Date: Sun, 13 Sep 2015 14:03:53 +0100 From: Steven Hartland <killing@multiplay.co.uk> To: freebsd-fs@freebsd.org Subject: Re: zfs_trim_enabled destroys zio_free() performance Message-ID: <55F57439.8060000@multiplay.co.uk> In-Reply-To: <55F308B7.3020302@FreeBSD.org> References: <CAJjvXiE2mRT4=kPMk3gwiT-3ykeAhaYBx6Tw6HgXhs2=XZWWFg@mail.gmail.com> <55F308B7.3020302@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Do you remember if this was this causing a deadlock or something similar that's easy to provoke? Regards Steve On 11/09/2015 18:00, Alexander Motin wrote: > Hi. > > The code in question was added by me at r253992. Commit message tells it > was made to decouple locks. I don't remember much more details, but may > be it can be redone somehow else. > > On 11.09.2015 19:07, Matthew Ahrens wrote: >> I discovered that when destroying a ZFS snapshot, we can end up using >> several seconds of CPU via this stack trace: >> >> kernel`spinlock_exit+0x2d >> kernel`taskqueue_enqueue+0x12c >> zfs.ko`zio_issue_async+0x7c >> zfs.ko`zio_execute+0x162 >> zfs.ko`dsl_scan_free_block_cb+0x15f >> zfs.ko`bpobj_iterate_impl+0x25d >> zfs.ko`bpobj_iterate_impl+0x46e >> zfs.ko`dsl_scan_sync+0x152 >> zfs.ko`spa_sync+0x5c1 >> zfs.ko`txg_sync_thread+0x3a6 >> kernel`fork_exit+0x9a >> kernel`0xffffffff80d0acbe >> 6558 ms >> >> This is not good for performance since, in addition to the CPU cost, it >> doesn't allow the sync thread to do anything else, and this is >> observable as periods where we don't do any write i/o to disk for >> several seconds. >> >> The problem is that when zfs_trim_enabled is set (which it is by >> default), zio_free_sync() always sets ZIO_STAGE_ISSUE_ASYNC, causing the >> free to be dispatched to a taskq. Since each task completes very >> quickly, there is a large locking and context switching overhead -- we >> would be better off just processing the free in the caller's context. >> >> I'm not sure exactly why we need to go async when trim is enabled, but >> it seems like at least we should not bother going async if trim is not >> actually being used (e.g. with an all-spinning-disk pool). It would >> also be worth investigating not going async even when trim is useful >> (e.g. on SSD-based pools). >> >> Here is the relevant code: >> >> zio_free_sync(): >> if (zfs_trim_enabled) >> stage |= ZIO_STAGE_ISSUE_ASYNC | ZIO_STAGE_VDEV_IO_START | >> ZIO_STAGE_VDEV_IO_ASSESS; >> /* >> * GANG and DEDUP blocks can induce a read (for the gang block >> header, >> * or the DDT), so issue them asynchronously so that this thread is >> * not tied up. >> */ >> else if (BP_IS_GANG(bp) || BP_GET_DEDUP(bp)) >> stage |= ZIO_STAGE_ISSUE_ASYNC; >> >> --matt >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?55F57439.8060000>