CCS C Software and Maintenance Offers
FAQFAQ   FAQForum Help   FAQOfficial CCS Support   SearchSearch  RegisterRegister 

ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

CCS does not monitor this forum on a regular basis.

Please do not post bug reports on this forum. Send them to support@ccsinfo.com

Schrodinger's Variable - SOLVED

 
Post new topic   Reply to topic    CCS Forum Index -> General CCS C Discussion
View previous topic :: View next topic  
Author Message
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

Schrodinger's Variable - SOLVED
PostPosted: Sat May 12, 2018 12:26 pm     Reply with quote

I have the following code snippets:

Code:
      
fprintf(lcd_putc,"_____________________________DEBUG - WIFI RESET FLAG:%u\r\n",WIFI_RESET_FLAG);
      Internet_Manager();            // Manages the connections and upload method


Where i print the variable "WIFI_RESET_FLAG"... Assume there has been a reset recently so it is 1.

Just as i finish printing it i call "Internet_Manager()"

Then:

Code:
void Internet_Manager()
{
   fprintf(lcd_putc,"_____________________________DEBUG - WIFI RESET FLAG:%u\r\n",WIFI_RESET_FLAG);

      if(GPRS_ONLY==FALSE)            // Default state is WIFI ON
      {   
               
         if(!WIFI_Upload_Data())         // Try Uploading via WIFI
         {


As soon as i enter "Internet_manager()" i print the Flag again... and right before the call it was 1 but inside the function its 0.

Assume all is good and i get to call "WIFI_Upload_Data() on the IF statement.

Then:

Code:
int1 WIFI_Upload_Data()
{
   fprintf(lcd_putc,"WTF1_____________________________DEBUG - WIFI RESET FLAG:%u\r\n",WIFI_RESET_FLAG);
   if(WIFI_RESET_FLAG==FALSE)
   {
   fprintf(lcd_putc,"WTF2_____________________________DEBUG - WIFI RESET FLAG:%u\r\n",WIFI_RESET_FLAG);



This IF checks out as TRUE!?!?!?!?!

BUT it gets wierder... as soon as i remove the DEBUG printf Before calling Internet_Manager() the IF fails as WIFI_RESET_FLAG shows FALSE!?!?!?!


So if i print to debug the variable my code works, if i dont print it the code fails.... WTFFFFF?!?!


WIFI_RESET_FLAG is a GLOBAL on the Wifi Dirver....
Internet_Manager() is part of the main application code.
WIFI_Upload_Data() is part of the Wifi Driver.


I appreciate your help.

G
_________________
CCS PCM 5.078 & CCS PCH 5.093


Last edited by Gabriel on Mon May 14, 2018 5:48 pm; edited 1 time in total
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Sat May 12, 2018 12:47 pm     Reply with quote

I have not updated my compiler yet as per my last post.. FYI.

This is pretty much the same code ive been using for the last 3 years.
This application has some diferences but none that change any of the WIFI functionality.

so im pretty much stuck at this moment.
_________________
CCS PCM 5.078 & CCS PCH 5.093
PCM programmer



Joined: 06 Sep 2003
Posts: 21708

View user's profile Send private message

PostPosted: Sat May 12, 2018 4:29 pm     Reply with quote

First, tell us your PIC and compiler version.

Then look at the .SYM file.

1. See if your flag has the same the address as any other variables.

2. Also, look at the variables that occur just before the address of your flag.
What are they ? Are they possibly a byte value that your code is treating
as a word, and thus your flag is over-written ? Or is it an array, and your
code is accidentally writing beyond the end of the array and over-writing
your flag.


Look at the .LST file.

1. What are the stack levels ? Do the levels used (worst case) exceed the
levels available in the PIC ?

2. What about RAM usage ? Is it near the limits ?

3. Look at the fuse list at the end of the .LST file. Make sure it has
NOXINST (if you're using Pic18F).
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Sat May 12, 2018 7:07 pm     Reply with quote

PIC18F87J50
PCH 5.013

69% RAM worst case

The variable in question is a "int1" global.
Its only written to in 2 places, but read across multiple parts of the code.

Will revert back after i look at your suggestions.

