Problems with clock on 1.1-CURRENT

[Justin C. Sherrill]

apollo.backplane.com> <410a584a$0$206$415eb37d at xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Organization: Time Warner Cable
User-Agent: KNode/0.7.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7Bit
X-cow-says: Moo
Lines: 27
NNTP-Posting-Host: 24.93.1.61
X-Trace: 1091560145 crater_reader.dragonflybsd.org 206 24.93.1.61
Xref: crater_reader.dragonflybsd.org dragonfly.bugs:1983


Just to bring up this problem again:  On my computer where the clock runs
too fast, I had this in my /boot/loader.conf:

vesa_load="YES"
splash_bmp_load="YES"
bitmap_load="YES"
bitmap_name="dfly.bmp"

I deleted that, and put in: 

debug.acpi.disabled="acpi"

And the clock ran at a normal speed, for about half an hour.  At that point,
I put the vesa_load and etc. lines back in, removed the ACPI line, and
rebooted.  After about an hour and a half, the clock had skewed forward
some two (additional) hours.

I deleted the vesa_load and etc. lines again, went back to having
debug.acpi.disabled="acpi" only in loader.conf, and so far (2 hours later),
the clock is still correct.  

The only other kernel module used is:
  2    1 0xcc5c8000 7000     logo_saver.ko

It looks like having vesa.ko loaded or perhaps ACPI on is what causes it. 
Can someone with the too-fast clock problem duplicate this?