Search
Support & Downloads
All of Support
This Category
Networking Connectivity
Intel® Network Adapters
PRO/100 NetWare* custom statistics

TxGoodFrames
Frames actually transmitted.

TxMaxCollisions
Collisions that exceed the 16 retries. May increase on saturated network (need to segment).

TxLateCollisions
Collisions encountered outside the normal collision window. Very rare - indicates problem with cables, distance, or duplex mismatch.

TxUnderrunErrors
Not transmitted or retransmitted due to DMA Underrun. Occurs because a lack of PCI bandwidth resulting in the Tx FIFO running dry during the transmission of a frame. Adaptive Tuning should keep this number under control. To correct, load the driver with TxThreshold=200 on the command line. See CurrentThreshold below.

TxLostCrsErrors
Transmitted despite the fact that it detected loss of link during the transmission.

TxDeferred
Deferred transmission due to activity on the link. This is a normal condition in a busy network, and can grow quite large.

TxSingleCollisions
Successfully sent after one collision. Normal in busy networks.

TxMultipleCollisions
Sent after more than 1 (less than 16) collisions. This only gets incremented once, even though one frame made have had several collisions before being successfully sent. Fix - .Segment network or increase IFS.

TxTotalCollisions
Includes late collisions and Max collisions.

RxGoodFrames
Number actually received and stored in memory.

RxCrcErrors
Frames discarded because of a CRC error. CRC Errors are usually caused by late collisions were another station has sent data in the middle of a frame. This number should be very low.

RxAlignmentErrors
Misaligned and CRC error frames.

RxResourceErrors
Good frames discarded because there were no resources available. Can be reduced by increasing the number of receive buffers or increasing the performance of the host machines processing power.

RxDmaOverrunErrors
Frames lost because receive FIFO overflowed. This can occur if the device is unable to get the necessary bandwidth on the system (PCI) bus. If the overflow condition persists for more than one frame, the frames that follow the first could also be lost. Remedies: a faster bus machine, or fewer devices on the bus.

RxColiisionDetectedErrors
Frames that encountered collisions during frame reception.

RxShortFrameErrors
Received shorter than the minimum frame length.

Flow Control TX Pause
Xon and Xoff frames transmitted.

Flow Control RX Pause
Xon and Xoff frames received.

Flow Control Unsupported
MAC Control frames received that are not Flow Control Pause frames. These frames contain the predefined MAX control type value, but contain an unsupported opcode.

TCO Rx (or Tx) Frames
Incremented at each received/transmitted TCO packet.

TxTimeouts & TxDynamicChaining
Not used

CurrentIFSValue
Current Interframe Spacing value. (wait after collision or to avoid collisions), Higher value indicates higher collisions and can lessen throughput.

CurrentThreshold
To lessen underruns , the software can change the wait time before starting to transfer data onto the hardware buffers. Higher value is longer. Poor bus performance, lots of traffic or lots of devices on the same bus can cause this to rise. If this number is always above 170, consider loading the driver with the TXThreshold parameter set to 200 (eliminates early sends).

RstartCnt
Normal to be non zero. Also known as RSC, it stands for Re-Start Count.

This applies to:
Intel® PRO/100 M Desktop Adapter
Intel® PRO/100 S Desktop Adapter
Intel® PRO/100 S Dual Port Server Adapter
Intel® PRO/100 S Management Adapter
Intel® PRO/100 S Server Adapter
Intel® PRO/100+ Adapter
Intel® PRO/100+ Dual Port Server Adapter
Intel® PRO/100+ Management Adapter
Intel® PRO/100+ Server Adapter

Solution ID: CS-008451
Date Created: 18-Jan-2004
Last Modified: 30-Nov-2007
Back to Top