On Sun, 7 Jun 2020 15:27:49 +0800 kernel test robot <[email protected]>
wrote:
> Greeting,
>
> FYI, we noticed the following commit (built with gcc-9):
>
> commit: 24256c140e59c3431af4918fd4ec892102cbc2f3 ("x86: mm: ptdump: calculate
> effective permissions correctly")
> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
>
> in testcase: boot
>
> on test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 8G
>
> caused below changes (please refer to attached dmesg/kmsg for entire
> log/backtrace):
>
>
>
>
> If you fix the issue, kindly add following tag
> Reported-by: kernel test robot <[email protected]>
>
>
> [ 8.785395] UBSAN: array-index-out-of-bounds in
> arch/x86/mm/dump_pagetables.c:285:27
> [ 8.787718] index -1 is out of range for type 'pgprotval_t [5]'
> [ 8.789278] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 5.7.0-rc7-00027-g24256c140e59c #1
> [ 8.791350] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
> 1.12.0-1 04/01/2014
> [ 8.793505] Call Trace:
> [ 8.794221] dump_stack+0x6c/0x94
> [ 8.795140] ubsan_epilogue+0x8/0x3e
> [ 8.796132] __ubsan_handle_out_of_bounds+0x48/0x53
> [ 8.797362] ? is_visible+0x2b/0x32
> [ 8.798231] ? atomic_try_cmpxchg+0x4c/0x57
> [ 8.799203] note_page+0x37/0x80b
> [ 8.800138] ? __up_read+0x352/0x37b
> [ 8.801124] ptdump_walk_pgd+0x57/0x60
> [ 8.802145] ptdump_walk_pgd_level_core+0x72/0x9f
> [ 8.803369] ? ptdump_walk_pgd_level_debugfs+0x18/0x18
> [ 8.804613] ? ptdump_walk_pgd_level_core+0x9f/0x9f
> [ 8.805779] ptdump_walk_pgd_level_checkwx+0x1d/0x21
> [ 8.806873] mark_rodata_ro+0x9f/0xa6
> [ 8.807821] ? rest_init+0x7f/0x164
> [ 8.808808] kernel_init+0x44/0xf7
> [ 8.809735] ? rest_init+0x164/0x164
> [ 8.810708] ret_from_fork+0x19/0x24
Yup, thanks, the new note_page() doesn't handle level==-1, which is
"unknown" according to the comment in include/linux/ptdump.h.
Steven, could you please take a look?
