From sbissonnette at microsyl.com  Thu Mar  1 17:25:25 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Thu Mar  1 17:35:32 2007
Subject: [Icc-avr] Problem compiling bootloader
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net>
Message-ID: <002901c75c69$a7bbc250$0301a8c0@Sylvain>

Hi,

    I have a MegaLoad user who have problem compiling the bootloader and I 
don't understand
what is the problem.  If someone could help me I will be please.  Here is 
the ICC information

> C:\iccv7avr\bin\imakew -f M.mak
> iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g -Mavr_enhanced_small 
> C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\main.c
> iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g -Mavr_enhanced_small 
>  -Wa-g C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\assembly.s
> iccavr -o 
> M -g -e:0x1000 -ucrtboot8k.o -bvector:0xc00.0x1000 -bfunc_lit:0xc34.0x1000 
>  -dram_end:0x2ff -bdata:0x100.0x2ff -dhwstk_size:16 -beeprom:0.256 -fihx_coff 
>  -S2 @M.lk
> !E <library>(57): Address out of range 65535 is larger than max address 
> 4096
> C:\iccv7avr\bin\imakew.exe: Error code 1
> Done: there are error(s). Exit code: 1. Thu Mar 01 19:53:44 2007
>

Yours truly
Sylvain Bissonnette

From richard-lists at imagecraft.com  Thu Mar  1 18:00:15 2007
From: richard-lists at imagecraft.com (Richard)
Date: Thu Mar  1 18:08:10 2007
Subject: [Icc-avr] Problem compiling bootloader
In-Reply-To: <002901c75c69$a7bbc250$0301a8c0@Sylvain>
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net>
	<002901c75c69$a7bbc250$0301a8c0@Sylvain>
Message-ID: <6.1.0.6.2.20070301175944.07918e00@192.168.100.42>

It's a bug with the new -e: flag and some other interaction. I am working 
on it. Sorry.

At 05:25 PM 3/1/2007, Sylvain Bissonnette wrote:
>Hi,
>
>    I have a MegaLoad user who have problem compiling the bootloader and I 
> don't understand
>what is the problem.  If someone could help me I will be please.  Here is 
>the ICC information
>
>>C:\iccv7avr\bin\imakew -f M.mak
>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g 
>>-Mavr_enhanced_small C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\main.c
>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g 
>>-Mavr_enhanced_small  -Wa-g C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\assembly.s
>>iccavr -o M -g -e:0x1000 -ucrtboot8k.o -bvector:0xc00.0x1000 
>>-bfunc_lit:0xc34.0x1000  -dram_end:0x2ff -bdata:0x100.0x2ff 
>>-dhwstk_size:16 -beeprom:0.256 -fihx_coff  -S2 @M.lk
>>!E <library>(57): Address out of range 65535 is larger than max address 4096
>>C:\iccv7avr\bin\imakew.exe: Error code 1
>>Done: there are error(s). Exit code: 1. Thu Mar 01 19:53:44 2007

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From sbissonnette at microsyl.com  Thu Mar  1 18:31:41 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Thu Mar  1 18:41:48 2007
Subject: [Icc-avr] Problem compiling bootloader
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net> <002901c75c69$a7bbc250$0301a8c0@Sylvain>
	<6.1.0.6.2.20070301175944.07918e00@192.168.100.42>
Message-ID: <001101c75c72$e9a91d30$0301a8c0@Sylvain>

Thanks Richard,

    I stay tune for news on this problem,

Sylvain
----- Original Message ----- 
From: "Richard" <richard-lists@imagecraft.com>
To: <icc-avr@imagecraft.com>
Sent: Thursday, March 01, 2007 9:00 PM
Subject: Re: [Icc-avr] Problem compiling bootloader


> It's a bug with the new -e: flag and some other interaction. I am working 
> on it. Sorry.
>
> At 05:25 PM 3/1/2007, Sylvain Bissonnette wrote:
>>Hi,
>>
>>    I have a MegaLoad user who have problem compiling the bootloader and I 
>> don't understand
>>what is the problem.  If someone could help me I will be please.  Here is 
>>the ICC information
>>
>>>C:\iccv7avr\bin\imakew -f M.mak
>>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g -Mavr_enhanced_small 
>>>C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\main.c
>>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g -Mavr_enhanced_small 
>>>  -Wa-g C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\assembly.s
>>>iccavr -o 
>>>M -g -e:0x1000 -ucrtboot8k.o -bvector:0xc00.0x1000 -bfunc_lit:0xc34.0x1000 
>>>  -dram_end:0x2ff -bdata:0x100.0x2ff -dhwstk_size:16 -beeprom:0.256 -fihx_coff 
>>>  -S2 @M.lk
>>>!E <library>(57): Address out of range 65535 is larger than max address 
>>>4096
>>>C:\iccv7avr\bin\imakew.exe: Error code 1
>>>Done: there are error(s). Exit code: 1. Thu Mar 01 19:53:44 2007
>
> // richard (This email is for mailing lists. To reach me directly, please 
> use richard at imagecraft.com)
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
> 

From richard-lists at imagecraft.com  Thu Mar  1 18:53:06 2007
From: richard-lists at imagecraft.com (Richard)
Date: Thu Mar  1 19:01:04 2007
Subject: 7.12A compiler was: Re: [Icc-avr] Problem compiling bootloader
In-Reply-To: <002901c75c69$a7bbc250$0301a8c0@Sylvain>
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net>
	<002901c75c69$a7bbc250$0301a8c0@Sylvain>
Message-ID: <6.1.0.6.2.20070301184719.07917c00@192.168.100.42>

Due to a bug where the check was done, this error will be emitted for any 
bootloader build for devices with less than 64K of memory. Sorry.

This fixes it: http://www.dragonsgate.net/pub/Betas/iccomavr.exe

Put it in your bin directory. I will formally release 7.12A soon.

At 05:25 PM 3/1/2007, Sylvain Bissonnette wrote:
>Hi,
>
>    I have a MegaLoad user who have problem compiling the bootloader and I 
> don't understand
>what is the problem.  If someone could help me I will be please.  Here is 
>the ICC information
>
>>C:\iccv7avr\bin\imakew -f M.mak
>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g 
>>-Mavr_enhanced_small C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\main.c
>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g 
>>-Mavr_enhanced_small  -Wa-g C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\assembly.s
>>iccavr -o M -g -e:0x1000 -ucrtboot8k.o -bvector:0xc00.0x1000 
>>-bfunc_lit:0xc34.0x1000  -dram_end:0x2ff -bdata:0x100.0x2ff 
>>-dhwstk_size:16 -beeprom:0.256 -fihx_coff  -S2 @M.lk
>>!E <library>(57): Address out of range 65535 is larger than max address 4096
>>C:\iccv7avr\bin\imakew.exe: Error code 1
>>Done: there are error(s). Exit code: 1. Thu Mar 01 19:53:44 2007

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From sbissonnette at microsyl.com  Thu Mar  1 19:05:35 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Thu Mar  1 19:15:42 2007
Subject: 7.12A compiler was: Re: [Icc-avr] Problem compiling bootloader
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net> <002901c75c69$a7bbc250$0301a8c0@Sylvain>
	<6.1.0.6.2.20070301184719.07917c00@192.168.100.42>
Message-ID: <000d01c75c77$a59e5830$0301a8c0@Sylvain>

Hi Richard

    Your path is not good http://www.dragonsgate.net/pub/Betas/iccomavr.exe

Sylvain

----- Original Message ----- 
From: "Richard" <richard-lists@imagecraft.com>
To: "Discussion list for ICCAVR and ICCtiny Users. You do NOT need 
tosubscribe to icc-announce if you are a member of this." 
<icc-avr@imagecraft.com>; "Discussion list for ICCAVR and ICCtiny Users. You 
do NOT needtosubscribeto icc-announce if you are a member of this." 
<icc-avr@imagecraft.com>
Sent: Thursday, March 01, 2007 9:53 PM
Subject: 7.12A compiler was: Re: [Icc-avr] Problem compiling bootloader


> Due to a bug where the check was done, this error will be emitted for any 
> bootloader build for devices with less than 64K of memory. Sorry.
>
> This fixes it: http://www.dragonsgate.net/pub/Betas/iccomavr.exe
>
> Put it in your bin directory. I will formally release 7.12A soon.
>
> At 05:25 PM 3/1/2007, Sylvain Bissonnette wrote:
>>Hi,
>>
>>    I have a MegaLoad user who have problem compiling the bootloader and I 
>> don't understand
>>what is the problem.  If someone could help me I will be please.  Here is 
>>the ICC information
>>
>>>C:\iccv7avr\bin\imakew -f M.mak
>>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g -Mavr_enhanced_small 
>>>C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\main.c
>>>iccavr -c -e -D__ICC_VERSION="7.12" -D_EE_EXTIO -DATMega48 -l -g -Mavr_enhanced_small 
>>>  -Wa-g C:\DOCUME~1\MACIEJ~1\Pulpit\MEGALO~1\assembly.s
>>>iccavr -o 
>>>M -g -e:0x1000 -ucrtboot8k.o -bvector:0xc00.0x1000 -bfunc_lit:0xc34.0x1000 
>>>  -dram_end:0x2ff -bdata:0x100.0x2ff -dhwstk_size:16 -beeprom:0.256 -fihx_coff 
>>>  -S2 @M.lk
>>>!E <library>(57): Address out of range 65535 is larger than max address 
>>>4096
>>>C:\iccv7avr\bin\imakew.exe: Error code 1
>>>Done: there are error(s). Exit code: 1. Thu Mar 01 19:53:44 2007
>
> // richard (This email is for mailing lists. To reach me directly, please 
> use richard at imagecraft.com)
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
> 

From richard-lists at imagecraft.com  Thu Mar  1 20:16:44 2007
From: richard-lists at imagecraft.com (Richard)
Date: Thu Mar  1 20:24:41 2007
Subject: 7.12A compiler was: Re: [Icc-avr] Problem compiling
  bootloader
In-Reply-To: <000d01c75c77$a59e5830$0301a8c0@Sylvain>
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net>
	<002901c75c69$a7bbc250$0301a8c0@Sylvain>
	<6.1.0.6.2.20070301184719.07917c00@192.168.100.42>
	<000d01c75c77$a59e5830$0301a8c0@Sylvain>
Message-ID: <6.1.0.6.2.20070301201614.07891d18@192.168.100.42>

http://www.dragonsgate.net/pub/BETAS/iccomavr.exe

At 07:05 PM 3/1/2007, you wrote:
>Hi Richard
>
>    Your path is not good http://www.dragonsgate.net/pub/Betas/iccomavr.exe
>
>Sylvain
>

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From richard-lists at imagecraft.com  Fri Mar  2 01:15:33 2007
From: richard-lists at imagecraft.com (Richard)
Date: Fri Mar  2 01:23:30 2007
Subject: 7.12A compiler was: Re: [Icc-avr] Problem compiling
  bootloader
In-Reply-To: <6.1.0.6.2.20070301201614.07891d18@192.168.100.42>
References: <07Feb26.154535cet.38958@gateway.pijnenburg.nl>
	<45E34881.8000801@virgin.net>
	<002901c75c69$a7bbc250$0301a8c0@Sylvain>
	<6.1.0.6.2.20070301184719.07917c00@192.168.100.42>
	<000d01c75c77$a59e5830$0301a8c0@Sylvain>
	<6.1.0.6.2.20070301201614.07891d18@192.168.100.42>
Message-ID: <6.1.0.6.2.20070302011521.07974268@192.168.100.42>

I dropped my brain in the middle of the Pacific on route home :-) I meant
http://www.dragonsgate.net/pub/BETAS/ilinkavr.exe

Hmm... in case you deleted your V7.12 iccomavr.exe, it's here:
http://www.dragonsgate.net/pub/BETAS/iccomavr.exe

At 08:16 PM 3/1/2007, Richard wrote:
>http://www.dragonsgate.net/pub/BETAS/iccomavr.exe
>
>At 07:05 PM 3/1/2007, you wrote:
>>Hi Richard
>>
>>    Your path is not good http://www.dragonsgate.net/pub/Betas/iccomavr.exe
>>
>>Sylvain
>
>// richard (This email is for mailing lists. To reach me directly, please 
>use richard at imagecraft.com)
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From richard at imagecraft.com  Fri Mar  2 21:18:45 2007
From: richard at imagecraft.com (Richard)
Date: Fri Mar  2 21:26:38 2007
Subject: [Icc-avr] (no subject)
Message-ID: <6.1.0.6.2.20070302183943.076c7438@192.168.100.42>

2007 is setting out to be an interesting year at ImageCraft. We will 
continue to enhance our existing products of course:

- new 8 bit optimizations for AVR and M8C (PRO version)
- support for new devices in the AVR, MSP430, ARM, and S12X families. 
Including extended addressing for 430X.

In addition, we have decided on the following new projects:

- Microchip PIC24 port. The PIC24 is a new 16 bit family from Microchip. It 
has many advantages that comes with being a Microchip device without some 
of the earlier architectural limitations. As PIC16 users upgrading to 
larger parts, the PIC24 is a natural step up.

- Propeller C. Propeller is a unique device from http://www.parallax.com. 8 
32-bit CPUs on one chip. Think "no interrupt." Propeller C gives us an 
opportunity to

- ARM7 Educational Kits. We are producing a complete education kit bundle 
with board, compiler, debugger, and course materials for some large 
universities. I expect variants of this kit will be made to address 
different unverisities' needs.

- More flexible dongle scheme. We sold over 500 dongles last year. I have 
ideas on making them more flexible.

Anyway, my brain is still somewhere out in the Pacific Ocean. That's for 
now. Thank you for your support.

// richard <http://www.imagecraft.com> 
<http://www.dragonsgate.net/mailman/listinfo> 

From sl at ecpower.dk  Fri Mar  2 22:55:16 2007
From: sl at ecpower.dk (Steven Lose)
Date: Fri Mar  2 23:08:13 2007
Subject: SV: [Icc-avr] (no subject)
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1655C8@seattle.ecpower.dk>

HI Richard.

Sounds god, can't wait to see what the 8 bit optimizations can do. ;o)

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk

www.ecpower.dk

-----Oprindelig meddelelse-----
Fra: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] P? vegne af Richard
Sendt: 3. marts 2007 06:19
Til: icc-announce@imagecraft.com; icc-avr@imagecraft.com; icc-mot@imagecraft.com; icc-430@imagecraft.com; icc-arm@imagecraft.com
Emne: [Icc-avr] (no subject)

2007 is setting out to be an interesting year at ImageCraft. We will 
continue to enhance our existing products of course:

- new 8 bit optimizations for AVR and M8C (PRO version)
- support for new devices in the AVR, MSP430, ARM, and S12X families. 
Including extended addressing for 430X.

In addition, we have decided on the following new projects:

- Microchip PIC24 port. The PIC24 is a new 16 bit family from Microchip. It 
has many advantages that comes with being a Microchip device without some 
of the earlier architectural limitations. As PIC16 users upgrading to 
larger parts, the PIC24 is a natural step up.

- Propeller C. Propeller is a unique device from http://www.parallax.com. 8 
32-bit CPUs on one chip. Think "no interrupt." Propeller C gives us an 
opportunity to

- ARM7 Educational Kits. We are producing a complete education kit bundle 
with board, compiler, debugger, and course materials for some large 
universities. I expect variants of this kit will be made to address 
different unverisities' needs.

