Hugh Brown
12/08/2008 11:40 AM
post18139
|
Please try the latest Net.e1000 driver that I have attached.
-----Original Message-----
From: s h [mailto:community-noreply@qnx.com]
Sent: Monday, December 08, 2008 10:19 AM
To: qnx4-community
Subject: Net.i82540
Hello,
We have 2 problems with Net.i82540 driver:
1) Network stops working and does not recover.
Machine have 2 cards, and running Net.ether82557 and Net.i82540. While
Net.i82540 is not working, the Net.ether82557 continues to work
properly. We noticed that this problem occurs when the card is receiving
many packets per second. At about 55000 interrupts per second the driver
stops working. Driver does not crash, just stops providing functions.
Chipset model: Intel 82547. This problem has two components:
1a) The driver should continue working and system should return to
normal behaviour when the load drops down.
1b) The driver should not configure the card to allow it to generate
excessive number of interrupts per second.
netinfo: Please, can you provide netinfo or explanation of codes
291,292,293,1099?
11:15:25 1 Status 29980 (291)
11:15:25 1 Status 29560 (292)
11:15:25 1 FF7F1F 000000 (293)
11:15:25 1 001D92 3E18E9 (3928) i82540 ( rx) some packets missed,
txing NACK
11:15:25 1 Status 29980 (1099)
11:15:25 1 FF7F1F 000000 (1099)
11:15:26 1 001D92 3E18E9 (3928) i82540 ( rx) some packets missed,
txing NACK
11:15:26 1 Status 29980 (1099)
11:15:26 1 FFFF1E 000000 (1099)
11:15:27 1 001D92 E08AA8 (3928) i82540 ( rx) some packets missed,
txing NACK
11:15:27 1 Status 29980 (291)
11:15:27 1 Status 29560 (292)
11:15:27 1 FFAF1A 000000 (293)
11:15:27 1 001D92 3E18E9 (3928) i82540 ( rx) some packets missed,
txing NACK
11:15:27 1 Status 18435 (291)
11:15:27 1 Status 29560 (292)
11:15:27 1 FF8B18 000000 (293)
11:15:27 1 001D92 E08A82 (3928) i82540 ( rx) some packets missed,
txing NACK
11:15:27 1 Status 4823 (1099)
11:15:27 1 FF0E00 000000 (1099)
11:15:27 1 Status 29980 (1099)
11:15:27 1 FFAF1A 000000 (1099)
11:15:27 1 Status 18435 (1099)
11:15:27 1 FF8B18 000000 (1099)
2) CRC errors.
(This problem is not on the same machines as problem 1)
Network consist of several identical qnx machine plus several other
machines and devices. The machines need virtual circuits, but they are
being broken from time to time which cause problems in ipc over network.
Chipset model: Intel 82547. The system shows "CRC errors" in "netinfo
-l" output:
Driver Slot 0: Driver Pid 32 Logical Net 1 Network Card: Ethernet/
INTEL 82540 Gigabit (Copper) Ethernet Controller
Vendor ID ........................ 0x8086
Device ID ........................ 0x1075
Revision ......................... 0x0
Physical Node ID ................. 00187D 0051DD
Media Rate ....................... 1000Mb/s FDX
Mtu .............................. 1514
I/O Port Range ................... 0xF8400000 -> 0xF840FFFE
Hardware Interrupt ............... 11
Total Packets Txd OK ............. 61549689
Total Packets Txd Bad ............ 0
Tx Collision Errors .............. 0
Tx Collision Errors (aborted) .... 0
Carrier Sense Lost on Tx ......... 0
FIFO Underruns During Tx ......... 0
Tx deferred ...................... 0
Out of Window Collisions ......... 0
Total Packets Rxd OK ............. 107442967
Total Rx Errors .................. 0
FIFO Overruns During Rx .......... 0
Alignment errors ................. 0
CRC errors ....................... 8402
Notes:
- We also tried the experimental Net.e1000 but even though the driver
reports 1000FDX its effective speed is only about 90kB/s.
- Also tried boards with Broadcoms and have problems with the
Net.bcm57xx which should auto-detect 1000FDX but detects 100HDX. And the
parameters are even worse when trying to force a speed and duplex mode.
Chipset model: Dual Broadcom BCM5787M...
View Full Message
|
|
|