[PATCH] Input: Fix double equality sign in uevent

[PATCH] Input: Fix double equality sign in uevent

[Henrik Rydberg]

Looking at the uevent stream for input devices, all properties
are on the form "A=B" except the bitmap values, which are on the
form "A==B". Remedied with this patch.

Cc: stable@kernel.org
Signed-off-by: Henrik Rydberg <rydberg@euromail.se>
---
Hi Dmitry,

Unless the input bitmap properties should really behave differently,
this patch ought to be a candidate for 2.6.37.

Thanks,
Henrik

 drivers/input/input.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/input/input.c b/drivers/input/input.c
index c3e5fc7..bed725f 100644
--- a/drivers/input/input.c
+++ b/drivers/input/input.c
@@ -1471,7 +1471,7 @@ static int input_add_uevent_bm_var(struct kobj_uevent_env *env,
 {
 	int len;
 
-	if (add_uevent_var(env, "%s=", name))
+	if (add_uevent_var(env, "%s", name))
 		return -ENOMEM;
 
 	len = input_print_bitmap(&env->buf[env->buflen - 1],
-- 
1.7.2.3

Re: [PATCH] Input: Fix double equality sign in uevent

[Dmitry Torokhov]

On Sat, Dec 18, 2010 at 08:31:57PM +0100, Henrik Rydberg wrote:
> Looking at the uevent stream for input devices, all properties
> are on the form "A=B" except the bitmap values, which are on the
> form "A==B". Remedied with this patch.
> 
> Cc: stable@kernel.org
> Signed-off-by: Henrik Rydberg <rydberg@euromail.se>
> ---
> Hi Dmitry,
> 
> Unless the input bitmap properties should really behave differently,
> this patch ought to be a candidate for 2.6.37.
> 

Hmm, I do not think so because:

- either nobody uses bitmap variables and then it is not priority
- users worked around the problem and by applying fix we'd be
  breaking their setups.

I think we should apply it but at the beginning of .38 cycle to give us
a chance to evaluate and maybe revert.

Does this make sense?

-- 
Dmitry

Re: [PATCH] Input: Fix double equality sign in uevent

[Henrik Rydberg]

On Sun, Dec 19, 2010 at 12:04:28AM -0800, Dmitry Torokhov wrote:
> On Sat, Dec 18, 2010 at 08:31:57PM +0100, Henrik Rydberg wrote:
> > Looking at the uevent stream for input devices, all properties
> > are on the form "A=B" except the bitmap values, which are on the
> > form "A==B". Remedied with this patch.
> > 
> > Cc: stable@kernel.org
> > Signed-off-by: Henrik Rydberg <rydberg@euromail.se>
> > ---
> > Hi Dmitry,
> > 
> > Unless the input bitmap properties should really behave differently,
> > this patch ought to be a candidate for 2.6.37.
> > 
> 
> Hmm, I do not think so because:
> 
> - either nobody uses bitmap variables and then it is not priority
> - users worked around the problem and by applying fix we'd be
>   breaking their setups.
> 
> I think we should apply it but at the beginning of .38 cycle to give us
> a chance to evaluate and maybe revert.
> 
> Does this make sense?

Indeed, since the problem has been around at least since 2007, it makes perfect sense.

Thanks,
Henrik