diff --git a/docs/system/devices/nvme.rst b/docs/system/devices/nvme.rst index 30f841ef62..a8bb8d729c 100644 --- a/docs/system/devices/nvme.rst +++ b/docs/system/devices/nvme.rst @@ -212,6 +212,41 @@ The namespace may be configured with additional parameters the minimum memory page size (CAP.MPSMIN). The default value (``0``) has this property inherit the ``mdts`` value. +Flexible Data Placement +----------------------- + +The device may be configured to support TP4146 ("Flexible Data Placement") by +configuring it (``fdp=on``) on the subsystem:: + + -device nvme-subsys,id=nvme-subsys-0,nqn=subsys0,fdp=on,fdp.nruh=16 + +The subsystem emulates a single Endurance Group, on which Flexible Data +Placement will be supported. Also note that the device emulation deviates +slightly from the specification, by always enabling the "FDP Mode" feature on +the controller if the subsystems is configured for Flexible Data Placement. + +Enabling Flexible Data Placement on the subsyste enables the following +parameters: + +``fdp.nrg`` (default: ``1``) + Set the number of Reclaim Groups. + +``fdp.nruh`` (default: ``0``) + Set the number of Reclaim Unit Handles. This is a mandatory paramater and + must be non-zero. + +``fdp.runs`` (default: ``96M``) + Set the Reclaim Unit Nominal Size. Defaults to 96 MiB. + +Namespaces within this subsystem may requests Reclaim Unit Handles:: + + -device nvme-ns,drive=nvm-1,fdp.ruhs=RUHLIST + +The ``RUHLIST`` is a semicolon separated list (i.e. ``0;1;2;3``) and may +include ranges (i.e. ``0;8-15``). If no reclaim unit handle list is specified, +the controller will assign the controller-specified reclaim unit handle to +placement handle identifier 0. + Metadata -------- @@ -320,4 +355,4 @@ controller are: .. code-block:: console - echo 0000:01:00.1 > /sys/bus/pci/drivers/nvme/bind \ No newline at end of file + echo 0000:01:00.1 > /sys/bus/pci/drivers/nvme/bind diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c index fd917fcda1..355668bdf8 100644 --- a/hw/nvme/ctrl.c +++ b/hw/nvme/ctrl.c @@ -43,7 +43,14 @@ * subsys= * -device nvme-ns,drive=,bus=,nsid=,\ * zoned=, \ - * subsys=,detached= + * subsys=,shared=, \ + * detached=, \ + * zoned.zone_size=, \ + * zoned.zone_capacity=, \ + * zoned.descr_ext_size=, \ + * zoned.max_active=, \ + * zoned.max_open=, \ + * zoned.cross_read= * * Note cmb_size_mb denotes size of CMB in MB. CMB is assumed to be at * offset 0 in BAR2 and supports only WDS, RDS and SQS for now. By default, the @@ -1748,6 +1755,7 @@ static void nvme_aio_err(NvmeRequest *req, int ret) case NVME_CMD_WRITE: case NVME_CMD_WRITE_ZEROES: case NVME_CMD_ZONE_APPEND: + case NVME_CMD_COPY: status = NVME_WRITE_FAULT; break; default: @@ -2847,6 +2855,25 @@ static void nvme_copy_source_range_parse(void *ranges, int idx, uint8_t format, } } +static inline uint16_t nvme_check_copy_mcl(NvmeNamespace *ns, + NvmeCopyAIOCB *iocb, uint16_t nr) +{ + uint32_t copy_len = 0; + + for (int idx = 0; idx < nr; idx++) { + uint32_t nlb; + nvme_copy_source_range_parse(iocb->ranges, idx, iocb->format, NULL, + &nlb, NULL, NULL, NULL); + copy_len += nlb + 1; + } + + if (copy_len > ns->id_ns.mcl) { + return NVME_CMD_SIZE_LIMIT | NVME_DNR; + } + + return NVME_SUCCESS; +} + static void nvme_copy_out_completed_cb(void *opaque, int ret) { NvmeCopyAIOCB *iocb = opaque; @@ -3159,6 +3186,11 @@ static uint16_t nvme_copy(NvmeCtrl *n, NvmeRequest *req) } } + status = nvme_check_copy_mcl(ns, iocb, nr); + if (status) { + goto invalid; + } + iocb->req = req; iocb->ret = 0; iocb->nr = nr; diff --git a/hw/nvme/ns.c b/hw/nvme/ns.c index 547c0b1543..44aba8f4d9 100644 --- a/hw/nvme/ns.c +++ b/hw/nvme/ns.c @@ -400,8 +400,9 @@ static bool nvme_ns_init_fdp(NvmeNamespace *ns, Error **errp) NvmeRuHandle *ruh; uint8_t lbafi = NVME_ID_NS_FLBAS_INDEX(ns->id_ns.flbas); g_autofree unsigned int *ruhids = NULL; - unsigned int *ruhid; - char *r, *p, *token; + unsigned int n, m, *ruhid; + const char *endptr, *token; + char *r, *p; uint16_t *ph; if (!ns->params.fdp.ruhs) { @@ -438,23 +439,55 @@ static bool nvme_ns_init_fdp(NvmeNamespace *ns, Error **errp) /* parse the placement handle identifiers */ while ((token = qemu_strsep(&p, ";")) != NULL) { - ns->fdp.nphs += 1; - if (ns->fdp.nphs > NVME_FDP_MAXPIDS || - ns->fdp.nphs == endgrp->fdp.nruh) { - error_setg(errp, "too many placement handles"); - free(r); - return false; - } - - if (qemu_strtoui(token, NULL, 0, ruhid++) < 0) { + if (qemu_strtoui(token, &endptr, 0, &n) < 0) { error_setg(errp, "cannot parse reclaim unit handle identifier"); free(r); return false; } + + m = n; + + /* parse range */ + if (*endptr == '-') { + token = endptr + 1; + + if (qemu_strtoui(token, NULL, 0, &m) < 0) { + error_setg(errp, "cannot parse reclaim unit handle identifier"); + free(r); + return false; + } + + if (m < n) { + error_setg(errp, "invalid reclaim unit handle identifier range"); + free(r); + return false; + } + } + + for (; n <= m; n++) { + if (ns->fdp.nphs++ == endgrp->fdp.nruh) { + error_setg(errp, "too many placement handles"); + free(r); + return false; + } + + *ruhid++ = n; + } } free(r); + /* verify that the ruhids are unique */ + for (unsigned int i = 0; i < ns->fdp.nphs; i++) { + for (unsigned int j = i + 1; j < ns->fdp.nphs; j++) { + if (ruhids[i] == ruhids[j]) { + error_setg(errp, "duplicate reclaim unit handle identifier: %u", + ruhids[i]); + return false; + } + } + } + ph = ns->fdp.phs = g_new(uint16_t, ns->fdp.nphs); ruhid = ruhids; diff --git a/hw/nvme/subsys.c b/hw/nvme/subsys.c index 24ddec860e..d30bb8bfd5 100644 --- a/hw/nvme/subsys.c +++ b/hw/nvme/subsys.c @@ -158,8 +158,10 @@ static bool nvme_subsys_setup_fdp(NvmeSubsystem *subsys, Error **errp) endgrp->fdp.nrg = subsys->params.fdp.nrg; - if (!subsys->params.fdp.nruh) { - error_setg(errp, "fdp.nruh must be non-zero"); + if (!subsys->params.fdp.nruh || + subsys->params.fdp.nruh > NVME_FDP_MAXPIDS) { + error_setg(errp, "fdp.nruh must be non-zero and less than %u", + NVME_FDP_MAXPIDS); return false; }