On 29/10/2024 11:30, Yu Kuai wrote:
Or is it guaranteed that IO error by atomic write won't hurt anyone,
user will handle this error and retry with non atomic write?
Yes, I think that the user could retry non-atomically for the same
write. Maybe returning a special error code could be useful for this.
And can you update the above error path comment when you support raid1
and raid10?
Sure, can do. I am not sure on a special error code value. I will think
about it.
And I will send update for
https://lore.kernel.org/linux-raid/20240903150748.2179966-1-john.g.garry@xxxxxxxxxx/T/#m5daa8d32d825d74422bbff272c9b25b6c4fc2788
soon with this suggestion.
Thanks,
John
