cvs commit: src/sys/dev/fdc fdc_isa.c

[Poul-Henning Kamp]

In message <20040929070455.GB905 at freebsd3.cimlogic.com.au>, John Birrell writes:
>On Wed, Sep 29, 2004 at 08:58:49AM +0200, Poul-Henning Kamp wrote:
>> This again smells like the drive change line not being sensed correctly at 3f7:
>> 
>> It would happen if we think there is a floppy in the drive, but in fact there isn't.
>
>Do you want me to get more accurate output via serial console? I'll try to
>increase the dmesg buffer site and disable the debug sysctls in sysctl.conf in the
>hope that I can capture the boot output without having to go out in search of a serial
>cable.
>
>The desktop machine I typing this message on hangs beyond the 2 minute
>period. I had to disable the floppy disk in the BIOS to get it to boot.

Can you try first to use this patch to find the resource allocation:

Index: src/sys/dev/fdc/fdc.c
===================================================================
RCS file: /usr/cvs/src/sys/dev/fdc/fdc.c,v
retrieving revision 1.283.2.1
diff -u -r1.283.2.1 fdc.c
--- src/sys/dev/fdc/fdc.c       18 Sep 2004 04:57:55 -0000      1.283.2.1
+++ src/sys/dev/fdc/fdc.c       28 Sep 2004 21:44:57 -0000
@@ -1684,6 +1684,14 @@
        int     error;

        fdc = device_get_softc(dev);
+
+       printf("Contents of fdc_data at start of fdc_attach:\n");
+       printf("  portt=0x%0x   porth=0x%0x   stst=0x%0x   stsh=0x%0x\n",
+           fdc->portt, fdc->porth, fdc->stst, fdc->stsh);
+       printf("  ctlt=0x%0x    ctlh=0x%0x\n", fdc->ctlt, fdc->ctlh);
+       printf("  port_off=0x%0x    ctl_off=0x%0x    sts_off=0x%0x\n",
+           fdc->port_off, fdc->ctl_off, fdc->sts_off);
+
        fdc->fdc_dev = dev;
        error = fdc_initial_reset(dev, fdc);
        if (error) {


-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.