From: Tang Junhui <tang.junhui@xxxxxxxxxx> Hello Eric > > > > From my personal opinion, updating bucket_in_use is for acting garbage > > collection. If number of bucket in use is not updated in time, garbage > > collection won't start due to old bucket_in_use still beyond > > CUTOFF_WRITEBACK_SYNC. > > > > We may maintain an atomic counter per-cache set for dirty buckets, and > > update it at some locations when allocating or reclaiming bucket. This > > counter is unnecessary to be very accurate, just accurate enough for > > should_writeback() working correctly. > > > > I am also looking at it for a better solution as well. > > Hi Coli & Tang, > > Have either of you had a chance to come up with a solution to this? This patch is updated by [PATCH] [PATCH V2] bcache: update bucket_in_use in real time and it has been applied to Mike's tree. Regards, Tang
