---
name: "memtool"
section: "1"
platform: "linux"
package: "memtool"
version: "2018.03.0-1"
release: "Debian 13 (Trixie)"
platforms: ["linux"]
url: "https://new.carta.tech/memtool"
---

# memtool(1)

display and modify memory

## NAME

memtool - display and modify memory

## SYNOPSIS

**memtool -V**\
 **memtool md** \[**-b**\|**-w**\|**-l**\|**-q**\] \[**-x**\] \[**-s** *filename*\] *region*\
 **memtool mw** \[**-b**\|**-w**\|**-l**\|**-q**\] \[**-d** *filename*\] *start* *data...*

## DESCRIPTION

memtool allows one to read and write regions of files. When applied to */dev/mem* (which is the default file) the regions represent memory mapped registers.

There are currently two subcommands: **mw** to write to memory/a file; and **md** to read from memory/a file.

Usually memtool operates on files (regular or devices) using [mmap(2)](/mmap/2.md). If *filename* is of the form mdio:*ethname*.*id* with *ethname* being the name of an ethernet device and *id* being an MDIO address, the phy with address *id* on the MDIO bus related to the ethernet device *ethname* is accessed instead. To prevent ambiguities when using the mmap access method, use mmap:*filename* as parameter.

Note that on some machines there are alignment restrictions that forbid for example to read a word from an address that is not word aligned. memtool doesn't try to be smart here but simply tries what is requested by the caller. This might result in all kind of errors which are not considered to be a bug of the tool, but either the machine or the caller.

## OPTIONS

- **-V**

  Dump memtool version and exit

## COMMON OPTIONS FOR SUBCOMMANDS

- **-b**

  Use byte wise (8-bit) access

- **-w**

  Use word wise (16-bit) access

- **-l**

  Use long wise (32-bit) access

- **-q**

  Use quad wise (64-bit) access

- **-d** *filename*

  *Write to* *filename*.

- **-s** *filename*

  *Read from* *filename*.

- **-x**

  *Swap bytes at output*

## REGIONS

*Memory regions can be specified in two different forms:*

- *start*\[**+***size*\]

  - *start***-***end*

  *start* specifies the byte offset of the region's beginning, i.e. the first included address. *size* specifies the size of the region in bytes and defaults to 0x100. Alternatively *end* specifies the offset of the last included address. Offsets can be specified in decimal or hexadecimal with a prefix 0x. Additionally you can use suffixes **G**, **M**, and **k**, which multiply by 1024^3, 1024^2, and 1024 respectively.