From: Pavel Machek <pavel@ucw.cz>
To: "David S. Miller" <davem@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>,
Arnd Bergmann <arnd@arndb.de>,
linux-kernel@vger.kernel.org
Subject: Re: ioctl cleanups: enable sg_io and serial stuff to be shared
Date: Thu, 8 May 2003 22:33:14 +0200 [thread overview]
Message-ID: <20030508203313.GA2787@elf.ucw.cz> (raw)
In-Reply-To: <1052323484.9817.14.camel@rth.ninka.net>
Hi!
> > Not sure if we are not too close to stable release to do that? And I
> > see no incremental way how to get there. Moving compatibility stuff
> > closer to drivers can be done close to stable release...
>
> You can define it as follows:
>
> 1) If entry exists in COMPAT or TRANSLATE table, invoke
> fops->ioctl(), else
>
> 2) If ->compat_ioctl() exists, invoke that, else
>
> 3) Fail.
>
> The COMPAT tables are sort of valuable, in that it eliminates
> the need to duplicate code when all of a drivers ioctls need
> no translation.
>
> BTW, need to extend this to netdev->do_ioctl as well for the
> handling of SIOCDEVPRIVATE based stuff. Oh goody, we can finally
> fix up that crap :))))
There's a *lot* of structs that contain *ioctl:
pavel@amd:/usr/src/linux-test/include/linux$ grep "*ioctl" *
pavel@amd:/usr/src/linux-test/include/linux$ grep "*ioctl" *
atmdev.h: int (*ioctl)(struct atm_dev *dev,unsigned int cmd,void *arg);
atmdev.h: int (*ioctl)(struct atm_dev *dev,unsigned int cmd,void *arg);
fs.h: int (*ioctl) (struct inode *, struct file *, unsigned, unsigned long);
fs.h: int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
hdlc.h: int (*ioctl)(struct net_device *dev, struct ifreq *ifr, int cmd);
hdlcdrv.h: int (*ioctl)(struct net_device *, struct ifreq *,
ide.h: int (*ioctl)(ide_drive_t *, struct inode *, struct file *, unsigned int, unsigned long);
if_bridge.h:extern void brioctl_set(int (*ioctl_hook)(unsigned long));
if_pppox.h: int (*ioctl)(struct socket *sock, unsigned int cmd,
ioctl32.h:typedef int (*ioctl_trans_handler_t)(unsigned int, unsigned int, unsigned long, struct file *);
loop.h: int (*ioctl)(struct loop_device *, int cmd,
loop.h: int (*ioctl)(struct loop_device *, int cmd, unsigned long arg);
net.h: int (*ioctl) (struct socket *sock, unsigned int cmd,
ppp_channel.h: int (*ioctl)(struct ppp_channel *, unsigned int, unsigned long);
serial_core.h: int (*ioctl)(struct uart_port *, unsigned int, unsigned long);
tty_driver.h: * int (*ioctl)(struct tty_struct *tty, struct file * file,
tty_driver.h: int (*ioctl)(struct tty_struct *tty, struct file * file,
tty_ldisc.h: * int (*ioctl)(struct tty_struct * tty, struct file * file,
tty_ldisc.h: int (*ioctl)(struct tty_struct * tty, struct file * file,
usb.h: int (*ioctl) (struct usb_interface *intf, unsigned int code, void *buf);
wanrouter.h: int (*ioctl) (struct wan_device *wandev, unsigned cmd,
pavel@amd:/usr/src/linux-test/include/linux$
What about this one: redefine it to (*ioctl)( ...., unsigned *long*,
unsinged long). That means we can add
#define IOCTL_COMPAT 0x1 0000 0000
and avoid adding new field to each such structure. Also I will not
have to duplicate lots of middle-level code (I will have to modify
unsigned int -> unsigned long, but no second copies of everything). It
means that architecture with CONFIG_COMPAT needs to have unsigned long
> 32 bits, but I guess we can live with that.
What do you think?
Pavel
--
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
next prev parent reply other threads:[~2003-05-08 20:24 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20030507104008$12ba@gated-at.bofh.it>
2003-05-07 11:51 ` ioctl cleanups: enable sg_io and serial stuff to be shared Arnd Bergmann
2003-05-07 12:41 ` Pavel Machek
2003-05-07 12:56 ` Christoph Hellwig
2003-05-07 14:39 ` David S. Miller
2003-05-07 15:12 ` Jeff Garzik
2003-05-07 14:07 ` David S. Miller
2003-05-08 10:46 ` Gerd Knorr
2003-05-08 15:16 ` Ben Collins
2003-05-08 15:51 ` Christoph Hellwig
2003-05-08 15:37 ` Ben Collins
2003-05-08 19:34 ` Pavel Machek
2003-05-08 19:27 ` Ben Collins
2003-05-08 20:06 ` Pavel Machek
2003-05-08 19:47 ` Ben Collins
2003-05-08 20:09 ` David S. Miller
2003-05-08 20:26 ` David Mosberger
2003-05-08 20:33 ` Arjan van de Ven
2003-05-08 21:11 ` David Mosberger
2003-05-08 15:23 ` Arnd Bergmann
2003-05-07 15:28 ` Pavel Machek
2003-05-07 16:04 ` David S. Miller
2003-05-07 19:13 ` Arnd Bergmann
2003-05-07 18:12 ` David S. Miller
2003-05-07 23:50 ` Arnd Bergmann
2003-05-08 10:35 ` David S. Miller
2003-05-07 22:20 ` Alan Cox
2003-05-08 20:33 ` Pavel Machek [this message]
2003-05-08 20:43 ` David S. Miller
2003-05-08 23:03 ` Pavel Machek
2003-05-08 23:35 ` Arnd Bergmann
2003-05-08 22:48 ` Arnd Bergmann
2003-05-08 23:22 ` Pavel Machek
2003-05-07 13:18 ` Arnd Bergmann
2003-05-07 14:01 ` Carl-Daniel Hailfinger
2003-05-07 15:16 ` Pavel Machek
2003-05-07 15:46 ` Arnd Bergmann
2003-05-07 16:07 ` Jens Axboe
2003-05-07 16:20 ` Arnd Bergmann
2003-05-07 10:27 Pavel Machek
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20030508203313.GA2787@elf.ucw.cz \
--to=pavel@ucw.cz \
--cc=arnd@arndb.de \
--cc=davem@redhat.com \
--cc=hch@infradead.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox