On Tue, Nov 19, 2019 at 04:57:30PM +0800, Jay Fang wrote: > On 2019/11/15 6:39, Bjorn Helgaas wrote: > > On Fri, Sep 06, 2019 at 10:33:58AM +0800, Jay Fang wrote: > >> A deadlock triggered by a NONFATAL AER event during a sysfs "sriov_numvfs" > >> operation: > > > > How often does this happen? Always? Only when an AER event races > > with the sysfs write? > Although not very frequent,the impact is fatal. This bug is very > necessary to be fixed. Absolutely; I didn't mean to suggest that we shouldn't fix it. I only wanted to know how to prioritize it against other issues. Bjorn
