Help with cross-compiling tools?

[brianinmaine]

I've been reading and reading and... I just don't seem to understand what I should do here:

I've looked at CodeSourcery and Linero and Scratchbox2 and Buildroot?? I've downloaded tons of stuff and am still trying to figure out what is best.

What is the easiest and/or most used path here?

Should I use the filesystem from the Kindle as a skeleton or use ???

I have a debian arm chroot working with qemu-arm-static so I can apt-get update in it, I thought I could compile in that? I guess not?

Many of the pages linked on this site for development are 404'd. Just some general guidance on what you use, and or maybe a config file or 2?

Thanks, as I have no more hair to pull out!

[ixtab]

I'm not cross-compiling much, but I've used Linaro because it's trivial to install and "just works" - in my case, for experimenting with u-boot (I wrote that for 10.04, but it continues to work after the upgrade to 12.04).

I guess others (twobob, knc1, NiLuJe...) will be more knowledgeable than I am, so you may want to wait for their opinion

[knc1]

Originally Posted by brianinmaine

Many of the pages linked on this site for development are 404'd.

There I go, buying food with my money rather paying the server bill.
I just thought that food would be a better long-term investment.

Originally Posted by brianinmaine

Just some general guidance on what you use, and or maybe a config file or 2?

Try this thread -
https://www.mobileread.com/forums/sho...d.php?t=222850
Describes three different approaches.

[twobob]

hmm...

It's up to you. if the project is BIG then you will NEED to cross compile.

if the project is TINY you could use something like tcc on the device.

in the "middle" is the grey area. The osmo, galculator and a few other bits (star-dict) I have actually built ON THE KINDLE touch.

This has many advantages to the layman, easy to include lib headers (apt-get libwhatever-dev), stable, plenty of "native arm" versions of possibly complicated libs to compile, dependency resolution.

The final choice I suppose would be the toolchain (or something like Buildroot that provides a Toolchain) and there there are few options now... Assuming you are focusing down binaries for your device and later only... The Linaro stuff or Code Sourcery and probably an awful lot more...

okay... All that said, let's presume you want an on-device solution...


AN ON-DEVICE SOLUTION::

If you are LOOKING for an "on device" solution I would suggest:
Debian Wheezy via grip (important as SID is now NOT making stuff our libc can run)
(I have a few sizes of this already built 500MB, 1Gb and 2Gb)

root@kindle:/# apt-get install libtool build-essential ccache automake
I would guess that this would, with the debian grip guts, weigh in around 320 MB)

(then inside debian)
root@kindle:/# mkdir -p /mnt/us/extensions
(on my system that becomes /mnt/us/debian/mnt/us/extensions)

build stuff like:
root@kindle:/# ./configure --prefix=/mnt/us/extensions/{thing}/usr <other options>

root@kindle:/# make clean; PATH=/usr/lib/ccache:$PATH make ; make install

then MOVE(/COPY) the /mnt/us/debian/mnt/us/extensions stuff to /mnt/us/extensions from outside the chroot.

and run it to test it.

perhaps KNC1 will be kind enough to let me update those debian images somewhere to replace the older ones I believe he is currently hosting.

Hope that helps.

[knc1]

Originally Posted by twobob

Perhaps KNC1 will be kind enough to let me update those debian images somewhere to replace the older ones I believe he is currently hosting.

Hope that helps.

Yes - we can deal with that during the (my) second, (your) third, work shift today.

[twobob]

Thanks. I think I have thought of a solution that WONT take up your space...

I've shoved a 500MB version (now zipped to 160mb) on dropbox and post INSTRUCTIONS to make it LARGER.
(not smaller)

EDIT: and here is a 7zipped (110Mb) link to the same file... 7 zipped 500Mb debian

You can find 7zip for the kindle here https://www.mobileread.com/forums/sho...d.php?t=223212


the following example will make the seed 500MB into a 2 Gb partition

FILENAME=/mnt/base-us/deb.5gb.ext3
dd if=/dev/zero of=$FILENAME bs=1M count=1500 oflag=append conv=notrunc
e2fsck -f $FILENAME
resize2fs $FILENAME


(You can choose ANY size just swap the 1500 for however many MB you want ADDED)
SEE THIS POST FOR A KINDLE DD THAT SUPPORTS THE RIGHT OPTIONS, (or do it on your PC)

That seems an efficient solution.


