Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Are tracepoints redundant in Linux kernel after kprobes support for ftrace?

What are the use-cases for using tracepoint events when kprobe events support for ftrace is available in Linux kernel? It seems everything that is possible to be done using 'tracepoint events' is possible using kprobe events, since one can set up a kprobe event to at the same point where tracepoint event is available.

Am I missing something?

like image 241
gabhijit Avatar asked Aug 10 '17 15:08

gabhijit


People also ask

How does Kprobes work?

kprobes (kernel probes) is Linux kernel mechanism for dynamic tracing. It allows you to insert breakpoint at almost any kernel function, invoke your handler and then continue executing. It works by runtime patching kernel image with so-called kernel probe/kprobe - see struct kprobe .

What is a kernel probe?

Kernel probes are a set of tools to collect Linux kernel debugging and performance information. Developers and system administrators usually use them either to debug the kernel, or to find system performance bottlenecks. The reported data can then be used to tune the system for better performance.

How does a Kprobe and return probe works?

A kprobe can be inserted on virtually any instruction in the kernel. A return probe fires when a specified function returns. In the typical case, Kprobes-based instrumentation is packaged as a kernel module. The module's init function installs (“registers”) one or more probes, and the exit function unregisters them.

What is Kprobe and Uprobe?

Uprobes are kind of like kprobes, except that instead of instrumenting a kernel function you're instrumenting userspace functions (like malloc).


1 Answers

Since kprobes trace arbitrary functions, tools relying on them may easily break from one Linux version to the next. For instance, the name of the function or one of its arguments can be changed, or the whole function might be removed. This sort of change happens frequently and can break kprobe-based tools.

Conversely, tracepoints are more stable. They should remain mostly the same and provide the same information. In addition, they are documented; you can find the type and location of information provided by tracepoints in /sys/kernel/debug/tracing:

# cat /sys/kernel/debug/tracing/events/skb/kfree_skb/format 
name: kfree_skb
ID: 1122
format:
    field:unsigned short common_type;   offset:0;   size:2; signed:0;
    field:unsigned char common_flags;   offset:2;   size:1; signed:0;
    field:unsigned char common_preempt_count;   offset:3;   size:1; signed:0;
    field:int common_pid;   offset:4;   size:4; signed:1;

    field:void * skbaddr;   offset:8;   size:8; signed:0;
    field:void * location;  offset:16;  size:8; signed:0;
    field:unsigned short protocol;  offset:24;  size:2; signed:0;

print fmt: "skbaddr=%p protocol=%u location=%p", REC->skbaddr, REC->protocol, REC->location
like image 146
pchaigno Avatar answered Sep 25 '22 06:09

pchaigno