pci bus 0x0008 cardnum 0x00 function 0x00: vendor 0x10ec device 0x8168
 Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet 
controller


I have three different hosts with the same chipset. One is running build 81 one 
is running build 86 and one is running build 113. 
Build 81 is a Sager Laptop that has run with the rge driver for over a year. It 
has run without problems for that time
Build 86 is a on motherboard ECS 770 motherboard that has been working for 
almost a year without problems.
Build 113 is a ECS 790 am3 motherboard that I have just installed. The rge 
driver will only configure the interface as 10/half or 1000/full.
Under 10/half the interface is slow with many collisions but functions.
Under 1000/Full the interface functions for a couple of minutes and then data 
does not seem to flow.  A snoop of the interface when data slows shows that the 
negotiated packet payloads drop down to "0".  I have seen this problem before 
on sparc based platform doing high loads to a Microsoft iis server. I found 
bugid 6276464 for that issue and even through the bug reported that it was not 
affected in solaris 10 I found that 
tcpInPastWinSegs    = 47652     < should be zero.

Once I set the workaround TCP parameter tcp_co_min = 1500 I was able to get 
around the problem.

Now I was thinking that I would like to see if I could work around the rge 
problem in the same way since the packet payload problem was looked the same. 
So I went to set this TCP parameter "tcp_co_min" but it does not exist in build 
113! It does exist in build 81 and 86.

So Was the tcp_co_min parameter removed from the tcp stack and could this be 
causing the problem that I am seeing?

Thanks
<mike>
-- 
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to