Alternately as pointed out by Icantbebothered:
FILENAME=/mnt/base-us/deb.5gb.ext3
dd if=/dev/zero bs=1M count=512 >> $FILENAME
e2fsck -f $FILENAME
resize2fs $FILENAME

The above is OKAY because the ext3 file should never be "being written to" anyways during this process.. so yeah. Good point by him.



deb.5gb.ext3 is a general purpose image that WILL need the additional steps outlined above doing to make it a "build system", when you are bored of it or you screw it up just replace it again. (EDIT: I did this just last night)

Enjoy


The script to mount this partition is available on the DEBIAN ON A PAPERWHITE thread (and ill post it again below)

[twobob]

#!/bin/sh
unset LD_PRELOAD
rm /dev/loop/debian
mknod -m0660 /dev/loop/debian b 7 250
echo "[*] Mounting Rootfs..."
/bin/mount -o loop=/dev/loop/debian,noatime -t ext3 /mnt/base-us/debian.ext3 /mnt/us/debian
echo "[*] Preparing Filesystem..."
/bin/mount -o bind /dev /mnt/us/debian/dev
/bin/mount -o bind /proc /mnt/us/debian/proc
/bin/mount -o bind /sys /mnt/us/debian/sys
/bin/mount -o bind /dev/pts /mnt/us/debian/dev/pts

#optionally make the kindle available inside debian...
#mount -o bind /mnt/base-us /mnt/us/debian/kindle/mnt/us
#mount -o bind /mnt/base-us /mnt/us/debian/kindle/mnt/base-us

echo "[*] Preparing Network Connections..."
cp /etc/hosts /mnt/us/debian/etc/hosts
cp /etc/resolv.conf /mnt/us/debian/etc/resolv.conf
echo "[*] Starting Debian shell..."
chroot /mnt/us/debian /bin/bash $@
#chroot /mnt/debian 
echo
echo "[*] Unmounting Rootfs..."

#optionally remove that kindle inside the debian stuff
#umount /mnt/us/debian/kindle/mnt/us
#umount /mnt/us/debian/kindle/mnt/base-us

/bin/umount /mnt/us/debian/dev/pts
/bin/umount /mnt/us/debian/dev
/bin/umount /mnt/us/debian/proc
/bin/umount /mnt/us/debian/sys
/bin/umount -d /mnt/us/debian
echo "[*] Debian unmounted."

Please note that this script EXPECTS the debian image to be called debian.ext3 and to be on the FAT ROOT like /mnt/us/debian.ext3

so rename your image to that and put it there...

put this script wherever you would like and run it directly via SSH.

[knc1]

Ah, you could:

* run the increase size stuff on your copy -
* gzip (or zip) the result -
All of those unused blocks will compress nicely, might even be smaller that the present non-compressed image file.

[knc1]

Originally Posted by twobob

Thanks. I think I have thought of a solution that WONT take up your space...

We (or at least I) need to talk - met you at the usual place.

[twobob]

one more thing... the busybox dd does NOT support these append modes...

Here is one that does. Enjoy.

[root@kindle debian]# /mnt/us/dd --help
Usage: /mnt/us/dd [OPERAND]...
or: /mnt/us/dd OPTION
Copy a file, converting and formatting according to the operands.

bs=BYTES read and write up to BYTES bytes at a time
cbs=BYTES convert BYTES bytes at a time
conv=CONVS convert the file as per the comma separated symbol list
count=BLOCKS copy only BLOCKS input blocks
ibs=BYTES read up to BYTES bytes at a time (default: 512)
if=FILE read from FILE instead of stdin
iflag=FLAGS read as per the comma separated symbol list
obs=BYTES write BYTES bytes at a time (default: 512)
of=FILE write to FILE instead of stdout
oflag=FLAGS write as per the comma separated symbol list
seek=BLOCKS skip BLOCKS obs-sized blocks at start of output
skip=BLOCKS skip BLOCKS ibs-sized blocks at start of input
status=noxfer suppress transfer statistics

BLOCKS and BYTES may be followed by the following multiplicative suffixes:
c =1, w =2, b =512, kB =1000, K =1024, MB =1000*1000, M =1024*1024, xM =M
GB =1000*1000*1000, G =1024*1024*1024, and so on for T, P, E, Z, Y.

Each CONV symbol may be:

ascii from EBCDIC to ASCII
ebcdic from ASCII to EBCDIC
ibm from ASCII to alternate EBCDIC
block pad newline-terminated records with spaces to cbs-size
unblock replace trailing spaces in cbs-size records with newline
lcase change upper case to lower case
ucase change lower case to upper case
swab swap every pair of input bytes
sync pad every input block with NULs to ibs-size; when used
with block or unblock, pad with spaces rather than NULs
excl fail if the output file already exists
nocreat do not create the output file
notrunc do not truncate the output file
noerror continue after read errors
fdatasync physically write output file data before finishing
fsync likewise, but also write metadata

Each FLAG symbol may be:

append append mode (makes sense only for output; conv=notrunc suggested)
direct use direct I/O for data
directory fail unless a directory
dsync use synchronized I/O for data
sync likewise, but also for metadata
fullblock accumulate full blocks of input (iflag only)
nonblock use non-blocking I/O
noatime do not update access time
nocache discard cached data
noctty do not assign controlling terminal from file
nofollow do not follow symlinks

Sending a USR1 signal to a running `dd' process makes it
print I/O statistics to standard error and then resume copying.

$ dd if=/dev/zero of=/dev/null& pid=$!
$ kill -USR1 $pid; sleep 1; kill $pid
18335302+0 records in
18335302+0 records out
9387674624 bytes (9.4 GB) copied, 34.6279 seconds, 271 MB/s

Options are:

--help display this help and exit
--version output version information and exit

Report dd bugs to bug-coreutils@gnu.org
GNU coreutils home page: <http://www.gnu.org/software/coreutils/>
General help using GNU software: <http://www.gnu.org/gethelp/>
For complete documentation, run: info coreutils 'dd invocation'

[NiLuJe]

And provided you want to go the cross way, like ixtab said, a Linaro or CodeSourcery TC is probably the easiest thing you can get your hands on, and, if you're only targetting the K4/K5, it usually works well enough without too many hair-pulling experiences .

I personally am building two different TCs with the help of crosstool-ng, one targetting the K2/K3, and the other the K5. I'm using bleeding edge GCC versions, so there is some hoop jumping to be done, especially for the K3 toolchain. My configs & build-scripts are available in the kindle-x-tc-YYYY.MM.tar.gz attachement of the gigantic K2/K3 hack threads.

There are a *lot* of different ways to do it, and each of us tends to do our "own" stuff, so there isn't a really universal 'right way' to do it. For a cross TC, I'd say the closest to it would be the CodeSourcery TC, a recent one for the K5, and the truly awfully old one from 2009q7 (I think?) for the K3.

If you have more specific questions, don't hesitate to ask, or to come looking for us on IRC .

[twobob]

Originally Posted by knc1

Ah, you could:

* run the increase size stuff on your copy -
* gzip (or zip) the result -
All of those unused blocks will compress nicely, might even be smaller that the present non-compressed image file.

Thanks mate.

Done. Good call.

[brianinmaine]

Originally Posted by knc1

There I go, buying food with my money rather paying the server bill.
I just thought that food would be a better long-term investment.

I REALLY APPRECIATE all the help I get on here! It's so much fun to see what has been posted every day. I'm not complaining about you personally, I just noticed that many of the links pointed to in other threads about setups were no longer there (not just yours about KEK).

Thank you all for the help!

PS: I belong to a wonderful motorcycle forum where members can ask ANYTHING, even how to change your oil, and many will chime in with help and photos of where and what to do. No mocking, just good natured fun. This forum reminds me of that kind of comraderie. Thanks.

[brianinmaine]

Originally Posted by NiLuJe

For a cross TC, I'd say the closest to it would be the CodeSourcery TC, a recent one for the K5

If you have more specific questions, don't hesitate to ask, or to come looking for us on IRC .

I installed the latest Code Sourcery TC and can compile very simple programs. Now off to learn how to actually USE it.

I also resized the twobob image to 2G and installed lots of stuff in that image on my Kindle. Looks very promising!

Thanks!

[knc1]

Originally Posted by brianinmaine

PS: I belong to a wonderful motorcycle forum where members can ask ANYTHING, even how to change your oil, and many will chime in with help and photos of where and what to do. No mocking, just good natured fun. This forum reminds me of that kind of comraderie. Thanks.

Sorry about the tone of my statement.
But . . . .
The situation developed where many of my sites where only serving to web 'bots, not people.
It reached such proportions, that I was getting "resource limits exceeded" messages more than once a day.

To continue the way I was going, I would have needed to implement (by monthly purchase) a much higher resource system.

I have been retired for almost 20 years -
To do that would have made serious in-roads into my food budget -
I decided in favor of food. (Literally, no humor intended.)

I know, that is a voluntary choice of lifestyle, of which I am solely responsible, but the choice had to be made.
So many of my "free" support services had to go.