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093
PCM programmer



Joined: 06 Sep 2003
Posts: 21708

View user's profile Send private message

PostPosted: Sat May 12, 2018 7:30 pm     Reply with quote

According to Ttelmah, he had a problem with variable
corruption in vs. 5.013 and similar versions:
http://www.ccsinfo.com/forum/viewtopic.php?t=51455

In this thread, he's using vs. 4.013, but has a similar problem:
http://www.ccsinfo.com/forum/viewtopic.php?t=28678
Ttelmah



Joined: 11 Mar 2010
Posts: 19195

View user's profile Send private message

PostPosted: Sun May 13, 2018 12:19 am     Reply with quote

The fix I found was to declare any variables like this to be 'static'.
Examining the .sym was the way to find this.
I must admit I have not seen this on 'recent' versions of the compiler, so I suspect it has been fixed now. However the version being referred to by the original poster is around where I found this.
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Sun May 13, 2018 10:35 am     Reply with quote

So, As Per PCM's Suggestions ive been looking at the SYM file.

found the following:
Quote:
200.0 Button_Pressed
200.1 Long_Button_Press
200.2 doneFlag
200.3 WIFI_RESET_FLAG
200.4 ISR_MODE
200.5 Long_Message
200.6 RX_DONE
200.7 Backlight_Off_Flag


Looking at the variable before "WIFI_RESET_FLAG" i found the following on the DS18B20 Driver that I got from this forum:

Code:
doneFlag = (lastDiscrep == 0);


Which seems like an extra '=' . or maybe its a use of the '==' operator that i was not aware of or forgot.


Will compile, test and revert to you again.

Thank you all for your help so far.
G
_________________
CCS PCM 5.078 & CCS PCH 5.093
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Sun May 13, 2018 10:42 am     Reply with quote

Well that didnt work jajajaj the DS18B20 driver failed to recognize any devices on the bus.

Better check that C book for that '==' use.
_________________
CCS PCM 5.078 & CCS PCH 5.093
PCM programmer



Joined: 06 Sep 2003
Posts: 21708

View user's profile Send private message

PostPosted: Sun May 13, 2018 10:47 am     Reply with quote

Gabriel, what do you mean by "revert to you" ?

To me, revert means
1. Go back to the previous compiler version.
2. Go back to the previous code.
Ttelmah



Joined: 11 Mar 2010
Posts: 19195

View user's profile Send private message

PostPosted: Sun May 13, 2018 10:56 am     Reply with quote

Code:

doneFlag = (lastDiscrep == 0);


Is perfectly reasonable.

It makes 'doneFlag' into the logical result of testing if lastDiscrep is equal to zero. This is what '==' does in C.

The thing to do on the sym, is see if any other variables are mapped at address 200.

The other thing to look at carefully is if you make any use of pointers. It'd be easy for you to accidentally write to the wrong address with these.
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Sun May 13, 2018 11:08 am     Reply with quote

Revert to you = Get back to you.

So i have 2 functions on my code that have the EXACT same code, but the function call thats inside after a couple of ifs is what changes.
one is for uploading and the other for downloading data.

Uploading fails, downloading reads the WIFI_RESET_FLAG correctly.

In any case, looking at disassembly listing file:

DOWNLOADING (WORKS):
Code:
33:                int1 WIFI_Download_Data()
34:                {
35:                   if(WIFI_RESET_FLAG==FALSE)
 11CE4    0102     MOVLB 0x2
 11CE6    B700     BTFSC 0, 0x3, BANKED
 11CE8    D016     BRA 0x11d16
36:                   {
37:                      if(Connect_to_Cloud())
 11CEA    0100     MOVLB 0
 11CEC    EC15     CALL 0xf82a, 0
 11CF0    5201     MOVF 0x1, F, ACCESS
 11CF2    E00C     BZ 0x11d0c
38:                      {
39:                         fprintf(lcd_putc,"WIFI - Initiated Data Download\r\n");



UPLOADING (FAILS);
Code:
17:                int1 WIFI_Upload_Data()
18:                {
19:                   if(WIFI_RESET_FLAG==FALSE)
 0FDC8    B400     BTFSC 0, 0x2, ACCESS
 0FDCA    D013     BRA 0xfdf2
20:                   {
21:                      if(Connect_to_Cloud())
 0FDCC    DD2E     RCALL 0xf82a
 0FDCE    5201     MOVF 0x1, F, ACCESS
 0FDD0    E00C     BZ 0xfdea
22:                      {
23:                         fprintf(lcd_putc,"WIFI - Initiated Data Upload\r\n");



There are some clear differences on what each code is doing.
I have to admit, I dont remember my assembly well enough, so im lost.

