[ECOS] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[ECOS] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Miklos]

Hi

I am trying to port Redboot to an Embest LPCEB2000. I am using the Olimex 
H2294 port as the base. The main diferences between them are
                    Embest                    Olimex
Ext SRAM         256kx16                256kx32
Ext Flash            1Mx16                    2Mx16
Ethernet            none                            yes
LCD                none                            yes
uP                    lpc2292                    lpc2294


I changed the linker script to reflect the external ram diference (target.ld 
and the rom.ldi file)

I have edited the hal_platform_setup.h, registers BCFG0,1 and PINSEL2 values 
changed for my board

I removed ethernet drivers using the configtool

I have edited the mlt_arm_lpc2xxxxolpce2294_rom.h  for my ram size

I have changed the clock and pll multiplier (10Mhz x 6) using the configtool

However Redboot does not work. I have checked with scope to see if the Uart 
is transmitting at the wrong
baud rate, but there is no activity on bootup

Are there any other changes that I need to make?


 Miklos Pflancer



-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Sergei Gavrikov]

On Fri, 15 Sep 2006, Miklos wrote:

> Hi
>
> I am trying to port Redboot to an Embest LPCEB2000. I am using the Olimex 
> H2294 port as the base. The main diferences between them are
>                   Embest                    Olimex
> Ext SRAM         256kx16                256kx32
> Ext Flash            1Mx16                    2Mx16
> Ethernet            none                            yes
> LCD                none                            yes
> uP                    lpc2292                    lpc2294

Here you pointed on Olimex LPC-E2294 HW things. That board does have
Ethernet controller and LCD display. In reality, it's better to use eCos
HAL for the Olimex LPC-H2294 board as start point. It is a simple header
board which contains the external SRAM (1Mb) and external Intel flash
(4Mb) parts only. Use OLPH2294 eCos HAL as prototype to get eCos for
your board.  There are the ecos packages for the some Olimex LPC2294
based boards there: http://sgs.gomel.by/sg/downloads/olimex/prebuilt.
Get olpch2294-1.0.epk from an olpch2294 folder.

> I changed the linker script to reflect the external ram diference (target.ld 
> and the rom.ldi file)
>
> I have edited the hal_platform_setup.h, registers BCFG0,1 and PINSEL2 values 
> changed for my board

I suggest to care about SRAM bus timing values, the Olimex LPC-H2294
board does have fast 12nS external SRAM, but possible you have more slow
memory, therefore, check WST1, WST2, RBLE values for your this one. Be
sure what you set bus wide properly too.

> I removed ethernet drivers using the configtool
>
> I have edited the mlt_arm_lpc2xxxxolpce2294_rom.h  for my ram size

More code, more mistakes, re-engine your work. Use the olpch2294 HAL.
It has less of code lines than the olpce2294 HAL.

> I have changed the clock and pll multiplier (10Mhz x 6) using the configtool

To do new HAL port with the configtool isn't good idea ;). Setup own
entry in $ECOS_REPOSITORY/hal/arm/lpc2xxx/<your short target name>. Copy
there the olpch2294 HAL files, rename some of them properly and register
your own target in eCos database ($ECOS_REPOSITORY/ecos.db). Then you
will be able to use the configtool for your target as you do it as well.

> However Redboot does not work. I have checked with scope to see if the Uart 
> is transmitting at the wrong
> baud rate, but there is no activity on bootup

You have to be able to run the less complex things than RedBoot before.
You have to know what your HAL code works as you expect. SRAM and UART
are only 2 HW parts which need to run the RedBoot.

> Are there any other changes that I need to make?

I don't know about your flash parts too. Possible, you will need in own
flash driver too, may be, there is not a needed thing in the
$ECOS_REPOSITOTY/devs/flash/* branches. Greep ecos.db to kwnow it.

There is the bestseller for every eCos newcomer:
http://ecos.sourceware.org/docs-latest/ref/hal-porting-guide.html on
eCos official web site.

Sergei

-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Miklos.Pflancer]

Hi Segei

Thanks for your help 

I have taken your advice and used the olpch2294 as the base for my HAL port.

I have now got a Redboot prompt on my console. However I have the following error

+FLASH: driver init failed: Driver does not support device
Sorry, FLASH config exceeds available space in FIS directory

RedBoot(tm) bootstrap and debug environment [ROM]
Non-certified release, version UNKNOWN - built 11:37:54, Sep 18 2006

Platform: Embest LPCEB2000 development board (ARM7TDMI)
Copyright (C) 2000, 2001, 2002, 2003, 2004 Red Hat, Inc.
Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited

RAM: 0x81000000-0x81080000, [0x81006f20-0x8107d000] available
FLASH: 0x00000000 - 0x1, 0 blocks of 0x00000000 bytes each.
RedBoot>

At the moment I only want to have driver support for the processors internal flash. How do I achieve this?

Best Regards

Miklos Pflancer
Electronics Engineer

--------------------------------------------------
email : miklos@dbresearch.co.uk
--------------------------------------------------
dB Research Ltd
Concept House
Merton Road
Bootle
Liverpool L20 3BG
UK

Tel (+44)(0)151 330 0800
Fax (+44)(0)151 330 0808
--------------------------------------------------

-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Sergei Gavrikov]

On Mon, 18 Sep 2006, Miklos.Pflancer@dBResearch.co.uk wrote:

> Hi Segei
>
> Thanks for your help
>
> I have taken your advice and used the olpch2294 as the base for my HAL port.
>
> I have now got a Redboot prompt on my console. However I have the following error
>
> +FLASH: driver init failed: Driver does not support device
> Sorry, FLASH config exceeds available space in FIS directory
>
> RedBoot(tm) bootstrap and debug environment [ROM]
> Non-certified release, version UNKNOWN - built 11:37:54, Sep 18 2006
>
> Platform: Embest LPCEB2000 development board (ARM7TDMI)
> Copyright (C) 2000, 2001, 2002, 2003, 2004 Red Hat, Inc.
> Copyright (C) 2003, 2004, 2005, 2006 eCosCentric Limited
>
> RAM: 0x81000000-0x81080000, [0x81006f20-0x8107d000] available
> FLASH: 0x00000000 - 0x1, 0 blocks of 0x00000000 bytes each.
> RedBoot>

I think, it's normal. My redboot_ROM.ecm config for olpch2294 contains
support for an Intel 28F320C3 part:

  	package -hardware CYGPKG_DEVS_FLASH_INTEL_28FXXX v1_0 ;
  	cdl_option CYGHWR_DEVS_FLASH_INTEL_28F320C3 {
  	    inferred_value 1
  	};

but, it looks like that's unneeded for your hardware. You have to write
own flash package instead my CYGPKG_DEVS_FLASH_ARM_OLPCH2294 and remove
the reference on these things from your own target:

  	CYGPKG_DEVS_FLASH_INTEL_28FXXX
  	CYGPKG_DEVS_FLASH_ARM_OLPCH2294

But just now, build own minimal RedBoot image and the redboot_ROM.ecm
file for own hardware (without Flash support):

  	ecosconfig new <your_target_name> redboot
  	cat >redboot_ROM.ecm<<__eof
  	cdl_component CYG_HAL_STARTUP {user_value ROM};
  	__eof
  	ecosconfig import redboot_ROM.ecm
  	ecosconfig resolve
  	ecosconfig export redboot_ROM.ecm
  	ecosconfig tree
  	make

With such a command set, you'll build minimal redboot.elf (without FIS
support).  Burn and use it, and don't forget to copy that exported
`redboot_ROM.ecm' file in your `misc' directory to use it a long while.

> At the moment I only want to have driver support for the processors internal flash. How do I achieve this?

So far, it's better to hold the RedBoot there (i.e. in the on-chip flash
memory) and using the RedBoot, load your apps in RAM via x/y modem or
using a GDB. Next what you have to achive, it's to setup up the FIS for
your external flash memory parts. If you are sorry about rest of the
on-chip flash place now, in fact, when you'll finish a debugging your
application in RAM, you'll be able to place its code in the on-chip
flash memory then.

Sergei


-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Miklos.Pflancer]

Hi Segei

Once again thanks for your help, I now have Redboot working without flash support. However I would like to implement flash support for my board. My board contains Atmel AT49 flash memory. I have tried using the configtools to include this in my ecos project but I get the following message from configtool

'Add and remove hardware packages by selecting a new hardware template'

Any ideas on how I can do this?

Thanks

Miklos Pflancer
Electronics Engineer

--------------------------------------------------
email : miklos@dbresearch.co.uk
--------------------------------------------------
dB Research Ltd
Concept House
Merton Road
Bootle
Liverpool L20 3BG
UK

Tel (+44)(0)151 330 0800
Fax (+44)(0)151 330 0808
--------------------------------------------------

-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Andrew Lunn]

On Tue, Sep 19, 2006 at 09:30:16AM +0100, Miklos.Pflancer@dBResearch.co.uk wrote:
> Hi Segei
> 
> Once again thanks for your help, I now have Redboot working without flash support. However I would like to implement flash support for my board. My board contains Atmel AT49 flash memory. I have tried using the configtools to include this in my ecos project but I get the following message from configtool
> 
> 'Add and remove hardware packages by selecting a new hardware template'
> 
> Any ideas on how I can do this?

You need to add the flash package to your target in the ecos.db file. eg
The AT91 eb40 target is:

target eb40a {
        alias { "Atmel evaluation board (EB40A)" at91_eb40a }
        packages { CYGPKG_HAL_ARM
                   CYGPKG_HAL_ARM_AT91
                   CYGPKG_HAL_ARM_AT91_EB40A
                   CYGPKG_DEVS_FLASH_EB40A
                   CYGPKG_DEVS_FLASH_ATMEL_AT49XXXX
                   CYGPKG_IO_SERIAL_ARM_AT91
                   CYGPKG_DEVICES_WATCHDOG_ARM_AT91
        }
        description "
        The eb40a target provides the packages needed to run eCos on an Atmel
        evaluation board (EB40A)."
}

Here you see the generic AT49XXXX package and the board specific
CYGPKG_DEVS_FLASH_EB40A package. You will need to write the board
specific package for your particular board. You might want to use the
EB40A code as an template, /packages/devs/flash/arm/eb40a/current/

     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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Miklos.Pflancer]

Thanks Andrew

I think I have FIS support now, thanks for your advice.

I still have a problems executing code loaded using GDB when interrupts are enabled. GDB hangs and so does my code.when it hits the line of code which enables interrupts

Any idea why this might be the case?

Best Regards


Miklos Pflancer
Electronics Engineer

--------------------------------------------------
email : miklos@dbresearch.co.uk
--------------------------------------------------
dB Research Ltd
Concept House
Merton Road
Bootle
Liverpool L20 3BG
UK

Tel (+44)(0)151 330 0800
Fax (+44)(0)151 330 0808
--------------------------------------------------

-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Andrew Lunn]

On Tue, Sep 19, 2006 at 12:30:56PM +0100, Miklos.Pflancer@dBResearch.co.uk wrote:
> Thanks Andrew
> 
> I think I have FIS support now, thanks for your advice.
> 
> I still have a problems executing code loaded using GDB when
> interrupts are enabled. GDB hangs and so does my code.when it hits
> the line of code which enables interrupts
> 
> Any idea why this might be the case?

Presumably some device is generating an interrupt. Check the AIC and
see which device has an interrupt pending. 

    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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Miklos.Pflancer]

Hi Andrew

Ive been looking in the hal_platform_setup and it seems as though 
Redboot copies the first 64 bytes of vectors to internal SRAM. An 
then remaps the vectors  to internal SRAM. 

What then happens if an interrupt occurs in my code, How is it 
handled by Redboot.

