void *buf;
unsigned int len;
unsigned int flags;
+ int retries;
+ int errors;
+ unsigned int tx_index;
+ unsigned int rx_index;
struct ntb_transport_qp *qp;
union {
struct ntb_payload_header __iomem *tx_hdr;
struct ntb_payload_header *rx_hdr;
};
- unsigned int index;
};
struct ntb_rx_info {
unsigned int rx_index;
unsigned int rx_max_entry;
unsigned int rx_max_frame;
+ unsigned int rx_alloc_entry;
dma_cookie_t last_cookie;
struct tasklet_struct rxc_db_work;
static void ntb_transport_rxc_db(unsigned long data);
static const struct ntb_ctx_ops ntb_transport_ops;
static struct ntb_client ntb_transport_client;
+static int ntb_async_tx_submit(struct ntb_transport_qp *qp,
+ struct ntb_queue_entry *entry);
+static void ntb_memcpy_tx(struct ntb_queue_entry *entry, void __iomem *offset);
+static int ntb_async_rx_submit(struct ntb_queue_entry *entry, void *offset);
+static void ntb_memcpy_rx(struct ntb_queue_entry *entry, void *offset);
+
static int ntb_transport_bus_match(struct device *dev,
struct device_driver *drv)
out_offset += snprintf(buf + out_offset, out_count - out_offset,
"rx_index - \t%u\n", qp->rx_index);
out_offset += snprintf(buf + out_offset, out_count - out_offset,
- "rx_max_entry - \t%u\n\n", qp->rx_max_entry);
+ "rx_max_entry - \t%u\n", qp->rx_max_entry);
+ out_offset += snprintf(buf + out_offset, out_count - out_offset,
+ "rx_alloc_entry - \t%u\n\n", qp->rx_alloc_entry);
out_offset += snprintf(buf + out_offset, out_count - out_offset,
"tx_bytes - \t%llu\n", qp->tx_bytes);
{
struct ntb_transport_qp *qp = &nt->qp_vec[qp_num];
struct ntb_transport_mw *mw;
+ struct ntb_dev *ndev = nt->ndev;
+ struct ntb_queue_entry *entry;
unsigned int rx_size, num_qps_mw;
unsigned int mw_num, mw_count, qp_count;
unsigned int i;
+ int node;
mw_count = nt->mw_count;
qp_count = nt->qp_count;
qp->rx_max_entry = rx_size / qp->rx_max_frame;
qp->rx_index = 0;
+ /*
+ * Checking to see if we have more entries than the default.
+ * We should add additional entries if that is the case so we
+ * can be in sync with the transport frames.
+ */
+ node = dev_to_node(&ndev->dev);
+ for (i = qp->rx_alloc_entry; i < qp->rx_max_entry; i++) {
+ entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
+ if (!entry)
+ return -ENOMEM;
+
+ entry->qp = qp;
+ ntb_list_add(&qp->ntb_rx_q_lock, &entry->entry,
+ &qp->rx_free_q);
+ qp->rx_alloc_entry++;
+ }
+
qp->remote_rx_info->entry = qp->rx_max_entry - 1;
/* setup the hdr offsets with 0's */
int node;
int rc, i;
+ mw_count = ntb_mw_count(ndev);
+ if (ntb_spad_count(ndev) < (NUM_MWS + 1 + mw_count * 2)) {
+ dev_err(&ndev->dev, "Not enough scratch pad registers for %s",
+ NTB_TRANSPORT_NAME);
+ return -EIO;
+ }
+
if (ntb_db_is_unsafe(ndev))
dev_dbg(&ndev->dev,
"doorbell is unsafe, proceed anyway...\n");
nt->ndev = ndev;
- mw_count = ntb_mw_count(ndev);
-
nt->mw_count = mw_count;
nt->mw_vec = kzalloc_node(mw_count * sizeof(*nt->mw_vec),
break;
entry->rx_hdr->flags = 0;
- iowrite32(entry->index, &qp->rx_info->entry);
+ iowrite32(entry->rx_index, &qp->rx_info->entry);
cb_data = entry->cb_data;
len = entry->len;
spin_unlock_irqrestore(&qp->ntb_rx_q_lock, irqflags);
}
-static void ntb_rx_copy_callback(void *data)
+static void ntb_rx_copy_callback(void *data,
+ const struct dmaengine_result *res)
{
struct ntb_queue_entry *entry = data;
+ /* we need to check DMA results if we are using DMA */
+ if (res) {
+ enum dmaengine_tx_result dma_err = res->result;
+
+ switch (dma_err) {
+ case DMA_TRANS_READ_FAILED:
+ case DMA_TRANS_WRITE_FAILED:
+ entry->errors++;
+ case DMA_TRANS_ABORTED:
+ {
+ struct ntb_transport_qp *qp = entry->qp;
+ void *offset = qp->rx_buff + qp->rx_max_frame *
+ qp->rx_index;
+
+ ntb_memcpy_rx(entry, offset);
+ qp->rx_memcpy++;
+ return;
+ }
+
+ case DMA_TRANS_NOERROR:
+ default:
+ break;
+ }
+ }
+
entry->flags |= DESC_DONE_FLAG;
ntb_complete_rxc(entry->qp);
/* Ensure that the data is fully copied out before clearing the flag */
wmb();
- ntb_rx_copy_callback(entry);
+ ntb_rx_copy_callback(entry, NULL);
}
-static void ntb_async_rx(struct ntb_queue_entry *entry, void *offset)
+static int ntb_async_rx_submit(struct ntb_queue_entry *entry, void *offset)
{
struct dma_async_tx_descriptor *txd;
struct ntb_transport_qp *qp = entry->qp;
int retries = 0;
len = entry->len;
-
- if (!chan)
- goto err;
-
- if (len < copy_bytes)
- goto err;
-
device = chan->device;
pay_off = (size_t)offset & ~PAGE_MASK;
buff_off = (size_t)buf & ~PAGE_MASK;
unmap->from_cnt = 1;
for (retries = 0; retries < DMA_RETRIES; retries++) {
- txd = device->device_prep_dma_memcpy(chan, unmap->addr[1],
+ txd = device->device_prep_dma_memcpy(chan,
+ unmap->addr[1],
unmap->addr[0], len,
DMA_PREP_INTERRUPT);
if (txd)
goto err_get_unmap;
}
- txd->callback = ntb_rx_copy_callback;
+ txd->callback_result = ntb_rx_copy_callback;
txd->callback_param = entry;
dma_set_unmap(txd, unmap);
qp->rx_async++;
- return;
+ return 0;
err_set_unmap:
dmaengine_unmap_put(unmap);
err_get_unmap:
dmaengine_unmap_put(unmap);
+err:
+ return -ENXIO;
+}
+
+static void ntb_async_rx(struct ntb_queue_entry *entry, void *offset)
+{
+ struct ntb_transport_qp *qp = entry->qp;
+ struct dma_chan *chan = qp->rx_dma_chan;
+ int res;
+
+ if (!chan)
+ goto err;
+
+ if (entry->len < copy_bytes)
+ goto err;
+
+ res = ntb_async_rx_submit(entry, offset);
+ if (res < 0)
+ goto err;
+
+ if (!entry->retries)
+ qp->rx_async++;
+
+ return;
+
err:
ntb_memcpy_rx(entry, offset);
qp->rx_memcpy++;
}
entry->rx_hdr = hdr;
- entry->index = qp->rx_index;
+ entry->rx_index = qp->rx_index;
if (hdr->len > entry->len) {
dev_dbg(&qp->ndev->pdev->dev,
}
}
-static void ntb_tx_copy_callback(void *data)
+static void ntb_tx_copy_callback(void *data,
+ const struct dmaengine_result *res)
{
struct ntb_queue_entry *entry = data;
struct ntb_transport_qp *qp = entry->qp;
struct ntb_payload_header __iomem *hdr = entry->tx_hdr;
+ /* we need to check DMA results if we are using DMA */
+ if (res) {
+ enum dmaengine_tx_result dma_err = res->result;
+
+ switch (dma_err) {
+ case DMA_TRANS_READ_FAILED:
+ case DMA_TRANS_WRITE_FAILED:
+ entry->errors++;
+ case DMA_TRANS_ABORTED:
+ {
+ void __iomem *offset =
+ qp->tx_mw + qp->tx_max_frame *
+ entry->tx_index;
+
+ /* resubmit via CPU */
+ ntb_memcpy_tx(entry, offset);
+ qp->tx_memcpy++;
+ return;
+ }
+
+ case DMA_TRANS_NOERROR:
+ default:
+ break;
+ }
+ }
+
iowrite32(entry->flags | DESC_DONE_FLAG, &hdr->flags);
ntb_peer_db_set(qp->ndev, BIT_ULL(qp->qp_num));
/* Ensure that the data is fully copied out before setting the flags */
wmb();
- ntb_tx_copy_callback(entry);
+ ntb_tx_copy_callback(entry, NULL);
}
-static void ntb_async_tx(struct ntb_transport_qp *qp,
