[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20090929234733.5b9fbc67@lxorguk.ukuu.org.uk>
Date: Tue, 29 Sep 2009 23:47:33 +0100
From: Alan Cox <alan@...rguk.ukuu.org.uk>
To: Alan Stern <stern@...land.harvard.edu>
Cc: Jason Wessel <jason.wessel@...driver.com>,
USB list <linux-usb@...r.kernel.org>,
Kernel development list <linux-kernel@...r.kernel.org>
Subject: Re: How to handle console devices
> I don't know what the best way is to accomplish this. Create dummy
> inode and file structs and pass them to the usual tty_open() routine?
> (But then what about hangup event handling?) What do you think?
What I long term envisioned was that
- Every tty (or at least every interesting tty) would have a tty_port
object for the hardware [done now for all consoles but vt]
- Every tty port object would have an "output" method
- The tty->termios would move to the tty_port
- The tty receive buffers would move to the tty_port (only needed for a
console that supports input)
At that point
- The tty lock/refcount isn't needed all the time for the receive data
paths which speeds it up a fair bit
- A console can be implemented without a tty_struct anywhere in sight
- Flow control and speed setting can be done on hardware generically on
resume paths
That fixes the lifetime and magic object invention issues that plague the
current console.
I still think that is the right way to fix it.
Alan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists