Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-02-05 Thread Ulf Hansson
On Tue, 5 Feb 2019 at 14:24, Geert Uytterhoeven wrote: > > Hi Ulf, > > On Wed, Jan 30, 2019 at 1:15 PM Ulf Hansson wrote: > > On Wed, 30 Jan 2019 at 12:16, Vincent Guittot > > wrote: > > > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > > The call path is: >

Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-02-05 Thread Geert Uytterhoeven
Hi Ulf, On Wed, Jan 30, 2019 at 1:15 PM Ulf Hansson wrote: > On Wed, 30 Jan 2019 at 12:16, Vincent Guittot > wrote: > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > The call path is: > > change_clocksource > > ... > > write_seqcount_begin > > ...

Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-01-30 Thread Ladislav Michl
On Wed, Jan 30, 2019 at 02:06:07PM +0100, Rafael J. Wysocki wrote: > On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot > wrote: > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > The call path is: > > change_clocksource > > ... > > write_seqcount_begin >

Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-01-30 Thread Ladislav Michl
On Wed, Jan 30, 2019 at 02:18:49PM +0100, Vincent Guittot wrote: > On Wed, 30 Jan 2019 at 14:06, Rafael J. Wysocki wrote: > > > > On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot > > wrote: > > > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > > The call path

Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-01-30 Thread Vincent Guittot
On Wed, 30 Jan 2019 at 14:06, Rafael J. Wysocki wrote: > > On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot > wrote: > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > The call path is: > > change_clocksource > > ... > > write_seqcount_begin > > ...

Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-01-30 Thread Rafael J. Wysocki
On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot wrote: > > A deadlock has been seen when swicthing clocksources which use PM runtime. > The call path is: > change_clocksource > ... > write_seqcount_begin > ... > timekeeping_update > ... > sh_cmt_clocksource_enable

Re: [PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-01-30 Thread Ulf Hansson
On Wed, 30 Jan 2019 at 12:16, Vincent Guittot wrote: > > A deadlock has been seen when swicthing clocksources which use PM runtime. > The call path is: > change_clocksource > ... > write_seqcount_begin > ... > timekeeping_update > ... > sh_cmt_clocksource_enable >

[PATCH v2 ] PM-runtime: fix deadlock with ktime

2019-01-30 Thread Vincent Guittot
A deadlock has been seen when swicthing clocksources which use PM runtime. The call path is: change_clocksource ... write_seqcount_begin ... timekeeping_update ... sh_cmt_clocksource_enable ... rpm_resume pm_runtime_mark_last_b