HP OpenVMS System Analysis Tools Manual

Contents

Index

2.3 Analyzing a System Dump

SDA performs certain tasks before bringing a dump into memory, presenting its initial displays, and accepting command input. These tasks include the following:

Verifying that the process invoking it is suitably privileged to read the dump file
Using RMS to read in pages from the dump file
Building the SDA symbol table from the files SDA$READ_DIR:SYS$BASE_IMAGE.EXE and SDA$READ_DIR:REQSYSDEF.STB
Executing the commands in the SDA initialization file

For detailed information on investigating system failures, see Section 2.7.

2.3.1 Requirements

To analyze a dump file, your process must have read access both to the file that contains the dump and to copies of SDA$READ_DIR:SYS$BASE_IMAGE.EXE and SDA$READ_DIR:REQSYSDEF.STB (the required subset of the symbols in the file SYSDEF.STB). SDA reads these tables by default.

2.3.2 Invoking SDA

If your process can access the files listed in Section 2.3.1, you can issue the DCL command ANALYZE/CRASH_DUMP to invoke SDA. If you do not specify the name of a dump file in the command, SDA prompts you:

$ ANALYZE/CRASH_DUMP _Dump File:

The default file specification is as follows:

SYS$DISK:[default-dir]SYSDUMP.DMP

SYS$DISK and [default-dir] represent the disk and directory specified in your last SET DEFAULT command.

If you are rebooting after a system failure, SDA is automatically invoked. See Section 2.2.3.

2.3.3 Mapping the Contents of the Dump File

SDA first attempts to map the contents of memory as stored in the specified dump file. To do this, it must first locate the page tables for system space among its contents. The system page tables contain one entry for each page of system virtual address space.

If SDA cannot find the system page tables in the dump file, it displays the following message:
%SDA-E-SPTNOTFND, system page table not found in dump file
If that error message is displayed, you cannot analyze the crash dump, but must take steps to ensure that any subsequent dump can be analyzed. To do this, you must either adjust the DUMPSTYLE system parameter as discussed in Section 2.2.1.1 or increase the size of the dump file as indicated in Section 2.2.1.3.

If SDA finds the system page tables in an incomplete dump, the following message is displayed:

%SDA-W-SHORTDUMP, dump file was n blocks too small when dump written; analysis may not be possible

Under certain conditions, some memory locations might not be saved in the system dump file. Additionally, if a bugcheck occurs during system initialization, the contents of the register display may be unreliable. The symptom of such a bugcheck is a SHOW SUMMARY display that shows no processes or only the swapper process.

If you use an SDA command to access a virtual address that has no corresponding physical address, SDA generates the following error message:

%SDA-E-NOTINPHYS, 'location': virtual data not in physical memory

When analyzing a selective dump file, if you use an SDA command to access a virtual address that has a corresponding physical address not saved in the dump file, SDA generates one of the following error messages:

%SDA-E-MEMNOTSVD, memory not saved in the dump file %SDA-E-NOREAD, unable to access location n

2.3.4 Building the SDA Symbol Table

After locating and reading the system dump file, SDA attempts to read the system symbol table file into the SDA symbol table. If SDA cannot find SDA$READ_DIR:SYS$BASE_IMAGE.EXE---or is given a file that is not a system symbol table in the /SYMBOL qualifier to the ANALYZE command---it displays a fatal error and exits. SDA also reads into its symbol table a subset of SDA$READ_DIR:SYSDEF.STB, called SDA$READ_DIR:REQSYSDEF.STB. This subset provides SDA with the information needed to access some of the data structures in the dump.

When SDA finishes building its symbol table, SDA displays a message identifying itself and the immediate cause of the system failure. In the following example, the cause of the system failure was the deallocation of a bad page file address.

OpenVMS Alpha System Dump Analyzer Dump taken on 27-MAR-1993 11:22:33.92 BADPAGFILD, Bad page file address deallocated

2.3.5 Executing the SDA Initialization File (SDA$INIT)

After displaying the system failure summary, SDA executes the commands in the SDA initialization file, if you have established one. SDA refers to its initialization file by using the logical name SDA$INIT. If SDA cannot find the file defined as SDA$INIT, it searches for the file SYS$LOGIN:SDA.INIT.

