Add reservation acquire, reservation register,
reservation release and reservation report commands
in the nvme device layer.
By introducing these commands, this enables the nvme
device to perform reservation-related tasks, including
querying keys, querying reservation status, registering
reservation keys, initiating and releasing reservations,
as well as clearing and preempting reservations held by
other keys.
These commands are crucial for management and control of
shared storage resources in a persistent manner.
Signed-off-by: Changqi Lu <luchangqi.123@bytedance.com>
Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
Acked-by: Klaus Jensen <k.jensen@samsung.com>
---
hw/nvme/ctrl.c | 330 +++++++++++++++++++++++++++++++++++++++++++
hw/nvme/nvme.h | 4 +
include/block/nvme.h | 37 +++++
3 files changed, 371 insertions(+)
diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
index ad212de723..1f8a5659c9 100644
--- a/hw/nvme/ctrl.c
+++ b/hw/nvme/ctrl.c
@@ -294,6 +294,10 @@ static const uint32_t nvme_cse_iocs_nvm[256] = {
[NVME_CMD_COMPARE] = NVME_CMD_EFF_CSUPP,
[NVME_CMD_IO_MGMT_RECV] = NVME_CMD_EFF_CSUPP,
[NVME_CMD_IO_MGMT_SEND] = NVME_CMD_EFF_CSUPP | NVME_CMD_EFF_LBCC,
+ [NVME_CMD_RESV_REGISTER] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_REPORT] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_ACQUIRE] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_RELEASE] = NVME_CMD_EFF_CSUPP,
};
static const uint32_t nvme_cse_iocs_zoned[256] = {
@@ -308,6 +312,10 @@ static const uint32_t nvme_cse_iocs_zoned[256] = {
[NVME_CMD_ZONE_APPEND] = NVME_CMD_EFF_CSUPP | NVME_CMD_EFF_LBCC,
[NVME_CMD_ZONE_MGMT_SEND] = NVME_CMD_EFF_CSUPP | NVME_CMD_EFF_LBCC,
[NVME_CMD_ZONE_MGMT_RECV] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_REGISTER] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_REPORT] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_ACQUIRE] = NVME_CMD_EFF_CSUPP,
+ [NVME_CMD_RESV_RELEASE] = NVME_CMD_EFF_CSUPP,
};
static void nvme_process_sq(void *opaque);
@@ -1747,6 +1755,13 @@ static void nvme_aio_err(NvmeRequest *req, int ret)
case NVME_CMD_READ:
status = NVME_UNRECOVERED_READ;
break;
+ case NVME_CMD_RESV_REPORT:
+ if (ret == -ENOTSUP) {
+ status = NVME_INVALID_OPCODE;
+ } else {
+ status = NVME_UNRECOVERED_READ;
+ }
+ break;
case NVME_CMD_FLUSH:
case NVME_CMD_WRITE:
case NVME_CMD_WRITE_ZEROES:
@@ -1754,6 +1769,15 @@ static void nvme_aio_err(NvmeRequest *req, int ret)
case NVME_CMD_COPY:
status = NVME_WRITE_FAULT;
break;
+ case NVME_CMD_RESV_REGISTER:
+ case NVME_CMD_RESV_ACQUIRE:
+ case NVME_CMD_RESV_RELEASE:
+ if (ret == -ENOTSUP) {
+ status = NVME_INVALID_OPCODE;
+ } else {
+ status = NVME_WRITE_FAULT;
+ }
+ break;
default:
status = NVME_INTERNAL_DEV_ERROR;
break;
@@ -2692,6 +2716,304 @@ static uint16_t nvme_verify(NvmeCtrl *n, NvmeRequest *req)
return NVME_NO_COMPLETE;
}
+typedef struct NvmeKeyInfo {
+ uint64_t cr_key;
+ uint64_t nr_key;
+} NvmeKeyInfo;
+
+static uint16_t nvme_resv_register(NvmeCtrl *n, NvmeRequest *req)
+{
+ int ret;
+ NvmeKeyInfo key_info;
+ NvmeNamespace *ns = req->ns;
+ uint32_t cdw10 = le32_to_cpu(req->cmd.cdw10);
+ bool ignore_key = cdw10 >> 3 & 0x1;
+ uint8_t action = cdw10 & 0x7;
+ uint8_t ptpl = cdw10 >> 30 & 0x3;
+ bool aptpl;
+
+ switch (ptpl) {
+ case NVME_RESV_PTPL_NO_CHANGE:
+ aptpl = (ns->id_ns.rescap & NVME_PR_CAP_PTPL) ? true : false;
+ break;
+ case NVME_RESV_PTPL_DISABLE:
+ aptpl = false;
+ break;
+ case NVME_RESV_PTPL_ENABLE:
+ aptpl = true;
+ break;
+ default:
+ return NVME_INVALID_FIELD;
+ }
+
+ ret = nvme_h2c(n, (uint8_t *)&key_info, sizeof(NvmeKeyInfo), req);
+ if (ret) {
+ return ret;
+ }
+
+ switch (action) {
+ case NVME_RESV_REGISTER_ACTION_REGISTER:
+ req->aiocb = blk_aio_pr_register(ns->blkconf.blk, 0,
+ key_info.nr_key, 0, aptpl,
+ ignore_key, nvme_misc_cb,
+ req);
+ break;
+ case NVME_RESV_REGISTER_ACTION_UNREGISTER:
+ req->aiocb = blk_aio_pr_register(ns->blkconf.blk, key_info.cr_key, 0,
+ 0, aptpl, ignore_key,
+ nvme_misc_cb, req);
+ break;
+ case NVME_RESV_REGISTER_ACTION_REPLACE:
+ req->aiocb = blk_aio_pr_register(ns->blkconf.blk, key_info.cr_key,
+ key_info.nr_key, 0, aptpl, ignore_key,
+ nvme_misc_cb, req);
+ break;
+ default:
+ return NVME_INVALID_FIELD;
+ }
+
+ return NVME_NO_COMPLETE;
+}
+
+static uint16_t nvme_resv_release(NvmeCtrl *n, NvmeRequest *req)
+{
+ int ret;
+ uint64_t cr_key;
+ NvmeNamespace *ns = req->ns;
+ uint32_t cdw10 = le32_to_cpu(req->cmd.cdw10);
+ uint8_t action = cdw10 & 0x7;
+ NvmeResvType type = cdw10 >> 8 & 0xff;
+
+ ret = nvme_h2c(n, (uint8_t *)&cr_key, sizeof(cr_key), req);
+ if (ret) {
+ return ret;
+ }
+
+ switch (action) {
+ case NVME_RESV_RELEASE_ACTION_RELEASE:
+ req->aiocb = blk_aio_pr_release(ns->blkconf.blk, cr_key,
+ nvme_pr_type_to_block(type),
+ nvme_misc_cb, req);
+ break;
+ case NVME_RESV_RELEASE_ACTION_CLEAR:
+ req->aiocb = blk_aio_pr_clear(ns->blkconf.blk, cr_key,
+ nvme_misc_cb, req);
+ break;
+ default:
+ return NVME_INVALID_FIELD;
+ }
+
+ return NVME_NO_COMPLETE;
+}
+
+static uint16_t nvme_resv_acquire(NvmeCtrl *n, NvmeRequest *req)
+{
+ int ret;
+ NvmeKeyInfo key_info;
+ NvmeNamespace *ns = req->ns;
+ uint32_t cdw10 = le32_to_cpu(req->cmd.cdw10);
+ uint8_t action = cdw10 & 0x7;
+ NvmeResvType type = cdw10 >> 8 & 0xff;
+
+ ret = nvme_h2c(n, (uint8_t *)&key_info, sizeof(NvmeKeyInfo), req);
+ if (ret) {
+ return ret;
+ }
+
+ switch (action) {
+ case NVME_RESV_ACQUIRE_ACTION_ACQUIRE:
