Affected: Linux

In the Linux kernel, the following vulnerability has been resolved:

HID: core: fix number/pointer type confusion on long items

When fetch_item() is called by hid_scan_report() on an item with HID_ITEM_TAG_LONG, it stores a pointer to the item data in item->data.longdata instead of storing a value directly in item->data.{u8/u16/u32}.

When item_udata() or item_sdata() encounters such an item, it incorrectly assumes that the item is in short format, and therefore returns the lower part of a kernel pointer reinterpreted as a number.

When a HID device is connected whose descriptor contains a HID_GLOBAL_ITEM_TAG_REPORT_SIZE encoded in long format with size=4, this causes the lower half of a kernel pointer to be printed into dmesg as a number, like this:

hid (null): invalid report_size 107953555

To fix it, let item_udata() and item_sdata() verify that the item is in short format.

Note that this bug only affects hid_scan_report(), while the main parsing pass hid_parse_collections() will always bail out when encountering a long item.

Sidenote: There are currently no users of data.longdata; maybe we should just remove any parsing of long-format descriptors as a follow-up.