Unpacking U-Boot image file

[Grant Edwards]

On 2014-04-10, Rhodri James <rhodri at wildebst.org.uk> wrote:
> On Wed, 09 Apr 2014 18:18:56 +0100, Rustom Mody <rustompmody at gmail.com> wrote:
>> On Wednesday, April 9, 2014 9:36:40 PM UTC+5:30, trewio wrote:
>>
>>> How to extract files from U-Boot image file, LZMA-compressed?
>>>
>>> Is there a Python script that can do this properly?
>>
>> For lzma theres this (recent) python library
>> https://docs.python.org/dev/library/lzma.html
>>
>> Though you might just be better off with the command-line xz unxz etc
>>
>> After that.. whats the U-boot format?
>
> The Fine Manual (http://www.denx.de/wiki/view/DULG/UBootImages) isn't very  
> forthcoming, sadly;
>
> "U-Boot operates on "image" files which can be basically anything,  
> preceeded by a special header; see the definitions in include/image.h for  
> details; basically, the header defines the following image properties
>
> * Target Operating System
> * Target CPU Architecture
> * Compression Type
> * Load Address
> * Entry Point
> * Image Name
> * Image Timestamp"
>
> I suspect taking this apart may be a little involved.

Not really.  For the format you're talking about (single-file uImage),
you can extract the data like so:

  data = open("foobar.uImage","rb").read()[64:]

Then just uncompress 'data' as desired.

If you want to look at individual fields in the heaer, here's the
structure:   

#define IH_MAGIC    0x27051956    /* Image Magic Number     */
#define IH_NMLEN    32            /* Image Name Length      */

typedef struct image_header 
{
  uint32_t        ih_magic;       /* Image Header Magic Number    */
  uint32_t        ih_hcrc;        /* Image Header CRC Checksum    */
  uint32_t        ih_time;        /* Image Creation Timestamp     */
  uint32_t        ih_size;        /* Image Data Size              */
  uint32_t        ih_load;        /* Data  Load  Address          */
  uint32_t        ih_ep;          /* Entry Point Address          */
  uint32_t        ih_dcrc;        /* Image Data CRC Checksum      */
  uint8_t         ih_os;          /* Operating System             */
  uint8_t         ih_arch;        /* CPU architecture             */
  uint8_t         ih_type;        /* Image Type                   */
  uint8_t         ih_comp;        /* Compression Type             */
  uint8_t         ih_name[IH_NMLEN];      /* Image Name           */
} image_header_t;

Header fields are all in network byte order (bit-endian).

However, there's also a "multi-file" uImage format that I suspect is
what the OP is referring to (since he says he needs to extract "files"
from a uImage file).

http://www.isysop.com/unpacking-and-repacking-u-boot-uimage-files/

In that case, the "Image Type" field will be set to 4, and the 'data'
field starts with a list of 32-bit image size values (network byte
order, terminated by 0x00000000).  That list is then followed by the
image data.

Here's a Python 2.7 program that reads the uImage header, prints some
stuff and then reads the data for the image(s).  What to _do_ with the
data for the image(s) is left as an exercise for the gentle reader.

-------------------------------------8<-------------------------------------
#!/usr/bin/python

import sys,struct

f = open(sys.argv[1],"rb")

(ih_magic, ih_hcrc, ih_time, ih_size, ih_load, ih_ep, ih_dcrc, 
 ih_os, ih_arch, ih_type, ih_comp, ih_name) \
    = struct.unpack("!IIIIIIIBBBB32s",f.read(64))

print "magic=%08x type=%d size=%d name='%s'" % (ih_magic, ih_type, ih_size, ih_name)

if ih_type == 4:
    # multi-file: read image lengths from data
    imagesizes = []
    while 1:
        size, = struct.unpack("!I",f.read(4))
        if size == 0:
            break
        imagesizes.append(size)
else:
    # single-file
    imagesizes = [ih_size]

print imagesizes

images = [f.read(size) for size in imagesizes]

print [len(image) for image in images]

extradata = f.read() 

if extradata:
    print "%d bytes of extra data" % len(extradata)

-------------------------------------8<-------------------------------------










-- 
Grant Edwards               grant.b.edwards        Yow! The Korean War must
                                  at               have been fun.
                              gmail.com