LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Alexey Dobriyan <adobriyan@openvz.org>
To: Al Viro <viro@ftp.linux.org.uk>
Cc: linux-kernel@vger.kernel.org, devel@openvz.org
Subject: Re: Racy /proc creations interfaces
Date: Thu, 28 Dec 2006 11:21:47 +0300	[thread overview]
Message-ID: <20061228082147.GA11127@localhost.sw.ru> (raw)
In-Reply-To: <20061227135624.GP17561@ftp.linux.org.uk>

On Wed, Dec 27, 2006 at 01:56:24PM +0000, Al Viro wrote:
> On Wed, Dec 27, 2006 at 04:42:23PM +0300, Alexey Dobriyan wrote:
> > 
> >    	struct proc_entry_raw foo_pe_raw = {
> > 		.owner = THIS_MODULE,
> > 		.name = "foo",
> > 		.mode = 0644,
> > 		.read_proc = foo_read_proc,
> > 		.data = foo_data,
> > 		.parent = foo_parent,
> > 	};
> > 
> > 	pde = create_proc_entry(&foo_pe_raw);
> > 	if (!pde)
> > 		return -ENOMEM;
> > 
> >    where "struct proc_entry_raw" is cut down version of "struct proc_dir_entry"
> 
> Ewwwwwwwwwwwwwww
> 
> Please, please no.  Especially not .parent.  If anything, let's add a
> helper saying "it's all set up now".  And turn create_proc_entry()
> into a macro that would pass THIS_MODULE to underlying function and
> call that helper, so that simple cases wouldn't have to bother at all.

People are setting ->data after create_proc_entry():

drivers/zorro/proc.c:
   110	static int __init zorro_proc_attach_device(u_int slot)
   111	{
   112		struct proc_dir_entry *entry;
   113		char name[4];
   114	
   115		sprintf(name, "%02x", slot);
   116		entry = create_proc_entry(name, 0, proc_bus_zorro_dir);
   117		if (!entry)
   118			return -ENOMEM;
   119		entry->proc_fops = &proc_bus_zorro_operations;
   120		entry->data = &zorro_autocon[slot];
   121		entry->size = sizeof(struct zorro_dev);

If create_proc_entry is a macro doing what you suggest (am I right?)

	#define create_proc_entry(name, mode, parent)
	({
		struct proc_dir_entry *pde;

		pde = __create_proc_entry(name, mode, parent, THIS_MODULE);
		if (pde)
			mark_proc_entry_ready(pde);
		pde;
	})

there is still a problem because we want it to be equivalent to

	pde = create_proc_entry(...);
	if (!pde)
		return -ENOMEM;
	pde->proc_fops = ...;
	pde->data = ...;
	mark_proc_entry_ready(pde);


  reply	other threads:[~2006-12-28  8:15 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-12-27 13:42 Alexey Dobriyan
2006-12-27 13:56 ` Al Viro
2006-12-28  8:21   ` Alexey Dobriyan [this message]
2007-01-15 15:39     ` [PATCH] Fix rmmod/read/write races in /proc entries Alexey Dobriyan
2007-01-23 20:58       ` Andrew Morton
2007-01-24 15:22         ` Alexey Dobriyan

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=20061228082147.GA11127@localhost.sw.ru \
    --to=adobriyan@openvz.org \
    --cc=devel@openvz.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=viro@ftp.linux.org.uk \
    --subject='Re: Racy /proc creations interfaces' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).