owfs and owhttpd

Hi Rob,
Sorry, I was busy to unexpeted work during this week and worked very late
everyday.
You have to use the firmware I sent earlier.

I would create a main process that evaluate start conditions and initate sub
process that handle the action until duration.

some comments in your code below.

I will write you a sample skeleton later today.

Pascal



2010/3/27 Rob Conway <rjconway@...>

>  Hi,
>
> I was hoping to get some feedback on the AE engine and creating some simple
> output timers based upon energizing outputs based on time of day.  I can
> already set the RTC from the HOST and it resets at midnight, so the RTC
> represents seconds since midnight.
>
> This is what I am trying to achieve on the stack
>
>     IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
>     then DUTY1,0
>     else DUTY1,10001
>
> This is some rough code I thought I could use however not tested, I wanted
> to know if I was on the correct track and if the stack can handle maths
> results that go negative.
>
yes,
for Byte, values above 127 are considered negative
for Word, values above 32767 are considered negative
for Long, values above 2147483647 are considered negative


>
//  ON / OFF TIMES  (Seconds)
> //   L_UserN  ON time for Relay1   Seconds since Midnight
> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
>
what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
instead variable.

>
>
> relay1_ctl:
> CMP.L    L_RTC,L_USERL
> BLE        relay1_off
>
> PUSH.L    L_RTC          //Get time since midnight
> PUSH.L   L_USERN      //  get user ON time in seconds since midnight
> SUB                             // Find out how many seconds difference
> between RTC and on time
> CMP.W     @-1, W_USERL    // See if difference is less than Duration, if so
> the output should be ON
>
as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
require 4 bytes
important to retore stack pointer to inital value  after stack operation AIX
-4

> BGR        relay_off
> BLE         relay_on
> END
>
> relay1_on
> SET.W    W_DUTY1,0
> BRA Relay2_ctl
>
> Relay1_off
>  SET.W    W_DUTY1,10001
>  BRA Relay2_ctl
>
>
>  relay2_ctl:
> ~ code similar to relay 1
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
>


-- 
Pascal
 http://www.brain4home.eu

I thought it would be easier having the ON time in seconds since midnight so
I can compare direct to RTC.  
The length of time the output is ON needs to be able to vary between 1
minutes to 12 hrs.  I know I could hard code the ON / OFF time into the
stack however whilst I still have user variables I may as well use them. 

-----Original Message-----
From: Pascal Baerten [mailto:pascal.baerten@...] 
Sent: Saturday, March 27, 2010 6:59 PM
To: OWFS (One-wire file system) discussion and help
Subject: Re: [Owfs-developers] BAE0910 AE Timed output code


Hi Rob,
Sorry, I was busy to unexpeted work during this week and worked very late
everyday. 
You have to use the firmware I sent earlier.

I would create a main process that evaluate start conditions and initate sub
process that handle the action until duration.

some comments in your code below.

I will write you a sample skeleton later today.

Pascal




2010/3/27 Rob Conway <rjconway@...>


Hi,
 
I was hoping to get some feedback on the AE engine and creating some simple
output timers based upon energizing outputs based on time of day.  I can
already set the RTC from the HOST and it resets at midnight, so the RTC
represents seconds since midnight.
 
This is what I am trying to achieve on the stack 
 
    IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
    then DUTY1,0
    else DUTY1,10001
 
This is some rough code I thought I could use however not tested, I wanted
to know if I was on the correct track and if the stack can handle maths
results that go negative.

yes, 
for Byte, values above 127 are considered negative
for Word, values above 32767 are considered negative
for Long, values above 2147483647 are considered negative



 

//  ON / OFF TIMES  (Seconds)
//   L_UserN  ON time for Relay1   Seconds since Midnight  
//   W_UserL  Duration for PIO       Duration for relay ON time, seconds

what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
instead variable. 


 
 
relay1_ctl:
CMP.L    L_RTC,L_USERL
BLE        relay1_off
 
PUSH.L    L_RTC          //Get time since midnight
PUSH.L   L_USERN      //  get user ON time in seconds since midnight
SUB                             // Find out how many seconds difference
between RTC and on time   
CMP.W     @-1, W_USERL    // See if difference is less than Duration, if so
the output should be ON

as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
require 4 bytes 
important to retore stack pointer to inital value  after stack operation AIX
-4 


BGR        relay_off         
BLE         relay_on
END
 
relay1_on
SET.W    W_DUTY1,0
BRA Relay2_ctl
 
Relay1_off

SET.W    W_DUTY1,10001

BRA Relay2_ctl
 
 

relay2_ctl:
~ code similar to relay 1
 

----------------------------------------------------------------------------
--
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Owfs-developers mailing list
Owfs-developers@...
https://lists.sourceforge.net/lists/listinfo/owfs-developers






-- 
Pascal
 http://www.brain4home.eu


No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00

here is the sample: I found this approach easier to have better control from
the master.
Don't hesitate for any questions.

// AutomationEngine assembly source.
> // Used for Temperature and Light Control of Reef Tank
> // March 2010 by Rob Conway
> //
> // BAE0910 Direct Connected to Futurlec 4 channel opto Realy Board
> // Channel 0 & 1 Relay has been replaced with a Mosfet IRF540
> //
> // RELAY Ch#   BAE0910 Ch#   Ribbon Cable    USAGE
> // Output 0    PWM1          (Brown)         Cooling Fan
> // Output 1    PWM2         (Orange)        Heater
> // Output 2    PIO           (Yellow)        Main Light
> // Output 3    OUT           (Blue)          Refugium Light
> //
> //
> // VARIABLES:
> //  ON / OFF TIMES  (Seconds)
> //   L_UserN  ON time for Relay1   Seconds since Midnight
> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
> //   W_USERI  decrementing variale with time remaining until realy1 is
> turned OFF
>
>
> #include "bae0910.inc"
>
> #eeprom 0, 0      // start_page, end_page
> #org  $00      // Code generated at address specified.
>
> begin:
>     NOP     // Allow auto start of process #0 on poweron, first byte of
> page0 should be a NOP
> init:
>
>     //your initialisation code here
>     SET.B    B_TPM1C,4    // Pre Scaler (1us resolution)
>     SET.W    W_PERIOD1,10000    // 10,000us = 100Hz
>     SET.W    W_DUTY1,10001    // Set PWM1 Output OFF
>     START 1, relay1process   // this independant process will monitor USERI
> for action
> main:
>     PUSH.L    L_RTC
>     PUSH.L  L_USERN
>     SUB.L             //compare current time with sheduled time
>     AIS    -4             //restore stack
>     BNE notnow1         // if strictly equal, start cycle (may be true
> twice within the same second, but this is harmless)
>     SET.W  W_USERI, W_USERL // set duration to decounting variable USERI,
>                             // this will invite process1 to turn on for
> duration
>                             // this approach allows master check remaining
> time
>                             // and also start/interrupt actions.
> notnow1:
>     //similar code for relay2
>
>
>     WAIT    10    //not too long to ensure that each seconds are seen by
> the loop
>     JMP main
>
> relay1process: // asynchronous process for relay1 control, the input is
> USERI variable representing duration in seconds
>     SET.W    W_DUTY1,10001  // turn off relay1
>     WAIT    1
>     CMP.W    W_USERI,0
>     BEQ        relay1process
>     CLR.W     W_DUTY1 // turn on relay1
> loop_till_end1:
>     WAIT     16          //wait one second
>     DEC.W    W_USERI
>     BGR        loop_till_end1  // master may set directly to 0 to stop
> relay action, in this case USERI may become negative
>     CLR.W    W_USERI // set to zero in case of -1
>     BRA        relay1process
>




2010/3/27 Rob Conway <rjconway@...>

>  I thought it would be easier having the ON time in seconds since midnight
> so I can compare direct to RTC.
> The length of time the output is ON needs to be able to vary between 1
> minutes to 12 hrs.  I know I could hard code the ON / OFF time into the
> stack however whilst I still have user variables I may as well use them.
>
>  -----Original Message-----
> *From:* Pascal Baerten [mailto:pascal.baerten@...]
> *Sent:* Saturday, March 27, 2010 6:59 PM
> *To:* OWFS (One-wire file system) discussion and help
> *Subject:* Re: [Owfs-developers] BAE0910 AE Timed output code
>
> Hi Rob,
> Sorry, I was busy to unexpeted work during this week and worked very late
> everyday.
> You have to use the firmware I sent earlier.
>
> I would create a main process that evaluate start conditions and initate
> sub process that handle the action until duration.
>
> some comments in your code below.
>
> I will write you a sample skeleton later today.
>
> Pascal
>
>
>
> 2010/3/27 Rob Conway <rjconway@...>
>
>>  Hi,
>>
>> I was hoping to get some feedback on the AE engine and creating some
>> simple output timers based upon energizing outputs based on time of day.  I
>> can already set the RTC from the HOST and it resets at midnight, so the RTC
>> represents seconds since midnight.
>>
>> This is what I am trying to achieve on the stack
>>
>>     IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
>>     then DUTY1,0
>>     else DUTY1,10001
>>
>> This is some rough code I thought I could use however not tested, I wanted
>> to know if I was on the correct track and if the stack can handle maths
>> results that go negative.
>>
> yes,
> for Byte, values above 127 are considered negative
> for Word, values above 32767 are considered negative
> for Long, values above 2147483647 are considered negative
>
>
>>
>  //  ON / OFF TIMES  (Seconds)
>> //   L_UserN  ON time for Relay1   Seconds since Midnight
>> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
>>
> what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
> instead variable.
>
>>
>>
>> relay1_ctl:
>> CMP.L    L_RTC,L_USERL
>> BLE        relay1_off
>>
>> PUSH.L    L_RTC          //Get time since midnight
>> PUSH.L   L_USERN      //  get user ON time in seconds since midnight
>> SUB                             // Find out how many seconds difference
>> between RTC and on time
>> CMP.W     @-1, W_USERL    // See if difference is less than Duration, if
>> so the output should be ON
>>
> as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
> require 4 bytes
> important to retore stack pointer to inital value  after stack operation
> AIX -4
>
>>  BGR        relay_off
>> BLE         relay_on
>> END
>>
>> relay1_on
>> SET.W    W_DUTY1,0
>> BRA Relay2_ctl
>>
>> Relay1_off
>>  SET.W    W_DUTY1,10001
>>  BRA Relay2_ctl
>>
>>
>>  relay2_ctl:
>> ~ code similar to relay 1
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> Owfs-developers mailing list
>> Owfs-developers@...
>> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>>
>>
>
>
> --
> Pascal
> http://www.brain4home.eu
>
> No virus found in this incoming message.
> Checked by AVG - http://www.avg.com
> Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
> 06:33:00
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
>


-- 
Pascal
 http://www.brain4home.eu

Pascal,  Thanks I understand the concept of starting another process to
handle the relay and duration.
 
 
I do have a couple of questions:
 
1)  I think though that whilst RTC is greater than the ON time the "main:"
will always be writing the duration to UserI thus never alowing
loop_till_end1: to decrement the value to 0 again ?
 
2) I have tried to read the manual again with regards to stack pointer and
why AIS -4, however I am still lost on this one.  
 
3) I assume AIX -4 is required in my first example as the CMP.L uses the
stack for calculations and the result of the compare is at position -5 on
the stack ?
 
 
rob

-----Original Message-----
From: Pascal Baerten [mailto:pascal.baerten@...] 
Sent: Saturday, March 27, 2010 9:18 PM
To: OWFS (One-wire file system) discussion and help
Subject: Re: [Owfs-developers] BAE0910 AE Timed output code


here is the sample: I found this approach easier to have better control from
the master.
Don't hesitate for any questions.



// AutomationEngine assembly source.
// Used for Temperature and Light Control of Reef Tank 
// March 2010 by Rob Conway
// 
// BAE0910 Direct Connected to Futurlec 4 channel opto Realy Board
// Channel 0 & 1 Relay has been replaced with a Mosfet IRF540
//
// RELAY Ch#   BAE0910 Ch#   Ribbon Cable    USAGE
// Output 0    PWM1          (Brown)         Cooling Fan
// Output 1    PWM2         (Orange)        Heater
// Output 2    PIO           (Yellow)        Main Light
// Output 3    OUT           (Blue)          Refugium Light
//
//
// VARIABLES:
//  ON / OFF TIMES  (Seconds)
//   L_UserN  ON time for Relay1   Seconds since Midnight 
//   W_UserL  Duration for PIO       Duration for relay ON time, seconds
//   W_USERI  decrementing variale with time remaining until realy1 is
turned OFF


#include "bae0910.inc"

#eeprom 0, 0      // start_page, end_page
#org  $00      // Code generated at address specified.

begin:
    NOP     // Allow auto start of process #0 on poweron, first byte of
page0 should be a NOP 
init:    

    //your initialisation code here
    SET.B    B_TPM1C,4    // Pre Scaler (1us resolution)
    SET.W    W_PERIOD1,10000    // 10,000us = 100Hz
    SET.W    W_DUTY1,10001    // Set PWM1 Output OFF
    START 1, relay1process   // this independant process will monitor USERI
for action
main:
    PUSH.L    L_RTC        
    PUSH.L  L_USERN
    SUB.L             //compare current time with sheduled time
    AIS    -4             //restore stack
    BNE notnow1         // if strictly equal, start cycle (may be true twice
within the same second, but this is harmless)
    SET.W  W_USERI, W_USERL // set duration to decounting variable USERI, 
                            // this will invite process1 to turn on for
duration
                            // this approach allows master check remaining
time
                            // and also start/interrupt actions.
notnow1:    
    //similar code for relay2
    
    
    WAIT    10    //not too long to ensure that each seconds are seen by the
loop
    JMP main

relay1process: // asynchronous process for relay1 control, the input is
USERI variable representing duration in seconds
    SET.W    W_DUTY1,10001  // turn off relay1
    WAIT    1
    CMP.W    W_USERI,0
    BEQ        relay1process
    CLR.W     W_DUTY1 // turn on relay1   
loop_till_end1:    
    WAIT     16          //wait one second 
    DEC.W    W_USERI
    BGR        loop_till_end1  // master may set directly to 0 to stop relay
action, in this case USERI may become negative
    CLR.W    W_USERI // set to zero in case of -1
    BRA        relay1process


    
    



2010/3/27 Rob Conway <rjconway@...>


I thought it would be easier having the ON time in seconds since midnight so
I can compare direct to RTC.  
The length of time the output is ON needs to be able to vary between 1
minutes to 12 hrs.  I know I could hard code the ON / OFF time into the
stack however whilst I still have user variables I may as well use them. 

-----Original Message-----
From: Pascal Baerten [mailto:pascal.baerten@...] 
Sent: Saturday, March 27, 2010 6:59 PM
To: OWFS (One-wire file system) discussion and help
Subject: Re: [Owfs-developers] BAE0910 AE Timed output code


Hi Rob,
Sorry, I was busy to unexpeted work during this week and worked very late
everyday. 
You have to use the firmware I sent earlier.

I would create a main process that evaluate start conditions and initate sub
process that handle the action until duration.

some comments in your code below.

I will write you a sample skeleton later today.

Pascal




2010/3/27 Rob Conway <rjconway@...>


Hi,
 
I was hoping to get some feedback on the AE engine and creating some simple
output timers based upon energizing outputs based on time of day.  I can
already set the RTC from the HOST and it resets at midnight, so the RTC
represents seconds since midnight.
 
This is what I am trying to achieve on the stack 
 
    IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
    then DUTY1,0
    else DUTY1,10001
 
This is some rough code I thought I could use however not tested, I wanted
to know if I was on the correct track and if the stack can handle maths
results that go negative.

yes, 
for Byte, values above 127 are considered negative
for Word, values above 32767 are considered negative
for Long, values above 2147483647 are considered negative



 

//  ON / OFF TIMES  (Seconds)
//   L_UserN  ON time for Relay1   Seconds since Midnight  
//   W_UserL  Duration for PIO       Duration for relay ON time, seconds

what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
instead variable. 


 
 
relay1_ctl:
CMP.L    L_RTC,L_USERL
BLE        relay1_off
 
PUSH.L    L_RTC          //Get time since midnight
PUSH.L   L_USERN      //  get user ON time in seconds since midnight
SUB                             // Find out how many seconds difference
between RTC and on time   
CMP.W     @-1, W_USERL    // See if difference is less than Duration, if so
the output should be ON

as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
require 4 bytes 
important to retore stack pointer to inital value  after stack operation AIX
-4 


BGR        relay_off         
BLE         relay_on
END
 
relay1_on
SET.W    W_DUTY1,0
BRA Relay2_ctl
 
Relay1_off

SET.W    W_DUTY1,10001

BRA Relay2_ctl
 
 

relay2_ctl:
~ code similar to relay 1
 

----------------------------------------------------------------------------
--
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Owfs-developers mailing list
Owfs-developers@...
https://lists.sourceforge.net/lists/listinfo/owfs-developers






-- 
Pascal
 http://www.brain4home.eu


No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00





----------------------------------------------------------------------------
--
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Owfs-developers mailing list
Owfs-developers@...
https://lists.sourceforge.net/lists/listinfo/owfs-developers






-- 
Pascal
 http://www.brain4home.eu


No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00

Oops sorry, I read BNE as branch if negative however on review its Brand Not
Equal so sorry for my Q1 below. And I assume AIX was a typo for AIS  
Could I replace the first couple of lines with CMP.L   RTC, L_USERN.
 
I would of liked to be able to still turn the output on even if I missed the
time when RTC=ON time.  I know this complicates matters however as my light
is on for 10-12 hours during the day if I reload the firmware or restart the
slave and miss the exact on time the output will not turn on for another
day.

-----Original Message-----
From: Rob Conway [mailto:rjconway@...] 
Sent: Saturday, March 27, 2010 10:39 PM
To: 'OWFS (One-wire file system) discussion and help'
Subject: Re: [Owfs-developers] BAE0910 AE Timed output code


Pascal,  Thanks I understand the concept of starting another process to
handle the relay and duration.
 
 
I do have a couple of questions:
 
1)  I think though that whilst RTC is greater than the ON time the "main:"
will always be writing the duration to UserI thus never alowing
loop_till_end1: to decrement the value to 0 again ?
 
2) I have tried to read the manual again with regards to stack pointer and
why AIS -4, however I am still lost on this one.  
 
3) I assume AIX -4 is required in my first example as the CMP.L uses the
stack for calculations and the result of the compare is at position -5 on
the stack ?
 
 
rob

-----Original Message-----
From: Pascal Baerten [mailto:pascal.baerten@...] 
Sent: Saturday, March 27, 2010 9:18 PM
To: OWFS (One-wire file system) discussion and help
Subject: Re: [Owfs-developers] BAE0910 AE Timed output code


here is the sample: I found this approach easier to have better control from
the master.
Don't hesitate for any questions.



