Message-ID: <520f0cf11002011350u3b541a0cxfb0ed882dca13afe@mail.gmail.com>
Date:	Mon, 1 Feb 2010 22:50:25 +0100
From:	John Kacur <jkacur@...il.com>
To:	Dmitry Torokhov <dmitry.torokhov@...il.com>
Cc:	Arnd Bergmann <arnd@...db.de>,
	Thadeu Lima de Souza Cascardo <cascardo@...oscopio.com>,
	linux-input@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] input: remove BKL from uinput open function

On Mon, Feb 1, 2010 at 10:21 PM, Dmitry Torokhov
<dmitry.torokhov@...il.com> wrote:
> On Mon, Feb 01, 2010 at 09:27:22PM +0100, John Kacur wrote:
>> On Mon, Feb 1, 2010 at 9:22 PM, John Kacur <jkacur@...hat.com> wrote:
>> > On Sun, Jan 31, 2010 at 6:29 AM, Dmitry Torokhov
>> > <dmitry.torokhov@...il.com> wrote:
>> >> On Sun, Jan 31, 2010 at 05:20:55AM +0100, Arnd Bergmann wrote:
>> >>> On Sunday 31 January 2010, John Kacur wrote:
>> >>> > > Sorry, I should have been clearer, but not implementing llseek
>> >>> > > is the problem I was referring to: When a driver has no explicit
>> >>> > > .llseek operation in its file operations and does not call
>> >>> > > nonseekable_open from its open operation, the VFS layer will
>> >>> > > implicitly use default_llseek, which takes the BKL. We're
>> >>> > > in the process of changing drivers not to do this, one by one
>> >>> > > so we can kill the BKL in the end.
>> >>> > >
>> >>> >
>> >>> > I know we've discussed this before, but why wouldn't the following
>> >>> > make more sense?
>> >>> >  .llseek         = no_llseek,
>> >>>
>> >>> That's one of the possible solutions. Assigning it to generic_file_llseek
>> >>> also gets rid of the BKL but keeps the current behaviour (calling seek
>> >>> returns success without having an effect, no_llseek returns -ESPIPE),
>> >>> while calling nonseekable_open has the other side-effect of making
>> >>> pread/pwrite fail with -ESPIPE, which is more consistent than
>> >>> only failing seek.
>> >>>
>> >>
>> >> OK, so how about the patch below (on top of Thadeu's patch)?
>> >>
>> >> --
>> >> Dmitry
>> >>
>> >> Input: uinput - use nonseekable_open
>> >>
>> >> Seeking does not make sense for uinput so let's use nonseekable_open
>> >> to mark the device non-seekable.
>> >>
>> >> Signed-off-by: Dmitry Torokhov <dtor@...l.ru>
>> >> ---
>> >>
>> >>  drivers/input/misc/uinput.c |    7 +++++++
>> >>  1 files changed, 7 insertions(+), 0 deletions(-)
>> >>
>> >>
>> >> diff --git a/drivers/input/misc/uinput.c b/drivers/input/misc/uinput.c
>> >> index 18206e1..7089151 100644
>> >> --- a/drivers/input/misc/uinput.c
>> >> +++ b/drivers/input/misc/uinput.c
>> >> @@ -278,6 +278,7 @@ static int uinput_create_device(struct uinput_device *udev)
>> >>  static int uinput_open(struct inode *inode, struct file *file)
>> >>  {
>> >>        struct uinput_device *newdev;
>> >> +       int error;
>> >>
>> >>        newdev = kzalloc(sizeof(struct uinput_device), GFP_KERNEL);
>> >>        if (!newdev)
>> >> @@ -291,6 +292,12 @@ static int uinput_open(struct inode *inode, struct file *file)
>> >>
>> >>        file->private_data = newdev;
>> >>
>> >> +       error = nonseekable_open(inode, file);
>> >> +       if (error) {
>> >> +               kfree(newdev);
>> >> +               return error;
>> >> +       }
>> >> +
>> >>        return 0;
>> >>  }
>> >>
>> >>
>> >
>> > Hmnn, if you look at nonseekable_open() it will always return 0. I
>> > think you can just do the following.
>
> It always returns 0 _now_ but I do not see any guarantees that it will
> never ever return anything but 0. If somebody would provide such
> garantee then we certainly would not need to handle errors.

Well, all it's doing is changing the f_mode. If anyone ever changes
that function
to return anything other than 0 it will be their responsibility to go
fix all the
uses of it. If you do a git grep of nonseekable_open, you'll see that this
is a very common paradigm. (to return 0). It makes your code shorter,
and more readable. Plus, you are writing speculative code based on
what might exist in the future? Also, then should uinput_release be called?
If it is called will kfree be called twice on the same memory. If it
isn't called, is
that a problem because you've already done most of the work that requires
a call to uinput_destroy_device ?

>
>> >
>> > diff --git a/drivers/input/misc/uinput.c b/drivers/input/misc/uinput.c
>> > index 18206e1..697c0a6 100644
>> > --- a/drivers/input/misc/uinput.c
>> > +++ b/drivers/input/misc/uinput.c
>> > @@ -291,7 +291,7 @@ static int uinput_open(struct inode *inode, struct file *fil
>> >
>> >        file->private_data = newdev;
>> >
>> > -       return 0;
>> > +       return nonseekable_open(inode, file);
>> >  }
>> >
>> > Signed-off-by: John Kacur <jkacur@...hat.com>
>> >
>>
>> Btw, Thadeu Lima de Souza Cascardo just combine that all into
>> one patch, no point really in making two patches out of that.
>
> I think these are 2 separate changes (the fact that nonseekable_open
> also gets rid of BKL invocation is a side-effect), that is not
> considering the fact that I already applied Thadeu's change and don't
> want to rewind my public branch unless really necessary.

Yeah, I agree, it's a PITA to rewind a public branch, in fact, you
should revert it
if it necessary. But, no worries, it's not necessary here.
However, generally, when you get rid of the BKL you do it in all
functions at once
as Arnd points out in another mail.

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/

Powered by blists - more mailing lists