- More flexible dongle scheme. We sold over 500 dongles last year. I have 
ideas on making them more flexible.

Anyway, my brain is still somewhere out in the Pacific Ocean. That's for 
now. Thank you for your support.

// richard <http://www.imagecraft.com> 
<http://www.dragonsgate.net/mailman/listinfo> 

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr


From scinthia_245 at rediffmail.com  Sat Mar  3 03:38:07 2007
From: scinthia_245 at rediffmail.com (cyn  thia)
Date: Sat Mar  3 03:49:28 2007
Subject: [Icc-avr] (no subject)
Message-ID: <20070303113807.12328.qmail@webmail103.rediffmail.com>

 ?
pls unsubscribe my account
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070303/e421ff17/attachment.html
From sbissonnette at microsyl.com  Mon Mar  5 13:54:30 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Mon Mar  5 14:02:39 2007
Subject: [Icc-avr] Is it possible to do this?
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
Message-ID: <000701c75f70$da89ac60$0301a8c0@Sylvain>

Hi,

    I'm not a #define expert, I want to know if it's possible to do this?

MINDELAY     (x - 0.148285449) / 0.016682113
x can be from 1 to 20

The result will always between 0 and 2000 (int)

Thanks
Sylvain Bissonnette

From richard at imagecraft.com  Mon Mar  5 18:12:16 2007
From: richard at imagecraft.com (Richard)
Date: Mon Mar  5 18:20:19 2007
Subject: [Icc-avr] V7.12A Beta0 
Message-ID: <6.1.0.6.2.20070305180724.05298990@192.168.100.42>

V7.12A
   IDE
     - The starting address for the 644P was incorrect (the only current "P"
       device with more interrupt vectors than the corresponding non-P devices.
     - For ISP with STK500, add a box for "No Erase." Use only if you
       pre-erase the chip.
   Linker
     - bootloader build for < 64K devices were broken WRT -e:<size> switch
       added in 7.10B
   Library
     - added the missing cstrpbrk and a new cstrtok

http://www.imagecraft.com/pub/iccv7avr_v712a_beta0.exe


// richard <http://www.imagecraft.com/>
<http://www.dragonsgate.net/mailman/listinfo> On-line orders, support, and 
listservers available on web site.
[ For technical support on ImageCraft products, please include all previous 
replies in your msgs. ] 

From r.rohde at indas.de  Tue Mar  6 04:09:54 2007
From: r.rohde at indas.de (Roman Rohde)
Date: Tue Mar  6 04:18:40 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <000701c75f70$da89ac60$0301a8c0@Sylvain>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
Message-ID: <45ED5A12.7060605@indas.de>

it should be possible, but try to avoid "/" and set the extra double 
brackets in the formula.

#define MINDELAY(x)    ( ((x)) - 0.148285449) * 59.94444469

Roman

Sylvain Bissonnette schrieb:
> Hi,
>
>    I'm not a #define expert, I want to know if it's possible to do this?
>
> MINDELAY     (x - 0.148285449) / 0.016682113
> x can be from 1 to 20
>
> The result will always between 0 and 2000 (int)
>
> Thanks
> Sylvain Bissonnette
>
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
>

From j_baraclough at zetnet.co.uk  Tue Mar  6 07:38:52 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Tue Mar  6 07:47:07 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <45ED5A12.7060605@indas.de>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de>
Message-ID: <E1HObkg-0004K2-7f@mailout.zetnet.co.uk>

What Roman suggests will certainly work but why do you want to do it 
that way? Five minutes work with an Excel spreadsheet gave me the 
following results:

const float MinDelay[20] =
{
     51.0555557920,  111.0000004796,  170.9444451671,  230.8888898547,
    290.8333345422,  350.7777792298,  410.7222239173,  470.6666686049,
    530.6111132924,  590.5555579800,  650.5000026675,  710.4444473551,
    770.3888920426,  830.3333367302,  890.2777814177,  950.2222261053,
   1010.1666707928, 1070.1111154804, 1130.0555601680, 1190.0000048555
};

OK that's only to ten decimal places, but with 8-bits it should be good enough.

All the best for now,
John


At 12:09 06/03/2007, you wrote:
>it should be possible, but try to avoid "/" and set the extra double 
>brackets in the formula.
>
>#define MINDELAY(x)    ( ((x)) - 0.148285449) * 59.94444469
>
>Roman
>
>Sylvain Bissonnette schrieb:
>>Hi,
>>
>>    I'm not a #define expert, I want to know if it's possible to do this?
>>
>>MINDELAY     (x - 0.148285449) / 0.016682113
>>x can be from 1 to 20
>>
>>The result will always between 0 and 2000 (int)
>>
>>Thanks
>>Sylvain Bissonnette
>>
>>_______________________________________________
>>Icc-avr mailing list
>>Icc-avr@imagecraft.com
>>http://dragonsgate.net/mailman/listinfo/icc-avr
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070306/3c4a2b7e/attachment.html
From sbissonnette at microsyl.com  Tue Mar  6 08:07:39 2007
From: sbissonnette at microsyl.com (sbissonnette@microsyl.com)
Date: Tue Mar  6 08:15:45 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <E1HObkg-0004K2-7f@mailout.zetnet.co.uk>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain> <45ED5A12.7060605@indas.de>
	<E1HObkg-0004K2-7f@mailout.zetnet.co.uk>
Message-ID: <34998.199.22.57.2.1173197259.squirrel@www.microsyl.com>

Hi,

   The reason is that my value will be from 1 to 20 but with decimal
it's the xtal speed of my MCU so it can be 3.579545

Thanks for all
Sylvain

> What Roman suggests will certainly work but why do you want to do it
> that way? Five minutes work with an Excel spreadsheet gave me the
> following results:
>
> const float MinDelay[20] =
> {
>      51.0555557920,  111.0000004796,  170.9444451671,  230.8888898547,
>     290.8333345422,  350.7777792298,  410.7222239173,  470.6666686049,
>     530.6111132924,  590.5555579800,  650.5000026675,  710.4444473551,
>     770.3888920426,  830.3333367302,  890.2777814177,  950.2222261053,
>    1010.1666707928, 1070.1111154804, 1130.0555601680, 1190.0000048555
> };
>
> OK that's only to ten decimal places, but with 8-bits it should be good
> enough.
>
> All the best for now,
> John
>
>
> At 12:09 06/03/2007, you wrote:
>>it should be possible, but try to avoid "/" and set the extra double
>>brackets in the formula.
>>
>>#define MINDELAY(x)    ( ((x)) - 0.148285449) * 59.94444469
>>
>>Roman
>>
>>Sylvain Bissonnette schrieb:
>>>Hi,
>>>
>>>    I'm not a #define expert, I want to know if it's possible to do
>>> this?
>>>
>>>MINDELAY     (x - 0.148285449) / 0.016682113
>>>x can be from 1 to 20
>>>
>>>The result will always between 0 and 2000 (int)
>>>
>>>Thanks
>>>Sylvain Bissonnette
>>>
>>>_______________________________________________
>>>Icc-avr mailing list
>>>Icc-avr@imagecraft.com
>>>http://dragonsgate.net/mailman/listinfo/icc-avr
>>
>>_______________________________________________
>>Icc-avr mailing list
>>Icc-avr@imagecraft.com
>>http://dragonsgate.net/mailman/listinfo/icc-avr
>>
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
>


From sbissonnette at microsyl.com  Tue Mar  6 08:10:33 2007
From: sbissonnette at microsyl.com (sbissonnette@microsyl.com)
Date: Tue Mar  6 08:18:39 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <45ED5A12.7060605@indas.de>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain> <45ED5A12.7060605@indas.de>
Message-ID: <46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>

Hi,

  could you told me why #define MINDELAY(x) "what () do?"
and is it the same for ( ((x)) - 0.148285449) * 59.94444469
                                     ^^^

Thanks
Sylvain


From j_baraclough at zetnet.co.uk  Tue Mar  6 09:59:44 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Tue Mar  6 10:07:59 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de>
	<46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>
Message-ID: <E1HOdx0-000579-Br@mailout.zetnet.co.uk>

Hi Sylvain,

Parentheses are required to keep the macro arguments together and 
there can be no space between the macro name and the opening 
parenthesis. As an example, consider the macro:

#define SQUARE(x)    x*x

When invoked as:

y = SQUARE(z + 1);

the resulting calculation after preprocessor substitution will be :

y = z*z + 1;

Because of operator precedence, that does not return the correct 
value, which should be:

y = (z + 1)*(z + 1);

Writing the macro as:

#define SQUARE(x)  (x*x)

will make the correct substitution.

The parentheses are even more important when the macro has multiple 
arguments, as in:

#define MAX(a, b)  ((a) > (b) ? (a) : (b))

which when  invoked as:

y = MAX(p+q, r+s);

will result in:

y = ((p+q) > (r+s) ? (p+q) : (r+s));


This is a good general reference for the C language:

  http://www.its.strath.ac.uk/courses/c/

and more specific detail can be found here:

  http://www.its.strath.ac.uk/courses/c/subsection3_13_2.html#SECTION00013200000000000000

Hope that helps.

All the best for now,
John


At 16:10 06/03/2007, you wrote:
>Hi,
>
>   could you told me why #define MINDELAY(x) "what () do?"
>and is it the same for ( ((x)) - 0.148285449) * 59.94444469
>                                      ^^^
>
>Thanks
>Sylvain
>
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070306/918aa1f3/attachment.html
From james.hatley at comcast.net  Tue Mar  6 08:40:28 2007
From: james.hatley at comcast.net (James Hatley)
Date: Tue Mar  6 12:33:15 2007
Subject: [Icc-avr] Is it possible to do this?
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com><ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com><E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk><000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de><E1HObkg-0004K2-7f@mailout.zetnet.co.uk>
	<34998.199.22.57.2.1173197259.squirrel@www.microsyl.com>
Message-ID: <001301c7600e$3222d260$6901a8c0@hsd1.wa.comcast.net>

I think you want and integer result so...

How about using 60*x-9 rather than the floating point?

Jim

----- Original Message ----- 
From: <sbissonnette@microsyl.com>
To: "Discussion list for ICCAVR and ICCtiny Users. You do NOT need
tosubscribe to icc-announce if you are a member of this."
<icc-avr@imagecraft.com>
Sent: Tuesday, March 06, 2007 8:07 AM
Subject: Re: [Icc-avr] Is it possible to do this?


> Hi,
>
>    The reason is that my value will be from 1 to 20 but with decimal
> it's the xtal speed of my MCU so it can be 3.579545
>
> Thanks for all
> Sylvain
>
> > What Roman suggests will certainly work but why do you want to do it
> > that way? Five minutes work with an Excel spreadsheet gave me the
> > following results:
> >
> > const float MinDelay[20] =
> > {
> >      51.0555557920,  111.0000004796,  170.9444451671,  230.8888898547,
> >     290.8333345422,  350.7777792298,  410.7222239173,  470.6666686049,
> >     530.6111132924,  590.5555579800,  650.5000026675,  710.4444473551,
> >     770.3888920426,  830.3333367302,  890.2777814177,  950.2222261053,
> >    1010.1666707928, 1070.1111154804, 1130.0555601680, 1190.0000048555
> > };
> >
> > OK that's only to ten decimal places, but with 8-bits it should be good
> > enough.
> >
> > All the best for now,
> > John
> >
> >
> > At 12:09 06/03/2007, you wrote:
> >>it should be possible, but try to avoid "/" and set the extra double
> >>brackets in the formula.
> >>
> >>#define MINDELAY(x)    ( ((x)) - 0.148285449) * 59.94444469
> >>
> >>Roman
> >>
> >>Sylvain Bissonnette schrieb:
> >>>Hi,
> >>>
> >>>    I'm not a #define expert, I want to know if it's possible to do
> >>> this?
> >>>
> >>>MINDELAY     (x - 0.148285449) / 0.016682113
> >>>x can be from 1 to 20
> >>>
> >>>The result will always between 0 and 2000 (int)
> >>>
> >>>Thanks
> >>>Sylvain Bissonnette
> >>>
> >>>_______________________________________________
> >>>Icc-avr mailing list
> >>>Icc-avr@imagecraft.com
> >>>http://dragonsgate.net/mailman/listinfo/icc-avr
> >>
> >>_______________________________________________
> >>Icc-avr mailing list
> >>Icc-avr@imagecraft.com
> >>http://dragonsgate.net/mailman/listinfo/icc-avr
> >>
> > _______________________________________________
> > Icc-avr mailing list
> > Icc-avr@imagecraft.com
> > http://dragonsgate.net/mailman/listinfo/icc-avr
> >
>
>
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
>


From sbissonnette at microsyl.com  Tue Mar  6 13:24:03 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Tue Mar  6 13:32:14 2007
Subject: [Icc-avr] Is it possible to do this?
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain> <45ED5A12.7060605@indas.de>
Message-ID: <000601c76035$c392cee0$0301a8c0@Sylvain>

Hi,

    I try it, but It's don't work, with calculator it's work I got the good
value, but not with ICC. Any idea?

> it should be possible, but try to avoid "/" and set the extra double 
> brackets in the formula.
> 
> #define MINDELAY(x)    ( ((x)) - 0.148285449) * 59.94444469

Thanks
Sylvain

From johan at edab.nu  Wed Mar  7 00:03:36 2007
From: johan at edab.nu (johan@edab.nu)
Date: Wed Mar  7 00:12:25 2007
Subject: [Icc-avr] Is it possible to do this?
Message-ID: <200703070803.l2783aCT001515@gigamail.giganet.se>



If you use that expression in the following way: 
y = (int)MINDELAY(x);
you're gonna get the wrong result - you need brackets around the whole statement otherwise the cast to (int) will cast the first part for the expression before multiplying with 59.94...

That's why the expression should look like this:
#define MINDELAY(x) ((x-0.148285449) * 59.94444469)

Likewise you should always write brackets around defines that consist of an expression, like:
#define ADD_ONE(x) (x+1) 

There should be no problem however (as someone claimed) to write a define and use spaces between the terms when using it, if the define is correctly written. This is an example that works in all cases:

#define SQUARE(x) ((x)*(x))

so it'll be possible to do
y = SQUARE(x + 1);

I see no use for the double brackets, what's up with those anyway?

regards 
Johan Wallstr?m
Elektronik Design AB


On 2007-03-06 Sylvain Bissonnette  wrote:

Hi,
>
>    I try it, but It's don't work, with calculator it's work I got the good
>value, but not with ICC. Any idea?
>
>> it should be possible, but try to avoid "/" and set the extra double 
>> brackets in the formula.
>> 
>> #define MINDELAY(x)    ( ((x)) - 0.148285449) * 59.94444469
>
>Thanks
>Sylvain
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>
>



From schaefer at mabel.info  Wed Mar  7 02:09:46 2007
From: schaefer at mabel.info (=?ISO-8859-15?Q?Thomas_Sch=E4fer?=)
Date: Wed Mar  7 02:18:08 2007
Subject: [Icc-avr] Problems w/ watchdog ATM2561
Message-ID: <45EE8F6A.2080907@mabel.info>

Hello,

I'm struggling with the watchdog in a ATM2561.
After setting up the entire system in Main(), I activate the watchdog with

/// \brief enables the watchdog timer.
/// \remarks Mode: System Reset
/// \remarks Timeout: Twdt = 8s
/// \warning After execution the WDR() macro has to be called repeatedly 
within Twdt
void watchdog_enable(void)
{
   CLI();
   WDR();
   WDTCSR |= (1<<WDCE) | (1<<WDE);
   WDTCSR = (1<<WDE) | (1<<WDP3) | (1<<WDP0);
   SEI();
}

right before I enter the main loop. As long as I issue WDR's inside the 
main loop, everything works fine. Now I tried to test the watchdog 
functionality by commenting out the WDR macro, awaiting a system reset 
8s after entering the main loop. But the system seems just to stop and 
remain in an undefined state.