// AutomationEngine assembly source.
// Used for Temperature and Light Control of Reef Tank 
// March 2010 by Rob Conway
// 
// BAE0910 Direct Connected to Futurlec 4 channel opto Realy Board
// Channel 0 & 1 Relay has been replaced with a Mosfet IRF540
//
// RELAY Ch#   BAE0910 Ch#   Ribbon Cable    USAGE
// Output 0    PWM1          (Brown)         Cooling Fan
// Output 1    PWM2         (Orange)        Heater
// Output 2    PIO           (Yellow)        Main Light
// Output 3    OUT           (Blue)          Refugium Light
//
//
// VARIABLES:
//  ON / OFF TIMES  (Seconds)
//   L_UserN  ON time for Relay1   Seconds since Midnight 
//   W_UserL  Duration for PIO       Duration for relay ON time, seconds
//   W_USERI  decrementing variale with time remaining until realy1 is
turned OFF


#include "bae0910.inc"

#eeprom 0, 0      // start_page, end_page
#org  $00      // Code generated at address specified.

begin:
    NOP     // Allow auto start of process #0 on poweron, first byte of
page0 should be a NOP 
init:    

    //your initialisation code here
    SET.B    B_TPM1C,4    // Pre Scaler (1us resolution)
    SET.W    W_PERIOD1,10000    // 10,000us = 100Hz
    SET.W    W_DUTY1,10001    // Set PWM1 Output OFF
    START 1, relay1process   // this independant process will monitor USERI
for action
main:
    PUSH.L    L_RTC        
    PUSH.L  L_USERN
    SUB.L             //compare current time with sheduled time
    AIS    -4             //restore stack
    BNE notnow1         // if strictly equal, start cycle (may be true twice
within the same second, but this is harmless)
    SET.W  W_USERI, W_USERL // set duration to decounting variable USERI, 
                            // this will invite process1 to turn on for
duration
                            // this approach allows master check remaining
time
                            // and also start/interrupt actions.
notnow1:    
    //similar code for relay2
    
    
    WAIT    10    //not too long to ensure that each seconds are seen by the
loop
    JMP main

relay1process: // asynchronous process for relay1 control, the input is
USERI variable representing duration in seconds
    SET.W    W_DUTY1,10001  // turn off relay1
    WAIT    1
    CMP.W    W_USERI,0
    BEQ        relay1process
    CLR.W     W_DUTY1 // turn on relay1   
loop_till_end1:    
    WAIT     16          //wait one second 
    DEC.W    W_USERI
    BGR        loop_till_end1  // master may set directly to 0 to stop relay
action, in this case USERI may become negative
    CLR.W    W_USERI // set to zero in case of -1
    BRA        relay1process


    
    



2010/3/27 Rob Conway <rjconway@...>


I thought it would be easier having the ON time in seconds since midnight so
I can compare direct to RTC.  
The length of time the output is ON needs to be able to vary between 1
minutes to 12 hrs.  I know I could hard code the ON / OFF time into the
stack however whilst I still have user variables I may as well use them. 

-----Original Message-----
From: Pascal Baerten [mailto:pascal.baerten@...] 
Sent: Saturday, March 27, 2010 6:59 PM
To: OWFS (One-wire file system) discussion and help
Subject: Re: [Owfs-developers] BAE0910 AE Timed output code


Hi Rob,
Sorry, I was busy to unexpeted work during this week and worked very late
everyday. 
You have to use the firmware I sent earlier.

I would create a main process that evaluate start conditions and initate sub
process that handle the action until duration.

some comments in your code below.

I will write you a sample skeleton later today.

Pascal




2010/3/27 Rob Conway <rjconway@...>


Hi,
 
I was hoping to get some feedback on the AE engine and creating some simple
output timers based upon energizing outputs based on time of day.  I can
already set the RTC from the HOST and it resets at midnight, so the RTC
represents seconds since midnight.
 
This is what I am trying to achieve on the stack 
 
    IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
    then DUTY1,0
    else DUTY1,10001
 
This is some rough code I thought I could use however not tested, I wanted
to know if I was on the correct track and if the stack can handle maths
results that go negative.

yes, 
for Byte, values above 127 are considered negative
for Word, values above 32767 are considered negative
for Long, values above 2147483647 are considered negative



 

//  ON / OFF TIMES  (Seconds)
//   L_UserN  ON time for Relay1   Seconds since Midnight  
//   W_UserL  Duration for PIO       Duration for relay ON time, seconds

what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
instead variable. 


 
 
relay1_ctl:
CMP.L    L_RTC,L_USERL
BLE        relay1_off
 
PUSH.L    L_RTC          //Get time since midnight
PUSH.L   L_USERN      //  get user ON time in seconds since midnight
SUB                             // Find out how many seconds difference
between RTC and on time   
CMP.W     @-1, W_USERL    // See if difference is less than Duration, if so
the output should be ON

as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
require 4 bytes 
important to retore stack pointer to inital value  after stack operation AIX
-4 


BGR        relay_off         
BLE         relay_on
END
 
relay1_on
SET.W    W_DUTY1,0
BRA Relay2_ctl
 
Relay1_off

SET.W    W_DUTY1,10001

BRA Relay2_ctl
 
 

relay2_ctl:
~ code similar to relay 1
 

----------------------------------------------------------------------------
--
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Owfs-developers mailing list
Owfs-developers@...
https://lists.sourceforge.net/lists/listinfo/owfs-developers






-- 
Pascal
 http://www.brain4home.eu


No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00





----------------------------------------------------------------------------
--
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Owfs-developers mailing list
Owfs-developers@...
https://lists.sourceforge.net/lists/listinfo/owfs-developers






-- 
Pascal
 http://www.brain4home.eu


No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00




No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00

see below

2010/3/27 Rob Conway <rjconway@...>

>  Pascal,  Thanks I understand the concept of starting another process to
> handle the relay and duration.
>
>
> I do have a couple of questions:
>
> 1)  I think though that whilst RTC is greater than the ON time the "main:"
> will always be writing the duration to UserI thus never alowing
> loop_till_end1: to decrement the value to 0 again ?
>
the SUB.W instruction set Z flag to 0 when substraction result is not null
else set to 1.
the BNE instruction do a branch if Z flag==0.
so the affectation of useri is only done precisely at RTC==usern not before,
not after.


> 2) I have tried to read the manual again with regards to stack pointer and
> why AIS -4, however I am still lost on this one.
>
to restore the stack pointer at previous position, you have to pop every
byte/word/longs pushed before.
the code pushed 2 longs before sub.l
the sub.l pop the two needed operands, calculate and push back the result
long.
after such operation, whe have to pop.l the result to a user variable to
free the stack and store the result, however here we don't need the use the
result, only the flags resulting from the opreation are important. So to
avoid consuming a user variable to pop the result, we simply adjust stack
pointer with ais -4.



> 3) I assume AIX -4 is required in my first example as the CMP.L uses the
> stack for calculations and the result of the compare is at position -5 on
> the stack ?
>
AIS not AIX (my bad)
SP (stack pointer) is always pointing to next avail location.
SP -4 is the position of the last long in the stack: use 4 bytes from SP-4
to and including  SP-1.
CMP is a SUB that does not write result back, only flags updated.


