On Sun, 2018-09-09 at 20:58 +-0800, Ming Lei wrote: +AD4 Now percpu+AF8-ref+AF8-reinit() can only be done on one percpu refcounter +AD4 when it drops zero. And the limit shouldn't be so strict, and it +AD4 is quite straightforward that percpu+AF8-ref+AF8-reinit() can be done when +AD4 this counter is at atomic mode. +AD4 +AD4 This patch relaxes the limit, so we may avoid extra change+AFs-1+AF0 for NVMe +AD4 timeout's requirement. +AD4 +AD4 +AFs-1+AF0 https://marc.info/?l+AD0-linux-kernel+ACY-m+AD0-153612052611020+ACY-w+AD0-2 Is the NVMe driver the only block driver that hangs if it is attempted to freeze its request queue when a request times out? If so, can this hang be fixed by modifying the NVMe driver instead of by modifying the percpu refcount implementation? Thanks, Bart.
