|BUS_SETUP_INTR(9)||Kernel Developer's Manual||BUS_SETUP_INTR(9)|
bus_teardown_intr — create,
attach and teardown an interrupt handler
device_t child, struct resource
*irq, int flags, driver_filter_t
*filter, driver_intr_t *ithread,
void *arg, void **cookiep);
struct resource *r, int flags,
driver_filter_t filter, driver_intr_t
ithread, void *arg, void
device_t child, struct resource
*irq, void *cookiep);
dev, struct resource
method will create and attach an interrupt handler to an interrupt
previously allocated by the resource manager's
BUS_ALLOC_RESOURCE(9) method. The
flags are found in
<sys/bus.h>, and give the
broad category of interrupt. The flags also tell the
interrupt handlers about certain device driver characteristics.
INTR_EXCL marks the handler as being an exclusive
handler for this interrupt.
INTR_MPSAFE 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
INTR_ENTROPY 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 Filter Routines section below for information on writing a filter. Otherwise, use the ithread argument. The defined handler will be called with the value arg as its only argument. See the ithread Routines section below for more information on writing an interrupt handler.
The cookiep argument
is a pointer to a void * that
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
The interrupt handler will be detached by
The cookie needs to be passed to
in order to tear down the correct interrupt handler. Once
BUS_TEARDOWN_INTR() returns, it is guaranteed that
the interrupt function is not active and will no longer be called.
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 passing
when initializing the mutex). wakeup(9) and similar
routines can be called. Atomic operations from
machine/atomic may be used. Reads and writes to
hardware through bus_space(9) 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
ithread routine to read the state from the hardware
and then reenable interrupts. The
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 taskqueue(9) 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 increase performance.
You should not malloc(9) from inside a filter. You may not call anything that uses a normal mutex. Witness may complain about these.
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) is sleeping. Calling any function that does any of these things is sleeping.
Zero is returned on success, otherwise an appropriate error is returned.
|November 3, 2010||Debian|