Re: More SATA disk detection issues with MCP65 et al

[Dieter Baron <dillo%danbala.tuwien.ac.at@localhost>]

On Sat, Aug 02, 2008 at 07:47:08PM +0200, Johan Ihren wrote:
> As reported yesterday a fix was committed that forces ahcisata to  
> attach to the Nvidia MCP65 as it doesn't work with viaide. So for so  
> good, viaide doesn't work for me either with the MCP65.
> 
> My problem is that even with this fix there's a bit of randomness as  
> to how many disks are detected, and particularly annoying is that if  
> there is more than one disk attached then usually the last disk to be  
> detected is completely wrongly matched. I.e. in a system with 4  
> identical Seagate 160GB disks I get this:
> 
> ...
> ahcisata0 port 0: device present, speed: 1.5Gb/s
> ahcisata0 port 1: device present, speed: 1.5Gb/s
> ahcisata0 port 2: device present, speed: 1.5Gb/s
> ahcisata0 port 3: device present, speed: 1.5Gb/s
> wd0 at atabus0 drive 1: <ST3160023A>             # this is another  
> disk, on an IDE bus, don't worry about it
> wd0: ...
> wd0: ...
> wd1 at atabus1 drive 0: <ST3160827AS>            # first SATA disk
> wd1: quirks 2<FORCE_LBA48>
> wd1: 149 GB, 310101 cyl, 16 head, 63 sec, 512 bytes/sect x 31258108  
> sectors
> wd2: at atabus1 drive 0: <ST506>                 # second SATA disk
> wd2: 69632 KB, 1024 cyl, 8 head, 17 sec, 512 bytes/sect x 139264 sectors
> wd2d: device timeout reading fsbn 0 (wd1 bn 0; cn 0 tn 0 sn 0), retrying
> 
> and that's the end of it (we never get to wd3 and wd4). If I reboot  
> without changing ANYTHING I sometimes get wd1, wd2 and wd3 correctly  
> matched and then the bogus ST506 match on wd4 (and the same hang at  
> "retrying" as above).
> 
> I've managed to get the system up with all four SATA disks attached a  
> few times, but that is unusual, most attempts have ended with a hang  
> at "retrying" on a wrongly matched disk "wdN" for varying values of "N".
> 
> Anyone else seen anything like this? Known problem? Any suggestions?

  Do you also have revision 1.17 of sys/dev/ic/ahcisata_core.c (search
for ``before actually starting''.  If the delay is 500ms you do, if
it is 100ms you don't).  Wihtout this fix, I had similar problems with
my MCP67.

                                        yours,
                                        dillo