User:Kangie/systemtap

From Gentoo Wiki
Jump to:navigation Jump to:search

SystemTap (stap) is a powerful tool that provides an infrastructure to simplify the gathering of information about the running Linux kernel or userspace programs[1]. It allows users to write and reuse simple scripts to deeply examine the activities of a running Linux system. These scripts can be designed to extract data, filter it, and summarize it quickly (and safely), enabling the diagnosis of complex performance (or even functional) problems.[2]

How it Works

SystemTap scripts are written in the SystemTap scripting language, are then compiled to C-code kernel modules and inserted into the kernel. This allows the scripts to instrument the execution of functions or statements in the kernel or user-space.

Usage

SystemTap provides a command line interface and a scripting language to examine the activities of a running Linux system, particularly the kernel, in fine detail.

Kernel

As SystemTap taps into the kernel at a low level, it requires that debug symbols be enabled (DWARF5, specifically) - for Gentoo this means reconfiguring the kernel[3].

For sys-kernel/gentoo-sources:

KERNEL debug sybmol generation
-> Kernel hacking                                                                                                                                                          
          -> Compile-time checks and compiler options
               -> Debug information (Generate DWARF Version 5 debuginfo)

For sys-kernel/gentoo-kernel:

FILE FILE /etc/kernel/config.d/99-debugsyms.config
CONFIG_DEBUG_INFO=y
CONFIG_DEBUG_INFO_DWARF5=y

Additional options that are probably already enabled: CONFIG_KPROBES, CONFIG_RELAY, CONFIG_DEBUG_FS, CONFIG_MODULES, CONFIG_MODULE_UNLOAD, CONFIG_UPROBES.

Users should try to reduce the number of modules / enabled options for an instrumented kernel where possible—CONFIG_DEBUG_INFO can multiply disk space usage. Be sure to leave CONFIG_DEBUG_INFO_SPLIT disabled; SystemTap doesn't handle split debuginfo yet.

Installation

root #emerge --ask dev-util/systemtap

Basic Usage

After installation, a basic probe to read the VFS can be performed to validate SystemTap functionality:

root #stap -v -e 'probe vfs.read {printf("read performed\n"); exit()}'
Pass 1: parsed user script and 45 library script(s) in 340usr/0sys/358real ms.
Pass 2: analyzed script: 1 probe(s), 1 function(s), 0 embed(s), 0 global(s) in 290usr/260sys/568real ms.
Pass 3: translated to C into "/tmp/stapiArgLX/stap_e5886fa50499994e6a87aacdc43cd392_399.c" in 490usr/430sys/938real ms.
Pass 4: compiled C into "stap_e5886fa50499994e6a87aacdc43cd392_399.ko" in 3310usr/430sys/3714real ms.
Pass 5: starting run.
read performed
Pass 5: run completed in 10usr/40sys/73real ms.

This command instructs SystemTap to print read performed and then exit properly once a virtual file system read is detected. If the SystemTap deployment was successful, it prints output similar to the above; the last three lines of the output (beginning with Pass 5) indicate that SystemTap was able to successfully create the instrumentation to probe the kernel, run the instrumentation, detect the event being probed (in this case, a virtual file system read), and execute a valid handler (print text then close it with no errors)[4].

Viewing Kernel Information

SystemTap can be used to view information about the kernel in various ways. For example, it can be used to identify the top system calls used by the system. It can also be used to determine which processes are performing the highest volume of system calls, providing more data in investigating systems for polling processes and other resource hogs.

Real-world Usage

This example describes using SystemTap to view the `inet_getname` function which was identified as the source of the following nfsd errors: nfsd: peername failed (err 107)!

Note
inet6_getname may cause failures if ipv6 is not enabled or loaded as a module. In that case just remove the line.
FILE nfsd_peername.stp
probe kernel.function("inet_getname").call,
      module("ipv6").function("inet6_getname").call
{
	if (execname() != "nfsd")
		next
	if ($peer == 1) {
		printf("%s %s -> %s addr: %s port: %d state: %s\n",
			tz_ctime(gettimeofday_s()),
			execname(),
			ppfunc(),
			format_ipaddr(__ip_sock_daddr($sock->sk), __ip_sock_family($sock->sk)),
			__tcp_sock_dport($sock->sk),
			tcp_sockstate_str(tcp_ts_get_info_state($sock->sk)));
	}
}

probe kernel.function("inet_getname").return,
      module("ipv6").function("inet6_getname").return
{
	if (execname() != "nfsd")
		next
	if ($peer == 1) {
		printf("%s %s <- %s ret: %d\n",
			tz_ctime(gettimeofday_s()),
			execname(),
			ppfunc(),
			$return)
	}
}

When run, the above script will log calls made to inet_gentame from binaries named nfsd, as well as the return value:

root #vim nfsd_peername.stp
root #stap -v nfsd_peername.stp
Pass 1: parsed user script and 114 library scripts using 57340virt/40276res/5700shr/35220data kb, in 120usr/10sys/130real ms.
Pass 2: analyzed script: 6 probes, 24 functions, 9 embeds, 3 globals using 228712virt/213276res/7376shr/206592data kb, in 2270usr/570sys/2737real ms.
Pass 3: translated to C into "/tmp/stapkilbvn/stap_aaaa6994e39808fec232416d081ab400_33413_src.c" using 228712virt/213468res/7568shr/206592data kb, in 10usr/20sys/32real ms.
Pass 4: compiled C into "stap_aaaa6994e39808fec232416d081ab400_33413.ko" in 7670usr/1440sys/8936real ms.
Pass 5: starting run.
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.84 port: 750 state: TCP_CLOSE_WAIT
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: 0
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.76 port: 940 state: TCP_CLOSE
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: -107
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.72 port: 671 state: TCP_CLOSE
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: -107
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.82 port: 742 state: TCP_CLOSE
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: -107
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.79 port: 749 state: TCP_CLOSE
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: -107
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.62 port: 886 state: TCP_ESTABLISHED
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: 0
Thu Dec  7 15:19:49 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.93 port: 861 state: TCP_ESTABLISHED
Thu Dec  7 15:19:49 2023 AEST nfsd <- inet_getname ret: 0
Thu Dec  7 15:19:50 2023 AEST nfsd -> inet_getname addr: 10.xxx.xxx.76 port: 940 state: TCP_ESTABLISHED
Thu Dec  7 15:19:50 2023 AEST nfsd <- inet_getname ret: 0

References

See Also