avr_interrupts(3avr)

NAME¶

avr_interrupts - <avr/interrupt.h>: Interrupts

SYNOPSIS¶

Global manipulation of the interrupt flag¶

The global interrupt flag is maintained in the I bit of the status register (SREG).

Handling interrupts frequently requires attention regarding atomic access to objects that could be altered by code running within an interrupt context, see <util/atomic.h>.

Frequently, interrupts are being disabled for periods of time in order to perform certain operations without being disturbed; see Problems with reordering code for things to be taken into account with respect to compiler optimizations.
#define sei() __asm__ __volatile__ ('sei' ::: 'memory')
#define cli() __asm__ __volatile__ ('cli' ::: 'memory')

Macros for writing interrupt handler functions¶

#define ISR(vector, attributes)
#define SIGNAL(vector)
#define EMPTY_INTERRUPT(vector)
#define ISR_ALIAS(vector, target_vector)
#define reti() __asm__ __volatile__ ('reti' ::: 'memory')
#define BADISR_vect

ISR attributes¶

#define ISR_BLOCK
#define ISR_NOBLOCK
#define ISR_NAKED
#define ISR_FLATTEN
#define ISR_NOICF
#define ISR_NOGCCISR
#define ISR_ALIASOF(target_vector)

Detailed Description¶

Note

This discussion of interrupts was originally taken from Rich Neswold's document. See Acknowledgments.

Introduction to AVR-LibC's interrupt handling¶

It's nearly impossible to find compilers that agree on how to handle interrupt code. Since the C language tries to stay away from machine dependent details, each compiler writer is forced to design their method of support.

In the AVR-GCC environment, the vector table is predefined to point to interrupt routines with predetermined names. By using the appropriate name, your routine will be called when the corresponding interrupt occurs. The device library provides a set of default interrupt routines, which will get used if you don't define your own.

Patching into the vector table is only one part of the problem. The compiler uses, by convention, a set of registers when it's normally executing compiler-generated code. It's important that these registers, as well as the status register, get saved and restored.

These details seem to make interrupt routines a little messy, but all these details are handled by the Interrupt API. An interrupt routine is defined with ISR(). This macro register and mark the routine as an interrupt handler for the specified peripheral. The following is an example definition of a handler for the ADC interrupt.

#include <avr/interrupt.h>
ISR(ADC_vect)
{


    // user code here
}

Refer to the chapter explaining assembler programming for an explanation about interrupt routines written solely in assembly.

Catch-all interrupt vector¶

If an unexpected interrupt occurs (interrupt is enabled but no handler is installed, which usually indicates a bug), then the default action is to reset the device by jumping to the reset vector. You can override this by supplying a function named BADISR_vect which should be defined with ISR() as such. The name BADISR_vect is actually an alias for __vector_default. The latter must be used inside assembly code in case <avr/interrupt.h> is not included.

#include <avr/interrupt.h>
ISR(BADISR_vect)
{


    // user code here
}

Nested interrupts¶

The AVR hardware clears the global interrupt flag in SREG when an interrupt request is serviced. Thus, normally interrupts will remain disabled inside the handler until the handler exits, where the RETI instruction (that is emitted by the compiler as part of the normal function epilogue for an ISR) will eventually re-enable further interrupts. For that reason, interrupt handlers normally do not nest. For most interrupt handlers, this is the desired behaviour, for some it is even required in order to prevent infinitely recursive interrupts (like UART interrupts, or level-triggered external interrupts). In rare circumstances though it might be desired to re-enable the global interrupt flag as early as possible in the interrupt handler, in order to not defer any other interrupt more than absolutely needed. This could be done using an sei() instruction right at the beginning of the interrupt handler, but this still leaves few instructions inside the compiler-generated function prologue to run with global interrupts disabled. The compiler can be instructed to insert a SEI instruction right at the beginning of an interrupt handler by declaring the handler the following way:

ISR(XXX_vect, ISR_NOBLOCK)
{


  ...
}

where XXX_vect is the name of a valid interrupt vector for the MCU type in question, as explained below.

Two vectors sharing the same code¶

In some circumstances, the actions to be taken upon two different interrupts might be completely identical so a single implementation for the ISR would suffice. For example, pin-change interrupts arriving from two different ports could logically signal an event that is independent from the actual port (and thus interrupt vector) where it happened. Sharing interrupt vector code can be accomplished using the ISR_ALIASOF() attribute to the ISR macro:

ISR(PCINT0_vect)
{


  ...


  // Code to handle the event.
}
ISR(PCINT1_vect, ISR_ALIASOF(PCINT0_vect));

Note

There is no body to the aliased ISR.

Note that the ISR_ALIASOF() feature requires GCC 4.2 or above (or a patched version of GCC 4.1.x). See the documentation of the ISR_ALIAS() macro for an implementation which is less elegant but could be applied to all compiler versions.

Empty interrupt service routines¶

In rare circumstances, in interrupt vector does not need any code to be implemented at all. The vector must be declared anyway, so when the interrupt triggers it won't execute the BADISR_vect code (which by default restarts the application).

This could for example be the case for interrupts that are solely enabled for the purpose of getting the controller out of sleep_mode().

A handler for such an interrupt vector can be declared using the EMPTY_INTERRUPT() macro:

EMPTY_INTERRUPT(ADC_vect);

Note

There is no body to this macro.

Manually defined ISRs¶

In some circumstances, the compiler-generated prologue and epilogue of the ISR might not be optimal for the job, and a manually defined ISR could be considered particularly to speedup the interrupt handling.

One solution to this could be to implement the entire ISR as manual assembly code in a separate (assembly) file. See Combining C and assembly source files for an example of how to implement it that way.

Another solution is to still implement the ISR in C language but take over the compiler's job of generating the prologue and epilogue. This can be done using the ISR_NAKED attribute to the ISR() macro. Note that the compiler does not generate anything as prologue or epilogue, so the final reti() must be provided by the actual implementation. SREG must be manually saved if the ISR code modifies it, and the compiler-implied assumption of __zero_reg__ always being 0 could be wrong (e. g. when interrupting right after of a MUL instruction).

Warning

According to the GCC documentation, only inline assembly is supported in naked functions, like with ISR_NAKED.

ISR(TIMER1_OVF_vect,  ISR_NAKED)
{


  PORTB |= _BV(0);  // results in SBI which does not affect SREG


  reti();
}

Choosing the vector: Interrupt vector names¶

The interrupt is chosen by supplying one of the vector names in the following table.

There are currently two different styles present for naming the vectors.

Starting with AVR-LibC v1.4, the style of interrupt vector names is a short phrase for the vector description followed by _vect. The short phrase matches the vector name as described in the datasheet of the respective device (and in the hardware manufacturer's XML/ATDF files), with spaces replaced by an underscore and other non-alphanumeric characters dropped. Using the suffix _vect is intented to improve portability to other C compilers available for the AVR that use a similar naming convention.
A deprecated form that uses names starting with SIG_, followed by a relatively verbose but arbitrarily chosen name describing the interrupt vector. This has been the only available style in AVR-LibC up to version 1.2.x. This historical naming style is not recommended for new projects, and some headers require that the macro __AVR_LIBC_DEPRECATED_ENABLE__ is defined so that the SIG_ names ISR names are available.

Note

The ISR() macro cannot really spell-check the argument passed to them. Thus, by misspelling one of the names below used in ISR(), a function will be created that, while possibly being usable as an interrupt function, is not actually wired into the interrupt vector table. The compiler will generate a warning if it detects a suspiciously looking name of an ISR() function (i.e. one that after macro replacement does not start with '__vector_').

Apart from the NAME_vect macros listed below, for each such ISR name there is also a macro NAME_vect_num defined which resolves to the IRQ number. This is the index into the vector table, where indices start at index 0 (the reset vector).

See also What ISR names are available for my device? in the FAQ for how find all the IRQ names for a specific device.

Vector Name Description Applicable for Device

Note

For the following devices, only the deprecated SIG_ names are available: AT43USB320, AT43USB355, AT76C711, AT90C8534, AT94K, M3000.

Macro Definition Documentation¶

#define BADISR_vect¶

#include <avr/interrupt.h>

This is a vector which is aliased to __vector_default, the vector executed when an IRQ fires with no accompanying ISR handler. This may be used along with the ISR() macro to create a catch-all for undefined but used ISRs for debugging purposes.

#define cli() asm volatile ('cli' ::: 'memory')¶

Disables all interrupts by clearing the global interrupt mask. This function actually compiles into a single line of assembly, so there is no function call overhead. However, the macro also implies a memory barrier which can cause additional loss of optimization.

In order to implement atomic access to multi-byte objects, consider using the macros from <util/atomic.h>, rather than implementing them manually with cli() and sei().

#define EMPTY_INTERRUPT(vector)¶

Defines an empty interrupt handler function. This will not generate any prolog or epilog code and will only return from the ISR. Do not define a function body as this will define it for you. Example:

EMPTY_INTERRUPT(ADC_vect);

#define ISR(vector, attributes)¶

Introduces an interrupt handler function (interrupt service routine) that runs with global interrupts initially disabled by default with no attributes specified.

The attributes are optional and alter the behaviour and resultant generated code of the interrupt routine. Multiple attributes may be used for a single function, with a space seperating each attribute.

Valid attributes are ISR_BLOCK, ISR_NOBLOCK, ISR_NAKED, ISR_FLATTEN, ISR_NOICF, ISR_NOGCCISR and ISR_ALIASOF(vect).

vector must be one of the interrupt vector names that are valid for the particular MCU type.

#define ISR_ALIAS(vector, target_vector)¶

Aliases a given vector to another one in the same manner as the ISR_ALIASOF attribute for the ISR() macro. Unlike the ISR_ALIASOF attribute macro however, this is compatible for all versions of GCC rather than just GCC version 4.2 onwards.

Note

This macro creates a trampoline function for the aliased macro. This will result in a two cycle penalty for the aliased vector compared to the ISR the vector is aliased to, due to the JMP/RJMP opcode used.

Deprecated

For new code, the use of ISR(..., ISR_ALIASOF(...)) is recommended.

Example:

ISR(INT0_vect)
{


    PORTB = 42;
}
ISR_ALIAS(INT1_vect, INT0_vect);

#define ISR_ALIASOF(target_vector)¶

The ISR is linked to another ISR, specified by the vect parameter. This is compatible with GCC 4.2 and greater only.

Use this attribute in the attributes parameter of the ISR macro. Example:

ISR (INT0_vect)
{


    PORTB = 42;
}
ISR (INT1_vect, ISR_ALIASOF (INT0_vect));

#define ISR_BLOCK¶

Identical to an ISR with no attributes specified. Global interrupts are initially disabled by the AVR hardware when entering the ISR, without the compiler modifying this state.