mirror of
https://github.com/frida/frida-tools.git
synced 2026-11-07 09:35:59 -10:00
It now writes all events to the ringbuffer, so the file no longer contains start and compile messages. Change its magic to "ITR2", and drop the basic block count from the progress display.
1.4 KiB
1.4 KiB
Frida CLI tools
CLI tools for Frida.
Installing Fish completions
Currently there is no mechanism to install Fish completions through the setup.py
script so if you want to have completions in Fish you will have to install it
manually. Unless you've changed your XDG_CONFIG_HOME location, you should just
copy the completion file into ~/.config/fish/completions like so:
cp completions/frida.fish ~/.config/fish/completions
frida-itrace file format
File starts with a 4-byte magic: "ITR2"
Then, following that, there are two different types of records, MESSAGE and CHUNK. Each record starts with a big-endian uint32 specifying the type of record, where 1 means MESSAGE, 2 means CHUNK.
MESSAGE
length: uint32 (big-endian)message: JSON, UTF-8 encodeddata_size: uint32 (big-endian)data_values: uint8[data_size]
There are two different kinds of MESSAGEs:
- "itrace:lost": Signals that records were dropped because the buffer was full. The JSON portion contains the total number of records lost so far.
- "itrace:end": Signals that the endpoint was reached, when specifying a range with an end address included.
CHUNK
size: uint32 (big-endian)data: uint8[size]
The CHUNK records combine to the stream of trace records and events written to frida-itrace's ringbuffer, as documented in its buffer format.