usb: gadget: udc: fotg210-udc: Fix a sleep-in-atomic-context bug in fotg210_get_status()
commit5e57274c63b332cf3a73277db7fc48ada598c2df
authorJia-Ju Bai <baijiaju1990@gmail.com>
Sat, 15 Sep 2018 03:04:40 +0000 (15 11:04 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 24 Nov 2019 07:22:55 +0000 (24 08:22 +0100)
tree058851380259ae4bd790fdb49cbf5a57f589da74
parent78e2b71efdb2eecf216fdc40ea55d6d1a8716169
usb: gadget: udc: fotg210-udc: Fix a sleep-in-atomic-context bug in fotg210_get_status()

[ Upstream commit 2337a77c1cc86bc4e504ecf3799f947659c86026 ]

The driver may sleep in an interrupt handler.
The function call path (from bottom to top) in Linux-4.17 is:

[FUNC] fotg210_ep_queue(GFP_KERNEL)
drivers/usb/gadget/udc/fotg210-udc.c, 744:
fotg210_ep_queue in fotg210_get_status
drivers/usb/gadget/udc/fotg210-udc.c, 768:
fotg210_get_status in fotg210_setup_packet
drivers/usb/gadget/udc/fotg210-udc.c, 949:
fotg210_setup_packet in fotg210_irq (interrupt handler)

To fix this bug, GFP_KERNEL is replaced with GFP_ATOMIC.
If possible, spin_unlock() and spin_lock() around fotg210_ep_queue()
can be also removed.

This bug is found by my static analysis tool DSAC.

Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/usb/gadget/udc/fotg210-udc.c