create, attach and teardown an interrupt
, device_t child
struct resource *irq
, struct resource *r
, device_t child
struct resource *irq
() method will create and
attach an interrupt handler to an interrupt previously allocated by the
resource manager's BUS_ALLOC_RESOURCE(9)
are found in
and give the broad category of interrupt. The
also tell the interrupt handlers about
certain device driver characteristics.
marks the handler as being an
exclusive handler for this interrupt.
tells the scheduler that the
interrupt handler is well behaved in a preemptive environment (``SMP safe''),
and does not need to be protected by the ``Giant Lock'' mutex.
marks the interrupt as being a
good source of entropy - this may be used by the entropy device
To define a time-critical handler that will not execute any potentially blocking
operation, use the filter
argument. See the
below for information on writing a filter. Otherwise, use the
argument. The defined handler will be
called with the value arg
as its only
argument. See the ithread
section below for more information on writing an interrupt
argument is a pointer to a
() will write a cookie for
the parent bus' use to if it is successful in establishing an interrupt.
Driver writers may assume that this cookie will be non-zero. The nexus driver
will write 0 on failure to cookiep
The interrupt handler will be detached by
(). The cookie needs to be
() in order to
tear down the correct interrupt handler. Once
() returns, it is
guaranteed that the interrupt function is not active and will no longer be
Mutexes are not allowed to be held across calls to these functions.
A filter runs in primary interrupt context. In this context, normal mutexes
cannot be used. Only the spin lock version of these can be used (specified by
() when initializing the mutex).
and similar routines can be called.
Atomic operations from machine/atomic
be used. Reads and writes to hardware through
may be used. PCI configuration
registers may be read and written. All other kernel interfaces cannot be used.
In this restricted environment, care must be taken to account for all races. A
careful analysis of races should be done as well. It is generally cheaper to
take an extra interrupt, for example, than to protect variables with
spinlocks. Read, modify, write cycles of hardware registers need to be
carefully analyzed if other threads are accessing the same registers.
Generally, a filter routine will use one of two strategies. The first strategy
is to simply mask the interrupt in hardware and allow the
routine to read the state from the
hardware and then reenable interrupts. The
also acknowledges the interrupt
before re-enabling the interrupt source in hardware. Most PCI hardware can
mask its interrupt source.
The second common approach is to use a filter with multiple
tasks. In this case, the filter
acknowledges the interrupts and queues the work to the appropriate taskqueue.
Where one has to multiplex different kinds of interrupt sources, like a
network card's transmit and receive paths, this can reduce lock contention and
You should not malloc(9)
from inside a filter. You
may not call anything that uses a normal mutex. Witness may complain about
You can do whatever you want in an ithread routine, except sleep. Care must be
taken not to sleep in an ithread. In addition, one should minimize lock
contention in an ithread routine because contested locks ripple over to all
other ithread routines on that interrupt.
Sleeping is voluntarily giving up control of your thread. All the sleep routine
found in msleep(9)
sleep. Waiting for a condition
variable described in condvar(9)
Calling any function that does any of these things is sleeping.
Zero is returned on success, otherwise an appropriate error is returned.
This manual page was written by Jeroen Ruigrok van
⟨asmodai@FreeBSD.org⟩ based on the manual
() written by