Small Device C Compiler suite

Is there any one can help me?

On Sat, Nov 26, 2005 at 12:37:53AM +0800, Mr.Kissinger wrote:
Re:Hi,
Re:	I am using SDCC on Winbond W78E516BP-40 chip.
Re:	When the target code (binary code) is shorter than 32K, it runs well.
Re:	But when the code length beyonds 32K, even 1K longer, it doesn't work at all.
Re:	Have any one resolved this problem?
Re:
Re:	My compiling option are followed:
Re:	sdcc  -mmcs51 --stack-auto --use-stdout --debug --opt-code-speed  -I/usr/local/share/sdcc/include -Iincludes  -D_DEBUG -DSDCC --model-large --code-size 65536 --iram-size 256 --xram-size 65536 -I/usr/local/share/sdcc/include
Re:	
Re:
Re:	Winbond W78E516BP-40 features (from official specification):
Re:
Re:Fully static design 8-bit CMOS microcontroller
Re:64K bytes of in-system programmable Flash EPROM for Application Program (APROM)
Re:4K bytes of auxiliary Flash EPROM for Loader Program (LDROM)
Re:512 bytes of on-chip RAM. (including 256 bytes of AUX-RAM, software selectable)
Re:64K bytes program memory address space and 64K bytes data memory address space
Re:Four 8-bit bi-directional ports
Re:One 4-bit multipurpose programmable port
Re:Three 16-bit timer/counters
Re:One full duplex serial port
Re:Eight-sources, two-level interrupt capability
Re:Built-in power management
Re:Code protection
Re:Packaged in
Re:  DIP 40: W78LE516-24
Re:  PLCC 44: W78LE516P-24
Re:  QFP 44: W78LE516F-24
Re:  LQFP 48: W78LE516D-24
Re:
Re:
Re:	Thanks.
Re:Hyacinth
Re:
Re:
Re:-------------------------------------------------------
Re:This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
Re:for problems?  Stop!  Download the new AJAX search engine that makes
Re:searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:_______________________________________________
Re:Sdcc-user mailing list
Re:Sdcc-user@...
Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user
Re:

Hello Mr.Kissinger,

MK> Is there any one can help me?

Yes, you! :-)

This is a bit like asking "Why is my PC crashing?". You need to write
some simple test code to explore the boundaries of the problem. And
you need to use what test gear is available to you, or what you can
string together. This can be as simple as a couple of LEDs.

Is it a hardware problem or is it a software problem? Does *all* the
code actually get flashed properly? etc. etc. Somebody may just have a
magic bullet i.e. have had a similar experience and offer an insight;
but you've painted things on such a broad canvas that it seems
unlikely - and that's painting-by-numbers and not art!

Analysing problems like this is what being an engineer is all about.
Rise to the challenge, and prove to yourself that you are one!

*J*

Saturday, November 26, 2005, 1:29:44 PM, you wrote:

