block/rnbd: private headers with rnbd protocol structs and helpers
authorJack Wang <jinpu.wang@cloud.ionos.com>
Mon, 11 May 2020 13:51:21 +0000 (15:51 +0200)
committerJason Gunthorpe <jgg@mellanox.com>
Sun, 17 May 2020 21:57:15 +0000 (18:57 -0300)
These are common private headers with rnbd protocol structures, logging,
sysfs and other helper functions, which are used on both client and server
sides.

Link: https://lore.kernel.org/r/20200511135131.27580-16-danil.kipnis@cloud.ionos.com
Signed-off-by: Danil Kipnis <danil.kipnis@cloud.ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Acked-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
drivers/block/rnbd/rnbd-common.c [new file with mode: 0644]
drivers/block/rnbd/rnbd-log.h [new file with mode: 0644]
drivers/block/rnbd/rnbd-proto.h [new file with mode: 0644]

diff --git a/drivers/block/rnbd/rnbd-common.c b/drivers/block/rnbd/rnbd-common.c
new file mode 100644 (file)
index 0000000..596c3f7
--- /dev/null
@@ -0,0 +1,23 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * RDMA Network Block Driver
+ *
+ * Copyright (c) 2014 - 2018 ProfitBricks GmbH. All rights reserved.
+ * Copyright (c) 2018 - 2019 1&1 IONOS Cloud GmbH. All rights reserved.
+ * Copyright (c) 2019 - 2020 1&1 IONOS SE. All rights reserved.
+ */
+#include "rnbd-proto.h"
+
+const char *rnbd_access_mode_str(enum rnbd_access_mode mode)
+{
+       switch (mode) {
+       case RNBD_ACCESS_RO:
+               return "ro";
+       case RNBD_ACCESS_RW:
+               return "rw";
+       case RNBD_ACCESS_MIGRATION:
+               return "migration";
+       default:
+               return "unknown";
+       }
+}
diff --git a/drivers/block/rnbd/rnbd-log.h b/drivers/block/rnbd/rnbd-log.h
new file mode 100644 (file)
index 0000000..136e7d6
--- /dev/null
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * RDMA Network Block Driver
+ *
+ * Copyright (c) 2014 - 2018 ProfitBricks GmbH. All rights reserved.
+ * Copyright (c) 2018 - 2019 1&1 IONOS Cloud GmbH. All rights reserved.
+ * Copyright (c) 2019 - 2020 1&1 IONOS SE. All rights reserved.
+ */
+#ifndef RNBD_LOG_H
+#define RNBD_LOG_H
+
+#include "rnbd-clt.h"
+#include "rnbd-srv.h"
+
+#define rnbd_clt_log(fn, dev, fmt, ...) (                              \
+               fn("<%s@%s> " fmt, (dev)->pathname,                     \
+               (dev)->sess->sessname,                                  \
+                  ##__VA_ARGS__))
+#define rnbd_srv_log(fn, dev, fmt, ...) (                              \
+                       fn("<%s@%s>: " fmt, (dev)->pathname,            \
+                          (dev)->sess->sessname, ##__VA_ARGS__))
+
+#define rnbd_clt_err(dev, fmt, ...)    \
+       rnbd_clt_log(pr_err, dev, fmt, ##__VA_ARGS__)
+#define rnbd_clt_err_rl(dev, fmt, ...) \
+       rnbd_clt_log(pr_err_ratelimited, dev, fmt, ##__VA_ARGS__)
+#define rnbd_clt_info(dev, fmt, ...) \
+       rnbd_clt_log(pr_info, dev, fmt, ##__VA_ARGS__)
+#define rnbd_clt_info_rl(dev, fmt, ...) \
+       rnbd_clt_log(pr_info_ratelimited, dev, fmt, ##__VA_ARGS__)
+
+#define rnbd_srv_err(dev, fmt, ...)    \
+       rnbd_srv_log(pr_err, dev, fmt, ##__VA_ARGS__)
+#define rnbd_srv_err_rl(dev, fmt, ...) \
+       rnbd_srv_log(pr_err_ratelimited, dev, fmt, ##__VA_ARGS__)
+#define rnbd_srv_info(dev, fmt, ...) \
+       rnbd_srv_log(pr_info, dev, fmt, ##__VA_ARGS__)
+#define rnbd_srv_info_rl(dev, fmt, ...) \
+       rnbd_srv_log(pr_info_ratelimited, dev, fmt, ##__VA_ARGS__)
+
+#endif /* RNBD_LOG_H */
diff --git a/drivers/block/rnbd/rnbd-proto.h b/drivers/block/rnbd/rnbd-proto.h
new file mode 100644 (file)
index 0000000..ca16624
--- /dev/null
@@ -0,0 +1,303 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * RDMA Network Block Driver
+ *
+ * Copyright (c) 2014 - 2018 ProfitBricks GmbH. All rights reserved.
+ * Copyright (c) 2018 - 2019 1&1 IONOS Cloud GmbH. All rights reserved.
+ * Copyright (c) 2019 - 2020 1&1 IONOS SE. All rights reserved.
+ */
+#ifndef RNBD_PROTO_H
+#define RNBD_PROTO_H
+
+#include <linux/types.h>
+#include <linux/blkdev.h>
+#include <linux/limits.h>
+#include <linux/inet.h>
+#include <linux/in.h>
+#include <linux/in6.h>
+#include <rdma/ib.h>
+
+#define RNBD_PROTO_VER_MAJOR 2
+#define RNBD_PROTO_VER_MINOR 0
+
+/* The default port number the RTRS server is listening on. */
+#define RTRS_PORT 1234
+
+/**
+ * enum rnbd_msg_types - RNBD message types
+ * @RNBD_MSG_SESS_INFO:        initial session info from client to server
+ * @RNBD_MSG_SESS_INFO_RSP:    initial session info from server to client
+ * @RNBD_MSG_OPEN:             open (map) device request
+ * @RNBD_MSG_OPEN_RSP:         response to an @RNBD_MSG_OPEN
+ * @RNBD_MSG_IO:               block IO request operation
+ * @RNBD_MSG_CLOSE:            close (unmap) device request
+ */
+enum rnbd_msg_type {
+       RNBD_MSG_SESS_INFO,
+       RNBD_MSG_SESS_INFO_RSP,
+       RNBD_MSG_OPEN,
+       RNBD_MSG_OPEN_RSP,
+       RNBD_MSG_IO,
+       RNBD_MSG_CLOSE,
+};
+
+/**
+ * struct rnbd_msg_hdr - header of RNBD messages
+ * @type:      Message type, valid values see: enum rnbd_msg_types
+ */
+struct rnbd_msg_hdr {
+       __le16          type;
+       __le16          __padding;
+};
+
+/**
+ * We allow to map RO many times and RW only once. We allow to map yet another
+ * time RW, if MIGRATION is provided (second RW export can be required for
+ * example for VM migration)
+ */
+enum rnbd_access_mode {
+       RNBD_ACCESS_RO,
+       RNBD_ACCESS_RW,
+       RNBD_ACCESS_MIGRATION,
+};
+
+/**
+ * struct rnbd_msg_sess_info - initial session info from client to server
+ * @hdr:               message header
+ * @ver:               RNBD protocol version
+ */
+struct rnbd_msg_sess_info {
+       struct rnbd_msg_hdr hdr;
+       u8              ver;
+       u8              reserved[31];
+};
+
+/**
+ * struct rnbd_msg_sess_info_rsp - initial session info from server to client
+ * @hdr:               message header
+ * @ver:               RNBD protocol version
+ */
+struct rnbd_msg_sess_info_rsp {
+       struct rnbd_msg_hdr hdr;
+       u8              ver;
+       u8              reserved[31];
+};
+
+/**
+ * struct rnbd_msg_open - request to open a remote device.
+ * @hdr:               message header
+ * @access_mode:       the mode to open remote device, valid values see:
+ *                     enum rnbd_access_mode
+ * @device_name:       device path on remote side
+ */
+struct rnbd_msg_open {
+       struct rnbd_msg_hdr hdr;
+       u8              access_mode;
+       u8              resv1;
+       s8              dev_name[NAME_MAX];
+       u8              reserved[3];
+};
+
+/**
+ * struct rnbd_msg_close - request to close a remote device.
+ * @hdr:       message header
+ * @device_id: device_id on server side to identify the device
+ */
+struct rnbd_msg_close {
+       struct rnbd_msg_hdr hdr;
+       __le32          device_id;
+};
+
+/**
+ * struct rnbd_msg_open_rsp - response message to RNBD_MSG_OPEN
+ * @hdr:               message header
+ * @device_id:         device_id on server side to identify the device
+ * @nsectors:          number of sectors in the usual 512b unit
+ * @max_hw_sectors:    max hardware sectors in the usual 512b unit
+ * @max_write_same_sectors: max sectors for WRITE SAME in the 512b unit
+ * @max_discard_sectors: max. sectors that can be discarded at once in 512b
+ * unit.
+ * @discard_granularity: size of the internal discard allocation unit in bytes
+ * @discard_alignment: offset from internal allocation assignment in bytes
+ * @physical_block_size: physical block size device supports in bytes
+ * @logical_block_size: logical block size device supports in bytes
+ * @max_segments:      max segments hardware support in one transfer
+ * @secure_discard:    supports secure discard
+ * @rotation:          is a rotational disc?
+ */
+struct rnbd_msg_open_rsp {
+       struct rnbd_msg_hdr     hdr;
+       __le32                  device_id;
+       __le64                  nsectors;
+       __le32                  max_hw_sectors;
+       __le32                  max_write_same_sectors;
+       __le32                  max_discard_sectors;
+       __le32                  discard_granularity;
+       __le32                  discard_alignment;
+       __le16                  physical_block_size;
+       __le16                  logical_block_size;
+       __le16                  max_segments;
+       __le16                  secure_discard;
+       u8                      rotational;
+       u8                      reserved[11];
+};
+
+/**
+ * struct rnbd_msg_io - message for I/O read/write
+ * @hdr:       message header
+ * @device_id: device_id on server side to find the right device
+ * @sector:    bi_sector attribute from struct bio
+ * @rw:                valid values are defined in enum rnbd_io_flags
+ * @bi_size:    number of bytes for I/O read/write
+ * @prio:       priority
+ */
+struct rnbd_msg_io {
+       struct rnbd_msg_hdr hdr;
+       __le32          device_id;
+       __le64          sector;
+       __le32          rw;
+       __le32          bi_size;
+       __le16          prio;
+};
+
+#define RNBD_OP_BITS  8
+#define RNBD_OP_MASK  ((1 << RNBD_OP_BITS) - 1)
+
+/**
+ * enum rnbd_io_flags - RNBD request types from rq_flag_bits
+ * @RNBD_OP_READ:           read sectors from the device
+ * @RNBD_OP_WRITE:          write sectors to the device
+ * @RNBD_OP_FLUSH:          flush the volatile write cache
+ * @RNBD_OP_DISCARD:        discard sectors
+ * @RNBD_OP_SECURE_ERASE:   securely erase sectors
