From: Heinz Mauelshagen <mauelshagen@redhat.com>
To: Jonathan Brassow <jbrassow@redhat.com>
Cc: heinzm@redhat.com, dm-devel@redhat.com, agk@redhat.com
Subject: Re: New read-balancing patch for dm-raid1.c
Date: Tue, 10 Oct 2006 21:02:44 +0200 [thread overview]
Message-ID: <20061010190244.GA4792@redhat.com> (raw)
In-Reply-To: <1160410439.30654.12.camel@hydrogen.msp.redhat.com>
Jon,
the "if (ms->read_count_reset)" should read "if (!ms->read_count_reset)"
to disable read balancing and the 2 while loops could be merged into one.
Heinz
On Mon, Oct 09, 2006 at 11:13:59AM -0500, Jonathan Brassow wrote:
> Looking for some opinions/comments on this read balancing patch. (Basic
> implementation details in patch header.)
>
> I haven't gone to too much trouble to make the read balancing code
> generic, since there is only one policy right now. In the future, I
> will likely have to swap out the three fields in the mirror set
> structure for a void pointer that will contain data for various other
> read balancing policies. I have, however, constructed the new table
> line arguments in such a way that the read-balancing arguments are easy
> to identify and pass to an initialization function without knowing too
> much about the actual arguments.
>
> One thing I'm not all that fond of is the order that the mirror is set
> up. The read-balancing arguments are parsed before 'alloc_context' is
> called - meaning that new read-balancing structures would be allocated
> first, then linked in later... and read-balancing is set up before the
> default mirror device is chosen. I don't really care for
> 'alloc_context' setting up 'ms->read_mirror', while other fields are
> initialized after 'alloc_context'. I don't think this is too much of a
> problem though.
>
> brassow
>
> This patch adds read balancing. The round-robin method is the first
> to be implemented, but provisions are made for others to be implemented
> in the future.
>
> The allowable mirror table arguments has been expanded. It is now
> as follows:
>
> <start> <length> mirror \
> <log-type> <# log params> <log params> \
> *new* [readbalance <# rb params> <rb params>] \
> <# mirrors> <device1> <offset1> ... <deviceN> <offsetN>
>
> The new read balancing arguments are optional, and the only
> currently valid read balancing arguments are:
> readbalance 2 roundrobin <count>
> Where 'count' is the number of I/Os that go to a device before
> switching to the next device.
>
> 'struct mirror *choose_mirror(struct mirror_set *ms)' is the
> function that chooses the read mirror based on read balancing
> policy. It should only be called when the region of the
> mirror being read from is known to be in-sync. 'choose_mirror'
> will avoid selecting devices with error_counts > 0 - returning
> NULL if no devices are available.
>
> Index: linux-2.6.18/drivers/md/dm-raid1.c
> ===================================================================
> --- linux-2.6.18.orig/drivers/md/dm-raid1.c 2006-10-05 13:38:27.000000000 -0500
> +++ linux-2.6.18/drivers/md/dm-raid1.c 2006-10-09 10:31:47.000000000 -0500
> @@ -135,6 +135,9 @@ struct mirror_set {
> struct mirror *default_mirror; /* Default mirror */
>
> unsigned int nr_mirrors;
> + unsigned int read_count_reset; /* number of reads before switching */
> + atomic_t read_count; /* Read counter for read balancing */
> + struct mirror *read_mirror; /* Last mirror read. */
> struct mirror mirror[0];
> };
>
> @@ -686,10 +689,59 @@ static void do_recovery(struct mirror_se
> /*-----------------------------------------------------------------
> * Reads
> *---------------------------------------------------------------*/
> -static struct mirror *choose_mirror(struct mirror_set *ms, sector_t sector)
> +
> +/* choose_mirror
> + * @ms: the mirror set
> + *
> + * This function is used for read balancing.
> + *
> + * Returns: chosen mirror, or NULL on failure
> + */
> +static struct mirror *choose_mirror(struct mirror_set *ms)
> {
> - /* FIXME: add read balancing */
> - return ms->default_mirror;
> + unsigned int i;
> + struct mirror *start_mirror = ms->read_mirror;
> +
> + /*
> + * If 'read_count_reset' is zero here, then read-balancing
> + * is disabled.
> + */
> + if (ms->read_count_reset)) {
> + do {
> + if (likely(!atomic_read(&ms->read_mirror->error_count)))
> + goto use_mirror;
> +
> + if (ms->read_mirror-- == ms->mirror)
> + ms->read_mirror += ms->nr_mirrors;
> + } while (ms->read_mirror != start_mirror);
> + return NULL;
> + }
> +
> + /*
> + * Perform ms->read_count_reset reads on each working mirror then
> + * advance to the next one. start_mirror stores
> + * the first we tried, so we know when we're done.
