On 09/20/2017 01:35 PM, Christoph Hellwig wrote: >> +/* >> + * When reading or writing the blktrace sysfs files, the references to the >> + * opened sysfs or device files should prevent the underlying block device >> + * from being removed. So no further delete protection is really needed. >> + * >> + * Protection from multiple readers and writers accessing blktrace data >> + * concurrently is still required. The bd_mutex was used for this purpose. >> + * That could lead to deadlock with concurrent block device deletion and >> + * sysfs access. As a result, a new blk_trace_mutex is now added to be >> + * used solely by the blktrace code. >> + */ > Comments about previous locking schemes really don't have a business > in the code - those are remarks for the commit logs. And in general > please explain the locking scheme near the data that they proctect > it, as locks should always protected data, not code and the comments > should follow that. It seems to be a general practice that we don't put detailed comments in the header files. The comment was put above the function with the first instance of the blk_trace_mutex. Yes, I agree that talking about the past history may not be applicable here. I will keep that in mind in the future. Thanks, Longman