>
> rob
>
>  -----Original Message-----
> *From:* Pascal Baerten [mailto:pascal.baerten@...]
> *Sent:* Saturday, March 27, 2010 9:18 PM
> *To:* OWFS (One-wire file system) discussion and help
> *Subject:* Re: [Owfs-developers] BAE0910 AE Timed output code
>
> here is the sample: I found this approach easier to have better control
> from the master.
> Don't hesitate for any questions.
>
> // AutomationEngine assembly source.
>> // Used for Temperature and Light Control of Reef Tank
>> // March 2010 by Rob Conway
>> //
>> // BAE0910 Direct Connected to Futurlec 4 channel opto Realy Board
>> // Channel 0 & 1 Relay has been replaced with a Mosfet IRF540
>> //
>> // RELAY Ch#   BAE0910 Ch#   Ribbon Cable    USAGE
>> // Output 0    PWM1          (Brown)         Cooling Fan
>> // Output 1    PWM2         (Orange)        Heater
>> // Output 2    PIO           (Yellow)        Main Light
>> // Output 3    OUT           (Blue)          Refugium Light
>> //
>> //
>> // VARIABLES:
>> //  ON / OFF TIMES  (Seconds)
>> //   L_UserN  ON time for Relay1   Seconds since Midnight
>> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
>> //   W_USERI  decrementing variale with time remaining until realy1 is
>> turned OFF
>>
>>
>> #include "bae0910.inc"
>>
>> #eeprom 0, 0      // start_page, end_page
>> #org  $00      // Code generated at address specified.
>>
>> begin:
>>     NOP     // Allow auto start of process #0 on poweron, first byte of
>> page0 should be a NOP
>> init:
>>
>>     //your initialisation code here
>>     SET.B    B_TPM1C,4    // Pre Scaler (1us resolution)
>>     SET.W    W_PERIOD1,10000    // 10,000us = 100Hz
>>     SET.W    W_DUTY1,10001    // Set PWM1 Output OFF
>>     START 1, relay1process   // this independant process will monitor
>> USERI for action
>> main:
>>     PUSH.L    L_RTC
>>     PUSH.L  L_USERN
>>     SUB.L             //compare current time with sheduled time
>>     AIS    -4             //restore stack
>>     BNE notnow1         // if strictly equal, start cycle (may be true
>> twice within the same second, but this is harmless)
>>     SET.W  W_USERI, W_USERL // set duration to decounting variable USERI,
>>                             // this will invite process1 to turn on for
>> duration
>>                             // this approach allows master check remaining
>> time
>>                             // and also start/interrupt actions.
>> notnow1:
>>     //similar code for relay2
>>
>>
>>     WAIT    10    //not too long to ensure that each seconds are seen by
>> the loop
>>     JMP main
>>
>> relay1process: // asynchronous process for relay1 control, the input is
>> USERI variable representing duration in seconds
>>     SET.W    W_DUTY1,10001  // turn off relay1
>>     WAIT    1
>>     CMP.W    W_USERI,0
>>     BEQ        relay1process
>>     CLR.W     W_DUTY1 // turn on relay1
>> loop_till_end1:
>>     WAIT     16          //wait one second
>>     DEC.W    W_USERI
>>     BGR        loop_till_end1  // master may set directly to 0 to stop
>> relay action, in this case USERI may become negative
>>     CLR.W    W_USERI // set to zero in case of -1
>>     BRA        relay1process
>>
>
>
>
>
> 2010/3/27 Rob Conway <rjconway@...>
>
>>  I thought it would be easier having the ON time in seconds since
>> midnight so I can compare direct to RTC.
>> The length of time the output is ON needs to be able to vary between 1
>> minutes to 12 hrs.  I know I could hard code the ON / OFF time into the
>> stack however whilst I still have user variables I may as well use them.
>>
>>   -----Original Message-----
>> *From:* Pascal Baerten [mailto:pascal.baerten@...]
>> *Sent:* Saturday, March 27, 2010 6:59 PM
>> *To:* OWFS (One-wire file system) discussion and help
>> *Subject:* Re: [Owfs-developers] BAE0910 AE Timed output code
>>
>> Hi Rob,
>> Sorry, I was busy to unexpeted work during this week and worked very late
>> everyday.
>> You have to use the firmware I sent earlier.
>>
>> I would create a main process that evaluate start conditions and initate
>> sub process that handle the action until duration.
>>
>> some comments in your code below.
>>
>> I will write you a sample skeleton later today.
>>
>> Pascal
>>
>>
>>
>> 2010/3/27 Rob Conway <rjconway@...>
>>
>>>  Hi,
>>>
>>> I was hoping to get some feedback on the AE engine and creating some
>>> simple output timers based upon energizing outputs based on time of day.  I
>>> can already set the RTC from the HOST and it resets at midnight, so the RTC
>>> represents seconds since midnight.
>>>
>>> This is what I am trying to achieve on the stack
>>>
>>>     IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
>>>     then DUTY1,0
>>>     else DUTY1,10001
>>>
>>> This is some rough code I thought I could use however not tested, I
>>> wanted to know if I was on the correct track and if the stack can handle
>>> maths results that go negative.
>>>
>> yes,
>> for Byte, values above 127 are considered negative
>> for Word, values above 32767 are considered negative
>> for Long, values above 2147483647 are considered negative
>>
>>
>>>
>>  //  ON / OFF TIMES  (Seconds)
>>> //   L_UserN  ON time for Relay1   Seconds since Midnight
>>> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
>>>
>> what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
>> instead variable.
>>
>>>
>>>
>>> relay1_ctl:
>>> CMP.L    L_RTC,L_USERL
>>> BLE        relay1_off
>>>
>>> PUSH.L    L_RTC          //Get time since midnight
>>> PUSH.L   L_USERN      //  get user ON time in seconds since midnight
>>> SUB                             // Find out how many seconds difference
>>> between RTC and on time
>>> CMP.W     @-1, W_USERL    // See if difference is less than Duration, if
>>> so the output should be ON
>>>
>> as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
>> require 4 bytes
>> important to retore stack pointer to inital value  after stack operation
>> AIX -4
>>
>>>  BGR        relay_off
>>> BLE         relay_on
>>> END
>>>
>>> relay1_on
>>> SET.W    W_DUTY1,0
>>> BRA Relay2_ctl
>>>
>>> Relay1_off
>>>  SET.W    W_DUTY1,10001
>>>  BRA Relay2_ctl
>>>
>>>
>>>  relay2_ctl:
>>> ~ code similar to relay 1
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Download Intel&#174; Parallel Studio Eval
>>> Try the new software tools for yourself. Speed compiling, find bugs
>>> proactively, and fine-tune applications for parallel performance.
>>> See why Intel Parallel Studio got high marks during beta.
>>> http://p.sf.net/sfu/intel-sw-dev
>>> _______________________________________________
>>> Owfs-developers mailing list
>>> Owfs-developers@...
>>> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>>>
>>>
>>
>>
>> --
>> Pascal
>> http://www.brain4home.eu
>>
>> No virus found in this incoming message.
>> Checked by AVG - http://www.avg.com
>> Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
>> 06:33:00
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> Owfs-developers mailing list
>> Owfs-developers@...
>> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>>
>>
>
>
> --
> Pascal
> http://www.brain4home.eu
>
> No virus found in this incoming message.
> Checked by AVG - http://www.avg.com
> Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
> 06:33:00
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
>


-- 
Pascal
 http://www.brain4home.eu

> see below
>
> 2010/3/27 Rob Conway <rjconway@...>
>
>>  Oops sorry, I read BNE as branch if negative however on review its Brand
>> Not Equal so sorry for my Q1 below. And I assume AIX was a typo for AIS
>> Could I replace the first couple of lines with CMP.L   RTC, L_USERN.
>>
> no, CMP require immediate operand2, you need to use the stack.
>
>
>
>>
> I would of liked to be able to still turn the output on even if I missed
>> the time when RTC=ON time.  I know this complicates matters however as my
>> light is on for 10-12 hours during the day if I reload the firmware or
>> restart the slave and miss the exact on time the output will not turn on for
>> another day.
>>
> if the chip is reset or powered cycled, RTC will restart to zero and will
> require master attention for RTC reinitialisation.
> You may during init, put an alarmed flag to allows master conditional
> search to know the the device require attention.
> On the master side, monitor alarmed directory to update chip if needed: RTC
> as well as force initial state.
> Alternatively, you may want to put some battery backup circuit.
>
> if you want to calculate an end time based on start time (long) + duration
> (word):
> PUSH.L L_USERN  //start time in the stack
> PUSH.W 0  //hi part of operand 2 is zero
> PUSH.W W_USERL // second part of operand 2
> ADD.L // pop both longs, add them and store resulting long in the stack
> then you may know if end time is reached or not:
> PUSH.L L_RTC
> SUB.L // calculate endduration-RTC
> AIS -4 // free stack
> B... skiplabel //conditional branch depending of last compare (SUB)
>
> hope this clarify...
>
-- 
Pascal
 http://www.brain4home.eu

see below

2010/3/27 Rob Conway <rjconway@...>

>  Oops sorry, I read BNE as branch if negative however on review its Brand
> Not Equal so sorry for my Q1 below. And I assume AIX was a typo for AIS
> Could I replace the first couple of lines with CMP.L   RTC, L_USERN.
>
no, CMP require immediate operand2, you need to use the stack.



>
I would of liked to be able to still turn the output on even if I missed the
> time when RTC=ON time.  I know this complicates matters however as my light
> is on for 10-12 hours during the day if I reload the firmware or restart the
> slave and miss the exact on time the output will not turn on for another
> day.
>
if the chip is reset or powered cycled, RTC will restart to zero and will
require master attention for RTC reinitialisation.
You may during init, put an alarmed flag to allows master conditional search
to know the the device require attention.
On the master side, monitor alarmed directory to update chip if needed: RTC
as well as force initial state.
Alternatively, you may want to put some battery backup circuit.

if you want to calculate an end time based on start time (long) + duration
(word):
PUSH.L L_USERN  //start time in the stack
PUSH.W 0  //hi part of operand 2 is zero
PUSH.W W_USERL // second part of operand 2
ADD.L // pop both longs, add them and store resulting long in the stack
then you may know if end time is reached or not:
PUSH.L L_RTC
SUB.L // calculate endduration-RTC
AIS -4 // free stack
B... skiplabel //conditional branch depending of last compare (SUB)

hope this clarify...

>  -----Original Message-----
> *From:* Rob Conway [mailto:rjconway@...]
> *Sent:* Saturday, March 27, 2010 10:39 PM
> *To:* 'OWFS (One-wire file system) discussion and help'
> *Subject:* Re: [Owfs-developers] BAE0910 AE Timed output code
>
> Pascal,  Thanks I understand the concept of starting another process to
> handle the relay and duration.
>
>
> I do have a couple of questions:
>
> 1)  I think though that whilst RTC is greater than the ON time the "main:"
> will always be writing the duration to UserI thus never alowing
> loop_till_end1: to decrement the value to 0 again ?
>
> 2) I have tried to read the manual again with regards to stack pointer and
> why AIS -4, however I am still lost on this one.
>
> 3) I assume AIX -4 is required in my first example as the CMP.L uses the
> stack for calculations and the result of the compare is at position -5 on
> the stack ?
>
>
> rob
>
>  -----Original Message-----
> *From:* Pascal Baerten [mailto:pascal.baerten@...]
> *Sent:* Saturday, March 27, 2010 9:18 PM
> *To:* OWFS (One-wire file system) discussion and help
> *Subject:* Re: [Owfs-developers] BAE0910 AE Timed output code
>
> here is the sample: I found this approach easier to have better control
> from the master.
> Don't hesitate for any questions.
>
> // AutomationEngine assembly source.
>> // Used for Temperature and Light Control of Reef Tank
>> // March 2010 by Rob Conway
>> //
>> // BAE0910 Direct Connected to Futurlec 4 channel opto Realy Board
>> // Channel 0 & 1 Relay has been replaced with a Mosfet IRF540
>> //
>> // RELAY Ch#   BAE0910 Ch#   Ribbon Cable    USAGE
>> // Output 0    PWM1          (Brown)         Cooling Fan
>> // Output 1    PWM2         (Orange)        Heater
>> // Output 2    PIO           (Yellow)        Main Light
>> // Output 3    OUT           (Blue)          Refugium Light
>> //
>> //
>> // VARIABLES:
>> //  ON / OFF TIMES  (Seconds)
>> //   L_UserN  ON time for Relay1   Seconds since Midnight
>> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
>> //   W_USERI  decrementing variale with time remaining until realy1 is
>> turned OFF
>>
>>
>> #include "bae0910.inc"
>>
>> #eeprom 0, 0      // start_page, end_page
>> #org  $00      // Code generated at address specified.
>>
>> begin:
>>     NOP     // Allow auto start of process #0 on poweron, first byte of
>> page0 should be a NOP
>> init:
>>
>>     //your initialisation code here
>>     SET.B    B_TPM1C,4    // Pre Scaler (1us resolution)
>>     SET.W    W_PERIOD1,10000    // 10,000us = 100Hz
>>     SET.W    W_DUTY1,10001    // Set PWM1 Output OFF
>>     START 1, relay1process   // this independant process will monitor
>> USERI for action
>> main:
>>     PUSH.L    L_RTC
>>     PUSH.L  L_USERN
>>     SUB.L             //compare current time with sheduled time
>>     AIS    -4             //restore stack
>>     BNE notnow1         // if strictly equal, start cycle (may be true
>> twice within the same second, but this is harmless)
>>     SET.W  W_USERI, W_USERL // set duration to decounting variable USERI,
>>                             // this will invite process1 to turn on for
>> duration
>>                             // this approach allows master check remaining
>> time
>>                             // and also start/interrupt actions.
>> notnow1:
>>     //similar code for relay2
>>
>>
>>     WAIT    10    //not too long to ensure that each seconds are seen by
>> the loop
>>     JMP main
>>
>> relay1process: // asynchronous process for relay1 control, the input is
>> USERI variable representing duration in seconds
>>     SET.W    W_DUTY1,10001  // turn off relay1
>>     WAIT    1
>>     CMP.W    W_USERI,0
>>     BEQ        relay1process
>>     CLR.W     W_DUTY1 // turn on relay1
>> loop_till_end1:
>>     WAIT     16          //wait one second
>>     DEC.W    W_USERI
>>     BGR        loop_till_end1  // master may set directly to 0 to stop
>> relay action, in this case USERI may become negative
>>     CLR.W    W_USERI // set to zero in case of -1
>>     BRA        relay1process
>>
>
>
>
>
> 2010/3/27 Rob Conway <rjconway@...>
>
>>  I thought it would be easier having the ON time in seconds since
>> midnight so I can compare direct to RTC.
>> The length of time the output is ON needs to be able to vary between 1
>> minutes to 12 hrs.  I know I could hard code the ON / OFF time into the
>> stack however whilst I still have user variables I may as well use them.
>>
>>   -----Original Message-----
>> *From:* Pascal Baerten [mailto:pascal.baerten@...]
>> *Sent:* Saturday, March 27, 2010 6:59 PM
>> *To:* OWFS (One-wire file system) discussion and help
>> *Subject:* Re: [Owfs-developers] BAE0910 AE Timed output code
>>
>> Hi Rob,
>> Sorry, I was busy to unexpeted work during this week and worked very late
>> everyday.
>> You have to use the firmware I sent earlier.
>>
>> I would create a main process that evaluate start conditions and initate
>> sub process that handle the action until duration.
>>
>> some comments in your code below.
>>
>> I will write you a sample skeleton later today.
>>
>> Pascal
>>
>>
>>
>> 2010/3/27 Rob Conway <rjconway@...>
>>
>>>  Hi,
>>>
>>> I was hoping to get some feedback on the AE engine and creating some
>>> simple output timers based upon energizing outputs based on time of day.  I
>>> can already set the RTC from the HOST and it resets at midnight, so the RTC
>>> represents seconds since midnight.
>>>
>>> This is what I am trying to achieve on the stack
>>>
>>>     IF RTC > ON_TIME and RTC <(ON-TIME+DURATION)
>>>     then DUTY1,0
>>>     else DUTY1,10001
>>>
>>> This is some rough code I thought I could use however not tested, I
>>> wanted to know if I was on the correct track and if the stack can handle
>>> maths results that go negative.
>>>
>> yes,
>> for Byte, values above 127 are considered negative
>> for Word, values above 32767 are considered negative
>> for Long, values above 2147483647 are considered negative
>>
>>
>>>
>>  //  ON / OFF TIMES  (Seconds)
>>> //   L_UserN  ON time for Relay1   Seconds since Midnight
>>> //   W_UserL  Duration for PIO       Duration for relay ON time, seconds
>>>
>> what's the max duration? ,maybe fit in 255 range or 65535  to use B or W
>> instead variable.
>>
>>>
>>>
>>> relay1_ctl:
>>> CMP.L    L_RTC,L_USERL
>>> BLE        relay1_off
>>>
>>> PUSH.L    L_RTC          //Get time since midnight
>>> PUSH.L   L_USERN      //  get user ON time in seconds since midnight
>>> SUB                             // Find out how many seconds difference
>>> between RTC and on time
>>> CMP.W     @-1, W_USERL    // See if difference is less than Duration, if
>>> so the output should be ON
>>>
>> as operands are L_ types, use CMP.L, also, @-4 needed for the L_ type who
>> require 4 bytes
>> important to retore stack pointer to inital value  after stack operation
>> AIX -4
>>
>>>  BGR        relay_off
>>> BLE         relay_on
>>> END
>>>
>>> relay1_on
>>> SET.W    W_DUTY1,0
>>> BRA Relay2_ctl
>>>
>>> Relay1_off
>>>  SET.W    W_DUTY1,10001
>>>  BRA Relay2_ctl
>>>
>>>
>>>  relay2_ctl:
>>> ~ code similar to relay 1
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Download Intel&#174; Parallel Studio Eval
>>> Try the new software tools for yourself. Speed compiling, find bugs
>>> proactively, and fine-tune applications for parallel performance.
>>> See why Intel Parallel Studio got high marks during beta.
>>> http://p.sf.net/sfu/intel-sw-dev
>>> _______________________________________________
>>> Owfs-developers mailing list
>>> Owfs-developers@...
>>> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>>>
>>>
>>
>>
>> --
>> Pascal
>> http://www.brain4home.eu
>>
>> No virus found in this incoming message.
>> Checked by AVG - http://www.avg.com
>> Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
>> 06:33:00
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> Owfs-developers mailing list
>> Owfs-developers@...
>> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>>
>>
>
>
> --
> Pascal
> http://www.brain4home.eu
>
> No virus found in this incoming message.
> Checked by AVG - http://www.avg.com
> Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
> 06:33:00
>
> No virus found in this incoming message.
> Checked by AVG - http://www.avg.com
> Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
> 06:33:00
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
>


-- 
Pascal
 http://www.brain4home.eu

On Fri, Mar 26, 2010 at 3:21 AM, Alex Shepherd <lists06@...> wrote:

> Just joined the list, so - hello everyone!
Welcome!
>
> Last year I've had some custom PCB's made (min order of 100 actually) for
> the DS28EA00 sensor to use in my new house to monitor the floor/room
> temperatures and control my underfloor heating system, that's powered by a

The DS28EA00 is supported in OWFS for temperature and PIO. Chain mode
hasn't been implemented, since there was no demand and since it fits a
little awkwardly in our filesystem metaphor.
>
> Having a background in industrial automation and control system I thought
> I'd try and be a bit smarter with the controls side of things and sprinkle a
> bunch of sensors around the house to have a better idea of the thermal load
> characteristics of the house and hopefully control it a bit better than with
> just a simple thermostat on the wall. However I did run some mains cables to
> where thermostats on the wall could go in case my grand plans fail... :(

There is a DS1821 thermostat chip. It's only sort of a 1-wire device,
with no ID and needing external circuitry to return to 1-wire control.
You would be better off using a microprocessor 1-wire slave.

> I have run 2 separate microLAN's using Cat5e UTP cable that will have about
> 20-30 sensors on each LAN. These LAN's run back to my central comms rack. I
> decided to keep the LAN's simple and linear as I need robustness and so I
> avoided going down the path of hubs etc to manage what would become a star
> network with it's associated impedance problems.

Very nice. Even if your wire runs didn't return to the same location,
you could have joined the 1-wire networks with owserver over TCP/IP.

> I have 3 x LinkUSB adaptors (1 spare for a while) and an old Compaq Thin
> Client with a 2.5" 80G disk and Debian LXDE Linux installed. I'll use this
> initially to get things going and host a web server to display the various
> room temperatures on a nice graphs to assist with commissioning the system.
> Hopefully I will be able to move some or all of this to a diskless NSLU2
> once things are all working.

The NSLU2 certainly works. Something more recent like the SheevaPlug
is the same price and power consumption and has 512MB ram vs 32MB.

> Initially I need to get my head around how to configure owfs to interface
> the 2 LinkUSB adaptors and get the temperatures stored in a database
> somewhere on my LAN and display the temperatures on graphs via a web
> browser. Once I get the measurement system in place I'll move on to
> controlling it.

I'd use owserver as the interface to the LINK adapters, and then
communicate with owserver using the various OWFS programs. This allows
ad-hoc queries and debugging, while also running a data collection and
control process separately.

The general syntax is
/opt/owfs/bin/owserver -d serial_port -p tcp_port (default tcp_port of
4304 is used if -p is omitted)

Linking the two adapters is effortless, just list them both on the command line.
/opt/owfs/bin/owserver --LINK /dev/ttyS0 --LINK /dev/ttyS1
(Here I used the LINKs in LINK mode rather than DS9097U emulation
mode, and used the default port of 4304)
Depending on serial port permissions, this may have to be run as root.

To see your network:
/opt/owfs/bin/owdir

To mount it as a filesystem on /mnt/1wire (with proper permission)
/opt/owfs/bin/owfs -s 4304 -m /mnt/1wire
Then get all your temperatures with
grep -r '.' /mnt/1wire/*/temperature

You should also look at the alias file entry to give each of the
sensor ID a human readable name (like the location of the sensor).

As for making web interface, and storing in a database, I'll let other
people chime in.

Paul Alfille

I've got a (very crude, basic) temperature logging/graphing system running at:

http://eme106-temperature.eecs.wsu.edu

It monitors and logs our temperatures for 11 "points".

We used DS18B20's for temperature sensing (which we also use in other
projects).  I highly recommend them for temperature sensing.  I also
strongly recommend avoiding the older, but compatible 18S20 -- they
are several times slower and are more finicky.  The DS18B20's cost in
the $2-3 range if I recall correctly.  They're a single-chip solution
in a TO-92 case.  One pin OW, one pin GND, and one pin either tied to
ground for parasitic mode, or hooked up to +5 for self-powered mode.
Its device family 28 for ID's.

We're running this particular system on a P3-700Mhz tower that was
laying around; for our main research projects we use the Sheeva Plugs,
which I do recommend...the sheeva's are a great device for things such
as this.  For logging, we are using a perl script and owperl to sleep
for 5 min, then read "fasttemp" from the uncached directory on all of
our sensors, storing the values into a mysql db.

After the collection run finishes, it calls a php script that
generates the main cart and temp table.  When you load it, there's a
bit of php code to ensure the data is current (the little check at the
top), but otherwise its a static page (it used to create all the
images on page load, but due to bad programming and weak hardware,
this was pretty hard on the server; the static content is much nicer,
and I now have no qualms about sending out the URL here...I wouldn't
have done that before!).

--Jim
Manager, WSU Artificial Intelligence Lab

On Fri, Mar 26, 2010 at 3:18 AM, Paul Alfille <paul.alfille@...> wrote:
> On Fri, Mar 26, 2010 at 3:21 AM, Alex Shepherd <lists06@...> wrote:
>
>> Just joined the list, so - hello everyone!
> Welcome!
>>
>> Last year I've had some custom PCB's made (min order of 100 actually) for
>> the DS28EA00 sensor to use in my new house to monitor the floor/room
>> temperatures and control my underfloor heating system, that's powered by a
>
> The DS28EA00 is supported in OWFS for temperature and PIO. Chain mode
> hasn't been implemented, since there was no demand and since it fits a
> little awkwardly in our filesystem metaphor.
>>
>> Having a background in industrial automation and control system I thought
>> I'd try and be a bit smarter with the controls side of things and sprinkle a
>> bunch of sensors around the house to have a better idea of the thermal load
>> characteristics of the house and hopefully control it a bit better than with
>> just a simple thermostat on the wall. However I did run some mains cables to
>> where thermostats on the wall could go in case my grand plans fail... :(
>
> There is a DS1821 thermostat chip. It's only sort of a 1-wire device,
> with no ID and needing external circuitry to return to 1-wire control.
> You would be better off using a microprocessor 1-wire slave.
>
>> I have run 2 separate microLAN's using Cat5e UTP cable that will have about
>> 20-30 sensors on each LAN. These LAN's run back to my central comms rack. I
>> decided to keep the LAN's simple and linear as I need robustness and so I
>> avoided going down the path of hubs etc to manage what would become a star
>> network with it's associated impedance problems.
>
> Very nice. Even if your wire runs didn't return to the same location,
> you could have joined the 1-wire networks with owserver over TCP/IP.
>
>> I have 3 x LinkUSB adaptors (1 spare for a while) and an old Compaq Thin
>> Client with a 2.5" 80G disk and Debian LXDE Linux installed. I'll use this
>> initially to get things going and host a web server to display the various
>> room temperatures on a nice graphs to assist with commissioning the system.
>> Hopefully I will be able to move some or all of this to a diskless NSLU2
>> once things are all working.
>
> The NSLU2 certainly works. Something more recent like the SheevaPlug
> is the same price and power consumption and has 512MB ram vs 32MB.
>
>> Initially I need to get my head around how to configure owfs to interface
>> the 2 LinkUSB adaptors and get the temperatures stored in a database
>> somewhere on my LAN and display the temperatures on graphs via a web
>> browser. Once I get the measurement system in place I'll move on to
>> controlling it.
>
> I'd use owserver as the interface to the LINK adapters, and then
> communicate with owserver using the various OWFS programs. This allows
> ad-hoc queries and debugging, while also running a data collection and
> control process separately.
>
> The general syntax is
> /opt/owfs/bin/owserver -d serial_port -p tcp_port (default tcp_port of
> 4304 is used if -p is omitted)
>
> Linking the two adapters is effortless, just list them both on the command line.
> /opt/owfs/bin/owserver --LINK /dev/ttyS0 --LINK /dev/ttyS1
> (Here I used the LINKs in LINK mode rather than DS9097U emulation
> mode, and used the default port of 4304)
> Depending on serial port permissions, this may have to be run as root.
>
> To see your network:
> /opt/owfs/bin/owdir
>
> To mount it as a filesystem on /mnt/1wire (with proper permission)
> /opt/owfs/bin/owfs -s 4304 -m /mnt/1wire
> Then get all your temperatures with
> grep -r '.' /mnt/1wire/*/temperature
>
> You should also look at the alias file entry to give each of the
> sensor ID a human readable name (like the location of the sensor).
>
> As for making web interface, and storing in a database, I'll let other
> people chime in.
>
> Paul Alfille
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>

Hi Paul,
 
> The DS28EA00 is supported in OWFS for temperature and PIO.

Yes, I noted that and it kinda encouraged me down this path as I knew that
having some extra I/O available at the sensing point would be a way to
achieve the control I expect to need if I don't go down the custom 1-wire
slave path initially. 
 
> Chain mode hasn't been implemented, since there was no demand 
> and since it fits a little awkwardly in our filesystem metaphor.

Yeah, I had stared at this long and hard and while it looks to be useful, I
figured that I'd just add a SMT LED to the board that I can control remotely
would be another way to identify sensors a bit like id done with racks of
servers and disks etc that can enable a special flashing LED etc so you can
find things when there are a lot of them that all look the same.

> There is a DS1821 thermostat chip. It's only sort of a 1-wire 
> device, with no ID and needing external circuitry to return 
> to 1-wire control.
> You would be better off using a microprocessor 1-wire slave.

Yeah, It's not really clear in my mind yet how best to do the actual control
aspect of my project. 

I have a number of sensors in the same areas as I have 19 under-floor
heating zones in total, that are fed from 3 manifold cabinets (4, 6, 9 zones
respectively). 

I've got at least 1 sensor in the concrete for each zone and several on the
walls in those same areas. I thought I'd also try and measure the return
temperatures for each zone and average that with the measured floor
temperature to get an idea of what that zone is doing in terms of providing
a heat source. Then consider the ambient room temperature before deciding to
what should happen next. So it's not just a simple on/off thermostat 1
sensor 1 control output problem.

Also as the system is now there is really only 1 heat on/off control per
manifold and you just tune the flows through each zone to hopefully balance
all the zones relative to each other so that overall the whole area is at a
comfortable heat. I have about 5 zones that I have to pay particular
attention to as the flooring material (laminated wood) must not change
temperature too fast or invalidate my warranty. So this was my main
motivation to consider all this. 

> Very nice. Even if your wire runs didn't return to the same 
> location, you could have joined the 1-wire networks with 
> owserver over TCP/IP.

Yes, but as everything cabling wise runs back to this central location and
as I found the LinkUSB adaptors I figured it would solve the potential star
network problem and also terminate each LAN into the one device.

> The NSLU2 certainly works. Something more recent like the 
> SheevaPlug is the same price and power consumption and has 
> 512MB ram vs 32MB.

Yes, although living in NZ means international shipping which can vary in
cost and the NSLU2 is close at hand. However, I have my eye on the GuruPlug
+ JTAG variant, but it's not out yet so I figured the NSLU2 should be
adequate in the meantime and may be good for the long term. I need this type
of control to be solid and reliable and if the NSLU2 can just sit there
doing the sensing and control part of the job reliably, that would be ideal.
I've got several other PC's and a NAS that can do the logging over time and
web serving etc.

> I'd use owserver as the interface to the LINK adapters, and 
> then communicate with owserver using the various OWFS 
> programs. This allows ad-hoc queries and debugging, while 
> also running a data collection and control process separately.

Ok

> Linking the two adapters is effortless, just list them both 
> on the command line.
> /opt/owfs/bin/owserver --LINK /dev/ttyS0 --LINK /dev/ttyS1 
> (Here I used the LINKs in LINK mode rather than DS9097U 
> emulation mode, and used the default port of 4304) Depending 
> on serial port permissions, this may have to be run as root.

Excellent! 

I was thinking I might have to run two owservers and then somehow merge the
readings in the owfs somehow but having owserver handle multiple LinkUSB's
makes it all simple again.

> To see your network:
> /opt/owfs/bin/owdir
> 
> To mount it as a filesystem on /mnt/1wire (with proper 
> permission) /opt/owfs/bin/owfs -s 4304 -m /mnt/1wire Then get 
> all your temperatures with grep -r '.' /mnt/1wire/*/temperature
> 
> You should also look at the alias file entry to give each of 
> the sensor ID a human readable name (like the location of the sensor).

