Messages in this thread | | | Date | Mon, 23 Feb 2015 12:33:13 -0500 (EST) | From | Nicolas Pitre <> | Subject | Re: [PATCH 32/35] clockevents: Fix cpu down race for hrtimer based broadcasting |
| |
On Mon, 23 Feb 2015, Nicolas Pitre wrote:
> On Mon, 23 Feb 2015, Peter Zijlstra wrote: > > > The reported function that fails: bL_switcher_restore_cpus() is called > > in the error paths of the former and the main path in the latter to make > > the 'stolen' cpus re-appear. > > > > The patch in question somehow makes that go boom. > > > > > > Now what all do you need to do to make it go boom? Just enable/disable > > the switcher once and it'll explode? Or does it need to do actual > > switches while it is enabled? > > It gets automatically enabled during boot. Then several switches are > performed while user space is brought up. If I manually disable it > via /sys then it goes boom.
OK. Forget the bL switcher. I configured it out of my kernel and then managed to get the same crash by simply hotplugging out one CPU and plugging it back in.
$ echo 0 > /sys/devices/system/cpu/cpu2/online [CPU2 gone] $ echo 1 > /sys/devices/system/cpu/cpu2/online [Boom!]
Nicolas
| |