On 4/26/22 13:38, Greg Kroah-Hartman wrote:
On Tue, Apr 26, 2022 at 12:24:24PM -0500, John Donnelly wrote:
On 4/26/22 12:52 PM, Greg Kroah-Hartman wrote:
On Tue, Apr 26, 2022 at 12:37:21PM -0500, john.p.donnelly@oracle.com wrote:
76723ed1fb89 2021-12-01 | locking/rwsem: Make handoff bit handling more consistent
In Linux 5.15.y.
That commit is in 5.15.6, released December 1, 2021. And this just now shows up? How is this related to 5.15.36-rc1?
Hi,
This was briefly discussed in :
Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent
Have a lore.kernel.org link? Why not continue it there?
https://lore.kernel.org/all/eae41639-cbca-4ea6-417f-f9b34a7138ea@oracle.com/
I tried and the thread is quiet.
Additional testing shows the rwsem hang still exists. It takes a 24hr fio soak test to show up.
It likely still exists in Linux 5.18.y too. We will be testing that in the future as time permits.
Can you test 5.17.y also as there is no 5.18.y yet :)
We did. It failed there too . ;-) We happened to have 5.18.rc2 in the work queue
thanks,
greg k-h