Oops in current tree in i2c
From: Linus Torvalds
Date: Sat Oct 27 2018 - 12:08:42 EST
On my laptop I'm getting a kernel page fault with the current git
tree, and I'm tentatively blaming commit
9ee3e06610fd ("HID: i2c-hid: override HID descriptors for certain devices")
but that's simply because it's the only thing that seems to touch this
particular area in this merge window.
The oops looks like this:
BUG: unable to handle kernel paging request at 000000007a25d598
PGD 0 P4D 0
Oops: 0000 [#1] SMP PTI
CPU: 1 PID: 888 Comm: systemd-udevd Not tainted 4.19.0-07715-g345671ea0f92 #4
Hardware name: Dell Inc. XPS 13 9350/09JHRY, BIOS 1.7.0 01/16/2018
where the code disassembly (and the register contents) shows that the
wild pointer is the first argument to "strstr()", which just has a
bogus value that is not a valid kernel pointer (RDI: 000000007a25d598
- which is obviously also the address of the page fault)
The call trace is:
which is why I suspect that new i2c_hid_get_dmi_hid_report_desc_override() code.
I *think* the problem is that the i2c_hid_dmi_desc_override_table
isn't terminated by a NULL entry, and I will test that next.
What makes me *very* unhappy about this is that if I'm right, I think
it means that code was literally not tested at all by anybody who
didn't have one of the entries in that list.