On Thu, 21 Sep 2017 17:38:53 +0200, Andrey Konovalov wrote:
Hi!
I've got the following report while fuzzing the kernel with syzkaller.
On commit ebb2c2437d8008d46796902ff390653822af6cc4 (Sep 18).
------------[ cut here ]------------ WARNING: CPU: 1 PID: 1846 at mm/page_alloc.c:3883 __alloc_pages_slowpath+0x1ef2/0x2d70 Modules linked in: CPU: 1 PID: 1846 Comm: kworker/1:2 Not tainted 4.14.0-rc1-42251-gebb2c2437d80 #215 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Workqueue: usb_hub_wq hub_event task: ffff8800643c18c0 task.stack: ffff88006b658000 RIP: 0010:trace_reclaim_retry_zone ./include/trace/events/oom.h:31 RIP: 0010:should_reclaim_retry mm/page_alloc.c:3783 RIP: 0010:__alloc_pages_slowpath+0x1ef2/0x2d70 mm/page_alloc.c:4039 RSP: 0018:ffff88006b65d938 EFLAGS: 00010246 RAX: 00000000ffffa666 RBX: 00000000014000c0 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000034 RDI: 000000000140c0c0 RBP: ffff88006b65e088 R08: 0000000000000000 R09: fffffffffff00f88 R10: 0000000000000000 R11: 0000000000000085 R12: ffff88006b65e130 R13: ffff88006b65e270 R14: ffff88006b65e0f0 R15: 000000000140c0c0 FS: 0000000000000000(0000) GS:ffff88006c900000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000020829000 CR3: 0000000063cce000 CR4: 00000000000006e0 Call Trace: __alloc_pages_nodemask+0x921/0xf70 mm/page_alloc.c:4217 alloc_pages_current+0xbb/0x1f0 mm/mempolicy.c:2035 alloc_pages ./include/linux/gfp.h:505 __get_free_pages+0x14/0x50 mm/page_alloc.c:4248 usb_stream_new+0x50f/0x9f0 sound/usb/usx2y/usb_stream.c:214
The warning itself should be harmless, indicating only that the driver tries to allocate too high-order memory pages. The error path handles the allocation error gracefully, so basically we can suppress the warning by adding __GFP_NOWARN, in addition to a sanity check in the caller side.
I've been traveling and will be traveling again in the next week, so I'll cook up once after back to work again.
thanks,
Takashi