Any hints?

regards,
Thomas

PS: I'm using V7.12 Professional.

From rohde at aragon-interactive.de  Wed Mar  7 02:44:20 2007
From: rohde at aragon-interactive.de (Roman Rohde)
Date: Wed Mar  7 02:52:29 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <E1HOdx0-000579-Br@mailout.zetnet.co.uk>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de>
	<46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>
	<E1HOdx0-000579-Br@mailout.zetnet.co.uk>
Message-ID: <op.toter61x55dwhn@pcgik.aragon-interactive.de>

Am Tue, 06 Mar 2007 18:59:44 +0100 schrieb John Baraclough  
<j_baraclough@zetnet.co.uk>:

> #define SQUARE(x)    x*x
> When invoked as:
> y = SQUARE(z + 1);
> the resulting calculation after preprocessor substitution will be :
> y = z*z + 1;
no it will be
y = z + 1*z + 1


From JonODonnell at moonlightconsulting.us  Wed Mar  7 04:02:13 2007
From: JonODonnell at moonlightconsulting.us (Jon O'Donnell)
Date: Wed Mar  7 04:08:24 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <op.toter61x55dwhn@pcgik.aragon-interactive.de>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com><ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com><E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk><000701c75f70$da89ac60$0301a8c0@Sylvain><45ED5A12.7060605@indas.de><46854.199.22.57.2.1173197433.squirrel@www.microsyl.com><E1HOdx0-000579-Br@mailout.zetnet.co.uk>
	<op.toter61x55dwhn@pcgik.aragon-interactive.de>
Message-ID: <003301c760b0$72659900$0201a8c0@ctc.parker.com>

The most dangerous part of this form of macros is

y = SQUARE(z++);

becomes

y = ((z++)*(z++));

This does not have the expected result no matter what combination of () are
used.  For this reason I avoid and calculation macros which make more than
one reference to an arg.

Jon

From rohde at aragon-interactive.de  Wed Mar  7 05:44:11 2007
From: rohde at aragon-interactive.de (Roman Rohde)
Date: Wed Mar  7 05:52:21 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <003301c760b0$72659900$0201a8c0@ctc.parker.com>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de>
	<46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>
	<E1HOdx0-000579-Br@mailout.zetnet.co.uk>
	<op.toter61x55dwhn@pcgik.aragon-interactive.de>
	<003301c760b0$72659900$0201a8c0@ctc.parker.com>
Message-ID: <op.totm3xmx55dwhn@pcgik.aragon-interactive.de>

Am Wed, 07 Mar 2007 13:02:13 +0100 schrieb Jon O'Donnell  
<JonODonnell@moonlightconsulting.us>:

> The most dangerous part of this form of macros is
> y = SQUARE(z++);
> becomes
> y = ((z++)*(z++));

Theres a simple workaroud: DON'T DO THAT!
never usr ++ or -- in a macro argument or function call.

just write
y = SQUARE(z);
z++;

From JonODonnell at moonlightconsulting.us  Wed Mar  7 08:37:47 2007
From: JonODonnell at moonlightconsulting.us (JonODonnell@moonlightconsulting.us)
Date: Wed Mar  7 08:46:06 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <op.totm3xmx55dwhn@pcgik.aragon-interactive.de>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de>
	<46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>
	<E1HOdx0-000579-Br@mailout.zetnet.co.uk>
	<op.toter61x55dwhn@pcgik.aragon-interactive.de>
	<003301c760b0$72659900$0201a8c0@ctc.parker.com>
	<op.totm3xmx55dwhn@pcgik.aragon-interactive.de>
Message-ID: <1173285467.45eeea5bec37a@webmail.moonlightconsulting.us>

Function calls as arguments can be OK.  My solution is
"Never write a macro which has side-effects which would differ from a
function."

Consequently, I end up with macros like this

#define Test(x) do { u8 v = (x); ....... } while (0)

This make the the utilization of Test the same as an inline function.  This does
not work for functions which return values.

Just personal preference.

Jon

Quoting Roman Rohde <rohde@aragon-interactive.de>:

> Am Wed, 07 Mar 2007 13:02:13 +0100 schrieb Jon O'Donnell  
> <JonODonnell@moonlightconsulting.us>:
> 
> > The most dangerous part of this form of macros is
> > y = SQUARE(z++);
> > becomes
> > y = ((z++)*(z++));
> 
> Theres a simple workaroud: DON'T DO THAT!
> never usr ++ or -- in a macro argument or function call.
> 
> just write
> y = SQUARE(z);
> z++;
> 
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
> 
> 



From j_baraclough at zetnet.co.uk  Wed Mar  7 13:25:32 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Wed Mar  7 13:33:49 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <op.toter61x55dwhn@pcgik.aragon-interactive.de>
References: <ef5d884e0702180447t6bd46a06nd977d283af93e89e@mail.gmail.com>
	<ef5d884e0702182344o1a6892e2g8b9d5ba6bd81606f@mail.gmail.com>
	<E1HJ6m0-0003GH-Kz@mailout.zetnet.co.uk>
	<000701c75f70$da89ac60$0301a8c0@Sylvain>
	<45ED5A12.7060605@indas.de>
	<46854.199.22.57.2.1173197433.squirrel@www.microsyl.com>
	<E1HOdx0-000579-Br@mailout.zetnet.co.uk>
	<op.toter61x55dwhn@pcgik.aragon-interactive.de>
Message-ID: <E1HP3di-0005du-MX@mailout.zetnet.co.uk>

Compile the following code and look at the result in Studio. I think 
you will be surprised.

#define SQUARE1(x) x * x
#define SQUARE2(x) (x) * (x)
#define SQUARE3(x) (x * x)
#define SQUARE4(x) ((x) * (x))

unsigned int y, z;

void main(void)
{
   y = 3;
   z = SQUARE1(y + 1);
   y = z;
   y = 3;
   z = SQUARE2(y + 1);
   y = z;
   y = 3;
   z = SQUARE3(y + 1);
   y = z;
   y = 3;
   z = SQUARE4(y + 1);
   y = z;

   while(1)
     ;
}


It's also interesting to look at the listing file.

John


At 10:44 07/03/2007, you wrote:
>Am Tue, 06 Mar 2007 18:59:44 +0100 schrieb John Baraclough
><j_baraclough@zetnet.co.uk>:
>
>>#define SQUARE(x)    x*x
>>When invoked as:
>>y = SQUARE(z + 1);
>>the resulting calculation after preprocessor substitution will be :
>>y = z*z + 1;
>no it will be
>y = z + 1*z + 1
>
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>


From sbissonnette at microsyl.com  Wed Mar  7 16:46:50 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Wed Mar  7 16:55:03 2007
Subject: [Icc-avr] Is it possible to do this?
References: <200703070803.l2783aCT001515@gigamail.giganet.se>
Message-ID: <003201c7611b$42ab91e0$0301a8c0@Sylvain>

Hi,

    Thanks for all the good informations. Now it's almost work with
#define MINDELAY(x) ((x-0.148285449) * 59.94444469)

But it's look like the computation don't take all the decimal because my 
result is a little bit
offset compare the the same calculation on my calculator.

Any input are again welcome
Sylvain Bissonnette 

From MDipperstein at CalAmp.com  Wed Mar  7 17:25:26 2007
From: MDipperstein at CalAmp.com (Michael Dipperstein)
Date: Wed Mar  7 17:33:35 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <E1HP3di-0005du-MX@mailout.zetnet.co.uk>
Message-ID: <DD8B2184B198D24E9F7F2C1D0AA849280109DA12@echinda.PD.CalAmpCorp.com>

When I'm not sure how a macro will be expanded, I look at the output of
the C Preprocessor.

ICC's preprocessor is icppw.exe.  In theory it shouldn't expand macros
any different than any other preprocessor, but that's just theory.

-Mike

-----Original Message-----
From: icc-avr-bounces@imagecraft.com
[mailto:icc-avr-bounces@imagecraft.com] On Behalf Of John Baraclough
Sent: Wednesday, March 07, 2007 1:26 PM
To: rohde@aragon-interactive.de; Discussion list for ICCAVR and ICCtiny
Users. You do NOT need tosubscribe to icc-announce if you are a member
of this.
Subject: Re: [Icc-avr] Is it possible to do this?

Compile the following code and look at the result in Studio. I think 
you will be surprised.

#define SQUARE1(x) x * x
#define SQUARE2(x) (x) * (x)
#define SQUARE3(x) (x * x)
#define SQUARE4(x) ((x) * (x))

unsigned int y, z;

void main(void)
{
   y = 3;
   z = SQUARE1(y + 1);
   y = z;
   y = 3;
   z = SQUARE2(y + 1);
   y = z;
   y = 3;
   z = SQUARE3(y + 1);
   y = z;
   y = 3;
   z = SQUARE4(y + 1);
   y = z;

   while(1)
     ;
}


It's also interesting to look at the listing file.

John


At 10:44 07/03/2007, you wrote:
>Am Tue, 06 Mar 2007 18:59:44 +0100 schrieb John Baraclough
><j_baraclough@zetnet.co.uk>:
>
>>#define SQUARE(x)    x*x
>>When invoked as:
>>y = SQUARE(z + 1);
>>the resulting calculation after preprocessor substitution will be :
>>y = z*z + 1;
>no it will be
>y = z + 1*z + 1
>
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>


_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr

From DRaymond at Bankspower.com  Wed Mar  7 17:32:09 2007
From: DRaymond at Bankspower.com (David Raymond)
Date: Wed Mar  7 17:40:24 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <003201c7611b$42ab91e0$0301a8c0@Sylvain>
Message-ID: <0E755D4F91B6D344B39791054F52A2A901220856@gbexchange.bankspower.local>

If you are expecting the AVR to calculate to this accuracy using floating point and Imagecraft, it is not going to happen. I never use floating point, but I have read that you only get 4 decimal points of accuracy using singles. If you are looking for an int result you might try scaling your equation to much higher numbers. Is your input always a float?
Example using float:
#define MINDELAY(x) (((6.74375 * (x))- 1.0) * 8.88889)
If you allow the preprocessor to do the calculations, you may gain accuracy. Richard will know if the preprocessor uses the full capability of the underlying computer to do the calculations. You can try this by defining another term using MINDELAY(x) cast as an int.
If your input can be a 32 bit value (unsigned long), you can scale this even higher.
Example using long:
#define MINDELAY(x) (unsigned int)((((1079L * (x))- 160) * 12800) / 9)

Dave Raymond 
Software Engineer 
Gale Banks Engineering 
Ph (626) 969-9600 x3301 
Fx (626) 334-2376 
Visit us on the Web: www.bankspower.com 
E-mail: draymond@bankspower.com 



-----Original Message-----
From: icc-avr-bounces@imagecraft.com
[mailto:icc-avr-bounces@imagecraft.com]On Behalf Of Sylvain Bissonnette
Sent: Wednesday, March 07, 2007 4:47 PM
To: Discussion list for ICCAVR and ICCtiny Users. You do NOT
needtosubscribetoicc-announce if you are a member of this.
Subject: Re: [Icc-avr] Is it possible to do this?


Hi,

    Thanks for all the good informations. Now it's almost work with
#define MINDELAY(x) ((x-0.148285449) * 59.94444469)

But it's look like the computation don't take all the decimal because my 
result is a little bit
offset compare the the same calculation on my calculator.

Any input are again welcome
Sylvain Bissonnette 

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr

From sbissonnette at microsyl.com  Wed Mar  7 18:56:04 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Wed Mar  7 19:04:18 2007
Subject: [Icc-avr] Is it possible to do this?
References: <0E755D4F91B6D344B39791054F52A2A901220856@gbexchange.bankspower.local>
Message-ID: <000801c7612d$5054b620$0301a8c0@Sylvain>

Hi,

    I'm not shure I understand, the way I think it's work the #define 
calculation is not made
by the ICC floating point?  The calculation is done by the PC before the 
compilation is done.
so imho it can be the AVR floating point who give me my error.

Thanks again
Sylvain Bissonnette

----- Original Message ----- 
From: "David Raymond" <DRaymond@Bankspower.com>
To: "Discussion list for ICCAVR and ICCtiny Users. You do NOT 
needtosubscribeto icc-announce if you are a member of this." 
<icc-avr@imagecraft.com>
Sent: Wednesday, March 07, 2007 8:32 PM
Subject: RE: [Icc-avr] Is it possible to do this?


> If you are expecting the AVR to calculate to this accuracy using floating 
> point and Imagecraft, it is not going to happen. I never use floating 
> point, but I have read that you only get 4 decimal points of accuracy 
> using singles. If you are looking for an int result you might try scaling 
> your equation to much higher numbers. Is your input always a float?
> Example using float:
> #define MINDELAY(x) (((6.74375 * (x))- 1.0) * 8.88889)
> If you allow the preprocessor to do the calculations, you may gain 
> accuracy. Richard will know if the preprocessor uses the full capability 
> of the underlying computer to do the calculations. You can try this by 
> defining another term using MINDELAY(x) cast as an int.
> If your input can be a 32 bit value (unsigned long), you can scale this 
> even higher.
> Example using long:
> #define MINDELAY(x) (unsigned int)((((1079L * (x))- 160) * 12800) / 9)
>
> Dave Raymond
> Software Engineer
> Gale Banks Engineering
> Ph (626) 969-9600 x3301
> Fx (626) 334-2376
> Visit us on the Web: www.bankspower.com
> E-mail: draymond@bankspower.com
>
>
>
> -----Original Message-----
> From: icc-avr-bounces@imagecraft.com
> [mailto:icc-avr-bounces@imagecraft.com]On Behalf Of Sylvain Bissonnette
> Sent: Wednesday, March 07, 2007 4:47 PM
> To: Discussion list for ICCAVR and ICCtiny Users. You do NOT
> needtosubscribetoicc-announce if you are a member of this.
> Subject: Re: [Icc-avr] Is it possible to do this?
>
>
> Hi,
>
>    Thanks for all the good informations. Now it's almost work with
> #define MINDELAY(x) ((x-0.148285449) * 59.94444469)
>
> But it's look like the computation don't take all the decimal because my
> result is a little bit
> offset compare the the same calculation on my calculator.
>
> Any input are again welcome
> Sylvain Bissonnette
>
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
>
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
> 

From richard-lists at imagecraft.com  Wed Mar  7 22:42:36 2007
From: richard-lists at imagecraft.com (Richard)
Date: Wed Mar  7 22:50:44 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <000801c7612d$5054b620$0301a8c0@Sylvain>
References: <0E755D4F91B6D344B39791054F52A2A901220856@gbexchange.bankspower.local>
	<000801c7612d$5054b620$0301a8c0@Sylvain>
Message-ID: <6.1.0.6.2.20070307223655.07842410@192.168.100.42>

Sorry I haven't followed all parts of the thread carefully, but let me 
summarize some salient points:

- the C Preprocessor is does textual substitution and nothing more (for the 
purpose of macro expansion). Hence the usual recommendation of using () 
inside the definition to minimize unintended results. This does not address 
the potential danger of calling a function like macro with expressions with 
side effect such as a++, as noted by some posters.

