OpenE2K
/
qemu-e2k
13
4
Fork 0
Code Issues 4 Pull Requests Projects 2 Releases Activity

hw/nvme fixes 

* new PMR test (Gollu Appalanaidu)
 * pmr/sgl mapping fix (Padmakar Kalghatgi)
 * hotplug fixes (me)
 * mmio out-of-bound read fix (me)
 * big-endian host fixes (me)
 -----BEGIN PGP SIGNATURE-----
 
 iQEzBAABCAAdFiEEUigzqnXi3OaiR2bATeGvMW1PDekFAmD/CnQACgkQTeGvMW1P
 DenrlQgAhiKhfKQAJPpP59noKaVNgHFCJySNFQdrPmnD8xDehn4wNRiP4pcdc4aL
 wpB28S5GN9LopzMpLtQBfQZaLygMPFa4QnuJ0ZFJnVoZZa5+T0yQUbBDt8xBJEM5
 gVXBeB4rjnP4TjyTipgtRhSUZt7wieul60z9oFUbJwg4ntt0Jp36DFRzSdXAhTXY
 MubNZy+jqf8p2R4icL97kZISI/phPbWBBMVDSmyqlU5E/pCnip/n5pdISJ/MTPgl
 amqBA93Zs9d2jzhZlhfiKMT+qtdUuItGUk8NV9L2IKxE7ZTdT+tujKfQ1c54JvGM
 J70NjthZApzV4EJwZyrqD0UXNYwFIg==
 =jXUw
 -----END PGP SIGNATURE-----

Merge remote-tracking branch 'remotes/nvme/tags/nvme-next-pull-request' into staging

hw/nvme fixes

* new PMR test (Gollu Appalanaidu)
* pmr/sgl mapping fix (Padmakar Kalghatgi)
* hotplug fixes (me)
* mmio out-of-bound read fix (me)
* big-endian host fixes (me)

# gpg: Signature made Mon 26 Jul 2021 20:18:12 BST
# gpg:                using RSA key 522833AA75E2DCE6A24766C04DE1AF316D4F0DE9
# gpg: Good signature from "Klaus Jensen <its@irrelevant.dk>" [unknown]
# gpg:                 aka "Klaus Jensen <k.jensen@samsung.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: DDCA 4D9C 9EF9 31CC 3468  4272 63D5 6FC5 E55D A838
#      Subkey fingerprint: 5228 33AA 75E2 DCE6 A247  66C0 4DE1 AF31 6D4F 0DE9

* remotes/nvme/tags/nvme-next-pull-request:
  tests/qtest/nvme-test: add mmio read test
  hw/nvme: fix mmio read
  hw/nvme: fix out-of-bounds reads
  hw/nvme: use symbolic names for registers
  hw/nvme: split pmrmsc register into upper and lower
  hw/nvme: fix controller hot unplugging
  tests/qtest/nvme-test: add persistent memory region test
  hw/nvme: error handling for too many mappings
  hw/nvme: unregister controller with subsystem at exit
  hw/nvme: mark nvme-subsys non-hotpluggable
  hw/nvme: remove NvmeCtrl parameter from ns setup/check functions

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This commit is contained in:
Peter Maydell 2021-07-27 13:24:56 +01:00
parent ca4b5ef371 9631a8ab21
commit 202abcd389
7 changed files with 402 additions and 207 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

