NAME crash - Analyze Linux crash dump data or a live system
SYNOPSIS crash [OPTION]... NAMELIST MEMORY-IMAGE[@ADDRESS] (dumpfile form) crash [OPTION]... [NAMELIST] (live system form)
DESCRIPTION Crash is a tool for interactively analyzing the state of the Linux sys- tem while it is running, or after a kernel crash has occurred and a core dump has been created by the netdump, diskdump, LKCD, kdump, xen- dump or kvmdump facilities. It is loosely based on the SVR4 UNIX crash command, but has been significantly enhanced by completely merging it with the gdb(1) debugger. The marriage of the two effectively combines the kernel-specific nature of the traditional UNIX crash utility with the source code level debugging capabilities of gdb(1).
In the dumpfile form, both a NAMELIST and a MEMORY-IMAGE argument must be entered. In the live system form, the NAMELIST argument must be entered if the kernels vmlinux file is not located in a known loca- tion, such as the /usr/lib/debug/lib/modules/<kernel-version> direc- tory.
The crash utility has also been extended to support the analysis of dumpfiles generated by a crash of the Xen hypervisor. In that case, the NAMELIST argument must be that of the xen-syms binary. Live system analysis is not supported for the Xen hypervisor.
The crash utility command set consists of common kernel core analysis tools such as kernel stack back traces of all processes, source code disassembly, formatted kernel structure and variable displays, virtual memory data, dumps of linked-lists, etc., along with several commands that delve deeper into specific kernel subsystems. Appropriate gdb commands may also be entered, which in turn are passed on to the gdb module for execution. If desired, commands may be placed in either a $HOME/.crashrc file and/or in a .crashrc file in the current directory. During initialization, the commands in $HOME/.crashrc are executed first, followed by those in the ./.crashrc file.
The crash utility is designed to be independent of Linux version depen- dencies. When new kernel source code impacts the correct functionality of crash and its command set, the utility will be updated to recognize new kernel code changes, while maintaining backwards compatibility with earlier releases.
OPTIONS NAMELIST This is a pathname to an uncompressed kernel image (a vmlinux file), or a Xen hypervisor image (a xen-syms file) which has been compiled with the "-g" option. If using the dumpfile form, a vmlinux file may be compressed in either gzip or bzip2 for- mats.
MEMORY-IMAGE[@ADDRESS] A kernel core dump file created by the netdump, diskdump, LKCD kdump, xendump or kvmdump facilities.
If a MEMORY-IMAGE argument is not entered, the session will be invoked on the live system, which typically requires root privi- leges because of the device file used to access system RAM. By default, /dev/crash will be used if it exists. If it does not exist, then /dev/mem will be used; but if the kernel has been configured with CONFIG_STRICT_DEVMEM, then /proc/kcore will be used. It is permissible to explicitly enter /dev/crash, /dev/mem or /proc/kcore.
An @ADDRESS value must be appended to the MEMORY-IMAGE if the dumpfile is a raw RAM dumpfile that has no header information describing the file contents. Multiple MEMORY-IMAGE@ADDRESS ordered pairs may be entered, with each dumpfile containing a contiguous block of RAM, where the ADDRESS value is the physical start address of the block expressed in hexadecimal. The physi- cal address value(s) will be used to create a temporary ELF header in /var/tmp, which will only exist during the crash ses- sion.
mapfile If the NAMELIST file is not the same kernel that is running (live system form), or the kernel that was running when the sys- tem crashed (dumpfile form), then the System.map file of the original kernel should be entered on the command line.
-h [option] --help [option] Without an option argument, display a crash usage help message. If the option argument is a crash command name, the help page for that command is displayed. If it is the string "input", a page describing the various crash command line input options is displayed. If it is the string "output", a page describing com- mand line output options is displayed. If it is the string "all", then all of the possible help messages are displayed. After the help message is displayed, crash exits.
-s Silently proceed directly to the "crash>" prompt without dis- playing any version, GPL, or crash initialization data during startup, and by default, runtime command output is not passed to any scrolling command.
-i file Execute the command(s) contained in file prior to displaying the "crash>" prompt for interactive user input.
-d num Set the internal debug level. The higher the number, the more debugging data will be printed when crash initializes and runs.
-S Use /boot/System.map as the mapfile.
-e vi | emacs Set the readline(3) command line editing mode to "vi" or "emacs". The default editing mode is "vi".
-f Force the usage of a compressed vmlinux file if its original name does not start with "vmlinux".
-k Indicate that the NAMELIST file is an LKCD "Kerntypes" debuginfo file.
-g [namelist] Determine if a vmlinux or xen-syms namelist file contains debug- ging data.
-t Display the system-crash timestamp and exit.
-L Attempt to lock all of its virtual address space into memory by calling mlockall(MCL_CURRENT|MCL_FUTURE) during initialization. If the system call fails, an error message will be displayed, but the session continues.
-c tty-device Open the tty-device as the console used for debug messages.
-p page-size If a processors page size cannot be determined by the dumpfile, and the processor default cannot be used, use page-size.
-o filename Only used with the MEMORY-IMAGE@ADDRESS format for raw RAM dump- files, specifies a filename of a new ELF vmcore that will be created and used as the dumpfile. It will be saved to allow future use as a standalone vmcore, replacing the original raw RAM dumpfile.
-m option=value --machdep option=value Pass an option and value pair to machine-dependent code. These architecture-specific option/pairs should only be required in very rare circumstances:
X86_64: phys_base=<physical-address> irq_eframe_link=<value> max_physmem_bits=<value> vm=orig (pre-2.6.11 virtual memory address ranges) vm=2.6.11 (2.6.11 and later virtual memory address ranges) vm=xen (Xen kernel virtual memory address ranges) vm=xen-rhel4 (RHEL4 Xen kernel virtual address ranges) PPC64: vm=orig vm=2.6.14 (4-level page tables) IA64: phys_start=<physical-address> init_stack_size=<size> vm=4l (4-level page tables) ARM: phys_base=<physical-address> ARM64: phys_offset=<physical-address>
-x Automatically load extension modules from a particular direc- tory. If a directory is specified in the CRASH_EXTENSIONS shell environment variable, then that directory will be used. Other- wise /usr/lib64/crash/extensions (64-bit architectures) or /usr/lib/crash/extensions (32-bit architectures) will be used; if they do not exist, then the ./extensions directory will be used.
--active Track only the active task on each cpu.
--buildinfo Display the crash binary s build date, the user ID of the builder, the hostname of the machine where the build was done, the target architecture, the version number, and the compiler version.
--memory_module modname Use the modname as an alternative kernel module to the crash.ko module that creates the /dev/crash device.
--memory_device device Use device as an alternative device to the /dev/crash, /dev/mem or /proc/kcore devices.
--log dumpfile Dump the contents of the kernel log buffer. A kernel namelist argument is not necessary, but the dumpfile must contain the VMCOREINFO data taken from the original /proc/vmcore ELF header.
--no_kallsyms Do not use kallsyms-generated symbol information contained within kernel module object files.
--no_modules Do not access or display any kernel module related information.
--no_ikconf Do not attempt to read configuration data that was built into kernels configured with CONFIG_IKCONFIG.
--no_data_debug Do not verify the validity of all structure member offsets and structure sizes that it uses.
--no_kmem_cache Do not initialize the kernel s slab cache infrastructure, and commands that use kmem_cache-related data will not work.
--no_elf_notes Do not use the registers from the ELF NT_PRSTATUS notes saved in a compressed kdump header for backtraces.
--kmem_cache_delay Delay the initialization of the kernels slab cache infrastruc- ture until it is required by a run-time command.
--readnow Pass this flag to the embedded gdb module, which will override its two-stage strategy that it uses for reading symbol tables from the NAMELIST.
--smp Specify that the system being analyzed is an SMP kernel.
-v --version Display the version of the crash utility, the version of the embedded gdb module, GPL information, and copyright notices.
--cpus number Specify the number of cpus in the SMP system being analyzed.
--osrelease dumpfile Display the OSRELEASE vmcoreinfo string from a kdump dumpfile header.
--hyper Force the session to be that of a Xen hypervisor.
--p2m_mfn pfn When a Xen Hypervisor or its dom0 kernel crashes, the dumpfile is typically analyzed with either the Xen hypervisor or the dom0 kernel. It is also possible to analyze any of the guest domU kernels if the pfn_to_mfn_list_list pfn value of the guest ker- nel is passed on the command line along with its NAMELIST and the dumpfile.
--xen_phys_start physical-address Supply the base physical address of the Xen hypervisors text and static data for older xendump dumpfiles that did not pass that information in the dumpfile header.
--zero_excluded If the makedumpfile(8) facility has filtered a compressed kdump dumpfile to exclude various types of non-essential pages, or has marked a compressed or ELF kdump dumpfile as incomplete due to an ENOSPC or other error during its creation, any attempt to read missing pages will fail. With this flag, reads from any of those pages will return zero-filled memory.
--no_panic Do not attempt to find the task that was running when the kernel crashed. Set the initial context to that of the "swapper" task on cpu 0.
--more Use /bin/more as the command output scroller, overriding the default of /usr/bin/less and any settings in either ./.crashrc or $HOME/.crashrc.
--less Use /usr/bin/less as the command output scroller, overriding any settings in either ./.crashrc or $HOME/.crashrc.
--hex Set the default command output radix to 16, overriding the default radix of 10, and any radix settings in either ./.crashrc or $HOME/.crashrc.
--dec Set the default command output radix to 10, overriding any radix settings in either ./.crashrc or $HOME/.crashrc. This is the default radix setting.
--CRASHPAGER Use the output paging command defined in the CRASHPAGER shell environment variable, overriding any settings in either ./.crashrc or $HOME/.crashrc.
--no_scroll Do not pass run-time command output to any scrolling command.
--no_strip Do not strip cloned kernel text symbol names.
--no_crashrc Do not execute the commands in either $HOME/.crashrc or ./.crashrc.
--mod directory When loading the debuginfo data of kernel modules with the mod -S command, search for their object files in directory instead of in the standard location.
--kaslr offset|auto If an x86_64 kernel was configured with CONFIG_RANDOMIZE_BASE, the offset value is equal to the difference between the symbol values compiled into the vmlinux file and their relocated KASLR values. If set to auto, the KASLR offset value will be automat- ically calculated.
--reloc size When analyzing live x86 kernels that were configured with a CON- FIG_PHYSICAL_START value that is larger than its CONFIG_PHYSI- CAL_ALIGN value, then it will be necessary to enter a relocation size equal to the difference between the two values.
--hash count Set the number of internal hash queue heads used for list gath- ering and verification. The default count is 32768.
--minimal Bring up a session that is restricted to the log, dis, rd, sym, eval, set and exit commands. This option may provide a way to extract some minimal/quick information from a corrupted or trun- cated dumpfile, or in situations where one of the several kernel subsystem initialization routines would abort the crash session.
--kvmhost [32|64] When examining an x86 KVM guest dumpfile, this option specifies that the KVM host that created the dumpfile was an x86 (32-bit) or an x86_64 (64-bit) machine, overriding the automatically determined value.
--kvmio <size> override the automatically-calculated KVM guest I/O hole size.
--offline [show|hide] Show or hide command output that is related to offline cpus. The default setting is show.
COMMANDS Each crash command generally falls into one of the following cate- gories:
Symbolic display Displays of kernel text/data, which take full advantage of the power of gdb to format and display data structures symbolically.
System state The majority of crash commands consist of a set of "kernel- aware" commands, which delve into various kernel subsystems on a system-wide or per-task basis.
Utility functions A set of useful helper commands serving various purposes, some simple, others quite powerful.
Session control Commands that control the crash session itself.
The following alphabetical list consists of a very simple overview of each crash command. However, since individual commands often have sev- eral options resulting in significantly different output, it is sug- gested that the full description of each command be viewed by executing crash -h <command>, or during a crash session by simply entering help command.
* "pointer to" is shorthand for either the struct or union com- mands. It displays the contents of a kernel structure or union.
alias creates a single-word alias for a command.
ascii displays an ascii chart or translates a numeric value into its ascii components.
bt displays a tasks kernel-stack backtrace. If it is given the -a option, it displays the stack traces of the active tasks on all CPUs. It is often used with the foreach command to display the backtraces of all tasks with one command.
btop translates a byte value (physical offset) to its page number.
dev displays data concerning the character and block device assign- ments, I/O port usage, I/O memory usage, and PCI device data.
dis disassembles memory, either entire kernel functions, from a location for a specified number of instructions, or from the start of a function up to a specified memory location.
eval evaluates an expression or numeric type and displays the result in hexadecimal, decimal, octal and binary.
exit causes crash to exit.
extend dynamically loads or unloads crash shared object extension mod- ules.
files displays information about open files in a context.
foreach repeats a specified command for the specified (or all) tasks in the system.
fuser displays the tasks using the specified file or socket.
gdb passes its argument to the embedded gdb module. It is useful for executing gdb commands that have the same name as crash com- mands.
help alone displays the command menu; if followed by a command name, a full description of a command, its options, and examples are displayed. Its output is far more complete and useful than this man page.
ipcs displays data about the System V IPC facilities.
irq displays data concerning interrupt request numbers and bottom- half interrupt handling.
kmem displays information about the use of kernel memory.
list displays the contents of a linked list.
log displays the kernel log_buf contents in chronological order.
mach displays data specific to the machine type.
mod displays information about the currently installed kernel mod- ules, or adds or deletes symbolic or debugging information about specified kernel modules.
mount displays information about the currently-mounted filesystems.
net display various network related data.
p passes its arguments to the gdb "print" command for evaluation and display.
ps displays process status for specified, or all, processes in the system.
pte translates the hexadecimal contents of a PTE into its physical page address and page bit settings.
ptob translates a page frame number to its byte value.
ptov translates a hexadecimal physical address into a kernel virtual address.
q is an alias for the "exit" command.
rd displays the contents of memory, with the output formatted in several different manners.
repeat repeats a command indefinitely, optionally delaying a given num- ber of seconds between each command execution.
runq displays the tasks on the run queue.
search searches a range of user or kernel memory space for given value.
set either sets a new context, or gets the current context for dis- play.
sig displays signal-handling data of one or more tasks.
struct displays either a structure definition or the contents of a ker- nel structure at a specified address.
swap displays information about each configured swap device.
sym translates a symbol to its virtual address, or a static kernel virtual address to its symbol -- or to a symbol-plus-offset value, if appropriate.
sys displays system-specific data.
task displays the contents of a task_struct.
tree displays the contents of a red-black tree or a radix tree.
timer displays the timer queue entries, both old- and new-style, in chronological order.
union is similar to the struct command, except that it works on kernel unions.
vm displays basic virtual memory information of a context.
vtop translates a user or kernel virtual address to its physical address.
waitq walks the wait queue list displaying the tasks which are blocked on the specified wait queue.
whatis displays the definition of structures, unions, typedefs or text/data symbols.
wr modifies the contents of memory on a live system. It can only be used if /dev/mem is the device file being used to access sys- tem RAM, and should obviously be used with great care.
When crash is invoked with a Xen hypervisor binary as the NAMELIST, the command set is slightly modified. The *, alias, ascii, bt, dis, eval, exit, extend, gdb, help, list, log, p, pte, rd, repeat, search, set, struct, sym, sys, union, whatis, wr and q commands are the same as above. The following commands are specific to the Xen hypervisor:
domain displays the contents of the domain structure for selected, or all, domains.
doms displays domain status for selected, or all, domains.
dumpinfo displays Xen dump information for selected, or all, cpus.
pcpus displays physical cpu information for selected, or all, cpus.
vcpus displays vcpu status for selected, or all, vcpus.
FILES .crashrc Initialization commands. The file can be located in the users HOME directory and/or the current directory. Commands found in the .crashrc file in the HOME directory are executed before those in the current directorys .crashrc file.
ENVIRONMENT EDITOR Command input is read using readline(3). If EDITOR is set to emacs or vi then suitable keybindings are used. If EDITOR is not set, then vi is used. This can be overridden by set vi or set emacs commands located in a .crashrc file, or by entering -e emacs on the crash command line.
CRASHPAGER If CRASHPAGER is set, its value is used as the name of the pro- gram to which command output will be sent. If not, then command output is sent to /usr/bin/less -E -X by default.
CRASH_MODULE_PATH Specifies an alternative directory tree to search for kernel module object files.
CRASH_EXTENSIONS Specifies a directory containing extension modules that will be loaded automatically if the -x command line option is used.
NOTES If crash does not work, look for a newer version: kernel evolution fre- quently makes crash updates necessary.
The command set scroll off will cause output to be sent directly to the terminal rather than through a paging program. This is useful, for example, if you are running crash in a window of emacs.
AUTHOR Dave Anderson <firstname.lastname@example.org> wrote crash.
Jay Fenlason <email@example.com> and Dave Anderson <anderson@red- hat.com> wrote this man page.
SEE ALSO The help command within crash provides more complete and accurate docu- mentation than this man page.
http://people.redhat.com/anderson - the home page of the crash utility.
netdump(8), gdb(1), makedumpfile(8)