hw/block/nvme: add simple copy command
v3 cover hw/block/nvme: add simple copy command
From: Klaus Jensen <[email protected]>
Add support for TP 4065 ("Simple Copy Command").
Changes for v3
* rebased on nvme-next
* changed the default msrc value to a more reasonable 127 from 255 to better align with the default mcl value of 128.
Changes for v2
* prefer style that aligns with existing NvmeIdCtrl field enums (Minwoo)
* swapped elbat/elbatm fields in copy source range. I've kept the R-b and A-b from Minwoo and Stefan since this is a non-functional change (the device does not use these fields at all).
Klaus Jensen (2):
nvme: updated shared header for copy command
hw/block/nvme: add simple copy command
hw/block/nvme-ns.h | 4 + hw/block/nvme.h | 1 + include/block/nvme.h | 45 ++++++++- hw/block/nvme-ns.c | 8 ++ hw/block/nvme.c | 224 +++++++++++++++++++++++++++++++++++++++++- hw/block/trace-events | 6 ++ 6 files changed, 285 insertions(+), 3 deletions(-)
Signature · 1 line
2.29.2
PATCH v3 1/2 nvme: updated shared header for copy command
From: Klaus Jensen <[email protected]>
Add new data structures and types for the Simple Copy command.
--- include/block/nvme.h | 45 ++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 43 insertions(+), 2 deletions(-)
include/block/nvme.h +43 -2
@@ -472,6 +472,7 @@ enum NvmeIoCommands {NVME_CMD_COMPARE =0x05 ,NVME_CMD_WRITE_ZEROES =0x08 ,NVME_CMD_DSM =0x09 , + NVME_CMD_COPY = 0x19, }; typedef struct QEMU_PACKED NvmeDeleteQ {
@@ -603,6 +604,35 @@ typedef struct QEMU_PACKED NvmeDsmRange {uint64_t slba; } NvmeDsmRange; +enum { +NVME_COPY_FORMAT_0 =0x0 , +}; + +typedef struct NvmeCopyCmd { +uint8_t opcode; +uint8_t flags; +uint16_t cid; +uint32_t nsid; +uint32_t rsvd2[4 ]; +NvmeCmdDptr dptr; +uint64_t sdlba; +uint32_t cdw12; +uint32_t cdw13; +uint32_t ilbrt; +uint16_t lbat; +uint16_t lbatm; +}NvmeCopyCmd ; + +typedef struct NvmeCopySourceRange { +uint8_t rsvd0[8 ]; +uint64_t slba; +uint16_t nlb; +uint8_t rsvd18[6 ]; +uint32_t eilbrt; +uint16_t elbat; +uint16_t elbatm; +}NvmeCopySourceRange ; +enum NvmeAsyncEventRequest {NVME_AER_TYPE_ERROR =0 ,NVME_AER_TYPE_SMART =1 ,
@@ -680,6 +710,7 @@ enum NvmeStatusCodes {NVME_CONFLICTING_ATTRS =0x0180 ,NVME_INVALID_PROT_INFO =0x0181 ,NVME_WRITE_TO_RO =0x0182 , +NVME_CMD_SIZE_LIMIT =0x0183 ,NVME_WRITE_FAULT =0x0280 ,NVME_UNRECOVERED_READ =0x0281 , NVME_E2E_GUARD_ERROR = 0x0282,
@@ -831,7 +862,7 @@ typedef struct QEMU_PACKED NvmeIdCtrl {uint8_t nvscc;uint8_t rsvd531;uint16_t acwu; -uint8_t rsvd534[2 ]; +uint16_t ocfs;uint32_t sgls;uint8_t rsvd540[228 ];uint8_t subnqn[256 ];
@@ -854,6 +885,11 @@ enum NvmeIdCtrlOncs {NVME_ONCS_FEATURES =1 <<4 ,NVME_ONCS_RESRVATIONS =1 <<5 ,NVME_ONCS_TIMESTAMP =1 <<6 , + NVME_ONCS_COPY = 1 << 8, +}; + +enum NvmeIdCtrlOcfs { +NVME_OCFS_COPY_FORMAT_0 =1 <<0 , }; enum NvmeIdCtrlFrmw {
@@ -995,7 +1031,10 @@ typedef struct QEMU_PACKED NvmeIdNs {uint16_t npdg;uint16_t npda;uint16_t nows; -uint8_t rsvd74[30 ]; +uint16_t mssrl; +uint32_t mcl; +uint8_t msrc; +uint8_t rsvd81[23 ];uint8_t nguid[16 ];uint64_t eui64;NvmeLBAF lbaf[16 ];
@@ -1059,6 +1098,7 @@ static inline void _nvme_check_size(void)QEMU_BUILD_BUG_ON (sizeof (NvmeAerResult) !=4 );QEMU_BUILD_BUG_ON (sizeof (NvmeCqe) !=16 );QEMU_BUILD_BUG_ON (sizeof (NvmeDsmRange) !=16 ); +QEMU_BUILD_BUG_ON (sizeof (NvmeCopySourceRange) !=32 );QEMU_BUILD_BUG_ON (sizeof (NvmeCmd) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeDeleteQ) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeCreateCq) !=64 );
@@ -1066,6 +1106,7 @@ static inline void _nvme_check_size(void)QEMU_BUILD_BUG_ON (sizeof (NvmeIdentify) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeRwCmd) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeDsmCmd) !=64 ); +QEMU_BUILD_BUG_ON (sizeof (NvmeCopyCmd) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeRangeType) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeErrorLog) !=64 );QEMU_BUILD_BUG_ON (sizeof (NvmeFwSlotInfoLog) !=512 );
Signature · 1 line
2.29.2
PATCH v3 2/2 hw/block/nvme: add simple copy command
In reply to Klaus JensenFrom: Klaus Jensen <[email protected]>
Add support for TP 4065a ("Simple Copy Command"), v2020.05.04 ("Ratified").
The implementation uses a bounce buffer to first read in the source logical blocks, then issue a write of that bounce buffer. The default maximum number of source logical blocks is 128, translating to 512 KiB for 4k logical blocks which aligns with the default value of MDTS.
--- hw/block/nvme-ns.h | 4 + hw/block/nvme.h | 1 + hw/block/nvme-ns.c | 8 ++ hw/block/nvme.c | 224 +++++++++++++++++++++++++++++++++++++++++- hw/block/trace-events | 6 ++ 5 files changed, 242 insertions(+), 1 deletion(-)
hw/block/nvme-ns.h +4 -0
@@ -21,6 +21,10 @@typedef struct NvmeNamespaceParams {uint32_t nsid; + +uint16_t mssrl; +uint32_t mcl; +uint8_t msrc; }NvmeNamespaceParams ; typedef struct NvmeNamespace {
hw/block/nvme.h +1 -0
@@ -62,6 +62,7 @@ static inline const char *nvme_io_opc_str(uint8_t opc)case NVME_CMD_READ :return "NVME_NVM_CMD_READ" ;case NVME_CMD_WRITE_ZEROES :return "NVME_NVM_CMD_WRITE_ZEROES" ;case NVME_CMD_DSM :return "NVME_NVM_CMD_DSM" ; +case NVME_CMD_COPY :return "NVME_NVM_CMD_COPY" ;default :return "NVME_NVM_CMD_UNKNOWN" ; } }
hw/block/nvme-ns.c +8 -0
@@ -59,6 +59,11 @@ static int nvme_ns_init(NvmeNamespace *ns, Error **errp) id_ns->npda = id_ns->npdg = npdg -1 ; +/* simple copy */ + id_ns->mssrl =cpu_to_le16 (ns->params.mssrl); + id_ns->mcl =cpu_to_le32 (ns->params.mcl); + id_ns->msrc = ns->params.msrc; +return 0 ; }
