The trace shows it's applying a 'discard' to the metadata and releasing blocks when the problem occurs.Right, given that the BUG trace shows active discards it could be that v3.8 commit e808807 ("dm thin: fix race between simultaneous io and discards to same block") may help. (though I never saw that race manifest with the BUG in question)
I am mounting ext4 with the discard option on my dm-thinp volumes. I've turned discard off for now hoping for stable operation.
-Eric -- Eric Wheeler www.globallinuxsecurity.pro -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
