Re: Can I run linux without a file system?

Re: Can I run linux without a file system?

[Tim Lai]

I am interested in both input/output operation
on the console. If I just set CONFIG_SERIAL_CONSOLE,
will I be able read input from the console?
The main application will be started from init(),
and the application will need to read and write
to the console. Are there are method to communicate
to the serial port other than open("/dev/ttyS0")?


>
> You don't need a filesystem to get output on the
> serial console
> you just need to enable the console with
> CONFIG_SERIAL_CONSOLE=y
> in your kernel configuration (atleast for mpc860
> that all)
> but you will have a hard time producing much more
> than a blinking
> cursor if you boot a Linux kernel and have no
> application that
> it then can run on the root-filesystem - what would
> be the point
> of such a setup - 1MB kernel code for a blinking
> cursor on a
> serial port seems expensive.
>
> hofrat


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re[2]: Can I run Linux without a file system?

[Ricardo Scop]

Tim,

Maybe initrd and linuxrc is enough for your system. Read the file
initrd.txt in the Linux source tree Documentation sub-directory.

[]'s, Scop                            mailto:scop@vanet.com.br

------------------------------------------------------------------
It might look like I'm doing nothing, but at the cellular level I'm
really quite busy.

Friday, June 21, 2002, 4:35:17 PM, you wrote:


TL> Thanks, Jason.
TL> I am new to linux kernel. I'll have the main
TL> application run from init(), so I wasn't planning
TL> to have a file system.

>> Yes. You will always have SOME kind of filesystem.
>> But this begs another
>> question. How much do you know about Linux, and what
>> are you really asking?

TL> If /proc and /dev is not really on any disk, what do
TL> I have to do to init or create /dev? Do I need ramdisk
TL> as a minumum requirement for linux?
TL> My main goal right now is to get the serial port
TL> to work, so I can do some debugging with the dumb
TL> terminal. After I do tty_register() in the serial
TL> driver, does linux assign /dev/ttyS to this device?

>> The /proc filesystem is not really on any disk, just
>> like /dev (I think)
>> isn't on any disk, though they look like to us users
>> that they are
>> filesystems.
>>

TL> Can you give me pointers on which file to read?

>>
>> Does this help?
>>

TL> Yes. Thank you very much. :)

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re[2]: Can I run linux without a file system?

[Ricardo Scop]

Tim,

See bellow...

[]'s, Scop                            mailto:scop@vanet.com.br

------------------------------------------------------------------
It might look like I'm doing nothing, but at the cellular level I'm
really quite busy.

Friday, June 21, 2002, 4:42:34 PM, you wrote:


TL> I am interested in both input/output operation
TL> on the console. If I just set CONFIG_SERIAL_CONSOLE,
TL> will I be able read input from the console?
TL> The main application will be started from init(),
TL> and the application will need to read and write
TL> to the console. Are there are method to communicate
TL> to the serial port other than open("/dev/ttyS0")?
Not AFAIK. The VFS (Virtual File System) is at the very heart of Linux
and _is_ the abstraction used to deal with I/O devices. You don't need
to try avoiding it. A simple initrd will do the job and can be as
light as you make it.

>>
>> You don't need a filesystem to get output on the
>> serial console
>> you just need to enable the console with
>> CONFIG_SERIAL_CONSOLE=y
>> in your kernel configuration (atleast for mpc860
>> that all)
>> but you will have a hard time producing much more
>> than a blinking
>> cursor if you boot a Linux kernel and have no
>> application that
>> it then can run on the root-filesystem - what would
>> be the point
>> of such a setup - 1MB kernel code for a blinking
>> cursor on a
>> serial port seems expensive.
you get the network protocol stacks, too...

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Jerry Van Baren]

You need at least a RAM file system for "/" and a bunch of subdirectories
such as /dev, /lib, etc.  The common way to do this on a minimalistic
system is to create a file system image in ROM (often compressed) and copy
it to RAM on start up.  Given the questions you are asking, I am very
confident creating a minimal RAM disk image will challenge you sufficiently
:-).  I'm not being snide, lots of people with lots of linux knowledge have
tried and failed.  Most people use someone else's pre-configured minimal
file systems and add/subtract (mostly add :-) programs to it.  This is
because it is very, very hard to create a minimal file system (that works,
that is).