379
hw/nvme/ctrl.c
View File

@ -439,10 +439,12 @@ static uint8_t nvme_sq_empty(NvmeSQueue *sq)
static void nvme_irq_check(NvmeCtrl *n) static void nvme_irq_check(NvmeCtrl *n)
{ {
uint32_t intms = ldl_le_p(&n->bar.intms);
if (msix_enabled(&(n->parent_obj))) { if (msix_enabled(&(n->parent_obj))) {
return; return;
} }
if (~n->bar.intms & n->irq_status) { if (~intms & n->irq_status) {
pci_irq_assert(&n->parent_obj); pci_irq_assert(&n->parent_obj);
} else { } else {
pci_irq_deassert(&n->parent_obj); pci_irq_deassert(&n->parent_obj);
@ -623,6 +625,10 @@ static uint16_t nvme_map_addr(NvmeCtrl *n, NvmeSg *sg, hwaddr addr, size_t len)
return NVME_INVALID_USE_OF_CMB | NVME_DNR; return NVME_INVALID_USE_OF_CMB | NVME_DNR;
} }
if (sg->iov.niov + 1 > IOV_MAX) {
goto max_mappings_exceeded;
}
if (cmb) { if (cmb) {
return nvme_map_addr_cmb(n, &sg->iov, addr, len); return nvme_map_addr_cmb(n, &sg->iov, addr, len);
} else { } else {
@ -634,9 +640,18 @@ static uint16_t nvme_map_addr(NvmeCtrl *n, NvmeSg *sg, hwaddr addr, size_t len)
return NVME_INVALID_USE_OF_CMB | NVME_DNR; return NVME_INVALID_USE_OF_CMB | NVME_DNR;
} }
if (sg->qsg.nsg + 1 > IOV_MAX) {
goto max_mappings_exceeded;
}
qemu_sglist_add(&sg->qsg, addr, len); qemu_sglist_add(&sg->qsg, addr, len);
return NVME_SUCCESS; return NVME_SUCCESS;
max_mappings_exceeded:
NVME_GUEST_ERR(pci_nvme_ub_too_many_mappings,
"number of mappings exceed 1024");
return NVME_INTERNAL_DEV_ERROR | NVME_DNR;
} }
static inline bool nvme_addr_is_dma(NvmeCtrl *n, hwaddr addr) static inline bool nvme_addr_is_dma(NvmeCtrl *n, hwaddr addr)
@ -1276,7 +1291,7 @@ static void nvme_post_cqes(void *opaque)
if (ret) { if (ret) {
trace_pci_nvme_err_addr_write(addr); trace_pci_nvme_err_addr_write(addr);
trace_pci_nvme_err_cfs(); trace_pci_nvme_err_cfs();
n->bar.csts = NVME_CSTS_FAILED; stl_le_p(&n->bar.csts, NVME_CSTS_FAILED);
break; break;
} }
QTAILQ_REMOVE(&cq->req_list, req, entry); QTAILQ_REMOVE(&cq->req_list, req, entry);
@ -4009,7 +4024,7 @@ static uint16_t nvme_create_sq(NvmeCtrl *n, NvmeRequest *req)
trace_pci_nvme_err_invalid_create_sq_sqid(sqid); trace_pci_nvme_err_invalid_create_sq_sqid(sqid);
return NVME_INVALID_QID | NVME_DNR; return NVME_INVALID_QID | NVME_DNR;
} }
if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) { if (unlikely(!qsize || qsize > NVME_CAP_MQES(ldq_le_p(&n->bar.cap)))) {
trace_pci_nvme_err_invalid_create_sq_size(qsize); trace_pci_nvme_err_invalid_create_sq_size(qsize);
return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR; return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR;
} }
@ -4195,7 +4210,7 @@ static uint16_t nvme_cmd_effects(NvmeCtrl *n, uint8_t csi, uint32_t buf_len,
return NVME_INVALID_FIELD | NVME_DNR; return NVME_INVALID_FIELD | NVME_DNR;
} }
switch (NVME_CC_CSS(n->bar.cc)) { switch (NVME_CC_CSS(ldl_le_p(&n->bar.cc))) {
case NVME_CC_CSS_NVM: case NVME_CC_CSS_NVM:
src_iocs = nvme_cse_iocs_nvm; src_iocs = nvme_cse_iocs_nvm;
/* fall through */ /* fall through */
@ -4357,7 +4372,7 @@ static uint16_t nvme_create_cq(NvmeCtrl *n, NvmeRequest *req)
trace_pci_nvme_err_invalid_create_cq_cqid(cqid); trace_pci_nvme_err_invalid_create_cq_cqid(cqid);
return NVME_INVALID_QID | NVME_DNR; return NVME_INVALID_QID | NVME_DNR;
} }
if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) { if (unlikely(!qsize || qsize > NVME_CAP_MQES(ldq_le_p(&n->bar.cap)))) {
trace_pci_nvme_err_invalid_create_cq_size(qsize); trace_pci_nvme_err_invalid_create_cq_size(qsize);
return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR; return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR;
} }
@ -5150,17 +5165,19 @@ static void nvme_update_dmrsl(NvmeCtrl *n)
static void nvme_select_iocs_ns(NvmeCtrl *n, NvmeNamespace *ns) static void nvme_select_iocs_ns(NvmeCtrl *n, NvmeNamespace *ns)
{ {
uint32_t cc = ldl_le_p(&n->bar.cc);
ns->iocs = nvme_cse_iocs_none; ns->iocs = nvme_cse_iocs_none;
switch (ns->csi) { switch (ns->csi) {
case NVME_CSI_NVM: case NVME_CSI_NVM:
if (NVME_CC_CSS(n->bar.cc) != NVME_CC_CSS_ADMIN_ONLY) { if (NVME_CC_CSS(cc) != NVME_CC_CSS_ADMIN_ONLY) {
ns->iocs = nvme_cse_iocs_nvm; ns->iocs = nvme_cse_iocs_nvm;
} }
break; break;
case NVME_CSI_ZONED: case NVME_CSI_ZONED:
if (NVME_CC_CSS(n->bar.cc) == NVME_CC_CSS_CSI) { if (NVME_CC_CSS(cc) == NVME_CC_CSS_CSI) {
ns->iocs = nvme_cse_iocs_zoned; ns->iocs = nvme_cse_iocs_zoned;
} else if (NVME_CC_CSS(n->bar.cc) == NVME_CC_CSS_NVM) { } else if (NVME_CC_CSS(cc) == NVME_CC_CSS_NVM) {
ns->iocs = nvme_cse_iocs_nvm; ns->iocs = nvme_cse_iocs_nvm;
} }
break; break;
@ -5497,7 +5514,7 @@ static void nvme_process_sq(void *opaque)
if (nvme_addr_read(n, addr, (void *)&cmd, sizeof(cmd))) { if (nvme_addr_read(n, addr, (void *)&cmd, sizeof(cmd))) {
trace_pci_nvme_err_addr_read(addr); trace_pci_nvme_err_addr_read(addr);
trace_pci_nvme_err_cfs(); trace_pci_nvme_err_cfs();
n->bar.csts = NVME_CSTS_FAILED; stl_le_p(&n->bar.csts, NVME_CSTS_FAILED);
break; break;
} }
nvme_inc_sq_head(sq); nvme_inc_sq_head(sq);
@ -5552,8 +5569,6 @@ static void nvme_ctrl_reset(NvmeCtrl *n)
n->aer_queued = 0; n->aer_queued = 0;
n->outstanding_aers = 0; n->outstanding_aers = 0;
n->qs_created = false; n->qs_created = false;
n->bar.cc = 0;
} }
static void nvme_ctrl_shutdown(NvmeCtrl *n) static void nvme_ctrl_shutdown(NvmeCtrl *n)
@ -5592,7 +5607,12 @@ static void nvme_select_iocs(NvmeCtrl *n)
static int nvme_start_ctrl(NvmeCtrl *n) static int nvme_start_ctrl(NvmeCtrl *n)
{ {
uint32_t page_bits = NVME_CC_MPS(n->bar.cc) + 12; uint64_t cap = ldq_le_p(&n->bar.cap);
uint32_t cc = ldl_le_p(&n->bar.cc);
uint32_t aqa = ldl_le_p(&n->bar.aqa);
uint64_t asq = ldq_le_p(&n->bar.asq);
uint64_t acq = ldq_le_p(&n->bar.acq);
uint32_t page_bits = NVME_CC_MPS(cc) + 12;
uint32_t page_size = 1 << page_bits; uint32_t page_size = 1 << page_bits;
if (unlikely(n->cq[0])) { if (unlikely(n->cq[0])) {
@ -5603,73 +5623,72 @@ static int nvme_start_ctrl(NvmeCtrl *n)
trace_pci_nvme_err_startfail_sq(); trace_pci_nvme_err_startfail_sq();
return -1; return -1;
} }
if (unlikely(!n->bar.asq)) { if (unlikely(!asq)) {
trace_pci_nvme_err_startfail_nbarasq(); trace_pci_nvme_err_startfail_nbarasq();
return -1; return -1;
} }
if (unlikely(!n->bar.acq)) { if (unlikely(!acq)) {
trace_pci_nvme_err_startfail_nbaracq(); trace_pci_nvme_err_startfail_nbaracq();
return -1; return -1;
} }
if (unlikely(n->bar.asq & (page_size - 1))) { if (unlikely(asq & (page_size - 1))) {
trace_pci_nvme_err_startfail_asq_misaligned(n->bar.asq); trace_pci_nvme_err_startfail_asq_misaligned(asq);
return -1; return -1;
} }
if (unlikely(n->bar.acq & (page_size - 1))) { if (unlikely(acq & (page_size - 1))) {
trace_pci_nvme_err_startfail_acq_misaligned(n->bar.acq); trace_pci_nvme_err_startfail_acq_misaligned(acq);
return -1; return -1;
} }
if (unlikely(!(NVME_CAP_CSS(n->bar.cap) & (1 << NVME_CC_CSS(n->bar.cc))))) { if (unlikely(!(NVME_CAP_CSS(cap) & (1 << NVME_CC_CSS(cc))))) {
trace_pci_nvme_err_startfail_css(NVME_CC_CSS(n->bar.cc)); trace_pci_nvme_err_startfail_css(NVME_CC_CSS(cc));
return -1; return -1;
} }
if (unlikely(NVME_CC_MPS(n->bar.cc) < if (unlikely(NVME_CC_MPS(cc) < NVME_CAP_MPSMIN(cap))) {
NVME_CAP_MPSMIN(n->bar.cap))) {
trace_pci_nvme_err_startfail_page_too_small( trace_pci_nvme_err_startfail_page_too_small(
NVME_CC_MPS(n->bar.cc), NVME_CC_MPS(cc),
