aboutsummaryrefslogtreecommitdiff
path: root/drivers/nvme/host/pci.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/nvme/host/pci.c')
-rw-r--r--drivers/nvme/host/pci.c189
1 files changed, 77 insertions, 112 deletions
diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
index f0f8027644bb..5b95c94ee40f 100644
--- a/drivers/nvme/host/pci.c
+++ b/drivers/nvme/host/pci.c
@@ -36,14 +36,15 @@
#define SQ_SIZE(q) ((q)->q_depth << (q)->sqes)
#define CQ_SIZE(q) ((q)->q_depth * sizeof(struct nvme_completion))
-#define SGES_PER_PAGE (PAGE_SIZE / sizeof(struct nvme_sgl_desc))
+#define SGES_PER_PAGE (NVME_CTRL_PAGE_SIZE / sizeof(struct nvme_sgl_desc))
/*
* These can be higher, but we need to ensure that any command doesn't
* require an sg allocation that needs more than a page of data.
*/
-#define NVME_MAX_KB_SZ 4096
-#define NVME_MAX_SEGS 127
+#define NVME_MAX_KB_SZ 8192
+#define NVME_MAX_SEGS 128
+#define NVME_MAX_NR_ALLOCATIONS 5
static int use_threaded_interrupts;
module_param(use_threaded_interrupts, int, 0444);
@@ -110,6 +111,7 @@ struct nvme_queue;
static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown);
static void nvme_delete_io_queues(struct nvme_dev *dev);
+static void nvme_update_attrs(struct nvme_dev *dev);
/*
* Represents an NVM Express device. Each nvme_dev is a PCI function.
@@ -144,9 +146,9 @@ struct nvme_dev {
mempool_t *iod_mempool;
/* shadow doorbell buffer support: */
- u32 *dbbuf_dbs;
+ __le32 *dbbuf_dbs;
dma_addr_t dbbuf_dbs_dma_addr;
- u32 *dbbuf_eis;
+ __le32 *dbbuf_eis;
dma_addr_t dbbuf_eis_dma_addr;
/* host memory buffer support: */
@@ -208,13 +210,18 @@ struct nvme_queue {
#define NVMEQ_SQ_CMB 1
#define NVMEQ_DELETE_ERROR 2
#define NVMEQ_POLLED 3
- u32 *dbbuf_sq_db;
- u32 *dbbuf_cq_db;
- u32 *dbbuf_sq_ei;
- u32 *dbbuf_cq_ei;
+ __le32 *dbbuf_sq_db;
+ __le32 *dbbuf_cq_db;
+ __le32 *dbbuf_sq_ei;
+ __le32 *dbbuf_cq_ei;
struct completion delete_done;
};
+union nvme_descriptor {
+ struct nvme_sgl_desc *sg_list;
+ __le64 *prp_list;
+};
+
/*
* The nvme_iod describes the data in an I/O.
*
@@ -224,7 +231,6 @@ struct nvme_queue {
struct nvme_iod {
struct nvme_request req;
struct nvme_command cmd;
- bool use_sgl;
bool aborted;
s8 nr_allocations; /* PRP list pool allocations. 0 means small
pool in use */
@@ -232,6 +238,7 @@ struct nvme_iod {
dma_addr_t first_dma;
dma_addr_t meta_dma;
struct sg_table sgt;
+ union nvme_descriptor list[NVME_MAX_NR_ALLOCATIONS];
};
static inline unsigned int nvme_dbbuf_size(struct nvme_dev *dev)
@@ -343,11 +350,11 @@ static inline int nvme_dbbuf_need_event(u16 event_idx, u16 new_idx, u16 old)
}
/* Update dbbuf and return true if an MMIO is required */
-static bool nvme_dbbuf_update_and_check_event(u16 value, u32 *dbbuf_db,
- volatile u32 *dbbuf_ei)
+static bool nvme_dbbuf_update_and_check_event(u16 value, __le32 *dbbuf_db,
+ volatile __le32 *dbbuf_ei)
{
if (dbbuf_db) {
- u16 old_value;
+ u16 old_value, event_idx;
/*
* Ensure that the queue is written before updating
@@ -355,8 +362,8 @@ static bool nvme_dbbuf_update_and_check_event(u16 value, u32 *dbbuf_db,
*/
wmb();
- old_value = *dbbuf_db;
- *dbbuf_db = value;
+ old_value = le32_to_cpu(*dbbuf_db);
+ *dbbuf_db = cpu_to_le32(value);
/*
* Ensure that the doorbell is updated before reading the event
@@ -366,7 +373,8 @@ static bool nvme_dbbuf_update_and_check_event(u16 value, u32 *dbbuf_db,
*/
mb();
- if (!nvme_dbbuf_need_event(*dbbuf_ei, value, old_value))
+ event_idx = le32_to_cpu(*dbbuf_ei);
+ if (!nvme_dbbuf_need_event(event_idx, value, old_value))
return false;
}
@@ -380,19 +388,9 @@ static bool nvme_dbbuf_update_and_check_event(u16 value, u32 *dbbuf_db,
*/
static int nvme_pci_npages_prp(void)
{
- unsigned nprps = DIV_ROUND_UP(NVME_MAX_KB_SZ + NVME_CTRL_PAGE_SIZE,
- NVME_CTRL_PAGE_SIZE);
- return DIV_ROUND_UP(8 * nprps, PAGE_SIZE - 8);
-}
-
-/*
- * Calculates the number of pages needed for the SGL segments. For example a 4k
- * page can accommodate 256 SGL descriptors.
- */
-static int nvme_pci_npages_sgl(void)
-{
- return DIV_ROUND_UP(NVME_MAX_SEGS * sizeof(struct nvme_sgl_desc),
- PAGE_SIZE);
+ unsigned max_bytes = (NVME_MAX_KB_SZ * 1024) + NVME_CTRL_PAGE_SIZE;
+ unsigned nprps = DIV_ROUND_UP(max_bytes, NVME_CTRL_PAGE_SIZE);
+ return DIV_ROUND_UP(8 * nprps, NVME_CTRL_PAGE_SIZE - 8);
}
static int nvme_admin_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
@@ -508,16 +506,10 @@ static void nvme_commit_rqs(struct blk_mq_hw_ctx *hctx)
spin_unlock(&nvmeq->sq_lock);
}
-static void **nvme_pci_iod_list(struct request *req)
-{
- struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
- return (void **)(iod->sgt.sgl + blk_rq_nr_phys_segments(req));
-}
-
-static inline bool nvme_pci_use_sgls(struct nvme_dev *dev, struct request *req)
+static inline bool nvme_pci_use_sgls(struct nvme_dev *dev, struct request *req,
+ int nseg)
{
struct nvme_queue *nvmeq = req->mq_hctx->driver_data;
- int nseg = blk_rq_nr_phys_segments(req);
unsigned int avg_seg_size;
avg_seg_size = DIV_ROUND_UP(blk_rq_payload_bytes(req), nseg);
@@ -539,7 +531,7 @@ static void nvme_free_prps(struct nvme_dev *dev, struct request *req)
int i;
for (i = 0; i < iod->nr_allocations; i++) {
- __le64 *prp_list = nvme_pci_iod_list(req)[i];
+ __le64 *prp_list = iod->list[i].prp_list;
dma_addr_t next_dma_addr = le64_to_cpu(prp_list[last_prp]);
dma_pool_free(dev->prp_page_pool, prp_list, dma_addr);
@@ -547,22 +539,6 @@ static void nvme_free_prps(struct nvme_dev *dev, struct request *req)
}
}
-static void nvme_free_sgls(struct nvme_dev *dev, struct request *req)
-{
- const int last_sg = SGES_PER_PAGE - 1;
- struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
- dma_addr_t dma_addr = iod->first_dma;
- int i;
-
- for (i = 0; i < iod->nr_allocations; i++) {
- struct nvme_sgl_desc *sg_list = nvme_pci_iod_list(req)[i];
- dma_addr_t next_dma_addr = le64_to_cpu((sg_list[last_sg]).addr);
-
- dma_pool_free(dev->prp_page_pool, sg_list, dma_addr);
- dma_addr = next_dma_addr;
- }
-}
-
static void nvme_unmap_data(struct nvme_dev *dev, struct request *req)
{
struct nvme_iod *iod = blk_mq_rq_to_pdu(req);
@@ -578,10 +554,11 @@ static void nvme_unmap_data(struct nvme_dev *dev, struct request *req)
dma_unmap_sgtable(dev->dev, &iod->sgt, rq_dma_dir(req), 0);
if (iod->nr_allocations == 0)
- dma_pool_free(dev->prp_small_pool, nvme_pci_iod_list(req)[0],
+ dma_pool_free(dev->prp_small_pool, iod->list[0].sg_list,
+ iod->first_dma);
+ else if (iod->nr_allocations == 1)
+ dma_pool_free(dev->prp_page_pool, iod->list[0].sg_list,
iod->first_dma);
- else if (iod->use_sgl)
- nvme_free_sgls(dev, req);
else
nvme_free_prps(dev, req);
mempool_free(iod->sgt.sgl, dev->iod_mempool);
@@ -612,7 +589,6 @@ static blk_status_t nvme_pci_setup_prps(struct nvme_dev *dev,
u64 dma_addr = sg_dma_address(sg);
int offset = dma_addr & (NVME_CTRL_PAGE_SIZE - 1);
__le64 *prp_list;
- void **list = nvme_pci_iod_list(req);
dma_addr_t prp_dma;
int nprps, i;
@@ -650,7 +626,7 @@ static blk_status_t nvme_pci_setup_prps(struct nvme_dev *dev,
iod->nr_allocations = -1;
return BLK_STS_RESOURCE;
}
- list[0] = prp_list;
+ iod->list[0].prp_list = prp_list;
iod->first_dma = prp_dma;
i = 0;
for (;;) {
@@ -659,7 +635,7 @@ static blk_status_t nvme_pci_setup_prps(struct nvme_dev *dev,
prp_list = dma_pool_alloc(pool, GFP_ATOMIC, &prp_dma);
if (!prp_list)
goto free_prps;
- list[iod->nr_allocations++] = prp_list;
+ iod->list[iod->nr_allocations++].prp_list = prp_list;
prp_list[0] = old_prp_list[i - 1];
old_prp_list[i - 1] = cpu_to_le64(prp_dma);
i = 1;
@@ -704,13 +680,8 @@ static void nvme_pci_sgl_set_seg(struct nvme_sgl_desc *sge,
dma_addr_t dma_addr, int entries)
{
sge->addr = cpu_to_le64(dma_addr);
- if (entries < SGES_PER_PAGE) {
- sge->length = cpu_to_le32(entries * sizeof(*sge));
- sge->type = NVME_SGL_FMT_LAST_SEG_DESC << 4;
- } else {
- sge->length = cpu_to_le32(PAGE_SIZE);
- sge->type = NVME_SGL_FMT_SEG_DESC << 4;
- }
+ sge->length = cpu_to_le32(entries * sizeof(*sge));
+ sge->type = NVME_SGL_FMT_LAST_SEG_DESC << 4;
}
static blk_status_t nvme_pci_setup_sgls(struct nvme_dev *dev,
@@ -746,34 +717,16 @@ static blk_status_t nvme_pci_setup_sgls(struct nvme_dev *dev,
return BLK_STS_RESOURCE;
}
- nvme_pci_iod_list(req)[0] = sg_list;
+ iod->list[0].sg_list = sg_list;
iod->first_dma = sgl_dma;
nvme_pci_sgl_set_seg(&cmd->dptr.sgl, sgl_dma, entries);
-
do {
- if (i == SGES_PER_PAGE) {
- struct nvme_sgl_desc *old_sg_desc = sg_list;
- struct nvme_sgl_desc *link = &old_sg_desc[i - 1];
-
- sg_list = dma_pool_alloc(pool, GFP_ATOMIC, &sgl_dma);
- if (!sg_list)
- goto free_sgls;
-
- i = 0;
- nvme_pci_iod_list(req)[iod->nr_allocations++] = sg_list;
- sg_list[i++] = *link;
- nvme_pci_sgl_set_seg(link, sgl_dma, entries);
- }
-
nvme_pci_sgl_set_data(&sg_list[i++], sg);
sg = sg_next(sg);
} while (--entries > 0);
return BLK_STS_OK;
-free_sgls:
- nvme_free_sgls(dev, req);
- return BLK_STS_RESOURCE;
}
static blk_status_t nvme_setup_prp_simple(struct nvme_dev *dev,
@@ -855,8 +808,7 @@ static blk_status_t nvme_map_data(struct nvme_dev *dev, struct request *req,
goto out_free_sg;
}
- iod->use_sgl = nvme_pci_use_sgls(dev, req);
- if (iod->use_sgl)
+ if (nvme_pci_use_sgls(dev, req, iod->sgt.nents))
ret = nvme_pci_setup_sgls(dev, req, &cmnd->rw);
else
ret = nvme_pci_setup_prps(dev, req, &cmnd->rw);
@@ -1361,7 +1313,7 @@ static enum blk_eh_timer_return nvme_timeout(struct request *req)
else
nvme_poll_irqdisable(nvmeq);
- if (blk_mq_request_completed(req)) {
+ if (blk_mq_rq_state(req) != MQ_RQ_IN_FLIGHT) {
dev_warn(dev->ctrl.device,
"I/O %d QID %d timeout, completion polled\n",
req->tag, nvmeq->qid);
@@ -1922,6 +1874,8 @@ static void nvme_map_cmb(struct nvme_dev *dev)
if ((dev->cmbsz & (NVME_CMBSZ_WDS | NVME_CMBSZ_RDS)) ==
(NVME_CMBSZ_WDS | NVME_CMBSZ_RDS))
pci_p2pmem_publish(pdev, true);
+
+ nvme_update_attrs(dev);
}
static int nvme_set_host_mem(struct nvme_dev *dev, u32 bits)
@@ -2208,6 +2162,11 @@ static const struct attribute_group *nvme_pci_dev_attr_groups[] = {
NULL,
};
+static void nvme_update_attrs(struct nvme_dev *dev)
+{
+ sysfs_update_group(&dev->ctrl.device->kobj, &nvme_pci_dev_attrs_group);
+}
+
/*
* nirqs is the number of interrupts available for write and read
* queues. The core already reserved an interrupt for the admin queue.
@@ -2332,10 +2291,12 @@ static int nvme_setup_io_queues(struct nvme_dev *dev)
if (dev->cmb_use_sqes) {
result = nvme_cmb_qdepth(dev, nr_io_queues,
sizeof(struct nvme_command));
- if (result > 0)
+ if (result > 0) {
dev->q_depth = result;
- else
+ dev->ctrl.sqsize = result - 1;
+ } else {
dev->cmb_use_sqes = false;
+ }
}
do {
@@ -2506,18 +2467,12 @@ static int nvme_pci_enable(struct nvme_dev *dev)
{
int result = -ENOMEM;
struct pci_dev *pdev = to_pci_dev(dev->dev);
- int dma_address_bits = 64;
if (pci_enable_device_mem(pdev))
return result;
pci_set_master(pdev);
- if (dev->ctrl.quirks & NVME_QUIRK_DMA_ADDRESS_BITS_48)
- dma_address_bits = 48;
- if (dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(dma_address_bits)))
- goto disable;
-
if (readl(dev->bar + NVME_REG_CSTS) == -1) {
result = -ENODEV;
goto disable;
@@ -2530,13 +2485,12 @@ static int nvme_pci_enable(struct nvme_dev *dev)
*/
result = pci_alloc_irq_vectors(pdev, 1, 1, PCI_IRQ_ALL_TYPES);
if (result < 0)
- return result;
+ goto disable;
dev->ctrl.cap = lo_hi_readq(dev->bar + NVME_REG_CAP);
dev->q_depth = min_t(u32, NVME_CAP_MQES(dev->ctrl.cap) + 1,
io_queue_depth);
- dev->ctrl.sqsize = dev->q_depth - 1; /* 0's based queue depth */
dev->db_stride = 1 << NVME_CAP_STRIDE(dev->ctrl.cap);
dev->dbs = dev->bar + 4096;
@@ -2577,15 +2531,20 @@ static int nvme_pci_enable(struct nvme_dev *dev)
dev_warn(dev->ctrl.device, "IO queue depth clamped to %d\n",
dev->q_depth);
}
-
+ dev->ctrl.sqsize = dev->q_depth - 1; /* 0's based queue depth */
nvme_map_cmb(dev);
pci_enable_pcie_error_reporting(pdev);
pci_save_state(pdev);
- return nvme_pci_configure_admin_queue(dev);
+ result = nvme_pci_configure_admin_queue(dev);
+ if (result)
+ goto free_irq;
+ return result;
+ free_irq:
+ pci_free_irq_vectors(pdev);
disable:
pci_disable_device(pdev);
return result;
@@ -2697,11 +2656,8 @@ static void nvme_release_prp_pools(struct nvme_dev *dev)
static int nvme_pci_alloc_iod_mempool(struct nvme_dev *dev)
{
- size_t npages = max(nvme_pci_npages_prp(), nvme_pci_npages_sgl());
- size_t alloc_size = sizeof(__le64 *) * npages +
- sizeof(struct scatterlist) * NVME_MAX_SEGS;
+ size_t alloc_size = sizeof(struct scatterlist) * NVME_MAX_SEGS;
- WARN_ON_ONCE(alloc_size > PAGE_SIZE);
dev->iod_mempool = mempool_create_node(1,
mempool_kmalloc, mempool_kfree,
(void *)alloc_size, GFP_KERNEL,
@@ -2963,7 +2919,7 @@ static struct nvme_dev *nvme_pci_alloc_dev(struct pci_dev *pdev,
dev = kzalloc_node(sizeof(*dev), GFP_KERNEL, node);
if (!dev)
- return NULL;
+ return ERR_PTR(-ENOMEM);
INIT_WORK(&dev->ctrl.reset_work, nvme_reset_work);
mutex_init(&dev->shutdown_lock);
@@ -2991,7 +2947,11 @@ static struct nvme_dev *nvme_pci_alloc_dev(struct pci_dev *pdev,
quirks);
if (ret)
goto out_put_device;
-
+
+ if (dev->ctrl.quirks & NVME_QUIRK_DMA_ADDRESS_BITS_48)
+ dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
+ else
+ dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
dma_set_min_align_mask(&pdev->dev, NVME_CTRL_PAGE_SIZE - 1);
dma_set_max_seg_size(&pdev->dev, 0xffffffff);
@@ -3024,8 +2984,8 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
int result = -ENOMEM;
dev = nvme_pci_alloc_dev(pdev, id);
- if (!dev)
- return -ENOMEM;
+ if (IS_ERR(dev))
+ return PTR_ERR(dev);
result = nvme_dev_map(dev);
if (result)
@@ -3095,6 +3055,7 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
nvme_start_ctrl(&dev->ctrl);
nvme_put_ctrl(&dev->ctrl);
+ flush_work(&dev->ctrl.scan_work);
return 0;
out_disable:
@@ -3415,6 +3376,8 @@ static const struct pci_device_id nvme_id_table[] = {
{ PCI_DEVICE(0x10ec, 0x5762), /* ADATA SX6000LNP */
.driver_data = NVME_QUIRK_IGNORE_DEV_SUBNQN |
NVME_QUIRK_BOGUS_NID, },
+ { PCI_DEVICE(0x10ec, 0x5763), /* ADATA SX6000PNP */
+ .driver_data = NVME_QUIRK_BOGUS_NID, },
{ PCI_DEVICE(0x1cc1, 0x8201), /* ADATA SX8200PNP 512GB */
.driver_data = NVME_QUIRK_NO_DEEPEST_PS |
NVME_QUIRK_IGNORE_DEV_SUBNQN, },
@@ -3493,7 +3456,8 @@ static const struct pci_device_id nvme_id_table[] = {
.driver_data = NVME_QUIRK_SINGLE_VECTOR |
NVME_QUIRK_128_BYTES_SQES |
NVME_QUIRK_SHARED_TAGS |
- NVME_QUIRK_SKIP_CID_GEN },
+ NVME_QUIRK_SKIP_CID_GEN |
+ NVME_QUIRK_IDENTIFY_CNS },
{ PCI_DEVICE_CLASS(PCI_CLASS_STORAGE_EXPRESS, 0xffffff) },
{ 0, }
};
@@ -3521,8 +3485,9 @@ static int __init nvme_init(void)
BUILD_BUG_ON(sizeof(struct nvme_create_sq) != 64);
BUILD_BUG_ON(sizeof(struct nvme_delete_queue) != 64);
BUILD_BUG_ON(IRQ_AFFINITY_MAX_SETS < 2);
- BUILD_BUG_ON(DIV_ROUND_UP(nvme_pci_npages_prp(), NVME_CTRL_PAGE_SIZE) >
- S8_MAX);
+ BUILD_BUG_ON(NVME_MAX_SEGS > SGES_PER_PAGE);
+ BUILD_BUG_ON(sizeof(struct scatterlist) * NVME_MAX_SEGS > PAGE_SIZE);
+ BUILD_BUG_ON(nvme_pci_npages_prp() > NVME_MAX_NR_ALLOCATIONS);
return pci_register_driver(&nvme_driver);
}
generated by cgit (git 2.34.1) at 2025-05-30 14:27:39 +0000