@@ -150,6 +155,9 @@ static void nvme_ns_realize(DeviceState *dev, Error **errp)static Property nvme_ns_props[] = {DEFINE_BLOCK_PROPERTIES (NvmeNamespace, blkconf),DEFINE_PROP_UINT32 ("nsid" , NvmeNamespace, params.nsid,0 ), +DEFINE_PROP_UINT16 ("mssrl" , NvmeNamespace, params.mssrl,128 ), +DEFINE_PROP_UINT32 ("mcl" , NvmeNamespace, params.mcl,128 ), +DEFINE_PROP_UINT8 ("msrc" , NvmeNamespace, params.msrc,127 ),DEFINE_PROP_END_OF_LIST (), };
hw/block/nvme.c +223 -1
250 lines not on this page · raw message
hw/block/trace-events +6 -0
@@ -42,6 +42,10 @@ pci_nvme_io_cmd(uint16_t cid, uint32_t nsid, uint16_t sqid, uint8_t opcode, cons pci_nvme_admin_cmd(uint16_t cid, uint16_t sqid, uint8_t opcode, const char *opname) "cid %"PRIu16" sqid %"PRIu16" opc 0x%"PRIx8" opname '%s'" pci_nvme_rw(uint16_t cid, const char *verb, uint32_t nsid, uint32_t nlb, uint64_t count, uint64_t lba) "cid %"PRIu16" opname '%s' nsid %"PRIu32" nlb %"PRIu32" count %"PRIu64" lba 0x%"PRIx64"" pci_nvme_rw_cb(uint16_t cid, const char *blkname) "cid %"PRIu16" blk '%s'" +pci_nvme_copy(uint16_t cid, uint32_t nsid, uint16_t nr, uint8_t format) "cid %"PRIu16" nsid %"PRIu32" nr %"PRIu16" format 0x%"PRIx8"" +pci_nvme_copy_source_range(uint64_t slba, uint32_t nlb) "slba 0x%"PRIx64" nlb %"PRIu32"" +pci_nvme_copy_in_complete(uint16_t cid) "cid %"PRIu16"" +pci_nvme_copy_cb(uint16_t cid) "cid %"PRIu16"" pci_nvme_write_zeroes(uint16_t cid, uint32_t nsid, uint64_t slba, uint32_t nlb) "cid %"PRIu16" nsid %"PRIu32" slba %"PRIu64" nlb %"PRIu32"" pci_nvme_block_status(int64_t offset, int64_t bytes, int64_t pnum, int ret, bool zeroed) "offset %"PRId64" bytes %"PRId64" pnum %"PRId64" ret 0x%x zeroed %d" pci_nvme_dsm(uint16_t cid, uint32_t nsid, uint32_t nr, uint32_t attr) "cid %"PRIu16" nsid %"PRIu32" nr %"PRIu32" attr 0x%"PRIx32""
@@ -49,6 +53,7 @@ pci_nvme_dsm_deallocate(uint16_t cid, uint32_t nsid, uint64_t slba, uint32_t nlb pci_nvme_compare(uint16_t cid, uint32_t nsid, uint64_t slba, uint32_t nlb) "cid %"PRIu16" nsid %"PRIu32" slba 0x%"PRIx64" nlb %"PRIu32"" pci_nvme_compare_cb(uint16_t cid) "cid %"PRIu16"" pci_nvme_aio_discard_cb(uint16_t cid) "cid %"PRIu16"" +pci_nvme_aio_copy_in_cb(uint16_t cid) "cid %"PRIu16"" pci_nvme_create_sq(uint64_t addr, uint16_t sqid, uint16_t cqid, uint16_t qsize, uint16_t qflags) "create submission queue, addr=0x%"PRIx64", sqid=%"PRIu16", cqid=%"PRIu16", qsize=%"PRIu16", qflags=%"PRIu16"" pci_nvme_create_cq(uint64_t addr, uint16_t cqid, uint16_t vector, uint16_t size, uint16_t qflags, int ien) "create completion queue, addr=0x%"PRIx64", cqid=%"PRIu16", vector=%"PRIu16", qsize=%"PRIu16", qflags=%"PRIu16", ien=%d" pci_nvme_del_sq(uint16_t qid) "deleting submission queue sqid=%"PRIu16""
