panic in 4.9-stable

[Mikhail Teterin]

The following panic hits seemingly at random, but persistent (once a
week on average):

	Fatal trap 12: page fault while in vm86 mode
	fault virtual address = 0x6400
	fault code            = user read, page not present
	instruction pointer   = 0x200:0x4400
	stack pointer         = 0x0:0xddb50ff8
	frame pointer         = 0x0:0x57
	code segment          = base 0xc01cc039, limit 0x99830, type 0x18
			      = DPL 0, pres 1, def32 0, gran 0
	processor eflags      = interrupt enabled, resume, vm86, IOPL = 0
	current process               = 3824 (XFree86)
	interrupt mask                = none
	trap number           = 12
	panic: page fault

	syncing disks... 36 9 5 3 
	done
	Uptime: 23m28s
	Terminate ACPI

The machine runs 4.9-stable from March 4th. It has a dual NVidia card
with two monitors at 1024x768 resolutions -- perhaps, some resource
shortage in NVidia's driver? The code segment seems to point to shmat(),
according to the sorted nm output fragment below (crash address inserted):

	[...]
	c01bfd08 t shm_delete_mapping
	c01bfd8c T shmdt
	c01bfe00 T shmat
-->	c01cc039 -- the code segment base from the panic above
	c01cc0e4 T sokqfilter
	c01cc15c t filt_sordetach
	c01cc1b4 t filt_soread
	c01cc214 t filt_sowdetach
	[...]

Any comments? Thanks!

	-mi