public inbox for linux-btrfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Nikolay Borisov <nborisov@suse.com>
To: Josef Bacik <josef@toxicpanda.com>,
	linux-btrfs@vger.kernel.org, kernel-team@fb.com
Subject: Re: [PATCH 13/23] btrfs: add the data transaction commit logic into may_commit_transaction
Date: Tue, 7 Jul 2020 17:39:12 +0300	[thread overview]
Message-ID: <216b0050-3684-9c74-16e6-8a776a23f41e@suse.com> (raw)
In-Reply-To: <20200630135921.745612-14-josef@toxicpanda.com>



On 30.06.20 г. 16:59 ч., Josef Bacik wrote:
> Data space flushing currently unconditionally commits the transaction
> twice in a row, and the last time it checks if there's enough pinned
> extents to satisfy it's reservation before deciding to commit the
> transaction for the 3rd and final time.
> 
> Encode this logic into may_commit_transaction().  In the next patch we
> will pass in U64_MAX for bytes_needed the first two times, and the final
> time we will pass in the actual bytes we need so the normal logic will
> apply.
> 
> This patch exists soley to make the logical changes I will make to the
> flushing state machine separate to make it easier to bisect any
> performance related regressions.
> 
> Reviewed-by: Nikolay Borisov <nborisov@suse.com>
> Tested-by: Nikolay Borisov <nborisov@suse.com>
> Signed-off-by: Josef Bacik <josef@toxicpanda.com>

On a second pass through I'm now somewhat reluctant to merge this code.
may_commit_transaction has grown more logic which pertain solely to
metadata. As such I think we should separate that logic (i.e current
may_commit_transaction) and any further adjustments we might want to
make for data space info. For example all the delayed refs/trans
reservation checks etc. make absolutely no sense for data space info.

> ---
>  fs/btrfs/space-info.c | 24 +++++++++++++++++++-----
>  1 file changed, 19 insertions(+), 5 deletions(-)
> 
> diff --git a/fs/btrfs/space-info.c b/fs/btrfs/space-info.c
> index e041b1d58e28..fb63ddc31540 100644
> --- a/fs/btrfs/space-info.c
> +++ b/fs/btrfs/space-info.c
> @@ -579,21 +579,33 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info,
>   * will return -ENOSPC.
>   */
>  static int may_commit_transaction(struct btrfs_fs_info *fs_info,
> -				  struct btrfs_space_info *space_info)
> +				  struct btrfs_space_info *space_info,
> +				  u64 bytes_needed)
>  {
>  	struct reserve_ticket *ticket = NULL;
>  	struct btrfs_block_rsv *delayed_rsv = &fs_info->delayed_block_rsv;
>  	struct btrfs_block_rsv *delayed_refs_rsv = &fs_info->delayed_refs_rsv;
>  	struct btrfs_block_rsv *trans_rsv = &fs_info->trans_block_rsv;
>  	struct btrfs_trans_handle *trans;
> -	u64 bytes_needed;
>  	u64 reclaim_bytes = 0;
>  	u64 cur_free_bytes = 0;
> +	bool do_commit = false;
>  
>  	trans = (struct btrfs_trans_handle *)current->journal_info;
>  	if (trans)
>  		return -EAGAIN;
>  
> +	/*
> +	 * If we are data and have passed in U64_MAX we just want to
> +	 * unconditionally commit the transaction to match the previous data
> +	 * flushing behavior.
> +	 */
> +	if ((space_info->flags & BTRFS_BLOCK_GROUP_DATA) &&
> +	   bytes_needed == U64_MAX) {
> +		do_commit = true;
> +		goto check_pinned;
> +	}
> +
>  	spin_lock(&space_info->lock);
>  	cur_free_bytes = btrfs_space_info_used(space_info, true);
>  	if (cur_free_bytes < space_info->total_bytes)
> @@ -607,7 +619,7 @@ static int may_commit_transaction(struct btrfs_fs_info *fs_info,
>  	else if (!list_empty(&space_info->tickets))
>  		ticket = list_first_entry(&space_info->tickets,
>  					  struct reserve_ticket, list);
> -	bytes_needed = (ticket) ? ticket->bytes : 0;
> +	bytes_needed = (ticket) ? ticket->bytes : bytes_needed;
>  
>  	if (bytes_needed > cur_free_bytes)
>  		bytes_needed -= cur_free_bytes;
> @@ -618,6 +630,7 @@ static int may_commit_transaction(struct btrfs_fs_info *fs_info,
>  	if (!bytes_needed)
>  		return 0;
>  
> +check_pinned:
>  	trans = btrfs_join_transaction(fs_info->extent_root);
>  	if (IS_ERR(trans))
>  		return PTR_ERR(trans);
> @@ -627,7 +640,8 @@ static int may_commit_transaction(struct btrfs_fs_info *fs_info,
>  	 * we have block groups that are going to be freed, allowing us to
>  	 * possibly do a chunk allocation the next loop through.
>  	 */
> -	if (test_bit(BTRFS_TRANS_HAVE_FREE_BGS, &trans->transaction->flags) ||
> +	if (do_commit ||
> +	    test_bit(BTRFS_TRANS_HAVE_FREE_BGS, &trans->transaction->flags) ||
>  	    __percpu_counter_compare(&space_info->total_bytes_pinned,
>  				     bytes_needed,
>  				     BTRFS_TOTAL_BYTES_PINNED_BATCH) >= 0)
> @@ -743,7 +757,7 @@ static void flush_space(struct btrfs_fs_info *fs_info,
>  		btrfs_wait_on_delayed_iputs(fs_info);
>  		break;
>  	case COMMIT_TRANS:
> -		ret = may_commit_transaction(fs_info, space_info);
> +		ret = may_commit_transaction(fs_info, space_info, num_bytes);
>  		break;
>  	default:
>  		ret = -ENOSPC;
> 

  reply	other threads:[~2020-07-07 14:39 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-30 13:58 [PATCH 00/23] Change data reservations to use the ticketing infra Josef Bacik
2020-06-30 13:58 ` [PATCH 01/23] btrfs: change nr to u64 in btrfs_start_delalloc_roots Josef Bacik
2020-06-30 13:59 ` [PATCH 02/23] btrfs: remove orig from shrink_delalloc Josef Bacik
2020-06-30 13:59 ` [PATCH 03/23] btrfs: handle U64_MAX for shrink_delalloc Josef Bacik
2020-06-30 13:59 ` [PATCH 04/23] btrfs: make shrink_delalloc take space_info as an arg Josef Bacik
2020-06-30 13:59 ` [PATCH 05/23] btrfs: make ALLOC_CHUNK use the space info flags Josef Bacik
2020-06-30 13:59 ` [PATCH 06/23] btrfs: call btrfs_try_granting_tickets when freeing reserved bytes Josef Bacik
2020-06-30 13:59 ` [PATCH 07/23] btrfs: call btrfs_try_granting_tickets when unpinning anything Josef Bacik
2020-06-30 13:59 ` [PATCH 08/23] btrfs: call btrfs_try_granting_tickets when reserving space Josef Bacik
2020-06-30 13:59 ` [PATCH 09/23] btrfs: use the btrfs_space_info_free_bytes_may_use helper for delalloc Josef Bacik
2020-06-30 13:59 ` [PATCH 10/23] btrfs: use btrfs_start_delalloc_roots in shrink_delalloc Josef Bacik
2020-06-30 13:59 ` [PATCH 11/23] btrfs: check tickets after waiting on ordered extents Josef Bacik
2020-06-30 13:59 ` [PATCH 12/23] btrfs: add flushing states for handling data reservations Josef Bacik
2020-06-30 13:59 ` [PATCH 13/23] btrfs: add the data transaction commit logic into may_commit_transaction Josef Bacik
2020-07-07 14:39   ` Nikolay Borisov [this message]
2020-07-07 14:54     ` Josef Bacik
2020-07-07 15:01     ` Josef Bacik
2020-06-30 13:59 ` [PATCH 14/23] btrfs: add btrfs_reserve_data_bytes and use it Josef Bacik
2020-06-30 13:59 ` [PATCH 15/23] btrfs: use ticketing for data space reservations Josef Bacik
2020-07-07 14:46   ` Nikolay Borisov
2020-07-07 14:56     ` Josef Bacik
2020-06-30 13:59 ` [PATCH 16/23] btrfs: serialize data reservations if we are flushing Josef Bacik
2020-06-30 13:59 ` [PATCH 17/23] btrfs: use the same helper for data and metadata reservations Josef Bacik
2020-06-30 13:59 ` [PATCH 18/23] btrfs: drop the commit_cycles stuff for data reservations Josef Bacik
2020-06-30 13:59 ` [PATCH 19/23] btrfs: don't force commit if we are data Josef Bacik
2020-06-30 13:59 ` [PATCH 20/23] btrfs: run delayed iputs before committing the transaction for data Josef Bacik
2020-06-30 13:59 ` [PATCH 21/23] btrfs: flush delayed refs when trying to reserve data space Josef Bacik
2020-06-30 13:59 ` [PATCH 22/23] btrfs: do async reclaim for data reservations Josef Bacik
2020-06-30 13:59 ` [PATCH 23/23] btrfs: add a comment explaining the data flush steps Josef Bacik
2020-07-03 16:30 ` [PATCH 00/23] Change data reservations to use the ticketing infra David Sterba
2020-07-07 15:28   ` Josef Bacik
  -- strict thread matches above, loose matches on Subject: below --
2020-07-21 14:22 [PATCH 00/23][v4] " Josef Bacik
2020-07-21 14:22 ` [PATCH 13/23] btrfs: add the data transaction commit logic into may_commit_transaction Josef Bacik
2020-08-13 16:35   ` David Sterba
2020-07-08 13:59 [PATCH 00/23][v3] Change data reservations to use the ticketing infra Josef Bacik
2020-07-08 14:00 ` [PATCH 13/23] btrfs: add the data transaction commit logic into may_commit_transaction Josef Bacik
2020-07-07 15:42 [PATCH 00/23][v2] Change data reservations to use the ticketing infra Josef Bacik
2020-07-07 15:42 ` [PATCH 13/23] btrfs: add the data transaction commit logic into may_commit_transaction Josef Bacik
2020-02-04 16:19 [PATCH 0/23][v4] Convert data reservations to the ticketing infrastructure Josef Bacik
2020-02-04 16:19 ` [PATCH 13/23] btrfs: add the data transaction commit logic into may_commit_transaction Josef Bacik
2020-02-04 16:56   ` Johannes Thumshirn

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=216b0050-3684-9c74-16e6-8a776a23f41e@suse.com \
    --to=nborisov@suse.com \
    --cc=josef@toxicpanda.com \
    --cc=kernel-team@fb.com \
    --cc=linux-btrfs@vger.kernel.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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox