Re: [f2fs-dev] [PATCH] f2fs: cover f2fs_lock_op in expand_inode_data case

[Chao Yu <yuchao0@huawei.com>]

On 2019/12/19 3:53, Jaegeuk Kim wrote:
> We were missing to cover f2fs_lock_op in this case.

Jaegeuk,

generic/269 will hang with this patch, and also bugzilla reports android
with last code will hang with the same stack.

Could you check this patch?

Thanks,

> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
> ---
>  fs/f2fs/file.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> index 4ea9bf9e8701..0b74f94ac8ee 100644
> --- a/fs/f2fs/file.c
> +++ b/fs/f2fs/file.c
> @@ -1646,12 +1646,13 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
>  			if (err && err != -ENODATA && err != -EAGAIN)
>  				goto out_err;
>  		}
> -
> +		f2fs_lock_op(sbi);
>  		down_write(&sbi->pin_sem);
>  		map.m_seg_type = CURSEG_COLD_DATA_PINNED;
>  		f2fs_allocate_new_segments(sbi, CURSEG_COLD_DATA);
>  		err = f2fs_map_blocks(inode, &map, 1, F2FS_GET_BLOCK_PRE_DIO);
>  		up_write(&sbi->pin_sem);
> +		f2fs_unlock_op(sbi);
>  
>  		done += map.m_len;
>  		len -= map.m_len;
> @@ -1661,7 +1662,9 @@ static int expand_inode_data(struct inode *inode, loff_t offset,
>  
>  		map.m_len = done;
>  	} else {
> +		f2fs_lock_op(sbi);
>  		err = f2fs_map_blocks(inode, &map, 1, F2FS_GET_BLOCK_PRE_AIO);
> +		f2fs_unlock_op(sbi);
>  	}
>  out_err:
>  	if (err) {
> 


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel