From: David Sterba <email@example.com> To: Damien Le Moal <Damien.LeMoal@wdc.com> Cc: "Josef Bacik" <firstname.lastname@example.org>, "Naohiro Aota" <Naohiro.Aota@wdc.com>, "email@example.com" <firstname.lastname@example.org>, "David Sterba" <email@example.com>, "Chris Mason" <firstname.lastname@example.org>, "Qu Wenruo" <email@example.com>, "Nikolay Borisov" <firstname.lastname@example.org>, "email@example.com" <firstname.lastname@example.org>, "Hannes Reinecke" <email@example.com>, "firstname.lastname@example.org" <email@example.com>, "Matias Bjørling" <firstname.lastname@example.org>, "Johannes Thumshirn" <email@example.com>, "Bart Van Assche" <firstname.lastname@example.org> Subject: Re: [PATCH 11/19] btrfs: introduce submit buffer Date: Tue, 18 Jun 2019 02:00:20 +0200 Message-ID: <20190618000020.GK19057@twin.jikos.cz> (raw) In-Reply-To: <BYAPR04MB5816E9FC012A289CA438E794E7EB0@BYAPR04MB5816.namprd04.prod.outlook.com> On Mon, Jun 17, 2019 at 03:16:05AM +0000, Damien Le Moal wrote: > Josef, > > On 2019/06/13 23:15, Josef Bacik wrote: > > On Fri, Jun 07, 2019 at 10:10:17PM +0900, Naohiro Aota wrote: > >> Sequential allocation is not enough to maintain sequential delivery of > >> write IOs to the device. Various features (async compress, async checksum, > >> ...) of btrfs affect ordering of the IOs. This patch introduces submit > >> buffer to sort WRITE bios belonging to a block group and sort them out > >> sequentially in increasing block address to achieve sequential write > >> sequences with __btrfs_map_bio(). > >> > >> Signed-off-by: Naohiro Aota <email@example.com> > > > > I hate everything about this. Can't we just use the plugging infrastructure for > > this and then make sure it re-orders the bios before submitting them? Also > > what's to prevent the block layer scheduler from re-arranging these io's? > > Thanks, > > The block I/O scheduler reorders requests in LBA order, but that happens for a > newly inserted request against pending requests. If there are no pending > requests because all requests were already issued, no ordering happen, and even > worse, if the drive queue is not full yet (e.g. there are free tags), then the > newly inserted request will be dispatched almost immediately, preventing > reordering with subsequent incoming write requests to happen. This would be good to add to the changelog. > > The other problem is that the mq-deadline scheduler does not track zone WP > position. Write request issuing is done regardless of the current WP value, > solely based on LBA ordering. This means that mq-deadline will not prevent > out-of-order, or rather, unaligned write requests. This seems to be the key point. > These will not be detected > and dispatched whenever possible. The reasons for this are that: > 1) the disk user (the FS) has to manage zone WP positions anyway. So duplicating > that management at the block IO scheduler level is inefficient. > 2) Adding zone WP management at the block IO scheduler level would also need a > write error processing path to resync the WP value in case of failed writes. But > the user/FS also needs that anyway. Again duplicated functionalities. > 3) The block layer will need a timeout to force issue or cancel pending > unaligned write requests. This is necessary in case the drive user stops issuing > writes (for whatever reasons) or the scheduler is being switched. This would > unnecessarily cause write I/O errors or cause deadlocks if the request queue > quiesce mode is entered at the wrong time (and I do not see a good way to deal > with that). > > blk-mq is already complicated enough. Adding this to the block IO scheduler will > unnecessarily complicate things further for no real benefits. I would like to > point out the dm-zoned device mapper and f2fs which are both already dealing > with write ordering and write error processing directly. Both are fairly > straightforward but completely different and each optimized for their own structure. So the question is where on which layer the decision logic is. The filesystem(s) or dm-zoned have enough information about the zones and the writes can be pre-sorted. This is what the patch proposes. From your explanation I get that the io scheduler can throw the wrench in the sequential ordering, for various reasons depending on state of internal structures od device queues. This is my simplified interpretation as I don't understand all the magic below filesystem layer. I assume there are some guarantees about the ordering, eg. within one plug, that apply to all schedulers (maybe not the noop one). Something like that should be the least common functionality that the filesystem layer can rely on. > Naohiro changes to btrfs IO scheduler have the same intent, that is, efficiently > integrate and handle write ordering "a la btrfs". Would creating a different > "hmzoned" btrfs IO scheduler help address your concerns ? IMHO this sounds both the same, all we care about is the sequential ordering, which in some sense is "scheduling", but I would not call it that way due to the simplicity. As implemented, it's a list of bios, but I'd suggest using rb-tree or xarray, the insertion is fast and submission is start to end traversal. I'm not sure that the loop in __btrfs_map_bio_zoned after label send_bios: has reasonable complexity, looks like an O(N^2).
next prev parent reply index Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-06-07 13:10 [PATCH v2 00/19] btrfs zoned block device support Naohiro Aota 2019-06-07 13:10 ` [PATCH 01/19] btrfs: introduce HMZONED feature flag Naohiro Aota 2019-06-07 13:10 ` [PATCH 02/19] btrfs: Get zone information of zoned block devices Naohiro Aota 2019-06-13 13:58 ` Josef Bacik 2019-06-18 6:04 ` Naohiro Aota 2019-06-13 13:58 ` Josef Bacik 2019-06-17 18:57 ` David Sterba 2019-06-18 6:42 ` Naohiro Aota 2019-06-27 15:11 ` David Sterba 2019-06-07 13:10 ` [PATCH 03/19] btrfs: Check and enable HMZONED mode Naohiro Aota 2019-06-13 13:57 ` Josef Bacik 2019-06-18 6:43 ` Naohiro Aota 2019-06-07 13:10 ` [PATCH 04/19] btrfs: disable fallocate in " Naohiro Aota 2019-06-07 13:10 ` [PATCH 05/19] btrfs: disable direct IO " Naohiro Aota 2019-06-13 14:00 ` Josef Bacik 2019-06-18 8:17 ` Naohiro Aota 2019-06-07 13:10 ` [PATCH 06/19] btrfs: align dev extent allocation to zone boundary Naohiro Aota 2019-06-07 13:10 ` [PATCH 07/19] btrfs: do sequential extent allocation in HMZONED mode Naohiro Aota 2019-06-13 14:07 ` Josef Bacik 2019-06-18 8:28 ` Naohiro Aota 2019-06-18 13:37 ` Josef Bacik 2019-06-17 22:30 ` David Sterba 2019-06-18 8:49 ` Naohiro Aota 2019-06-27 15:28 ` David Sterba 2019-06-07 13:10 ` [PATCH 08/19] btrfs: make unmirroed BGs readonly only if we have at least one writable BG Naohiro Aota 2019-06-13 14:09 ` Josef Bacik 2019-06-18 7:42 ` Naohiro Aota 2019-06-18 13:35 ` Josef Bacik 2019-06-07 13:10 ` [PATCH 09/19] btrfs: limit super block locations in HMZONED mode Naohiro Aota 2019-06-13 14:12 ` Josef Bacik 2019-06-18 8:51 ` Naohiro Aota 2019-06-17 22:53 ` David Sterba 2019-06-18 9:01 ` Naohiro Aota 2019-06-27 15:35 ` David Sterba 2019-06-28 3:55 ` Anand Jain 2019-06-28 6:39 ` Naohiro Aota 2019-06-28 6:52 ` Anand Jain 2019-06-07 13:10 ` [PATCH 10/19] btrfs: rename btrfs_map_bio() Naohiro Aota 2019-06-07 13:10 ` [PATCH 11/19] btrfs: introduce submit buffer Naohiro Aota 2019-06-13 14:14 ` Josef Bacik 2019-06-17 3:16 ` Damien Le Moal 2019-06-18 0:00 ` David Sterba [this message] 2019-06-18 4:04 ` Damien Le Moal 2019-06-18 13:33 ` Josef Bacik 2019-06-19 10:32 ` Damien Le Moal 2019-06-07 13:10 ` [PATCH 12/19] btrfs: expire submit buffer on timeout Naohiro Aota 2019-06-13 14:15 ` Josef Bacik 2019-06-17 3:19 ` Damien Le Moal 2019-06-07 13:10 ` [PATCH 13/19] btrfs: avoid sync IO prioritization on checksum in HMZONED mode Naohiro Aota 2019-06-13 14:17 ` Josef Bacik 2019-06-07 13:10 ` [PATCH 14/19] btrfs: redirty released extent buffers in sequential BGs Naohiro Aota 2019-06-13 14:24 ` Josef Bacik 2019-06-18 9:09 ` Naohiro Aota 2019-06-07 13:10 ` [PATCH 15/19] btrfs: reset zones of unused block groups Naohiro Aota 2019-06-07 13:10 ` [PATCH 16/19] btrfs: wait existing extents before truncating Naohiro Aota 2019-06-13 14:25 ` Josef Bacik 2019-06-07 13:10 ` [PATCH 17/19] btrfs: shrink delayed allocation size in HMZONED mode Naohiro Aota 2019-06-13 14:27 ` Josef Bacik 2019-06-07 13:10 ` [PATCH 18/19] btrfs: support dev-replace " Naohiro Aota 2019-06-13 14:33 ` Josef Bacik 2019-06-18 9:14 ` Naohiro Aota 2019-06-07 13:10 ` [PATCH 19/19] btrfs: enable to mount HMZONED incompat flag Naohiro Aota 2019-06-07 13:17 ` [PATCH 01/12] btrfs-progs: build: Check zoned block device support Naohiro Aota 2019-06-07 13:17 ` [PATCH 02/12] btrfs-progs: utils: Introduce queue_param Naohiro Aota 2019-06-07 13:17 ` [PATCH 03/12] btrfs-progs: add new HMZONED feature flag Naohiro Aota 2019-06-07 13:17 ` [PATCH 04/12] btrfs-progs: Introduce zone block device helper functions Naohiro Aota 2019-06-07 13:17 ` [PATCH 05/12] btrfs-progs: load and check zone information Naohiro Aota 2019-06-07 13:17 ` [PATCH 06/12] btrfs-progs: avoid writing super block to sequential zones Naohiro Aota 2019-06-07 13:17 ` [PATCH 07/12] btrfs-progs: support discarding zoned device Naohiro Aota 2019-06-07 13:17 ` [PATCH 08/12] btrfs-progs: volume: align chunk allocation to zones Naohiro Aota 2019-06-07 13:17 ` [PATCH 09/12] btrfs-progs: do sequential allocation Naohiro Aota 2019-06-07 13:17 ` [PATCH 10/12] btrfs-progs: mkfs: Zoned block device support Naohiro Aota 2019-06-07 13:17 ` [PATCH 11/12] btrfs-progs: device-add: support HMZONED device Naohiro Aota 2019-06-07 13:17 ` [PATCH 12/12] btrfs-progs: introduce support for dev-place " Naohiro Aota 2019-06-12 17:51 ` [PATCH v2 00/19] btrfs zoned block device support David Sterba 2019-06-13 4:59 ` Naohiro Aota 2019-06-13 13:46 ` David Sterba 2019-06-14 2:07 ` Naohiro Aota 2019-06-17 2:44 ` Damien Le Moal
Reply instructions: You may reply publically to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20190618000020.GK19057@twin.jikos.cz \ --firstname.lastname@example.org \ --cc=Damien.LeMoal@wdc.com \ --cc=Naohiro.Aota@wdc.com \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ --email@example.com \ --firstname.lastname@example.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
Linux-BTRFS Archive on lore.kernel.org Archives are clonable: git clone --mirror https://lore.kernel.org/linux-btrfs/0 linux-btrfs/git/0.git # If you have public-inbox 1.1+ installed, you may # initialize and index your mirror using the following commands: public-inbox-init -V2 linux-btrfs linux-btrfs/ https://lore.kernel.org/linux-btrfs \ email@example.com firstname.lastname@example.org public-inbox-index linux-btrfs Newsgroup available over NNTP: nntp://nntp.lore.kernel.org/org.kernel.vger.linux-btrfs AGPL code for this site: git clone https://public-inbox.org/ public-inbox