linux-kernel - Re: [PATCH] sony_pi: Remove the BKL from sonypi_misc

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 21 Oct 2009 23:41:07 +0200 (CEST)
From:	John Kacur <jkacur@...hat.com>
To:	Frederic Weisbecker <fweisbec@...il.com>
cc:	Arnd Bergmann <arndbergmann@...glemail.com>,
	linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>,
	Alan Cox <alan@...rguk.ukuu.org.uk>,
	Ingo Molnar <mingo@...e.hu>
Subject: Re: [PATCH] sony_pi: Remove the BKL from sonypi_misc_open



On Wed, 21 Oct 2009, Frederic Weisbecker wrote:

> On Tue, Oct 20, 2009 at 12:08:57AM +0200, Arnd Bergmann wrote:
> > On Tuesday 20 October 2009, Arnd Bergmann wrote:
> > > On Monday 19 October 2009, John Kacur wrote:
> > > > How does this look? (Version 2 of the patch follows)
> > > 
> > > Looks good now.
> > > 
> > 
> > A bit of background:
> > 
> > Doing only one of the two conversions is a correct patch as well
> > of course, I just want to make sure you don't have to go through all
> > the same files again once someone does a blind pushdown into the ioctl
> > and llseek functions, so once you prove that a specific driver doesn't
> > need the BKL, please always make sure to remove it from all three places.
> > 
> > I fear that the llseek part will get interesting as well, just because
> > we call default_llseek instead of no_ll by default currently.
> > It might be a good idea to add one of .llseek=no_llseek or 
> > .llseek=generic_file_llseek in any file_operations that you prove
> > to not require the BKL.
> > 
> > 	Arnd <><
> 
> 
> What about a pusdown of default_lseek attribution for these
> fops that don't have any llseek() (and rename it to
> deprecated_default_lseek() )
> 
> Because we can probably fix these fops one by one but what
> about the next drivers that will have no llseek() ?
> 
> We can't attribute default_llseek() by default anymore for
> further fops that are to come.
> 
> 

Frederic, I think it is still useful to explicity set to no_llseek, 
drivers that don't use llseek.

I also have to agree with you, that we should no longer be using a 
default_llseek that relies on the BKL.

That is a rather large effort though. All drivers that don't specify an 
llseek function, need to either set it to no_llseek, or as you are 
proposing a deprecated default_llseek that uses the bkl.

thinking of how to start this.

John
--
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/