Re: [PATCH 1/2] scsi_host: add support for request batching

[Paolo Bonzini <pbonzini@redhat.com>]

On 30/05/19 17:36, Bart Van Assche wrote:
> On 5/30/19 4:28 AM, Paolo Bonzini wrote:
>> +static const struct blk_mq_ops scsi_mq_ops_no_commit = {
>> +    .get_budget    = scsi_mq_get_budget,
>> +    .put_budget    = scsi_mq_put_budget,
>> +    .queue_rq    = scsi_queue_rq,
>> +    .complete    = scsi_softirq_done,
>> +    .timeout    = scsi_timeout,
>> +#ifdef CONFIG_BLK_DEBUG_FS
>> +    .show_rq    = scsi_show_rq,
>> +#endif
>> +    .init_request    = scsi_mq_init_request,
>> +    .exit_request    = scsi_mq_exit_request,
>> +    .initialize_rq_fn = scsi_initialize_rq,
>> +    .busy        = scsi_mq_lld_busy,
>> +    .map_queues    = scsi_map_queues,
>> +};
>> +
>> +static void scsi_commit_rqs(struct blk_mq_hw_ctx *hctx)
>> +{
>> +    struct request_queue *q = hctx->queue;
>> +    struct scsi_device *sdev = q->queuedata;
>> +    struct Scsi_Host *shost = sdev->host;
>> +
>> +    shost->hostt->commit_rqs(shost, hctx->queue_num);
>> +}
>> +
>>   static const struct blk_mq_ops scsi_mq_ops = {
>>       .get_budget    = scsi_mq_get_budget,
>>       .put_budget    = scsi_mq_put_budget,
>>       .queue_rq    = scsi_queue_rq,
>> +    .commit_rqs    = scsi_commit_rqs,
>>       .complete    = scsi_softirq_done,
>>       .timeout    = scsi_timeout,
>>   #ifdef CONFIG_BLK_DEBUG_FS
> 
> Hi Paolo,
> 
> Have you considered to modify the block layer such that a single
> scsi_mq_ops structure can be used for all SCSI LLD types?

Yes, but I don't think it's possible to do it in a nice way.
Any adjustment we make to the block layer to fit the SCSI subsystem's
desires would make all other block drivers uglier, so I chose to confine
the ugliness here.

The root issue is that the SCSI subsystem is unique in how it sits on
top of the block layer; this is the famous "adapter" (or "midlayer",
though that is confusing when talking about SCSI) design that Linux
usually tries to avoid.

Paolo