Re: [PATCH] device-dax: fix sysfs attribute deadlock

2017-05-02 Thread Dan Williams
On Tue, May 2, 2017 at 3:43 AM, Yi Zhang wrote: > Verified this patch on 4.11. > > Tested-by: Yi Zhang Thanks!

Re: [PATCH] device-dax: fix sysfs attribute deadlock

2017-05-02 Thread Dan Williams
On Tue, May 2, 2017 at 3:43 AM, Yi Zhang wrote: > Verified this patch on 4.11. > > Tested-by: Yi Zhang Thanks!

Re: [PATCH] device-dax: fix sysfs attribute deadlock

2017-05-02 Thread Yi Zhang
Zhang" <yiz...@redhat.com> Sent: Sunday, April 30, 2017 10:21:54 PM Subject: [PATCH] device-dax: fix sysfs attribute deadlock Usage of device_lock() for dax_region attributes is unnecessary and deadlock prone. It's unnecessary because the order of registration / un-registration guarant

Re: [PATCH] device-dax: fix sysfs attribute deadlock

2017-05-02 Thread Yi Zhang
ject: [PATCH] device-dax: fix sysfs attribute deadlock Usage of device_lock() for dax_region attributes is unnecessary and deadlock prone. It's unnecessary because the order of registration / un-registration guarantees that drvdata is always valid. It's deadlock prone because it sets up thi

[PATCH] device-dax: fix sysfs attribute deadlock

2017-04-30 Thread Dan Williams
Usage of device_lock() for dax_region attributes is unnecessary and deadlock prone. It's unnecessary because the order of registration / un-registration guarantees that drvdata is always valid. It's deadlock prone because it sets up this situation: ndctl D0 2170 2082 0x

[PATCH] device-dax: fix sysfs attribute deadlock

2017-04-30 Thread Dan Williams
Usage of device_lock() for dax_region attributes is unnecessary and deadlock prone. It's unnecessary because the order of registration / un-registration guarantees that drvdata is always valid. It's deadlock prone because it sets up this situation: ndctl D0 2170 2082 0x