From: Douglas Gilbert <dgilbert@interlog.com>
To: linux-scsi@vger.kernel.org
Cc: martin.petersen@oracle.com, jejb@linux.vnet.ibm.com,
hare@suse.de, bvanassche@acm.org
Subject: [PATCH v25 40/44] sg: no_dxfer: move to/from kernel buffers
Date: Sun, 23 Oct 2022 23:20:54 -0400 [thread overview]
Message-ID: <20221024032058.14077-41-dgilbert@interlog.com> (raw)
In-Reply-To: <20221024032058.14077-1-dgilbert@interlog.com>
When the NO_DXFER flag is use on a command/request, the data-in
and data-out buffers (if present) should not be ignored. Add
sg_rq_map_kern() function to handle this. Uses a single bio with
multiple bvec_s usually each holding multiple pages, if necessary.
The driver default element size is 32 KiB so if PAGE_SIZE is 4096
then get_order()==3 .
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Douglas Gilbert <dgilbert@interlog.com>
---
drivers/scsi/sg.c | 78 +++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 76 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index f6acb1cffc57..94f55ce27c64 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -2278,6 +2278,24 @@ static const struct vm_operations_struct sg_mmap_vm_ops = {
.close = sg_vma_close,
};
+static inline void
+sg_bio_get(struct bio *bio)
+{
+ bio_get(bio);
+}
+
+static void
+sg_bio_put(struct bio *bio)
+{
+ if (unlikely(bio_flagged(bio, BIO_REFFED))) {
+ WARN_ON(!atomic_read(&bio->__bi_cnt));
+ if (!atomic_dec_and_test(&bio->__bi_cnt))
+ return;
+ }
+ bio_uninit(bio);
+ kfree(bio);
+}
+
/*
* Entry point for mmap(2) system call. For mmap(2) to work, request's
* scatter gather list needs to be order 0 which it is unlikely to be
@@ -2792,6 +2810,59 @@ exit_sg(void)
idr_destroy(&sg_index_idr);
}
+static struct bio *
+sg_mk_kern_bio(int bvec_cnt)
+{
+ struct bio *biop;
+
+ if (bvec_cnt > BIO_MAX_VECS)
+ return NULL;
+ biop = bio_kmalloc(bvec_cnt, GFP_ATOMIC);
+ if (!biop)
+ return NULL;
+ return biop;
+}
+
+/*
+ * Setup to move data between kernel buffers managed by this driver and a SCSI device. Note that
+ * there is no corresponding 'unmap' call as is required by blk_rq_map_user() . Uses a single
+ * bio with an expanded appended bvec if necessary.
+ */
+static int
+sg_rq_map_kern(struct sg_request *srp, struct request_queue *q, struct request *rqq, int rw_ind)
+{
+ struct sg_scatter_hold *schp = &srp->sgat_h;
+ struct bio *bio;
+ int k, ln;
+ int op_flags = 0;
+ int num_sgat = schp->num_sgat;
+ int dlen = schp->dlen;
+ int pg_sz = 1 << (PAGE_SHIFT + schp->page_order);
+
+ SG_LOG(4, srp->parentfp, "%s: dlen=%d, pg_sz=%d\n", __func__, dlen, pg_sz);
+ if (num_sgat <= 0)
+ return 0;
+ if (rw_ind == WRITE)
+ op_flags = REQ_SYNC | REQ_IDLE;
+ bio = sg_mk_kern_bio(num_sgat);
+ if (!bio)
+ return -ENOMEM;
+ bio_init(bio, NULL, bio->bi_inline_vecs, num_sgat, (req_op(rqq) | op_flags));
+ bio->bi_end_io = sg_bio_put;
+
+ for (k = 0; k < num_sgat && dlen > 0; ++k, dlen -= ln) {
+ ln = min_t(int, dlen, pg_sz);
+ if (bio_add_pc_page(q, bio, schp->pages[k], ln, 0) < ln) {
+ sg_bio_put(bio);
+ return -EINVAL;
+ }
+ }
+ /* used blk_rq_append_bio() before but this is simpler */
+ blk_rq_bio_prep(rqq, bio, num_sgat);
+ rqq->nr_phys_segments = (1 << schp->page_order) * num_sgat;
+ return 0;
+}
+
static inline void
sg_set_map_data(const struct sg_scatter_hold *schp, bool up_valid,
struct rq_map_data *mdp)
@@ -2916,8 +2987,11 @@ sg_start_req(struct sg_request *srp, struct sg_comm_wr_t *cwrp, int dxfer_dir)
md->from_user = (dxfer_dir == SG_DXFER_TO_FROM_DEV);
}
- res = blk_rq_map_user_io(rqq, md, up, dxfer_len, GFP_ATOMIC,
- iov_count, iov_count, 1, r0w);
+ if (us_xfer)
+ res = blk_rq_map_user_io(rqq, md, up, dxfer_len, GFP_ATOMIC,
+ iov_count, iov_count, 1, r0w);
+ else /* transfer data to/from kernel buffers */
+ res = sg_rq_map_kern(srp, q, rqq, r0w);
fini:
if (unlikely(res)) { /* failure, free up resources */
WRITE_ONCE(srp->rqq, NULL);
--
2.37.3
next prev parent reply other threads:[~2022-10-24 3:26 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-24 3:20 [PATCH v25 00/44] sg: add v4 interface Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 01/44] sg: move functions around Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 02/44] sg: remove typedefs, type+formatting cleanup Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 03/44] sg: sg_log and is_enabled Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 04/44] sg: remove typedefs, type+formatting cleanup Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 05/44] sg: bitops in sg_device Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 06/44] sg: make open count an atomic Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 07/44] sg: move header to uapi section Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 08/44] sg: speed sg_poll and sg_get_num_waiting Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 09/44] sg: sg_allow_if_err_recovery and renames Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 10/44] sg: change rwlock to spinlock Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 11/44] sg: ioctl handling Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 12/44] sg: split sg_read Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 13/44] sg: sg_common_write add structure for arguments Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 14/44] sg: rework sg_vma_fault Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 15/44] sg: rework sg_mmap Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 16/44] sg: replace sg_allow_access Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 17/44] sg: rework scatter gather handling Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 18/44] sg: introduce request state machine Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 19/44] sg: sg_find_srp_by_id Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 20/44] sg: sg_fill_request_element Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 21/44] sg: printk change %p to %pK Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 22/44] sg: xarray for fds in device Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 23/44] sg: xarray for reqs in fd Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 24/44] sg: replace rq array with xarray Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 25/44] sg: sense buffer rework Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 26/44] sg: add sg v4 interface support Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 27/44] sg: rework debug info Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 28/44] sg: add 8 byte SCSI LUN to sg_scsi_id Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 29/44] sg: expand sg_comm_wr_t Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 30/44] sg: add sg_iosubmit_v3 and sg_ioreceive_v3 ioctls Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 31/44] sg: move procfs objects to avoid forward decls Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 32/44] sg: protect multiple receivers Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 33/44] sg: first debugfs support Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 34/44] sg: rework mmap support Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 35/44] sg: defang allow_dio Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 36/44] sg: warn v3 write system call users Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 37/44] sg: add mmap_sz tracking Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 38/44] sg: track lowest inactive and await indexes Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 39/44] sg: remove unit attention check for device changed Douglas Gilbert
2022-10-24 3:20 ` Douglas Gilbert [this message]
2022-10-24 3:20 ` [PATCH v25 41/44] sg: add bio_poll support Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 42/44] sg: add statistics similar to st Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 43/44] sg: rework command completion when removed device Douglas Gilbert
2022-10-24 3:20 ` [PATCH v25 44/44] sg: bump version to 4.0.14 Douglas Gilbert
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 \
--in-reply-to=20221024032058.14077-41-dgilbert@interlog.com \
--to=dgilbert@interlog.com \
--cc=bvanassche@acm.org \
--cc=hare@suse.de \
--cc=jejb@linux.vnet.ibm.com \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
/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;
as well as URLs for NNTP newsgroup(s).