On 6/30/21 12:29, Linus Torvalds wrote:
> On Tue, Jun 29, 2021 at 12:55 PM Jens Axboe <axboe@xxxxxxxxx> wrote:
>> Note that this will throw a trivial merge conflict in
>> drivers/nvme/host/fabrics.c due to the NVME_SC_HOST_PATH_ERROR addition.
> Grr.
>
> I don't mind the conflict. It's trivial, as you say.
>
> I *do* mind what the conflict unearthed: commit 63d20f54a3d0
> ("nvme-fabrics: remove extra new lines in the switch").
>
> That commit claims to remove new-lines. In fact, the full commit
> message explicitly says "No functionality change in this patch".
>
> Except what it does is not just non-functional whitespace cleanup. No,
> it adds that
>
> case NVME_SC_HOST_PATH_ERROR:
>
> thing that was added in mainline too by commit 4d9442bf263a
> ("nvme-fabrics: decode host pathing error for connect").
>
> THAT is not ok. Commit messages that explicitly say one thing, and
> then do something completely different are very bad.
>
> Linus
Indeed it can be misleading based on [1] 63d20f54a3d0and [2]4d9442bf263a.
Please let me know how I can help to fix this.
[1] http://lists.infradead.org/pipermail/linux-nvme/2021-May/025449.html
[2] http://lists.infradead.org/pipermail/linux-nvme/2021-May/025407.html
