Re: [PATCH v3 1/5] erofs-utils: introduce dump.erofs

[Gao Xiang <hsiangkao@linux.alibaba.com>]

Hi,

On Sun, Sep 26, 2021 at 11:21:42AM +0800, Gao Xiang wrote:
> Hi Xuenan and Qi,
> 
> On Wed, Sep 15, 2021 at 05:35:33PM +0800, Guo Xuenan wrote:
> > From: Wang Qi <mpiglet@outlook.com>
> > 
> > Add dump-tool for erofs to facilitate users directly
> > analyzing the erofs image file.
> > 
> > Signed-off-by: Guo Xuenan <guoxuenan@huawei.com>
> > Signed-off-by: Wang Qi <mpiglet@outlook.com>
> 
> I'm almost fine with the series, and I will merge some of the patches
> later.
> 
> Due to busy work, my original plan was to fix some nits by myself and
> apply. Anyway, I will reply some comments this evening...

I've merged the first 2 patches into dev branch with modification (so no
need to resend the first two patches).

The rest patches are still a bit messy, I've set up a new
experimental-dump branch, please check out:
https://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git/?h=experimental-dump

There are some stuffs needing to be resolved in advance:
 1) rename all "dumpfs_" prefix into "erofsdump_";
 2) I feel still uncomfortable when reading get_path_by_nid() and
    erofs_read_dir(). Could we refactor them by introducing
    erofs_for_each_dir() or likewise?
 3) file_category_types and the switch in dumpfs_print_inode() are
    duplicated to me. I think one of them can be removed instead.
 4) please help using "filefrag -v -b1" style when printing extent info
    in erofsdump_show_inode_phy(), like below:

 ext:     logical_offset:        physical_offset: length:   expected: flags:
   0:        0..   20479: 21788270592..21788291071:  20480:             last,eof

Thanks,
Gao Xiang