[ECOS] Memory footprint

[ECOS] Memory footprint

[Emmanuel Viollet]

Hi.

I can't find any information on this on the internet, so perhaps you can 
help me out...

I need to compile a kernel with PPP. Do you have any idea of the 
footprint of eCOS with lwIP and PPP?
I compiled a minimal program with PPP for my Olimex LPC2106 evaluation 
board, and it takes 560Kb.
Does this size seem normal?

In advance, thanks.

-- 
Emmanuel Viollet
Software engineer
Webdyn
e-mail : Emmanuel.Viollet@webdyn.com
Tel    : +33 (0)1 39 04 29 55


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Memory footprint

[Andrew Lunn]

On Wed, May 10, 2006 at 11:08:49AM +0200, Emmanuel Viollet wrote:
> Hi.
> 
> I can't find any information on this on the internet, so perhaps you can 
> help me out...
> 
> I need to compile a kernel with PPP. Do you have any idea of the 
> footprint of eCOS with lwIP and PPP?
> I compiled a minimal program with PPP for my Olimex LPC2106 evaluation 
> board, and it takes 560Kb.
> Does this size seem normal?

Is that the size of the elf or the size of the binary which is
actually loaded onto the target? 

what does arm-elf-size say?

     Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Memory footprint

[Emmanuel Viollet]

Andrew Lunn a écrit :
> On Wed, May 10, 2006 at 11:08:49AM +0200, Emmanuel Viollet wrote:
>> Hi.
>>
>> I can't find any information on this on the internet, so perhaps you can 
>> help me out...
>>
>> I need to compile a kernel with PPP. Do you have any idea of the 
>> footprint of eCOS with lwIP and PPP?
>> I compiled a minimal program with PPP for my Olimex LPC2106 evaluation 
>> board, and it takes 560Kb.
>> Does this size seem normal?
> 
> Is that the size of the elf or the size of the binary which is
> actually loaded onto the target? 
> 
> what does arm-elf-size say?

$ arm-elf-size.exe a.out
    text    data     bss     dec     hex filename
   41920    1512    8028   51460    c904 a.out

I suppose that this means the binary will be 51460Kb in size?

I am using a Phlips 2106 processor with 128Kb EEPROM and 64Kb RAM.
Do you think I can fit in LwIP + PPP + Zlib in this?

The linker is giving me errors :
$ arm-elf-gcc -I../../ecos_install/include hello.c 
-L../../ecos_install/lib -Ttarget.ld -nostdlib
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x2624c de a.out de la section .text n'est pas dans la région rom
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x2624c de a.out de la section .fini n'est pas dans la région rom
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x2aff1 de a.out de la section .rodata n'est pas dans la région
  rom
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x2aff4 de a.out de la section .rodata1 n'est pas dans la région rom
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x2aff4 de a.out de la section .fixup n'est pas dans la région rom
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x2aff4 de a.out de la section .gcc_except_table n'est pas dans la 
région rom
/usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
0x4001ca20 de a.out de la section .bss n'est pas dans la région ram

I'm sorry, but the error messages are in french.
 From what I gather, the code is too big for RAM and ROM?
"the address 0x2624c from a.out in section .text is not in the rom region"?

-- 
Emmanuel Viollet
Software engineer
Webdyn
e-mail : Emmanuel.Viollet@webdyn.com
Tel    : +33 (0)1 39 04 29 55

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Choice of Web Servers

[Laurie Gellatly]

Hi All,
I'm new to the list. Still selecting parts for my project.

I've been looking for a capable web server (IPv4/6, HTTPS, DAA, CGI,
threads...) and GoAhead seems to fit except for the logo on the end
product licensing requirement which rules it out. The standard server
unfortunately lacks required features and I searched through past posts
and found 'seminole'. Google found 'shttpd' which also looks promising
but there appears to be some level of porting required. Anyone suggest
an alternative and/or provide comments on these last two?

Cheers			...Laurie:{)


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Memory footprint

[Andrew Lunn]

On Wed, May 10, 2006 at 02:54:55PM +0200, Emmanuel Viollet wrote:
> Andrew Lunn a ?crit :
> >On Wed, May 10, 2006 at 11:08:49AM +0200, Emmanuel Viollet wrote:
> >>Hi.
> >>
> >>I can't find any information on this on the internet, so perhaps you can 
> >>help me out...
> >>
> >>I need to compile a kernel with PPP. Do you have any idea of the 
> >>footprint of eCOS with lwIP and PPP?
> >>I compiled a minimal program with PPP for my Olimex LPC2106 evaluation 
> >>board, and it takes 560Kb.
> >>Does this size seem normal?
> >
> >Is that the size of the elf or the size of the binary which is
> >actually loaded onto the target? 
> >
> >what does arm-elf-size say?
> 
> $ arm-elf-size.exe a.out
>    text    data     bss     dec     hex filename
>   41920    1512    8028   51460    c904 a.out
> 
> I suppose that this means the binary will be 51460Kb in size?
                                               
No. These are in units of bytes, not kilo bits. 51460Kb is about
6MBytes which makes no sense.

The binary will be 41920 + 1512 bytes. You also need 8028bytes of
statically allocated RAM. The size of the heap is not show here.

> I am using a Phlips 2106 processor with 128Kb EEPROM and 64Kb RAM.
> Do you think I can fit in LwIP + PPP + Zlib in this?

zlib will be a problem. In its default configuration it will allocate
64Kbytes from the heap..... You can make is use less memory, but it
will then no longer decompress data compressed with the default zlib
configuration. So i don't recommend zlib. lwip + PPP does seems
reasonable with 128Kb + 64Kb.

> The linker is giving me errors :
> $ arm-elf-gcc -I../../ecos_install/include hello.c 
> -L../../ecos_install/lib -Ttarget.ld -nostdlib
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x2624c de a.out de la section .text n'est pas dans la r?gion rom
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x2624c de a.out de la section .fini n'est pas dans la r?gion rom
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x2aff1 de a.out de la section .rodata n'est pas dans la r?gion
>  rom
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x2aff4 de a.out de la section .rodata1 n'est pas dans la r?gion rom
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x2aff4 de a.out de la section .fixup n'est pas dans la r?gion rom
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x2aff4 de a.out de la section .gcc_except_table n'est pas dans la 
> r?gion rom
> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
> 0x4001ca20 de a.out de la section .bss n'est pas dans la r?gion ram
> 
> I'm sorry, but the error messages are in french.
> From what I gather, the code is too big for RAM and ROM?

This should fit, so it suggests a problem with your linker script.
Is it trying to put everything in RAM?

   Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Memory footprint

[Emmanuel Viollet]

Andrew Lunn a écrit :
>> $ arm-elf-size.exe a.out
>>    text    data     bss     dec     hex filename
>>   41920    1512    8028   51460    c904 a.out
>>
>> I suppose that this means the binary will be 51460Kb in size?
>>     
>                                                
> No. These are in units of bytes, not kilo bits. 51460Kb is about
> 6MBytes which makes no sense.
>   
True. In fact, 51460Kb, is closer to 50Mb... ;)
I mistyped and added a "K". My mistake!! :)
> The binary will be 41920 + 1512 bytes. You also need 8028bytes of
> statically allocated RAM. The size of the heap is not show here.
>   
OK. Thanks.
>> I am using a Phlips 2106 processor with 128Kb EEPROM and 64Kb RAM.
>> Do you think I can fit in LwIP + PPP + Zlib in this?
>>     
> zlib will be a problem. In its default configuration it will allocate
> 64Kbytes from the heap..... You can make is use less memory, but it
> will then no longer decompress data compressed with the default zlib
> configuration. So i don't recommend zlib. lwip + PPP does seems
> reasonable with 128Kb + 64Kb.
>   
I only need to compress data for gzip using zlib. I don't need to 
uncompress files coming from some other place. Do you think zlib could 
work with only 32Kb? Do you know of any other gzip compatible libraries?
>> The linker is giving me errors :
>> $ arm-elf-gcc -I../../ecos_install/include hello.c 
>> -L../../ecos_install/lib -Ttarget.ld -nostdlib
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x2624c de a.out de la section .text n'est pas dans la r?gion rom
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x2624c de a.out de la section .fini n'est pas dans la r?gion rom
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x2aff1 de a.out de la section .rodata n'est pas dans la r?gion
>>  rom
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x2aff4 de a.out de la section .rodata1 n'est pas dans la r?gion rom
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x2aff4 de a.out de la section .fixup n'est pas dans la r?gion rom
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x2aff4 de a.out de la section .gcc_except_table n'est pas dans la 
>> r?gion rom
>> /usr/lib/gcc-lib/arm-elf/3.4.0/../../../../arm-elf/bin/ld: l'adresse 
>> 0x4001ca20 de a.out de la section .bss n'est pas dans la r?gion ram
>>
>> I'm sorry, but the error messages are in french.
>> >From what I gather, the code is too big for RAM and ROM?
>>     
>
> This should fit, so it suggests a problem with your linker script.
> Is it trying to put everything in RAM?
>
>   
The program I was trying to compile in the second part has nothing to do 
with the numbers I gave in the first part of this email.
This program contains lwIP + PPP + zlib + debug information + lots of 
other things.
I just wanted to check if the error messages I was seeing were due to 
code too large.

