Re: NFS workload leaves nfsd threads in D state

[Chuck Lever III <chuck.lever@xxxxxxxxxx>]

> On Jul 10, 2023, at 1:28 PM, Christoph Hellwig <hch@xxxxxx> wrote:
> 
> Only found a virtualized AHCI setup so far without much success.  Can
> you try this (from Chengming Zhou) in the meantime?
> 
> diff --git a/block/blk-flush.c b/block/blk-flush.c
> index dba392cf22bec6..5c392a277b9eb2 100644
> --- a/block/blk-flush.c
> +++ b/block/blk-flush.c
> @@ -443,7 +443,7 @@ bool blk_insert_flush(struct request *rq)
> * the post flush, and then just pass the command on.
> */
> blk_rq_init_flush(rq);
> - rq->flush.seq |= REQ_FSEQ_POSTFLUSH;
> + rq->flush.seq |= REQ_FSEQ_PREFLUSH;
> spin_lock_irq(&fq->mq_flush_lock);
> list_move_tail(&rq->flush.list, &fq->flush_data_in_flight);
> spin_unlock_irq(&fq->mq_flush_lock);

Thanks for the quick response. No change.

--
Chuck Lever