re: kern/54855 (New "ahcisata0 port 1: device present" messages with NetBSD 9)

[matthew green <mrg%eterna.com.au@localhost>]

Simon Burge writes:
> jdolecek%NetBSD.org@localhost wrote:
> 
> > State-Changed-From-To: open->feedback
> > State-Changed-By: jdolecek%NetBSD.org@localhost
> > State-Changed-When: Mon, 13 Jan 2020 21:21:57 +0000
> > State-Changed-Why:
> > -current with sys/dev/ata/wd.c rev.1.454 should now disable NCQ by default,
> > can you check that fixes the problem for you?
> 
> That diff applied cleanly to the netbsd-9 branch and I've built a new
> kernel.  I'll test this tonight when I get home.
> 
> https://www.samsung.com/semiconductor/minisite/ssd/product/consumer/860evo/
> says these are the models in the 860 EVO range:
> 
> 	250 GB, 500 GB, 1,000 GB, 2,000 GB, 4,000 GB
> 
> I saw mention of the 250GB model having problems in one of the forums.
> We should add the extra model sizes to the quirk list?

i'm glad we've figured out what is the trigger here it seems.
howver, i am not a fan of the current solution, and think it
should be restricted to known bad combos.

most people with 860 EVOs are successful with NCQ, i'm one
of them with a 500GB model, with netbsd-9 for months now,
and they're a wildly successful line i'm sure many others
also have working happily with NCQ.

we don't know what the real issue is.  perhaps it is a 
problem that the controller had that these drives trigger,
rather than a problem with them, so perhaps we should
disable it on the controller instead of the drive, if
we're going to pick one of the two triggers here.


.mrg.