[dm-devel] Revert "dm space maps: don't reset space map allocation cursor when committing"

Mike Snitzer snitzer at redhat.com
Mon Feb 28 15:37:08 UTC 2022


On Mon, Feb 28 2022 at  9:13P -0500,
Luo Meng <luomeng12 at huawei.com> wrote:

> This reverts commit 5faafc77f7de69147d1e818026b9a0cbf036a7b2.
> 
> This commit 5faafc77f7de ("dm space maps: don't reset space map allocation
> cursor when committing") change the way to find free block.
> 
> But when use ramdisk(not support discard) for thin-pool,and storage
> over-commitment. Then constantly create and delete file, can find block
> in thin-pool, but can't find block in ramdisk.
> 
> So need revert this patch.
> 
> Fixes: 5faafc77f7de ("dm space maps: don't reset space map allocation cursor when committing")
> Signed-off-by: Luo Meng <luomeng12 at huawei.com>

What you're saying doesn't make any sense.  Especially when you
consider this last part of the commit message:
"Fix these issues by leaving the cursor alone, only resetting when the
 search hits the end of the metadata device."

If you've found that the cursor isn't being reset to the beginning
once the search hits the end of the metadata device, then that needs
fixing... but a revert of the  commit in question isn't appropriate.

BTW, it shouldn't matter whether or not the underlying ramdisk supports
discards.  The thin-pool controls the allocation of underlying storage.

So all said:

Nacked-by: Mike Snitzer <snitzer at redhat.com>




More information about the dm-devel mailing list