Pointers to development systems with example RAM disk images:
   http://www.denx.de/solutions-en.html
     ftp://ftp.denx.de/pub/LinuxPPC/usr/src/SELF/
   http://www.mvista.com/
   (there are others, I'm just too lazy to do the google search for you)

Trying to run linux without a file system of any sort would require you to
rewrite of pretty much everything and the three or four things you didn't
rewrite, you would have to rebuild (static link, no shared
libraries).  There are a lot of software engineers and hackers that would
turn down the opportunity to do this at any price.  On the other hand, a
lot of naivety and a even more coffee sometimes generates remarkable
results :-).

You really want to look at eCOS or one of the other light weight tasking OSs.
   http://www.redhat.com/embedded/technologies/ecos/

gvb


At 12:42 PM 6/21/2002 -0700, Tim Lai wrote:

>I am interested in both input/output operation
>on the console. If I just set CONFIG_SERIAL_CONSOLE,
>will I be able read input from the console?
>The main application will be started from init(),
>and the application will need to read and write
>to the console. Are there are method to communicate
>to the serial port other than open("/dev/ttyS0")?
>
>
> >
> > You don't need a filesystem to get output on the
> > serial console
> > you just need to enable the console with
> > CONFIG_SERIAL_CONSOLE=y
> > in your kernel configuration (atleast for mpc860
> > that all)
> > but you will have a hard time producing much more
> > than a blinking
> > cursor if you boot a Linux kernel and have no
> > application that
> > it then can run on the root-filesystem - what would
> > be the point
> > of such a setup - 1MB kernel code for a blinking
> > cursor on a
> > serial port seems expensive.
> >
> > hofrat
>
>


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Tim Lai]

Thanks for all the replies. After reading all the
helpful response. I think I will try to use initrd,
and probably try ramdisk, too. I don't necessary need
to run with a minimum system. The hardware has 8M,
and I will try 1M for ramdisk. It seems like this
is the easiest way to get thing starts.


--- Jerry Van Baren <vanbaren_gerald@si.com> wrote:
>
> You need at least a RAM file system for "/" and a
> bunch of subdirectories
> such as /dev, /lib, etc.  The common way to do this
> on a minimalistic
> system is to create a file system image in ROM
> (often compressed) and copy
> it to RAM on start up.  Given the questions you are
> asking, I am very
> confident creating a minimal RAM disk image will
> challenge you sufficiently
> :-).  I'm not being snide, lots of people with lots
> of linux knowledge have
> tried and failed.  Most people use someone else's
> pre-configured minimal
> file systems and add/subtract (mostly add :-)
> programs to it.  This is
> because it is very, very hard to create a minimal
> file system (that works,
> that is).
>
> Pointers to development systems with example RAM
> disk images:
>    http://www.denx.de/solutions-en.html
>      ftp://ftp.denx.de/pub/LinuxPPC/usr/src/SELF/
>    http://www.mvista.com/
>    (there are others, I'm just too lazy to do the
> google search for you)
>
> Trying to run linux without a file system of any
> sort would require you to
> rewrite of pretty much everything and the three or
> four things you didn't
> rewrite, you would have to rebuild (static link, no
> shared
> libraries).  There are a lot of software engineers
> and hackers that would
> turn down the opportunity to do this at any price.
> On the other hand, a
> lot of naivety and a even more coffee sometimes
> generates remarkable
> results :-).
>
> You really want to look at eCOS or one of the other
> light weight tasking OSs.
>    http://www.redhat.com/embedded/technologies/ecos/
>
> gvb
>


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Dr. Craig Hollabaugh]

At 04:12 PM 6/21/2002 -0400, Jerry Van Baren wrote:
>
>You need at least a RAM file system for "/" and a bunch of subdirectories
>such as /dev, /lib, etc.  The common way to do this on a minimalistic
>system is to create a file system image in ROM (often compressed) and copy
>it to RAM on start up.  Given the questions you are asking, I am very
>confident creating a minimal RAM disk image will challenge you sufficiently
>:-).  I'm not being snide, lots of people with lots of linux knowledge have
>tried and failed.  Most people use someone else's pre-configured minimal
>file systems and add/subtract (mostly add :-) programs to it.  This is
>because it is very, very hard to create a minimal file system (that works,
>that is).

I disagree. Creating a minimal root filesystem is easy once you understand what
happens when init executes. (Shameless plug here) In my Embedded Linux book, I show
you exactly the files needed to get bash running with a network connection. You use ldd to find out lib dependencies. Then copy binaries you want with their libs and some config files and your done. Here's my script to build such a root filesystem (http://www.embeddedlinuxinterfacing.com/chapters/04/buildrootfilesystem/buildrootfilesystem). Look it over, it probably won't make much sense without reading chapter 4.