NVME_CAP_MPSMIN(n->bar.cap)); NVME_CAP_MPSMIN(cap));
return -1; return -1;
} }
if (unlikely(NVME_CC_MPS(n->bar.cc) > if (unlikely(NVME_CC_MPS(cc) >
NVME_CAP_MPSMAX(n->bar.cap))) { NVME_CAP_MPSMAX(cap))) {
trace_pci_nvme_err_startfail_page_too_large( trace_pci_nvme_err_startfail_page_too_large(
NVME_CC_MPS(n->bar.cc), NVME_CC_MPS(cc),
NVME_CAP_MPSMAX(n->bar.cap)); NVME_CAP_MPSMAX(cap));
return -1; return -1;
} }
if (unlikely(NVME_CC_IOCQES(n->bar.cc) < if (unlikely(NVME_CC_IOCQES(cc) <
NVME_CTRL_CQES_MIN(n->id_ctrl.cqes))) { NVME_CTRL_CQES_MIN(n->id_ctrl.cqes))) {
trace_pci_nvme_err_startfail_cqent_too_small( trace_pci_nvme_err_startfail_cqent_too_small(
NVME_CC_IOCQES(n->bar.cc), NVME_CC_IOCQES(cc),
NVME_CTRL_CQES_MIN(n->bar.cap)); NVME_CTRL_CQES_MIN(cap));
return -1; return -1;
} }
if (unlikely(NVME_CC_IOCQES(n->bar.cc) > if (unlikely(NVME_CC_IOCQES(cc) >
NVME_CTRL_CQES_MAX(n->id_ctrl.cqes))) { NVME_CTRL_CQES_MAX(n->id_ctrl.cqes))) {
trace_pci_nvme_err_startfail_cqent_too_large( trace_pci_nvme_err_startfail_cqent_too_large(
NVME_CC_IOCQES(n->bar.cc), NVME_CC_IOCQES(cc),
NVME_CTRL_CQES_MAX(n->bar.cap)); NVME_CTRL_CQES_MAX(cap));
return -1; return -1;
} }
if (unlikely(NVME_CC_IOSQES(n->bar.cc) < if (unlikely(NVME_CC_IOSQES(cc) <
NVME_CTRL_SQES_MIN(n->id_ctrl.sqes))) { NVME_CTRL_SQES_MIN(n->id_ctrl.sqes))) {
trace_pci_nvme_err_startfail_sqent_too_small( trace_pci_nvme_err_startfail_sqent_too_small(
NVME_CC_IOSQES(n->bar.cc), NVME_CC_IOSQES(cc),
NVME_CTRL_SQES_MIN(n->bar.cap)); NVME_CTRL_SQES_MIN(cap));
return -1; return -1;
} }
if (unlikely(NVME_CC_IOSQES(n->bar.cc) > if (unlikely(NVME_CC_IOSQES(cc) >
NVME_CTRL_SQES_MAX(n->id_ctrl.sqes))) { NVME_CTRL_SQES_MAX(n->id_ctrl.sqes))) {
trace_pci_nvme_err_startfail_sqent_too_large( trace_pci_nvme_err_startfail_sqent_too_large(
NVME_CC_IOSQES(n->bar.cc), NVME_CC_IOSQES(cc),
NVME_CTRL_SQES_MAX(n->bar.cap)); NVME_CTRL_SQES_MAX(cap));
return -1; return -1;
} }
if (unlikely(!NVME_AQA_ASQS(n->bar.aqa))) { if (unlikely(!NVME_AQA_ASQS(aqa))) {
trace_pci_nvme_err_startfail_asqent_sz_zero(); trace_pci_nvme_err_startfail_asqent_sz_zero();
return -1; return -1;
} }
if (unlikely(!NVME_AQA_ACQS(n->bar.aqa))) { if (unlikely(!NVME_AQA_ACQS(aqa))) {
trace_pci_nvme_err_startfail_acqent_sz_zero(); trace_pci_nvme_err_startfail_acqent_sz_zero();
return -1; return -1;
} }
@ -5677,12 +5696,10 @@ static int nvme_start_ctrl(NvmeCtrl *n)
n->page_bits = page_bits; n->page_bits = page_bits;
n->page_size = page_size; n->page_size = page_size;
n->max_prp_ents = n->page_size / sizeof(uint64_t); n->max_prp_ents = n->page_size / sizeof(uint64_t);
n->cqe_size = 1 << NVME_CC_IOCQES(n->bar.cc); n->cqe_size = 1 << NVME_CC_IOCQES(cc);
n->sqe_size = 1 << NVME_CC_IOSQES(n->bar.cc); n->sqe_size = 1 << NVME_CC_IOSQES(cc);
nvme_init_cq(&n->admin_cq, n, n->bar.acq, 0, 0, nvme_init_cq(&n->admin_cq, n, acq, 0, 0, NVME_AQA_ACQS(aqa) + 1, 1);
NVME_AQA_ACQS(n->bar.aqa) + 1, 1); nvme_init_sq(&n->admin_sq, n, asq, 0, 0, NVME_AQA_ASQS(aqa) + 1);
nvme_init_sq(&n->admin_sq, n, n->bar.asq, 0, 0,
NVME_AQA_ASQS(n->bar.aqa) + 1);
nvme_set_timestamp(n, 0ULL); nvme_set_timestamp(n, 0ULL);
@ -5695,22 +5712,33 @@ static int nvme_start_ctrl(NvmeCtrl *n)
static void nvme_cmb_enable_regs(NvmeCtrl *n) static void nvme_cmb_enable_regs(NvmeCtrl *n)
{ {
NVME_CMBLOC_SET_CDPCILS(n->bar.cmbloc, 1); uint32_t cmbloc = ldl_le_p(&n->bar.cmbloc);
NVME_CMBLOC_SET_CDPMLS(n->bar.cmbloc, 1); uint32_t cmbsz = ldl_le_p(&n->bar.cmbsz);
NVME_CMBLOC_SET_BIR(n->bar.cmbloc, NVME_CMB_BIR);
NVME_CMBSZ_SET_SQS(n->bar.cmbsz, 1); NVME_CMBLOC_SET_CDPCILS(cmbloc, 1);
NVME_CMBSZ_SET_CQS(n->bar.cmbsz, 0); NVME_CMBLOC_SET_CDPMLS(cmbloc, 1);
NVME_CMBSZ_SET_LISTS(n->bar.cmbsz, 1); NVME_CMBLOC_SET_BIR(cmbloc, NVME_CMB_BIR);
NVME_CMBSZ_SET_RDS(n->bar.cmbsz, 1); stl_le_p(&n->bar.cmbloc, cmbloc);
NVME_CMBSZ_SET_WDS(n->bar.cmbsz, 1);
NVME_CMBSZ_SET_SZU(n->bar.cmbsz, 2); /* MBs */ NVME_CMBSZ_SET_SQS(cmbsz, 1);
NVME_CMBSZ_SET_SZ(n->bar.cmbsz, n->params.cmb_size_mb); NVME_CMBSZ_SET_CQS(cmbsz, 0);
NVME_CMBSZ_SET_LISTS(cmbsz, 1);
NVME_CMBSZ_SET_RDS(cmbsz, 1);
NVME_CMBSZ_SET_WDS(cmbsz, 1);
NVME_CMBSZ_SET_SZU(cmbsz, 2); /* MBs */
NVME_CMBSZ_SET_SZ(cmbsz, n->params.cmb_size_mb);
stl_le_p(&n->bar.cmbsz, cmbsz);
} }
static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data, static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
unsigned size) unsigned size)
{ {
uint64_t cap = ldq_le_p(&n->bar.cap);
uint32_t cc = ldl_le_p(&n->bar.cc);
uint32_t intms = ldl_le_p(&n->bar.intms);
uint32_t csts = ldl_le_p(&n->bar.csts);
uint32_t pmrsts = ldl_le_p(&n->bar.pmrsts);
if (unlikely(offset & (sizeof(uint32_t) - 1))) { if (unlikely(offset & (sizeof(uint32_t) - 1))) {
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_misaligned32, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_misaligned32,
"MMIO write not 32-bit aligned," "MMIO write not 32-bit aligned,"
@ -5727,65 +5755,77 @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
} }
switch (offset) { switch (offset) {
case 0xc: /* INTMS */ case NVME_REG_INTMS:
if (unlikely(msix_enabled(&(n->parent_obj)))) { if (unlikely(msix_enabled(&(n->parent_obj)))) {
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_intmask_with_msix, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_intmask_with_msix,
"undefined access to interrupt mask set" "undefined access to interrupt mask set"
" when MSI-X is enabled"); " when MSI-X is enabled");
/* should be ignored, fall through for now */ /* should be ignored, fall through for now */
} }
n->bar.intms |= data & 0xffffffff; intms |= data;
stl_le_p(&n->bar.intms, intms);
n->bar.intmc = n->bar.intms; n->bar.intmc = n->bar.intms;
trace_pci_nvme_mmio_intm_set(data & 0xffffffff, n->bar.intmc); trace_pci_nvme_mmio_intm_set(data & 0xffffffff, intms);
nvme_irq_check(n); nvme_irq_check(n);
break; break;
case 0x10: /* INTMC */ case NVME_REG_INTMC:
if (unlikely(msix_enabled(&(n->parent_obj)))) { if (unlikely(msix_enabled(&(n->parent_obj)))) {
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_intmask_with_msix, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_intmask_with_msix,
"undefined access to interrupt mask clr" "undefined access to interrupt mask clr"
" when MSI-X is enabled"); " when MSI-X is enabled");
/* should be ignored, fall through for now */ /* should be ignored, fall through for now */
} }
n->bar.intms &= ~(data & 0xffffffff); intms &= ~data;
stl_le_p(&n->bar.intms, intms);
n->bar.intmc = n->bar.intms; n->bar.intmc = n->bar.intms;
trace_pci_nvme_mmio_intm_clr(data & 0xffffffff, n->bar.intmc); trace_pci_nvme_mmio_intm_clr(data & 0xffffffff, intms);
nvme_irq_check(n); nvme_irq_check(n);
break; break;
case 0x14: /* CC */ case NVME_REG_CC:
trace_pci_nvme_mmio_cfg(data & 0xffffffff); trace_pci_nvme_mmio_cfg(data & 0xffffffff);
/* Windows first sends data, then sends enable bit */ /* Windows first sends data, then sends enable bit */
if (!NVME_CC_EN(data) && !NVME_CC_EN(n->bar.cc) && if (!NVME_CC_EN(data) && !NVME_CC_EN(cc) &&
!NVME_CC_SHN(data) && !NVME_CC_SHN(n->bar.cc)) !NVME_CC_SHN(data) && !NVME_CC_SHN(cc))
{ {
n->bar.cc = data; cc = data;
} }
if (NVME_CC_EN(data) && !NVME_CC_EN(n->bar.cc)) { if (NVME_CC_EN(data) && !NVME_CC_EN(cc)) {
n->bar.cc = data; cc = data;
/* flush CC since nvme_start_ctrl() needs the value */
stl_le_p(&n->bar.cc, cc);
if (unlikely(nvme_start_ctrl(n))) { if (unlikely(nvme_start_ctrl(n))) {
trace_pci_nvme_err_startfail(); trace_pci_nvme_err_startfail();
n->bar.csts = NVME_CSTS_FAILED; csts = NVME_CSTS_FAILED;
} else { } else {
trace_pci_nvme_mmio_start_success(); trace_pci_nvme_mmio_start_success();
n->bar.csts = NVME_CSTS_READY; csts = NVME_CSTS_READY;
} }
} else if (!NVME_CC_EN(data) && NVME_CC_EN(n->bar.cc)) { } else if (!NVME_CC_EN(data) && NVME_CC_EN(cc)) {
trace_pci_nvme_mmio_stopped(); trace_pci_nvme_mmio_stopped();
nvme_ctrl_reset(n); nvme_ctrl_reset(n);
n->bar.csts &= ~NVME_CSTS_READY; cc = 0;
csts &= ~NVME_CSTS_READY;
} }
if (NVME_CC_SHN(data) && !(NVME_CC_SHN(n->bar.cc))) {
if (NVME_CC_SHN(data) && !(NVME_CC_SHN(cc))) {
trace_pci_nvme_mmio_shutdown_set(); trace_pci_nvme_mmio_shutdown_set();
nvme_ctrl_shutdown(n); nvme_ctrl_shutdown(n);
n->bar.cc = data; cc = data;
n->bar.csts |= NVME_CSTS_SHST_COMPLETE; csts |= NVME_CSTS_SHST_COMPLETE;
} else if (!NVME_CC_SHN(data) && NVME_CC_SHN(n->bar.cc)) { } else if (!NVME_CC_SHN(data) && NVME_CC_SHN(cc)) {
trace_pci_nvme_mmio_shutdown_cleared(); trace_pci_nvme_mmio_shutdown_cleared();
n->bar.csts &= ~NVME_CSTS_SHST_COMPLETE; csts &= ~NVME_CSTS_SHST_COMPLETE;
n->bar.cc = data; cc = data;
} }
stl_le_p(&n->bar.cc, cc);
stl_le_p(&n->bar.csts, csts);
break; break;
case 0x1c: /* CSTS */ case NVME_REG_CSTS:
if (data & (1 << 4)) { if (data & (1 << 4)) {
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_ssreset_w1c_unsupported, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_ssreset_w1c_unsupported,
"attempted to W1C CSTS.NSSRO" "attempted to W1C CSTS.NSSRO"
@ -5796,7 +5836,7 @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
" of controller status"); " of controller status");
} }
break; break;
case 0x20: /* NSSR */ case NVME_REG_NSSR:
if (data == 0x4e564d65) { if (data == 0x4e564d65) {
trace_pci_nvme_ub_mmiowr_ssreset_unsupported(); trace_pci_nvme_ub_mmiowr_ssreset_unsupported();
} else { } else {
@ -5804,53 +5844,53 @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
return; return;
} }
break; break;
case 0x24: /* AQA */ case NVME_REG_AQA:
n->bar.aqa = data & 0xffffffff; stl_le_p(&n->bar.aqa, data);
trace_pci_nvme_mmio_aqattr(data & 0xffffffff); trace_pci_nvme_mmio_aqattr(data & 0xffffffff);
break; break;
case 0x28: /* ASQ */ case NVME_REG_ASQ:
n->bar.asq = size == 8 ? data : stn_le_p(&n->bar.asq, size, data);
(n->bar.asq & ~0xffffffffULL) | (data & 0xffffffff);
trace_pci_nvme_mmio_asqaddr(data); trace_pci_nvme_mmio_asqaddr(data);
break; break;
case 0x2c: /* ASQ hi */ case NVME_REG_ASQ + 4:
n->bar.asq = (n->bar.asq & 0xffffffff) | (data << 32); stl_le_p((uint8_t *)&n->bar.asq + 4, data);
trace_pci_nvme_mmio_asqaddr_hi(data, n->bar.asq); trace_pci_nvme_mmio_asqaddr_hi(data, ldq_le_p(&n->bar.asq));
break; break;
case 0x30: /* ACQ */ case NVME_REG_ACQ:
trace_pci_nvme_mmio_acqaddr(data); trace_pci_nvme_mmio_acqaddr(data);
n->bar.acq = size == 8 ? data : stn_le_p(&n->bar.acq, size, data);
(n->bar.acq & ~0xffffffffULL) | (data & 0xffffffff);
break; break;
case 0x34: /* ACQ hi */ case NVME_REG_ACQ + 4:
n->bar.acq = (n->bar.acq & 0xffffffff) | (data << 32); stl_le_p((uint8_t *)&n->bar.acq + 4, data);
trace_pci_nvme_mmio_acqaddr_hi(data, n->bar.acq); trace_pci_nvme_mmio_acqaddr_hi(data, ldq_le_p(&n->bar.acq));
break; break;
case 0x38: /* CMBLOC */ case NVME_REG_CMBLOC:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_cmbloc_reserved, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_cmbloc_reserved,
"invalid write to reserved CMBLOC" "invalid write to reserved CMBLOC"
" when CMBSZ is zero, ignored"); " when CMBSZ is zero, ignored");
return; return;
case 0x3C: /* CMBSZ */ case NVME_REG_CMBSZ:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_cmbsz_readonly, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_cmbsz_readonly,
"invalid write to read only CMBSZ, ignored"); "invalid write to read only CMBSZ, ignored");
return; return;
case 0x50: /* CMBMSC */ case NVME_REG_CMBMSC:
if (!NVME_CAP_CMBS(n->bar.cap)) { if (!NVME_CAP_CMBS(cap)) {
return; return;
} }
n->bar.cmbmsc = size == 8 ? data : stn_le_p(&n->bar.cmbmsc, size, data);
(n->bar.cmbmsc & ~0xffffffff) | (data & 0xffffffff);
n->cmb.cmse = false; n->cmb.cmse = false;
if (NVME_CMBMSC_CRE(data)) { if (NVME_CMBMSC_CRE(data)) {
nvme_cmb_enable_regs(n); nvme_cmb_enable_regs(n);
if (NVME_CMBMSC_CMSE(data)) { if (NVME_CMBMSC_CMSE(data)) {
hwaddr cba = NVME_CMBMSC_CBA(data) << CMBMSC_CBA_SHIFT; uint64_t cmbmsc = ldq_le_p(&n->bar.cmbmsc);
hwaddr cba = NVME_CMBMSC_CBA(cmbmsc) << CMBMSC_CBA_SHIFT;
if (cba + int128_get64(n->cmb.mem.size) < cba) { if (cba + int128_get64(n->cmb.mem.size) < cba) {
NVME_CMBSTS_SET_CBAI(n->bar.cmbsts, 1); uint32_t cmbsts = ldl_le_p(&n->bar.cmbsts);
NVME_CMBSTS_SET_CBAI(cmbsts, 1);
stl_le_p(&n->bar.cmbsts, cmbsts);
return; return;
} }
@ -5863,53 +5903,57 @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
} }
return; return;
case 0x54: /* CMBMSC hi */ case NVME_REG_CMBMSC + 4:
n->bar.cmbmsc = (n->bar.cmbmsc & 0xffffffff) | (data << 32); stl_le_p((uint8_t *)&n->bar.cmbmsc + 4, data);
return; return;
case 0xe00: /* PMRCAP */ case NVME_REG_PMRCAP:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrcap_readonly, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrcap_readonly,
"invalid write to PMRCAP register, ignored"); "invalid write to PMRCAP register, ignored");
return; return;
case 0xe04: /* PMRCTL */ case NVME_REG_PMRCTL:
if (!NVME_CAP_PMRS(n->bar.cap)) { if (!NVME_CAP_PMRS(cap)) {
return; return;
} }
n->bar.pmrctl = data; stl_le_p(&n->bar.pmrctl, data);
if (NVME_PMRCTL_EN(data)) { if (NVME_PMRCTL_EN(data)) {
memory_region_set_enabled(&n->pmr.dev->mr, true); memory_region_set_enabled(&n->pmr.dev->mr, true);
n->bar.pmrsts = 0; pmrsts = 0;
} else { } else {
memory_region_set_enabled(&n->pmr.dev->mr, false); memory_region_set_enabled(&n->pmr.dev->mr, false);
NVME_PMRSTS_SET_NRDY(n->bar.pmrsts, 1); NVME_PMRSTS_SET_NRDY(pmrsts, 1);
n->pmr.cmse = false; n->pmr.cmse = false;
} }
stl_le_p(&n->bar.pmrsts, pmrsts);
return; return;
case 0xe08: /* PMRSTS */ case NVME_REG_PMRSTS:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrsts_readonly, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrsts_readonly,
"invalid write to PMRSTS register, ignored"); "invalid write to PMRSTS register, ignored");
return; return;
case 0xe0C: /* PMREBS */ case NVME_REG_PMREBS:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrebs_readonly, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrebs_readonly,
"invalid write to PMREBS register, ignored"); "invalid write to PMREBS register, ignored");
return; return;
case 0xe10: /* PMRSWTP */ case NVME_REG_PMRSWTP:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrswtp_readonly, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_pmrswtp_readonly,
"invalid write to PMRSWTP register, ignored"); "invalid write to PMRSWTP register, ignored");
return; return;
case 0xe14: /* PMRMSCL */ case NVME_REG_PMRMSCL:
if (!NVME_CAP_PMRS(n->bar.cap)) { if (!NVME_CAP_PMRS(cap)) {
return; return;
} }
n->bar.pmrmsc = (n->bar.pmrmsc & ~0xffffffff) | (data & 0xffffffff); stl_le_p(&n->bar.pmrmscl, data);
n->pmr.cmse = false; n->pmr.cmse = false;
if (NVME_PMRMSC_CMSE(n->bar.pmrmsc)) { if (NVME_PMRMSCL_CMSE(data)) {
hwaddr cba = NVME_PMRMSC_CBA(n->bar.pmrmsc) << PMRMSC_CBA_SHIFT; uint64_t pmrmscu = ldl_le_p(&n->bar.pmrmscu);
hwaddr cba = pmrmscu << 32 |
(NVME_PMRMSCL_CBA(data) << PMRMSCL_CBA_SHIFT);
if (cba + int128_get64(n->pmr.dev->mr.size) < cba) { if (cba + int128_get64(n->pmr.dev->mr.size) < cba) {
NVME_PMRSTS_SET_CBAI(n->bar.pmrsts, 1); NVME_PMRSTS_SET_CBAI(pmrsts, 1);
stl_le_p(&n->bar.pmrsts, pmrsts);
return; return;
} }
@ -5918,12 +5962,12 @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data,
} }
return; return;
case 0xe18: /* PMRMSCU */ case NVME_REG_PMRMSCU:
if (!NVME_CAP_PMRS(n->bar.cap)) { if (!NVME_CAP_PMRS(cap)) {
return; return;
} }
n->bar.pmrmsc = (n->bar.pmrmsc & 0xffffffff) | (data << 32); stl_le_p(&n->bar.pmrmscu, data);
return; return;
default: default:
NVME_GUEST_ERR(pci_nvme_ub_mmiowr_invalid, NVME_GUEST_ERR(pci_nvme_ub_mmiowr_invalid,
@ -5938,7 +5982,6 @@ static uint64_t nvme_mmio_read(void *opaque, hwaddr addr, unsigned size)
{ {
NvmeCtrl *n = (NvmeCtrl *)opaque; NvmeCtrl *n = (NvmeCtrl *)opaque;
uint8_t *ptr = (uint8_t *)&n->bar; uint8_t *ptr = (uint8_t *)&n->bar;
uint64_t val = 0;
trace_pci_nvme_mmio_read(addr, size); trace_pci_nvme_mmio_read(addr, size);
@ -5954,24 +5997,25 @@ static uint64_t nvme_mmio_read(void *opaque, hwaddr addr, unsigned size)
/* should RAZ, fall through for now */ /* should RAZ, fall through for now */
} }
if (addr < sizeof(n->bar)) { if (addr > sizeof(n->bar) - size) {
/*
* When PMRWBM bit 1 is set then read from
* from PMRSTS should ensure prior writes
* made it to persistent media
*/
if (addr == 0xe08 &&
(NVME_PMRCAP_PMRWBM(n->bar.pmrcap) & 0x02)) {
memory_region_msync(&n->pmr.dev->mr, 0, n->pmr.dev->size);
}
memcpy(&val, ptr + addr, size);
} else {
NVME_GUEST_ERR(pci_nvme_ub_mmiord_invalid_ofs, NVME_GUEST_ERR(pci_nvme_ub_mmiord_invalid_ofs,
"MMIO read beyond last register," "MMIO read beyond last register,"
" offset=0x%"PRIx64", returning 0", addr); " offset=0x%"PRIx64", returning 0", addr);
return 0;
} }
return val; /*
* When PMRWBM bit 1 is set then read from
* from PMRSTS should ensure prior writes
* made it to persistent media
*/
if (addr == NVME_REG_PMRSTS &&
(NVME_PMRCAP_PMRWBM(ldl_le_p(&n->bar.pmrcap)) & 0x02)) {
memory_region_msync(&n->pmr.dev->mr, 0, n->pmr.dev->size);
}
return ldn_le_p(ptr + addr, size);
} }
static void nvme_process_db(NvmeCtrl *n, hwaddr addr, int val) static void nvme_process_db(NvmeCtrl *n, hwaddr addr, int val)
@ -6229,6 +6273,7 @@ static void nvme_init_state(NvmeCtrl *n)
static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev) static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev)
{ {
uint64_t cmb_size = n->params.cmb_size_mb * MiB; uint64_t cmb_size = n->params.cmb_size_mb * MiB;
uint64_t cap = ldq_le_p(&n->bar.cap);
n->cmb.buf = g_malloc0(cmb_size); n->cmb.buf = g_malloc0(cmb_size);
memory_region_init_io(&n->cmb.mem, OBJECT(n), &nvme_cmb_ops, n, memory_region_init_io(&n->cmb.mem, OBJECT(n), &nvme_cmb_ops, n,
@ -6238,7 +6283,8 @@ static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev)
PCI_BASE_ADDRESS_MEM_TYPE_64 | PCI_BASE_ADDRESS_MEM_TYPE_64 |
PCI_BASE_ADDRESS_MEM_PREFETCH, &n->cmb.mem); PCI_BASE_ADDRESS_MEM_PREFETCH, &n->cmb.mem);
NVME_CAP_SET_CMBS(n->bar.cap, 1); NVME_CAP_SET_CMBS(cap, 1);
stq_le_p(&n->bar.cap, cap);
if (n->params.legacy_cmb) { if (n->params.legacy_cmb) {
nvme_cmb_enable_regs(n); nvme_cmb_enable_regs(n);
@ -6248,14 +6294,17 @@ static void nvme_init_cmb(NvmeCtrl *n, PCIDevice *pci_dev)
static void nvme_init_pmr(NvmeCtrl *n, PCIDevice *pci_dev) static void nvme_init_pmr(NvmeCtrl *n, PCIDevice *pci_dev)
{ {
NVME_PMRCAP_SET_RDS(n->bar.pmrcap, 1); uint32_t pmrcap = ldl_le_p(&n->bar.pmrcap);
NVME_PMRCAP_SET_WDS(n->bar.pmrcap, 1);
NVME_PMRCAP_SET_BIR(n->bar.pmrcap, NVME_PMR_BIR);
/* Turn on bit 1 support */
NVME_PMRCAP_SET_PMRWBM(n->bar.pmrcap, 0x02);
NVME_PMRCAP_SET_CMSS(n->bar.pmrcap, 1);
pci_register_bar(pci_dev, NVME_PMRCAP_BIR(n->bar.pmrcap), NVME_PMRCAP_SET_RDS(pmrcap, 1);
NVME_PMRCAP_SET_WDS(pmrcap, 1);
NVME_PMRCAP_SET_BIR(pmrcap, NVME_PMR_BIR);
/* Turn on bit 1 support */
NVME_PMRCAP_SET_PMRWBM(pmrcap, 0x02);
NVME_PMRCAP_SET_CMSS(pmrcap, 1);
stl_le_p(&n->bar.pmrcap, pmrcap);
pci_register_bar(pci_dev, NVME_PMR_BIR,
PCI_BASE_ADDRESS_SPACE_MEMORY | PCI_BASE_ADDRESS_SPACE_MEMORY |
PCI_BASE_ADDRESS_MEM_TYPE_64 | PCI_BASE_ADDRESS_MEM_TYPE_64 |
PCI_BASE_ADDRESS_MEM_PREFETCH, &n->pmr.dev->mr); PCI_BASE_ADDRESS_MEM_PREFETCH, &n->pmr.dev->mr);
@ -6345,6 +6394,7 @@ static void nvme_init_ctrl(NvmeCtrl *n, PCIDevice *pci_dev)
{ {
NvmeIdCtrl *id = &n->id_ctrl; NvmeIdCtrl *id = &n->id_ctrl;
uint8_t *pci_conf = pci_dev->config; uint8_t *pci_conf = pci_dev->config;
uint64_t cap = ldq_le_p(&n->bar.cap);
id->vid = cpu_to_le16(pci_get_word(pci_conf + PCI_VENDOR_ID)); id->vid = cpu_to_le16(pci_get_word(pci_conf + PCI_VENDOR_ID));
id->ssvid = cpu_to_le16(pci_get_word(pci_conf + PCI_SUBSYSTEM_VENDOR_ID)); id->ssvid = cpu_to_le16(pci_get_word(pci_conf + PCI_SUBSYSTEM_VENDOR_ID));
@ -6423,17 +6473,18 @@ static void nvme_init_ctrl(NvmeCtrl *n, PCIDevice *pci_dev)
id->cmic |= NVME_CMIC_MULTI_CTRL; id->cmic |= NVME_CMIC_MULTI_CTRL;
} }
NVME_CAP_SET_MQES(n->bar.cap, 0x7ff); NVME_CAP_SET_MQES(cap, 0x7ff);
NVME_CAP_SET_CQR(n->bar.cap, 1); NVME_CAP_SET_CQR(cap, 1);
NVME_CAP_SET_TO(n->bar.cap, 0xf); NVME_CAP_SET_TO(cap, 0xf);
NVME_CAP_SET_CSS(n->bar.cap, NVME_CAP_CSS_NVM); NVME_CAP_SET_CSS(cap, NVME_CAP_CSS_NVM);
NVME_CAP_SET_CSS(n->bar.cap, NVME_CAP_CSS_CSI_SUPP); NVME_CAP_SET_CSS(cap, NVME_CAP_CSS_CSI_SUPP);
NVME_CAP_SET_CSS(n->bar.cap, NVME_CAP_CSS_ADMIN_ONLY); NVME_CAP_SET_CSS(cap, NVME_CAP_CSS_ADMIN_ONLY);
NVME_CAP_SET_MPSMAX(n->bar.cap, 4); NVME_CAP_SET_MPSMAX(cap, 4);
NVME_CAP_SET_CMBS(n->bar.cap, n->params.cmb_size_mb ? 1 : 0); NVME_CAP_SET_CMBS(cap, n->params.cmb_size_mb ? 1 : 0);
NVME_CAP_SET_PMRS(n->bar.cap, n->pmr.dev ? 1 : 0); NVME_CAP_SET_PMRS(cap, n->pmr.dev ? 1 : 0);
stq_le_p(&n->bar.cap, cap);
n->bar.vs = NVME_SPEC_VER; stl_le_p(&n->bar.vs, NVME_SPEC_VER);
n->bar.intmc = n->bar.intms = 0; n->bar.intmc = n->bar.intms = 0;
} }
@ -6498,7 +6549,7 @@ static void nvme_realize(PCIDevice *pci_dev, Error **errp)
ns = &n->namespace; ns = &n->namespace;
ns->params.nsid = 1; ns->params.nsid = 1;
if (nvme_ns_setup(n, ns, errp)) { if (nvme_ns_setup(ns, errp)) {
return; return;
} }
@ -6514,13 +6565,15 @@ static void nvme_exit(PCIDevice *pci_dev)
nvme_ctrl_reset(n); nvme_ctrl_reset(n);
for (i = 1; i <= NVME_MAX_NAMESPACES; i++) { if (n->subsys) {
ns = nvme_ns(n, i); for (i = 1; i <= NVME_MAX_NAMESPACES; i++) {
if (!ns) { ns = nvme_ns(n, i);
continue; if (ns) {
ns->attached--;
}
} }
nvme_ns_cleanup(ns); nvme_subsys_unregister_ctrl(n->subsys, n);
} }
g_free(n->cq); g_free(n->cq);
@ -6582,7 +6635,7 @@ static void nvme_set_smart_warning(Object *obj, Visitor *v, const char *name,
cap = NVME_SMART_SPARE | NVME_SMART_TEMPERATURE | NVME_SMART_RELIABILITY cap = NVME_SMART_SPARE | NVME_SMART_TEMPERATURE | NVME_SMART_RELIABILITY
| NVME_SMART_MEDIA_READ_ONLY | NVME_SMART_FAILED_VOLATILE_MEDIA; | NVME_SMART_MEDIA_READ_ONLY | NVME_SMART_FAILED_VOLATILE_MEDIA;
if (NVME_CAP_PMRS(n->bar.cap)) { if (NVME_CAP_PMRS(ldq_le_p(&n->bar.cap))) {
cap |= NVME_SMART_PMR_UNRELIABLE; cap |= NVME_SMART_PMR_UNRELIABLE;
} }

