SBus Expansion Subsystem Testing with SunDiag

This appendix is divided into two parts. The first half provides a brief overview of the SunDiag diagnostic program and explains how to start and run tests. The second part provides specific information on running xbtest which tests the SBus Expansion Subsystem.

An Overview of SunDiag

SunDiag is a system exerciser; its primary purpose is to "stress" Sun hardware devices and record any errors that result from this testing. SunDiag allows multiprocessing systems to run any number of tests on any or all processors.

By nature, this program also tests system configurations. SunDiag automatically probes for hardware devices when it is started. If the probe does not find a hardware device that is connected to the system under test, that indicates a problem.

System administrators and hardware maintenance personnel use this program the most. The SunDiag interface is easy to use however, and it can be run by anyone who wants to "stress-test" a Sun system.

This Appendix provides a brief overview of the SunDiag program. For a detailed description of the program and the individual tests, see the latest version of the SunDiag User's Guide, shipped with SunOS documentation.

Testing Modes

You can run the SunDiag System shipped with your SBus Expansion Subsystem in two different modes. These modes vary according to the configuration of your hardware components. The two modes of testing are transparent and nontransparent. These modes are mutually exclusive and you cannot run them both at the same time.

Transparent Mode

Select this testing mode whenever you have SBus cards plugged into the SBus Expansion Subsystem.

Nontransparent Mode

Use this testing mode when you do not have any SBus cards plugged into the SBus Expansion Subsystem. It will test only the SBus Expansion Subsystem. This mode is normally only used by the production locations and the field service representatives.

How to Run SunDiag

The primary and easiest way to run SunDiag test is from the SunDiag SUNVIEWŽ window. This is the window interface shown in Figure C-1, and the only one described in detail in this Appendix.

SunDiag also has a TTY interface to run tests from a terminal, and individual tests can be run from the command line of a SunOS shell. See the latest version of the SunDiag User's Guide for details.

Hardware and Software Requirements

SunDiag runs on any SPARCstation or SPARCsystem installed with SunOS version 4.1.1 or later. The operating system kernel must be configured to support all of the peripherals to be tested.

The amount of swap space required varies widely with the type and amount of hardware being tested. SunDiag tests require a minimum of 8 Mbytes of swap space for SPARCstations, and SPARCsystem servers should have enough swap space to equal twice the amount of on-board memory. In addition, frame buffer tests require additional swap space, and the virtual memory test has special swap space considerations. See the SunDiag User's Guide for more detailed information on swap space requirements.

The SunDiag Window

When you start SunDiag in SUNVIEW, the SunDiag window appears on your screen (see Figure C-1). The window is divided into four sections:

A system status panel, located on the upper-left of the screen. Test results are displayed here.

A performance monitor panel, located in the upper-middle of the screen. System activity levels are tracked and displayed here.

A control panel, located in the upper-right of the screen. The panel has buttons, settings, and window buttons that allow you to manipulate test parameters and select test options.

A console window, located at the bottom-left of the screen. This window displays operating system messages and test messages, and also allows you (as superuser) to use SunOS commands.

Figure C-1 The SunDiag SUNVIEW window for the SBus Expansion Subsystem

Starting the SunDiag SUNVIEW Window

Follow these steps to run the SunDiag SunView user interface.

1. Become superuser. Type su and press Return

2. Type the following command on a terminal or on the command line of a shell to start the SunDiag program and bring up the window interface:

--------------------------------------
penguin# /usr/diag/sundiag/sundiag --------------------------------------

This command starts the basic SunDiag program. There are a wide variety of options to this command; see the SunDiag User's Guide shipped with SunOS documentation, for details on testing other than SBus Expansion Subsystem configurations.

3. Check the control panel on the right side of the SunDiag Window (see Figure C-2).

The bottom set of tests are dedicated to testing Sbus cards inserted into an SBus Expansion Subsystem. For each slot, the system provides a transparent mode check box and a non transparent mode check box. When you check a box in the Control Panel, the system will automatically display the device status in the System Status Panel.

4. To activate a test for a particular slot in the SBus Expansion Subsystem, move your cursor to the box and click the select button of your mouse.

You should not select transparent mode if the SBus Expansion slots are empty and you cannot select the nontransparent mode if you have one or more SBus cards in the SBus Expansion slots. These two types of modes are mutually exclusive and cannot be tested at the same time.

If a device is missing from the control panel, make sure that the device is properly installed.

Figure C-2 Typical SunDiag Control Panel

Note - SunDiag enables all available tests when it is invoked. Starting all available tests slows down your system considerably. You can avoid most problems by disabling the virtual memory (vmem) test.

5. After you select the conditions you want to test, click the Apply button and a check mark appears in the box. Then go to Step 7.

6. If you need to change the user tests, move the cursor to the User Tests box you want to discontinue and click SELECT to remove the check. Then move the cursor to the User Tests box you want to enable and click SELECT. To restore the default option settings, click the Reset button.

Note - You only need to change the options settings once. SunDiag continues to use your selections each time you run SunDiag.

7. Press the Start button at the top left of the control panel to start testing. Once testing starts, the Start button becomes the Stop button. Click Stop to halt testing.

SBus Expansion Subsystems Test-specific Information

The test name, xbtest, must match the name that appears on the SunDiag System Status panel.

xbtest Test Description

This test is divided into two modes of testing: transparent and nontransparent.

1. Use the transparent mode to test the system when you have devices, such as SBus cards installed in the SBus Expansion Subsystem.

