2.5.4c I downloaded the commit r65 and have compiled by myself meanwhile

Same kernel panic here. I would compile it myself but since a very kind fellow here has provided it I don't have to.

I found UDP csum does not work in 10.8.5.

I uploaded 2.5.4d, in which the function is disabled.

The update causes things like Skype to have *incredible* outgoing packet loss, to the point where it's impossible to understand voice. It might not be related but reverting fixes the issue in it's entirety so I'm mildly assuming it is. I had tried both the precompiled and compiled-myself latest versions and the same thing happens in both. I'm using 10.9 DP8.

 

Additionally, enabling PCI Express Gen-3 on an MSI X79A-GD45 8D board causes this specific driver to panic. I find that odd but not too outrageous. If you need me to do some debugging I can.

 

Here it is.

<3

Share this post


Link to post
Share on other sites

I checked Debug builds with WireShark and found that check sum calculation is sometimes not requested by system when the packet is UDP.

A really strange behavior change...

 

When UDP hardware tx-check sum is enabled in 10.8.5:

*IP (or UDP) csum becomes corrupted when UDP payload size > 64.

*The length of ip packet length becomes strange ( not constant ), too.

 

It seems I need to check the 10.8.5 xnu source when it becomes available.

Edited by hnak

Share this post


Link to post
Share on other sites

I just did an svn update, recompiled, and it seems to be working. I'm not quite sure when the update was pushed to svn but if it fixes the issue I'll edit this post. If there is anything that immediately needs doing or investigating let me know since I have the hardware (Intel 82579V).

 

EDIT: it's working with skype. packet loss is *minimal*.

Share this post


Link to post
Share on other sites

I checked Debug builds with WireShark and found that check sum calculation is sometimes not requested by system when the packet is UDP.

A really strange behavior change...

 

When UDP hardware tx-check sum is enabled in 10.8.5:

*IP (or UDP) csum becomes corrupted when UDP payload size > 64.

*The length of ip packet length becomes strange ( not constant ), too.

 

It seems I need to check the 10.8.5 xnu source when it becomes available.

I just upgraded to 2.5.4d and I'm not seeing the same nslookup problems as before and my initial test with Skype seems to be working fine now.  Thanks!

Share this post


Link to post
Share on other sites

I will give it a test and report back... Many thanks...

 

EDIT: Tried 2.5.4d. Resulted at a hangup at the 10 gig mark of a 65 gig copy as it did previously with 2.5.4c. Reverted back to 2.4.14. Retested to confirm no such issue exists with that version.

I'm having a similar issue. Updated to 2.5.4d, started a rather large FTP upload on my intranet, and after approximately 10gb or less of data the connection goes into a black hole and a reboot is then required in order to regain network connectivity. Happened about four times within a 30min period, each time requiring a reboot. System log is filled with incessant repetitions of the following:

 

10/1/13 8:07:15.000 PM kernel[0]: AppleIntelE1000e(Err): Detected Hardware Unit Hang:

10/1/13 8:07:15.000 PM kernel[0]: TDH                  <62>

10/1/13 8:07:15.000 PM kernel[0]: TDT                  <41>

10/1/13 8:07:15.000 PM kernel[0]: next_to_use          <41>

10/1/13 8:07:15.000 PM kernel[0]: next_to_clean        <61>

10/1/13 8:07:15.000 PM kernel[0]: buffer_info[next_to_clean]:

10/1/13 8:07:15.000 PM kernel[0]: time_stamp           <422b>

10/1/13 8:07:15.000 PM kernel[0]: next_to_watch        <6f>

10/1/13 8:07:15.000 PM kernel[0]: next_to_watch.status <0>

10/1/13 8:07:15.000 PM kernel[0]: MAC Status             <40080083>

10/1/13 8:07:15.000 PM kernel[0]: PHY Status             <796d>

10/1/13 8:07:15.000 PM kernel[0]: PHY 1000BASE-T Status  <3800>

10/1/13 8:07:15.000 PM kernel[0]: PHY Extended Status    <3000>

10/1/13 8:07:15.000 PM kernel[0]: PCI Status             <10>

 

Reverted back to 2.4.14 and it's behaving much better.

 

System is a Z77X-UP5-TH on 10.8.5.

Share this post


Link to post
Share on other sites

Try this one. It does not resolve hang,  but may call a reset handler when detected.

Tested. Same problem, hanging up at around the 10 gig mark:

 

 

10/2/13 9:04:48.000 AM kernel[0]: AppleIntelE1000e(Err): Detected Hardware Unit Hang:

Share this post


Link to post
Share on other sites

I'm stress testing it to a NAS ATM - so far the latest isn't crashing instantly when dealing with gigabit transfers - and it's getting gigabit rather than 800 megabit which is an improvement, although that might be just chance. If you did something, yeah.

 

If you have an idea on how to crash it I'm game, I'm stress testing it, it's gone through probably 10GB of data both ways so far.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

×