On 20 Jun 2024, at 11:48, Chuck Lever wrote:
> On Thu, Jun 20, 2024 at 11:45:02AM -0400, Benjamin Coddington wrote:
>> On 20 Jun 2024, at 10:15, Christoph Hellwig wrote:
>>
>>> On Thu, Jun 20, 2024 at 09:52:59AM -0400, Benjamin Coddington wrote:
>>>> On 20 Jun 2024, at 1:06, Christoph Hellwig wrote:
>>>>
>>>>> On Wed, Jun 19, 2024 at 01:39:33PM -0400, cel@xxxxxxxxxx wrote:
>>>>>> - if (test_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags) == 0)
>>>>>> + if (test_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags) == 0) {
>>>>>
>>>>> It might be worth to invert this and keep the unavailable handling in
>>>>> the branch as that's the exceptional case. That code is also woefully
>>>>> under-documented and could have really used a comment.
>>>>
>>>> The transient device handling in general, or just this bit of it?
>>>
>>> Basically the code behind this NFS_DEVICEID_UNAVAILABLE check here.
>>
>> How about..
>
> Let's leave this as a separate patch. IMO this is dealing with an
> entirely orthogonal issue.
Agree - just wanted feedback on what's appropriate for comments in here. I
can send something after your work.
Ben