+ req->aiocb = blk_aio_pr_reserve(ns->blkconf.blk, key_info.cr_key,
+ nvme_pr_type_to_block(type),
+ nvme_misc_cb, req);
+ break;
+ case NVME_RESV_ACQUIRE_ACTION_PREEMPT:
+ req->aiocb = blk_aio_pr_preempt(ns->blkconf.blk,
+ key_info.cr_key, key_info.nr_key,
+ nvme_pr_type_to_block(type),
+ false, nvme_misc_cb, req);
+ break;
+ case NVME_RESV_ACQUIRE_ACTION_PREEMPT_AND_ABORT:
+ req->aiocb = blk_aio_pr_preempt(ns->blkconf.blk, key_info.cr_key,
+ key_info.nr_key, type, true,
+ nvme_misc_cb, req);
+ break;
+ default:
+ return NVME_INVALID_FIELD;
+ }
+
+ return NVME_NO_COMPLETE;
+}
+
+typedef struct NvmeResvKeys {
+ uint32_t generation;
+ uint32_t num_keys;
+ uint64_t *keys;
+ NvmeRequest *req;
+} NvmeResvKeys;
+
+typedef struct NvmeReadReservation {
+ uint32_t generation;
+ uint64_t key;
+ BlockPrType type;
+ NvmeRequest *req;
+ NvmeResvKeys *keys_info;
+} NvmeReadReservation;
+
+static int nvme_read_reservation_cb(NvmeReadReservation *reservation)
+{
+ int rc;
+ NvmeReservationStatus *nvme_status;
+ NvmeRequest *req = reservation->req;
+ NvmeCtrl *n = req->sq->ctrl;
+ NvmeResvKeys *keys_info = reservation->keys_info;
+ int len = sizeof(NvmeReservationStatusHeader) +
+ sizeof(NvmeRegisteredCtrl) * keys_info->num_keys;
+
+ nvme_status = g_malloc(len);
+ nvme_status->header.gen = reservation->generation;
+ nvme_status->header.rtype = block_pr_type_to_nvme(reservation->type);
+ nvme_status->header.regctl = keys_info->num_keys;
+ for (int i = 0; i < keys_info->num_keys; i++) {
+ nvme_status->regctl_ds[i].cntlid = nvme_ctrl(req)->cntlid;
+ nvme_status->regctl_ds[i].rkey = keys_info->keys[i];
+ nvme_status->regctl_ds[i].rcsts = keys_info->keys[i] ==
+ reservation->key ? 1 : 0;
+ /* hostid is not supported currently */
+ memset(&nvme_status->regctl_ds[i].hostid, 0, 8);
+ }
+
+ rc = nvme_c2h(n, (uint8_t *)nvme_status, len, req);
+ g_free(nvme_status);
+ return rc;
+}
+
+static int nvme_read_reservation_ext_cb(NvmeReadReservation *reservation)
+{
+ int rc;
+ NvmeReservationStatusExt *nvme_status_ext;
+ NvmeRequest *req = reservation->req;
+ NvmeCtrl *n = req->sq->ctrl;
+ NvmeResvKeys *keys_info = reservation->keys_info;
+ int len = sizeof(NvmeReservationStatusHeader) +
+ sizeof(uint8_t) * 40 +
+ sizeof(NvmeRegisteredCtrlExt) * keys_info->num_keys;
+
+ nvme_status_ext = g_malloc(len);
+ nvme_status_ext->header.gen = cpu_to_be32(reservation->generation);
+ nvme_status_ext->header.rtype = block_pr_type_to_nvme(reservation->type);
+ nvme_status_ext->header.regctl = cpu_to_be16(keys_info->num_keys);
+
+ for (int i = 0; i < keys_info->num_keys; i++) {
+ uint16_t ctnlid = nvme_ctrl(req)->cntlid;
+ nvme_status_ext->regctl_eds[i].cntlid = cpu_to_be16(ctnlid);
+ nvme_status_ext->regctl_eds[i].rkey = cpu_to_be64(keys_info->keys[i]);