After you invoke the xbtest the system will fork into two processes. The first, the parent process, will wait for the exit of the child process and then it will exit. The child process will issue an XAC_WAIT_FOR_ERROR ioctl call to the SBus Expansion Subsystem device driver and then enter a sleep mode.

When the device driver receives an error, it wakes up and passes an error packet to the child process. The child process dumps the contents of the error packet and then it exits also.

2. Use the nontransparent mode of the xbtest when there are no SBus card installed in the SBus Expansion Subsystem. It should never be run when there is an SBus card in the SBus Expansion Subsystem. As soon as you invoke the nontransparent mode, xbtest will call the following tests:

a. TEST_NON_TRANSPARENT

This test will perform the following steps:

i. switch to nontransparent mode

ii. check error status packet for any errors

iii. check xac and xbc registers and compare with expected values

iv. switch back to transparent mode

b. TEST_XAC_DVMA

This test will perform the following steps:

i. causes a DVMA transfer by asserting DVTE in control register 1 of the XAC

ii. compares error status packet with expected values

c. TEST_XBC_DVMA

This test will perform the following steps:

i. causes a DVMA transfer by asserting DVTE in control register 1 of XBC

ii. compares error status packet with expected values

d. TEST_XAC_DVMA_INT

This test will perform the following steps:

i. causes a DVMA transfer by asserting DVTE + INTT in control register 1 of XAC

ii. waits for interrupt

iii. compares error status packet with expected values

e. TEST_XBC_DVMA_INT

This test will perform the following steps:

i. causes a dvma transfer by asserting DVTE + INTT in control register 1 of XBC

ii. waits for interrupt

iii. compares error status packet with expected values

The system will repeat each test ten times. Time-out checks are included to avoid indefinite hangs. Failure of any test should not result in panics although full recovery is not guaranteed.

Options

You can change the time option settings on the option menu. The default time is set to 15 minutes. However, you can change this setting to a different value by entering the number in minutes.

xbtest On-line Command Arguments

You can run SunDiag in a single test mode. To do this, enter the following command in any shell.

--------------------------------------------------------------
/usr/diag/sundiag/xbtest D=devicename WE T=test_period SD standard_arguments --------------------------------------------------------------

----------------------------------------------------------------------------------
Arguments ----------------------------------------------------------------------------------

D=devicename devicename specifies the name of the device to be tested. Up to three SBus Expansion Subsystems are supported, so the choices are: xbox0, xbox1, and xbox2. WE Wait for Error. This option directs xbtest to run in transparent mode. This option will run xbtest continually until an error is returned, or until the time interval specified with the T=test_period option has ended. This argument cannot be used with the SD argument T=test_period Used only with the WE option. test_period specifies how long the test will run unless an error is returned, or the test is stopped manually. Substitute test_period with the number of minutes you want the test to run. The default is 15 minutes. SD Self Diagnostic test. This options directs xbtest to run in non transparent mode. This argument cannot be used with the WE argument. ----------------------------------------------------------------------------------

Note - The SunDiag standard arguments are described in Chapter 4 of the SunDiag User's Guide

.usertest File Command Line

The following are samples of .usertestcommands using the command arguments:

--------------------------------------------------
xbox0 transparent mode, xbtest, S D=xbox0 WE T=1 xbox1 nontransparent mode, xbtest, S D=xbox0 SD --------------------------------------------------

Error Messages

Whenever there are errors generated by the system, the SunDiag system will display the following error message in the lower left half of the SunDiag SUNVIEW window:

-----------------------------
error status dirty bit n -----------------------------

The system will return a hex character to indicate the error incurred. The hex number will replace the n with the character shown in the Bit Number column of table C-1.

Table C-1 SBus Expansion Subsystem Error Status Type Bit

--------------------------------------------------------
Error Message Mnemonic Bit Number --------------------------------------------------------

cable resend timeout error (dpr1) crlt1 e cable parity error (dpr1) cadp1 d expansion sbus read error (err ack) xrea c expansion sbus read error (rsvd ack) xrra b expansion sbus read error (late error) xrle a expansion sbus timeout error xbto 9 write 0 read error wrzr 8 buffer write error (err ack) bwea 7 buffer write error (rsvd ack) bwra 6 buffer write error (late error) bwle 5 cable resend timeout error (dpr0) crtl 4 cable ack timeout error cato 3 cable parity error (dpr0) cadp0 2 cable serial interrupt parity error csip 1 child not ready error cnrd 0 --------------------------------------------------------

The system may return one or more of the following error messages:

Table C-2 Miscellaneous SunDiag Errors

----------------------------------------------------------
probe xbox device failed

action_on_error ioctl command fail xbox status ioctl command fail can not switch to non transparent mode self diag test failed in TEST_NON_TRANSPARENT count %d self diag test failed in TEST_XAC_DVMA count %d self diag test failed in TEST_XBC_DVMA count %d self diag test failed in TEST_XAC_DVMA.int count %d self diag test failed in TEST_XBC_DVMA.int count %d XAC_TRANSPARANT ioctl command fail XAC_NON_TRANSPARANT ioctl command fail (device name): main:fork (device name) XAC_WAIT_FOR_ERROR ioctl command fail XAC_CLEAR_WAIT_FOR_ERROR ioctl command fail ----------------------------------------------------------

Note - Please refer to the SunDiag 2.4 User's Guide for specific information on testing SBus cards and disk drives. Test information may also be included in the documentation accompanying the particular device.