Again excellent. This seems like it might be easier than I had thought -
however past experience tells me that there is _always_ some issue that will
crop-up along the way. I just don't know what it is yet.

> As for making web interface, and storing in a database, I'll 
> let other people chime in.

Cool

Alex

Hi Jim,

> I've got a (very crude, basic) temperature logging/graphing 
> system running at:
> 
> http://eme106-temperature.eecs.wsu.edu
> 
> It monitors and logs our temperatures for 11 "points".

Ok. Looks adequate for what you're needing. Any other plans?

I've found several other projects but I've not got my mind around them yet
to know how they work and if they would be suitable for my needs.

> We're running this particular system on a P3-700Mhz tower 
> that was laying around; for our main research projects we use 
> the Sheeva Plugs, which I do recommend...the sheeva's are a 
> great device for things such as this.

Yes, they certainly look cool. I'm waiting for the GuruPlug variant to
become available. Until then I have several PC based options, but I don't
perceive their reliability to be as good as these low power, silent and
diskless/fanless options.

> For logging, we are 
> using a perl script and owperl to sleep for 5 min, then read 
> "fasttemp" from the uncached directory on all of our sensors, 
> storing the values into a mysql db.

Yeah, I expect to do the same into a MySQL database that my NAS can provide.

> After the collection run finishes, it calls a php script that 
> generates the main cart and temp table.  When you load it, 
> there's a bit of php code to ensure the data is current (the 
> little check at the top), but otherwise its a static page (it 
> used to create all the images on page load, but due to bad 
> programming and weak hardware, this was pretty hard on the 
> server; the static content is much nicer, and I now have no 
> qualms about sending out the URL here...I wouldn't have done 
> that before!).

Ok, that sounds like a good idea.

Anyway thanks for the tips, I really need to make some fast progress as
spare time is at a premium and the air temperature is getting chilly...

Alex

Hi Alex,

I used an NSLU2 for at least 1 year with good result for that hardware
platform.  I got about 4~8 weeks at a time before reboot.  I was using OWFS
and RRDTools for logging and trending.  I can remember that I could achieve
a higher uptime with the NSLU2 without using fuse.  My scripts simply used
owserver with s USB adapter

If you only had a couple of loops to actually automate I would be looking at
the BAE0910 if it can handle the I/O.  I am about to use two of these
devices and although there are alternatives they just seem a neat solution
for 1wire control and for those of us that are "time poor" they are easy to
deploy.

rob



-----Original Message-----
From: Alex Shepherd [mailto:lists06@...] 
Sent: Saturday, March 27, 2010 8:45 AM
To: 'OWFS (One-wire file system) discussion and help'
Subject: Re: [Owfs-developers] Hello and AVR based slave device


Hi Jim,

> I've got a (very crude, basic) temperature logging/graphing
> system running at:
> 
> http://eme106-temperature.eecs.wsu.edu
> 
> It monitors and logs our temperatures for 11 "points".

Ok. Looks adequate for what you're needing. Any other plans?

I've found several other projects but I've not got my mind around them yet
to know how they work and if they would be suitable for my needs.

> We're running this particular system on a P3-700Mhz tower
> that was laying around; for our main research projects we use 
> the Sheeva Plugs, which I do recommend...the sheeva's are a 
> great device for things such as this.

Yes, they certainly look cool. I'm waiting for the GuruPlug variant to
become available. Until then I have several PC based options, but I don't
perceive their reliability to be as good as these low power, silent and
diskless/fanless options.

> For logging, we are
> using a perl script and owperl to sleep for 5 min, then read 
> "fasttemp" from the uncached directory on all of our sensors, 
> storing the values into a mysql db.

Yeah, I expect to do the same into a MySQL database that my NAS can provide.

> After the collection run finishes, it calls a php script that
> generates the main cart and temp table.  When you load it, 
> there's a bit of php code to ensure the data is current (the 
> little check at the top), but otherwise its a static page (it 
> used to create all the images on page load, but due to bad 
> programming and weak hardware, this was pretty hard on the 
> server; the static content is much nicer, and I now have no 
> qualms about sending out the URL here...I wouldn't have done 
> that before!).

Ok, that sounds like a good idea.

Anyway thanks for the tips, I really need to make some fast progress as
spare time is at a premium and the air temperature is getting chilly...

Alex 


----------------------------------------------------------------------------
--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance. See why
Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Owfs-developers mailing list Owfs-developers@...
https://lists.sourceforge.net/lists/listinfo/owfs-developers
No virus found in this incoming message.
Checked by AVG - http://www.avg.com 
Version: 9.0.791 / Virus Database: 271.1.1/2772 - Release Date: 03/27/10
06:33:00

Ok, I'll figure this out. The problem may be in some refactoring of
ow.i (the swig template) that occurred in that time frame. Thanks for
the report and your work pinpointing the problem.

Paul Alfille

On Wed, Mar 24, 2010 at 4:38 PM, Jacob Joseph <jacob@...> wrote:
> Hi.
>
> I'm unsure of whether this was discussed in the recent mentions of
> pyowfs, which I'm not currently using.  If it has been, I don't see
> where.
>
> I've recently tried to upgrade from owfs 2.7_p21 to something more
> recent, especially to test my BAE device.  I have a few python programs
> that use the 'ow' module, and have found these to not work with newer
> versions.  Here's an example:
>
> import ow
> ow.init('192.168.99.16:4304')
> ow._get("28.65CA0A010000/temperature")
>
> I expect, e.g.,:
> Out[8]: '       10.75'
>
> This has always worked, and I've double-checked versions p{21, 26, 28,
> 29}.
>
> In versions 2.7_p{32, 31, 30}, very often, I receive the exception:
> exUnknownSensor: '28.65CA0A010000/temperature'
>
> At all other times, I receive a single whitespace character, and never
> a temperature reading:
> Out[4]: ' '
>
> Accessing the 1-wire bus through the mounted FUSE filesystem always
> works.
>
> --------
>
> As a separate matter, with the mounted filesystem, I see some
> inconsistency.  In particular, the first (and maybe after some
> timeout) time I do a directory listing after restarting owserver,
> an ls returns instantly, but contains only "bus.0 settings  statistics
> structure system uncached".  If I type ls again, it pauses briefly,
> but shows all of my devices.
>
> Thanks.
> ~Jacob
>
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>
>

Great work, I'll try to address the issues you brought up:

1. I added your writeup to the website. http://owfs.org/index.php?page=macintosh
2. Great news about FUSE. Made the change.
3. Can you give more information about the owhttpd problem? There are
no unusual dependencies as far as I know -- should be straight C code.
Perhaps some issue with thread or signal handling?
4. I'll have to investigate the Bonjour issue. It's always available
in OSX so I probabbly am not invoking it in the correct manner for
that platform.
5. Adding <sys/types.h> before the 2 instances of <usb.h> (in
module/owlib/src/c/ow_ds9490.c) causes no problems with linux
function. I don't know if it helps with Mac.
6. Concerning _DARWIN_C_SOURCE I found this example
http://code.google.com/p/libnfc/source/browse/trunk/configure.ac?spec=svn315&r=315
and tried it in OWFS.and it doesn't interfere with Linux functioning.

These changes are in the CVS repository. Once I figure out the python
problem I'd apprecite it if you could test and give me feedback.

Thanks

Paul Alfille



On Wed, Mar 24, 2010 at 12:18 PM, Peter Radcliffe <pir@...> wrote:
> I was recently introduced to 1-wire sensors after setting up power
> monitoring from a currentcost unit via a usb interface cable to my mac
> mini running osx 10.5. I want to use the same machine for 1-wire since
> it's the only machine I have at home that is on all the time.
>
> Since I want to use switches (mostly as on/off sensors for a couple of
> things, like my heating system) I understood I couldn't use digitemp
> (which does now build on osx by default) so owfs it is.
>
> After... some work... I got owfs to build and even work with macfuse
> from macports (fuse isn't, as the owfs web pages claim, linux specific
> any more). The python examples (I want to continue to use python since
> that's what I wrote my currentcost stuff in) didn't work however.
>
> Luckily for me some searching found this message posted before I
> joined the list:
>  http://article.gmane.org/gmane.comp.file-systems.owfs.devel/6544
> so I rolled back to p29 and python access works, owfs works, owhttpd
> doesn't build but I don't intend to use it anyway.
>
>
> In building I had to turn off zeroconf (it caused link errors that I
> didn't track down since I don't want to use it anyway), install
> macfuse, python25 (I didn't test with the system standard python),
> gsed and libusb-compat from macports and fix some linuxisms and
> libtool's config was broken (it had the mac specific config variables
> in the file... immediately followed by linux x86 config variables
> which overrode them).
>
> What fixed most source issues was:
>  #define _DARWIN_C_SOURCE 1
> which should probably go into configure and be #ifdefed in config.h.
>
> For libusb all uses of usb.h should include <sys/types.h> first...
> I found it easier to just add <sys/types.h> to /opt/local/include/usb.h
> because i was being lazy.
>
> I haven't found proper fixes for a lot of these problems, thus far
> I've just automated getting it to the point of building and working
> and thought others may find it useful if something similar hasn't been
> posted already (I couldn't find a guide to getting recent owfs working
> on OSX). My fix/build script is here:
>
>  http://p.pir.net/hacks/fix-owfs.sh
>
> P.
>
> --
> pir
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>

Paul Alfille <paul.alfille@...> probably said:
>Great work, I'll try to address the issues you brought up:

Thanks.

