Date: Wed, 30 Nov 2011 16:59:06 +0100 From: Baptiste Daroussin <bapt@freebsd.org> To: Kostik Belousov <kostikbel@gmail.com> Cc: freebsd-current@freebsd.org Subject: Re: [patch] turning devctl into a "multiple openable" device Message-ID: <20111130155906.GB1621@azathoth.lan> In-Reply-To: <20111130154636.GX50300@deviant.kiev.zoral.com.ua> References: <20111130124320.GA1449@azathoth.lan> <201111301005.11938.jhb@freebsd.org> <20111130154636.GX50300@deviant.kiev.zoral.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
--KFztAG8eRSV9hGtP Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, Nov 30, 2011 at 05:46:36PM +0200, Kostik Belousov wrote: > On Wed, Nov 30, 2011 at 10:05:11AM -0500, John Baldwin wrote: > > On Wednesday, November 30, 2011 7:43:20 am Baptiste Daroussin wrote: > > > Hi all, > > >=20 > > > With the help of cognet, I wrote a patch to turn devctl into a multip= le openable > > > device, that mean that it will allow to open /dev/devctl in multiple = programs, > > > for example hald and everythings that want to receive notification fr= om the > > > device won't need to depend on haveing devd running. > > >=20 > > > here is the patch:=20 > > > http://people.freebsd.org/~bapt/devctl_multi_open.diff > >=20 > > Shouldn't devctl_queue_data_f() use the requested malloc() flags instea= d of > > hardcoding M_NOWAIT? > This is an obvious fallback of holding mutex around the call to > per_devctl_queue_data_f(), which caused the author a trouble to use > M_WAITOK. >=20 > Having n readers causes the patch to queue each message n times, that loo= ks > like a waste. >=20 > I wonder why the waiting_threads stuff is needed at all. The cv could > be woken up unconditionally everytime. What is the reason for the cv_wait > call in cdevpriv data destructor ? You cannot have a thread doing e.g. > read on the file descriptor while destructor is run. >=20 > >=20 > > Also, I know that it was an intentional design decisison by Warner to h= ave > > the multiplexing of devctl data done in userland via devd rather than i= n the > > kernel. (I think he envisioned devd providing a UNIX domain socket or = some > > such for other daemons to use to listen to events.) Have you asked him= about > > this change? > And I fully agree that doing multiplexing in user mode is the right appro= ach. > Not least because you could apply some advanced access control and provide > filtering for the consumers. I agree that in most cases this is better, but being able to have multiple readers is anyway useful, having the futur libudev or alike not depends on = devd running would be great imho. I have boxes that do not have devd and won't have it would be useless but r= un programs that needs to get notification for some hardware. I would love to remove devd on those boxes (they are boxes where the FS size is limited) regards, Bapt --KFztAG8eRSV9hGtP Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (FreeBSD) iEYEARECAAYFAk7WUsoACgkQ8kTtMUmk6EzkwACfUTu9h2FY/9WFnxMH1wk4iKrm 9ecAnReKgkHD10fwKhd6EvDsUOguuIl7 =e8tO -----END PGP SIGNATURE----- --KFztAG8eRSV9hGtP--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20111130155906.GB1621>