On Mon 24-06-19 05:58:56, Tejun Heo wrote: > Hello, Jan. > > On Mon, Jun 24, 2019 at 10:21:30AM +0200, Jan Kara wrote: > > OK, now I understand. Just one more question: So effectively, you are using > > wbc->no_wbc_acct to pass information from btrfs code to btrfs code telling > > it whether IO should or should not be accounted with wbc_account_io(). > > Yes. > > > Wouldn't it make more sense to just pass this information internally > > within btrfs? Granted, if this mechanism gets more widespread use by other > > filesystems, then probably using wbc flag makes more sense. But I'm not > > sure if this isn't a premature generalization... > > The btrfs async issuers end up using generic writeback path and uses > the generic wbc owner mechanisms so that ios are attached to the right > cgroup too. So, I kinda prefer to provide a generic mechanism from > wbc side. OK, I can live with that. We just have to be kind of careful so that people just don't sprinkle no_wbc_acct writeback around because they don't know better. Maybe you could at least add comment to no_wbc_acct mentioning that this is for the cases where writeback has already been accounted for? > That said, the names are a bit misleading and I think it'd > be better to rename them to something more explicit, e.g. sth along > the line of wbc_update_cgroup_owner() and wbc->no_cgroup_owner. What > do you think? Yeah, renaming would probably make things clearer as well. Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR
