Hi,

在 2024/01/09 7:38, Song Liu 写道:

On Tue, Sep 26, 2023 at 11:19 PM Yu Kuai [off-list ref] wrote:

quoted

From: Yu Kuai <redacted>

Before this patch, the implementation is hacky and hard to understand:

1) md_seq_start set pos to 1;
2) md_seq_show found pos is 1, then print Personalities;
3) md_seq_next found pos is 1, then it update pos to the first mddev;
4) md_seq_show found pos is not 1 or 2, show mddev;
5) md_seq_next found pos is not 1 or 2, update pos to next mddev;
6) loop 4-5 until the last mddev, then md_seq_next update pos to 2;
7) md_seq_show found pos is 2, then print unused devices;
8) md_seq_next found pos is 2, stop;

This patch remove the magic value and use seq_list_start/next/stop()
directly, and move printing "Personalities" to md_seq_start(),
"unsed devices" to md_seq_stop():

1) md_seq_start print Personalities, and then set pos to first mddev;
2) md_seq_show show mddev;
3) md_seq_next update pos to next mddev;
4) loop 2-3 until the last mddev;
5) md_seq_stop print unsed devices;

Signed-off-by: Yu Kuai <redacted>

Just realized this introduced a behavior change:

When there is not md devices, before this patch, we have

[root@eth50-1 ~]# cat /proc/mdstat
Personalities : [raid0] [raid1] [raid10] [raid6] [raid5] [raid4]
unused devices: <none>

After this patch, "cat /proc/mdstat" returns nothing. This causes
some confusion for users who want to read "Personalities" line,
for example, the mdadm test suite reads it.

I haven't figured out the best fix yet.

Yes, that's a problem. And after reviewing seq_read_iter() in detail, I
realize that I also can't use seq_printf() in m->op->start() directly,
because if seq buffer overflowed, md_seq_start() can be called more than
once.

I'll fix these problems soon.

Thanks,
Kuai

Thanks,
Song

.