- C does not say that constant floating point expressions have to be done 
at compile time. In fact, BECAUSE the compile time floating point (FP) 
constant folding will probably give different results than runtime 
evaluation (accuracy of PC math vs. 8 bit micro math), the compiler usually 
does not fold FP constants unless it needs to.

- Single precision gives ~7+ decimal digits of meaning values.

- Any FP calculation done by the compiler uses the native compiler 
behaviors, which is most likely in the native 80 bits intermediate format.

At 06:56 PM 3/7/2007, Sylvain Bissonnette wrote:
>Hi,
>
>    I'm not shure I understand, the way I think it's work the #define 
> calculation is not made
>by the ICC floating point?  The calculation is done by the PC before the 
>compilation is done.
>so imho it can be the AVR floating point who give me my error.
>
>Thanks again
>Sylvain Bissonnette

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From j_baraclough at zetnet.co.uk  Thu Mar  8 02:03:11 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Thu Mar  8 02:11:27 2007
Subject: [Icc-avr] Is it possible to do this?
In-Reply-To: <000801c7612d$5054b620$0301a8c0@Sylvain>
References: <0E755D4F91B6D344B39791054F52A2A901220856@gbexchange.bankspower.local>
	<000801c7612d$5054b620$0301a8c0@Sylvain>
Message-ID: <E1HPFSt-0002C3-Ck@mailout.zetnet.co.uk>

Hi Sylvain,

When something isn't working correctly, it's always useful to look at 
the listing output to see what is happening.

Compiling this little snippet of code:

#define MINDELAY1(x) ((x-0.148285449) * 59.94444469)
#define MINDELAY2(x) (((6.74375 * (x))- 1.0) * 8.88889)

float f, g;

void main(void)
{
   f = MINDELAY1(5);
   g = f;
   g = 5.0;
   f = MINDELAY1(g);
   g = f;
   f = MINDELAY2(5);
   g = f;
   g = 5.0;
   f = MINDELAY2(g);
   g = f;

   while(1)
     ;
}

Gives this output:

.
.
.
(0021) void main(void)
(0022) {
(0023)   f = MINDELAY1(5);
_main:
       97 EE08      LDI  R16,0xE8
       98 E010      LDI  R17,0
       99 940E 0141 CALL elpm32
       9B 9310 0205 STS  f+1,R17
       9D 9300 0204 STS  f,R16
       9F 9330 0207 STS  f+3,R19
       A1 9320 0206 STS  f+2,R18
(0024)   g = f;
       A3 0118      MOVW R2,R16
       A4 0129      MOVW R4,R18
       A5 9230 0201 STS  g+1,R3
       A7 9220 0200 STS  g,R2
       A9 9250 0203 STS  g+3,R5
       AB 9240 0202 STS  g+2,R4
(0025)   g = 5.0;
       AD EE04      LDI  R16,0xE4
       AE E010      LDI  R17,0
       AF 940E 0141 CALL elpm32
       B1 9310 0201 STS  g+1,R17
       B3 9300 0200 STS  g,R16
       B5 9330 0203 STS  g+3,R19
       B7 9320 0202 STS  g+2,R18
(0026)   f = MINDELAY1(g);
       B9 EE00      LDI  R16,0xE0
       BA E010      LDI  R17,0
       BB 940E 0141 CALL elpm32
       BD 0118      MOVW R2,R16
       BE 0129      MOVW R4,R18
       BF 9080 0202 LDS  R8,g+2
       C1 9090 0203 LDS  R9,g+3
       C3 9060 0200 LDS  R6,g
       C5 9070 0201 LDS  R7,g+1
       C7 ED0C      LDI  R16,0xDC
       C8 E010      LDI  R17,0
       C9 940E 0141 CALL elpm32
       CB 933A      ST   R19,-Y
       CC 932A      ST   R18,-Y
       CD 931A      ST   R17,-Y
       CE 930A      ST   R16,-Y
       CF 0183      MOVW R16,R6
       D0 0194      MOVW R18,R8
       D1 940E 01C6 CALL fpsub2x
       D3 0181      MOVW R16,R2
       D4 0192      MOVW R18,R4
       D5 940E 02B6 CALL fpmule2
       D7 9310 0205 STS  f+1,R17
       D9 9300 0204 STS  f,R16
       DB 9330 0207 STS  f+3,R19
       DD 9320 0206 STS  f+2,R18
.
.
.

As you can see from that, in the first invocation of MINDELAY1(), the 
host computer is doing the hard work and the code is simply 
recovering a stored value from Flash, however in the second it needs 
to do the calculation as the value is not known at compile time. Thus 
the second result is not the same as the first. Having said that, 
when run through AvrStudio, the first invocation returns 290.83334 
and the second 290.83331, which is only a 1.03e-5 percentage error.

A quick look at your own listing file will reveal how the compiler is 
treating your case.

All the best for now,
John


At 02:56 08/03/2007, you wrote:
>Hi,
>
>    I'm not shure I understand, the way I think it's work the 
> #define calculation is not made
>by the ICC floating point?  The calculation is done by the PC before 
>the compilation is done.
>so imho it can be the AVR floating point who give me my error.
>
>Thanks again
>Sylvain Bissonnette
>
>----- Original Message ----- From: "David Raymond" <DRaymond@Bankspower.com>
>To: "Discussion list for ICCAVR and ICCtiny Users. You do NOT 
>needtosubscribeto icc-announce if you are a member of this." 
><icc-avr@imagecraft.com>
>Sent: Wednesday, March 07, 2007 8:32 PM
>Subject: RE: [Icc-avr] Is it possible to do this?
>
>
>>If you are expecting the AVR to calculate to this accuracy using 
>>floating point and Imagecraft, it is not going to happen. I never 
>>use floating point, but I have read that you only get 4 decimal 
>>points of accuracy using singles. If you are looking for an int 
>>result you might try scaling your equation to much higher numbers. 
>>Is your input always a float?
>>Example using float:
>>#define MINDELAY(x) (((6.74375 * (x))- 1.0) * 8.88889)
>>If you allow the preprocessor to do the calculations, you may gain 
>>accuracy. Richard will know if the preprocessor uses the full 
>>capability of the underlying computer to do the calculations. You 
>>can try this by defining another term using MINDELAY(x) cast as an int.
>>If your input can be a 32 bit value (unsigned long), you can scale 
>>this even higher.
>>Example using long:
>>#define MINDELAY(x) (unsigned int)((((1079L * (x))- 160) * 12800) / 9)
>>
>>Dave Raymond
>>Software Engineer
>>Gale Banks Engineering
>>Ph (626) 969-9600 x3301
>>Fx (626) 334-2376
>>Visit us on the Web: www.bankspower.com
>>E-mail: draymond@bankspower.com
>>
>>
>>
>>-----Original Message-----
>>From: icc-avr-bounces@imagecraft.com
>>[mailto:icc-avr-bounces@imagecraft.com]On Behalf Of Sylvain Bissonnette
>>Sent: Wednesday, March 07, 2007 4:47 PM
>>To: Discussion list for ICCAVR and ICCtiny Users. You do NOT
>>needtosubscribetoicc-announce if you are a member of this.
>>Subject: Re: [Icc-avr] Is it possible to do this?
>>
>>
>>Hi,
>>
>>    Thanks for all the good informations. Now it's almost work with
>>#define MINDELAY(x) ((x-0.148285449) * 59.94444469)
>>
>>But it's look like the computation don't take all the decimal because my
>>result is a little bit
>>offset compare the the same calculation on my calculator.
>>
>>Any input are again welcome
>>Sylvain Bissonnette
>>
>>_______________________________________________
>>Icc-avr mailing list
>>Icc-avr@imagecraft.com
>>http://dragonsgate.net/mailman/listinfo/icc-avr
>>
>>_______________________________________________
>>Icc-avr mailing list
>>Icc-avr@imagecraft.com
>>http://dragonsgate.net/mailman/listinfo/icc-avr
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070308/bef60d96/attachment-0001.html
From sbissonnette at microsyl.com  Thu Mar  8 17:30:09 2007
From: sbissonnette at microsyl.com (Sylvain Bissonnette)
Date: Thu Mar  8 17:38:23 2007
Subject: [Icc-avr] Is it possible to do this?
References: <0E755D4F91B6D344B39791054F52A2A901220856@gbexchange.bankspower.local>
	<000801c7612d$5054b620$0301a8c0@Sylvain>
	<6.1.0.6.2.20070307223655.07842410@192.168.100.42>
Message-ID: <001b01c761ea$79b476a0$0301a8c0@Sylvain>

Hi,

    I what to thanks all of you who contribuate on this topic,  now I'm fix 
on what #define does for real

Yours truly
Sylvain Bissonnette

----- Original Message ----- 
From: "Richard" <richard-lists@imagecraft.com>
To: "Discussion list for ICCAVR and ICCtiny Users. You do NOT need 
tosubscribe to icc-announce if you are a member of this." 
<icc-avr@imagecraft.com>; "Discussion list for ICCAVR and ICCtiny Users. You 
do NOT needtosubscribeto icc-announce if you are a member of this." 
<icc-avr@imagecraft.com>
Sent: Thursday, March 08, 2007 1:42 AM
Subject: Re: [Icc-avr] Is it possible to do this?


> Sorry I haven't followed all parts of the thread carefully, but let me 
> summarize some salient points:
>
> - the C Preprocessor is does textual substitution and nothing more (for 
> the purpose of macro expansion). Hence the usual recommendation of using 
> () inside the definition to minimize unintended results. This does not 
> address the potential danger of calling a function like macro with 
> expressions with side effect such as a++, as noted by some posters.
>
> - C does not say that constant floating point expressions have to be done 
> at compile time. In fact, BECAUSE the compile time floating point (FP) 
> constant folding will probably give different results than runtime 
> evaluation (accuracy of PC math vs. 8 bit micro math), the compiler 
> usually does not fold FP constants unless it needs to.
>
> - Single precision gives ~7+ decimal digits of meaning values.
>
> - Any FP calculation done by the compiler uses the native compiler 
> behaviors, which is most likely in the native 80 bits intermediate format.
>
> At 06:56 PM 3/7/2007, Sylvain Bissonnette wrote:
>>Hi,
>>
>>    I'm not shure I understand, the way I think it's work the #define 
>> calculation is not made
>>by the ICC floating point?  The calculation is done by the PC before the 
>>compilation is done.
>>so imho it can be the AVR floating point who give me my error.
>>
>>Thanks again
>>Sylvain Bissonnette
>
> // richard (This email is for mailing lists. To reach me directly, please 
> use richard at imagecraft.com)
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
> 

From Svenn at symetrics.no  Fri Mar  9 00:39:30 2007
From: Svenn at symetrics.no (=?iso-8859-1?Q?Svenn_Dahlstr=F8m?=)
Date: Fri Mar  9 00:47:43 2007
Subject: SV: [Icc-avr] Problems w/ watchdog ATM2561
Message-ID: <23CAB4A1878E854994AB398BC4CF5B0E034F6B@SERVER.symetrics.local>

Hi Thomas

I had the same problem.
Found out, if I disable the watchdog before setting it up and enable it, then it worked fine...

//Disable watchdog reset
 WDR();
 MCUSR &= ~(1<<WDRF);			//clear watchdog reset flag...
 WDTCSR = (1<<WDCE) | (1<<WDE);	//Write logical one to WDCE and WDE
 WDTCSR = 0;					//all zero,  Turn off WDT

Svenn :)

> 
> Hello,
> 
> I'm struggling with the watchdog in a ATM2561.
> After setting up the entire system in Main(), I activate the watchdog with
> 
> /// \brief enables the watchdog timer.
> /// \remarks Mode: System Reset
> /// \remarks Timeout: Twdt = 8s
> /// \warning After execution the WDR() macro has to be called repeatedly
> within Twdt
> void watchdog_enable(void)
> {
>    CLI();
>    WDR();
>    WDTCSR |= (1<<WDCE) | (1<<WDE);
>    WDTCSR = (1<<WDE) | (1<<WDP3) | (1<<WDP0);
>    SEI();
> }
> 
> right before I enter the main loop. As long as I issue WDR's inside the
> main loop, everything works fine. Now I tried to test the watchdog
> functionality by commenting out the WDR macro, awaiting a system reset
> 8s after entering the main loop. But the system seems just to stop and
> remain in an undefined state.
> 
> Any hints?
> 
> regards,
> Thomas
> 
> PS: I'm using V7.12 Professional.
> 
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr
> 
> 


From benra at imt.liu.se  Fri Mar  9 02:59:31 2007
From: benra at imt.liu.se (Bengt Ragnemalm)
Date: Fri Mar  9 03:40:15 2007
Subject: SV: [Icc-avr] Problems w/ watchdog ATM2561
In-Reply-To: <23CAB4A1878E854994AB398BC4CF5B0E034F6B@SERVER.symetrics.local>
References: <23CAB4A1878E854994AB398BC4CF5B0E034F6B@SERVER.symetrics.local>
Message-ID: <011a01c7623a$043c3bb0$b160ec82@Shagrat>

I think the problem is that if you do a lot of things before setting up the
watchdog, you have no idea of how long the watchdog prescaler and counter
have counted with a potential risk of getting a reset imediately.

What is the normal behaviour for interrupt flags? Will the flag always get
set but nothing happened if the interrupt are off or is the flag not set at
all? I think it is the first. And if you have an interrupt flag set and
after that, enable the interrupt, is the normal behaviour that you in that
case trigger the interrupt immediately?

Best regards,
Bengt

> -----Ursprungligt meddelande-----
> Fr?n: icc-avr-bounces@imagecraft.com [mailto:icc-avr-
> bounces@imagecraft.com] F?r Svenn Dahlstr?m
> Skickat: den 9 mars 2007 09:40
> Till: Discussion list for ICCAVR and ICCtiny Users. You do NOT
> needtosubscribeto icc-announce if you are a member of this.
> ?mne: SV: [Icc-avr] Problems w/ watchdog ATM2561
> 
> Hi Thomas
> 
> I had the same problem.
> Found out, if I disable the watchdog before setting it up and enable it,
> then it worked fine...
> 
> //Disable watchdog reset
>  WDR();
>  MCUSR &= ~(1<<WDRF);			//clear watchdog reset flag...
>  WDTCSR = (1<<WDCE) | (1<<WDE);	//Write logical one to WDCE and WDE
>  WDTCSR = 0;					//all
> zero,  Turn off WDT
> 
> Svenn :)
> 
> >
> > Hello,
> >
> > I'm struggling with the watchdog in a ATM2561.
> > After setting up the entire system in Main(), I activate the watchdog
> with
> >
> > /// \brief enables the watchdog timer.
> > /// \remarks Mode: System Reset
> > /// \remarks Timeout: Twdt = 8s
> > /// \warning After execution the WDR() macro has to be called repeatedly
> > within Twdt
> > void watchdog_enable(void)
> > {
> >    CLI();
> >    WDR();
> >    WDTCSR |= (1<<WDCE) | (1<<WDE);
> >    WDTCSR = (1<<WDE) | (1<<WDP3) | (1<<WDP0);
> >    SEI();
> > }
> >
> > right before I enter the main loop. As long as I issue WDR's inside the
> > main loop, everything works fine. Now I tried to test the watchdog
> > functionality by commenting out the WDR macro, awaiting a system reset
> > 8s after entering the main loop. But the system seems just to stop and
> > remain in an undefined state.
> >
> > Any hints?
> >
> > regards,
> > Thomas
> >
> > PS: I'm using V7.12 Professional.
> >
> > _______________________________________________
> > Icc-avr mailing list
> > Icc-avr@imagecraft.com
> > http://dragonsgate.net/mailman/listinfo/icc-avr
> >
> >
> 
> 
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr


From graham at altongate.co.uk  Fri Mar  9 06:53:14 2007
From: graham at altongate.co.uk (Graham Whyte)
Date: Fri Mar  9 07:22:10 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
Message-ID: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>


If I have a switch statement with multiple cases like

int a;
a=1;
Switch (a)
	{
	case:1
	case:2
		do_something();
		break;

	default:
		break;
	}

Question:
Does do_something() get executed if a=1 ?
I suspect not but it will if a=2
I want it to execute if a=1 and if a=2

If I am correct, how do I get do_something() to execute for a=1 or a=2
without having separate case: statements ?

