When calling into hid_map_usage(), the passed event code is blindly stored as is, even if it doesn't fit in the associated bitmap.
This event code can come from a variety of sources, including devices masquerading as input devices, only a bit more "programmable".
Instead of taking the event code at face value, check that it actually fits the corresponding bitmap, and if it doesn't: - spit out a warning so that we know which device is acting up - NULLify the bitmap pointer so that we catch unexpected uses
Code paths that can make use of untrusted inputs can now check that the mapping was indeed correct and bail out if not.
Cc: stable@vger.kernel.org Signed-off-by: Marc Zyngier maz@kernel.org --- * From v1: - Dropped the input.c changes, and turned hid_map_usage() into the validation primitive. - Handle mapping failures in hidinput_configure_usage() and mt_touch_input_mapping() (on top of hid_map_usage_clear() which was already handled)
drivers/hid/hid-input.c | 4 ++++ drivers/hid/hid-multitouch.c | 2 ++ include/linux/hid.h | 40 +++++++++++++++++++++++++----------- 3 files changed, 34 insertions(+), 12 deletions(-)
diff --git a/drivers/hid/hid-input.c b/drivers/hid/hid-input.c index b8eabf206e74..88e19996427e 100644 --- a/drivers/hid/hid-input.c +++ b/drivers/hid/hid-input.c @@ -1132,6 +1132,10 @@ static void hidinput_configure_usage(struct hid_input *hidinput, struct hid_fiel }
mapped: + /* Mapping failed, bail out */ + if (!bit) + return; + if (device->driver->input_mapped && device->driver->input_mapped(device, hidinput, field, usage, &bit, &max) < 0) { diff --git a/drivers/hid/hid-multitouch.c b/drivers/hid/hid-multitouch.c index 3f94b4954225..e3152155c4b8 100644 --- a/drivers/hid/hid-multitouch.c +++ b/drivers/hid/hid-multitouch.c @@ -856,6 +856,8 @@ static int mt_touch_input_mapping(struct hid_device *hdev, struct hid_input *hi, code = BTN_0 + ((usage->hid - 1) & HID_USAGE);
hid_map_usage(hi, usage, bit, max, EV_KEY, code); + if (!*bit) + return -1; input_set_capability(hi->input, EV_KEY, code); return 1;
diff --git a/include/linux/hid.h b/include/linux/hid.h index 875f71132b14..ff4ccf7ba694 100644 --- a/include/linux/hid.h +++ b/include/linux/hid.h @@ -959,34 +959,49 @@ static inline void hid_device_io_stop(struct hid_device *hid) { * @max: maximal valid usage->code to consider later (out parameter) * @type: input event type (EV_KEY, EV_REL, ...) * @c: code which corresponds to this usage and type + * + * The value pointed to by @bit will be set to NULL if either @type is + * an unhandled event type, or if @c is out of range for @type. This + * can be used as an error condition. */ static inline void hid_map_usage(struct hid_input *hidinput, struct hid_usage *usage, unsigned long **bit, int *max, __u8 type, __u16 c) { struct input_dev *input = hidinput->input; - - usage->type = type; - usage->code = c; + unsigned long *bmap = NULL; + u16 limit = 0;
switch (type) { case EV_ABS: - *bit = input->absbit; - *max = ABS_MAX; + bmap = input->absbit; + limit = ABS_MAX; break; case EV_REL: - *bit = input->relbit; - *max = REL_MAX; + bmap = input->relbit; + limit = REL_MAX; break; case EV_KEY: - *bit = input->keybit; - *max = KEY_MAX; + bmap = input->keybit; + limit = KEY_MAX; break; case EV_LED: - *bit = input->ledbit; - *max = LED_MAX; + bmap = input->ledbit; + limit = LED_MAX; break; } + + if (unlikely(c > limit || !bmap)) { + pr_warn_ratelimited("%s: Invalid code %d type %d\n", + input->name, c, type); + *bit = NULL; + return; + } + + usage->type = type; + usage->code = c; + *max = limit; + *bit = bmap; }
/** @@ -1000,7 +1015,8 @@ static inline void hid_map_usage_clear(struct hid_input *hidinput, __u8 type, __u16 c) { hid_map_usage(hidinput, usage, bit, max, type, c); - clear_bit(c, *bit); + if (*bit) + clear_bit(usage->code, *bit); }
/**
On Wed, 26 Aug 2020, Marc Zyngier wrote:
When calling into hid_map_usage(), the passed event code is blindly stored as is, even if it doesn't fit in the associated bitmap.
This event code can come from a variety of sources, including devices masquerading as input devices, only a bit more "programmable".
Instead of taking the event code at face value, check that it actually fits the corresponding bitmap, and if it doesn't:
- spit out a warning so that we know which device is acting up
- NULLify the bitmap pointer so that we catch unexpected uses
Code paths that can make use of untrusted inputs can now check that the mapping was indeed correct and bail out if not.
Cc: stable@vger.kernel.org Signed-off-by: Marc Zyngier maz@kernel.org
- From v1:
- Dropped the input.c changes, and turned hid_map_usage() into the validation primitive.
- Handle mapping failures in hidinput_configure_usage() and mt_touch_input_mapping() (on top of hid_map_usage_clear() which was already handled)
Benjamin, could you please run this through your regression testing machinery?
It's a non-trivial core change, at the same time I'd like not to postpone it for 5.10 due to its nature.
Thanks,
On 2020-08-27 10:33, Jiri Kosina wrote:
On Wed, 26 Aug 2020, Marc Zyngier wrote:
When calling into hid_map_usage(), the passed event code is blindly stored as is, even if it doesn't fit in the associated bitmap.
This event code can come from a variety of sources, including devices masquerading as input devices, only a bit more "programmable".
Instead of taking the event code at face value, check that it actually fits the corresponding bitmap, and if it doesn't:
- spit out a warning so that we know which device is acting up
- NULLify the bitmap pointer so that we catch unexpected uses
Code paths that can make use of untrusted inputs can now check that the mapping was indeed correct and bail out if not.
Cc: stable@vger.kernel.org Signed-off-by: Marc Zyngier maz@kernel.org
- From v1:
- Dropped the input.c changes, and turned hid_map_usage() into the validation primitive.
- Handle mapping failures in hidinput_configure_usage() and mt_touch_input_mapping() (on top of hid_map_usage_clear() which was already handled)
Benjamin, could you please run this through your regression testing machinery?
It's a non-trivial core change, at the same time I'd like not to postpone it for 5.10 due to its nature.
I found yet another nit that this patch doesn't quite catch. v3 going out in a minute.
Thanks,
M.
On Thu, Aug 27, 2020 at 11:33 AM Jiri Kosina jikos@kernel.org wrote:
On Wed, 26 Aug 2020, Marc Zyngier wrote:
When calling into hid_map_usage(), the passed event code is blindly stored as is, even if it doesn't fit in the associated bitmap.
This event code can come from a variety of sources, including devices masquerading as input devices, only a bit more "programmable".
Instead of taking the event code at face value, check that it actually fits the corresponding bitmap, and if it doesn't:
- spit out a warning so that we know which device is acting up
- NULLify the bitmap pointer so that we catch unexpected uses
Code paths that can make use of untrusted inputs can now check that the mapping was indeed correct and bail out if not.
Cc: stable@vger.kernel.org Signed-off-by: Marc Zyngier maz@kernel.org
- From v1:
- Dropped the input.c changes, and turned hid_map_usage() into the validation primitive.
- Handle mapping failures in hidinput_configure_usage() and mt_touch_input_mapping() (on top of hid_map_usage_clear() which was already handled)
Benjamin, could you please run this through your regression testing machinery?
It's a non-trivial core change, at the same time I'd like not to postpone it for 5.10 due to its nature.
OK, just passed the v4 to the testsuite, and this seems good. It won't break touchscreens nor keyboards/mice that needed to be added in the past.
So this is a go for me.
Cheers, Benjamin
Thanks,
-- Jiri Kosina SUSE Labs
linux-stable-mirror@lists.linaro.org