public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jens Axboe <[email protected]>
To: Kanchan Joshi <[email protected]>,
	Chaitanya Kulkarni <[email protected]>
Cc: "[email protected]" <[email protected]>,
	"[email protected]" <[email protected]>,
	"[email protected]" <[email protected]>,
	"[email protected]" <[email protected]>,
	"[email protected]" <[email protected]>,
	"[email protected]" <[email protected]>,
	"[email protected]" <[email protected]>,
	Anuj Gupta <[email protected]>
Subject: Re: [PATCH for-next v5 3/4] block: add helper to map bvec iterator for passthrough
Date: Thu, 8 Sep 2022 08:46:45 -0600	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <20220908105200.GB15034@test-zns>

On 9/8/22 4:52 AM, Kanchan Joshi wrote:
> On Wed, Sep 07, 2022 at 03:32:26PM +0000, Chaitanya Kulkarni wrote:
>> On 9/5/22 23:27, Kanchan Joshi wrote:
>>> Add blk_rq_map_user_bvec which maps the bvec iterator into a bio and
>>> places that into the request. This helper will be used in nvme for
>>> uring-passthrough with fixed-buffer.
>>> While at it, create another helper bio_map_get to reduce the code
>>> duplication.
>>>
>>> Signed-off-by: Kanchan Joshi <[email protected]>
>>> Signed-off-by: Anuj Gupta <[email protected]>
>>> ---
>>>   block/blk-map.c        | 94 +++++++++++++++++++++++++++++++++++++-----
>>>   include/linux/blk-mq.h |  1 +
>>>   2 files changed, 85 insertions(+), 10 deletions(-)
>>>
>>> diff --git a/block/blk-map.c b/block/blk-map.c
>>> index f3768876d618..e2f268167342 100644
>>> --- a/block/blk-map.c
>>> +++ b/block/blk-map.c
>>> @@ -241,17 +241,10 @@ static void bio_map_put(struct bio *bio)
>>>       }
>>>   }
>>>
>>> -static int bio_map_user_iov(struct request *rq, struct iov_iter *iter,
>>> +static struct bio *bio_map_get(struct request *rq, unsigned int nr_vecs,
>>>           gfp_t gfp_mask)
>>>   {
>>> -    unsigned int max_sectors = queue_max_hw_sectors(rq->q);
>>> -    unsigned int nr_vecs = iov_iter_npages(iter, BIO_MAX_VECS);
>>>       struct bio *bio;
>>> -    int ret;
>>> -    int j;
>>> -
>>> -    if (!iov_iter_count(iter))
>>> -        return -EINVAL;
>>>
>>>       if (rq->cmd_flags & REQ_POLLED) {
>>>           blk_opf_t opf = rq->cmd_flags | REQ_ALLOC_CACHE;
>>> @@ -259,13 +252,31 @@ static int bio_map_user_iov(struct request *rq, struct iov_iter *iter,
>>>           bio = bio_alloc_bioset(NULL, nr_vecs, opf, gfp_mask,
>>>                       &fs_bio_set);
>>>           if (!bio)
>>> -            return -ENOMEM;
>>> +            return NULL;
>>>       } else {
>>>           bio = bio_kmalloc(nr_vecs, gfp_mask);
>>>           if (!bio)
>>> -            return -ENOMEM;
>>> +            return NULL;
>>>           bio_init(bio, NULL, bio->bi_inline_vecs, nr_vecs, req_op(rq));
>>>       }
>>> +    return bio;
>>> +}
>>> +
>>> +static int bio_map_user_iov(struct request *rq, struct iov_iter *iter,
>>> +        gfp_t gfp_mask)
>>> +{
>>> +    unsigned int max_sectors = queue_max_hw_sectors(rq->q);
>>> +    unsigned int nr_vecs = iov_iter_npages(iter, BIO_MAX_VECS);
>>> +    struct bio *bio;
>>> +    int ret;
>>> +    int j;
>>> +
>>> +    if (!iov_iter_count(iter))
>>> +        return -EINVAL;
>>> +
>>> +    bio = bio_map_get(rq, nr_vecs, gfp_mask);
>>> +    if (bio == NULL)
>>> +        return -ENOMEM;
>>>
>>>       while (iov_iter_count(iter)) {
>>>           struct page **pages, *stack_pages[UIO_FASTIOV];
>>> @@ -612,6 +623,69 @@ int blk_rq_map_user(struct request_queue *q, struct request *rq,
>>>   }
>>>   EXPORT_SYMBOL(blk_rq_map_user);
>>>
>>> +/* Prepare bio for passthrough IO given an existing bvec iter */
>>> +int blk_rq_map_user_bvec(struct request *rq, struct iov_iter *iter)
>>> +{
>>> +    struct request_queue *q = rq->q;
>>> +    size_t iter_count, nr_segs;
>>> +    struct bio *bio;
>>> +    struct bio_vec *bv, *bvec_arr, *bvprvp = NULL;
>>> +    struct queue_limits *lim = &q->limits;
>>> +    unsigned int nsegs = 0, bytes = 0;
>>> +    int ret, i;
>>> +
>>
>> consider this (untested), it also sets the variable i data type same
>> as it comparison variable in nr_segs the loop i.e. size_t :-
>>
>> +       struct bio_vec *bv, *bvec_arr, *bvprvp = NULL;
>> +       struct request_queue *q = rq->q;
>> +       struct queue_limits *lim = &q->limits;
>> +       unsigned int nsegs = 0, bytes = 0;
>> +       size_t iter_count, nr_segs, i;
>> +       struct bio *bio;
>> +       int ret;
>>
>>
>>> +    iter_count = iov_iter_count(iter);
>>> +    nr_segs = iter->nr_segs;
>>> +
>>> +    if (!iter_count || (iter_count >> 9) > queue_max_hw_sectors(q))
>>> +        return -EINVAL;
>>
>> can we remove braces for iter_count >> 9 without impacting the intended
>> functionality?
> 
> I think removing that make it hard to read.
> I will fold all other changes you mentioned in v6.

