ACPI Resume returns g_vfs_done() error and kernel panic...
Dr. Gary E. RAFE
drgerlists at gmail.com
Sun Dec 17 21:21:41 PST 2006
Background: Toshiba Satellite Pro 6100
BIOS Version 1.90 (latest, dated 11 Aug 2003)
FreeBSD 6.1-RELEASE with custom kernel(s)
APM Suspend/Resume works without trouble
on this configuration
(*with* USB devices compiled into the kernel).
Currently: After reconfiguring my wife's Toshiba Satellite Pro 4600
recently with FreeBSD 6.1-RELEASE and finding ACPI Suspend/Resume
appears to work mostly as it should
(USB still complains a bit, but it does work OK),
I decided to re-visit ACPI on my 6100.
The first ACPI suspend attempt on the 6100 with a (new)
USB-free kernel appeared to suspend the system OK
with "acpiconf -s S3".
Lid-open resume on this first attempt returned
a corrupted text console that came back OK
following an Fn-F5 cycle.
Many "g_vfs_done():ad0s2a..." error messages
on the text console suggested trouble, however,
so "init 6" was issued, which immediately initiated
a kernel panic.
Following the reboot, I discovered that my /etc/rc.local
file was corrupted (truncated) -- darn.
A complete "verbose" dmesg(8) output from a subsequent ACPI
suspend/resume cycle is posted here:
Note that setting "sysctl hw.acpi.video_reset=0" fixed
the corrupted text console on this Toshiba.
By way of contrast, dmesg(8) output from a working APM
suspend/resume is here:
For completeness, hardware ACPI-related sysctl(8) information
And the ASL dump from acpidump(8):
Since APM Suspend/Resume is known to work,
this is not a show-stopper for me.
I have been interested, however, in getting ACPI services
to work on this Toshiba.
Guidance & direction as to where to look next will be
Dr Gary E RAFE: drgerlists at gmail dot com
More information about the freebsd-mobile