[gPXE] Transfer speed with e1000 and pcnet32 in VMware

[Joakim Schicht]

I have further investigated the issue and come to a conclusion that it may be a vmware related issue only. The stall (hang) behaviour only occurs when no physical nic is connected to host (regardless of guest connection being host-only or nat). Using bridged connection in vmware shows good connection. Keeping physical nic connected and testing over host-only connection also shows good connection. And there is no difference in using e1000/pcnet32 regarding the issue.

Without physical nic and stall behaviour, the transfer will average on 1950 Kb/s. With physical nic and without stall behaviour, the transfer will average on 3100 Kb/s.

I can only acknowledge the fact that this happens, but cannot understand why, or see any logic in it.

But using another stack than gPXE in the same environment (even with no physical nic connected) will not trigger this stall behaviour. Which leads me to think that it could be some weird combination. 

Either way, it is definately a vmware-related issue.

I am not familiar with other virtualization software and did not test further.

Sorry for wasting your time on this issue. I honestly thought it was gPXE related.

Btw, my host testing machine is an sx280 with XP SP3 and with vmware workstation 6.5.3.

Thanks for looking into this.

Joakim Schicht
 		 	   		  
_________________________________________________________________
Få nye Windows Live™ Messenger.
http://download.live.com/messenger
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://etherboot.org/pipermail/gpxe/attachments/20091130/71c68dd2/attachment.html