55
hw/nvme/ns.c
View File

@ -346,8 +346,7 @@ static void nvme_zoned_ns_shutdown(NvmeNamespace *ns)
assert(ns->nr_open_zones == 0); assert(ns->nr_open_zones == 0);
} }
static int nvme_ns_check_constraints(NvmeCtrl *n, NvmeNamespace *ns, static int nvme_ns_check_constraints(NvmeNamespace *ns, Error **errp)
Error **errp)
{ {
if (!ns->blkconf.blk) { if (!ns->blkconf.blk) {
error_setg(errp, "block backend not configured"); error_setg(errp, "block backend not configured");
@ -366,20 +365,6 @@ static int nvme_ns_check_constraints(NvmeCtrl *n, NvmeNamespace *ns,
return -1; return -1;
} }
if (!n->subsys) {
if (ns->params.detached) {
error_setg(errp, "detached requires that the nvme device is "
"linked to an nvme-subsys device");
return -1;
}
if (ns->params.shared) {
error_setg(errp, "shared requires that the nvme device is "
"linked to an nvme-subsys device");
return -1;
}
}
if (ns->params.zoned) { if (ns->params.zoned) {
if (ns->params.max_active_zones) { if (ns->params.max_active_zones) {
if (ns->params.max_open_zones > ns->params.max_active_zones) { if (ns->params.max_open_zones > ns->params.max_active_zones) {
@ -411,9 +396,9 @@ static int nvme_ns_check_constraints(NvmeCtrl *n, NvmeNamespace *ns,
return 0; return 0;
} }
int nvme_ns_setup(NvmeCtrl *n, NvmeNamespace *ns, Error **errp) int nvme_ns_setup(NvmeNamespace *ns, Error **errp)
{ {
if (nvme_ns_check_constraints(n, ns, errp)) { if (nvme_ns_check_constraints(ns, errp)) {
return -1; return -1;
} }
@ -456,6 +441,15 @@ void nvme_ns_cleanup(NvmeNamespace *ns)
} }
} }
static void nvme_ns_unrealize(DeviceState *dev)
{
NvmeNamespace *ns = NVME_NS(dev);
nvme_ns_drain(ns);
nvme_ns_shutdown(ns);
nvme_ns_cleanup(ns);
}
static void nvme_ns_realize(DeviceState *dev, Error **errp) static void nvme_ns_realize(DeviceState *dev, Error **errp)
{ {
NvmeNamespace *ns = NVME_NS(dev); NvmeNamespace *ns = NVME_NS(dev);
@ -465,7 +459,29 @@ static void nvme_ns_realize(DeviceState *dev, Error **errp)
uint32_t nsid = ns->params.nsid; uint32_t nsid = ns->params.nsid;
int i; int i;
if (nvme_ns_setup(n, ns, errp)) { if (!n->subsys) {
if (ns->params.detached) {
error_setg(errp, "detached requires that the nvme device is "
"linked to an nvme-subsys device");
return;
}
if (ns->params.shared) {
error_setg(errp, "shared requires that the nvme device is "
"linked to an nvme-subsys device");
return;
}
} else {
/*
* If this namespace belongs to a subsystem (through a link on the
* controller device), reparent the device.
*/
if (!qdev_set_parent_bus(dev, &subsys->bus.parent_bus, errp)) {
return;
}
}
if (nvme_ns_setup(ns, errp)) {
return; return;
} }
@ -553,6 +569,7 @@ static void nvme_ns_class_init(ObjectClass *oc, void *data)
dc->bus_type = TYPE_NVME_BUS; dc->bus_type = TYPE_NVME_BUS;
dc->realize = nvme_ns_realize; dc->realize = nvme_ns_realize;
dc->unrealize = nvme_ns_unrealize;
device_class_set_props(dc, nvme_ns_props); device_class_set_props(dc, nvme_ns_props);
dc->desc = "Virtual NVMe namespace"; dc->desc = "Virtual NVMe namespace";
} }

18
hw/nvme/nvme.h
View File

@ -33,12 +33,20 @@ QEMU_BUILD_BUG_ON(NVME_MAX_NAMESPACES > NVME_NSID_BROADCAST - 1);
typedef struct NvmeCtrl NvmeCtrl; typedef struct NvmeCtrl NvmeCtrl;
typedef struct NvmeNamespace NvmeNamespace; typedef struct NvmeNamespace NvmeNamespace;
#define TYPE_NVME_BUS "nvme-bus"
OBJECT_DECLARE_SIMPLE_TYPE(NvmeBus, NVME_BUS)
typedef struct NvmeBus {
BusState parent_bus;
} NvmeBus;
#define TYPE_NVME_SUBSYS "nvme-subsys" #define TYPE_NVME_SUBSYS "nvme-subsys"
#define NVME_SUBSYS(obj) \ #define NVME_SUBSYS(obj) \
OBJECT_CHECK(NvmeSubsystem, (obj), TYPE_NVME_SUBSYS) OBJECT_CHECK(NvmeSubsystem, (obj), TYPE_NVME_SUBSYS)
typedef struct NvmeSubsystem { typedef struct NvmeSubsystem {
DeviceState parent_obj; DeviceState parent_obj;
NvmeBus bus;
uint8_t subnqn[256]; uint8_t subnqn[256];
NvmeCtrl *ctrls[NVME_MAX_CONTROLLERS]; NvmeCtrl *ctrls[NVME_MAX_CONTROLLERS];
@ -50,6 +58,7 @@ typedef struct NvmeSubsystem {
} NvmeSubsystem; } NvmeSubsystem;
int nvme_subsys_register_ctrl(NvmeCtrl *n, Error **errp); int nvme_subsys_register_ctrl(NvmeCtrl *n, Error **errp);
void nvme_subsys_unregister_ctrl(NvmeSubsystem *subsys, NvmeCtrl *n);
static inline NvmeCtrl *nvme_subsys_ctrl(NvmeSubsystem *subsys, static inline NvmeCtrl *nvme_subsys_ctrl(NvmeSubsystem *subsys,
uint32_t cntlid) uint32_t cntlid)
@ -246,7 +255,7 @@ static inline void nvme_aor_dec_active(NvmeNamespace *ns)
} }
void nvme_ns_init_format(NvmeNamespace *ns); void nvme_ns_init_format(NvmeNamespace *ns);
int nvme_ns_setup(NvmeCtrl *n, NvmeNamespace *ns, Error **errp); int nvme_ns_setup(NvmeNamespace *ns, Error **errp);
void nvme_ns_drain(NvmeNamespace *ns); void nvme_ns_drain(NvmeNamespace *ns);
void nvme_ns_shutdown(NvmeNamespace *ns); void nvme_ns_shutdown(NvmeNamespace *ns);
void nvme_ns_cleanup(NvmeNamespace *ns); void nvme_ns_cleanup(NvmeNamespace *ns);
@ -364,13 +373,6 @@ typedef struct NvmeCQueue {
QTAILQ_HEAD(, NvmeRequest) req_list; QTAILQ_HEAD(, NvmeRequest) req_list;
} NvmeCQueue; } NvmeCQueue;
#define TYPE_NVME_BUS "nvme-bus"
#define NVME_BUS(obj) OBJECT_CHECK(NvmeBus, (obj), TYPE_NVME_BUS)
typedef struct NvmeBus {
BusState parent_bus;
} NvmeBus;
#define TYPE_NVME "nvme" #define TYPE_NVME "nvme"
#define NVME(obj) \ #define NVME(obj) \
OBJECT_CHECK(NvmeCtrl, (obj), TYPE_NVME) OBJECT_CHECK(NvmeCtrl, (obj), TYPE_NVME)

9
hw/nvme/subsys.c
View File

@ -32,6 +32,11 @@ int nvme_subsys_register_ctrl(NvmeCtrl *n, Error **errp)
return cntlid; return cntlid;
} }
void nvme_subsys_unregister_ctrl(NvmeSubsystem *subsys, NvmeCtrl *n)
{
subsys->ctrls[n->cntlid] = NULL;
}
static void nvme_subsys_setup(NvmeSubsystem *subsys) static void nvme_subsys_setup(NvmeSubsystem *subsys)
{ {
const char *nqn = subsys->params.nqn ? const char *nqn = subsys->params.nqn ?
@ -45,6 +50,9 @@ static void nvme_subsys_realize(DeviceState *dev, Error **errp)
{ {
NvmeSubsystem *subsys = NVME_SUBSYS(dev); NvmeSubsystem *subsys = NVME_SUBSYS(dev);
qbus_create_inplace(&subsys->bus, sizeof(NvmeBus), TYPE_NVME_BUS, dev,
dev->id);
nvme_subsys_setup(subsys); nvme_subsys_setup(subsys);
} }
@ -61,6 +69,7 @@ static void nvme_subsys_class_init(ObjectClass *oc, void *data)
dc->realize = nvme_subsys_realize; dc->realize = nvme_subsys_realize;
dc->desc = "Virtual NVMe subsystem"; dc->desc = "Virtual NVMe subsystem";
dc->hotpluggable = false;
device_class_set_props(dc, nvme_subsystem_props); device_class_set_props(dc, nvme_subsystem_props);
} }

1
hw/nvme/trace-events
View File

@ -199,3 +199,4 @@ pci_nvme_ub_db_wr_invalid_cqhead(uint32_t qid, uint16_t new_head) "completion qu
pci_nvme_ub_db_wr_invalid_sq(uint32_t qid) "submission queue doorbell write for nonexistent queue, sqid=%"PRIu32", ignoring" pci_nvme_ub_db_wr_invalid_sq(uint32_t qid) "submission queue doorbell write for nonexistent queue, sqid=%"PRIu32", ignoring"
pci_nvme_ub_db_wr_invalid_sqtail(uint32_t qid, uint16_t new_tail) "submission queue doorbell write value beyond queue size, sqid=%"PRIu32", new_head=%"PRIu16", ignoring" pci_nvme_ub_db_wr_invalid_sqtail(uint32_t qid, uint16_t new_tail) "submission queue doorbell write value beyond queue size, sqid=%"PRIu32", new_head=%"PRIu16", ignoring"
pci_nvme_ub_unknown_css_value(void) "unknown value in cc.css field" pci_nvme_ub_unknown_css_value(void) "unknown value in cc.css field"
pci_nvme_ub_too_many_mappings(void) "too many prp/sgl mappings"

60
include/block/nvme.h
View File

@ -9,7 +9,7 @@ typedef struct QEMU_PACKED NvmeBar {
uint32_t cc; uint32_t cc;
uint8_t rsvd24[4]; uint8_t rsvd24[4];
uint32_t csts; uint32_t csts;
uint32_t nssrc; uint32_t nssr;
uint32_t aqa; uint32_t aqa;
uint64_t asq; uint64_t asq;
uint64_t acq; uint64_t acq;
@ -26,10 +26,38 @@ typedef struct QEMU_PACKED NvmeBar {
uint32_t pmrsts; uint32_t pmrsts;
uint32_t pmrebs; uint32_t pmrebs;
uint32_t pmrswtp; uint32_t pmrswtp;
uint64_t pmrmsc; uint32_t pmrmscl;
uint32_t pmrmscu;
uint8_t css[484]; uint8_t css[484];
} NvmeBar; } NvmeBar;
enum NvmeBarRegs {
NVME_REG_CAP = offsetof(NvmeBar, cap),
NVME_REG_VS = offsetof(NvmeBar, vs),
NVME_REG_INTMS = offsetof(NvmeBar, intms),
NVME_REG_INTMC = offsetof(NvmeBar, intmc),
NVME_REG_CC = offsetof(NvmeBar, cc),
NVME_REG_CSTS = offsetof(NvmeBar, csts),
NVME_REG_NSSR = offsetof(NvmeBar, nssr),
NVME_REG_AQA = offsetof(NvmeBar, aqa),
NVME_REG_ASQ = offsetof(NvmeBar, asq),
NVME_REG_ACQ = offsetof(NvmeBar, acq),
NVME_REG_CMBLOC = offsetof(NvmeBar, cmbloc),
NVME_REG_CMBSZ = offsetof(NvmeBar, cmbsz),
NVME_REG_BPINFO = offsetof(NvmeBar, bpinfo),
NVME_REG_BPRSEL = offsetof(NvmeBar, bprsel),
NVME_REG_BPMBL = offsetof(NvmeBar, bpmbl),
NVME_REG_CMBMSC = offsetof(NvmeBar, cmbmsc),
NVME_REG_CMBSTS = offsetof(NvmeBar, cmbsts),
NVME_REG_PMRCAP = offsetof(NvmeBar, pmrcap),
NVME_REG_PMRCTL = offsetof(NvmeBar, pmrctl),
NVME_REG_PMRSTS = offsetof(NvmeBar, pmrsts),
NVME_REG_PMREBS = offsetof(NvmeBar, pmrebs),
NVME_REG_PMRSWTP = offsetof(NvmeBar, pmrswtp),
NVME_REG_PMRMSCL = offsetof(NvmeBar, pmrmscl),
NVME_REG_PMRMSCU = offsetof(NvmeBar, pmrmscu),
};
enum NvmeCapShift { enum NvmeCapShift {
CAP_MQES_SHIFT = 0, CAP_MQES_SHIFT = 0,
CAP_CQR_SHIFT = 16, CAP_CQR_SHIFT = 16,
@ -475,25 +503,25 @@ enum NvmePmrswtpMask {
#define NVME_PMRSWTP_SET_PMRSWTV(pmrswtp, val) \ #define NVME_PMRSWTP_SET_PMRSWTV(pmrswtp, val) \
(pmrswtp |= (uint64_t)(val & PMRSWTP_PMRSWTV_MASK) << PMRSWTP_PMRSWTV_SHIFT) (pmrswtp |= (uint64_t)(val & PMRSWTP_PMRSWTV_MASK) << PMRSWTP_PMRSWTV_SHIFT)
enum NvmePmrmscShift { enum NvmePmrmsclShift {
PMRMSC_CMSE_SHIFT = 1, PMRMSCL_CMSE_SHIFT = 1,
PMRMSC_CBA_SHIFT = 12, PMRMSCL_CBA_SHIFT = 12,
}; };
enum NvmePmrmscMask { enum NvmePmrmsclMask {
PMRMSC_CMSE_MASK = 0x1, PMRMSCL_CMSE_MASK = 0x1,
PMRMSC_CBA_MASK = 0xfffffffffffff, PMRMSCL_CBA_MASK = 0xfffff,
}; };
#define NVME_PMRMSC_CMSE(pmrmsc) \ #define NVME_PMRMSCL_CMSE(pmrmscl) \
((pmrmsc >> PMRMSC_CMSE_SHIFT) & PMRMSC_CMSE_MASK) ((pmrmscl >> PMRMSCL_CMSE_SHIFT) & PMRMSCL_CMSE_MASK)
#define NVME_PMRMSC_CBA(pmrmsc) \ #define NVME_PMRMSCL_CBA(pmrmscl) \
((pmrmsc >> PMRMSC_CBA_SHIFT) & PMRMSC_CBA_MASK) ((pmrmscl >> PMRMSCL_CBA_SHIFT) & PMRMSCL_CBA_MASK)
#define NVME_PMRMSC_SET_CMSE(pmrmsc, val) \ #define NVME_PMRMSCL_SET_CMSE(pmrmscl, val) \
(pmrmsc |= (uint64_t)(val & PMRMSC_CMSE_MASK) << PMRMSC_CMSE_SHIFT) (pmrmscl |= (uint32_t)(val & PMRMSCL_CMSE_MASK) << PMRMSCL_CMSE_SHIFT)
#define NVME_PMRMSC_SET_CBA(pmrmsc, val) \ #define NVME_PMRMSCL_SET_CBA(pmrmscl, val) \
(pmrmsc |= (uint64_t)(val & PMRMSC_CBA_MASK) << PMRMSC_CBA_SHIFT) (pmrmscl |= (uint32_t)(val & PMRMSCL_CBA_MASK) << PMRMSCL_CBA_SHIFT)
enum NvmeSglDescriptorType { enum NvmeSglDescriptorType {
NVME_SGL_DESCR_TYPE_DATA_BLOCK = 0x0, NVME_SGL_DESCR_TYPE_DATA_BLOCK = 0x0,

87
tests/qtest/nvme-test.c
View File

@ -13,6 +13,7 @@
#include "libqos/libqtest.h" #include "libqos/libqtest.h"
#include "libqos/qgraph.h" #include "libqos/qgraph.h"
#include "libqos/pci.h" #include "libqos/pci.h"
#include "include/block/nvme.h"
typedef struct QNvme QNvme; typedef struct QNvme QNvme;
@ -66,12 +67,89 @@ static void nvmetest_oob_cmb_test(void *obj, void *data, QGuestAllocator *alloc)
g_assert_cmpint(qpci_io_readl(pdev, bar, cmb_bar_size - 1), !=, 0x44332211); g_assert_cmpint(qpci_io_readl(pdev, bar, cmb_bar_size - 1), !=, 0x44332211);
} }
static void nvmetest_reg_read_test(void *obj, void *data, QGuestAllocator *alloc)
{
QNvme *nvme = obj;
QPCIDevice *pdev = &nvme->dev;
QPCIBar bar;
uint32_t cap_lo, cap_hi;
uint64_t cap;
qpci_device_enable(pdev);
bar = qpci_iomap(pdev, 0, NULL);
cap_lo = qpci_io_readl(pdev, bar, 0x0);
g_assert_cmpint(NVME_CAP_MQES(cap_lo), ==, 0x7ff);
cap_hi = qpci_io_readl(pdev, bar, 0x4);
g_assert_cmpint(NVME_CAP_MPSMAX((uint64_t)cap_hi << 32), ==, 0x4);
cap = qpci_io_readq(pdev, bar, 0x0);
g_assert_cmpint(NVME_CAP_MQES(cap), ==, 0x7ff);
g_assert_cmpint(NVME_CAP_MPSMAX(cap), ==, 0x4);
qpci_iounmap(pdev, bar);
}
static void nvmetest_pmr_reg_test(void *obj, void *data, QGuestAllocator *alloc)
{
QNvme *nvme = obj;
QPCIDevice *pdev = &nvme->dev;
QPCIBar pmr_bar, nvme_bar;
uint32_t pmrcap, pmrsts;
qpci_device_enable(pdev);
pmr_bar = qpci_iomap(pdev, 4, NULL);
/* Without Enabling PMRCTL check bar enablemet */
qpci_io_writel(pdev, pmr_bar, 0, 0xccbbaa99);
g_assert_cmpint(qpci_io_readb(pdev, pmr_bar, 0), !=, 0x99);
g_assert_cmpint(qpci_io_readw(pdev, pmr_bar, 0), !=, 0xaa99);
/* Map NVMe Bar Register to Enable the Mem Region */
nvme_bar = qpci_iomap(pdev, 0, NULL);
pmrcap = qpci_io_readl(pdev, nvme_bar, 0xe00);
g_assert_cmpint(NVME_PMRCAP_RDS(pmrcap), ==, 0x1);
g_assert_cmpint(NVME_PMRCAP_WDS(pmrcap), ==, 0x1);
g_assert_cmpint(NVME_PMRCAP_BIR(pmrcap), ==, 0x4);
g_assert_cmpint(NVME_PMRCAP_PMRWBM(pmrcap), ==, 0x2);
g_assert_cmpint(NVME_PMRCAP_CMSS(pmrcap), ==, 0x1);
/* Enable PMRCTRL */
qpci_io_writel(pdev, nvme_bar, 0xe04, 0x1);
qpci_io_writel(pdev, pmr_bar, 0, 0x44332211);
g_assert_cmpint(qpci_io_readb(pdev, pmr_bar, 0), ==, 0x11);
g_assert_cmpint(qpci_io_readw(pdev, pmr_bar, 0), ==, 0x2211);
g_assert_cmpint(qpci_io_readl(pdev, pmr_bar, 0), ==, 0x44332211);
pmrsts = qpci_io_readl(pdev, nvme_bar, 0xe08);
g_assert_cmpint(NVME_PMRSTS_NRDY(pmrsts), ==, 0x0);
/* Disable PMRCTRL */
qpci_io_writel(pdev, nvme_bar, 0xe04, 0x0);
qpci_io_writel(pdev, pmr_bar, 0, 0x88776655);
g_assert_cmpint(qpci_io_readb(pdev, pmr_bar, 0), !=, 0x55);
g_assert_cmpint(qpci_io_readw(pdev, pmr_bar, 0), !=, 0x6655);
g_assert_cmpint(qpci_io_readl(pdev, pmr_bar, 0), !=, 0x88776655);
pmrsts = qpci_io_readl(pdev, nvme_bar, 0xe08);
g_assert_cmpint(NVME_PMRSTS_NRDY(pmrsts), ==, 0x1);
qpci_iounmap(pdev, nvme_bar);
qpci_iounmap(pdev, pmr_bar);
}
static void nvme_register_nodes(void) static void nvme_register_nodes(void)
{ {
QOSGraphEdgeOptions opts = { QOSGraphEdgeOptions opts = {
.extra_device_opts = "addr=04.0,drive=drv0,serial=foo", .extra_device_opts = "addr=04.0,drive=drv0,serial=foo",
.before_cmd_line = "-drive id=drv0,if=none,file=null-co://," .before_cmd_line = "-drive id=drv0,if=none,file=null-co://,"
"file.read-zeroes=on,format=raw", "file.read-zeroes=on,format=raw "
"-object memory-backend-ram,id=pmr0,"
"share=on,size=8",
}; };
add_qpci_address(&opts, &(QPCIAddress) { .devfn = QPCI_DEVFN(4, 0) }); add_qpci_address(&opts, &(QPCIAddress) { .devfn = QPCI_DEVFN(4, 0) });
@ -83,6 +161,13 @@ static void nvme_register_nodes(void)
qos_add_test("oob-cmb-access", "nvme", nvmetest_oob_cmb_test, &(QOSGraphTestOptions) { qos_add_test("oob-cmb-access", "nvme", nvmetest_oob_cmb_test, &(QOSGraphTestOptions) {
.edge.extra_device_opts = "cmb_size_mb=2" .edge.extra_device_opts = "cmb_size_mb=2"
}); });
qos_add_test("pmr-test-access", "nvme", nvmetest_pmr_reg_test,
&(QOSGraphTestOptions) {
.edge.extra_device_opts = "pmrdev=pmr0"
});
qos_add_test("reg-read", "nvme", nvmetest_reg_read_test, NULL);
} }
libqos_init(nvme_register_nodes); libqos_init(nvme_register_nodes);