releases/4.19.53/hid-multitouch-handle-faulty-elo-touch-device.patch

   1 From 81bcbad53bab4bf9f200eda303d7a05cdb9bd73b Mon Sep 17 00:00:00 2001
   2 From: Benjamin Tissoires <benjamin.tissoires@redhat.com>
   3 Date: Tue, 21 May 2019 15:38:31 +0200
   4 Subject: HID: multitouch: handle faulty Elo touch device
   5
   6 From: Benjamin Tissoires <benjamin.tissoires@redhat.com>
   7
   8 commit 81bcbad53bab4bf9f200eda303d7a05cdb9bd73b upstream.
   9
  10 Since kernel v5.0, one single win8 touchscreen device failed.
  11 And it turns out this is because it reports 2 InRange usage per touch.
  12
  13 It's a first, and I *really* wonder how this was allowed by Microsoft in
  14 the first place. But IIRC, Breno told me this happened *after* a firmware
  15 upgrade...
  16
  17 Anyway, better be safe for those crappy devices, and make sure we have
  18 a full slot before jumping to the next.
  19 This won't prevent all crappy devices to fail here, but at least we will
  20 have a safeguard as long as the contact ID and the X and Y coordinates
  21 are placed in the report after the grabage.
  22
  23 Fixes: 01eaac7e5713 ("HID: multitouch: remove one copy of values")
  24 CC: stable@vger.kernel.org # v5.0+
  25 Reported-and-tested-by: Breno Leitao <leitao@debian.org>
  26 Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
  27 Signed-off-by: Jiri Kosina <jkosina@suse.cz>
  28 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
  29
  30 diff --git a/drivers/hid/hid-multitouch.c b/drivers/hid/hid-multitouch.c
  31 index c02d4cad1893..1565a307170a 100644
  32 --- a/drivers/hid/hid-multitouch.c
  33 +++ b/drivers/hid/hid-multitouch.c
  34 @@ -641,6 +641,13 @@ static void mt_store_field(struct hid_device *hdev,
  35         if (*target != DEFAULT_TRUE &&
  36             *target != DEFAULT_FALSE &&
  37             *target != DEFAULT_ZERO) {
  38 +               if (usage->contactid == DEFAULT_ZERO ||
  39 +                   usage->x == DEFAULT_ZERO ||
  40 +                   usage->y == DEFAULT_ZERO) {
  41 +                       hid_dbg(hdev,
  42 +                               "ignoring duplicate usage on incomplete");
  43 +                       return;
  44 +               }
  45                 usage = mt_allocate_usage(hdev, application);
  46                 if (!usage)
  47                         return;