-- 
Emmanuel Viollet
Software engineer
Webdyn
e-mail : Emmanuel.Viollet@webdyn.com
Tel    : +33 (0)1 39 04 29 55


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Memory footprint

[Andrew Lunn]

> >zlib will be a problem. In its default configuration it will allocate
> >64Kbytes from the heap..... You can make is use less memory, but it
> >will then no longer decompress data compressed with the default zlib
> >configuration. So i don't recommend zlib. lwip + PPP does seems
> >reasonable with 128Kb + 64Kb.
> >  
> I only need to compress data for gzip using zlib. I don't need to 
> uncompress files coming from some other place. Do you think zlib could 
> work with only 32Kb? Do you know of any other gzip compatible libraries?

Sorry, no idea. I only needed to decompress data, not compress it.

       Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Re: Memory footprint

[Grant Edwards]

>> what does arm-elf-size say?
>
> $ arm-elf-size.exe a.out
>     text    data     bss     dec     hex filename
>    41920    1512    8028   51460    c904 a.out
>
> I suppose that this means the binary will be 51460Kb in size?

That means the binary will require 43432 bytes of ROM
(41920+1512) and 9540 bytes of RAM (8028+1512).

-- 
Grant Edwards                   grante             Yow!  I'm pretending that
                                  at               we're all watching PHIL
                               visi.com            SILVERS instead of RICARDO
                                                   MONTALBAN!

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Re: Memory footprint

[Andrew Lunn]

On Wed, May 10, 2006 at 09:53:42AM -0500, Grant Edwards wrote:
> >> what does arm-elf-size say?
> >
> > $ arm-elf-size.exe a.out
> >     text    data     bss     dec     hex filename
> >    41920    1512    8028   51460    c904 a.out
> >
> > I suppose that this means the binary will be 51460Kb in size?
> 
> That means the binary will require 43432 bytes of ROM
> (41920+1512) and 9540 bytes of RAM (8028+1512).

I suppose i should read the man page, but couldn't part of data be
rodata and so will not be copied to RAM at startup?

[Goes and reads the man page]

There is no mention of the rodata. Hum...

        Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Re: Memory footprint

[Gary Thomas]

On Wed, 2006-05-10 at 17:19 +0200, Andrew Lunn wrote:
> On Wed, May 10, 2006 at 09:53:42AM -0500, Grant Edwards wrote:
> > >> what does arm-elf-size say?
> > >
> > > $ arm-elf-size.exe a.out
> > >     text    data     bss     dec     hex filename
> > >    41920    1512    8028   51460    c904 a.out
> > >
> > > I suppose that this means the binary will be 51460Kb in size?
> > 
> > That means the binary will require 43432 bytes of ROM
> > (41920+1512) and 9540 bytes of RAM (8028+1512).
> 
> I suppose i should read the man page, but couldn't part of data be
> rodata and so will not be copied to RAM at startup?
> 
> [Goes and reads the man page]
> 
> There is no mention of the rodata. Hum...

The read-only data section will be included in the "text" segment
as printed by this tool.  If you want to see the details of how
things are really laid out, use objdump -h.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Re: Memory footprint

[Grant Edwards]

In gmane.os.ecos.general, you wrote:
> On Wed, May 10, 2006 at 09:53:42AM -0500, Grant Edwards wrote:
>> >> what does arm-elf-size say?
>> >
>> > $ arm-elf-size.exe a.out
>> >     text    data     bss     dec     hex filename
>> >    41920    1512    8028   51460    c904 a.out
>> >
>> > I suppose that this means the binary will be 51460Kb in size?
>> 
>> That means the binary will require 43432 bytes of ROM
>> (41920+1512) and 9540 bytes of RAM (8028+1512).
>
> I suppose i should read the man page, but couldn't part of data be
> rodata and so will not be copied to RAM at startup?
>
> [Goes and reads the man page]
>
> There is no mention of the rodata. Hum...

I'm pretty sure the "text" number included rodata.

-- 
Grant Edwards                   grante             Yow!  I just had a NOSE
                                  at               JOB!!
                               visi.com            

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Re: Memory footprint

[Grant Edwards]

In gmane.os.ecos.general, you wrote:
>>> what does arm-elf-size say?
>>
>> $ arm-elf-size.exe a.out
>>     text    data     bss     dec     hex filename
>>    41920    1512    8028   51460    c904 a.out
>>
>> I suppose that this means the binary will be 51460Kb in size?
>
> That means the binary will require 43432 bytes of ROM
> (41920+1512) and 9540 bytes of RAM (8028+1512).

I forgot to include stack space, so you'll need to add that in.

-- 
Grant Edwards                   grante             Yow!  Is this ANYWHERE,
                                  at               USA?
                               visi.com            

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Memory Footprint

[Reto.Bucher@elektrobit.ch]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 729 bytes --]

Hello!
 
I'm currently evaluating different operating 
systems for embedded systems. Therefore, memorysize is one major 
concern.
 
+ What minimal size of memory footprint can I 
expect when I need to include a TCP/IP stack?
 
+ is eCos a hard real time system?
 
 
bye
Reto
BEGIN:VCARD
VERSION:2.1
N:Bucher;Reto
FN:Bucher Reto
ORG:Elektrobit AG
TEL;WORK;VOICE:+41 (0)55 253 20 83
TEL;WORK;FAX:+41 (0)55 253 20 70
ADR;WORK;ENCODING=QUOTED-PRINTABLE:;;Rosswiesstrasse 29=0D=0A;Bubikon;;8608;Switzerland
LABEL;WORK;ENCODING=QUOTED-PRINTABLE:Rosswiesstrasse 29=0D=0A=0D=0ABubikon 8608=0D=0ASwitzerland
URL:
URL: http://www.elektrobit.ch
EMAIL;PREF;INTERNET:reto.bucher@elektrobit.ch
REV:20010724T055402Z
END:VCARD

[ECOS] Memory Footprint

[Reto.Bucher@elektrobit.ch]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 729 bytes --]

Hello!
 
I'm currently evaluating different operating 
systems for embedded systems. Therefore, memorysize is one major 
concern.
 
+ What minimal size of memory footprint can I 
expect when I need to include a TCP/IP stack?
 
+ is eCos a hard real time system?
 
 
bye
Reto
BEGIN:VCARD
VERSION:2.1
N:Bucher;Reto
FN:Bucher Reto
ORG:Elektrobit AG
TEL;WORK;VOICE:+41 (0)55 253 20 83
TEL;WORK;FAX:+41 (0)55 253 20 70
ADR;WORK;ENCODING=QUOTED-PRINTABLE:;;Rosswiesstrasse 29=0D=0A;Bubikon;;8608;Switzerland
LABEL;WORK;ENCODING=QUOTED-PRINTABLE:Rosswiesstrasse 29=0D=0A=0D=0ABubikon 8608=0D=0ASwitzerland
URL:
URL: http://www.elektrobit.ch
EMAIL;PREF;INTERNET:reto.bucher@elektrobit.ch
REV:20010724T055402Z
END:VCARD

Re: [ECOS] Memory Footprint

[Jonathan Larmour]

> "Reto.Bucher@elektrobit.ch" wrote:
> 
> Hello!
> 
> I'm currently evaluating different operating systems for embedded
> systems. Therefore, memorysize is one major concern.
> 
> + What minimal size of memory footprint can I expect when I need to
> include a TCP/IP stack?

Here's something I wrote before for the ARM:

-=-=-=-=-=-=-=-
From some simple measurements, the difference between including and not
including the stack appears to be about 121944 bytes code, 3944 bytes
initialized data, and 296940 bytes BSS, 256K of which is networking
buffers.

The buffer size cannot be shrunk down to near 0 though. Hugo would know
better what the minimal requirements for this are. I've quoted below some
things he wrote before that indicate that 128K is almost certainly the
absolute minimum, otherwise you may never receive 64K packets. So you can
subtract 128K off that 296940 bytes BSS to give the minimum. And as he
says, there will potentially be more overruns/lost data even with that
minimum.

This is for an app that includes read, write and select. Values may differ
more with extra functions as large sections of the stack may be brought in
due to selective linking. But probably not.
-=-=-=-=-=-=-=-

eCos needs very little above that.
 
> + is eCos a hard real time system?

No, it is soft real time. The reasoning is that a lot of people believe
they want a hard real time system, when they actually don't require it
(hard real time systems while more deterministic, tend to be a lot slower).

Jifl
-- 
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
Maybe this world is another planet's Hell -Aldous Huxley || Opinions==mine
Come to the Red Hat TechWorld open source conference in Brussels!
Keynotes, techie talks and exhibitions    http://www.redhat-techworld.com/