Hi Dave,

Yes, I am very careful not to use the GPL code.  I look at it for reference 
only since the data sheets don't tell the entire picture with the Intel NICs.  

Thanks,
Matt

-----Original Message-----
From: Ertman, DavidX M [mailto:[email protected]] 
Sent: Monday, May 19, 2014 1:45 PM
To: Matt Edwards; [email protected]
Subject: RE: 82579 Transmit hang with C200 and ME

> -----Original Message-----
> From: Matt Edwards [mailto:[email protected]]
> Sent: Monday, May 19, 2014 8:59 AM
> To: Ertman, DavidX M; [email protected]
> Subject: RE: 82579 Transmit hang with C200 and ME
> 
> Hi Dave,
> 
> The stock driver under Linux works OK on the system.  The embedded 
> system is a real time extension to the Windows OS.  The extensions 
> take over the NIC such that Windows never sees it.  This works well on 
> many other Intel NICs.  Over the years we have come to rely on the 
> Intel source more than the data sheets for the NICs.
> 
> We have always had problems when the system has AMT/ME.  Is there 
> anything extra at a system level that we need to do when a PC has AMT?  
> An extra driver or something?  We have ME disabled in the BIOS.  I 
> also do all the checks to see if ME has control of the PHY and toggle 
> a bit to force the MAC/PHY link to go back to PCIe.
> 
> I realize this system is sort of off-topic for the e1000-devel forum 
> but I think even through premier support it would come back to you guys.
> 
> Thanks,
> Matt
> 
> -----Original Message-----
> From: Ertman, DavidX M [mailto:[email protected]]
> Sent: Monday, May 19, 2014 10:50 AM
> To: Matt Edwards; [email protected]
> Subject: RE: 82579 Transmit hang with C200 and ME
> 
> > -----Original Message-----
> > From: Matt Edwards [mailto:[email protected]]
> > Sent: Friday, May 16, 2014 7:18 AM
> > To: Ertman, DavidX M; [email protected]
> > Subject: RE: 82579 Transmit hang with C200 and ME
> >
> > Hi Dave,
> >
> > Thanks for the reply.
> >
> > I meant to say the E1000e.  I started with version 2.5.4 and merged 
> > in anything new I found in version 3.04.
> >
> > Matt Edwards
> >
> >
> > -----Original Message-----
> > From: Ertman, DavidX M [mailto:[email protected]]
> > Sent: Thursday, May 15, 2014 5:45 PM
> > To: Matt Edwards; [email protected]
> > Subject: RE: 82579 Transmit hang with C200 and ME
> >
> > > -----Original Message-----
> > > From: Matt Edwards [mailto:[email protected]]
> > > Sent: Wednesday, May 07, 2014 10:04 AM
> > > To: [email protected]
> > > Subject: [E1000-devel] 82579 Transmit hang with C200 and ME
> > >
> > > Hello,
> > >
> > > I have modified the current E1000 for a small embedded system.  
> > > The driver works well on systems that do not have Intel ME.  On 
> > > systems that have Intel ME the transmit hangs.  BTW the only 
> > > systems I have tested with ME are
> > > C200 PCH boards.  The problem could also be the result of 
> > > differences between the C200 and C202.
> > >
> > > On transmit I have verified that transmit descriptor head and tail 
> > > pointer registers are correct after the driver loads up the DMA 
> > > buffers but the MAC never sends the frames and never updates the 
> > > head, and never sets the DD bit.  This code works with other NIC's 
> > > and on other 82579 NIC's.  Receive DMA works fine.
> > >
> > > I've had transmit hangs when the PHY throws data away, but when 
> > > that happens the MAC updates the descriptors as thought the 
> > > packets were transmitted.
> > >
> > > Is there something at the system level that must be running with 
> > > Intel
> ME.
> > > I've gone the E1000 driver to find all ME specific code and I 
> > > think I have included it.
> > >
> > > Any help is very much appreciated, Best Regards, Matt
> > >
> >
> > Good afternoon Matt,
> >
> > You started with the e1000 driver?  Why not the e1000e driver?  
> > There are quite a few workarounds for 82579 in the e1000e driver 
> > that do not exist in e1000.
> >
> > Dave Ertman
> 
> Matt,
> 
> That clears up a few things.  You mentioned that this was for an 
> embedded system - what OS / kernel are you using?  Does this problem 
> reproduce with the in-kernel or SourceForge stock drivers?
> 
> Dave Ertman

Matt,

I am not a GPL lawyer by any means, but mixing GPL code and the Windows OS 
seems like a very dangerous place to be.  Since you have access to a premier 
support channel, I think this would be your best avenue to a possible 
resolution. 

As far as anything extra to do at the system level when ME is present?  I can 
only state to look at the current SourceForge driver and see what is being done 
there.  You probably already know this, but maybe focus in on things that use a 
MAC check for pch2lan or PHY type of 82579, as that was a major change in 
architecture and resulted in the origination of several workarounds.

Also, if you find issue with using the GPL code, you may want to look at our 
FreeBSD driver which is licensed under a BSD license.  It may fit better with 
your work.

Dave Ertman


------------------------------------------------------------------------------
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos.
Get unparalleled scalability from the best Selenium testing platform available
Simple to use. Nothing to install. Get started now for free."
http://p.sf.net/sfu/SauceLabs
_______________________________________________
E1000-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit 
http://communities.intel.com/community/wired

Reply via email to