On Thursday 24 April 2008 10:12:38 pm M. Warner Losh wrote: > In message: <[EMAIL PROTECTED]> > > John Baldwin <[EMAIL PROTECTED]> writes: > : --- //depot/projects/smpng/sys/dev/sio/sio.c 2008/01/21 18:58:30 > : +++ //depot/user/jhb/intr/dev/sio/sio.c 2008/04/24 14:51:42 > : @@ -85,6 +85,11 @@ > : #endif > : #include <dev/ic/ns16550.h> > : > : +#if defined(__i386__) || defined(__amd64__) > : +#define PROBE_IRQ > : +#include <machine/intr_machdep.h> > : +#endif > : + > : #define LOTS_OF_EVENTS 64 /* helps separate urgent events from > input */ > : > : #ifdef COM_MULTIPORT > > noprobe is true for all busses except isa and cbus. So really, you > could make this patch a lot simpler. > > Index: sio.c > =================================================================== > RCS file: /cache/ncvs/src/sys/dev/sio/sio.c,v > retrieving revision 1.473 > diff -u -r1.473 sio.c > --- sio.c 25 Dec 2007 17:51:57 -0000 1.473 > +++ sio.c 25 Apr 2008 02:11:08 -0000 > @@ -430,8 +430,10 @@ > int fn; > device_t idev; > Port_t iobase; > +#ifdef DEV_ISA > intrmask_t irqmap[4]; > intrmask_t irqs; > +#endif > u_char mcr_image; > int result; > u_long xirq; > @@ -700,9 +702,10 @@ > device_set_softc(dev, NULL); > free(com, M_DEVBUF); > } > - return (result == 0 ? BUS_PROBE_DEFAULT + 1 : result); > + return (result == 0 ? BUS_PROBE_DEFAULT - 1 : result); > } > > +#ifdef DEV_ISA > /* > * Check that > * o the CFCR, IER and MCR in UART hold the values written to them > @@ -777,7 +780,10 @@ > device_set_softc(dev, NULL); > free(com, M_DEVBUF); > } > - return (result == 0 ? BUS_PROBE_DEFAULT + 1 : result); > + return (result == 0 ? BUS_PROBE_DEFAULT - 1 : result); > +#else > + return (BUS_PROBE_DEFAULT - 1); > +#endif > } > > #ifdef COM_ESP > > > Warner
I had pondered making it not do all the checks if the IRQ wasn't an ISA irq. We can also probably not bother with the check if the IRQ came from ACPI as that tends to be correct (I've seen PNPBIOS be wrong OTOH). Probably this needs to only be done in the sio_isa_probe() routine (and sio_cbus_probe()) and moved out of sioprobe() and have a separate sio_acpi_probe() for ACPI that doesn't do it. -- John Baldwin _______________________________________________ cvs-all@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/cvs-all To unsubscribe, send any mail to "[EMAIL PROTECTED]"
