linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Nikolay Borisov <nborisov@suse.com>
To: Johannes Thumshirn <Johannes.Thumshirn@wdc.com>,
	Christoph Hellwig <hch@lst.de>, Chris Mason <clm@fb.com>,
	Josef Bacik <josef@toxicpanda.com>,
	David Sterba <dsterba@suse.com>
Cc: "linux-btrfs@vger.kernel.org" <linux-btrfs@vger.kernel.org>
Subject: Re: [PATCH 10/11] btrfs: make the btrfs_io_context allocation in __btrfs_map_block optional
Date: Fri, 15 Jul 2022 11:30:11 +0300	[thread overview]
Message-ID: <c92d13d8-67cc-fe22-54e8-b53686d08ec6@suse.com> (raw)
In-Reply-To: <PH0PR04MB741647058825EF24A1E89CAA9B899@PH0PR04MB7416.namprd04.prod.outlook.com>



On 13.07.22 г. 12:58 ч., Johannes Thumshirn wrote:
> On 13.07.22 08:14, Christoph Hellwig wrote:
>> +	/*
>> +	 * If this I/O maps to a single device, try to return the device and
>> +	 * physical block information on the stack instead of allocating an
>> +	 * I/O context structure.
>> +	 */
>> +	if (smap && num_alloc_stripes == 1 &&
>> +	    !((map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) && mirror_num > 1) &&
>> +	    (!need_full_stripe(op) || !dev_replace_is_ongoing ||
>> +	     !dev_replace->tgtdev)) {
>> +		if (unlikely(patch_the_first_stripe_for_dev_replace)) {
>> +			smap->dev = dev_replace->tgtdev;
>> +			smap->physical = physical_to_patch_in_first_stripe;
>> +			*mirror_num_p = map->num_stripes + 1;
>> +		} else {
>> +			set_stripe(smap, map, stripe_index, stripe_offset,
>> +				   stripe_nr);
>> +			*mirror_num_p = mirror_num;
>> +		}
>> +		*bioc_ret = NULL;
>> +		ret = 0;
>> +		goto out;
>> +	}
>> +
> 
> 
> Could be my changes on top, but in order to get RAID0 with a raid-stripe-tree
> working I needed the following change:
> 
> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
> index a7886e421153..344d2cf941a7 100644
> --- a/fs/btrfs/volumes.c
> +++ b/fs/btrfs/volumes.c
> @@ -6506,7 +6506,7 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info,
>           * physical block information on the stack instead of allocating an
>           * I/O context structure.
>           */
> -       if (smap && num_alloc_stripes == 1 &&
> +       if (smap && num_alloc_stripes == 1 && !(map->type & BTRFS_BLOCK_GROUP_STRIPE_MASK) &&

nit: BLOCK_GROUP_STRIPE_MASK already contains BLOCK_GROUP_RAID56_MASK so 
BTRFS_BLOCK_GROUP_RAID56_MASK can be replaced. But I agree with hch that 
this change should come with your series.

>              !((map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) && mirror_num > 1) &&
>              (!need_full_stripe(op) || !dev_replace_is_ongoing ||
>               !dev_replace->tgtdev)) {
> 

  parent reply	other threads:[~2022-07-15  8:30 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-13  6:13 btrfs I/O completion cleanup and single device I/O optimizations v2 Christoph Hellwig
2022-07-13  6:13 ` [PATCH 01/11] btrfs: don't call bioset_integrity_create for btrfs_bioset Christoph Hellwig
2022-07-13  6:13 ` [PATCH 02/11] btrfs: move btrfs_bio allocation to volumes.c Christoph Hellwig
2022-07-13  6:13 ` [PATCH 03/11] btrfs: pass the operation to btrfs_bio_alloc Christoph Hellwig
2022-07-13  6:13 ` [PATCH 04/11] btrfs: don't take a bio_counter reference for cloned bios Christoph Hellwig
2022-07-13  6:27   ` Johannes Thumshirn
2022-07-13  6:13 ` [PATCH 05/11] btrfs: remove bioc->stripes_pending Christoph Hellwig
2022-07-13  6:31   ` Johannes Thumshirn
2022-07-13  6:41   ` Johannes Thumshirn
2022-07-14 12:15   ` Nikolay Borisov
2022-07-13  6:13 ` [PATCH 06/11] btrfs: properly abstract the parity raid bio handling Christoph Hellwig
2022-07-13  6:13 ` [PATCH 07/11] btrfs: give struct btrfs_bio a real end_io handler Christoph Hellwig
2022-07-13  6:47   ` Johannes Thumshirn
2022-07-13  6:51     ` Christoph Hellwig
2022-07-13  6:13 ` [PATCH 08/11] btrfs: split submit_stripe_bio Christoph Hellwig
2022-07-13  6:46   ` Johannes Thumshirn
2022-07-13  6:50     ` Christoph Hellwig
2022-07-13  6:52       ` Johannes Thumshirn
2022-07-14 14:02   ` Nikolay Borisov
2022-07-13  6:13 ` [PATCH 09/11] btrfs: simplify the submit_stripe_bio calling convention Christoph Hellwig
2022-07-13  6:49   ` Johannes Thumshirn
2022-07-14 14:30   ` Nikolay Borisov
2022-07-13  6:13 ` [PATCH 10/11] btrfs: make the btrfs_io_context allocation in __btrfs_map_block optional Christoph Hellwig
2022-07-13  9:58   ` Johannes Thumshirn
2022-07-13 11:18     ` Christoph Hellwig
2022-07-15  8:30     ` Nikolay Borisov [this message]
2022-07-13  6:13 ` [PATCH 11/11] btrfs: stop allocation a btrfs_io_context for simple I/O Christoph Hellwig
2022-09-06 13:12 ` btrfs I/O completion cleanup and single device I/O optimizations v2 David Sterba
2022-09-07  9:08   ` Christoph Hellwig
2022-09-09 12:34     ` David Sterba
  -- strict thread matches above, loose matches on Subject: below --
2022-08-06  8:03 btrfs I/O completion cleanup and single device I/O optimizations v3 Christoph Hellwig
2022-08-06  8:03 ` [PATCH 10/11] btrfs: make the btrfs_io_context allocation in __btrfs_map_block optional Christoph Hellwig
2022-08-20 11:34   ` Anand Jain

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=c92d13d8-67cc-fe22-54e8-b53686d08ec6@suse.com \
    --to=nborisov@suse.com \
    --cc=Johannes.Thumshirn@wdc.com \
    --cc=clm@fb.com \
    --cc=dsterba@suse.com \
    --cc=hch@lst.de \
    --cc=josef@toxicpanda.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;
as well as URLs for NNTP newsgroup(s).