Re: [PATCH v4 1/7] block: Prevent the use of REQ_FUA with read operations

From: Damien Le Moal <damien.lemoal@opensource.wdc.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: linux-ide@vger.kernel.org, linux-block@vger.kernel.org,
	Jens Axboe <axboe@kernel.dk>,
	"Maciej S . Szmigiero" <mail@maciej.szmigiero.name>,
	Hannes Reinecke <hare@suse.de>
Subject: Re: [PATCH v4 1/7] block: Prevent the use of REQ_FUA with read operations
Date: Wed, 2 Nov 2022 07:05:35 +0900	[thread overview]
Message-ID: <3af6895b-b776-cf0d-fe1e-866ce5e6b0b0@opensource.wdc.com> (raw)
In-Reply-To: <Y2E2wFnbeUzAPjo0@infradead.org>

On 11/2/22 00:09, Christoph Hellwig wrote:
> On Mon, Oct 31, 2022 at 11:26:36AM +0900, Damien Le Moal wrote:
>> +	/*
>> +	 * REQ_FUA does not apply to read requests because:
>> +	 * - There is no way to reliably force media access for read operations
>> +	 *   with a block device that does not support FUA.
>> +	 * - Not all block devices support FUA for read operations (e.g. ATA
>> +	 *   devices with NCQ support turned off).
>> +	 */
>> +	if (!op_is_write(rq->cmd_flags) && (rq->cmd_flags & REQ_FUA)) {
>> +		blk_mq_end_request(rq, BLK_STS_NOTSUPP);
> 
> How could this even happen?  If we want a debug check,  I think it
> should be in submit_bio and a WARN_ON_ONCE.

I have not found any code that issues a FUA read. So I do not think this
can happen at all currently. The check is about making sure that it
*never* happens.

I thought of having the check higher up in the submit path but I wanted to
avoid adding yet another check in the very hot path. But if you are OK
with that, I will move it.

-- 
Damien Le Moal
Western Digital Research