+ nvme_status_ext->regctl_eds[i].rcsts = keys_info->keys[i] ==
+ reservation->key ? 1 : 0;
+ /* hostid is not supported currently */
+ memset(&nvme_status_ext->regctl_eds[i].hostid, 0, 16);
+ }
+
+ rc = nvme_c2h(n, (uint8_t *)nvme_status_ext, len, req);
+ g_free(nvme_status_ext);
+ return rc;
+}
+
+static void nvme_resv_read_reservation_cb(void *opaque, int ret)
+{
+ NvmeReadReservation *reservation = opaque;
+ NvmeRequest *req = reservation->req;
+ bool eds = req->cmd.cdw11 & 0x1;
+ NvmeResvKeys *keys_info = reservation->keys_info;
+
+ if (ret < 0) {
+ goto out;
+ }
+
+ if (eds) {
+ ret = nvme_read_reservation_ext_cb(reservation);
+ } else {
+ ret = nvme_read_reservation_cb(reservation);
+ }
+
+out:
+ g_free(keys_info->keys);
+ g_free(keys_info);
+ g_free(reservation);
+ nvme_misc_cb(req, ret);
+}
+
+static void nvme_resv_read_keys_cb(void *opaque, int ret)
+{
+ NvmeResvKeys *keys_info = opaque;
+ NvmeRequest *req = keys_info->req;
+ NvmeNamespace *ns = req->ns;
+ NvmeReadReservation *reservation;
+
+ if (ret < 0) {
+ goto out;
+ }
+
+ keys_info->num_keys = MIN(ret, keys_info->num_keys);
+ reservation = g_new0(NvmeReadReservation, 1);
+ memset(reservation, 0, sizeof(*reservation));
+ reservation->req = req;
+ reservation->keys_info = keys_info;
+
+ req->aiocb = blk_aio_pr_read_reservation(ns->blkconf.blk,
+ &reservation->generation, &reservation->key,
+ &reservation->type, nvme_resv_read_reservation_cb,
+ reservation);
+ return;
+
+out:
+ g_free(keys_info->keys);
+ g_free(keys_info);
+ nvme_misc_cb(req, ret);
+}
+
+
+static uint16_t nvme_resv_report(NvmeCtrl *n, NvmeRequest *req)
+{
+ int num_keys;
+ uint32_t cdw10 = req->cmd.cdw10;
+ uint32_t cdw11 = req->cmd.cdw11;
+ int buflen = (cdw10 + 1) * sizeof(uint32_t);
+ bool eds = cdw11 & 0x1;
+ NvmeNamespace *ns = req->ns;
+ NvmeResvKeys *keys_info;
+
+ if (eds) {
+ if (buflen < sizeof(NvmeReservationStatusHeader) +
+ sizeof(uint8_t) * 40) {
+ return NVME_INVALID_FIELD;
+ }
+
+ num_keys = (buflen - sizeof(NvmeReservationStatusHeader) -
+ sizeof(uint8_t) * 40) /
+ sizeof(struct NvmeRegisteredCtrlExt);
+ } else {
+ if (buflen < sizeof(NvmeReservationStatusHeader)) {
+ return NVME_INVALID_FIELD;
+ }
+
+ num_keys = (buflen - sizeof(NvmeReservationStatusHeader)) /
+ sizeof(struct NvmeRegisteredCtrl);
+ }
+
+ keys_info = g_new0(NvmeResvKeys, 1);
+ keys_info->generation = 0;
+ /* num_keys is the maximum number of keys that can be transmitted */
+ keys_info->num_keys = num_keys;
+ keys_info->keys = g_malloc(sizeof(uint64_t) * num_keys);
+ keys_info->req = req;
+
+ req->aiocb = blk_aio_pr_read_keys(ns->blkconf.blk, &keys_info->generation,
+ keys_info->num_keys, keys_info->keys,
+ nvme_resv_read_keys_cb, keys_info);
+
+ return NVME_NO_COMPLETE;
+}
+
typedef struct NvmeCopyAIOCB {
BlockAIOCB common;
BlockAIOCB *aiocb;
@@ -4469,6 +4791,14 @@ static uint16_t nvme_io_cmd(NvmeCtrl *n, NvmeRequest *req)
return nvme_dsm(n, req);
case NVME_CMD_VERIFY:
return nvme_verify(n, req);
+ case NVME_CMD_RESV_REGISTER:
+ return nvme_resv_register(n, req);
+ case NVME_CMD_RESV_REPORT:
+ return nvme_resv_report(n, req);
+ case NVME_CMD_RESV_ACQUIRE:
+ return nvme_resv_acquire(n, req);
+ case NVME_CMD_RESV_RELEASE:
+ return nvme_resv_release(n, req);
case NVME_CMD_COPY:
return nvme_copy(n, req);
case NVME_CMD_ZONE_MGMT_SEND:
diff --git a/hw/nvme/nvme.h b/hw/nvme/nvme.h
index 6d0e456348..fdd919c2b1 100644
--- a/hw/nvme/nvme.h
+++ b/hw/nvme/nvme.h
@@ -470,6 +470,10 @@ static inline const char *nvme_io_opc_str(uint8_t opc)
case NVME_CMD_ZONE_MGMT_SEND: return "NVME_ZONED_CMD_MGMT_SEND";
case NVME_CMD_ZONE_MGMT_RECV: return "NVME_ZONED_CMD_MGMT_RECV";
case NVME_CMD_ZONE_APPEND: return "NVME_ZONED_CMD_ZONE_APPEND";
+ case NVME_CMD_RESV_REGISTER: return "NVME_CMD_RESV_REGISTER";
+ case NVME_CMD_RESV_REPORT: return "NVME_CMD_RESV_REPORT";
+ case NVME_CMD_RESV_ACQUIRE: return "NVME_CMD_RESV_ACQUIRE";
+ case NVME_CMD_RESV_RELEASE: return "NVME_CMD_RESV_RELEASE";
default: return "NVME_NVM_CMD_UNKNOWN";
}
}
diff --git a/include/block/nvme.h b/include/block/nvme.h
index 9b9eaeb3a7..2f24570d4a 100644
--- a/include/block/nvme.h
+++ b/include/block/nvme.h
@@ -702,6 +702,43 @@ typedef enum NVMEPrCap {
NVME_PR_CAP_EX_AC_AR),
} NvmePrCap;
+typedef struct QEMU_PACKED NvmeRegisteredCtrl {
+ uint16_t cntlid;
+ uint8_t rcsts;
+ uint8_t rsvd3[5];
+ uint8_t hostid[8];
+ uint64_t rkey;
+} NvmeRegisteredCtrl;
+
+typedef struct QEMU_PACKED NvmeRegisteredCtrlExt {
+ uint16_t cntlid;
+ uint8_t rcsts;
+ uint8_t rsvd3[5];
+ uint64_t rkey;
+ uint8_t hostid[16];
+ uint8_t rsvd32[32];
+} NvmeRegisteredCtrlExt;
+
+typedef struct QEMU_PACKED NvmeReservationStatusHeader {
+ uint32_t gen;
+ uint8_t rtype;
+ uint16_t regctl;
+ uint16_t resv5;
+ uint8_t ptpls;
+ uint8_t resv10[14];
+} NvmeReservationStatusHeader;
+
+typedef struct QEMU_PACKED NvmeReservationStatus {
+ struct NvmeReservationStatusHeader header;
+ struct NvmeRegisteredCtrl regctl_ds[];
+} NvmeReservationStatus;
+
+typedef struct QEMU_PACKED NvmeReservationStatusExt {
+ struct NvmeReservationStatusHeader header;
+ uint8_t rsvd24[40];
+ struct NvmeRegisteredCtrlExt regctl_eds[];
+} NvmeReservationStatusExt;
+
typedef struct QEMU_PACKED NvmeDeleteQ {
uint8_t opcode;
uint8_t flags;
--
2.20.1On Aug 29 16:42, Changqi Lu wrote:
> Add reservation acquire, reservation register,
> reservation release and reservation report commands
> in the nvme device layer.
