Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1571200 > unrolled thread

[PATCH] Input: uinput - fix crash when mixing old and new init style

Back to article view | Back to linux.kernel

Contents

  [PATCH] Input: uinput - fix crash when mixing old and new init style Dmitry Torokhov <dmitry.torokhov@gmail.com> - 2017-02-01 00:20 +0100
    Re: [PATCH] Input: uinput - fix crash when mixing old and new init  style Benjamin Tissoires <benjamin.tissoires@redhat.com> - 2017-02-01 09:50 +0100
      Re: [PATCH] Input: uinput - fix crash when mixing old and new init  style Dmitry Torokhov <dmitry.torokhov@gmail.com> - 2017-02-01 18:30 +0100

#1571200 — [PATCH] Input: uinput - fix crash when mixing old and new init style

If user tries to initialize uinput device mixing old and new style
initialization (i.e. using old UI_SET_ABSBIT instead of UI_ABS_SETUP,
we forget to allocate input->absinfo and will crash when trying to send
absolute events:

        ioctl(ui, UI_DEV_SETUP, &us);
        ioctl(ui, UI_SET_EVBIT, EV_ABS);
        ioctl(ui, UI_SET_ABSBIT, ABS_X);
        ioctl(ui, UI_SET_ABSBIT, ABS_Y);
        ioctl(ui, UI_DEV_CREATE, 0);

Reported-by: Rodrigo Rivas Costa <rodrigorivascosta@gmail.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=191811
Fixes: fbae10db0940 ("Input: uinput - rework ABS validation")
Cc: stable@vger.kernel.org
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
 drivers/input/misc/uinput.c | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/drivers/input/misc/uinput.c b/drivers/input/misc/uinput.c
index 92595b98e7ed..022be0e22eba 100644
--- a/drivers/input/misc/uinput.c
+++ b/drivers/input/misc/uinput.c
@@ -263,13 +263,21 @@ static int uinput_create_device(struct uinput_device *udev)
 		return -EINVAL;
 	}
 
-	if (test_bit(ABS_MT_SLOT, dev->absbit)) {
-		nslot = input_abs_get_max(dev, ABS_MT_SLOT) + 1;
-		error = input_mt_init_slots(dev, nslot, 0);
-		if (error)
+	if (test_bit(EV_ABS, dev->evbit)) {
+		input_alloc_absinfo(dev);
+		if (!dev->absinfo) {
+			error = -EINVAL;
 			goto fail1;
-	} else if (test_bit(ABS_MT_POSITION_X, dev->absbit)) {
-		input_set_events_per_packet(dev, 60);
+		}
+
+		if (test_bit(ABS_MT_SLOT, dev->absbit)) {
+			nslot = input_abs_get_max(dev, ABS_MT_SLOT) + 1;
+			error = input_mt_init_slots(dev, nslot, 0);
+			if (error)
+				goto fail1;
+		} else if (test_bit(ABS_MT_POSITION_X, dev->absbit)) {
+			input_set_events_per_packet(dev, 60);
+		}
 	}
 
 	if (test_bit(EV_FF, dev->evbit) && !udev->ff_effects_max) {
-- 
2.11.0.483.g087da7b7c-goog

-- 
Dmitry

[toc] | [next] | [standalone]

#1571332 — Re: [PATCH] Input: uinput - fix crash when mixing old and new init style

On Jan 31 2017 or thereabouts, Dmitry Torokhov wrote:
> If user tries to initialize uinput device mixing old and new style
> initialization (i.e. using old UI_SET_ABSBIT instead of UI_ABS_SETUP,
> we forget to allocate input->absinfo and will crash when trying to send
> absolute events:
> 
>         ioctl(ui, UI_DEV_SETUP, &us);
>         ioctl(ui, UI_SET_EVBIT, EV_ABS);
>         ioctl(ui, UI_SET_ABSBIT, ABS_X);
>         ioctl(ui, UI_SET_ABSBIT, ABS_Y);
>         ioctl(ui, UI_DEV_CREATE, 0);

Shouldn't we prevent completely the mix of the 2 versions? If the user
used UI_DEV_SETUP, we should probably prevent the use of UI_SET_ABSBIT
afterwards, or at least warn it somehow.