- struct ntb_queue_entry *entry)
+static int ntb_async_tx_submit(struct ntb_transport_qp *qp,
+ struct ntb_queue_entry *entry)
{
- struct ntb_payload_header __iomem *hdr;
struct dma_async_tx_descriptor *txd;
struct dma_chan *chan = qp->tx_dma_chan;
struct dma_device *device;
+ size_t len = entry->len;
+ void *buf = entry->buf;
size_t dest_off, buff_off;
struct dmaengine_unmap_data *unmap;
dma_addr_t dest;
dma_cookie_t cookie;
- void __iomem *offset;
- size_t len = entry->len;
- void *buf = entry->buf;
int retries = 0;
- offset = qp->tx_mw + qp->tx_max_frame * qp->tx_index;
- hdr = offset + qp->tx_max_frame - sizeof(struct ntb_payload_header);
- entry->tx_hdr = hdr;
-
- iowrite32(entry->len, &hdr->len);
- iowrite32((u32)qp->tx_pkts, &hdr->ver);
-
- if (!chan)
- goto err;
-
- if (len < copy_bytes)
- goto err;
-
device = chan->device;
- dest = qp->tx_mw_phys + qp->tx_max_frame * qp->tx_index;
+ dest = qp->tx_mw_phys + qp->tx_max_frame * entry->tx_index;
buff_off = (size_t)buf & ~PAGE_MASK;
dest_off = (size_t)dest & ~PAGE_MASK;
unmap->to_cnt = 1;
for (retries = 0; retries < DMA_RETRIES; retries++) {
- txd = device->device_prep_dma_memcpy(chan, dest, unmap->addr[0],
- len, DMA_PREP_INTERRUPT);
+ txd = device->device_prep_dma_memcpy(chan, dest,
+ unmap->addr[0], len,
+ DMA_PREP_INTERRUPT);
if (txd)
break;
goto err_get_unmap;
}
- txd->callback = ntb_tx_copy_callback;
+ txd->callback_result = ntb_tx_copy_callback;
txd->callback_param = entry;
dma_set_unmap(txd, unmap);
dmaengine_unmap_put(unmap);
dma_async_issue_pending(chan);
- qp->tx_async++;
- return;
+ return 0;
err_set_unmap:
dmaengine_unmap_put(unmap);
err_get_unmap:
dmaengine_unmap_put(unmap);
+err:
+ return -ENXIO;
+}
+
+static void ntb_async_tx(struct ntb_transport_qp *qp,
+ struct ntb_queue_entry *entry)
+{
+ struct ntb_payload_header __iomem *hdr;
+ struct dma_chan *chan = qp->tx_dma_chan;
+ void __iomem *offset;
+ int res;
+
+ entry->tx_index = qp->tx_index;
+ offset = qp->tx_mw + qp->tx_max_frame * entry->tx_index;
+ hdr = offset + qp->tx_max_frame - sizeof(struct ntb_payload_header);
+ entry->tx_hdr = hdr;
+
+ iowrite32(entry->len, &hdr->len);
+ iowrite32((u32)qp->tx_pkts, &hdr->ver);
+
+ if (!chan)
+ goto err;
+
+ if (entry->len < copy_bytes)
+ goto err;
+
+ res = ntb_async_tx_submit(qp, entry);
+ if (res < 0)
+ goto err;
+
+ if (!entry->retries)
+ qp->tx_async++;
+
+ return;
+
err:
ntb_memcpy_tx(entry, offset);
qp->tx_memcpy++;
ntb_list_add(&qp->ntb_rx_q_lock, &entry->entry,
&qp->rx_free_q);
}
+ qp->rx_alloc_entry = NTB_QP_DEF_NUM_ENTRIES;
- for (i = 0; i < NTB_QP_DEF_NUM_ENTRIES; i++) {
+ for (i = 0; i < qp->tx_max_entry; i++) {
entry = kzalloc_node(sizeof(*entry), GFP_ATOMIC, node);
if (!entry)
goto err2;
while ((entry = ntb_list_rm(&qp->ntb_tx_free_q_lock, &qp->tx_free_q)))
kfree(entry);
err1:
+ qp->rx_alloc_entry = 0;
while ((entry = ntb_list_rm(&qp->ntb_rx_q_lock, &qp->rx_free_q)))
kfree(entry);
if (qp->tx_dma_chan)
entry->buf = data;
entry->len = len;
entry->flags = 0;
+ entry->retries = 0;
+ entry->errors = 0;
+ entry->rx_index = 0;
ntb_list_add(&qp->ntb_rx_q_lock, &entry->entry, &qp->rx_pend_q);
entry->buf = data;
entry->len = len;
entry->flags = 0;
+ entry->errors = 0;
+ entry->retries = 0;
+ entry->tx_index = 0;
rc = ntb_process_tx(qp, entry);
if (rc)