2020-07-08 13:43:00

by Lukas Wunner

[permalink] [raw]
Subject: [PATCH 0/3] Fix races on device removal

Prevent dynamic SPI device addition below a controller which is
being removed. To do so, set the controller's "dead" flag using
kill_device() (patch [3/3]).

Serialize access to a device's "dead" flag with a newly introduced
rw_semaphore in lieu of the device_lock to avoid deadlocks occurring
with the new use case (patch [2/3]).

Add a missing check for the "dead" flag upon driver binding
(patch [1/3]).


Lukas Wunner (3):
driver core: Avoid binding drivers to dead devices
driver core: Use rwsem for kill_device() serialization
driver core: Avoid adding children below a dead parent

drivers/base/base.h | 2 ++
drivers/base/core.c | 49 ++++++++++++++++++++++++++++++--------------
drivers/base/dd.c | 12 ++++++++++-
drivers/nvdimm/bus.c | 8 +-------
drivers/spi/spi.c | 3 +++
5 files changed, 51 insertions(+), 23 deletions(-)

--
2.27.0


2020-07-30 06:55:51

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: [PATCH 0/3] Fix races on device removal

On Wed, Jul 08, 2020 at 03:27:00PM +0200, Lukas Wunner wrote:
> Prevent dynamic SPI device addition below a controller which is
> being removed. To do so, set the controller's "dead" flag using
> kill_device() (patch [3/3]).

Why is the SPI bus allowing this to happen? Don't you have a per-bus
lock that can serialize all of this?

thanks,

greg k-h