>3. Can you give more information about the owhttpd problem? There are
>no unusual dependencies as far as I know -- should be straight C code.
>Perhaps some issue with thread or signal handling?

With p29:

libtool: link: gcc -I../include -I../../../owlib/src/include -fexceptions
 -Wall -W -Wundef -Wshadow -Wpointer-arith -Wcast-qual -Wcast-align
 -Wstrict-prototypes -Wredundant-decls -D_FILE_OFFSET_BITS=64
 -D_BSD_SOURCE=1 -D_ISOC99_SOURCE=1 -D_POSIX_C_SOURCE=200112L -D_THREAD_SAFE
 -I/opt/local/include -I/usr/local/include -DNO_NESTED_FUNCTIONS
 -D_BSD_SOURCE=1 -D_ISOC99_SOURCE=1 -D_POSIX_C_SOURCE=200112L
 -o .libs/owhttpd owhttpd.o owhttpd_handler.o owhttpd_present.o
 owhttpd_write.o owhttpd_read.o owhttpd_dir.o owhttpd_favicon.o
 -Wl,-framework -Wl,IOKit -Wl,-framework -Wl,CoreFoundation 
 -L/Users/pir/owfs-2.7p29/module/owlib/src/c -L/opt/local/lib
 -L/usr/local/lib /Users/pir/owfs-2.7p29/module/owlib/src/c/.libs/libow.so
 /opt/local/lib/libusb.dylib /opt/local/lib/libusb-1.0.dylib -pthread
Undefined symbols:
  "_getline", referenced from:
      _handle_POST in owhttpd_handler.o
      _ReadToCRLF in owhttpd_handler.o
      _GetPostPath in owhttpd_handler.o
      _GetPostData in owhttpd_handler.o
ld: symbol(s) not found
collect2: ld returned 1 exit status
gmake[4]: *** [owhttpd] Error 1
gmake[4]: Leaving directory `/Users/pir/owfs-2.7p29/module/owhttpd/src/c'
gmake[3]: *** [all-recursive] Error 1
gmake[3]: Leaving directory `/Users/pir/owfs-2.7p29/module/owhttpd/src'
gmake[2]: *** [all-recursive] Error 1
gmake[2]: Leaving directory `/Users/pir/owfs-2.7p29/module/owhttpd'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory `/Users/pir/owfs-2.7p29/module'
gmake: *** [all-recursive] Error 1


It does link and build with p32, so I assumed this case had already
been fixed.

libtool: link: gcc -I../include -I../../../owlib/src/include -fexceptions
 -Wall -W -Wundef -Wshadow -Wpointer-arith -Wcast-qual -Wcast-align
 -Wstrict-prototypes -Wredundant-decls -D_FILE_OFFSET_BITS=64
 -D_BSD_SOURCE=1 -D_ISOC99_SOURCE=1 -D_POSIX_C_SOURCE=200112L -D_THREAD_SAFE
 -I/opt/local/include -I/usr/local/include -DNO_NESTED_FUNCTIONS
 -D_BSD_SOURCE=1 -D_ISOC99_SOURCE=1 -D_POSIX_C_SOURCE=200112L
 -o .libs/owhttpd owhttpd.o owhttpd_handler.o owhttpd_present.o
 owhttpd_write.o owhttpd_read.o owhttpd_dir.o owhttpd_favicon.o
 -Wl,-framework -Wl,IOKit -Wl,-framework -Wl,CoreFoundation 
 -L/Users/pir/owfs-2.7p32/module/owlib/src/c -L/opt/local/lib
 -L/usr/local/lib /Users/pir/owfs-2.7p32/module/owlib/src/c/.libs/libow.so
 /opt/local/lib/libusb.dylib /opt/local/lib/libusb-1.0.dylib -pthread
gmake[4]: Leaving directory `/Users/pir/owfs-2.7p32/module/owhttpd/src/c'

>4. I'll have to investigate the Bonjour issue. It's always available
>in OSX so I probabbly am not invoking it in the correct manner for
>that platform.

When trying to link owlib (the link command is several screens long so
I won't paste it in unless you particularly want to see what it ends
up being on OSX):

ld: duplicate symbol _avahi_client_new in .libs/ow_alias.o and .libs/ow_dnssd.o
collect2: ld returned 1 exit status
gmake[4]: *** [libow.la] Error 1
gmake[4]: Leaving directory `/Users/pir/owfs-2.7p32/module/owlib/src/c'

I poked at it a little, enough to work out what the avahi client was
and what, roughly, was being declared... then set
configure --disable-zero and moved on.

>These changes are in the CVS repository. Once I figure out the python
>problem I'd apprecite it if you could test and give me feedback.

Great, happy to test things. I'll take a look once you've figured that out.

Thanks!
P.

-- 
pir

Ahh yes. getline is a GNU extension and not available in all
libraries. I put in a compatible implementation for other platforms
that seems to work well.

Paul

On Thu, Mar 25, 2010 at 4:23 AM, Peter Radcliffe <pir@...> wrote:
> Paul Alfille <paul.alfille@...> probably said:
>>Great work, I'll try to address the issues you brought up:
>
> Thanks.
>
>>3. Can you give more information about the owhttpd problem? There are
>>no unusual dependencies as far as I know -- should be straight C code.
>>Perhaps some issue with thread or signal handling?
>
> With p29:
>
> libtool: link: gcc -I../include -I../../../owlib/src/include -fexceptions
>  -Wall -W -Wundef -Wshadow -Wpointer-arith -Wcast-qual -Wcast-align
>  -Wstrict-prototypes -Wredundant-decls -D_FILE_OFFSET_BITS=64
>  -D_BSD_SOURCE=1 -D_ISOC99_SOURCE=1 -D_POSIX_C_SOURCE=200112L -D_THREAD_SAFE
>  -I/opt/local/include -I/usr/local/include -DNO_NESTED_FUNCTIONS
>  -D_BSD_SOURCE=1 -D_ISOC99_SOURCE=1 -D_POSIX_C_SOURCE=200112L
>  -o .libs/owhttpd owhttpd.o owhttpd_handler.o owhttpd_present.o
>  owhttpd_write.o owhttpd_read.o owhttpd_dir.o owhttpd_favicon.o
>  -Wl,-framework -Wl,IOKit -Wl,-framework -Wl,CoreFoundation
>  -L/Users/pir/owfs-2.7p29/module/owlib/src/c -L/opt/local/lib
>  -L/usr/local/lib /Users/pir/owfs-2.7p29/module/owlib/src/c/.libs/libow.so
>  /opt/local/lib/libusb.dylib /opt/local/lib/libusb-1.0.dylib -pthread
> Undefined symbols:
>  "_getline", referenced from:

So as I understand OWFS virtually mounts a filesystem from the onewire
device, after it I can access to its properties, like files


> Well, the easiest way to explore OWFS would be to use the web interface.
>
> # start the web server on port 4444
> # "-u" assumes you are using the USB bus master
> # "-d /dev/ttyS0" for a serial adapter, etc
> sudo /opt/owfs/bin/owhttpd -u -p 4444
>
> then point your web browser to http://localhost:4444
>
> ----------
>
> On the other hand, to show the usage in an email message, we'll use
> the filesystem interface (uses FUSE)
>
> # make a plce to mount the virtual filesystem
> mkdir 1wire
>
> # use simulated devices (fake) in this case
> # --allow_other lets a regular user access the directory
> sudo /opt/owfs/bin/owfs --fake=10,10 -m 1wire --allow_other
>
> # see the directory with 2 DS18S20 devices (10.xxx)
> ls 1wire/
> 10.4AEC29CDBAAB  alarm  settings      statistics  system
> 10.67C6697351FF  bus.0  simultaneous  structure   uncached
>
> # show their available properties
> ls 1wire/10.*
> 1wire/10.4AEC29CDBAAB:
> address  crc8    family  locator  present    r_id       temperature
> templow
> alias    errata  id      power    r_address  r_locator  temphigh     type
>
> 1wire/10.67C6697351FF:
> address  crc8    family  locator  present    r_id       temperature
> templow
> alias    errata  id      power    r_address  r_locator  temphigh     type
>
> # read the temperatures using a grep trick
> grep -r "." 1wire/10.*/temperature
> 1wire/10.4AEC29CDBAAB/temperature:     14.3879
> 1wire/10.67C6697351FF/temperature:     52.1366
>
> Paul Alfille
>
> On Tue, Mar 23, 2010 at 2:54 AM, kako <kako@...> wrote:
>> Hi,
>>
>> I've just found the owfs tools. Downloaded, compiled successfully but I
>> don't know, how to use it.
>> I simply want to read out the temperature data out of two different
>> addressed DS1825 (0 and 1 :-) ) periodically. So my question is "how to
>> start?". I've read the manuals from the source tree, but did'n find the
>> holy grail:-)
>>
>> Thanks
>> Istvan Kassai
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Owfs-developers mailing list
> Owfs-developers@...
> https://lists.sourceforge.net/lists/listinfo/owfs-developers
>

On Tue, Mar 23, 2010 at 04:06:40PM +0100, kako@... wrote:
> So as I understand OWFS virtually mounts a filesystem from the onewire
> device, after it I can access to its properties, like files

Mostly :)

owfs actually creates a virtual filesystem to expose properties it can get
from the 1-wire devices and then lets you mount it.
One big thing is that it caches the 1-wire calls for a certain number of
seconds (like 15 for temperatures) so that you can read it more quickly and
repeatetly without hanging.

Marc
-- 
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems & security ....
                                      .... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/