[rtems-libbsd commit] NVD(4): Add RTEMS-specific NVMe disk driver
Sebastian Huber
sebh at rtems.org
Thu Nov 14 06:05:46 UTC 2019
Module: rtems-libbsd
Branch: 5-freebsd-12
Commit: e7291f7ac34011e6cf0812054298dda9a8d34a76
Changeset: http://git.rtems.org/rtems-libbsd/commit/?id=e7291f7ac34011e6cf0812054298dda9a8d34a76
Author: Sebastian Huber <sebastian.huber at embedded-brains.de>
Date: Wed Sep 11 11:10:21 2019 +0200
NVD(4): Add RTEMS-specific NVMe disk driver
Update #3821.
---
libbsd.py | 6 +
rtemsbsd/sys/dev/nvd/nvd.c | 341 +++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 347 insertions(+)
diff --git a/libbsd.py b/libbsd.py
index ee4a87b..44b76ed 100644
--- a/libbsd.py
+++ b/libbsd.py
@@ -1650,6 +1650,12 @@ class nvme(builder.Module):
],
mm.generator['source']()
)
+ self.addRTEMSSourceFiles(
+ [
+ 'sys/dev/nvd/nvd.c',
+ ],
+ mm.generator['source']()
+ )
#
# Networking
diff --git a/rtemsbsd/sys/dev/nvd/nvd.c b/rtemsbsd/sys/dev/nvd/nvd.c
new file mode 100644
index 0000000..897b0af
--- /dev/null
+++ b/rtemsbsd/sys/dev/nvd/nvd.c
@@ -0,0 +1,341 @@
+#include <machine/rtems-bsd-kernel-space.h>
+
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
+ *
+ * Copyright (C) 2012-2016 Intel Corporation
+ * All rights reserved.
+ * Copyright (C) 2018 Alexander Motin <mav at FreeBSD.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+#include <sys/param.h>
+#include <sys/lock.h>
+#include <sys/mutex.h>
+#include <sys/kernel.h>
+#include <sys/malloc.h>
+#include <sys/module.h>
+#include <sys/queue.h>
+#include <sys/systm.h>
+
+#include <dev/nvme/nvme_private.h>
+
+#include <stdatomic.h>
+#include <rtems/blkdev.h>
+
+#define NVD_STR "nvd"
+
+struct nvd_controller;
+
+static int nvd_load(void);
+static void *nvd_new_disk(struct nvme_namespace *ns, void *ctrlr);
+static void *nvd_new_controller(struct nvme_controller *ctrlr);
+static void nvd_controller_fail(void *ctrlr);
+
+MALLOC_DEFINE(M_NVD, "nvd", "nvd(4) allocations");
+
+struct nvd_disk {
+ struct nvd_controller *ctrlr;
+ struct nvme_namespace *ns;
+ uint32_t lb_per_media_block;
+ u_int unit;
+ TAILQ_ENTRY(nvd_disk) global_tailq;
+ TAILQ_ENTRY(nvd_disk) ctrlr_tailq;
+};
+
+struct nvd_controller {
+ TAILQ_ENTRY(nvd_controller) tailq;
+ TAILQ_HEAD(, nvd_disk) disk_head;
+};
+
+static struct mtx nvd_lock;
+static TAILQ_HEAD(, nvd_controller) ctrlr_head;
+static TAILQ_HEAD(disk_list, nvd_disk) disk_head;
+
+static int
+nvd_modevent(module_t mod, int type, void *arg)
+{
+ int error = 0;
+
+ switch (type) {
+ case MOD_LOAD:
+ error = nvd_load();
+ break;
+ default:
+ break;
+ }
+
+ return (error);
+}
+
+moduledata_t nvd_mod = {
+ NVD_STR,
+ nvd_modevent,
+ 0
+};
+