MK> On Sat, Nov 26, 2005 at 12:37:53AM +0800, Mr.Kissinger wrote:
MK> Re:Hi, Re:	I am using SDCC on Winbond W78E516BP-40 chip. Re:	When
MK> the target code (binary code) is shorter than 32K, it runs well.
MK> Re:	But when the code length beyonds 32K, even 1K longer, it
MK> doesn't work at all. Re:	Have any one resolved this problem? Re:
MK> Re:	My compiling option are followed: Re:	sdcc  -mmcs51
MK> --stack-auto --use-stdout --debug --opt-code-speed
MK> -I/usr/local/share/sdcc/include -Iincludes  -D_DEBUG -DSDCC
MK> --model-large --code-size 65536 --iram-size 256 --xram-size 65536
MK> -I/usr/local/share/sdcc/include Re:	 Re: Re:	Winbond W78E516BP-40
MK> features (from official specification): Re: Re:Fully static design
MK> 8-bit CMOS microcontroller Re:64K bytes of in-system programmable
MK> Flash EPROM for Application Program (APROM) Re:4K bytes of
MK> auxiliary Flash EPROM for Loader Program (LDROM) Re:512 bytes of
MK> on-chip RAM. (including 256 bytes of AUX-RAM, software selectable)
MK> Re:64K bytes program memory address space and 64K bytes data
MK> memory address space Re:Four 8-bit bi-directional ports Re:One
MK> 4-bit multipurpose programmable port Re:Three 16-bit
MK> timer/counters Re:One full duplex serial port Re:Eight-sources,
MK> two-level interrupt capability Re:Built-in power management
MK> Re:Code protection Re:Packaged in Re:  DIP 40: W78LE516-24 Re:
MK> PLCC 44: W78LE516P-24 Re:  QFP 44: W78LE516F-24 Re:  LQFP 48:
MK> W78LE516D-24 Re: Re: Re:	Thanks. Re:Hyacinth Re: Re:
MK> Re:------------------------------------------------------- Re:This
MK> SF.net email is sponsored by: Splunk Inc. Do you grep through log
MK> files Re:for problems?  Stop!  Download the new AJAX search engine
MK> that makes Re:searching your log files as easy as surfing the
MK> web.  DOWNLOAD SPLUNK!
MK> Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
MK> Re:_______________________________________________ Re:Sdcc-user
MK> mailing list Re:Sdcc-user@...
MK> Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user Re:


MK> ------------------------------------------------------- This
MK> SF.net email is sponsored by: Splunk Inc. Do you grep through log
MK> files for problems?  Stop!  Download the new AJAX search engine
MK> that makes searching your log files as easy as surfing the  web.
MK> DOWNLOAD SPLUNK!
MK> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
MK> _______________________________________________ Sdcc-user mailing
MK> list Sdcc-user@...
MK> https://lists.sourceforge.net/lists/listinfo/sdcc-user

Mr.Kissinger wrote:
> Re:	I am using SDCC on Winbond W78E516BP-40 chip.
> Re:	When the target code (binary code) is shorter than 32K, it runs well.
> Re:	But when the code length beyonds 32K, even 1K longer, it doesn't work at all.
> Re:	Have any one resolved this problem?

Hi,
don't know whether it helps, but there's no limitiation
at 32k imposed by SDCC.

You can put that limit you're seeing under a closer test
by using a "less than 32k running well version" and
inserting some NOPs with

_asm
   NOP
   NOP
   NOP
   NOP
   ; many more nops to get a binary larger than 32k
_endasm;

in the main() routine.

Also check whether the 1K additional code you're adding
doesn't need too much additional data/idata space (so
that the stack still is large enough).

Greetings,
Frieder

I added a lot of "NOP" at the beginning of main().
When the code size is less than 32K, or even 32769 (32K+1) bytes, the code works.
When I added more "NOP", the code became to 32862 (32K+94) bytes, it run failed.

On Sun, Nov 27, 2005 at 02:12:38PM +0100, Frieder Ferlemann wrote:
Re:Mr.Kissinger wrote:
Re:>Re:	I am using SDCC on Winbond W78E516BP-40 chip.
Re:>Re:	When the target code (binary code) is shorter than 32K, it runs well.
Re:>Re:	But when the code length beyonds 32K, even 1K longer, it doesn't 
Re:>work at all.
Re:>Re:	Have any one resolved this problem?
Re:
Re:Hi,
Re:don't know whether it helps, but there's no limitiation
Re:at 32k imposed by SDCC.
Re:
Re:You can put that limit you're seeing under a closer test
Re:by using a "less than 32k running well version" and
Re:inserting some NOPs with
Re:
Re:_asm
Re:  NOP
Re:  NOP
Re:  NOP
Re:  NOP
Re:  ; many more nops to get a binary larger than 32k
Re:_endasm;
Re:
Re:in the main() routine.
Re:
Re:Also check whether the 1K additional code you're adding
Re:doesn't need too much additional data/idata space (so
Re:that the stack still is large enough).
Re:
Re:Greetings,
Re:Frieder
Re:
Re:
Re:-------------------------------------------------------
Re:This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
Re:for problems?  Stop!  Download the new AJAX search engine that makes
Re:searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:_______________________________________________
Re:Sdcc-user mailing list
Re:Sdcc-user@...
Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user
Re:

Yes, I tried hard before I asked.
I cannot even initialized the LEDs.

