The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: Gao Xiang <xiang@kernel.org>
To: Friendy Su <friendy.su@sony.com>
Cc: xiang@kernel.org, chao@kernel.org, jack@suse.cz,
	amir73il@gmail.com, linux-erofs@lists.ozlabs.org,
	linux-kernel@vger.kernel.org, Yuezhang Mo <Yuezhang.Mo@sony.com>,
	Daniel Palmer <daniel.palmer@sony.com>
Subject: Re: [PATCH v2] erofs: set SB_I_ALLOW_HSM for bdev-based mounts
Date: Thu, 19 Nov 2026 22:40:15 +0100	[thread overview]
Message-ID: <av9svyT3otgEEDgX@MacBookPro.fritz.box> (raw)
In-Reply-To: <20261023021635.2852395-1-friendy.su@sony.com>

Hi Friendy,

On Fri, Oct 23, 2026 at 10:16:35AM +0800, Friendy Su wrote:
> SB_I_ALLOW_HSM makes erofs support fanotify FAN_PRE_ACCESS event
> for bdev-based mounts. file-backed mounts do not support this.
> 
> This is for use case that erofs file data is not available on
> local block device at mount, but need to be filled at fanotify
> FAN_PRE_ACCESS event handler.
> 
> E.g. format erofs to blob mode, meta data on /dev/pmem0, file data
> on blob device /dev/pmem1. The file content in blob device is dummy.
> 
> mount as below:
> 
> $ sudo mount -o dax,device=/dev/pmem1 /dev/pmem0 /mnt/erofs/
> 
> In FAN_PRE_ACCESS handler, actual file content is fetched from remote
> then filled into the DAX mounted block device /dev/pmem1.


Set SB_I_ALLOW_HSM to enable fanotify FAN_PRE_ACCESS events for
bdev-backed mounts. File-backed mounts are not supported.

This allows EROFS to handle cases where file data is not available
locally at mount time and needs to be fetched on demand by a
FAN_PRE_ACCESS event handler.

For example, an EROFS image can use blob mode with metadata stored
on /dev/pmem0 and file data on /dev/pmem1, where the latter initially
contains dummy data.

The filesystem can be mounted with:

    mount -o dax,device=/dev/pmem1 /dev/pmem0 /mnt/erofs

On receiving a FAN_PRE_ACCESS event, the handler fetches the actual
file data from a remote source and fills the corresponding blocks
on /dev/pmem1 before access proceeds.


> 
> Signed-off-by: Friendy Su <friendy.su@sony.com>
> Reviewed-by: Yuezhang Mo <Yuezhang.Mo@sony.com>
> Reviewed-by: Daniel Palmer <daniel.palmer@sony.com>
> ---
>  fs/erofs/super.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
> index 0967aace9f7f..be8728b8d116 100644
> --- a/fs/erofs/super.c
> +++ b/fs/erofs/super.c
> @@ -667,6 +667,15 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
>  		errorfc(fc, "FSDAX is not allowed when inode_share is on");
>  		return -EINVAL;
>  	}
> +	if (!erofs_is_fileio_mode(sbi)) {
> +		/*
> +		 * Only enable HSM to support fanotify FAN_PRE_ACCESS event
> +		 * under block device backed mounts.
> +		 * For file-backed mounts, do not enable since erofs-utils
> +		 * has already supported FAN_PRE_ACCESS event on backend image.
> +		 */

Could you just change it to

		 * Follow other filesystems by supporting FAN_PRE_ACCESS events
		 * only on bdev-backed mount. Don't bother with file-backed mounts.

I hope we could avoid mentioning user-space details
in the kernel codebase directly.

Could you send a new version as I suggested, otherwise
it looks good to me,

Reviewed-by: Gao Xiang <xiang@kernel.org>

Thanks,
Gao Xiang

> +		sb->s_iflags |= SB_I_ALLOW_HSM;
> +	}
>  
>  	sbi->blkszbits = PAGE_SHIFT;
>  	if (erofs_is_fileio_mode(sbi)) {
> -- 
> 2.34.1
> 
> 

      parent reply	other threads:[~2026-11-19 21:40 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-23  2:16 [PATCH v2] erofs: set SB_I_ALLOW_HSM for bdev-based mounts Friendy Su
2026-10-23  2:32 ` Friendy.Su
2026-11-19 21:40 ` Gao Xiang [this message]

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=av9svyT3otgEEDgX@MacBookPro.fritz.box \
    --to=xiang@kernel.org \
    --cc=Yuezhang.Mo@sony.com \
    --cc=amir73il@gmail.com \
    --cc=chao@kernel.org \
    --cc=daniel.palmer@sony.com \
    --cc=friendy.su@sony.com \
    --cc=jack@suse.cz \
    --cc=linux-erofs@lists.ozlabs.org \
    --cc=linux-kernel@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