1.\" Copyright (c) 2000 Jeroen Ruigrok van der Werven 2.\" All rights reserved. 3.\" 4.\" Redistribution and use in source and binary forms, with or without 5.\" modification, are permitted provided that the following conditions 6.\" are met: 7.\" 1. Redistributions of source code must retain the above copyright 8.\" notice, this list of conditions and the following disclaimer. 9.\" 2. Redistributions in binary form must reproduce the above copyright 10.\" notice, this list of conditions and the following disclaimer in the 11.\" documentation and/or other materials provided with the distribution. 12.\" 13.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 14.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 15.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 16.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 17.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 18.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 19.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 20.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 21.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 22.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 23.\" SUCH DAMAGE. 24.\" 25.Dd November 3, 2010 26.Dt BUS_SETUP_INTR 9 27.Os 28.Sh NAME 29.Nm BUS_SETUP_INTR , 30.Nm bus_setup_intr , 31.Nm BUS_TEARDOWN_INTR , 32.Nm bus_teardown_intr 33.Nd create, attach and teardown an interrupt handler 34.Sh SYNOPSIS 35.In sys/param.h 36.In sys/bus.h 37.Ft int 38.Fo BUS_SETUP_INTR 39.Fa "device_t dev" "device_t child" "struct resource *irq" "int flags" 40.Fa "driver_filter_t *filter" "driver_intr_t *ithread" "void *arg" 41.Fa "void **cookiep" 42.Fc 43.Ft int 44.Fo bus_setup_intr 45.Fa "device_t dev" "struct resource *r" "int flags" 46.Fa "driver_filter_t filter" "driver_intr_t ithread" "void *arg" 47.Fa "void **cookiep" 48.Fc 49.Ft int 50.Fo BUS_TEARDOWN_INTR 51.Fa "device_t dev" "device_t child" "struct resource *irq" "void *cookiep" 52.Fc 53.Ft int 54.Fn bus_teardown_intr "device_t dev" "struct resource *r" "void *cookiep" 55.Sh DESCRIPTION 56The 57.Fn BUS_SETUP_INTR 58method 59will create and attach an interrupt handler to an interrupt 60previously allocated by the resource manager's 61.Xr BUS_ALLOC_RESOURCE 9 62method. 63The 64.Fa flags 65are found in 66.In sys/bus.h , 67and give the broad category of interrupt. 68The 69.Fa flags 70also tell the interrupt handlers about certain 71device driver characteristics. 72.Dv INTR_EXCL 73marks the handler as being 74an exclusive handler for this interrupt. 75.Dv INTR_MPSAFE 76tells the scheduler that the interrupt handler 77is well behaved in a preemptive environment 78(``SMP safe''), 79and does not need 80to be protected by the ``Giant Lock'' mutex. 81.Dv INTR_ENTROPY 82marks the interrupt as being a good source of entropy - 83this may be used by the entropy device 84.Pa /dev/random . 85.Pp 86To define a time-critical handler that will not execute any potentially 87blocking operation, use the 88.Fa filter 89argument. 90See the 91.Sx "Filter Routines" 92section below for information on writing a filter. 93Otherwise, use the 94.Fa ithread 95argument. 96The defined handler 97will be called with the value 98.Fa arg 99as its only argument. 100See the 101.Sx "ithread Routines" 102section below for more information on writing an interrupt handler. 103.Pp 104The 105.Fa cookiep 106argument is a pointer to a 107.Vt "void *" 108that 109.Fn BUS_SETUP_INTR 110will write a cookie for the parent bus' use to if it is successful in 111establishing an interrupt. 112Driver writers may assume that this cookie will be non-zero. 113The nexus driver will write 0 on failure to 114.Fa cookiep . 115.Pp 116The interrupt handler will be detached by 117.Fn BUS_TEARDOWN_INTR . 118The cookie needs to be passed to 119.Fn BUS_TEARDOWN_INTR 120in order to tear down the correct interrupt handler. 121Once 122.Fn BUS_TEARDOWN_INTR 123returns, it is guaranteed that the interrupt function is not active and 124will no longer be called. 125.Pp 126Mutexes are not allowed to be held across calls to these functions. 127.Ss "Filter Routines" 128A filter runs in primary interrupt context. 129In this context, normal mutexes cannot be used. 130Only the spin lock version of these can be used (specified by passing 131.Dv MTX_SPIN 132to 133.Fn mtx_init 134when initializing the mutex). 135.Xr wakeup 9 136and similar routines can be called. 137Atomic operations from 138.Pa machine/atomic 139may be used. 140Reads and writes to hardware through 141.Xr bus_space 9 142may be used. 143PCI configuration registers may be read and written. 144All other kernel interfaces cannot be used. 145.Pp 146In this restricted environment, care must be taken to account for all 147races. 148A careful analysis of races should be done as well. 149It is generally cheaper to take an extra interrupt, for example, than 150to protect variables with spinlocks. 151Read, modify, write cycles of hardware registers need to be carefully 152analyzed if other threads are accessing the same registers. 153.Pp 154Generally, a filter routine will use one of two strategies. 155The first strategy is to simply mask the interrupt in hardware and 156allow the 157.Dv ithread 158routine to read the state from the hardware and then reenable 159interrupts. 160The 161.Dv ithread 162also acknowledges the interrupt before re-enabling the interrupt 163source in hardware. 164Most PCI hardware can mask its interrupt source. 165.Pp 166The second common approach is to use a filter with multiple 167.Xr taskqueue 9 168tasks. 169In this case, the filter acknowledges the interrupts and queues the 170work to the appropriate taskqueue. 171Where one has to multiplex different kinds of interrupt sources, like 172a network card's transmit and receive paths, this can reduce lock 173contention and increase performance. 174.Pp 175You should not 176.Xr malloc 9 177from inside a filter. 178You may not call anything that uses a normal mutex. 179Witness may complain about these. 180.Ss "ithread Routines" 181You can do whatever you want in an ithread routine, except sleep. 182Care must be taken not to sleep in an ithread. 183In addition, one should minimize lock contention in an ithread routine 184because contested locks ripple over to all other ithread routines on 185that interrupt. 186.Ss "Sleeping" 187Sleeping is voluntarily giving up control of your thread. 188All the sleep routine found in 189.Xr msleep 9 190sleep. 191Waiting for a condition variable described in 192.Xr condvar 9 193is sleeping. 194Calling any function that does any of these things is sleeping. 195.Sh RETURN VALUES 196Zero is returned on success, 197otherwise an appropriate error is returned. 198.Sh SEE ALSO 199.Xr random 4 , 200.Xr device 9 , 201.Xr driver 9 , 202.Xr locking 9 203.Sh AUTHORS 204.An -nosplit 205This manual page was written by 206.An Jeroen Ruigrok van der Werven Aq Mt asmodai@FreeBSD.org 207based on the manual pages for 208.Fn BUS_CREATE_INTR 209and 210.Fn BUS_CONNECT_INTR 211written by 212.An Doug Rabson Aq Mt dfr@FreeBSD.org . 213