Can I write case:(1 || 2) ?

Sorry for the probably dumb question
I have been writing C code for many years but surprisingly have never come
across this

Graham

From Michael.Kremer at adc.com  Fri Mar  9 07:35:27 2007
From: Michael.Kremer at adc.com (Kremer, Michael)
Date: Fri Mar  9 07:43:43 2007
Subject: [Icc-avr] C question off topic
Message-ID: <5792D2F8C9A6134BBB5363709C83B23701E5E5A1@mn01exch02.adc.com>

Case statements always fall though to the next unless you 
explicitly place a break in the statement. So for a=1 
do_something() will be run. 

Note: the ':' needs to be on the other side of the number

switch(a)
{
   case 1:
   case 2:
      do_something();
      break;
   default:
      break;
}

-----Original Message-----
From: icc-avr-bounces@imagecraft.com
[mailto:icc-avr-bounces@imagecraft.com] On Behalf Of Graham Whyte
Sent: Friday, March 09, 2007 8:53 AM
To: benra@imt.liu.se; Discussion list for ICCAVR and ICCtiny Users. You
do NOT needtosubscribeto icc-announce if you are a member of this.
Subject: [Icc-avr] C question off topic


If I have a switch statement with multiple cases like

int a;
a=1;
Switch (a)
	{
	case:1
	case:2
		do_something();
		break;

	default:
		break;
	}

Question:
Does do_something() get executed if a=1 ?
I suspect not but it will if a=2
I want it to execute if a=1 and if a=2

If I am correct, how do I get do_something() to execute for a=1 or a=2
without having separate case: statements ?

Can I write case:(1 || 2) ?

Sorry for the probably dumb question
I have been writing C code for many years but surprisingly have never
come across this

Graham

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr

From sidprice at softtools.com  Fri Mar  9 07:42:31 2007
From: sidprice at softtools.com (Sid Price)
Date: Fri Mar  9 07:51:10 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
Message-ID: <007f01c76261$93912060$0200a8c0@SID004>

Yes the code will execute since the "case:1" does not have a "break"
statement the processor "falls through" to execute the code after "case:2".

Sid.

Sid Price Software Design
http://www.softtools.com
 
-----Original Message-----
From: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com]
On Behalf Of Graham Whyte
Sent: Friday, March 09, 2007 7:53 AM
To: benra@imt.liu.se; Discussion list for ICCAVR and ICCtiny Users. You do
NOT needtosubscribeto icc-announce if you are a member of this.
Subject: [Icc-avr] C question off topic


If I have a switch statement with multiple cases like

int a;
a=1;
Switch (a)
	{
	case:1
	case:2
		do_something();
		break;

	default:
		break;
	}

Question:
Does do_something() get executed if a=1 ?
I suspect not but it will if a=2
I want it to execute if a=1 and if a=2

If I am correct, how do I get do_something() to execute for a=1 or a=2
without having separate case: statements ?

Can I write case:(1 || 2) ?

Sorry for the probably dumb question
I have been writing C code for many years but surprisingly have never come
across this

Graham

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr

From obparham at jpl.nasa.gov  Fri Mar  9 07:49:51 2007
From: obparham at jpl.nasa.gov (Bruce Parham)
Date: Fri Mar  9 07:58:07 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
References: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
Message-ID: <45F1821F.5060503@jpl.nasa.gov>

Graham Whyte wrote:

> If I have a switch statement with multiple cases like
> 
> int a;
> a=1;
> Switch (a)
> 	{
> 	case:1
> 	case:2
> 		do_something();
> 		break;
> 
> 	default:
> 		break;
> 	}
> 
> Question:
> Does do_something() get executed if a=1 ?
> I suspect not but it will if a=2
> I want it to execute if a=1 and if a=2
> 
> If I am correct, how do I get do_something() to execute for a=1 or a=2
> without having separate case: statements ?
> 
> Can I write case:(1 || 2) ?
> 
> Sorry for the probably dumb question
> I have been writing C code for many years but surprisingly have never come
> across this
> 
> Graham

Hi Graham,

The switch statement is a "computed goto" and the case phrase is a label so,
if a is 1 or 2, the do_something() gets done.

Bruce



From j_baraclough at zetnet.co.uk  Fri Mar  9 07:57:27 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Fri Mar  9 08:05:43 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
Message-ID: <E1HPhTG-0008MD-1c@mailout.zetnet.co.uk>

Hi Graham,

Yes you can. It's probably just a typo in your message, but you have 
written the 'case' statements incorrectly. It should be:

switch(a)
{
   case 1:
   case 2:
     do_something();
     break;

   default:
     do_anotherthing();
     break;
}

You should be aware that multiple entry case statements are now 
deprecated (isn't that a wonderful word). They should not generate an 
error, but may well be non-portable.

If you are in doubt about something, the first step is to look at the 
listing file and see what the compiler is making of your source.

All the best for now,
John


At 14:53 09/03/2007, you wrote:

>If I have a switch statement with multiple cases like
>
>int a;
>a=1;
>Switch (a)
>         {
>         case:1
>         case:2
>                 do_something();
>                 break;
>
>         default:
>                 break;
>         }
>
>Question:
>Does do_something() get executed if a=1 ?
>I suspect not but it will if a=2
>I want it to execute if a=1 and if a=2
>
>If I am correct, how do I get do_something() to execute for a=1 or a=2
>without having separate case: statements ?
>
>Can I write case:(1 || 2) ?
>
>Sorry for the probably dumb question
>I have been writing C code for many years but surprisingly have never come
>across this
>
>Graham
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070309/7a4aa2ee/attachment-0001.html
From stephan.daub at sap.com  Fri Mar  9 07:55:36 2007
From: stephan.daub at sap.com (Daub, Stephan)
Date: Fri Mar  9 08:06:35 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
Message-ID: <F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>

Hi Graham,

>Question:
>Does do_something() get executed if a=1 ?

Answer:
Yes; unless you break the case chain by using "break".

// sample 1
switch (a) {
	case 1: {
   			do_something_else();   //will beexecuted if a==1

			break;
		  }	
	case 2: {
   			do_something();   //will beexecuted if a==2
			break;
		  }	

	default:   // is useless in this case
		break; // is useless in this case
}

// sample 2
switch (a) {
	case 1:   // will "fall through" to the next case...
	case 2: {
   			do_something();   //will beexecuted if a==1 or
a==2
			break;
		  }	
}

// Sample 2 with <if> (which is more readable in my opinion)
// but it's creating more or less the same code
	
	if (a==1 || a==2) {
		do_something();   //will beexecuted if a==1 or a==2
	};

Best, Stephan



 

-----Original Message-----
From: icc-avr-bounces@imagecraft.com
[mailto:icc-avr-bounces@imagecraft.com] On Behalf Of Graham Whyte
Sent: Friday, March 09, 2007 3:53 PM
To: benra@imt.liu.se; Discussion list for ICCAVR and ICCtiny Users. You
do NOT needtosubscribeto icc-announce if you are a member of this.
Subject: [Icc-avr] C question off topic


If I have a switch statement with multiple cases like

int a;
a=1;
Switch (a)
	{
	case:1
	case:2
		do_something();
		break;

	default:
		break;
	}

Question:
Does do_something() get executed if a=1 ?
I suspect not but it will if a=2
I want it to execute if a=1 and if a=2

If I am correct, how do I get do_something() to execute for a=1 or a=2
without having separate case: statements ?

Can I write case:(1 || 2) ?

Sorry for the probably dumb question
I have been writing C code for many years but surprisingly have never
come
across this

Graham

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr

From obparham at jpl.nasa.gov  Fri Mar  9 08:14:59 2007
From: obparham at jpl.nasa.gov (Bruce Parham)
Date: Fri Mar  9 08:23:14 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
References: <MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
Message-ID: <45F18803.6020408@jpl.nasa.gov>

Graham Whyte wrote:

> If I have a switch statement with multiple cases like
> 
> int a;
> a=1;
> Switch (a)
> 	{
> 	case:1
> 	case:2
> 		do_something();
> 		break;
> 
> 	default:
> 		break;
> 	}
> 
> Question:
> Does do_something() get executed if a=1 ?
> I suspect not but it will if a=2
> I want it to execute if a=1 and if a=2
> 
> If I am correct, how do I get do_something() to execute for a=1 or a=2
> without having separate case: statements ?
> 
> Can I write case:(1 || 2) ?
> 
> Sorry for the probably dumb question
> I have been writing C code for many years but surprisingly have never come
> across this
> 
> Graham
> 

Also the case label should end with the colon:

     Switch (a)
     {
         case 1:
         case 2: do_something();
                 break;

         default: break;
     }

The value between the case word and the colon can be any int type value so you can use
chars or two char ints. The following example uses a switch to parse command line options:

     while(argc > 1)  // handle command line arg's
     {
         switch(toupper(argv[1][0]))
         {
             case 'B':   sscanf(argv[1]+1, "%d", &bits);
                         break;

             case 'M':   sscanf(argv[1]+1, "%d", &m);
                         break;

             case 'N':   sscanf(argv[1]+1, "%d", &n);
                         break;

             case 'O':   sscanf(argv[1]+1, "%s", name);
                         break;

             case 'S':   sscanf(argv[1]+1, "%d", &s);
                         break;

             case 'P':   sscanf(argv[1]+1, "%d", &gsize);
                         break;

             case 'G':   g = 0;
                         break;

             default:    fprintf(stderr, "Unknown Option: %s\n\n", argv[1]);
                         exit(1);
         }
         --argc; ++argv;
     }

HTH,

Bruce


From obparham at jpl.nasa.gov  Fri Mar  9 09:48:58 2007
From: obparham at jpl.nasa.gov (Bruce Parham)
Date: Fri Mar  9 09:57:10 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <E1HPhTG-0008MD-1c@mailout.zetnet.co.uk>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<E1HPhTG-0008MD-1c@mailout.zetnet.co.uk>
Message-ID: <45F19E0A.9090000@jpl.nasa.gov>

John Baraclough wrote:

.
.
.
> You should be aware that multiple entry case statements are now 
> deprecated (isn't that a wonderful word). They should not generate an 
> error, but may well be non-portable.
.
.
.
   http://dictionary.reference.com/browse/deprecate

Why? The Bible (K&R 2nd Ed.) pg 58 states:

   "Each case is labeled by one or more integer-valued constants or constant expressions. ..."

If you change the rules now, you'll be breaking a lot of existing code.

Bruce


From j_baraclough at zetnet.co.uk  Fri Mar  9 10:16:46 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Fri Mar  9 10:25:02 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap. corp>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>
Message-ID: <E1HPje5-0001bS-4O@mailout.zetnet.co.uk>

The other reason for choosing the 'if(){} else if(){} else{}' 
construct over 'switch(){}', is that there is a limited amount of 
code you can put between 'case:' statements. Sometimes a 'switch(){}' 
will generate an 'Out of range' (or something like that) error 
message, in which case an 'if()' is the only solution.

All the best for now,
John


At 15:55 09/03/2007, you wrote:
>Hi Graham,
>
> >Question:
> >Does do_something() get executed if a=1 ?
>
>Answer:
>Yes; unless you break the case chain by using "break".
>
>// sample 1
>switch (a) {
>         case 1: {
>                         do_something_else();   //will beexecuted if a==1
>
>                         break;
>                   }
>         case 2: {
>                         do_something();   //will beexecuted if a==2
>                         break;
>                   }
>
>         default:   // is useless in this case
>                 break; // is useless in this case
>}
>
>// sample 2
>switch (a) {
>         case 1:   // will "fall through" to the next case...
>         case 2: {
>                         do_something();   //will beexecuted if a==1 or
>a==2
>                         break;
>                   }
>}
>
>// Sample 2 with <if> (which is more readable in my opinion)
>// but it's creating more or less the same code
>
>         if (a==1 || a==2) {
>                 do_something();   //will beexecuted if a==1 or a==2
>         };
>
>Best, Stephan


From j_baraclough at zetnet.co.uk  Fri Mar  9 11:51:57 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Fri Mar  9 12:00:13 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <45F19E0A.9090000@jpl.nasa.gov>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<E1HPhTG-0008MD-1c@mailout.zetnet.co.uk>
	<45F19E0A.9090000@jpl.nasa.gov>
Message-ID: <E1HPl8C-0006dm-A6@mailout.zetnet.co.uk>

Hi Bruce,

I particularly like definition 4 there: 'To pray for deliverance 
from'. Maybe we can deprecate global warming, but that's another story.

It's page 55 in my copy (K&R 1st Ed. c1978), but it also says on p56:

"... Falling through from one case to another is not robust, being 
prone to disintegration when the program is modified."

I've been there and done that. However I have also made use of the 
following construct:

   case 1:
     do_a_case1_only_action();

   case 2:
     do_a_case1_and_2_action();
     break;

which is great when you write it, but a real pig when you come back 
to it some months later and think WTF is that all about.

New versions of existing compilers must not break existing code by 
rejecting multiple case entries, however newer compilers (such as C#) 
are very specific about 'switch(){}'. Each 'case x:' must be single 
entry and must have a 'break;' and there must be a 'default:' with a 
'break;' even if it's empty. To my mind, that's probably a good 
discipline to adopt. Even though it does lead to a certain amount of 
code bloat, it is totally portable.

All the best for now,
John


At 17:48 09/03/2007, you wrote:
>John Baraclough wrote:
>
>.
>.
>.
>>You should be aware that multiple entry case statements are now 
>>deprecated (isn't that a wonderful word). They should not generate 
>>an error, but may well be non-portable.
>.
>.
>.
>   http://dictionary.reference.com/browse/deprecate
>
>Why? The Bible (K&R 2nd Ed.) pg 58 states:
>
>   "Each case is labeled by one or more integer-valued constants or 
> constant expressions. ..."
>
>If you change the rules now, you'll be breaking a lot of existing code.
>
>Bruce
>
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070309/57895c2b/attachment.html
From richard-lists at imagecraft.com  Fri Mar  9 12:40:34 2007
From: richard-lists at imagecraft.com (Richard)
Date: Fri Mar  9 12:48:36 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <E1HPje5-0001bS-4O@mailout.zetnet.co.uk>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>
	<E1HPje5-0001bS-4O@mailout.zetnet.co.uk>
Message-ID: <6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>

No it should not have any limit between cases. If it does, then it's a 
compiler bug and should be fixed.

At 10:16 AM 3/9/2007, John Baraclough wrote:
>The other reason for choosing the 'if(){} else if(){} else{}' construct 
>over 'switch(){}', is that there is a limited amount of code you can put 
>between 'case:' statements. Sometimes a 'switch(){}' will generate an 'Out 
>of range' (or something like that) error message, in which case an 'if()' 
>is the only solution.

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From jassenbaum at htp-tel.de  Fri Mar  9 12:55:59 2007
From: jassenbaum at htp-tel.de (Johannes Assenbaum)
Date: Fri Mar  9 13:02:39 2007
Subject: [Icc-avr] C question off topic
References: <E1HPhTG-0008MD-1c@mailout.zetnet.co.uk>
	<45F19E0A.9090000@jpl.nasa.gov>
	<E1HPl8C-0006dm-A6@mailout.zetnet.co.uk>
Message-ID: <POdd7ap5obWdHgDjVezmm1ALMrgTExyPOHG73JCStZe@akmail>

Ahm, I'd just say, that's, why I don't like cases in C so much :-)

Johannes


> Hi Bruce,

> I particularly like definition 4 there: 'To pray for deliverance
> from'. Maybe we can deprecate global warming, but that's another story.

> It's page 55 in my copy (K&R 1st Ed. c1978), but it also says on p56:

> "... Falling through from one case to another is not robust, being
> prone to disintegration when the program is modified."

> I've been there and done that. However I have also made use of the
> following construct:

>    case 1:
>      do_a_case1_only_action();

>    case 2:
>      do_a_case1_and_2_action();
>      break;

> which is great when you write it, but a real pig when you come back
> to it some months later and think WTF is that all about.

> New versions of existing compilers must not break existing code by
> rejecting multiple case entries, however newer compilers (such as C#)
> are very specific about 'switch(){}'. Each 'case x:' must be single
> entry and must have a 'break;' and there must be a 'default:' with a
> 'break;' even if it's empty. To my mind, that's probably a good
> discipline to adopt. Even though it does lead to a certain amount of
> code bloat, it is totally portable.

> All the best for now,
> John


> At 17:48 09/03/2007, you wrote:
>>John Baraclough wrote:
>>
>>.
>>.
>>.
>>>You should be aware that multiple entry case statements are now
>>>deprecated (isn't that a wonderful word). They should not generate
>>>an error, but may well be non-portable.
>>.
>>.
>>.
>>   http://dictionary.reference.com/browse/deprecate
>>
>>Why? The Bible (K&R 2nd Ed.) pg 58 states:
>>
>>   "Each case is labeled by one or more integer-valued constants or
>> constant expressions. ..."
>>
>>If you change the rules now, you'll be breaking a lot of existing code.
>>
>>Bruce
>>
>>
>>_______________________________________________
>>Icc-avr mailing list
>>Icc-avr@imagecraft.com
>>http://dragonsgate.net/mailman/listinfo/icc-avr
>>


From j_baraclough at zetnet.co.uk  Fri Mar  9 15:44:14 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Fri Mar  9 15:52:31 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>
	<E1HPje5-0001bS-4O@mailout.zetnet.co.uk>
	<6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>
Message-ID: <E1HPol0-0007fW-5M@mailout.zetnet.co.uk>

Hi Richard,

I've never seen this in V7, but did have it occur on a few occasions 
with V6 (at least two years ago, possibly more). I overcame the 
problem then by moving in-line code into functions and reducing the 
span of the jump. I guess it has been fixed already.

All the best for now,
John


At 20:40 09/03/2007, you wrote:
>No it should not have any limit between cases. If it does, then it's 
>a compiler bug and should be fixed.
>
>At 10:16 AM 3/9/2007, John Baraclough wrote:
>>The other reason for choosing the 'if(){} else if(){} else{}' 
>>construct over 'switch(){}', is that there is a limited amount of 
>>code you can put between 'case:' statements. Sometimes a 
>>'switch(){}' will generate an 'Out of range' (or something like 
>>that) error message, in which case an 'if()' is the only solution.
>
>// richard (This email is for mailing lists. To reach me directly, 
>please use richard at imagecraft.com)
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
>


From richard-lists at imagecraft.com  Fri Mar  9 16:00:32 2007
From: richard-lists at imagecraft.com (Richard)
Date: Fri Mar  9 16:08:34 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <E1HPol0-0007fW-5M@mailout.zetnet.co.uk>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>
	<E1HPje5-0001bS-4O@mailout.zetnet.co.uk>
	<6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>
	<E1HPol0-0007fW-5M@mailout.zetnet.co.uk>
Message-ID: <6.1.0.6.2.20070309160013.0a365408@192.168.100.42>

I do recall fixing something similar..

At 03:44 PM 3/9/2007, you wrote:
>Hi Richard,
>
>I've never seen this in V7, but did have it occur on a few occasions with 
>V6 (at least two years ago, possibly more). I overcame the problem then by 
>moving in-line code into functions and reducing the span of the jump. I 
>guess it has been fixed already.
>
>All the best for now,
>John
>

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From jwbacon at comcast.net  Fri Mar  9 16:37:11 2007
From: jwbacon at comcast.net (Jim Bacon)
Date: Fri Mar  9 16:45:50 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>
	<E1HPje5-0001bS-4O@mailout.zetnet.co.uk>
	<6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>
Message-ID: <200703100045.l2A0jmSH078282@dragonsgate2.imagecraft.com>

I'd really like to agree with that, Richard, but it seems to me that 
'removing limits' like that could generate some "issues".
We're not dealing with 32 bit flash space, nor are we looking at 
'unlimited stack sizes' like we're used to on the PC.
My preference would be to just declare a compiler limit publicly,
and/or issue a caveat that this type of program structure can 
possibly be placed too near a 64K limit.

At 01:40 PM 3/9/2007, you wrote:
>No it should not have any limit between cases. If it does, then it's 
>a compiler bug and should be fixed.
>
>At 10:16 AM 3/9/2007, John Baraclough wrote:
>>The other reason for choosing the 'if(){} else if(){} else{}' 
>>construct over 'switch(){}', is that there is a limited amount of 
>>code you can put between 'case:' statements. Sometimes a 
>>'switch(){}' will generate an 'Out of range' (or something like 
>>that) error message, in which case an 'if()' is the only solution.
>
>// richard (This email is for mailing lists. To reach me directly, 
>please use richard at imagecraft.com)
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr


*         Give me ambiguity or give me something else.          *
*****************************************************************
Jim Bacon
jwbacon@comcast.net
303-666-9455 (H)
http://adams22.homeip.net
jbacon@boulder.mellesgriot.com (W)
720-494-4938 ext 329 (W) ... 720-494-9432 (FAX)


From pladow at gmail.com  Fri Mar  9 16:50:34 2007
From: pladow at gmail.com (Peter LaDow)
Date: Fri Mar  9 16:58:47 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <200703100045.l2A0jmSH078282@dragonsgate2.imagecraft.com>
References: <011a01c7623a$043c3bb0$b160ec82@Shagrat>
	<MCBBIDEBGICJKCBFKELHIEGPCNAA.graham@altongate.co.uk>
	<F21A48426DE8E6439E780AA5F69C82CA038EB6E0@dewdfe13.wdf.sap.corp>
	<E1HPje5-0001bS-4O@mailout.zetnet.co.uk>
	<6.1.0.6.2.20070309123929.0a3300a0@192.168.100.42>
	<200703100045.l2A0jmSH078282@dragonsgate2.imagecraft.com>
Message-ID: <9e264e780703091650x76ef27dfn79fcaf5754a06be6@mail.gmail.com>

On 3/9/07, Jim Bacon <jwbacon@comcast.net> wrote:
>
> I'd really like to agree with that, Richard, but it seems to me that
> 'removing limits' like that could generate some "issues".
> We're not dealing with 32 bit flash space, nor are we looking at
> 'unlimited stack sizes' like we're used to on the PC.
> My preference would be to just declare a compiler limit publicly,
> and/or issue a caveat that this type of program structure can
> possibly be placed too near a 64K limit.


I think Richard is correct.  Switch statements need not be implemented as
relative branches.  In fact, the GNU tools generate jump tables if the
number of switch targets is large or the the size of each case statement is
large.  And if the jump table can hold any absolute target address, there is
no limit to how large a case statement is (within the memory limits of
course).

And I suspect that is exactly how Richard does things if whatever heuristic
he uses determines that a relative branch should not be used.
-- 
--
"To love for the sake of being loved is human;  to love for the sake of
loving is Angelic."  -- Alphonse de Lamartine
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070309/a6daf407/attachment.html
From ivanv at realtimedesigns.com.au  Fri Mar  9 19:10:51 2007
From: ivanv at realtimedesigns.com.au (Ivan Vernot)
Date: Fri Mar  9 19:19:07 2007
Subject: [Icc-avr] switch/case (was C question off topic)
Message-ID: <00c801c762c1$b615f150$4200a8c0@harry>

Hello Richard,

Speaking  of the implementation of switch/case..

If I am not mistaken, ICCAVR V7  implements a switch/case block as a series 
of int (16bit) compares.
It there any compelling reason why the 'smallest possible' type size is not 
used in the compare?
This would not only increase speed but decrease code size as well?

I find that most of my switch/case constructs use a byte sizes 'state 
variable' and would benefit a lot from this type of optimisation.

Kind Regards
Ivan Vernot



On 10/03/2007 11:00:32 AM, Richard (richard-lists@imagecraft.com) wrote:
> I do recall fixing something similar..
>
> At 03:44 PM 3/9/2007, you wrote:
> >Hi Richard,
> >
> >I've never seen this in V7, but did have it occur on a few occasions with
> >V6 (at least two years ago, possibly more). I overcame the problem then 
> >by
> >moving in-line code into functions and reducing the span of the jump. I
> >guess it has been fixed already.
> >
> >All the best for now,
> >John
> >
>
> // richard (This email is for mailing lists. To reach me directly, please
> use richard at imagecraft.com)
>
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr


From j_baraclough at zetnet.co.uk  Sat Mar 10 02:16:29 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Sat Mar 10 02:24:41 2007
Subject: [Icc-avr] switch/case (was C question off topic)
In-Reply-To: <00c801c762c1$b615f150$4200a8c0@harry>
References: <00c801c762c1$b615f150$4200a8c0@harry>
Message-ID: <E1HPycs-00075V-9P@mailout.zetnet.co.uk>

Hi Ivan,

This one has been around here for a long time. It's probably down to 
about #197 on Richard's 'To Do' list >:-}. Unfortunately the C 
specification calls for 'case x:' constant values to be of size 'int' 
and it would require an diversion from the spec to fix it. Something like:

#pragma switch_size char
#pragma switch_size int

would be useful and save a lot of unnecessary compares with zero. It 
would definitely be an advantage for the devices with smaller Flash 
sizes. Maybe it could be an option in the code compression algorithm.

All the best for now,
John


At 03:10 10/03/2007, you wrote:
>Hello Richard,
>
>Speaking  of the implementation of switch/case..
>
>If I am not mistaken, ICCAVR V7  implements a switch/case block as a 
>series of int (16bit) compares.
>It there any compelling reason why the 'smallest possible' type size 
>is not used in the compare?
>This would not only increase speed but decrease code size as well?
>
>I find that most of my switch/case constructs use a byte sizes 
>'state variable' and would benefit a lot from this type of optimisation.
>
>Kind Regards
>Ivan Vernot


From richard-lists at imagecraft.com  Sat Mar 10 02:46:20 2007
From: richard-lists at imagecraft.com (Richard)
Date: Sat Mar 10 02:54:16 2007
Subject: [Icc-avr] switch/case (was C question off topic)
In-Reply-To: <E1HPycs-00075V-9P@mailout.zetnet.co.uk>
References: <00c801c762c1$b615f150$4200a8c0@harry>
	<E1HPycs-00075V-9P@mailout.zetnet.co.uk>
Message-ID: <6.1.0.6.2.20070310024345.0b31ef08@192.168.100.42>

No, not #197 :-) It's part of the 8 bit optimization coming real soon now 
(PRO only). I don't know if the first release will have the switch 
optimization in, but there should be some good performance improvements coming!

At 02:16 AM 3/10/2007, John Baraclough wrote:
>Hi Ivan,
>
>This one has been around here for a long time. It's probably down to about 
>#197 on Richard's 'To Do' list >:-}. Unfortunately the C specification 
>calls for 'case x:' constant values to be of size 'int' and it would 
>require an diversion from the spec to fix it. Something like:
>
>#pragma switch_size char
>#pragma switch_size int
>
>would be useful and save a lot of unnecessary compares with zero. It would 
>definitely be an advantage for the devices with smaller Flash sizes. Maybe 
>it could be an option in the code compression algorithm.
>
>All the best for now,
>John
>
>
>At 03:10 10/03/2007, you wrote:
>>Hello Richard,
>>
>>Speaking  of the implementation of switch/case..
>>
>>If I am not mistaken, ICCAVR V7  implements a switch/case block as a 
>>series of int (16bit) compares.
>>It there any compelling reason why the 'smallest possible' type size is 
>>not used in the compare?
>>This would not only increase speed but decrease code size as well?
>>
>>I find that most of my switch/case constructs use a byte sizes 'state 
>>variable' and would benefit a lot from this type of optimisation.
>>
>>Kind Regards
>>Ivan Vernot
>
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr

// richard (This email is for mailing lists. To reach me directly, please 
use richard at imagecraft.com) 

From schaefer at mabel.info  Sat Mar 10 09:04:53 2007
From: schaefer at mabel.info (=?ISO-8859-1?Q?Thomas_Sch=E4fer?=)
Date: Sat Mar 10 09:12:52 2007
Subject: SV: [Icc-avr] Problems w/ watchdog ATM2561
In-Reply-To: <23CAB4A1878E854994AB398BC4CF5B0E034F6B@SERVER.symetrics.local>
References: <23CAB4A1878E854994AB398BC4CF5B0E034F6B@SERVER.symetrics.local>
Message-ID: <45F2E535.4040004@mabel.info>

Hi Svenn,

That's it! Thanks a lot.

regards,
Thomas


> Hi Thomas
> 
> I had the same problem.
> Found out, if I disable the watchdog before setting it up and enable it, then it worked fine...
> 
> //Disable watchdog reset
>  WDR();
>  MCUSR &= ~(1<<WDRF);			//clear watchdog reset flag...
>  WDTCSR = (1<<WDCE) | (1<<WDE);	//Write logical one to WDCE and WDE
>  WDTCSR = 0;					//all zero,  Turn off WDT
> 
> Svenn :)
> 
>> Hello,
>>
>> I'm struggling with the watchdog in a ATM2561.
>> After setting up the entire system in Main(), I activate the watchdog with
>>
>> /// \brief enables the watchdog timer.
>> /// \remarks Mode: System Reset
>> /// \remarks Timeout: Twdt = 8s
>> /// \warning After execution the WDR() macro has to be called repeatedly
>> within Twdt
>> void watchdog_enable(void)
>> {
>>    CLI();
>>    WDR();
>>    WDTCSR |= (1<<WDCE) | (1<<WDE);
>>    WDTCSR = (1<<WDE) | (1<<WDP3) | (1<<WDP0);
>>    SEI();
>> }
>>
>> right before I enter the main loop. As long as I issue WDR's inside the
>> main loop, everything works fine. Now I tried to test the watchdog
>> functionality by commenting out the WDR macro, awaiting a system reset
>> 8s after entering the main loop. But the system seems just to stop and
>> remain in an undefined state.
>>
>> Any hints?
>>
>> regards,
>> Thomas
>>
>> PS: I'm using V7.12 Professional.
>>
>> _______________________________________________
>> Icc-avr mailing list
>> Icc-avr@imagecraft.com
>> http://dragonsgate.net/mailman/listinfo/icc-avr
>>
>>
> 
> 
> _______________________________________________
> Icc-avr mailing list
> Icc-avr@imagecraft.com
> http://dragonsgate.net/mailman/listinfo/icc-avr

From richard at imagecraft.com  Sat Mar 10 18:29:01 2007
From: richard at imagecraft.com (Richard)
Date: Sat Mar 10 18:36:57 2007
Subject: [Icc-avr] 7.12A released
Message-ID: <6.1.0.6.2.20070310182815.0b1ecba0@192.168.100.42>

V7.12A - March 11th, 2007
   IDE
     - The starting address for the 644P was incorrect (the only current "P"
       device with more interrupt vectors than the corresponding non-P devices.
     - For ISP with STK500, add a box for "No Erase." Use only if you
       pre-erase the chip.
   Linker
     - bootloader build for < 64K devices were broken WRT -e:<size> switch
       added in 7.10B
   Library
     - added the missing cstrpbrk and the new cstrtok
     - the crt* startup files now always initialize RAMPZ to the appropriate
       value based on the project type. This is so that if you are building
       an application that is invoked by a bootloader, the startup file for
       the main app will initialize RAMPZ to zero, ensuring correct program
       behaviors.