This initialization file can contain SDA commands that read symbols into SDA's symbol table, define keys, establish a log of SDA commands and output, or perform other tasks. For instance, you may want to use an SDA initialization file to augment SDA's symbol table with definitions helpful in locating system code. If you issue the following command, SDA includes those symbols that define many of the system's data structures, including those in the I/O database:

READ SDA$READ_DIR:filename

You may also find it helpful to define those symbols that identify the modules in the images that make up the executive by issuing the following command:

READ/EXECUTIVE SDA$READ_DIR:

After SDA has executed the commands in the initialization file, it displays its prompt as follows:

SDA>

This prompt indicates that you can use SDA interactively and enter SDA commands.

An SDA initialization file may invoke a command procedure with the @ command. However, such command procedures cannot invoke other command procedures.

2.4 Analyzing a Running System

Occasionally, OpenVMS encounters an internal problem that hinders system performance without causing a system failure. By allowing you to examine the running system, SDA enables you to search for the solution without disturbing the operating system. For example, you may be able to use SDA to examine the stack and memory of a process that is stalled in a scheduler state, such as a miscellaneous wait (MWAIT) or a suspended (SUSP) state.

If your process has change-mode-to-kernel (CMKRNL) privilege, you can invoke SDA to examine the system. Use the following DCL command:

$ ANALYZE/SYSTEM

SDA attempts to load SDA$READ_DIR:SYS$BASE_IMAGE.EXE and SDA$READ_DIR:REQSYSDEF.STB. It then executes the contents of any existing SDA initialization file, as it does when invoked to analyze a crash dump (see Sections 2.3.4 and 2.3.5, respectively). SDA subsequently displays its identification message and prompt, as follows:

OpenVMS Alpha System Analyzer SDA>

This prompt indicates that you can use SDA interactively and enter SDA commands. When analyzing a running system, SDA sets its process context to that of the process running SDA.

If you are analyzing a running system, consider the following:

When used in this mode, SDA does not map the entire system, but instead retrieves only the information it needs to process each individual command. To update any given display, you must reissue the previous command.

Caution:

When using SDA to analyze a running system, carefully interpret its displays. Because system states change frequently, it is possible that the information SDA displays may be inconsistent with the current state of the system.

Certain SDA commands are illegal in this mode, such as SET CPU. Use of these commands results in the following error message:
%SDA-E-CMDNOTVLD, command not valid on the running system
The SHOW CRASH command, although valid, does not display the contents of any of the processor's set of hardware registers.

2.5 SDA Context

When you invoke SDA to analyze either a crash dump or a running system, SDA establishes a default context for itself from which it interprets certain commands.

When you are analyzing a uniprocessor system, SDA's context is solely process context, which means SDA can interpret its process-specific commands in the context of either the process current on the uniprocessor or some other process in another scheduling state. When SDA is initially invoked to analyze a crash dump, SDA's process context defaults to that of the process that was current at the time of the system failure. When you invoke SDA to analyze a running system, SDA's process context defaults to that of the current process, that is, the one executing SDA. To change SDA's process context, issue any of the following commands:

SET PROCESS process-name
SET PROCESS/ADDRESS=pcb-address
SET PROCESS/INDEX=nn
SET PROCESS/NEXT
SET PROCESS/SYSTEM
SHOW PROCESS process-name
SHOW PROCESS/ADDRESS=pcb-address
SHOW PROCESS/INDEX=nn
SHOW PROCESS/SYSTEM
SHOW PROCESS/NEXT

When you invoke SDA to analyze a crash dump from a multiprocessing system with more than one active CPU, SDA maintains a second dimension of context---its CPU context---that allows it to display certain processor-specific information. This information includes the reason for the bugcheck exception, the currently executing process, the current IPL, and the spinlocks owned by the processor. When you invoke SDA to analyze a multiprocessor's crash dump, its CPU context defaults to that of the processor that induced the system failure. When you are analyzing a running system, CPU context is not accessible to SDA. Therefore, the SET CPU command is not permitted.

You can change the SDA CPU context by using any of the following commands:

SET CPU cpu-id
SHOW CPU cpu-id
SHOW CRASH
SHOW MACHINE_CHECK cpu-id

Changing CPU context involves an implicit change in process context in either of the following ways:

If there is a current process on the CPU made current, SDA process context is changed to that of that CPU's current process.
If there is no current process on the CPU made current, SDA process context is undefined and no process-specific information is available until SDA process context is set to that of a specific process.