Agree - if you have to think about operator precedence, then that's a
sign that the code is less readable and more fragile.

-- 
Jens Axboe



  reply	other threads:[~2022-09-08 14:47 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CGME20220906063719epcas5p3157e79583a5412a3be81f3d96f8aaadd@epcas5p3.samsung.com>
2022-09-06  6:27 ` [PATCH for-next v5 0/4] fixed-buffer for uring-cmd/passthru Kanchan Joshi
     [not found]   ` <CGME20220906063723epcas5p23946fd33031aee591210af1c3cd2d574@epcas5p2.samsung.com>
2022-09-06  6:27     ` [PATCH for-next v5 1/4] io_uring: introduce io_uring_cmd_import_fixed Kanchan Joshi
     [not found]   ` <CGME20220906063726epcas5p42f764b4c01b841dd1fc34abebcab02e6@epcas5p4.samsung.com>
2022-09-06  6:27     ` [PATCH for-next v5 2/4] io_uring: introduce fixed buffer support for io_uring_cmd Kanchan Joshi
     [not found]   ` <CGME20220906063729epcas5p1bf05e6873de0f7246234380d66c21fb9@epcas5p1.samsung.com>
2022-09-06  6:27     ` [PATCH for-next v5 3/4] block: add helper to map bvec iterator for passthrough Kanchan Joshi
2022-09-07 15:32       ` Chaitanya Kulkarni
2022-09-08 10:52         ` Kanchan Joshi
2022-09-08 14:46           ` Jens Axboe [this message]
2022-09-08 15:11           ` Pankaj Raghav
     [not found]   ` <CGME20220906063733epcas5p22984174bd6dbb2571152fea18af90924@epcas5p2.samsung.com>
2022-09-06  6:27     ` [PATCH for-next v5 4/4] nvme: wire up fixed buffer support for nvme passthrough Kanchan Joshi
2022-09-07 14:51       ` Chaitanya Kulkarni
2022-09-08 10:47         ` Kanchan Joshi
2022-09-08 14:50           ` Jens Axboe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox