Date: Sat, 16 Apr 2016 10:48:08 +0200 From: "O. Hartmann" <ohartman@zedat.fu-berlin.de> To: Warner Losh <imp@bsdimp.com> Cc: Maxim Sobolev <sobomax@freebsd.org>, "Conrad E. Meyer" <cem@freebsd.org>, FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: Heads up Message-ID: <20160416104808.7228783b.ohartman@zedat.fu-berlin.de> In-Reply-To: <CANCZdfpKmk4%2BmS-xNN0cdCyOb=GN2P5%2BESZpOWwJCTZPkWpm8w@mail.gmail.com> References: <CANCZdfpnYnVrvhNagYUT9RhAuC1AMCrxh=GCt8RKT0bqxuJybw@mail.gmail.com> <CAH7qZftspY=i5MaQ6sNg5L5XXBN7gfCuAMO-NZt=7Qp5zo6u_w@mail.gmail.com> <CAG6CVpXKZJFi7f=uxoXbc-MoHwd6JqZMwh-6m%2BGZvJ9JT95ZOA@mail.gmail.com> <CAH7qZfuA56CxMo9B6-1%2B7hhvugGiAMJG0_O3qW57chesSA_1Kw@mail.gmail.com> <CANCZdfpKmk4%2BmS-xNN0cdCyOb=GN2P5%2BESZpOWwJCTZPkWpm8w@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--Sig_/L0rk=nwW+fkMajLQe1K3nwT Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable Am Fri, 15 Apr 2016 11:30:05 -0600 Warner Losh <imp@bsdimp.com> schrieb: > On Fri, Apr 15, 2016 at 11:25 AM, Maxim Sobolev <sobomax@freebsd.org> wro= te: >=20 > > Conrad, yes, you can, but sticking it into FreeBSD source tree IMHO > > restricts your options somewhat. If it's your private code you can easi= ly > > put F-world into it and nobody obviously cares apart from your co-worke= rs > > and your boss. Would probably be considered highly inappropriate by mos= t to > > merge it into a public one, so "whatever your like" rule has its limits. > > =20 >=20 > The name isn't obscene or offensive. >=20 >=20 > > Also, apart from being not very suggestive as to what this option actua= lly > > does, let's not forget that "Netflix" is probably registered as a trade= mark > > and whatnot in North America at least, if not worldwide. I am not a law= yer, > > but I suspect using it might have some ramifications for our ability to > > name things using this name. > > =20 >=20 > Probably not. This is a weak argument. It's a descriptive use of the word, > and not being used to market or sell video streaming services or DVD rent= al > by mail. >=20 >=20 > > If "SSDNG" is not a good name, I'd suggest coming up with some other na= me > > that is neutral and also has some hints as to its functionality. For > > example, CAM_IOCHED_LOW_LAT or something. > > =20 >=20 > Also a horrible name. It's a generic I/O scheduler. It can do lots of > things. I keep saying that, and categorically refuse to name the more > expansive scheduler anything that's so limiting. >=20 > Warner Why the not simply CAM_IOSCHED_EXT for an extension of the already present = i/o scheduling? Or without the _EXT? >=20 >=20 > > On Fri, Apr 15, 2016 at 9:22 AM, Conrad Meyer <cem@freebsd.org> wrote: > > =20 > >> Max, > >> > >> If you implement a new IO scheduler you can name it whatever you like. > >> "NG" isn't any more meaningful than "Netflix." > >> > >> Best, > >> Conrad > >> > >> On Fri, Apr 15, 2016 at 9:13 AM, Maxim Sobolev <sobomax@freebsd.org> > >> wrote: =20 > >> > Great, work Warner, thanks! Small note, though. The CAM_IOSCHED_NETF= LIX > >> > seems like a quite poor name for a kernel option. IMHO there is no g= ood > >> > reason for polluting it with the name of the company that sponsored = the > >> > development. I don't think we have any precedents of doing this unle= ss =20 > >> the =20 > >> > option is related to a piece of hardware that the company makes, and= =20 > >> it's =20 > >> > not the case here. Apart from "coolness" factor as far as I understa= nd =20 > >> that =20 > >> > _NETFLIX suffix does not give any tangible benefit for anybody readi= ng > >> > kernel config and trying to understand what this option actually doe= s. > >> > CAM_IOSCHED_SSDNG or something would be better IMHO. Just my $0.02. > >> > > >> > -Max > >> > > >> > On Thu, Apr 14, 2016 at 3:42 PM, Warner Losh <imp@bsdimp.com> wrote: > >> > =20 > >> >> The CAM I/O scheduler has been committed to current. This work is = =20 > >> described =20 > >> >> in https://people.freebsd.org/~imp/bsdcan2015/iosched-v3.pdf though= =20 > >> the =20 > >> >> default scheduler doesn't change the default (old) behavior. > >> >> > >> >> One possible issue, however, is that it also enables NCQ Trims on a= da =20 > >> SSDs. =20 > >> >> There are a few rogue drives that claim support for this feature, b= ut > >> >> actually implement data corrupt instead of queued trims. The list o= f =20 > >> known =20 > >> >> rogues is believed to be complete, but some caution is in order. > >> >> > >> >> Warner > >> >> _______________________________________________ > >> >> freebsd-current@freebsd.org mailing list > >> >> https://lists.freebsd.org/mailman/listinfo/freebsd-current > >> >> To unsubscribe, send any mail to " =20 > >> freebsd-current-unsubscribe@freebsd.org" =20 > >> >> > >> >> =20 > >> > _______________________________________________ > >> > freebsd-current@freebsd.org mailing list > >> > https://lists.freebsd.org/mailman/listinfo/freebsd-current > >> > To unsubscribe, send any mail to " =20 > >> freebsd-current-unsubscribe@freebsd.org" > >> > >> =20 > > =20 > _______________________________________________ > freebsd-current@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org" --Sig_/L0rk=nwW+fkMajLQe1K3nwT Content-Type: application/pgp-signature Content-Description: OpenPGP digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAEBCAAGBQJXEfxIAAoJEOgBcD7A/5N82g0H+gP730HgUUJt4ezg/6VEAqZj B14s+YmbsnX6i5y2WxXZNWQzPv4Pfkjuww4x9vihy5YEB3f6+WApA/zzbWbSrRCC OpPQebjB4SVer78vur1Qccs4H7Zbb4ocxMOHkbJfO23Au7Dn1YBOX1ClE6EqSqm5 qwOycugcU8KcBc7DLobsD7vapINgGTWOwP1E11aMINqlv4IOH/4SCrkeZY7pzTWf MN9ZEXNvfB2gxZIaBPhmVZI/n2ZyLUetdJKzU0hVId+TF0gdyq4XjFb5Kq9F6Lar DUMunL5Qa+0atPmQu/TKtEbhYcXJbTTQ1jjSK1puZr9ibduRaX1nk8/ZsAwVFyY= =LqBh -----END PGP SIGNATURE----- --Sig_/L0rk=nwW+fkMajLQe1K3nwT--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160416104808.7228783b.ohartman>