On Sun, Nov 27, 2005 at 12:25:00PM +0000, Tipe Def wrote:
Re:Hello Mr.Kissinger,
Re:
Re:MK> Is there any one can help me?
Re:
Re:Yes, you! :-)
Re:
Re:This is a bit like asking "Why is my PC crashing?". You need to write
Re:some simple test code to explore the boundaries of the problem. And
Re:you need to use what test gear is available to you, or what you can
Re:string together. This can be as simple as a couple of LEDs.
Re:
Re:Is it a hardware problem or is it a software problem? Does *all* the
Re:code actually get flashed properly? etc. etc. Somebody may just have a
Re:magic bullet i.e. have had a similar experience and offer an insight;
Re:but you've painted things on such a broad canvas that it seems
Re:unlikely - and that's painting-by-numbers and not art!
Re:
Re:Analysing problems like this is what being an engineer is all about.
Re:Rise to the challenge, and prove to yourself that you are one!
Re:
Re:*J*
Re:
Re:Saturday, November 26, 2005, 1:29:44 PM, you wrote:
Re:
Re:MK> On Sat, Nov 26, 2005 at 12:37:53AM +0800, Mr.Kissinger wrote:
Re:MK> Re:Hi, Re:	I am using SDCC on Winbond W78E516BP-40 chip. Re:	When
Re:MK> the target code (binary code) is shorter than 32K, it runs well.
Re:MK> Re:	But when the code length beyonds 32K, even 1K longer, it
Re:MK> doesn't work at all. Re:	Have any one resolved this problem? Re:
Re:MK> Re:	My compiling option are followed: Re:	sdcc  -mmcs51
Re:MK> --stack-auto --use-stdout --debug --opt-code-speed
Re:MK> -I/usr/local/share/sdcc/include -Iincludes  -D_DEBUG -DSDCC
Re:MK> --model-large --code-size 65536 --iram-size 256 --xram-size 65536
Re:MK> -I/usr/local/share/sdcc/include Re:	 Re: Re:	Winbond W78E516BP-40
Re:MK> features (from official specification): Re: Re:Fully static design
Re:MK> 8-bit CMOS microcontroller Re:64K bytes of in-system programmable
Re:MK> Flash EPROM for Application Program (APROM) Re:4K bytes of
Re:MK> auxiliary Flash EPROM for Loader Program (LDROM) Re:512 bytes of
Re:MK> on-chip RAM. (including 256 bytes of AUX-RAM, software selectable)
Re:MK> Re:64K bytes program memory address space and 64K bytes data
Re:MK> memory address space Re:Four 8-bit bi-directional ports Re:One
Re:MK> 4-bit multipurpose programmable port Re:Three 16-bit
Re:MK> timer/counters Re:One full duplex serial port Re:Eight-sources,
Re:MK> two-level interrupt capability Re:Built-in power management
Re:MK> Re:Code protection Re:Packaged in Re:  DIP 40: W78LE516-24 Re:
Re:MK> PLCC 44: W78LE516P-24 Re:  QFP 44: W78LE516F-24 Re:  LQFP 48:
Re:MK> W78LE516D-24 Re: Re: Re:	Thanks. Re:Hyacinth Re: Re:
Re:MK> Re:------------------------------------------------------- Re:This
Re:MK> SF.net email is sponsored by: Splunk Inc. Do you grep through log
Re:MK> files Re:for problems?  Stop!  Download the new AJAX search engine
Re:MK> that makes Re:searching your log files as easy as surfing the
Re:MK> web.  DOWNLOAD SPLUNK!
Re:MK> Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:MK> Re:_______________________________________________ Re:Sdcc-user
Re:MK> mailing list Re:Sdcc-user@...
Re:MK> Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user Re:
Re:
Re:
Re:MK> ------------------------------------------------------- This
Re:MK> SF.net email is sponsored by: Splunk Inc. Do you grep through log
Re:MK> files for problems?  Stop!  Download the new AJAX search engine
Re:MK> that makes searching your log files as easy as surfing the  web.
Re:MK> DOWNLOAD SPLUNK!
Re:MK> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:MK> _______________________________________________ Sdcc-user mailing
Re:MK> list Sdcc-user@...
Re:MK> https://lists.sourceforge.net/lists/listinfo/sdcc-user
Re:
Re:
Re:
Re:
Re:-------------------------------------------------------
Re:This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
Re:for problems?  Stop!  Download the new AJAX search engine that makes
Re:searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:_______________________________________________
Re:Sdcc-user mailing list
Re:Sdcc-user@...
Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user
Re:

Hello Mr.Kissinger,

If you can execute good code out of the bottom 32KB then you're more
than half-way to diagnosing the problem. Write some code to dump the
flash to the UART or a soft UART (toggling any old spare I/O pin).
You'll need a TTL/RS232 level changer to dive a PC's serial port, and
old dumb terminal (VDU) can be a godsend here. Maybe your hardware
already has RS232 or a display you can press into service?

Are the ROM contents *both* above and below the 32KB boundary exactly
what you'd expect. You can get rather good at reading Intel hex doing
this! :-)

What are you using to flash the chips? Do you have an alternative
programmer - preferably of different design? How many chips have you
tried? Does the programmer have a chip verify/chip dump mode? Can you
trust it? Have you anything hanging on A15 - isolate it?

Often it's not a waste of code space to write a ROM checksum routine that
runs on a cold start. Do it in assembler to minimise the resources
that need to work to get a result.

Small logical steps and you'll get there!

Regards
*J*

Tuesday, November 29, 2005, 11:39:51 AM, you wrote:

MK> Yes, I tried hard before I asked.
MK> I cannot even initialized the LEDs.

MK> On Sun, Nov 27, 2005 at 12:25:00PM +0000, Tipe Def wrote:
MK> Re:Hello Mr.Kissinger,
MK> Re:
Re:MK>> Is there any one can help me?
MK> Re:
MK> Re:Yes, you! :-)
MK> Re:
MK> Re:This is a bit like asking "Why is my PC crashing?". You need to write
MK> Re:some simple test code to explore the boundaries of the problem. And
MK> Re:you need to use what test gear is available to you, or what you can
MK> Re:string together. This can be as simple as a couple of LEDs.
MK> Re:
MK> Re:Is it a hardware problem or is it a software problem? Does *all* the
MK> Re:code actually get flashed properly? etc. etc. Somebody may just have a
MK> Re:magic bullet i.e. have had a similar experience and offer an insight;
MK> Re:but you've painted things on such a broad canvas that it seems
MK> Re:unlikely - and that's painting-by-numbers and not art!
MK> Re:
MK> Re:Analysing problems like this is what being an engineer is all about.
MK> Re:Rise to the challenge, and prove to yourself that you are one!
MK> Re:
MK> Re:*J*
MK> Re:
MK> Re:Saturday, November 26, 2005, 1:29:44 PM, you wrote:
MK> Re:
Re:MK>> On Sat, Nov 26, 2005 at 12:37:53AM +0800, Mr.Kissinger wrote:
Re:MK>> Re:Hi, Re:	I am using SDCC on Winbond W78E516BP-40 chip. Re:	When
Re:MK>> the target code (binary code) is shorter than 32K, it runs well.
Re:MK>> Re:	But when the code length beyonds 32K, even 1K longer, it
Re:MK>> doesn't work at all. Re:	Have any one resolved this problem? Re:
Re:MK>> Re:	My compiling option are followed: Re:	sdcc  -mmcs51
Re:MK>> --stack-auto --use-stdout --debug --opt-code-speed
Re:MK>> -I/usr/local/share/sdcc/include -Iincludes  -D_DEBUG -DSDCC
Re:MK>> --model-large --code-size 65536 --iram-size 256 --xram-size 65536
Re:MK>> -I/usr/local/share/sdcc/include Re:	 Re: Re:	Winbond W78E516BP-40
