Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 14 Jul 2003 10:53:38 -0700
From:      Terry Lambert <tlambert2@mindspring.com>
To:        Petri Helenius <pete@he.iki.fi>
Cc:        Andrew Gallatin <gallatin@cs.duke.edu>
Subject:   Re: dumb KSE question
Message-ID:  <3F12EE22.A145A9B5@mindspring.com>
References:  <Pine.BSF.4.21.0307121222150.87910-100000@InterJet.elischer.org> <003501c348ad$a599fba0$812a40c1@PETEX31>

next in thread | previous in thread | raw e-mail | index | archive | help
Petri Helenius wrote:
> The cv I=B4m talking about would be signaled by a device driver which
> shares some memory with the netgraph device node and the question
> is about if in the ng_device like thing I can have the d_ioctl_t block
> on the cv.

People always want to do this sort of thing, but it's a really bogus
thing to want to do.  The problem with it is that you are crossing a
protection domain, and, in doing that, all bets are off, in terms of
allowable locking, etc..  The big issue you are going to hit is that
you will deadlock.

The proper thing to do is perform a blocking operation on the device
itself, and unblock it when the condition is true in order to signal
the user space process.

-- Terry



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3F12EE22.A145A9B5>