Changing process context requires a switch of CPU context as well. For instance, when you issue a SET PROCESS command, SDA automatically changes its CPU context to that of the CPU on which that process was most recently current. The following commands can have this effect:

SET PROCESS process-name
SET PROCESS/ADDRESS=pcb-address
SET PROCESS/INDEX=nn
SET PROCESS/NEXT
SHOW PROCESS process-name
SHOW PROCESS/ADDRESS=pcb-address
SHOW PROCESS/INDEX=nn
SHOW PROCESS/NEXT

2.6 SDA Command Format

The following sections describe the format of SDA commands and the expressions you can use with SDA commands.

SDA uses a command format similar to that used by the DCL interpreter. Issue commands in the following format:

command-name[/qualifier...] [parameter][/qualifier...] [!comment]

The command-name is an SDA command. Each command tells the utility to perform a function. Commands can consist of one or more words, and can be abbreviated to the number of characters that make the command unique. For example, SH stands for SHOW.

The parameter is the target of the command. For example, SHOW PROCESS RUSKIN tells SDA to display the context of the process RUSKIN. The command EXAMINE 80104CD0;40 displays the contents of 40 bytes of memory, beginning with location 80104CD0.

When you supply part of a file specification as a parameter, SDA assumes default values for the omitted portions of the specification. The default device is SYS$DISK, the device specified in your most recent SET DEFAULT command. The default directory is the directory specified in the most recent SET DEFAULT command. See the HP OpenVMS DCL Dictionary for a description of the DCL command SET DEFAULT.

The qualifier modifies the action of an SDA command. A qualifier is always preceded by a slash (/). Several qualifiers can follow a single parameter or command name, but each must be preceded by a slash. Qualifiers can be abbreviated to the shortest string of characters that uniquely identifies the qualifier.

The comment consists of text that describes the command; this comment is not actually part of the command. Comments are useful for documenting SDA command procedures. When executing a command, SDA ignores the exclamation point and all characters that follow it on the same line.

2.6.1 Using Expressions and Operators

You can use expressions as parameters for some SDA commands, such as SEARCH and EXAMINE. To create expressions, use any of the following elements:

Numerals
Radix operators
Arithmetic and logical operators
Precedence operators
Symbols

Numerals are one possible component of an expression. The following sections describe the use of the other components.

2.6.1.1 Radix Operators

Radix operators determine which numeric base SDA uses to evaluate expressions. You can use one of the three radix operators to specify the radix of the numeric expression that follows the operator:

^X (hexadecimal)
^O (octal)
^D (decimal)

The default radix is hexadecimal. SDA displays hexadecimal numbers with leading zeros and decimal numbers with leading spaces.

2.6.1.2 Arithmetic and Logical Operators

There are two types of arithmetic and logical operators, both of which are listed in Table 2-3.

Unary operators affect the value of the expression that follows them.
Binary operators combine the operands that precede and follow them.

In evaluating expressions containing binary operators, SDA performs logical AND, OR, and XOR operations, and multiplication, division, and arithmetic shifting before addition and subtraction. Note that the SDA arithmetic operators perform integer arithmetic on 64-bit operands.

Table 2-3 SDA Operators
Operator Action

Unary Operators

# Performs a logical NOT of the expression.

+ Makes the value of the expression positive.

-- Makes the value of the expression negative.

@ Evaluates the following expression as an address, then uses the contents of that address as its value.

^Q Specifies that the size of the field to be used as an address is a quadword when used with the unary operator @ ¹.

^L Specifies that the size of the field to be used as an address is a longword when used with the unary operator @ ¹.

^W Specifies that the size of the field to be used as an address is a word when used with the unary operator @ ¹.

^B Specifies that the size of the field to be used as an address is a byte when used with the unary operator @ ¹.

^P Specifies a physical address when used with the unary operator @ ¹.

^V Specifies a virtual address when used with the unary operator @ ¹.

G Adds FFFFFFFF 80000000 ₁₆ to the value of the expression ².

H Adds 7FFE0000 ₁₆ to the value of the expression ³.

I Fills the leading digits of the following hexadecimal number with hex value of F. For example:

SDA> eval i80000000
Hex = FFFFFFFF.80000000 Decimal = -2147483648 G
SYS$PUBLIC_VECTORS_NPRO

Binary Operators

+ Addition

-- Subtraction

* Multiplication

& Logical AND

| Logical OR

\ Logical XOR

/ Division ⁴

@ Arithmetic shifting

"." Catenates two 32-bit values into a 64-bit value. For example:
SDA> eval fe.50000
Hex = 000000FE00050000 Decimal = 1090922020864

**Table 2-3 SDA Operators**
Operator	Action
Unary Operators
#	Performs a logical NOT of the expression.
+	Makes the value of the expression positive.
--	Makes the value of the expression negative.
@	Evaluates the following expression as an address, then uses the contents of that address as its value.
^Q	Specifies that the size of the field to be used as an address is a quadword when used with the unary operator @ ¹.
^L	Specifies that the size of the field to be used as an address is a longword when used with the unary operator @ ¹.
^W	Specifies that the size of the field to be used as an address is a word when used with the unary operator @ ¹.
^B	Specifies that the size of the field to be used as an address is a byte when used with the unary operator @ ¹.
^P	Specifies a physical address when used with the unary operator @ ¹.
^V	Specifies a virtual address when used with the unary operator @ ¹.
G	Adds FFFFFFFF 80000000 ₁₆ to the value of the expression ².
H	Adds 7FFE0000 ₁₆ to the value of the expression ³.
I	Fills the leading digits of the following hexadecimal number with hex value of F. For example:
SDA> eval i80000000 Hex = FFFFFFFF.80000000 Decimal = -2147483648 G SYS$PUBLIC_VECTORS_NPRO
Binary Operators
+	Addition
--	Subtraction
*	Multiplication
&	Logical AND
\|	Logical OR
\	Logical XOR
/	Division ⁴
@	Arithmetic shifting
"."	Catenates two 32-bit values into a 64-bit value. For example: SDA> eval fe.50000 Hex = 000000FE00050000 Decimal = 1090922020864

¹The command SET FETCH can be used to change the default FETCH size and/or access method. See the SET FETCH command description in Chapter 4 for more details and examples.
²The unary operator G corresponds to the first virtual address in S0 system space. For example, the expression GD40 can be used to represent the address FFFFFFFF 80000D40₁₆.
³The unary operator H corresponds to a convenient base address in P1 space (7FFE0000₁₆). You can therefore refer to an address such as 7FFE2A64₁₆ as H2A64.
⁴In division, SDA truncates the quotient to an integer, if necessary, and does not retain a remainder.

2.6.1.3 Precedence Operators

SDA uses parentheses as precedence operators. Expressions enclosed in parentheses are evaluated first. SDA evaluates nested parenthetical expressions from the innermost to the outermost pairs of parentheses.

2.6.1.4 Symbols

An SDA symbol can represent several value types. It can represent a constant, a data address, a procedure descriptor address, or a routine address. Constants are usually offsets of a particular field in a data structure; however, they can also represent constant values such as the BUG$_xxx symbols.

Symbols are composed of up to 31 letters and numbers, and can include the dollar sign ($) and underscore (_) characters. When you invoke SDA, it reads in the global symbols from the symbols table section of SYS$BASE_IMAGE.EXE, and from REQSYSDEF.STB, a required subset of the symbols in the file SYSDEF.STB. You can add other symbols to SDA's symbol table by using the DEFINE and READ commands.

All address symbols identify memory locations. SDA generally does not distinguish among different types of address symbols. However, for a symbol identified as the name of a procedure descriptor, SDA takes an additional step of creating an associated symbol to name the code entry point address of the procedure. It forms the code entry point symbol name by appending _C to the name of the procedure descriptor.

Also, SDA substitutes the code entry point symbol name for the procedure descriptor symbol when you enter the following command:

SDA> EXAMINE/INSTRUCTION procedure descriptor

For example, enter the following command:

SDA> EXAMINE/INSTRUCTION SCH$QAST

SDA displays the following information:

SCH$QAST_C: SUBQ SP,#X40,SP

Now enter the EXAMINE command but do not specify the /INSTRUCTION qualifier, as follows:

SDA> EXAMINE SCH$QAST

SDA displays the following information:

SCH$QAST: 0000002C.00003009 ".0..,..."

This display shows the contents of the first two longwords of the procedure descriptor.

Note that there are no routine address symbols on Alpha systems, except for those in MACRO-64 assembly language modules. Therefore, SDA creates a routine address symbol for every procedure descriptor it has in its symbol table. The new symbol name is the same as for the procedure descriptor except that it has an _C appended to the end of the name.

Sources for SDA Symbols

SDA obtains its information from the following:

Images (.EXE files)
Image symbol table files (.STB files)
Object files

SDA also defines symbols to access registers and to access common data structures.

The only images with symbols are shareable images and executive images. These images contain only universal symbols, such as constants and addresses.

The image symbol table files are produced by the linker with the /SYMBOLS qualifier. These files normally only contain universal symbols, as do the executable images. However, if the SYMBOL_TABLE=GLOBALS linker option is specified, the .STB file also contains all global symbols defined in the image. See the HP OpenVMS Linker Utility Manual for more information.

Object files can contain global constant values. An object file used with SDA typically contains symbol definitions for data structure fields. Such an object file can be generated by compiling a MACRO-32 source module that invokes specific macros. The macros, which are typically defined in SYS$LIBRARY:LIB.MLB or STARLET.MLB, define symbols that correspond to data structure field offsets. The macro $UCBDEF, for example, defines offsets for fields within a unit control block (UCB). OpenVMS Alpha and I64 provide several such object modules in SDA$READ_DIR, as listed in Table 2-4. For compatibility with OpenVMS VAX, the modules' file types have been renamed to .STB.

Table 2-4 Modules Containing SDA Global Symbols and Data Structures
File Contents

DCLDEF.STB Symbols for the DCL interpreter

DECDTMDEF.STB Symbols for transaction processing

GLXDEF.STB Symbols for OpenVMS Galaxy data structures

IMGDEF.STB Symbols for the image activator

IODEF.STB I/O database structure symbols

NETDEF.STB Symbols for DECnet data structures

REQSYSDEF.STB Required symbols for SDA

RMSDEF.STB Symbols that define RMS internal and user data structures and RMS$_ xxx completion codes

SCSDEF.STB Symbols that define data structures for system communications services

SYSDEF.STB Symbols that define system data structures, including the I/O database

TCPIP$NET_GLOBALS.STB ¹ Data structure definitions for TCP/IP internet driver, execlet, and ACP data structures

TCPIP$NFS_GLOBALS.STB ¹ Data structure definitions for TCP/IP NFS server

TCPIP$PROXY_GLOBALS.STB ¹ Data structure definitions for TCP/IP proxy execlet

TCPIP$PWIP_GLOBALS.STB ¹ Data structure definitions for TCP/IP PWIP driver, and ACP data structures

TCPIP$TN_GLOBALS.STB ¹ Data structure definitions for TCP/IP TELNET/RLOGIN server driver data structures

**Table 2-4 Modules Containing SDA Global Symbols and Data Structures**
File	Contents
DCLDEF.STB	Symbols for the DCL interpreter
DECDTMDEF.STB	Symbols for transaction processing
GLXDEF.STB	Symbols for OpenVMS Galaxy data structures
IMGDEF.STB	Symbols for the image activator
IODEF.STB	I/O database structure symbols
NETDEF.STB	Symbols for DECnet data structures
REQSYSDEF.STB	Required symbols for SDA
RMSDEF.STB	Symbols that define RMS internal and user data structures and RMS$_ xxx completion codes
SCSDEF.STB	Symbols that define data structures for system communications services
SYSDEF.STB	Symbols that define system data structures, including the I/O database
TCPIP$NET_GLOBALS.STB ¹	Data structure definitions for TCP/IP internet driver, execlet, and ACP data structures
TCPIP$NFS_GLOBALS.STB ¹	Data structure definitions for TCP/IP NFS server
TCPIP$PROXY_GLOBALS.STB ¹	Data structure definitions for TCP/IP proxy execlet
TCPIP$PWIP_GLOBALS.STB ¹	Data structure definitions for TCP/IP PWIP driver, and ACP data structures
TCPIP$TN_GLOBALS.STB ¹	Data structure definitions for TCP/IP TELNET/RLOGIN server driver data structures

¹Only available if TCP/IP has been installed. These are found in SYS$SYSTEM, so that all files are not automatically read in when you issue a READ/EXEC command.

Contents

Index