+DECLARE_MODULE(nvd, nvd_mod, SI_SUB_DRIVERS, SI_ORDER_ANY);
+MODULE_VERSION(nvd, 1);
+MODULE_DEPEND(nvd, nvme, 1, 1, 1);
+
+static int
+nvd_load(void)
+{
+
+ mtx_init(&nvd_lock, "nvd_lock", NULL, MTX_DEF);
+ TAILQ_INIT(&ctrlr_head);
+ TAILQ_INIT(&disk_head);
+
+ nvme_register_consumer(nvd_new_disk,
+ nvd_new_controller, NULL, nvd_controller_fail);
+
+ return (0);
+}
+
+static void *
+nvd_new_controller(struct nvme_controller *ctrlr)
+{
+ struct nvd_controller *nvd_ctrlr;
+
+ nvd_ctrlr = malloc(sizeof(*nvd_ctrlr), M_NVD,
+ M_ZERO | M_WAITOK);
+
+ TAILQ_INIT(&nvd_ctrlr->disk_head);
+ mtx_lock(&nvd_lock);
+ TAILQ_INSERT_TAIL(&ctrlr_head, nvd_ctrlr, tailq);
+ mtx_unlock(&nvd_lock);
+
+ return (nvd_ctrlr);
+}
+
+#define NVD_BUFNUM_SHIFT 16
+
+#define NVD_BUFNUM_DEC (UINT32_C(1) << NVD_BUFNUM_SHIFT)
+
+static void
+nvd_request_done(rtems_blkdev_request *req)
+{
+ uint32_t prev;
+
+ prev = atomic_fetch_sub_explicit(&req->bufnum, NVD_BUFNUM_DEC,
+ memory_order_relaxed);
+
+ if ((prev >> NVD_BUFNUM_SHIFT) == 1) {
+ rtems_blkdev_request_done(req, req->status);
+ }
+}
+
+static void
+nvd_completion(void *arg, const struct nvme_completion *status)
+{
+ rtems_blkdev_request *req;
+
+ req = arg;
+
+ if (nvme_completion_is_error(status)) {
+ if (req->status == RTEMS_SUCCESSFUL) {
+ req->status = RTEMS_IO_ERROR;
+ }
+ }
+
+ nvd_request_done(req);
+}
+
+static int
+nvd_request(struct nvd_disk *ndisk, rtems_blkdev_request *req,
+ uint32_t media_blocks_per_block)
+{
+ uint32_t i;
+ uint32_t lb_count;
+ uint32_t bufnum;
+
+ BSD_ASSERT(req->req == RTEMS_BLKDEV_REQ_READ ||
+ req->req == RTEMS_BLKDEV_REQ_WRITE);
+ BSD_ASSERT(rtems_event_transient_receive(RTEMS_NO_WAIT, 0) == RTEMS_UNSATISFIED);
+ BSD_ASSERT(req->bufnum < NVD_BUFNUM_DEC);
+
+ req->status = RTEMS_SUCCESSFUL;
+ bufnum = req->bufnum;
+ req->bufnum |= bufnum << NVD_BUFNUM_SHIFT;
+ lb_count = media_blocks_per_block * ndisk->lb_per_media_block;
+
+ for (i = 0; i < bufnum; ++i) {
+ rtems_blkdev_sg_buffer *sg;
+ int error;
+
+ sg = &req->bufs[i];
+
+ if (req->req == RTEMS_BLKDEV_REQ_READ) {
+ error = nvme_ns_cmd_read(ndisk->ns, sg->buffer,
+ sg->block * ndisk->lb_per_media_block, lb_count,
+ nvd_completion, req);
+ } else {
+ error = nvme_ns_cmd_write(ndisk->ns, sg->buffer,
+ sg->block * ndisk->lb_per_media_block, lb_count,
+ nvd_completion, req);
+ }
+
+ if (error != 0) {
+ req->status = RTEMS_NO_MEMORY;
+ nvd_request_done(req);
+ }
+ }
+
+ return (0);
+}
+
+static void
+nvd_sync_completion(void *arg, const struct nvme_completion *status)
+{
+ rtems_status_code sc;
+
+ if (nvme_completion_is_error(status)) {
+ sc = RTEMS_IO_ERROR;
+ } else {
+ sc = RTEMS_SUCCESSFUL;
+ }
+
+ rtems_blkdev_request_done(arg, sc);
+}
+
+static int
+nvd_sync(struct nvd_disk *ndisk, rtems_blkdev_request *req)
+{
+ int error;
+
+ error = nvme_ns_cmd_flush(ndisk->ns, nvd_sync_completion, req);
+ if (error != 0) {
+ rtems_blkdev_request_done(req, RTEMS_NO_MEMORY);
+ }
+
+ return (0);
+}
+
+static int
+nvd_ioctl(rtems_disk_device *dd, uint32_t req, void *arg)
+{
+ struct nvd_disk *ndisk;
+
+ ndisk = rtems_disk_get_driver_data(dd);
+
+ if (req == RTEMS_BLKIO_REQUEST) {
+ return (nvd_request(ndisk, arg, dd->media_blocks_per_block));
+ }
+
+ if (req == RTEMS_BLKDEV_REQ_SYNC) {
+ return (nvd_sync(ndisk, arg));
+ }
+
+ if (req == RTEMS_BLKIO_CAPABILITIES) {
+ *(uint32_t *)arg = RTEMS_BLKDEV_CAP_SYNC;
+ return (0);
+ }
+
+ if (req == RTEMS_BLKIO_DELETED) {
+ panic("nvd_ioctl");
+ return (0);
+ }
+
+ return (rtems_blkdev_ioctl(dd, req, arg));
+}
+
+static void *
+nvd_new_disk(struct nvme_namespace *ns, void *arg)
+{
+ char path[64];
+ struct nvd_disk *ndisk;
+ struct nvd_disk *tnd;
+ struct nvd_controller *ctrlr;
+ int unit;
+ rtems_status_code sc;
+ uint32_t block_size;
+ uint32_t min_page_size;
+ rtems_blkdev_bnum block_count;
+
+ ctrlr = arg;
+ ndisk = malloc(sizeof(*ndisk), M_NVD, M_ZERO | M_WAITOK);
+ ndisk->ctrlr = ctrlr;
+ ndisk->ns = ns;
+
+ mtx_lock(&nvd_lock);
+ unit = 0;
+ TAILQ_FOREACH(tnd, &disk_head, global_tailq) {
+ if (tnd->unit > unit) {
+ break;
+ }
+ unit = tnd->unit + 1;
+ }
+ ndisk->unit = unit;
+ if (tnd != NULL) {
+ TAILQ_INSERT_BEFORE(tnd, ndisk, global_tailq);
+ } else {
+ TAILQ_INSERT_TAIL(&disk_head, ndisk, global_tailq);
+ }
+ TAILQ_INSERT_TAIL(&ctrlr->disk_head, ndisk, ctrlr_tailq);
+ mtx_unlock(&nvd_lock);
+
+ min_page_size = ndisk->ns->ctrlr->min_page_size;
+ block_size = nvme_ns_get_sector_size(ns);
+
+ if (block_size < min_page_size) {
+ ndisk->lb_per_media_block = min_page_size / block_size;
+ block_size = min_page_size;
+ } else {
+ ndisk->lb_per_media_block = 1;
+ }
+
+ block_count = nvme_ns_get_size(ns) / block_size;
+ snprintf(path, sizeof(path), "/dev/nvd%i", unit);
+ sc = rtems_blkdev_create(path, block_size, block_count, nvd_ioctl,
+ ndisk);
+ if (sc != RTEMS_SUCCESSFUL) {
+ panic("nvd_new_disk");
+ }
+
+ return (ndisk);
+}
+
+static void
+nvd_gone(struct nvd_disk *ndisk)
+{
+
+ panic("nvd_gone");
+}
+
+static void
+nvd_controller_fail(void *ctrlr_arg)
+{
+ struct nvd_controller *ctrlr = ctrlr_arg;
+ struct nvd_disk *ndisk;
+
+ mtx_lock(&nvd_lock);
+ TAILQ_REMOVE(&ctrlr_head, ctrlr, tailq);
+ TAILQ_FOREACH(ndisk, &ctrlr->disk_head, ctrlr_tailq)
+ nvd_gone(ndisk);
+ while (!TAILQ_EMPTY(&ctrlr->disk_head))
+ msleep(&ctrlr->disk_head, &nvd_lock, 0, "nvd_fail", 0);
+ mtx_unlock(&nvd_lock);
+ free(ctrlr, M_NVD);
+}
+
More information about the vc
mailing list