linux-kernel - Re: Re: [PATCH 2/2] btrfs: zoned: wake up cleaner sooner if needed

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Mon, 22 Jan 2024 15:26:45 +0000
From: Naohiro Aota <Naohiro.Aota@....com>
To: Johannes Thumshirn <Johannes.Thumshirn@....com>
CC: Josef Bacik <josef@...icpanda.com>, David Sterba <dsterba@...e.com>,
	"linux-btrfs@...r.kernel.org" <linux-btrfs@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>, Damien Le Moal
	<dlemoal@...nel.org>
Subject: Re: Re: [PATCH 2/2] btrfs: zoned: wake up cleaner sooner if needed

On Mon, Jan 22, 2024 at 02:43:03PM +0000, Johannes Thumshirn wrote:
> On 22.01.24 15:39, Naohiro Aota wrote:
> >>>
> >>> Also, looking into btrfs_zoned_should_reclaim(), it sums device->bytes_used
> >>> for each fs_devices->devices. And, device->bytes_used is set at
> >>> create_chunk() or at btrfs_remove_chunk(). Isn't it feasible to do the
> >>> calculation only there?
> >>
> >> Oh sh*t! Right we should check bytes_used from all space_infos in
> >> btrfs_zoned_should_reclaim() and compare that to the disk total bytes.
> > 
> > You mean device->bytes_used? space_info->bytes_used does not count free
> > space and zone_unusable in BGs, so using that changes the behavior. Even,
> > it won't kick the thread if there are many zone_unusable but small used
> > space.
> > 
> 
> I did mean btrfs_space_info_used():
> 
> diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
> index b7e7b5a5a6fa..d5242c96c97c 100644
> --- a/fs/btrfs/zoned.c
> +++ b/fs/btrfs/zoned.c
> @@ -2414,6 +2414,7 @@ bool btrfs_zoned_should_reclaim(struct 
> btrfs_fs_info *fs_info)
>   {
>          struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
>          struct btrfs_device *device;
> +       struct btrfs_space_info *space_info;
>          u64 used = 0;
>          u64 total = 0;
>          u64 factor;
> @@ -2429,10 +2430,15 @@ bool btrfs_zoned_should_reclaim(struct 
> btrfs_fs_info *fs_info)
>                          continue;
> 
>                  total += device->disk_total_bytes;
> -               used += device->bytes_used;
>          }
>          rcu_read_unlock();
> 
> +       list_for_each_entry(space_info, &fs_info->space_info, list) {
> +               spin_lock(&space_info->lock);
> +               used += btrfs_space_info_used(space_info, true);
> +               spin_unlock(&space_info->lock);
> +       }
> +
>          factor = div64_u64(used * 100, total);
>          return factor >= fs_info->bg_reclaim_threshold;
>   }
> 

This changes the behavior. btrfs_space_info_used() excludes unallocated
space.

Also, if we calculate it with device->disk_total_bytes, it screws up on
DUP/RAID* profile because btrfs_space_info_used() counts the logical space
vs disk_total_bytes counts the physical space.