@@ -99,6 +104,7 @@ pci_nvme_err_addr_read(uint64_t addr) "addr 0x%"PRIx64"" pci_nvme_err_addr_write(uint64_t addr) "addr 0x%"PRIx64"" pci_nvme_err_cfs(void) "controller fatal status" pci_nvme_err_aio(uint16_t cid, const char *errname, uint16_t status) "cid %"PRIu16" err '%s' status 0x%"PRIx16"" +pci_nvme_err_copy_invalid_format(uint8_t format) "format 0x%"PRIx8"" pci_nvme_err_invalid_sgld(uint16_t cid, uint8_t typ) "cid %"PRIu16" type 0x%"PRIx8"" pci_nvme_err_invalid_num_sgld(uint16_t cid, uint8_t typ) "cid %"PRIu16" type 0x%"PRIx8"" pci_nvme_err_invalid_sgl_excess_length(uint16_t cid) "cid %"PRIu16""
Signature · 1 line
2.29.2
Quoted 2 lines · Klaus Jensen
hw/block/nvme.c +2
-0+static uint16_t nvme_copy (NvmeCtrl *n,NvmeRequest *req) +{
<snip>
Quoted 22 lines
hw/block/nvme.c +22
-0+for (i =0 ; i < nr; i++) { +uint32_t _nlb =le16_to_cpu (range[i].nlb) +1 ; +if (_nlb >le16_to_cpu (ns->id_ns.mssrl)) { +return NVME_CMD_SIZE_LIMIT |NVME_DNR ; + } + + nlb += _nlb; + } + +if (nlb >le32_to_cpu (ns->id_ns.mcl)) { +return NVME_CMD_SIZE_LIMIT |NVME_DNR ; + } + + bounce = bouncep =g_malloc (nvme_l2b (ns, nlb)); + +for (i =0 ; i < nr; i++) { +uint64_t slba =le64_to_cpu (range[i].slba); +uint32_t nlb =le16_to_cpu (range[i].nlb) +1 ; + + status =nvme_check_bounds (ns, slba, nlb); +if (status) { +trace_pci_nvme_err_invalid_lba_range (slba, nlb, ns->id_ns.nsze);
hw/block/nvme.c static uint16_t nvme_dsm(NvmeCtrl *n, NvmeRequest *req) · line 1289
Quoted 10 lines
hw/block/nvme.c +10
-0+ goto free_bounce; + } + +if (NVME_ERR_REC_DULBE (ns->features.err_rec)) { + status =nvme_check_dulbe (ns, slba, nlb); +if (status) { + goto free_bounce; + } + } + }
Only comment I have is that these two for-loops look like they can be collaped into one, which also simplifies how you account for the bounce buffer when error'ing out.
Quoted 34 lines · Keith Busch
Quoted 2 lines · Klaus Jensen
hw/block/nvme.c +2
-0+static uint16_t nvme_copy (NvmeCtrl *n,NvmeRequest *req) +{<snip>
Quoted 28 lines
hw/block/nvme.c +28
-0+for (i =0 ; i < nr; i++) { +uint32_t _nlb =le16_to_cpu (range[i].nlb) +1 ; +if (_nlb >le16_to_cpu (ns->id_ns.mssrl)) { +return NVME_CMD_SIZE_LIMIT |NVME_DNR ; + } + + nlb += _nlb; + } + +if (nlb >le32_to_cpu (ns->id_ns.mcl)) { +return NVME_CMD_SIZE_LIMIT |NVME_DNR ; + } + + bounce = bouncep =g_malloc (nvme_l2b (ns, nlb)); + +for (i =0 ; i < nr; i++) { +uint64_t slba =le64_to_cpu (range[i].slba); +uint32_t nlb =le16_to_cpu (range[i].nlb) +1 ; + + status =nvme_check_bounds (ns, slba, nlb); +if (status) { +trace_pci_nvme_err_invalid_lba_range (slba, nlb, ns->id_ns.nsze); + goto free_bounce; + } + +if (NVME_ERR_REC_DULBE (ns->features.err_rec)) { + status =nvme_check_dulbe (ns, slba, nlb); +if (status) {
Quoted 9 lines
hw/block/nvme.c static uint16_t nvme_dsm(NvmeCtrl *n, NvmeRequest *req) · line 1288
Quoted 4 lines
hw/block/nvme.c +4
-0+ goto free_bounce; + } + } + }Only comment I have is that these two for-loops look like they can be collaped into one, which also simplifies how you account for the bounce buffer when error'ing out.
Yeah. And the shadowing of nlb is not good either. I'll fix it up.