2018-06-20 02:50:57

by Jia-Ju Bai

Subject: [BUG] xen: Two possible sleep-in-atomic-context bugs in bind_evtchn_to_irqhandler()

The driver may sleep with holding a spinlock.
The function call paths (from bottom to top) in Linux-4.16.7 are:

[FUNC] mutex_lock_nested --> can sleep
drivers/xen/events/events_base.c, 839: mutex_lock_nested in
bind_evtchn_to_irq
drivers/xen/events/events_base.c, 1030: bind_evtchn_to_irq in
bind_evtchn_to_irqhandler
drivers/xen/pvcalls-front.c, 371: bind_evtchn_to_irqhandler in create_active
drivers/xen/pvcalls-front.c, 417: create_active in pvcalls_front_connect
drivers/xen/pvcalls-front.c, 410: spin_lock in pvcalls_front_connect

[FUNC] request_irq --> can sleep
drivers/xen/events/events_base.c, 1003: request_irq in
bind_evtchn_to_irqhandler
drivers/xen/pvcalls-front.c, 371: bind_evtchn_to_irqhandler in create_active
drivers/xen/pvcalls-front.c, 417: create_active in pvcalls_front_connect
drivers/xen/pvcalls-front.c, 410: spin_lock in pvcalls_front_connect

These bugs are found by my static analysis tool (DSAC-2) and checked by my
code review.

I do not know how to correctly fix these bugs, so I just report them.
Maybe create_active() should not be called with holding a spinlock.


Best wishes,
Jia-Ju Bai

2018-06-20 09:15:48

by Jürgen Groß

Subject: Re: [BUG] xen: Two possible sleep-in-atomic-context bugs in bind_evtchn_to_irqhandler()

On 20/06/18 04:49, Jia-Ju Bai wrote:
> The driver may sleep with holding a spinlock.
> The function call paths (from bottom to top) in Linux-4.16.7 are:
>
> [FUNC] mutex_lock_nested --> can sleep
> drivers/xen/events/events_base.c, 839: mutex_lock_nested in
> bind_evtchn_to_irq
> drivers/xen/events/events_base.c, 1030: bind_evtchn_to_irq in
> bind_evtchn_to_irqhandler
> drivers/xen/pvcalls-front.c, 371: bind_evtchn_to_irqhandler in
> create_active
> drivers/xen/pvcalls-front.c, 417: create_active in pvcalls_front_connect
> drivers/xen/pvcalls-front.c, 410: spin_lock in pvcalls_front_connect
>
> [FUNC] request_irq --> can sleep
> drivers/xen/events/events_base.c, 1003: request_irq in
> bind_evtchn_to_irqhandler
> drivers/xen/pvcalls-front.c, 371: bind_evtchn_to_irqhandler in
> create_active
> drivers/xen/pvcalls-front.c, 417: create_active in pvcalls_front_connect
> drivers/xen/pvcalls-front.c, 410: spin_lock in pvcalls_front_connect
>
> These bugs are found by my static analysis tool (DSAC-2) and checked by my
> code review.
>
> I do not know how to correctly fix these bugs, so I just report them.
> Maybe create_active() should not be called with holding a spinlock.

Right, I think calling create_active() should be done before taking the
lock. Stefano, what do you think?


Juergen