
16.5 Receiving a Trace Dump
This section describes the steps and bit settings for receiving a trace
dump. Trace dumps are used for detailed fault analysis and include a
history of SCSI and servo command activity for the CTS.
To receive a trace dump, follow these steps:
1. Issue a SEND DIAGNOSTIC command with the SelfTest,
DevOfl, and UntOfl bits set to 0 and the MD bit set to 1. The
Parameter List Length field is ignored and can have any value.
2. Issue a RECEIVE DIAGNOSTIC RESULTS command with the
Allocation Length field set to 500h (1,280 bytes) and the TD bit
set to 1.
Note: To ensure that data in the trace dump is valid, be sure that the
CTS is reserved for the initiator’s exclusive use.
As shown in Table 16-5, 1,280 bytes of information are returned in the
trace dump.
Bytes... Contain this data...
0 to 255 SCSI Instruction Trace Table
256 to 511 SCSI State Trace Table
512 to 1,279 Servo Instruction, Status, Extended Status Tables
As described in the next section, you can use the SCSI instruction
trace table (bytes 0 to 255) to analyze SCSI command execution
sequences. Interpreting the other two sections of the trace dump
(bytes 256 to 1279), however, requires more sophisticated analysis
procedures that are beyond the scope of this manual.
Table 16-5 Bytes returned in a trace dump
May 1994 EXB-8205 and EXB-8505
(Standard and XL)
16-7
16 RECEIVE DIAGNOSTIC RESULTS (1Ch)
Comentarios a estos manuales