Browse Teach Curate

Reading a Kernel Oops and Decoding a Stack Trace

When kernel code does something impossible, the kernel prints a page of registers and addresses called an oops, then may kill the task or stop entirely. Those raw addresses mean little until a script turns them back into file names and line numbers using your build. Reading an oops tells you which function died and how it got there. Without decoding it you are left with hexadecimal and a guess.

Questions this Concept answers

Why do the raw addresses in an oops need decoding before they are useful?

jeremy Video

Linux Kernel Debugging Tools: Kernel Oops, KGDB, and Ftrace

Kernel debugging rests on matching a bug's symptom class (crash, lockup, logic error, resource leak, performance regression) to the debugging technique best suited to it — background knowledge, post-…