Post by thread:[PIC] Quick C question

Im a fool... Sorry to bother. Mauricio {Quote hidden}> -----Original Message----- > From: spam_OUTpiclist-bouncesTakeThisOuTmit.edu > [.....piclist-bouncesKILLspam@spam@mit.edu] On Behalf Of Mauricio Jancic > Sent: Tuesday, March 06, 2007 20:29 > To: 'Microcontroller discussion list - Public.' > Subject: [PIC] Quick C question > > Hi, > I modifying a C code for an ATMEL chip, and I found > this statement: > > if(!(PINB & (1<<PB1))) > > Whats that? Specially the (1<<PB1) that I cant seem to > figure out... > > Regards, > > Mauricio > > -

On 3/6/07, Mauricio Jancic <infoKILLspamjanso.com.ar> wrote: > > Hi, > I modifying a C code for an ATMEL chip, and I found this statement: > > if(!(PINB & (1<<PB1))) > > Whats that? Specially the (1<<PB1) that I cant seem to figure > out... That's a bit shift, a common way to address individual pins. They are masking the read of PINB against a byte with the PB1 bit set. So, the statement is conditional on gettting a zero on the PB1 bit of PINB.

I'm guessing the PB1 has been defined somewhere as a constant - probably 1 since it's PB1), Then 1<<PB1 means "shift 1 left PB1 (i.e. 1) times. [ - The result is 2] So the statement tests if the input of port B, pin 1 is high. (or whichever pin PB1 actually points to). RP On 07/03/07, Mauricio Jancic <EraseMEinfospam_OUTTakeThisOuTjanso.com.ar> wrote: {Quote hidden}> Hi, > I modifying a C code for an ATMEL chip, and I found this statement: > > if(!(PINB & (1<<PB1))) > > Whats that? Specially the (1<<PB1) that I cant seem to figure out... > > Regards, > > Mauricio > > -

On 3/7/07, Mauricio Jancic <infospam_OUTjanso.com.ar> wrote: > if(!(PINB & (1<<PB1))) > > Whats that? Specially the (1<<PB1) that I cant seem to figure out... The statement "x << y" means "take the number x and rotate its bits y times left". If PB1 in your example is 5, then "1<<PB1" would mean "1 << 5", which in binary is "00000001 << 5", which finally gives the result "00010000". Next step (going from the innermost parenthesis and out) is "PINB & (the above)", which is a logical AND-function. AND returns 1 at those bitpositions where both are 1s. "111000 & 000111" is zero, while "111000 & 001000" is 8 (1000 binary). For the sake of the argument, if the value of PINB has a "1" at the same position as the result from above will be the same as (1 << PB1). And finally, the "!" inverts the whole parenthesis. What the if does is checking if the particular bit in PINB is set or not. If it's not set, the code will proceed into the if's then-clause. -- - Rikard.

> Whats that? Specially the (1<<PB1) that I cant seem to figure out. Nobody has been wrong yet, I think. But the way I'd put it: (1<<BITNUM) is a common C construct for converting a bit NUMBER to a bit MASK. It's especially common in microcontroller projects because of the existence of machine instructions that take the bit number directly (and you really don't want to have separate definitions for the mask for a bit and the bit number for a bit.) It works by shifting a single 1 bit ("1") left one position for each bit number (for systems that number their bits from the LSB.) Your confusion demonstrates the usual problem with short symbol names (especially) in global include files. But no one wants to type 1<<PORTB_BITNUM_B0 BillW

> > The thing is PB1 is a constant indicating pin number, and it was shifting > the 1 PB1 times to the left and then ANDing it to PINB. > > Just a missreading > > THANKS!! > Which proves the fact that in order to write maintainable code it is very important to choose a meaningful name for variables and constants. The original author actually did manage half way since he in fact provided a way to easily change both the port and pin for a function in one place - the header file. I would have used macros, though. /Ruben ============================== Ruben Jönsson AB Liros Electronic Box 9124, 200 39 Malmö, Sweden TEL INT +46 40142078 FAX INT +46 40947388 @spam@rubenKILLspampp.sbbs.se ==============================

Though the code author's not really done anything wrong, since PINB and PB1 are both defined in the standard header file provided by Atmel, and are standard constants. Arguably he should have used PINB1 instead of PB1 for better clarity. On 3/7/07, Ruben Jönsson <KILLspamrubenKILLspampp.sbbs.se> wrote: {Quote hidden}> > > > > The thing is PB1 is a constant indicating pin number, and it was > shifting > > the 1 PB1 times to the left and then ANDing it to PINB. > > > > Just a missreading > > > > THANKS!! > > > > Which proves the fact that in order to write maintainable code it is very > important to choose a meaningful name for variables and constants. > > The original author actually did manage half way since he in fact provided > a > way to easily change both the port and pin for a function in one place - > the > header file. > > I would have used macros, though. > > /Ruben > ============================== > Ruben Jönsson > AB Liros Electronic > Box 9124, 200 39 Malmö, Sweden > TEL INT +46 40142078 > FAX INT +46 40947388 > RemoveMErubenTakeThisOuTpp.sbbs.se > ============================== > >

He should have used a describing name for the input and bit, like ALARM_PORT and ALARM_BIT and defined these in a header file to PINB1 and PINB. This way he could have changed the definition for ALARM_PORT and ALARM_BIT at one place only if he wanted to use another input pin. The source code would also directly tell what he wanted to do. I don't know the AVR but I take it that as he wrote it, is a general way to check for a bit in a port using the predefined names for ports and bits. If this is the case I got it wrong from the beginning since he not only didn't provide a way to change the port pin for a specific use in one place but also didn't use an understandable name that easily can be redefined. /Ruben {Quote hidden}> Though the code author's not really done anything wrong, since PINB and PB1 are > both defined in the standard header file provided by Atmel, and are standard > constants. Arguably he should have used PINB1 instead of PB1 for better clarity. > > On 3/7/07, Ruben Jönsson <spamBeGonerubenspamBeGonepp.sbbs.se> wrote: > > > > > > > > The thing is PB1 is a constant indicating pin number, and it was > > shifting > > > the 1 PB1 times to the left and then ANDing it to PINB. > > > > > > Just a missreading > > > > > > THANKS!! > > > > > > > Which proves the fact that in order to write maintainable code it is very > > important to choose a meaningful name for variables and constants. > > > > The original author actually did manage half way since he in fact provided a > > way to easily change both the port and pin for a function in one place - the > > header file. > > > > I would have used macros, though. > > > > /Ruben > > ============================== > > Ruben Jönsson > > AB Liros Electronic > > Box 9124, 200 39 Malmö, Sweden > > TEL INT +46 40142078 > > FAX INT +46 40947388 > > TakeThisOuTrubenEraseMEspam_OUTpp.sbbs.se > > ============================== > > > > --