But the patch in itself doesn't hurt, so:
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>

Cheers,
Benjamin

> 
> Reported-by: Rodrigo Rivas Costa <rodrigorivascosta@gmail.com>
> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=191811
> Fixes: fbae10db0940 ("Input: uinput - rework ABS validation")
> Cc: stable@vger.kernel.org
> Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
> ---
>  drivers/input/misc/uinput.c | 20 ++++++++++++++------
>  1 file changed, 14 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/input/misc/uinput.c b/drivers/input/misc/uinput.c
> index 92595b98e7ed..022be0e22eba 100644
> --- a/drivers/input/misc/uinput.c
> +++ b/drivers/input/misc/uinput.c
> @@ -263,13 +263,21 @@ static int uinput_create_device(struct uinput_device *udev)
>  		return -EINVAL;
>  	}
>  
> -	if (test_bit(ABS_MT_SLOT, dev->absbit)) {
> -		nslot = input_abs_get_max(dev, ABS_MT_SLOT) + 1;
> -		error = input_mt_init_slots(dev, nslot, 0);
> -		if (error)
> +	if (test_bit(EV_ABS, dev->evbit)) {
> +		input_alloc_absinfo(dev);
> +		if (!dev->absinfo) {
> +			error = -EINVAL;
>  			goto fail1;
> -	} else if (test_bit(ABS_MT_POSITION_X, dev->absbit)) {
> -		input_set_events_per_packet(dev, 60);
> +		}
> +
> +		if (test_bit(ABS_MT_SLOT, dev->absbit)) {
> +			nslot = input_abs_get_max(dev, ABS_MT_SLOT) + 1;
> +			error = input_mt_init_slots(dev, nslot, 0);
> +			if (error)
> +				goto fail1;
> +		} else if (test_bit(ABS_MT_POSITION_X, dev->absbit)) {
> +			input_set_events_per_packet(dev, 60);
> +		}
>  	}
>  
>  	if (test_bit(EV_FF, dev->evbit) && !udev->ff_effects_max) {
> -- 
> 2.11.0.483.g087da7b7c-goog
> 
> -- 
> Dmitry

[toc] | [prev] | [next] | [standalone]

#1571810 — Re: [PATCH] Input: uinput - fix crash when mixing old and new init style

On Wed, Feb 01, 2017 at 09:49:25AM +0100, Benjamin Tissoires wrote:
> On Jan 31 2017 or thereabouts, Dmitry Torokhov wrote:
> > If user tries to initialize uinput device mixing old and new style
> > initialization (i.e. using old UI_SET_ABSBIT instead of UI_ABS_SETUP,
> > we forget to allocate input->absinfo and will crash when trying to send
> > absolute events:
> > 
> >         ioctl(ui, UI_DEV_SETUP, &us);
> >         ioctl(ui, UI_SET_EVBIT, EV_ABS);
> >         ioctl(ui, UI_SET_ABSBIT, ABS_X);
> >         ioctl(ui, UI_SET_ABSBIT, ABS_Y);
> >         ioctl(ui, UI_DEV_CREATE, 0);
> 
> Shouldn't we prevent completely the mix of the 2 versions? If the user
> used UI_DEV_SETUP, we should probably prevent the use of UI_SET_ABSBIT
> afterwards, or at least warn it somehow.

Why? If user does not wish to set up parameters of an axis then just
using UI_SET_ABSBIT is fine. In fact, I think it would be weird if we
allowed all UI_SET_*BIT except for ABS.

> 
> But the patch in itself doesn't hurt, so:
> Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>

Thanks.

-- 
Dmitry

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web