Re: [PATCH] erofs: handle 48-bit blocks_hi for compressed inodes

[Gao Xiang]

Hi Xusheng,

On 2026/6/22 15:34, Zhan Xusheng wrote:
> Combine i_nb.blocks_hi with i_u.blocks_lo when computing
> inode->i_blocks for compressed inodes, mirroring the startblk_hi
> handling for unencoded inodes a few lines above.  Also evaluate
> the shift in u64 to avoid truncation.
>
> Fixes: 2e1473d5195f ("erofs: implement 48-bit block addressing for unencoded inodes")

It's not the right fix, it should be efb2aef569b3 since
the pcluster layout only allows 32-bit addresssing for now.

> Signed-off-by: Zhan Xusheng <zhanxusheng@xxxxxxxxxx>
> ---
>   fs/erofs/inode.c | 5 +++--
>   1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
> index a188c570087a..cf2f00e13cae 100644
> --- a/fs/erofs/inode.c
> +++ b/fs/erofs/inode.c
> @@ -191,8 +191,9 @@ static int erofs_read_inode(struct inode *inode)
>   		err = -EFSCORRUPTED;
>   		goto err_out;
>   	} else {
> -		inode->i_blocks = le32_to_cpu(copied.i_u.blocks_lo) <<
> -				(sb->s_blocksize_bits - 9);
> +		inode->i_blocks = ((u64)le16_to_cpu(copied.i_nb.blocks_hi) << 32 |
> +				   le32_to_cpu(copied.i_u.blocks_lo)) <<
> +				  (sb->s_blocksize_bits - 9);

I hope it could be:
		inode->i_blocks = (le32_to_cpu(copied.i_u.blocks_lo) |
			((u64)le16_to_cpu(copied.i_nb.blocks_hi) << 32)) <<
				(sb->s_blocksize_bits - 9);

to explicitly use "()" and avoid overly long lines.

Thanks,
Gao Xiang

>   	}
>   
>   	if (vi->datalayout == EROFS_INODE_CHUNK_BASED) {