Date: Fri, 22 Jun 2007 12:05:33 -0700 From: Scott Long <scottl@samsco.org> To: "M. Warner Losh" <imp@bsdimp.com> Cc: scsi@freebsd.org Subject: Re: DA_Q_NO_SYN_CACHE Message-ID: <467C1D7D.7010501@samsco.org> In-Reply-To: <20070622.113538.-749248556.imp@bsdimp.com> References: <20070622.110918.-345494230.imp@bsdimp.com> <467C055A.5090008@samsco.org> <20070622.113538.-749248556.imp@bsdimp.com>
next in thread | previous in thread | raw e-mail | index | archive | help
M. Warner Losh wrote: > In message: <467C055A.5090008@samsco.org> > Scott Long <scottl@samsco.org> writes: > : M. Warner Losh wrote: > : > Is the DA_Q_NO_SYN_CACHE quirk still needed? there's suggestions in > : > some PRs I'm looking at to never send this command for UFI devices. > : > Was that ever implemented, and is that a good idea? > : > > : > : Are you suggesting modifying umass to universally disable the sync cache > : for UFI devices? That's something that should be looked at, but I don't > : think that the 7.0 code freeze is the best time to experiment with that. > > My question was more "has this already been done," since if it has, I > can close about 15PRs with no action required in the tree. If it has > not been done, then we're past what I feel comfortable doing during a > freeze as well, unless one can get a lot of verified testing done with > ~100 devices or more. > > So it sounds like the answer is 'We need to keep adding the quirks for > those devices that hang the system' that can be found in the PR > database. That I do feel comfortable doing during a code freeze. I don't see anything in the umass driver that suggests it gets automatically handled. So yeah, the right thing to do at this point is to just add more quirks (make sure you read Nate's posting on adding quirks first), and hopefully we'll improve the state of the art in the future. Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?467C1D7D.7010501>