Repeatable lockup, and return of SATA drive-size problems

[Scott Ellis <scotte%warped.com@localhost>]

On my amd64 system, updating from May 28th to June 10th -current shows
two peculiar issues that have appeared.  First, the system will hang 
after ~5 hours of uptime.  Last night it hung running the nightly 
scripts, and I can get it to hang quicker (it seems) by stressing the 
filesystem (read/write using Samba, for example).

Since I'm running a kernel w/o debug, the
behavior is simply that the kernel keeps running ('ping' to the box
works), but userland doesn't (no new connections can get established,
existing applications are frozen).  My next step is a debug kernel so I
can (hopefully) catch what it's doing in the debugger.

But...the more interesting this IMHO is that the problem where the disks
will "sometimes" probe with the wrong size (and show up as ST506) is
back.  This is identical to PR#38548.  It seems that if I shutdown
the system fully and cold-boot, the drives probe properly.  Rebooting 
otherwise more often than not, exhibits the problem.  I'm not sure where 
to look to help debug this part.

Just wanted to make a note of it on this list, in case others are 
experiencing the same problems.

        ScottE