new device support for forcedeth.c
PATCH 1/1 new device support for forcedeth.c
Hi Brian,
could you check the full duplex handling? I'm testing my new nforce-250 Gb (epox 8KDA3J) with a normal 100 MBit autonegotiation link partner and I'm getting bad performance (30 kB/sec) when sending. Half duplex works. The nic reports late collisions:
nv_tx_done: looking at packet 233, Flags 0x28200.
-> packet not yet completed nv_tx_done: looking at packet 233, Flags 0x1a810. bits 16, 15, 13, 11, 4 error, lastpacket, late collision, deferred, retryerror. nv_tx_done: looking at packet 234, Flags 0x8000. -> successful nv_tx_done: looking at packet 235, Flags 0x28200. -> not yet completed.
Autonegotiation seems to work correctly: np->linkspeed is set to 65636, np->duplex to 1. lspci reports:
00:05.0 Bridge: nVidia Corporation: Unknown device 00df (rev a2)
Subsystem: Unknown device 1695:100c
Flags: bus master, 66Mhz, fast devsel, latency 0, IRQ 11
Memory at ec000000 (32-bit, non-prefetchable)
I/O ports at b400 [size=8]
Capabilities: [44] Power Management version 200: de 10 df 00 07 00 b0 00 a2 00 80 06 00 00 00 00 10: 00 00 00 ec 01 b4 00 00 00 00 00 00 00 00 00 00 20: 00 00 00 00 00 00 00 00 00 00 00 00 95 16 0c 10 30: 00 00 00 00 44 00 00 00 00 00 00 00 0b 01 01 14
Could you give me a hint what's wrong? If you need further register values, just ask.
Thanks,
--
Manfred