I kinda get out of this that one code is actually accessing the flag, while the other takes a "Banked" version of it??? which i dont know how to resolve.

G
_________________
CCS PCM 5.078 & CCS PCH 5.093
Ttelmah



Joined: 11 Mar 2010
Posts: 19195

View user's profile Send private message

PostPosted: Sun May 13, 2018 12:07 pm     Reply with quote

The second routine is assuming that the bank register is already set to '2'. You'd have to check where it was called from to verify this is the case. However big difference is it is checking bit 2, rather than bit 3. The symbol file shows this as being doneFlag, not WIFI_RESET_FLAG.
Was it working before?.
What has triggered it to go wrong?.
5.012, is an early enough V5 compiler that it did have some significant issues.
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Mon May 14, 2018 2:59 pm     Reply with quote

So... Following up on this

The was one variable(bit) on that 200 address that was not being used - as per a warning on the compile output file.

I removed the variable and 1 or 2 others that were unused to see if that helped.
(a warning i usually dismiss as i run several versions of the same code for different devices which may or may not use a variable depending on setup...)

obviously the SYM file changed and now i have the following:

Code:
17:                int1 WIFI_Upload_Data()
18:                {
19:                   if(WIFI_RESET_FLAG==FALSE)
 0FDDC    0101     MOVLB 0x1
 0FDDE    B5FF     BTFSC 0xff, 0x2, BANKED
 0FDE0    D015     BRA 0xfe0c
20:                   {
21:                      if(Connect_to_Cloud())
 0FDE2    0100     MOVLB 0
 0FDE4    DD2C     RCALL 0xf83e
 0FDE6    5201     MOVF 0x1, F, ACCESS
 0FDE8    E00C     BZ 0xfe02
22:                      {



and

Code:
33:                int1 WIFI_Download_Data()
34:                {
35:                   if(WIFI_RESET_FLAG==FALSE)
 11D0E    0101     MOVLB 0x1
 11D10    B5FF     BTFSC 0xff, 0x2, BANKED
 11D12    D016     BRA 0x11d40
36:                   {
37:                      if(Connect_to_Cloud())
 11D14    0100     MOVLB 0
 11D16    EC1F     CALL 0xf83e, 0
 11D1A    5201     MOVF 0x1, F, ACCESS
 11D1C    E00C     BZ 0x11d36
38:                      {


Code:
1FF.0   Button_Pressed
1FF.1   doneFlag
1FF.2   WIFI_RESET_FLAG
1FF.3   ISR_MODE
1FF.4   Long_Message
1FF.5   RX_DONE
1FF.6   Backlight_Off_Flag
1FF.7   Backlight_State



The major change was removing that "Long_Button_Pressed" variable
I'm not in a position to test the code, but from the LST file i can see that now both functions indeed access the right bit.

Now that I've "fixed" this leak, i hope another hole does'nt open up elsewhere.

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093
Gabriel



Joined: 03 Aug 2009
Posts: 1067
Location: Panama

View user's profile Send private message

PostPosted: Mon May 14, 2018 5:48 pm     Reply with quote

And that's all folks...

I guess unused 1bit variables mess things up when packed in to a byte by the compiler?

... anyways... compiled, tested on hardware, solved.

Thanks for your guidance on this, I would have never figured it out if you all had not pointed me in the right direction.

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093
temtronic



Joined: 01 Jul 2010
Posts: 9081
Location: Greensville,Ontario

View user's profile Send private message

PostPosted: Tue May 15, 2018 5:38 am     Reply with quote

hmm... now IF you have any hair left, I'd like to know IF you could just put a line of code say 'variable.unused_bit=variable.unused_bit;' in your code, would it compile properly?
This way the unused bit IS 'used', though as I type this...hmm will the compiler says 'that does nothing,so I won't code it'?
Jay
Display posts from previous:   
Post new topic   Reply to topic    CCS Forum Index -> General CCS C Discussion All times are GMT - 6 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group