I can say that a minimum RedHat install uses 29,296 files – 382.020MB, debian 10,734 files - 67.428MB, my minimal root filesystem 82 files and 4.8MB.


Here's a listing of tftproot directory

root@tbdev1[513]: cd /tftpboot/powerpc-rootfs/
root@tbdev1[514]: du -s
4824    .
root@tbdev1[515]: find .
.
./dev
./dev/tty
./dev/tty0
./dev/ttyS0
./dev/console
./dev/null
./dev/ram
./dev/initctl
./dev/mem
./dev/pts
./dev/ptyp0
./dev/ttyp0
./etc
./etc/init.d
./etc/init.d/rcS
./etc/init.d/umountfs
./etc/protocols
./etc/services
./etc/fstab
./etc/inittab
./etc/resolv.conf
./etc/mtab
./bin
./bin/bash
./bin/cat
./bin/ls
./bin/mount
./bin/umount
./bin/ps
./bin/df
./bin/kill
./bin/ping
./bin/chmod
./bin/touch
./bin/rm
./bin/echo
./bin/sh
./sbin
./sbin/init
./sbin/ifconfig
./sbin/route
./sbin/depmod
./sbin/insmod
./sbin/lsmod
./sbin/rmmod
./lib
./lib/ld-2.2.3.so
./lib/ld.so.1
./lib/libc-2.2.3.so
./lib/libc.so.6
./lib/libutil-2.2.3.so
./lib/libutil.so.1
./lib/libncurses.so.5
./lib/libncurses.so.5.2
./lib/libdl-2.2.3.so
./lib/libdl.so.2
./lib/libnss_dns-2.2.3.so
./lib/libnss_dns.so.2
./lib/libnss_files-2.2.3.so
./lib/libnss_files.so.2
./lib/libresolv-2.2.3.so
./lib/libresolv.so.2
./lib/libproc.so.2.0.7
./lib/librt-2.2.3.so
./lib/librt.so.1
./lib/libpthread-0.9.so
./lib/libpthread.so.0
./lib/libm-2.2.3.so
./lib/libm.so.6
./lib/libstdc++-3-libc6.1-2-2.10.0.so
./lib/libstdc++-libc6.1-2.so.3
./usr
./usr/bin
./usr/bin/telnet
./usr/bin/gdbserver
./usr/lib
./proc
./tmp


root@tbdev1[516]: find . | wc
     82      82    1212


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Tim Lai]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=us-ascii, Size: 335 bytes --]


>
> I can say that a minimum RedHat install uses 29,296
> files – 382.020MB, debian 10,734 files - 67.428MB,
> my minimal root filesystem 82 files and 4.8MB.
>
>

Do I need to install all these files in the file
system? Can I have "/", with no file in it?


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[John W. Linville]

Tim Lai wrote:
>
> Do I need to install all these files in the file
> system? Can I have "/", with no file in it?

Respectfully, I think perhaps you should call WindRiver or some other
traditional RTOS vendor.

If you insist on using Linux, I highly recommend you go to Amazon and
order Craig's book "Embedded Linux".  Make sure you get the one by Craig
Hollabaugh, as there is another book by the same title that is, well,
less useful...

Good luck.

John
--
John W. Linville
LVL7 Systems, Inc.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Dr. Craig Hollabaugh]

At 01:49 PM 6/21/2002 -0700, Tim Lai wrote:
>
>>
>> I can say that a minimum RedHat install uses 29,296
>> files – 382.020MB, debian 10,734 files - 67.428MB,
>> my minimal root filesystem 82 files and 4.8MB.
>>
>>
>
>Do I need to install all these files in the file
>system? Can I have "/", with no file in it?


What do you want to do? Do you want a shell, run a program, load a device driver? What is it exactly?

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Jerry Van Baren]

OK, you and John Kerl caught me fear-mongering :-).  Thank you for the link
and shameless plug, it looks very, very useful.  4.8MB is very respectable
but not what most people think about when they say "minimal file
system".  Tim wants less than 1MByte, care to take up THAT challenge
:-)?  That would require shrinking the libraries to just the necessary
functions, which is a big jump in difficulty.

gvb


At 02:31 PM 6/21/2002 -0600, Dr. Craig Hollabaugh wrote:
>At 04:12 PM 6/21/2002 -0400, Jerry Van Baren wrote:
> >
> >You need at least a RAM file system for "/" and a bunch of subdirectories
> >such as /dev, /lib, etc.  The common way to do this on a minimalistic
> >system is to create a file system image in ROM (often compressed) and copy
> >it to RAM on start up.  Given the questions you are asking, I am very
> >confident creating a minimal RAM disk image will challenge you sufficiently
> >:-).  I'm not being snide, lots of people with lots of linux knowledge have
> >tried and failed.  Most people use someone else's pre-configured minimal
> >file systems and add/subtract (mostly add :-) programs to it.  This is
> >because it is very, very hard to create a minimal file system (that works,
> >that is).
>
>I disagree. Creating a minimal root filesystem is easy once you understand
>what
>happens when init executes. (Shameless plug here) In my Embedded Linux
>book, I show
>you exactly the files needed to get bash running with a network
>connection. You use ldd to find out lib dependencies. Then copy binaries
>you want with their libs and some config files and your done. Here's my
>script to build such a root filesystem
>(http://www.embeddedlinuxinterfacing.com/chapters/04/buildrootfilesystem/buildrootfilesystem).
>Look it over, it probably won't make much sense without reading chapter 4.
>
>I can say that a minimum RedHat install uses 29,296 files ­ 382.020MB,
>debian 10,734 files - 67.428MB, my minimal root filesystem 82 files and 4.8MB.

[snip]


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Wolfgang Denk]

In message <3.0.3.32.20020621143133.006d37a8@hollabaugh.com> you wrote:
>
> I can say that a minimum RedHat install uses 29,296 files – 382.020MB, debian 10,734 files - 67.428MB, my minimal root filesystem 82 files and 4.8MB.

You cannot really call this a "minimal root filesystem".

A small (not even this is minimal, nor optimized) image can be found at
ftp://ftp.denx.de/pub/LinuxPPC/usr/src/mini-2048kB-ramdisk.image.gz

This has:
	-> find .
	.
	./lost+found
	./bin
	./bin/sh
	./bin/ls
	./dev
	./dev/ttyS0
	./dev/tty
	./dev/console
	./lib
	./lib/ld.so.1
	./lib/libc-1.99.so
	./lib/libc.so.6
	./lib/libtermcap.so.2.0.8
	./lib/libtermcap.so.2
	-> du -sk .
	1367    .

Obvious optimizations: rmdir lost+found, use busybox instead of  bash
+ ls, ...


Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Were there fewer fools, knaves would starve.              - Anonymous

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[Tim Lai]

> >
> >Do I need to install all these files in the file
> >system? Can I have "/", with no file in it?
>
>
> What do you want to do? Do you want a shell, run a
> program, load a device driver? What is it exactly?
>

I 'll run my application which is build in the kernel.

No shell in the beginning. But I will need the shell
later when more NVRAM is available. With the hardware
that I have in the future, I can build my application
seperately and store it in the file system. The
application do some network processing, I'll not
have any display other then the console and maybe
telnet session.

Where or how do I get these lib for ppc?
         ./lib/ld.so.1
	./lib/libc-1.99.so
	./lib/libc.so.6
	./lib/libtermcap.so.2.0.8
	./lib/libtermcap.so.2

BTW, I just ordered the Dr.Hollabaugh's book.
Thanks for the advice.


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Can I run linux without a file system?

[David Blythe]

use uclibc and busybox.  it is pretty straightforward to get a root file
system smaller than 1M with lots of useful stuff, either with or without
shared libraries (shared uclibc is < 300K including ld.so on my ppc
system, busybox is 200K including ash,mv,cp,ls,...).  The downside is
you have to build all of the stuff yourself, but if you want to optimize
for space, sooner or later you'll end up building it all yourself.

	david


Jerry Van Baren wrote:
>
> OK, you and John Kerl caught me fear-mongering :-).  Thank you for the link
> and shameless plug, it looks very, very useful.  4.8MB is very respectable
> but not what most people think about when they say "minimal file
> system".  Tim wants less than 1MByte, care to take up THAT challenge
> :-)?  That would require shrinking the libraries to just the necessary
> functions, which is a big jump in difficulty.
>
> gvb
>
>
> At 02:31 PM 6/21/2002 -0600, Dr. Craig Hollabaugh wrote:
>
...


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/