+ * @RNBD_OP_WRITE_SAME:     write the same sectors many times
+
+ * @RNBD_F_SYNC:            request is sync (sync write or read)
+ * @RNBD_F_FUA:             forced unit access
+ */
+enum rnbd_io_flags {
+
+       /* Operations */
+
+       RNBD_OP_READ            = 0,
+       RNBD_OP_WRITE           = 1,
+       RNBD_OP_FLUSH           = 2,
+       RNBD_OP_DISCARD = 3,
+       RNBD_OP_SECURE_ERASE    = 4,
+       RNBD_OP_WRITE_SAME      = 5,
+
+       RNBD_OP_LAST,
+
+       /* Flags */
+
+       RNBD_F_SYNC  = 1<<(RNBD_OP_BITS + 0),
+       RNBD_F_FUA   = 1<<(RNBD_OP_BITS + 1),
+
+       RNBD_F_ALL   = (RNBD_F_SYNC | RNBD_F_FUA)
+
+};
+
+static inline u32 rnbd_op(u32 flags)
+{
+       return flags & RNBD_OP_MASK;
+}
+
+static inline u32 rnbd_flags(u32 flags)
+{
+       return flags & ~RNBD_OP_MASK;
+}
+
+static inline bool rnbd_flags_supported(u32 flags)
+{
+       u32 op;
+
+       op = rnbd_op(flags);
+       flags = rnbd_flags(flags);
+
+       if (op >= RNBD_OP_LAST)
+               return false;
+       if (flags & ~RNBD_F_ALL)
+               return false;
+
+       return true;
+}
+
+static inline u32 rnbd_to_bio_flags(u32 rnbd_opf)
+{
+       u32 bio_opf;
+
+       switch (rnbd_op(rnbd_opf)) {
+       case RNBD_OP_READ:
+               bio_opf = REQ_OP_READ;
+               break;
+       case RNBD_OP_WRITE:
+               bio_opf = REQ_OP_WRITE;
+               break;
+       case RNBD_OP_FLUSH:
+               bio_opf = REQ_OP_FLUSH | REQ_PREFLUSH;
+               break;
+       case RNBD_OP_DISCARD:
+               bio_opf = REQ_OP_DISCARD;
+               break;
+       case RNBD_OP_SECURE_ERASE:
+               bio_opf = REQ_OP_SECURE_ERASE;
+               break;
+       case RNBD_OP_WRITE_SAME:
+               bio_opf = REQ_OP_WRITE_SAME;
+               break;
+       default:
+               WARN(1, "Unknown RNBD type: %d (flags %d)\n",
+                    rnbd_op(rnbd_opf), rnbd_opf);
+               bio_opf = 0;
+       }
+
+       if (rnbd_opf & RNBD_F_SYNC)
+               bio_opf |= REQ_SYNC;
+
+       if (rnbd_opf & RNBD_F_FUA)
+               bio_opf |= REQ_FUA;
+
+       return bio_opf;
+}
+
+static inline u32 rq_to_rnbd_flags(struct request *rq)
+{
+       u32 rnbd_opf;
+
+       switch (req_op(rq)) {
+       case REQ_OP_READ:
+               rnbd_opf = RNBD_OP_READ;
+               break;
+       case REQ_OP_WRITE:
+               rnbd_opf = RNBD_OP_WRITE;
+               break;
+       case REQ_OP_DISCARD:
+               rnbd_opf = RNBD_OP_DISCARD;
+               break;
+       case REQ_OP_SECURE_ERASE:
+               rnbd_opf = RNBD_OP_SECURE_ERASE;
+               break;
+       case REQ_OP_WRITE_SAME:
+               rnbd_opf = RNBD_OP_WRITE_SAME;
+               break;
+       case REQ_OP_FLUSH:
+               rnbd_opf = RNBD_OP_FLUSH;
+               break;
+       default:
+               WARN(1, "Unknown request type %d (flags %llu)\n",
+                    req_op(rq), (unsigned long long)rq->cmd_flags);
+               rnbd_opf = 0;
+       }
+
+       if (op_is_sync(rq->cmd_flags))
+               rnbd_opf |= RNBD_F_SYNC;
+
+       if (op_is_flush(rq->cmd_flags))
+               rnbd_opf |= RNBD_F_FUA;
+
+       return rnbd_opf;
+}
+
+const char *rnbd_access_mode_str(enum rnbd_access_mode mode);
+
+#endif /* RNBD_PROTO_H */