Debugging in Linux

One way to debug in Linux would be to use gdb. This program can run and debug most programs you can write.

setting up to debug a program

First, you need to build the binary in such a way that allows the program to be debugged. This can be done by passing gcc or g++ the -g flag.

attach

attach [PID]

This command allows gdb to attach itself to another process.

[c]ontinue

This continues and stops at the next break point.

[b]ack[t]race

This prints the back trace (stack). The call

[b]reak

break [function|label|filename:number|linenumber|address]

• filename:number - This will set a break point at the line number in a givin file.
• label - This will set a break point at theline fallowing the label
• filename:number - This will set a break point at the line number in a givin file.

[d]elete

delete [break_point_number]

This deletes a break point number.

disassemble

disassemble [label]

This command disassembles the assembly code at a certain label.

[i]nfo

info [name]

• [b]reak - list all break points
• [fu]nctions - List all of the function signatures
• [f]rame - Lists the current stack frame information
• [r]egisters - list all of the registers
• [s]tack - This prints all of the text

layout

layout [layout name]

This changes the current layout. There are 4 different layouts src, asm, split and regs

• src - soure code split with the gdb command prompt
• asm - the ows the assembly and the gdb command prompt
• split - This shows the both the source code, assemby and the gdb command prompt.
• regs - THis displlays all of the CPU registers

[n]ext

next [count]

This runs how ever many lines of code is specified. This will step over any functions.

[r]un

run args1 arg2 ... < file_in.txt > file_out.txt

This command runt the binary. Pass any arguments to this and it passes them on to the child program

[s]tep

step count

This single steps to the next line or the next n instructions