>
> By introducing these commands, this enables the nvme
> device to perform reservation-related tasks, including
> querying keys, querying reservation status, registering
> reservation keys, initiating and releasing reservations,
> as well as clearing and preempting reservations held by
> other keys.
>
> These commands are crucial for management and control of
> shared storage resources in a persistent manner.
> Signed-off-by: Changqi Lu <luchangqi.123@bytedance.com>
> Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
> Acked-by: Klaus Jensen <k.jensen@samsung.com>
> ---
> hw/nvme/ctrl.c | 330 +++++++++++++++++++++++++++++++++++++++++++
> hw/nvme/nvme.h | 4 +
> include/block/nvme.h | 37 +++++
> 3 files changed, 371 insertions(+)
>
> diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
> index ad212de723..1f8a5659c9 100644
> --- a/hw/nvme/ctrl.c
> +++ b/hw/nvme/ctrl.c
> @@ -1747,6 +1755,13 @@ static void nvme_aio_err(NvmeRequest *req, int ret)
> case NVME_CMD_READ:
> status = NVME_UNRECOVERED_READ;
> break;
> + case NVME_CMD_RESV_REPORT:
> + if (ret == -ENOTSUP) {
> + status = NVME_INVALID_OPCODE;
> + } else {
> + status = NVME_UNRECOVERED_READ;
> + }
> + break;
> case NVME_CMD_FLUSH:
> case NVME_CMD_WRITE:
> case NVME_CMD_WRITE_ZEROES:
> @@ -1754,6 +1769,15 @@ static void nvme_aio_err(NvmeRequest *req, int ret)
> case NVME_CMD_COPY:
> status = NVME_WRITE_FAULT;
> break;
> + case NVME_CMD_RESV_REGISTER:
> + case NVME_CMD_RESV_ACQUIRE:
> + case NVME_CMD_RESV_RELEASE:
> + if (ret == -ENOTSUP) {
> + status = NVME_INVALID_OPCODE;
> + } else {
> + status = NVME_WRITE_FAULT;
> + }
> + break;
> default:
> status = NVME_INTERNAL_DEV_ERROR;
> break;
This wasn't exactly what I had in mind ;) See below.
> @@ -2692,6 +2716,304 @@ static uint16_t nvme_verify(NvmeCtrl *n, NvmeRequest *req)
> return NVME_NO_COMPLETE;
> }
>
> +typedef struct NvmeKeyInfo {
> + uint64_t cr_key;
> + uint64_t nr_key;
> +} NvmeKeyInfo;
> +
> +static uint16_t nvme_resv_register(NvmeCtrl *n, NvmeRequest *req)
> +{
> + int ret;
> + NvmeKeyInfo key_info;
> + NvmeNamespace *ns = req->ns;
> + uint32_t cdw10 = le32_to_cpu(req->cmd.cdw10);
> + bool ignore_key = cdw10 >> 3 & 0x1;
> + uint8_t action = cdw10 & 0x7;
> + uint8_t ptpl = cdw10 >> 30 & 0x3;
> + bool aptpl;
> +
Here, check ns->id_ns.rescap for reservations support and return
NVME_INVALID_OPCODE. There is no need to call the block layer otherwise.
Same for the below commands.
© 2016 - 2024 Red Hat, Inc.