Re: [PATCH v3] NFSv4: replace seqcount_t with a rw_semaphore

From: Trond Myklebust
Date: Wed Nov 02 2016 - 13:39:23 EST

> On Nov 2, 2016, at 13:11, Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> wrote:
> On 2016-10-31 16:11:02 [+0000], Trond Myklebust wrote:
>> Yes, and yes. We canât rely on the list pointers remaining correct, so we restart the list scan and we use the ops->state_flag_bit to signal whether or not state has been recovered for the entry being scanned.
> but this is tested at the top of the loop and by then you look at
> lists' ->next pointer which might be invalid.

No. We ensure we restart the list scan if we release the spinlock. Itâs safeâ