Re:MK>> features (from official specification): Re: Re:Fully static design
Re:MK>> 8-bit CMOS microcontroller Re:64K bytes of in-system programmable
Re:MK>> Flash EPROM for Application Program (APROM) Re:4K bytes of
Re:MK>> auxiliary Flash EPROM for Loader Program (LDROM) Re:512 bytes of
Re:MK>> on-chip RAM. (including 256 bytes of AUX-RAM, software selectable)
Re:MK>> Re:64K bytes program memory address space and 64K bytes data
Re:MK>> memory address space Re:Four 8-bit bi-directional ports Re:One
Re:MK>> 4-bit multipurpose programmable port Re:Three 16-bit
Re:MK>> timer/counters Re:One full duplex serial port Re:Eight-sources,
Re:MK>> two-level interrupt capability Re:Built-in power management
Re:MK>> Re:Code protection Re:Packaged in Re:  DIP 40: W78LE516-24 Re:
Re:MK>> PLCC 44: W78LE516P-24 Re:  QFP 44: W78LE516F-24 Re:  LQFP 48:
Re:MK>> W78LE516D-24 Re: Re: Re:	Thanks. Re:Hyacinth Re: Re:
Re:MK>> Re:------------------------------------------------------- Re:This
Re:MK>> SF.net email is sponsored by: Splunk Inc. Do you grep through log
Re:MK>> files Re:for problems?  Stop!  Download the new AJAX search engine
Re:MK>> that makes Re:searching your log files as easy as surfing the
Re:MK>> web.  DOWNLOAD SPLUNK!
Re:MK>> Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:MK>> Re:_______________________________________________ Re:Sdcc-user
Re:MK>> mailing list Re:Sdcc-user@...
Re:MK>> Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user Re:
MK> Re:
MK> Re:
Re:MK>> ------------------------------------------------------- This
Re:MK>> SF.net email is sponsored by: Splunk Inc. Do you grep through log
Re:MK>> files for problems?  Stop!  Download the new AJAX search engine
Re:MK>> that makes searching your log files as easy as surfing the  web.
Re:MK>> DOWNLOAD SPLUNK!
Re:MK>> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
Re:MK>> _______________________________________________ Sdcc-user mailing
Re:MK>> list Sdcc-user@...
Re:MK>> https://lists.sourceforge.net/lists/listinfo/sdcc-user
MK> Re:
MK> Re:
MK> Re:
MK> Re:
MK> Re:-------------------------------------------------------
MK> Re:This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
MK> Re:for problems?  Stop!  Download the new AJAX search engine that makes
MK> Re:searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
MK> Re:http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
MK> Re:_______________________________________________
MK> Re:Sdcc-user mailing list
MK> Re:Sdcc-user@...
MK> Re:https://lists.sourceforge.net/lists/listinfo/sdcc-user
MK> Re:


MK> -------------------------------------------------------
MK> This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
MK> for problems?  Stop!  Download the new AJAX search engine that makes
MK> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
MK> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
MK> _______________________________________________
MK> Sdcc-user mailing list
MK> Sdcc-user@...
MK> https://lists.sourceforge.net/lists/listinfo/sdcc-user