> + */
> + do {
> + if (likely(!atomic_read(&ms->read_mirror->error_count)) &&
> + !atomic_dec_and_test(&ms->read_count))
> + goto use_mirror;
> +
> + atomic_set(&ms->read_count, ms->read_count_reset);
> +
> + if (ms->read_mirror-- == ms->mirror)
> + ms->read_mirror += ms->nr_mirrors;
> + } while (ms->read_mirror != start_mirror);
> +
> + /*
> + * We've rejected every mirror.
> + * Confirm the start_mirror can be used.
> + */
> + if (unlikely(atomic_read(&ms->read_mirror->error_count)))
> + return NULL;
> +
> +use_mirror:
> + return ms->read_mirror;
> }
>
> /*
> @@ -714,7 +766,7 @@ static void do_reads(struct mirror_set *
> * We can only read balance if the region is in sync.
> */
> if (rh_in_sync(&ms->rh, region, 0))
> - m = choose_mirror(ms, bio->bi_sector);
> + m = choose_mirror(ms);
> else
> m = ms->default_mirror;
>
> @@ -907,6 +959,7 @@ static struct mirror_set *alloc_context(
> ms->nr_mirrors = nr_mirrors;
> ms->nr_regions = dm_sector_div_up(ti->len, region_size);
> ms->in_sync = 0;
> + ms->read_mirror = &ms->mirror[DEFAULT_MIRROR];
> ms->default_mirror = &ms->mirror[DEFAULT_MIRROR];
>
> if (rh_init(&ms->rh, ms, dl, region_size, ms->nr_regions)) {
> @@ -1028,6 +1081,7 @@ static struct dirty_log *create_dirty_lo
> static int mirror_ctr(struct dm_target *ti, unsigned int argc, char **argv)
> {
> int r;
> + unsigned int read_count_reset = 0, read_balance_args;
> unsigned int nr_mirrors, m, args_used;
> struct mirror_set *ms;
> struct dirty_log *dl;
> @@ -1039,6 +1093,29 @@ static int mirror_ctr(struct dm_target *
> argv += args_used;
> argc -= args_used;
>
> + if (!argc) {
> + ti->error = "Invalid number of arguments";
> + dm_destroy_dirty_log(dl);
> + return -EINVAL;
> + } else if (!strcmp("readbalance", argv[0]) &&
> + (sscanf(argv[1], "%u", &read_balance_args) == 1)) {
> + /*
> + * When there is more than one read-balancing policy,
> + * we will push this next if statement into an
> + * initialization function.
> + */
> + if ((read_balance_args == 2) &&
> + !strcmp("roundrobin", argv[2]) &&
> + (sscanf(argv[3], "%u", &read_count_reset) == 1)) {
> + argv += 4;
> + argc -= 4;
> + } else {
> + ti->error = "Invalid read-balancing arguments";
> + dm_destroy_dirty_log(dl);
> + return -EINVAL;
> + }
> + }
> +
> if (!argc || sscanf(argv[0], "%u", &nr_mirrors) != 1 ||
> nr_mirrors < 2 || nr_mirrors > KCOPYD_MAX_REGIONS + 1) {
> ti->error = "Invalid number of mirrors";
> @@ -1060,6 +1137,8 @@ static int mirror_ctr(struct dm_target *
> return -ENOMEM;
> }
>
> + ms->read_count_reset = read_count_reset;
> +
> /* Get the mirror parameter sets */
> for (m = 0; m < nr_mirrors; m++) {
> r = get_mirror(ms, ti, m, argv);
> @@ -1147,7 +1226,7 @@ static int mirror_map(struct dm_target *
> return 0;
> }
>
> - m = choose_mirror(ms, bio->bi_sector);
> + m = choose_mirror(ms);
> if (!m)
> return -EIO;
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Heinz Mauelshagen Red Hat GmbH
Consulting Development Engineer Am Sonnenhang 11
Storage Development 56242 Marienrachdorf
Germany
Mauelshagen@RedHat.com PHONE +49 171 7803392
FAX +49 2626 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
next prev parent reply other threads:[~2006-10-10 19:02 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-10-09 16:13 New read-balancing patch for dm-raid1.c Jonathan Brassow
2006-10-10 19:02 ` Heinz Mauelshagen [this message]
-- strict thread matches above, loose matches on Subject: below --
2006-10-11 15:33 Jonathan Brassow
2006-10-12 14:46 ` Alasdair G Kergon
2006-10-16 15:16 ` Jonathan E Brassow
Reply instructions:
You may reply publicly 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=20061010190244.GA4792@redhat.com \
--to=mauelshagen@redhat.com \
--cc=agk@redhat.com \
--cc=dm-devel@redhat.com \
--cc=heinzm@redhat.com \
--cc=jbrassow@redhat.com \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.