// richard <http://www.imagecraft.com/>
<http://www.dragonsgate.net/mailman/listinfo> On-line orders, support, and 
listservers available on web site.
[ For technical support on ImageCraft products, please include all previous 
replies in your msgs. ] 

From t.jaspers at cpseurope.com  Mon Mar 12 01:03:02 2007
From: t.jaspers at cpseurope.com (Jaspers, Ton)
Date: Mon Mar 12 01:11:17 2007
Subject: [Icc-avr] C question off topic
Message-ID: <7B0EB27CF1CC93439B5CFB7526E5D74C3BF5FE@mickey.PBNV.local>

From: John Baraclough
	
> I've been there and done that. However I have also made use of the
following construct:
> 	
>   case 1:
>     do_a_case1_only_action();
> 
>   case 2:
>     do_a_case1_and_2_action();
>     break;
> 	
> which is great when you write it, but a real pig when you come back to
it some months later and think WTF is that all about.
	

That is why I added to our in-house style guide that such cases have a
madatory comment:

  case 1:
    do_a_case1_only_action();

    /* Fall through is intentional */

  case 2:
    do_a_case1_and_2_action();
    break;

I would be the first to admit that this is not a perfect solution, but
it does help a little.


Cheers,
Ton Jaspers
CPS Europe BV
 


From sl at ecpower.dk  Mon Mar 12 05:14:20 2007
From: sl at ecpower.dk (Steven Lose)
Date: Mon Mar 12 05:27:44 2007
Subject: [Icc-avr] SPI master and slave
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1A4A7A@seattle.ecpower.dk>

Hi.

 

In a new design, I want a Master (atmega 128) to talk to a Slave (atmega8).

But do I run into problems during programming with the ISP?

For example if I want to program the Slave and the master is talking on the SPI port, I probably run into trouble.

Do I need resistors on the wires? How big?

I would expect Atmel to have some application notes on this, but did not find any.

 

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/4a4239f4/attachment.html
From bobgardner at aol.com  Mon Mar 12 05:32:13 2007
From: bobgardner at aol.com (bobgardner@aol.com)
Date: Mon Mar 12 05:40:28 2007
Subject: [Icc-avr] SPI master and slave
In-Reply-To: <072D96786BFC014AAEBA9EB07A8070EA1A4A7A@seattle.ecpower.dk>
References: <072D96786BFC014AAEBA9EB07A8070EA1A4A7A@seattle.ecpower.dk>
Message-ID: <8C932C1FAABC4B0-88C-961E@webmail-de08.sysops.aol.com>

Programming pins on a 128 are rx0 and tx0 instead of mosi and miso, so maybe no prob? 
 
-----Original Message-----
From: sl@ecpower.dk
Sent: Mon, 12 Mar 2007 9:14 AM
Subject: [Icc-avr] SPI master and slave


Hi.
 
In a new design, I want a Master (atmega 128) to talk to a Slave (atmega8).
But do I run into problems during programming with the ISP?
For example if I want to program the Slave and the master is talking on the SPI port, I probably run into trouble.
Do I need resistors on the wires? How big?
I would expect Atmel to have some application notes on this, but did not find any.
 
Med venlig hilsen / Best regards / mit freundlichen Gr??en
EC POWER A/S
Steven Lose
Software Ingeni?r
Tlf.: +45 87434100
Direkte tlf. +45 58286608
Email: sl@ecpower.dk
www.ecpower.dk
 
_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr
________________________________________________________________________
AOL now offers free email to everyone.  Find out more about what's free from AOL at AOL.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/eed22381/attachment.html
From Thorsten.Geffers at ATSonline.de  Mon Mar 12 05:45:59 2007
From: Thorsten.Geffers at ATSonline.de (Thorsten Geffers)
Date: Mon Mar 12 05:54:15 2007
Subject: AW: [Icc-avr] SPI master and slave
Message-ID: <A5442E33E67598488DDDFDC13B3B5FC53CEE7A@mail-10-4.ATS-Intranet.local>

Hi Steven,
 
maybe you should have a look at the Atmel application note AVR042: "AVR Hardware Design Considerations".
There you find a part "Connecting ISP Lines". Maybe there you can find some useful information.
 
Thorsten

________________________________

Von: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] Im Auftrag von Steven Lose
Gesendet: Montag, 12. M?rz 2007 14:14
An: Discussion list for ICCAVR and ICCtiny Users. You do NOT need tosubscribeto icc-announce if you are a member of this.
Betreff: [Icc-avr] SPI master and slave



Hi.

 

In a new design, I want a Master (atmega 128) to talk to a Slave (atmega8).

But do I run into problems during programming with the ISP?

For example if I want to program the Slave and the master is talking on the SPI port, I probably run into trouble.

Do I need resistors on the wires? How big?

I would expect Atmel to have some application notes on this, but did not find any.

 

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/a79b1ac3/attachment-0001.html
From tim at sabretechnology.co.uk  Mon Mar 12 06:11:48 2007
From: tim at sabretechnology.co.uk (Tim Mitchell)
Date: Mon Mar 12 06:20:01 2007
Subject: [Icc-avr] SPI master and slave
Message-ID: <04671BB8D269034BBC4BB6BA89486726061697@sserver.SabreTechnology.local>

bobgardner@aol.com wrote:
> 
> -----Original Message-----
> From: sl@ecpower.dk
> 
> Hi.
> 
> In a new design, I want a Master (atmega 128) to talk to a Slave
> (atmega8). 
> But do I run into problems during programming with the ISP?
> For example if I want to program the Slave and the master is talking
> on the SPI port, I probably run into trouble. 
> Do I need resistors on the wires? How big?
> I would expect Atmel to have some application notes on this, but did
> not find any. 
> 
> Med venlig hilsen / Best regards / mit freundlichen Gr??en EC POWER
> A/S Steven Lose Software Ingeni?r 


> Programming pins on a 128 are rx0 and tx0 instead of mosi and miso,
> so maybe no prob? 


Would not help if programming the mega8 slave, the mega128 would be trying to drive the lines.

When I had to do this, I tied the reset lines together on both chips, so any attempt to program one would hold them both in reset.


-- 
Tim Mitchell
tim@sabretechnology.co.uk http://www.sabretechnology.co.uk
Sabre Technology (Hull) Ltd, 3a Newlands Science Park, Hull HU6 7TQ
Registered in England and Wales no.3131504
t:01482 801003 f:01482 801078

From sl at ecpower.dk  Mon Mar 12 07:19:00 2007
From: sl at ecpower.dk (Steven Lose)
Date: Mon Mar 12 07:32:29 2007
Subject: SV: [Icc-avr] SPI master and slave
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1A4AA3@seattle.ecpower.dk>

Hi.

Thanks for the quick answers.

Thorsten: I just read the note, looks god but does not say anything on size of the resistors, but did tell me that resistors is to be used on all 3 lines.

I also thought on coupling the reset lines together, but as I understand it, the reset is only hold until programming mode is entered, so when reset is gone the one that is not in programming mode will boot and will make trouble on the lines, but maybe it's only a question on resistor size versus communication speed.


Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk

www.ecpower.dk


-----Oprindelig meddelelse-----
Fra: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] P? vegne af Tim Mitchell
Sendt: 12. marts 2007 15:12
Til: Discussion list for ICCAVR and ICCtiny Users. You do NOT needtosubscribeto icc-announce if you are a member of this.
Emne: RE: [Icc-avr] SPI master and slave

bobgardner@aol.com wrote:
> 
> -----Original Message-----
> From: sl@ecpower.dk
> 
> Hi.
> 
> In a new design, I want a Master (atmega 128) to talk to a Slave
> (atmega8). 
> But do I run into problems during programming with the ISP?
> For example if I want to program the Slave and the master is talking
> on the SPI port, I probably run into trouble. 
> Do I need resistors on the wires? How big?
> I would expect Atmel to have some application notes on this, but did
> not find any. 
> 
> Med venlig hilsen / Best regards / mit freundlichen Gr??en EC POWER
> A/S Steven Lose Software Ingeni?r 


> Programming pins on a 128 are rx0 and tx0 instead of mosi and miso,
> so maybe no prob? 


Would not help if programming the mega8 slave, the mega128 would be trying to drive the lines.

When I had to do this, I tied the reset lines together on both chips, so any attempt to program one would hold them both in reset.


-- 
Tim Mitchell
tim@sabretechnology.co.uk http://www.sabretechnology.co.uk
Sabre Technology (Hull) Ltd, 3a Newlands Science Park, Hull HU6 7TQ
Registered in England and Wales no.3131504
t:01482 801003 f:01482 801078

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr


From j_baraclough at zetnet.co.uk  Mon Mar 12 07:42:30 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Mon Mar 12 07:50:43 2007
Subject: [Icc-avr] SPI master and slave
In-Reply-To: <072D96786BFC014AAEBA9EB07A8070EA1A4A7A@seattle.ecpower.dk>
References: <072D96786BFC014AAEBA9EB07A8070EA1A4A7A@seattle.ecpower.dk>
Message-ID: <E1HQmfQ-0008V8-R5@mailout.zetnet.co.uk>

Connect the ISP header pins for the slave 
directly to the mega8 MISO, MOSI & SCK and 
isolate these from the other processor with 1k resistors.

You should do the same with the RXD0 pin on the 
mega128. TXD0 is OK without a resistor and SCK 
will be isolated by the resistor that goes to the mega8 pins.

A picture is worth 1k words so view this in a fixed pitch font such as Courier.

                                    ISP           ISP

                         ___        | | |         | | |
                 RXD0  -|___|-------)-)-+         | | |
                                    | |           | | |
                 TXD0  -------------)-+           | | |
                                    |             | | |
                                    |       ___   | | |
                 MOSI  -------------)------|___|--)-o-)----  MISO
                                    |             |   |
                                    |       ___   |   |
          M128   MISO  -------------)------|___|--)---o----  MOSI  M8
                                    |             |
                                    |       ___   |
                 SCK  -------------o------|___|--o--------  SCK


(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)


All the best for now,
John


At 13:14 12/03/2007, you wrote:
>Content-class: urn:content-classes:message
>Content-Type: multipart/alternative;
>         boundary="----_=_NextPart_001_01C764A8.57BB54C1"
>
>Hi.
>
>In a new design, I want a Master (atmega 128) to talk to a Slave (atmega8).
>But do I run into problems during programming with the ISP?
>For example if I want to program the Slave and 
>the master is talking on the SPI port, I probably run into trouble.
>Do I need resistors on the wires? How big?
>I would expect Atmel to have some application 
>notes on this, but did not find any.
>
>
>Med venlig hilsen / Best regards / mit freundlichen Gr??en
>
>EC POWER A/S
>
>Steven Lose
>
>Software Ingeni?r
>
>Tlf.: +45 87434100
>
>Direkte tlf. +45 58286608
>
>Email: <blocked::mailto:bsl@ecpower.dk>sl@ecpower.dk
>
>www.ecpower.dk
>
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/99289d3f/attachment.html
From sl at ecpower.dk  Mon Mar 12 08:06:12 2007
From: sl at ecpower.dk (Steven Lose)
Date: Mon Mar 12 08:23:20 2007
Subject: SV: [Icc-avr] SPI master and slave
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1A4AAB@seattle.ecpower.dk>

Thanks John.

 

What is the reason for RXD0 resistor? Haven't heard about this one before.

How did you end up with the 1K value?

 

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk

________________________________

Fra: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] P? vegne af John Baraclough
Sendt: 12. marts 2007 16:42
Til: Discussion list for ICCAVR and ICCtiny Users. You do NOT need tosubscribe to icc-announce if you are a member of this.
Emne: Re: [Icc-avr] SPI master and slave

 

Connect the ISP header pins for the slave directly to the mega8 MISO, MOSI & SCK and isolate these from the other processor with 1k resistors.

You should do the same with the RXD0 pin on the mega128. TXD0 is OK without a resistor and SCK will be isolated by the resistor that goes to the mega8 pins.

A picture is worth 1k words so view this in a fixed pitch font such as Courier.

                                   ISP           ISP

                        ___        | | |         | | |
                RXD0  -|___|-------)-)-+         | | |
                                   | |           | | |
                TXD0  -------------)-+           | | |
                                   |             | | |
                                   |       ___   | | |
                MOSI  -------------)------|___|--)-o-)----  MISO
                                   |             |   |
                                   |       ___   |   |
         M128   MISO  -------------)------|___|--)---o----  MOSI  M8
                                   |             |
                                   |       ___   |
                SCK  -------------o------|___|--o--------  SCK


(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)


All the best for now,
John


At 13:14 12/03/2007, you wrote:



Content-class: urn:content-classes:message
Content-Type: multipart/alternative;
         boundary="----_=_NextPart_001_01C764A8.57BB54C1"

Hi.
 
In a new design, I want a Master (atmega 128) to talk to a Slave (atmega8).
But do I run into problems during programming with the ISP?
For example if I want to program the Slave and the master is talking on the SPI port, I probably run into trouble.
Do I need resistors on the wires? How big?
I would expect Atmel to have some application notes on this, but did not find any.
 

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk <http://www.ecpower.dk/> 
 
_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/9112054f/attachment-0001.html
From j_baraclough at zetnet.co.uk  Mon Mar 12 08:32:21 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Mon Mar 12 08:40:32 2007
Subject: [Icc-avr] SPI master and slave
In-Reply-To: <E1HQmfQ-0008V8-R5@mailout.zetnet.co.uk>
References: <072D96786BFC014AAEBA9EB07A8070EA1A4A7A@seattle.ecpower.dk>
	<E1HQmfQ-0008V8-R5@mailout.zetnet.co.uk>
Message-ID: <E1HQnRe-0003sm-PP@mailout.zetnet.co.uk>

Correction to diagram below. I put the RXD0 
resistor in the wrong place. All resistors are 1k.

John


>A picture is worth 1k words so view this in a 
>fixed pitch font such as Courier.
>
>                                ISP         RS232    ISP
>
>                                | | |    ___  |      | | |
>                 RXD0  ---------)-)-+---|___|-+      | | |
>                                | |                  | | |
>                 TXD0  ---------)-+                  | | |
>                                |                    | | |
>                                |         ___        | | |
>                 MOSI  ---------)--------|___|-------)-o-)----  MISO
>                                |                    |   |
>                                |         ___        |   |
>          M128   MISO  ---------)--------|___|-------)---o----  MOSI  M8
>                                |                    |
>                                |         ___        |
>                 SCK  ----------o--------|___|-------o--------  SCK
>
>
>(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)
>
>
>All the best for now,
>John
>
>
>At 13:14 12/03/2007, you wrote:
>>Content-class: urn:content-classes:message
>>Content-Type: multipart/alternative;
>>         boundary="----_=_NextPart_001_01C764A8.57BB54C1"
>>
>>Hi.
>>
>>In a new design, I want a Master (atmega 128) to talk to a Slave (atmega8).
>>But do I run into problems during programming with the ISP?
>>For example if I want to program the Slave and 
>>the master is talking on the SPI port, I probably run into trouble.
>>Do I need resistors on the wires? How big?
>>I would expect Atmel to have some application 
>>notes on this, but did not find any.
>>
>>
>>Med venlig hilsen / Best regards / mit freundlichen Gr??en
>>
>>EC POWER A/S
>>
>>Steven Lose
>>
>>Software Ingeni?r
>>
>>Tlf.: +45 87434100
>>
>>Direkte tlf. +45 58286608
>>
>>Email: <blocked::mailto:bsl@ecpower.dk>sl@ecpower.dk
>>
>>www.ecpower.dk
>>
>>_______________________________________________
>>Icc-avr mailing list
>>Icc-avr@imagecraft.com
>>http://dragonsgate.net/mailman/listinfo/icc-avr
>_______________________________________________
>Icc-avr mailing list
>Icc-avr@imagecraft.com
>http://dragonsgate.net/mailman/listinfo/icc-avr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/bafd5995/attachment.html
From j_baraclough at zetnet.co.uk  Mon Mar 12 08:57:03 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Mon Mar 12 09:05:14 2007
Subject: SV: [Icc-avr] SPI master and slave
In-Reply-To: <072D96786BFC014AAEBA9EB07A8070EA1A4AAB@seattle.ecpower.dk>
References: <072D96786BFC014AAEBA9EB07A8070EA1A4AAB@seattle.ecpower.dk>
Message-ID: <E1HQnpZ-0005Sp-53@mailout.zetnet.co.uk>

Hi Steven,

Our posts have crossed so there's a later one 
with a corrected diagram. The resistor on RXD0 
ensures that no signal on the RS232 receiver will 
interfere with the ISP when it is operating. The 
value is sufficient to ensure isolation but not 
large enough to interfere with normal operation when the ISP isn't active.

It's a simple trick that I picked up from one of 
the very early Atmel data sheets. They don't put 
that particular diagram in the data sheets any more.

There's a similar trick to be aware of when using 
SPI memory on AVR's that use the SPI signals for 
programming. The SPI memory MISO signal can 
become active when the chip select is floating 
during programming and it can cause a programming 
failure. It needs either a 10k pull-up on the #CS 
line or a 1k isolating resistor in the MISO line 
between the memory chip and the processor. Again 
the ISP must be connected directly to the processor pins.

All the best for now,
John


At 16:06 12/03/2007, you wrote:
>Content-class: urn:content-classes:message
>Content-Type: multipart/alternative;
>         boundary="----_=_NextPart_001_01C764C0.DFF02675"
>
>Thanks John.
>
>What is the reason for RXD0 resistor? Haven?t heard about this one before.
>How did you end up with the 1K value?
>
>
>Med venlig hilsen / Best regards / mit freundlichen Gr??en
>
>EC POWER A/S
>
>Steven Lose
>
>Software Ingeni?r
>
>Tlf.: +45 87434100
>
>Direkte tlf. +45 58286608
>
>Email: <blocked::mailto:bsl@ecpower.dk>sl@ecpower.dk
>
>www.ecpower.dk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/46c0bc83/attachment.html
From sl at ecpower.dk  Mon Mar 12 10:55:42 2007
From: sl at ecpower.dk (Steven Lose)
Date: Mon Mar 12 11:09:26 2007
Subject: SV: SV: [Icc-avr] SPI master and slave
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1A4AAD@seattle.ecpower.dk>

Thanks a lot.

 

I guess this diagram also solves the ISP upgrade failure problem, when AVR STUDIO says that ISP need upgrade, - tries and fails on verify.

 

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk

________________________________

Fra: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] P? vegne af John Baraclough
Sendt: 12. marts 2007 17:57
Til: Discussion list for ICCAVR and ICCtiny Users. You do NOT need tosubscribe to icc-announce if you are a member of this.
Emne: Re: SV: [Icc-avr] SPI master and slave

 

Hi Steven,

Our posts have crossed so there's a later one with a corrected diagram. The resistor on RXD0 ensures that no signal on the RS232 receiver will interfere with the ISP when it is operating. The value is sufficient to ensure isolation but not large enough to interfere with normal operation when the ISP isn't active.

It's a simple trick that I picked up from one of the very early Atmel data sheets. They don't put that particular diagram in the data sheets any more.

There's a similar trick to be aware of when using SPI memory on AVR's that use the SPI signals for programming. The SPI memory MISO signal can become active when the chip select is floating during programming and it can cause a programming failure. It needs either a 10k pull-up on the #CS line or a 1k isolating resistor in the MISO line between the memory chip and the processor. Again the ISP must be connected directly to the processor pins.

All the best for now,
John  


At 16:06 12/03/2007, you wrote:



Content-class: urn:content-classes:message
Content-Type: multipart/alternative;
         boundary="----_=_NextPart_001_01C764C0.DFF02675"

Thanks John.
 
What is the reason for RXD0 resistor? Haven't heard about this one before.
How did you end up with the 1K value?
 

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk <http://www.ecpower.dk/>  

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/de9ae5a9/attachment.html
From albert.vanveen at pertronic.co.nz  Mon Mar 12 11:31:38 2007
From: albert.vanveen at pertronic.co.nz (Albert van Veen)
Date: Mon Mar 12 11:39:58 2007
Subject: [Icc-avr] C question off topic
In-Reply-To: <7B0EB27CF1CC93439B5CFB7526E5D74C3BF5FE@mickey.PBNV.local>
Message-ID: <EIELLGLNIEFIOEAOLPOIGEMOCFAA.albert.vanveen@pertronic.co.nz>

You can make this intention more obvious (what I would do) by writing it
like this:

  case 1:
  case 2:
    if (case 1)
	do_a_case1_only_action();
    do_a_case1_and_2_action();
    break;

Would give hardly (if any) more code.

Albert.


-----Original Message-----
From: Jaspers, Ton [mailto:t.jaspers@cpseurope.com]
Sent: Monday, 12 March 2007 10:03 p.m.
To: Discussion list for ICCAVR and ICCtiny Users. You do NOT need
tosubscribe to icc-announce if you are a member of this.
Subject: RE: [Icc-avr] C question off topic


From: John Baraclough

> I've been there and done that. However I have also made use of the
following construct:
>
>   case 1:
>     do_a_case1_only_action();
>
>   case 2:
>     do_a_case1_and_2_action();
>     break;
>
> which is great when you write it, but a real pig when you come back to
it some months later and think WTF is that all about.


That is why I added to our in-house style guide that such cases have a
madatory comment:

  case 1:
    do_a_case1_only_action();

    /* Fall through is intentional */

  case 2:
    do_a_case1_and_2_action();
    break;

I would be the first to admit that this is not a perfect solution, but
it does help a little.


Cheers,
Ton Jaspers
CPS Europe BV



_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr



From j_baraclough at zetnet.co.uk  Mon Mar 12 11:46:43 2007
From: j_baraclough at zetnet.co.uk (John Baraclough)
Date: Mon Mar 12 11:54:55 2007
Subject: SV: SV: [Icc-avr] SPI master and slave
In-Reply-To: <072D96786BFC014AAEBA9EB07A8070EA1A4AAD@seattle.ecpower.dk>
References: <072D96786BFC014AAEBA9EB07A8070EA1A4AAD@seattle.ecpower.dk>
Message-ID: <E1HQqTm-00069Z-DF@mailout.zetnet.co.uk>

Hi Steven,

I'm not sure about that. Atmel recommends that 
the signal pins are all unconnected for 
upgrading. I have a 10-pin ISP connector with 
only +5V and GND connected, that works for ISP upgrading every time.

Incidentally I recently upgraded Studio from 4.11 
to 4.12 SP4 and when I tried the ISP it told me I 
needed to upgrade from version 2.01 to version 
2.00!!! That tells me there was something wrong 
with version 2.01, but AFAIK nothing has been said.

All the best for now,
John


At 18:55 12/03/2007, you wrote:
>Content-class: urn:content-classes:message
>Content-Type: multipart/alternative;
>         boundary="----_=_NextPart_001_01C764D8.106A897F"
>
>Thanks a lot.
>
>I guess this diagram also solves the ISP upgrade 
>failure problem, when AVR STUDIO says that ISP 
>need upgrade, - tries and fails on verify.
>
>
>Med venlig hilsen / Best regards / mit freundlichen Gr??en
>
>EC POWER A/S
>
>Steven Lose
>
>Software Ingeni?r
>
>Tlf.: +45 87434100
>
>Direkte tlf. +45 58286608
>
>Email: <blocked::mailto:bsl@ecpower.dk>sl@ecpower.dk
>
>www.ecpower.dk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://dragonsgate.net/pipermail/icc-avr/attachments/20070312/3e14faec/attachment.html
From robertrade at yahoo.com  Mon Mar 12 14:44:11 2007
From: robertrade at yahoo.com (Robert Rademacher)
Date: Mon Mar 12 14:52:19 2007
Subject: [Icc-avr] Re: Icc-avr Digest, Vol 32, Issue 12
In-Reply-To: <200703121645.l2CGj1Ve017404@dragonsgate2.imagecraft.com>
Message-ID: <444700.87448.qm@web60411.mail.yahoo.com>

Hi,

Have you ever considered using I2C/TWI bus for Mega128master-Mega8slave
connection, if 400kb speed is sufficient for you?

I used it on my systems, which works very nicely.

Robert
===========================

Connect the ISP header pins for the slave 
directly to the mega8 MISO, MOSI & SCK and 
isolate these from the other processor with 1k resistors.

You should do the same with the RXD0 pin on the 
mega128. TXD0 is OK without a resistor and SCK 
will be isolated by the resistor that goes to the mega8 pins.

A picture is worth 1k words so view this in a fixed pitch font such as 
Courier.

                                    ISP           ISP

                         ___        | | |         | | |
                 RXD0  -|___|-------)-)-+         | | |
                                    | |           | | |
                 TXD0  -------------)-+           | | |
                                    |             | | |
                                    |       ___   | | |
                 MOSI  -------------)------|___|--)-o-)----  MISO
                                    |             |   |
                                    |       ___   |   |
          M128   MISO  -------------)------|___|--)---o----  MOSI  M8
                                    |             |
                                    |       ___   |
                 SCK  -------------o------|___|--o--------  SCK


(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)


All the best for now,
John


 
____________________________________________________________________________________
Don't pick lemons.
See all the new 2007 cars at Yahoo! Autos.
http://autos.yahoo.com/new_cars.html 

From sl at ecpower.dk  Mon Mar 12 15:08:27 2007
From: sl at ecpower.dk (Steven Lose)
Date: Mon Mar 12 15:21:54 2007
Subject: SV: [Icc-avr] Re: Icc-avr Digest, Vol 32, Issue 12
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1A4AAF@seattle.ecpower.dk>

Hi Robert.

Would be a good solution to, but in this case I already have these pins in use with other SPI devices and I'm out of pins.

Thanks


Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk

www.ecpower.dk

-----Oprindelig meddelelse-----
Fra: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] P? vegne af Robert Rademacher
Sendt: 12. marts 2007 23:44
Til: icc-avr@imagecraft.com
Emne: [Icc-avr] Re: Icc-avr Digest, Vol 32, Issue 12

Hi,

Have you ever considered using I2C/TWI bus for Mega128master-Mega8slave
connection, if 400kb speed is sufficient for you?

I used it on my systems, which works very nicely.

Robert
===========================

Connect the ISP header pins for the slave 
directly to the mega8 MISO, MOSI & SCK and 
isolate these from the other processor with 1k resistors.

You should do the same with the RXD0 pin on the 
mega128. TXD0 is OK without a resistor and SCK 
will be isolated by the resistor that goes to the mega8 pins.

A picture is worth 1k words so view this in a fixed pitch font such as 
Courier.

                                    ISP           ISP

                         ___        | | |         | | |
                 RXD0  -|___|-------)-)-+         | | |
                                    | |           | | |
                 TXD0  -------------)-+           | | |
                                    |             | | |
                                    |       ___   | | |
                 MOSI  -------------)------|___|--)-o-)----  MISO
                                    |             |   |
                                    |       ___   |   |
          M128   MISO  -------------)------|___|--)---o----  MOSI  M8
                                    |             |
                                    |       ___   |
                 SCK  -------------o------|___|--o--------  SCK


(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)


All the best for now,
John


 
____________________________________________________________________________________
Don't pick lemons.
See all the new 2007 cars at Yahoo! Autos.
http://autos.yahoo.com/new_cars.html 

_______________________________________________
Icc-avr mailing list
Icc-avr@imagecraft.com
http://dragonsgate.net/mailman/listinfo/icc-avr


From sl at ecpower.dk  Mon Mar 12 15:10:17 2007
From: sl at ecpower.dk (Steven Lose)
Date: Mon Mar 12 15:23:44 2007
Subject: SV: SV: SV: [Icc-avr] SPI master and slave
Message-ID: <072D96786BFC014AAEBA9EB07A8070EA1A4AB0@seattle.ecpower.dk>

Hi John.

 

Tried that to, from 2.07 to 2.xx something and it said downgrade, weird!

Med venlig hilsen / Best regards / mit freundlichen Gr??en

EC POWER A/S

Steven Lose

Software Ingeni?r

Tlf.: +45 87434100

Direkte tlf. +45 58286608

Email: sl@ecpower.dk <blocked::mailto:bsl@ecpower.dk> 

www.ecpower.dk

________________________________

Fra: icc-avr-bounces@imagecraft.com [mailto:icc-avr-bounces@imagecraft.com] P? vegne af John Baraclough
Sendt: 12. marts 2007 20:47
Til: Discussion list for ICCAVR and ICCtiny Users. You do NOT need tosubscribe to icc-announce if you are a member of this.
Emne: Re: SV: SV: [Icc-avr] SPI master and slave

 

Hi Steven,

I'm not sure about that. Atmel recommends that the signal pins are all unconnected for upgrading. I have a 10-pin ISP connector with only +5V and GND connected, that works for ISP upgrading every time.

Incidentally I recently upgraded Studio from 4.11 to 4.12 SP4 and when I tried the ISP it told me I needed to upgrade from version 2.01 to version 2.00!!! That tells me there was something wrong with version 2.01, but AFAIK nothing has been said.

All the best for now,
John


At 18:55 12/03/2007, you wrote:



Content-class: urn:content-classes:message
Content-Type: multipart/alternative;
         boundary="----_=_NextPart_001_01C764D8.106A897F"

Thanks a lot.
 
I guess this diagram also solves the ISP upgrade failure problem, when AVR STUDIO says that ISP need upgrade, -