DELL SAS5/E Controller bug

[Stephane LAPIE]

John Baldwin wrote:
> Gah, that should be the case that I ignore.  Can you replace the second 
> warnx() call I added with this:
> 
> 		warnx("mpt_read_ioc_page(6): %s (%x)", mpt_ioc_status(IOCStatus),
> 		    IOCStatus);

I now get the following message :
mptutil: mpt_read_ioc_page(6): Invalid configuration page (8022)

(Though I guess this doesn't tell anything that we did not know initially)

> I know that the rescan after removing a device is a bit messy (lots of 
> messages before daX actually goes away), but I don't recall it taking such a 
> long time.

Even without rescanning the bus, the device actually goes away on its
own after the same delay of three minutes.

> The documentation is not public.  The 0x12 and 0x16 messages are events that
> I have seen.  You can try talking to scottl@ as he has access to the docs.

I could contact Scott, and here are the relevant bits of his answer :
> The basic problem is that FreeBSD still sees all of this as parallel SCSI, subject to rescans and resets and timeouts.  It's fighting with the SAS controller.  I'll explain more below.

> I'm working on code that will make FreeBSD more aware of how SAS works.  It's several months from being done, though. 

Reposting here for reference the meaning of 0x12 and 0x16 events :
0x12 : SAS Link status changed
0x16 : SAS Discovery Event

I was wondering if using an Areca SAS controller could be a better
solution, but Scott's answer has me wondering if this is a common issue
to all SAS controllers on FreeBSD.
-- 
Stephane LAPIE, EPITA SRS, Promo 2005
"Even when they have digital readouts, I can't understand them."
--MegaTokyo

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 261 bytes
Desc: OpenPGP digital signature
Url : http://lists.freebsd.org/pipermail/freebsd-hardware/attachments/20100122/c76bf463/signature.pgp