On Tue, Nov 28, 2017 at 10:06:24AM +0100, Greg KH wrote: > On Mon, Nov 27, 2017 at 09:35:02AM -0800, Michael Lyle wrote: > > On Mon, Nov 27, 2017 at 9:01 AM, Jens Axboe <axboe@xxxxxxxxx> wrote: > > > > > On 11/27/2017 09:45 AM, Eddie Chapman wrote: > > > > On 27/11/17 16:06, gregkh@xxxxxxxxxxxxxxxxxxx wrote: > > > >> > > > >> This is a note to let you know that I've just added the patch titled > > > >> > > > >> bcache: only permit to recovery read error when cache device is > > > clean > > > > > > > > > > > Hi Greg, > > > > > > > > This patch is an important fix for a possible data corruption issue, but > > > > it also introduces a bug that can produce read errors under certain > > > > circumstances. The issue introduced by this patch is not as severe as > > > > the issue it fixes, but can lead to e.g. upper layer fs remounting read > > > > only. In my environment upper layer handled it badly and indirectly > > > > resulted in data loss (not directly bcache's fault of course). > > > > > > > > Michael Lyle CC'd the stable list with a follow up fix for the issue > > > > introduced by this patch, on 24th Nov, subject "bcache: recover data > > > > from backing when data is clean". > > > > > > > > However, the followup fix is not in Linus' tree yet, only in Michael's. > > > > I guess that means you can't pick it up yet. Never-the-less I felt it > > > > important to point this out here. > > > > > > It's commit e393aa2446150536929140739f09c6ecbcbea7f0 in my tree and will > > > go upstream shortly - but yes, probably should not add this one, before > > > both can be pulled in. > > > > > > Thanks everyone for the quick reply on this. I agree. > > Ok, I've dropped this patch from all of the stable queues for now. Both now queued up, thanks. greg k-h