Should I copy by vectors to this location?

I am a bit confused

Hope you can help

Best regards

Miklos Pflancer
Electronics Engineer

--------------------------------------------------
email : miklos@dbresearch.co.uk
--------------------------------------------------
dB Research Ltd
Concept House
Merton Road
Bootle
Liverpool L20 3BG
UK

Tel (+44)(0)151 330 0800
Fax (+44)(0)151 330 0808
--------------------------------------------------

-- 
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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Andrew Lunn]

On Tue, Sep 19, 2006 at 03:29:56PM +0100, Miklos.Pflancer@dBResearch.co.uk wrote:
> Hi Andrew
> 
> Ive been looking in the hal_platform_setup and it seems as though 
> Redboot copies the first 64 bytes of vectors to internal SRAM. An 
> then remaps the vectors  to internal SRAM. 
> 
> What then happens if an interrupt occurs in my code, How is it 
> handled by Redboot.
> 
> Should I copy by vectors to this location?
> 
> I am a bit confused

Is you application an eCos application? If so eCos will do this
itself.
 
However if you are writing a standalone application which does not use
eCos, you have to look after interrupts yourself.

      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] Is it possible to debug assembly code on source level?

[wang cui]

As I debug code by GNU Insight debugger, it works good when debug C/C++ 
code. But if I steps into assembly code segment, it only shows assembly 
without any cross information with original assembly source file(like 
vectors.s, etc.).
Althouth I can look over the assembly source outside the debugger and find 
out what is running. I think it is not convenience(since there are lots of 
#ifdef in the assembly source).
Can't Insight debug on assembly source level?



-- 
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] How to integrate two different 16x5x DUART hardware into a application?

[wang cui]

On my target board, I use LPC2220 as processor, there is another Phlips 
16C2550B DUART chip.
In my application, I should make use of 4 UART channels.

As I looked over the serial infrastructure of eCos, I think it is hard to 
use the "16x5x generic serial device drivers" for both chip concurrently.
1. The "ser_16x5x.c" rely on a "CYGDAT_IO_SERIAL_GENERIC_16X5X_INL", and 
this macro has been defined in "ARM LPC2XXX serial device drivers" package. 
How to add 16C2550B's description info to "ser_16x5x.c"?
2. "ser_16x5x.c" use "CYGPRI_IO_SERIAL_GENERIC_16X5X_STEP" macro to reflect 
delta of UART registers. In LPC2XXX, the UART's register step is 4 bytes, 
while on my target board, 16C2550B has been designed with register step 2 
bytes. How to solve this condition?

Thanks.



-- 
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] How to integrate two different 16x5x DUART hardware into a application?

[Andrew Lunn]

On Wed, Sep 20, 2006 at 09:47:42AM +0000, wang cui wrote:
> On my target board, I use LPC2220 as processor, there is another Phlips 
> 16C2550B DUART chip.
> In my application, I should make use of 4 UART channels.
> 
> As I looked over the serial infrastructure of eCos, I think it is hard to 
> use the "16x5x generic serial device drivers" for both chip concurrently.
> 1. The "ser_16x5x.c" rely on a "CYGDAT_IO_SERIAL_GENERIC_16X5X_INL", and 
> this macro has been defined in "ARM LPC2XXX serial device drivers" package. 
> How to add 16C2550B's description info to "ser_16x5x.c"?
> 2. "ser_16x5x.c" use "CYGPRI_IO_SERIAL_GENERIC_16X5X_STEP" macro to reflect 
> delta of UART registers. In LPC2XXX, the UART's register step is 4 bytes, 
> while on my target board, 16C2550B has been designed with register step 2 
> bytes. How to solve this condition?

A quick and dirty solution would be to duplicate the
CYGPKG_IO_SERIAL_GENERIC_16X5X package. I.E. make a second copy with
the name CYGPKG_IO_SERIAL_GENERIC_16X5X_DUP and rename all the CDL
variables in the second copy in the same way. 

Doing a clean solution so that one instance of the driver can drive
multiple 16x5 chips is going to take quiet a bit of work. 

         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] Port eCos Redboot Olimex LPC2294 to Embest LPCEB2000 (lpc2292)

[Miklos.Pflancer]

Hi Andrew

I still thinks there is a problem with the gdb stubs in redboot. I 
can load and run a  standalone application which uses interrupts into 
external ram  using gdb, however If I set breakpoints or try to 
single step the application stops and gdb becomes unresponsive. 

I can burn the same application into flash and load with Redboot and 
this works fine.

I was wondering if  there are options in the configtool which should 
be selected but are'nt. I am not familiar with many of these 
settings. 

Best Regards

Miklos



> On Tue, Sep 19, 2006 at 03:29:56PM +0100, 
Miklos.Pflancer@dBResearch.co.uk wrote:
> > Hi Andrew
> > 
> > Ive been looking in the hal_platform_setup and it seems as though 

> > Redboot copies the first 64 bytes of vectors to internal SRAM. An 

> > then remaps the vectors  to internal SRAM. 
> > 
> > What then happens if an interrupt occurs in my code, How is it 
> > handled by Redboot.
> > 
> > Should I copy by vectors to this location?
> > 
> > I am a bit confused
> 
> Is you application an eCos application? If so eCos will do this
> itself.
>  
> However if you are writing a standalone application which does not 
use
> eCos, you have to look after interrupts yourself.
> 
>       Andrew
> 


Miklos Pflancer
Electronics Engineer

--------------------------------------------------
email : miklos@dbresearch.co.uk
--------------------------------------------------
dB Research Ltd
Concept House
Merton Road
Bootle
Liverpool L20 3BG
UK

Tel (+44)(0)151 330 0800
Fax (+44)(0)151 330 0808
--------------------------------------------------

-- 
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] "RedBoot" configuration appears in "default" template?

[wang cui]

For example:
1.use "eCos Configuration Tool" to new a project
2.choose "Build->Templates..."
3."Hardware" = "Olimex evaluation board LPC-P2106 "; "Packages" = 
"default", then OK

In the configuration tree, there appears "Redboot for ARM options" and 
"Redboot HAL options". Of course, there are disabled.

Although it can be built successfuly, I think it is confused and 
unreasonable. :(



-- 
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] Can I also generate other outputs of gcc when build?

[wang cui]

As eCos Configuration Tool generate ".o" file for each source file 
defaultly. 
Manytimes I need to check the preprocessed output(sometimes compiled 
assembly output) of source file.

By now, I have to manually call gcc, in a bash shell, with a additional "-E 
-o xxx.i" option to generate the preprocessed output for check.

For example:
arm-elf-gcc -c  
-I/ecos-c/cygwin/opt/development/project/cpu_4c_lib/lib_install/include 
-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current 
-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src 
-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/tests 
-I. -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/ 
-finline-limit=7000 -mcpu=arm7tdmi -mno-short-load-words -Wall 
-Wpointer-arith -Wstrict-prototypes -Winline -Wundef  -g -O0 
-ffunction-sections -fdata-sections  -fno-exceptions   
-Wp,-MD,src/ser_16x5x.tmp -E -o src/devs_serial_generic_16x5x_ser_16x5x.i 
/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/ser_16x5x.c


So I am asking can I do something to let eCos Configuration Tool generate 
NOT ONLY ".o", BUT ALSO other outputs(.i, .s, etc.)?



-- 
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] Can I also generate other outputs of gcc when build?

[Gary Thomas]

wang cui wrote:
> As eCos Configuration Tool generate ".o" file for each source file 
> defaultly. Manytimes I need to check the preprocessed output(sometimes 
> compiled assembly output) of source file.
> 
> By now, I have to manually call gcc, in a bash shell, with a additional 
> "-E -o xxx.i" option to generate the preprocessed output for check.
> 
> For example:
> arm-elf-gcc -c  
> -I/ecos-c/cygwin/opt/development/project/cpu_4c_lib/lib_install/include 
> -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current 
> -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src 
> -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/tests 
> -I. 
> -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/ 
> -finline-limit=7000 -mcpu=arm7tdmi -mno-short-load-words -Wall 
> -Wpointer-arith -Wstrict-prototypes -Winline -Wundef  -g -O0 
> -ffunction-sections -fdata-sections  -fno-exceptions   
> -Wp,-MD,src/ser_16x5x.tmp -E -o 
> src/devs_serial_generic_16x5x_ser_16x5x.i 
> /opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/ser_16x5x.c 
> 
> 
> 
> So I am asking can I do something to let eCos Configuration Tool 
> generate NOT ONLY ".o", BUT ALSO other outputs(.i, .s, etc.)?

Just add "-save-outputs" to your global GCC options.

Note: I normally do this by hand for the file(s) I'm interested in as
turning it on in a build tree will create tons of files, some of which
can be quite large...

-- 
------------------------------------------------------------
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

Re: [ECOS] Can I also generate other outputs of gcc when build?

[wang cui]

Thanks, but "-save-outputs" seems unrecognized when invoke arm-elf-gcc.
For example:
arm-elf-gcc -c  
-I/ecos-c/cygwin/opt/development/project/cpu_4c_lib/lib_install/include 
-I/opt/development/ecos/packages/language/c/libc/string/current 
-I/opt/development/ecos/packages/language/c/libc/string/current/src 
-I/opt/development/ecos/packages/language/c/libc/string/current/tests -I. 
-I/opt/development/ecos/packages/language/c/libc/string/current/src/ 
-finline-limit=7000 -mcpu=arm7tdmi -mno-short-load-words -Wall 
-Wpointer-arith  -Winline -Wundef -Woverloaded-virtual -g -O0 
-ffunction-sections -fdata-sections -fno-rtti -fno-exceptions   
-save-outputs -Wp,-MD,src/strcoll.tmp -o 
src/language_c_libc_string_strcoll.o 
/opt/development/ecos/packages/language/c/libc/string/current/src/strcoll.cxx

arm-elf-gcc: unrecognized option `-save-outputs'

I am using the prebuilt arm-elf-gcc from eCosCentric, the version is 3.2.1



>From: Gary Thomas <gary@mlbassoc.com>
>To: wang cui <iucgnaw@msn.com>
>CC: ecos-discuss@ecos.sourceware.org
>Subject: Re: [ECOS] Can I also generate other outputs of gcc when build?
>Date: Fri, 22 Sep 2006 03:47:42 -0600
>
>wang cui wrote:
> > As eCos Configuration Tool generate ".o" file for each source file
> > defaultly. Manytimes I need to check the preprocessed output(sometimes
> > compiled assembly output) of source file.
> >
> > By now, I have to manually call gcc, in a bash shell, with a additional
> > "-E -o xxx.i" option to generate the preprocessed output for check.
> >
> > For example:
> > arm-elf-gcc -c
> > -I/ecos-c/cygwin/opt/development/project/cpu_4c_lib/lib_install/include
> > -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current
> > -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src
> > 
-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/tests
> > -I.
> > -I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/
> > -finline-limit=7000 -mcpu=arm7tdmi -mno-short-load-words -Wall
> > -Wpointer-arith -Wstrict-prototypes -Winline -Wundef  -g -O0
> > -ffunction-sections -fdata-sections  -fno-exceptions
> > -Wp,-MD,src/ser_16x5x.tmp -E -o
> > src/devs_serial_generic_16x5x_ser_16x5x.i
> > 
/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/ser_16x5x.c

> >
> >
> >
> > So I am asking can I do something to let eCos Configuration Tool
> > generate NOT ONLY ".o", BUT ALSO other outputs(.i, .s, etc.)?
>
>Just add "-save-outputs" to your global GCC options.
>
>Note: I normally do this by hand for the file(s) I'm interested in as
>turning it on in a build tree will create tons of files, some of which
>can be quite large...
>
>--
>------------------------------------------------------------
>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

Re: [ECOS] Can I also generate other outputs of gcc when build?

[wang cui]

However, I have done it by add make steps to the 
"\packages\pkgconf\rules.mak". Such as:

%.o.d : %.c
ifeq ($(HOST),CYGWIN)
	@mkdir -p `cygpath -w "$(dir $@)" | sed "s@\\\\\\\\@/@g"`
else
	@mkdir -p $(dir $@)
endif
	$(CC) -c $(INCLUDE_PATH) -I$(dir $<) $(ACTUAL_CFLAGS) 
-Wp,-MD,$(@:.o.d=.tmp) -E -o $(dir $@)$(OBJECT_PREFIX)_$(notdir 
$(@:.o.d=.i)) $<
	$(CC) -c $(INCLUDE_PATH) -I$(dir $<) $(ACTUAL_CFLAGS) 
-Wp,-MD,$(@:.o.d=.tmp) -o $(dir $@)$(OBJECT_PREFIX)_$(notdir $(@:.o.d=.o)) 
$<
	@sed -e '/^ *\\/d' -e "s#.*: #$@: #" $(@:.o.d=.tmp) > $@
	@rm $(@:.o.d=.tmp)

So it can generate ".i" file for every ".c".

Thanks!

>From: "wang cui" <iucgnaw@msn.com>
>To: gary@mlbassoc.com
>CC: ecos-discuss@ecos.sourceware.org
>Subject: Re: [ECOS] Can I also generate other outputs of gcc when build?
>Date: Mon, 25 Sep 2006 04:02:55 +0000
>
>Thanks, but "-save-outputs" seems unrecognized when invoke 
>arm-elf-gcc.
>For example:
>arm-elf-gcc -c  
>-I/ecos-c/cygwin/opt/development/project/cpu_4c_lib/lib_install/include 
>-I/opt/development/ecos/packages/language/c/libc/string/current 
>-I/opt/development/ecos/packages/language/c/libc/string/current/src 
>-I/opt/development/ecos/packages/language/c/libc/string/current/tests 
>-I. 
>-I/opt/development/ecos/packages/language/c/libc/string/current/src/ 
>-finline-limit=7000 -mcpu=arm7tdmi -mno-short-load-words -Wall 
>-Wpointer-arith  -Winline -Wundef -Woverloaded-virtual -g -O0 
>-ffunction-sections -fdata-sections -fno-rtti -fno-exceptions   
>-save-outputs -Wp,-MD,src/strcoll.tmp -o 
>src/language_c_libc_string_strcoll.o 
>/opt/development/ecos/packages/language/c/libc/string/current/src/strcoll.cxx

>
>arm-elf-gcc: unrecognized option `-save-outputs'
>
>I am using the prebuilt arm-elf-gcc from eCosCentric, the version is 
>3.2.1
>
>
>
>>From: Gary Thomas <gary@mlbassoc.com>
>>To: wang cui <iucgnaw@msn.com>
>>CC: ecos-discuss@ecos.sourceware.org
>>Subject: Re: [ECOS] Can I also generate other outputs of gcc when 
>>build?
>>Date: Fri, 22 Sep 2006 03:47:42 -0600
>>
>>wang cui wrote:
>> > As eCos Configuration Tool generate ".o" file for each source 
>>file
>> > defaultly. Manytimes I need to check the preprocessed 
>>output(sometimes
>> > compiled assembly output) of source file.
>> >
>> > By now, I have to manually call gcc, in a bash shell, with a 
>>additional
>> > "-E -o xxx.i" option to generate the preprocessed output for 
>>check.
>> >
>> > For example:
>> > arm-elf-gcc -c
>> > 
>>-I/ecos-c/cygwin/opt/development/project/cpu_4c_lib/lib_install/include
>> > 
>>-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current
>> > 
>>-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src
>> >
>-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/tests
>> > -I.
>> > 
>>-I/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/
>> > -finline-limit=7000 -mcpu=arm7tdmi -mno-short-load-words -Wall
>> > -Wpointer-arith -Wstrict-prototypes -Winline -Wundef  -g -O0
>> > -ffunction-sections -fdata-sections  -fno-exceptions
>> > -Wp,-MD,src/ser_16x5x.tmp -E -o
>> > src/devs_serial_generic_16x5x_ser_16x5x.i
>> >
>/opt/development/ecos/packages/devs/serial/generic/16x5x/current/src/ser_16x5x.c

>
>> >
>> >
>> >
>> > So I am asking can I do something to let eCos Configuration Tool
>> > generate NOT ONLY ".o", BUT ALSO other outputs(.i, .s, etc.)?
>>
>>Just add "-save-outputs" to your global GCC options.
>>
>>Note: I normally do this by hand for the file(s) I'm interested in 
>>as
>>turning it on in a build tree will create tons of files, some of 
>>which
>>can be quite large...
>>
>>--
>>------------------------------------------------------------
>>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
>



-- 
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] Is it possible to debug assembly code on source level?

[Bart Veer]

>>>>> " " == wang cui <iucgnaw@msn.com> writes:

     > As I debug code by GNU Insight debugger, it works good when
     > debug C/C++ code. But if I steps into assembly code segment, it
     > only shows assembly without any cross information with original
     > assembly source file(like vectors.s, etc.). Althouth I can look
     > over the assembly source outside the debugger and find out what
     > is running. I think it is not convenience(since there are lots
     > of #ifdef in the assembly source). Can't Insight debug on
     > assembly source level?

The debugger can only show information such as the source file and
line number if the ELF executable contains the appropriate debug
information. If you compile some C code with -S -g and look at the
assembler code produced you will see lots of debug-related directives
like .type, .file and .loc. It is these which end up allowing the
debugger to do its work.

For hand-written assembly code the assembler does not generate such
directives automatically. Doing so would be quite hard. Assembler
source code can be an arbitrary mixture of code and data interleaved
in various ways, so the assembler has no easy way of knowing what
debug info it should generate.

Instead if you want source-level debugging of assembler code then you
will need to add the appropriate debug directives into that code. Just
specifying .type on a function entry point seems to be enough to get
the filename into the executable. To get line numbers into the debug
info you will probably need a .debug_line section and .file and .loc
directives, but I do not know the details.

Bart

-- 
Bart Veer                       eCos Configuration Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts

-- 
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] CYG_ASSERTCLASS() fail when use wallclock->dallas->ds1307 device driver

[Wang Cui]

Hi, all
As I use the ds1307 driver on my target, I configure to use the generic
I2C API to access the ds1307 chip.
the application failed in a assertion. I debugged it and find the cause, see
the call stack below:
-cyg_hal_invoke_constructors()
-Cyg_WallClock::Cyg_WallClock()
-init_hw_seconds()
-init_ds_hwclock()
-DS_GET()
-cyg_i2c_transaction_begin()
-cyg_mutex_lock()
-Cyg_Mutex::lock()
-CYG_ASSERTCLASS( this, "Bad this pointer");
- Cyg_Mutex::check_this()
-line198:        if(  locked && owner == NULL ) return false;
Here, 'owner' of the mutex is NULL, cause assertion failed.

Since the driver is initialized in class construction stage before scheduler
runs, there is no thread context yet, then the owner is a 'NULL' thread.

So how to handle it?


-- 
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] Can I also generate other outputs of gcc when build?

[Simon]

On 22/09/2006, at 7:47 PM, Gary Thomas gary-at-mlbassoc.com |ecos| wrote:

> > 
> > So I am asking can I do something to let eCos Configuration Tool 
> > generate NOT ONLY ".o", BUT ALSO other outputs(.i, .s, etc.)?
> 
> Just add "-save-outputs" to your global > GCC options.

Actually, I think the option to do this in gcc is "-save-temps".

Regards,

Simon

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