i386/100142: /dev/smb0 device not available on systems with Intel PiiX4

Arthur Hartwig arthur.hartwig at nokia.com
Wed Jul 12 05:10:15 UTC 2006


>Number:         100142
>Category:       i386
>Synopsis:       /dev/smb0 device not available on systems with Intel PiiX4
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    freebsd-i386
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Jul 12 05:10:14 GMT 2006
>Closed-Date:
>Last-Modified:
>Originator:     Arthur Hartwig
>Release:        6.0
>Organization:
Nokia
>Environment:
FreeBSD oz-net-10.nes.nokia.com 6.0-RELEASE FreeBSD 6.0-RELEASE #3: Wed Mar  1 10:46:02 EST 2006     hartwig at oz-net-10.nes.nokia.com:/usr/src/sys/i386/compile/oz-net-10  i386
>Description:
System configured with devices intppm, smbb and smbus.

On startup smbus1 is reported but no smbus0 is reported. There is no file /dev/smb0

In sys/pci/intpm.c intpm_attach() executes 
    smbinterface = device_add_child(dev, ""intsmb", unit); 

    device_probe_and_attach(smbinterface);

device_probe_and_attach() calls device_probe_child() which ends us calling intsmb_probe() twice, once when search for the best matching driver and again, because intsmb_probe() returns BUS_PROBE_DEFAULT, immediately after the comment "Call the probe method again to make sure we have the right description."

Calling intsmb_probe() twice results in creation of devices smbus0 (on the first call to intsmb_probe) and smbus1 (on the second call to intsmb_probe). The second probe call results in When intsmb_attach(0 is called, sc->smbus contains a pointer to a device structure for smbus1 and consequently smbus1 is reported in the startup.


>How-To-Repeat:

>Fix:
Suggested fix:
In sys/pci/intpm move
        sc->smbus=device_add_child(dev, "smbus", -1);
        if (!sc->smbus)
                return (EINVAL);    /* XXX don't know what to return else */

from intsmb_probe() to intsmb_attach()

The suggested fix mirrors more closely the code in sys/pci/viapm.c and sys/pci/amdpm.c



>Release-Note:
>Audit-Trail:
>Unformatted:


More information about the freebsd-i386 mailing list