Date: Fri, 19 Jun 1998 07:49:12 -0700 (PDT) From: Julian Elischer <julian@whistle.com> To: "Kenneth D. Merry" <ken@plutotech.com> Cc: Joao Carlos Mendes Luis <jonny@jonny.eng.br>, ckempf@enigami.com, freebsd-scsi@FreeBSD.ORG Subject: Re: Rolling CAM in, what is still needed? Message-ID: <Pine.BSF.3.95.980619074704.11965B-100000@current1.whistle.com> In-Reply-To: <199806190414.WAA10410@panzer.plutotech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
surely both old and new interfaces can co-exist.. this is really quite important for netBSD binary compatibility.. Juat mention in the man pages that the old one is less prefered and slower (is it?) Surely the old interface is pretty generic.. (userland scsi that is) On Thu, 18 Jun 1998, Kenneth D. Merry wrote: > Joao Carlos Mendes Luis wrote... > > #define quoting(Kenneth D. Merry) > > // I know that changes in API can be a PITA, I had to port a very > > // large application from the old scsireq/SCIOCCOMMAND system to the new > > // passthrough driver. We can't keep old API's around forever for no > > // particular reason, so I'd rather go ahead and port things. > > > > Sorry for my ignorance, but is this interface FreeBSD only ? > > Nope, it's also used in NetBSD and OpenBSD, I think. > > > If it's somewhat generic (say, *BSD), I'd prefer to have both, > > and have immediate portability for "future" applications. > > Is there a big advantage on the new scheme ? > > Yeah, it doesn't involve needless translation, and you can send > more than just SCSI commands with it. You can send any type of CCB using > the standard passthrough ioctl. > > > At least, while CAM is in transition mode, being available only > > as patches, the compatibilty API would make easy to choose cam > > or not-cam during boot time (for those interfaces available in > > both modes, of course). I hope that's what you said above. > > Yeah, you could do it that way I suppose. But there's more than > just random SCSI utilities that will break if you run a CAM userland with a > regular -current kernel. The device statistics code is completely > different in CAM, and all the dk* stuff doesn't exist anymore. i.e. > systat, iostat, vmstat and rpc.rstatd compiled under CAM won't work right > with a regular -current kernel, and vice versa. > > > As always, JMHO, and not being an involved programmer, my vote > > is not much here. :) > > I'm not terribly inclined to do it myself, at the moment I don't > think it's really worth the effort. If someone else thinks it's important > to have scsireq support, and wants to go through the trouble to do it, I > may be able to fish out my (probably now broken) code to implement the > SCIOCCOMMAND in the passthrough driver. You'll still have to point the > applications to /dev/rpassn, though. > > > Ken > -- > Kenneth Merry > ken@plutotech.com > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-scsi" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95.980619074704.11965B-100000>