Re: [RFC PATCH v3 3/3] nvme: wire up support for async passthrough

[Jens Axboe <axboe@kernel.dk>]

On 3/17/21 2:52 AM, Christoph Hellwig wrote:
>> +/*
>> + * This is carved within the block_uring_cmd, to avoid dynamic allocation.
>> + * Care should be taken not to grow this beyond what is available.
>> + */
>> +struct uring_cmd_data {
>> +	union {
>> +		struct bio *bio;
>> +		u64 result; /* nvme cmd result */
>> +	};
>> +	void *meta; /* kernel-resident buffer */
>> +	int status; /* nvme cmd status */
>> +};
>> +
>> +inline u64 *ucmd_data_addr(struct io_uring_cmd *ioucmd)
>> +{
>> +	return &(((struct block_uring_cmd *)&ioucmd->pdu)->unused[0]);
>> +}
> 
> The whole typing is a mess, but this mostly goes back to the series
> you're basing this on.  Jens, can you send out the series so that
> we can do a proper review?

I will post it soon, only reason I haven't reposted is that I'm not that
happy with how the sqe split is done (and that it's done in the first
place). But I'll probably just post the current version for comments,
and hopefully we can get it to where it needs to be soon.

-- 
Jens Axboe