When we start from blk_cleanup_queue(), we put request queue in bypass mode, drain it (and service queues), and then destroy blkcgs (explicitly)
When we start from blk_release_queue(), we do not drain first and then destroy blkcgs. So if we destroy blkcg and then call (implicitly) and bail out of blk_drain_queue, we would not have drained the service queues which is not what we want. I do not see any harm in waiting till end to release blkcgs (as I understand). Regards, Shirish On Tue, Jun 24, 2014 at 4:09 PM, Tejun Heo <t...@kernel.org> wrote: > Hello, > > On Mon, Jun 23, 2014 at 01:52:25PM -0500, shirishpargaon...@gmail.com wrote: >> From: Shirish Pargaonkar <spargaon...@suse.com> >> >> Release blkg infrastructure only after last policy is deactivated >> (i.e. let blkg_destroy_all() be called only from blkcg_deactivate_policy()) >> >> Otherwise, module can oops because root_blkg gets assigned NULL during >> cleanup and we attempt draining the service queues via root_blkg afterwords. > > I'm not sure this fix makes sense. Cleanup path oopses on an already > freed resource. How can the solution be not freeing? Why not simply > make blkcg_drain_queue() bail if the blkgs are all destroyed? The > whole thing is fully synchronized with the queuelock, right? > > Can you please also cc Jens when you post the next iteration? > > Thanks. > > -- > tejun -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
