static int alloc_host_sq(struct c4iw_rdev *rdev, struct t4_sq *sq)
{
sq->queue = dma_alloc_coherent(&(rdev->lldi.pdev->dev), sq->memsize,
&(sq->dma_addr), GFP_KERNEL);
if (!sq->queue)
return -ENOMEM;
sq->phys_addr = virt_to_phys(sq->queue);
pci_unmap_addr_set(sq, mapping, sq->dma_addr);
return 0;
}
static int b44_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct b44 *bp = netdev_priv(dev);
dma_addr_t mapping;
u32 len, entry, ctrl;
len = skb->len;
spin_lock_irq(&bp->lock);
/* This is a hard error, log it. */
if (unlikely(TX_BUFFS_AVAIL(bp) < 1)) {
netif_stop_queue(dev);
spin_unlock_irq(&bp->lock);
printk(KERN_ERR PFX "%s: BUG! Tx Ring full when queue awake!\n",
dev->name);
return 1;
}
entry = bp->tx_prod;
mapping = pci_map_single(bp->pdev, skb->data, len, PCI_DMA_TODEVICE);
bp->tx_buffers[entry].skb = skb;
pci_unmap_addr_set(&bp->tx_buffers[entry], mapping, mapping);
ctrl = (len & DESC_CTRL_LEN);
ctrl |= DESC_CTRL_IOC | DESC_CTRL_SOF | DESC_CTRL_EOF;
if (entry == (B44_TX_RING_SIZE - 1))
ctrl |= DESC_CTRL_EOT;
bp->tx_ring[entry].ctrl = cpu_to_le32(ctrl);
bp->tx_ring[entry].addr = cpu_to_le32((u32) mapping+bp->dma_offset);
entry = NEXT_TX(entry);
bp->tx_prod = entry;
wmb();
bw32(bp, B44_DMATX_PTR, entry * sizeof(struct dma_desc));
if (bp->flags & B44_FLAG_BUGGY_TXPTR)
bw32(bp, B44_DMATX_PTR, entry * sizeof(struct dma_desc));
if (bp->flags & B44_FLAG_REORDER_BUG)
br32(bp, B44_DMATX_PTR);
if (TX_BUFFS_AVAIL(bp) < 1)
netif_stop_queue(dev);
spin_unlock_irq(&bp->lock);
dev->trans_start = jiffies;
return 0;
}
static int alloc_host_sq(struct c4iw_rdev *rdev, struct t4_sq *sq)
{
sq->queue = contigmalloc(sq->memsize, M_DEVBUF, M_NOWAIT, 0ul, ~0ul,
4096, 0);
if (sq->queue)
sq->dma_addr = vtophys(sq->queue);
else
return -ENOMEM;
sq->phys_addr = vtophys(sq->queue);
pci_unmap_addr_set(sq, mapping, sq->dma_addr);
CTR4(KTR_IW_CXGBE, "%s sq %p dma_addr %p phys_addr %p", __func__,
sq->queue, sq->dma_addr, sq->phys_addr);
return 0;
}
/* Works like this. This chip writes a 'struct rx_header" 30 bytes
* before the DMA address you give it. So we allocate 30 more bytes
* for the RX buffer, DMA map all of it, skb_reserve the 30 bytes, then
* point the chip at 30 bytes past where the rx_header will go.
*/
static int b44_alloc_rx_skb(struct b44 *bp, int src_idx, u32 dest_idx_unmasked)
{
struct dma_desc *dp;
struct ring_info *src_map, *map;
struct rx_header *rh;
struct sk_buff *skb;
dma_addr_t mapping;
int dest_idx;
u32 ctrl;
src_map = NULL;
if (src_idx >= 0)
src_map = &bp->rx_buffers[src_idx];
dest_idx = dest_idx_unmasked & (B44_RX_RING_SIZE - 1);
map = &bp->rx_buffers[dest_idx];
skb = dev_alloc_skb(RX_PKT_BUF_SZ);
if (skb == NULL)
return -ENOMEM;
skb->dev = bp->dev;
mapping = pci_map_single(bp->pdev, skb->data,
RX_PKT_BUF_SZ,
PCI_DMA_FROMDEVICE);
skb_reserve(skb, bp->rx_offset);
rh = (struct rx_header *)
(skb->data - bp->rx_offset);
rh->len = 0;
rh->flags = 0;
map->skb = skb;
pci_unmap_addr_set(map, mapping, mapping);
if (src_map != NULL)
src_map->skb = NULL;
ctrl = (DESC_CTRL_LEN & (RX_PKT_BUF_SZ - bp->rx_offset));
if (dest_idx == (B44_RX_RING_SIZE - 1))
ctrl |= DESC_CTRL_EOT;
dp = &bp->rx_ring[dest_idx];
dp->ctrl = cpu_to_le32(ctrl);
dp->addr = cpu_to_le32((u32) mapping + bp->rx_offset + bp->dma_offset);
return RX_PKT_BUF_SZ;
}
static int c2_alloc_cq_buf(struct c2_dev *c2dev, struct c2_mq *mq, int q_size,
int msg_size)
{
u8 *pool_start;
pool_start = dma_alloc_coherent(&c2dev->pcidev->dev, q_size * msg_size,
&mq->host_dma, GFP_KERNEL);
if (!pool_start)
return -ENOMEM;
c2_mq_rep_init(mq,
0, /* index (currently unknown) */
q_size,
msg_size,
pool_start,
NULL, /* peer (currently unknown) */
C2_MQ_HOST_TARGET);
pci_unmap_addr_set(mq, mapping, mq->host_dma);
return 0;
}
static int create_cq(struct c4iw_rdev *rdev, struct t4_cq *cq,
struct c4iw_dev_ucontext *uctx)
{
struct fw_ri_res_wr *res_wr;
struct fw_ri_res *res;
int wr_len;
int user = (uctx != &rdev->uctx);
struct c4iw_wr_wait wr_wait;
int ret;
struct sk_buff *skb;
cq->cqid = c4iw_get_cqid(rdev, uctx);
if (!cq->cqid) {
ret = -ENOMEM;
goto err1;
}
if (!user) {
cq->sw_queue = kzalloc(cq->memsize, GFP_KERNEL);
if (!cq->sw_queue) {
ret = -ENOMEM;
goto err2;
}
}
cq->queue = dma_alloc_coherent(&rdev->lldi.pdev->dev, cq->memsize,
&cq->dma_addr, GFP_KERNEL);
if (!cq->queue) {
ret = -ENOMEM;
goto err3;
}
pci_unmap_addr_set(cq, mapping, cq->dma_addr);
memset(cq->queue, 0, cq->memsize);
/* build fw_ri_res_wr */
wr_len = sizeof *res_wr + sizeof *res;
skb = alloc_skb(wr_len, GFP_KERNEL);
if (!skb) {
ret = -ENOMEM;
goto err4;
}
set_wr_txq(skb, CPL_PRIORITY_CONTROL, 0);
res_wr = (struct fw_ri_res_wr *)__skb_put(skb, wr_len);
memset(res_wr, 0, wr_len);
res_wr->op_nres = cpu_to_be32(
FW_WR_OP(FW_RI_RES_WR) |
V_FW_RI_RES_WR_NRES(1) |
FW_WR_COMPL(1));
res_wr->len16_pkd = cpu_to_be32(DIV_ROUND_UP(wr_len, 16));
res_wr->cookie = (unsigned long) &wr_wait;
res = res_wr->res;
res->u.cq.restype = FW_RI_RES_TYPE_CQ;
res->u.cq.op = FW_RI_RES_OP_WRITE;
res->u.cq.iqid = cpu_to_be32(cq->cqid);
res->u.cq.iqandst_to_iqandstindex = cpu_to_be32(
V_FW_RI_RES_WR_IQANUS(0) |
V_FW_RI_RES_WR_IQANUD(1) |
F_FW_RI_RES_WR_IQANDST |
V_FW_RI_RES_WR_IQANDSTINDEX(*rdev->lldi.rxq_ids));
res->u.cq.iqdroprss_to_iqesize = cpu_to_be16(
F_FW_RI_RES_WR_IQDROPRSS |
V_FW_RI_RES_WR_IQPCIECH(2) |
V_FW_RI_RES_WR_IQINTCNTTHRESH(0) |
F_FW_RI_RES_WR_IQO |
V_FW_RI_RES_WR_IQESIZE(1));
res->u.cq.iqsize = cpu_to_be16(cq->size);
res->u.cq.iqaddr = cpu_to_be64(cq->dma_addr);
c4iw_init_wr_wait(&wr_wait);
ret = c4iw_ofld_send(rdev, skb);
if (ret)
goto err4;
PDBG("%s wait_event wr_wait %p\n", __func__, &wr_wait);
ret = c4iw_wait_for_reply(rdev, &wr_wait, 0, 0, __func__);
if (ret)
goto err4;
cq->gen = 1;
cq->gts = rdev->lldi.gts_reg;
cq->rdev = rdev;
if (user) {
cq->ugts = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
(cq->cqid << rdev->cqshift);
cq->ugts &= PAGE_MASK;
}
return 0;
err4:
dma_free_coherent(&rdev->lldi.pdev->dev, cq->memsize, cq->queue,
pci_unmap_addr(cq, mapping));
err3:
kfree(cq->sw_queue);
err2:
c4iw_put_cqid(rdev, cq->cqid, uctx);
err1:
return ret;
}
/*
* Allocate and register buffer for WQEs. qp->rq.max, sq.max,
* rq.max_gs and sq.max_gs must all be assigned.
* mthca_alloc_wqe_buf will calculate rq.wqe_shift and
* sq.wqe_shift (as well as send_wqe_offset, is_direct, and
* queue)
*/
static int mthca_alloc_wqe_buf(struct mthca_dev *dev,
struct mthca_pd *pd,
struct mthca_qp *qp)
{
int size;
int i;
int npages, shift;
dma_addr_t t;
u64 *dma_list = NULL;
int err = -ENOMEM;
size = sizeof (struct mthca_next_seg) +
qp->rq.max_gs * sizeof (struct mthca_data_seg);
for (qp->rq.wqe_shift = 6; 1 << qp->rq.wqe_shift < size;
qp->rq.wqe_shift++)
; /* nothing */
size = sizeof (struct mthca_next_seg) +
qp->sq.max_gs * sizeof (struct mthca_data_seg);
if (qp->transport == MLX)
size += 2 * sizeof (struct mthca_data_seg);
else if (qp->transport == UD)
size += sizeof (struct mthca_ud_seg);
else /* bind seg is as big as atomic + raddr segs */
size += sizeof (struct mthca_bind_seg);
for (qp->sq.wqe_shift = 6; 1 << qp->sq.wqe_shift < size;
qp->sq.wqe_shift++)
; /* nothing */
qp->send_wqe_offset = ALIGN(qp->rq.max << qp->rq.wqe_shift,
1 << qp->sq.wqe_shift);
size = PAGE_ALIGN(qp->send_wqe_offset +
(qp->sq.max << qp->sq.wqe_shift));
qp->wrid = kmalloc((qp->rq.max + qp->sq.max) * sizeof (u64),
GFP_KERNEL);
if (!qp->wrid)
goto err_out;
if (size <= MTHCA_MAX_DIRECT_QP_SIZE) {
qp->is_direct = 1;
npages = 1;
shift = get_order(size) + PAGE_SHIFT;
if (0)
mthca_dbg(dev, "Creating direct QP of size %d (shift %d)\n",
size, shift);
qp->queue.direct.buf = pci_alloc_consistent(dev->pdev, size, &t);
if (!qp->queue.direct.buf)
goto err_out;
pci_unmap_addr_set(&qp->queue.direct, mapping, t);
memset(qp->queue.direct.buf, 0, size);
while (t & ((1 << shift) - 1)) {
--shift;
npages *= 2;
}
dma_list = kmalloc(npages * sizeof *dma_list, GFP_KERNEL);
if (!dma_list)
goto err_out_free;
for (i = 0; i < npages; ++i)
dma_list[i] = t + i * (1 << shift);
} else {
qp->is_direct = 0;
npages = size / PAGE_SIZE;
shift = PAGE_SHIFT;
if (0)
mthca_dbg(dev, "Creating indirect QP with %d pages\n", npages);
dma_list = kmalloc(npages * sizeof *dma_list, GFP_KERNEL);
if (!dma_list)
goto err_out;
qp->queue.page_list = kmalloc(npages *
sizeof *qp->queue.page_list,
GFP_KERNEL);
if (!qp->queue.page_list)
goto err_out;
for (i = 0; i < npages; ++i) {
qp->queue.page_list[i].buf =
pci_alloc_consistent(dev->pdev, PAGE_SIZE, &t);
if (!qp->queue.page_list[i].buf)
goto err_out_free;
memset(qp->queue.page_list[i].buf, 0, PAGE_SIZE);
pci_unmap_addr_set(&qp->queue.page_list[i], mapping, t);
dma_list[i] = t;
}
}
err = mthca_mr_alloc_phys(dev, pd->pd_num, dma_list, shift,
npages, 0, size,
MTHCA_MPT_FLAG_LOCAL_WRITE |
MTHCA_MPT_FLAG_LOCAL_READ,
&qp->mr);
if (err)
goto err_out_free;
kfree(dma_list);
return 0;
err_out_free:
if (qp->is_direct) {
pci_free_consistent(dev->pdev, size,
qp->queue.direct.buf,
pci_unmap_addr(&qp->queue.direct, mapping));
} else
for (i = 0; i < npages; ++i) {
if (qp->queue.page_list[i].buf)
pci_free_consistent(dev->pdev, PAGE_SIZE,
qp->queue.page_list[i].buf,
pci_unmap_addr(&qp->queue.page_list[i],
mapping));
}
err_out:
kfree(qp->wrid);
kfree(dma_list);
return err;
}
static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
struct t4_cq *rcq, struct t4_cq *scq,
struct c4iw_dev_ucontext *uctx)
{
int user = (uctx != &rdev->uctx);
struct fw_ri_res_wr *res_wr;
struct fw_ri_res *res;
int wr_len;
struct c4iw_wr_wait wr_wait;
struct sk_buff *skb;
int ret;
int eqsize;
wq->sq.qid = c4iw_get_qpid(rdev, uctx);
if (!wq->sq.qid)
return -ENOMEM;
wq->rq.qid = c4iw_get_qpid(rdev, uctx);
if (!wq->rq.qid)
goto err1;
if (!user) {
wq->sq.sw_sq = kzalloc(wq->sq.size * sizeof *wq->sq.sw_sq,
GFP_KERNEL);
if (!wq->sq.sw_sq)
goto err2;
wq->rq.sw_rq = kzalloc(wq->rq.size * sizeof *wq->rq.sw_rq,
GFP_KERNEL);
if (!wq->rq.sw_rq)
goto err3;
}
/*
* RQT must be a power of 2.
*/
wq->rq.rqt_size = roundup_pow_of_two(wq->rq.size);
wq->rq.rqt_hwaddr = c4iw_rqtpool_alloc(rdev, wq->rq.rqt_size);
if (!wq->rq.rqt_hwaddr)
goto err4;
if (user) {
if (alloc_oc_sq(rdev, &wq->sq) && alloc_host_sq(rdev, &wq->sq))
goto err5;
} else
if (alloc_host_sq(rdev, &wq->sq))
goto err5;
memset(wq->sq.queue, 0, wq->sq.memsize);
pci_unmap_addr_set(&wq->sq, mapping, wq->sq.dma_addr);
wq->rq.queue = dma_alloc_coherent(&(rdev->lldi.pdev->dev),
wq->rq.memsize, &(wq->rq.dma_addr),
GFP_KERNEL);
if (!wq->rq.queue)
goto err6;
PDBG("%s sq base va 0x%p pa 0x%llx rq base va 0x%p pa 0x%llx\n",
__func__, wq->sq.queue,
(unsigned long long)virt_to_phys(wq->sq.queue),
wq->rq.queue,
(unsigned long long)virt_to_phys(wq->rq.queue));
memset(wq->rq.queue, 0, wq->rq.memsize);
pci_unmap_addr_set(&wq->rq, mapping, wq->rq.dma_addr);
wq->db = rdev->lldi.db_reg;
wq->gts = rdev->lldi.gts_reg;
if (user) {
wq->sq.udb = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
(wq->sq.qid << rdev->qpshift);
wq->sq.udb &= PAGE_MASK;
wq->rq.udb = (u64)pci_resource_start(rdev->lldi.pdev, 2) +
(wq->rq.qid << rdev->qpshift);
wq->rq.udb &= PAGE_MASK;
}
wq->rdev = rdev;
wq->rq.msn = 1;
/* build fw_ri_res_wr */
wr_len = sizeof *res_wr + 2 * sizeof *res;
skb = alloc_skb(wr_len, GFP_KERNEL);
if (!skb) {
ret = -ENOMEM;
goto err7;
}
set_wr_txq(skb, CPL_PRIORITY_CONTROL, 0);
res_wr = (struct fw_ri_res_wr *)__skb_put(skb, wr_len);
memset(res_wr, 0, wr_len);
res_wr->op_nres = cpu_to_be32(
FW_WR_OP(FW_RI_RES_WR) |
V_FW_RI_RES_WR_NRES(2) |
FW_WR_COMPL(1));
res_wr->len16_pkd = cpu_to_be32(DIV_ROUND_UP(wr_len, 16));
res_wr->cookie = (unsigned long) &wr_wait;
res = res_wr->res;
res->u.sqrq.restype = FW_RI_RES_TYPE_SQ;
res->u.sqrq.op = FW_RI_RES_OP_WRITE;
/*
* eqsize is the number of 64B entries plus the status page size.
*/
eqsize = wq->sq.size * T4_SQ_NUM_SLOTS + T4_EQ_STATUS_ENTRIES;
res->u.sqrq.fetchszm_to_iqid = cpu_to_be32(
V_FW_RI_RES_WR_HOSTFCMODE(0) | /* no host cidx updates */
V_FW_RI_RES_WR_CPRIO(0) | /* don't keep in chip cache */
V_FW_RI_RES_WR_PCIECHN(0) | /* set by uP at ri_init time */
(t4_sq_onchip(&wq->sq) ? F_FW_RI_RES_WR_ONCHIP : 0) |
V_FW_RI_RES_WR_IQID(scq->cqid));
res->u.sqrq.dcaen_to_eqsize = cpu_to_be32(
V_FW_RI_RES_WR_DCAEN(0) |
V_FW_RI_RES_WR_DCACPU(0) |
V_FW_RI_RES_WR_FBMIN(2) |
V_FW_RI_RES_WR_FBMAX(2) |
V_FW_RI_RES_WR_CIDXFTHRESHO(0) |
V_FW_RI_RES_WR_CIDXFTHRESH(0) |
V_FW_RI_RES_WR_EQSIZE(eqsize));
res->u.sqrq.eqid = cpu_to_be32(wq->sq.qid);
res->u.sqrq.eqaddr = cpu_to_be64(wq->sq.dma_addr);
res++;
res->u.sqrq.restype = FW_RI_RES_TYPE_RQ;
res->u.sqrq.op = FW_RI_RES_OP_WRITE;
/*
* eqsize is the number of 64B entries plus the status page size.
*/
eqsize = wq->rq.size * T4_RQ_NUM_SLOTS + T4_EQ_STATUS_ENTRIES;
res->u.sqrq.fetchszm_to_iqid = cpu_to_be32(
V_FW_RI_RES_WR_HOSTFCMODE(0) | /* no host cidx updates */
V_FW_RI_RES_WR_CPRIO(0) | /* don't keep in chip cache */
V_FW_RI_RES_WR_PCIECHN(0) | /* set by uP at ri_init time */
V_FW_RI_RES_WR_IQID(rcq->cqid));
res->u.sqrq.dcaen_to_eqsize = cpu_to_be32(
V_FW_RI_RES_WR_DCAEN(0) |
V_FW_RI_RES_WR_DCACPU(0) |
V_FW_RI_RES_WR_FBMIN(2) |
V_FW_RI_RES_WR_FBMAX(2) |
V_FW_RI_RES_WR_CIDXFTHRESHO(0) |
V_FW_RI_RES_WR_CIDXFTHRESH(0) |
V_FW_RI_RES_WR_EQSIZE(eqsize));
res->u.sqrq.eqid = cpu_to_be32(wq->rq.qid);
res->u.sqrq.eqaddr = cpu_to_be64(wq->rq.dma_addr);
c4iw_init_wr_wait(&wr_wait);
ret = c4iw_ofld_send(rdev, skb);
if (ret)
goto err7;
ret = c4iw_wait_for_reply(rdev, &wr_wait, 0, wq->sq.qid, __func__);
if (ret)
goto err7;
PDBG("%s sqid 0x%x rqid 0x%x kdb 0x%p squdb 0x%llx rqudb 0x%llx\n",
__func__, wq->sq.qid, wq->rq.qid, wq->db,
(unsigned long long)wq->sq.udb, (unsigned long long)wq->rq.udb);
return 0;
err7:
dma_free_coherent(&(rdev->lldi.pdev->dev),
wq->rq.memsize, wq->rq.queue,
pci_unmap_addr(&wq->rq, mapping));
err6:
dealloc_sq(rdev, &wq->sq);
err5:
c4iw_rqtpool_free(rdev, wq->rq.rqt_hwaddr, wq->rq.rqt_size);
err4:
kfree(wq->rq.sw_rq);
err3:
kfree(wq->sq.sw_sq);
err2:
c4iw_put_qpid(rdev, wq->rq.qid, uctx);
err1:
c4iw_put_qpid(rdev, wq->sq.qid, uctx);
return -ENOMEM;
}
int mthca_buf_alloc(struct mthca_dev *dev, int size, int max_direct,
union mthca_buf *buf, int *is_direct, struct mthca_pd *pd,
int hca_write, struct mthca_mr *mr)
{
int err = -ENOMEM;
int npages, shift;
u64 *dma_list = NULL;
dma_addr_t t;
int i;
if (size <= max_direct) {
*is_direct = 1;
npages = 1;
shift = get_order(size) + PAGE_SHIFT;
buf->direct.buf = dma_alloc_coherent(&dev->pdev->dev,
size, &t, GFP_KERNEL);
if (!buf->direct.buf)
return -ENOMEM;
pci_unmap_addr_set(&buf->direct, mapping, t);
memset(buf->direct.buf, 0, size);
while (t & ((1 << shift) - 1)) {
--shift;
npages *= 2;
}
dma_list = kmalloc(npages * sizeof *dma_list, GFP_KERNEL);
if (!dma_list)
goto err_free;
for (i = 0; i < npages; ++i)
dma_list[i] = t + i * (1 << shift);
} else {
*is_direct = 0;
npages = (size + PAGE_SIZE - 1) / PAGE_SIZE;
shift = PAGE_SHIFT;
dma_list = kmalloc(npages * sizeof *dma_list, GFP_KERNEL);
if (!dma_list)
return -ENOMEM;
buf->page_list = kmalloc(npages * sizeof *buf->page_list,
GFP_KERNEL);
if (!buf->page_list)
goto err_out;
for (i = 0; i < npages; ++i)
buf->page_list[i].buf = NULL;
for (i = 0; i < npages; ++i) {
buf->page_list[i].buf =
dma_alloc_coherent(&dev->pdev->dev, PAGE_SIZE,
&t, GFP_KERNEL);
if (!buf->page_list[i].buf)
goto err_free;
dma_list[i] = t;
pci_unmap_addr_set(&buf->page_list[i], mapping, t);
memset(buf->page_list[i].buf, 0, PAGE_SIZE);
}
}
err = mthca_mr_alloc_phys(dev, pd->pd_num,
dma_list, shift, npages,
0, size,
MTHCA_MPT_FLAG_LOCAL_READ |
(hca_write ? MTHCA_MPT_FLAG_LOCAL_WRITE : 0),
mr);
if (err)
goto err_free;
kfree(dma_list);
return 0;
err_free:
mthca_buf_free(dev, size, buf, *is_direct, NULL);
err_out:
kfree(dma_list);
return err;
}
static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
struct t4_cq *rcq, struct t4_cq *scq,
struct c4iw_dev_ucontext *uctx)
{
struct adapter *sc = rdev->adap;
int user = (uctx != &rdev->uctx);
struct fw_ri_res_wr *res_wr;
struct fw_ri_res *res;
int wr_len;
struct c4iw_wr_wait wr_wait;
int ret;
int eqsize;
struct wrqe *wr;
wq->sq.qid = c4iw_get_qpid(rdev, uctx);
if (!wq->sq.qid)
return -ENOMEM;
wq->rq.qid = c4iw_get_qpid(rdev, uctx);
if (!wq->rq.qid)
goto err1;
if (!user) {
wq->sq.sw_sq = kzalloc(wq->sq.size * sizeof *wq->sq.sw_sq,
GFP_KERNEL);
if (!wq->sq.sw_sq)
goto err2;
wq->rq.sw_rq = kzalloc(wq->rq.size * sizeof *wq->rq.sw_rq,
GFP_KERNEL);
if (!wq->rq.sw_rq)
goto err3;
}
/* RQT must be a power of 2. */
wq->rq.rqt_size = roundup_pow_of_two(wq->rq.size);
wq->rq.rqt_hwaddr = c4iw_rqtpool_alloc(rdev, wq->rq.rqt_size);
if (!wq->rq.rqt_hwaddr)
goto err4;
if (alloc_host_sq(rdev, &wq->sq))
goto err5;
memset(wq->sq.queue, 0, wq->sq.memsize);
pci_unmap_addr_set(&wq->sq, mapping, wq->sq.dma_addr);
wq->rq.queue = contigmalloc(wq->rq.memsize,
M_DEVBUF, M_NOWAIT, 0ul, ~0ul, 4096, 0);
if (wq->rq.queue)
wq->rq.dma_addr = vtophys(wq->rq.queue);
else
goto err6;
CTR5(KTR_IW_CXGBE,
"%s sq base va 0x%p pa 0x%llx rq base va 0x%p pa 0x%llx", __func__,
wq->sq.queue, (unsigned long long)vtophys(wq->sq.queue),
wq->rq.queue, (unsigned long long)vtophys(wq->rq.queue));
memset(wq->rq.queue, 0, wq->rq.memsize);
pci_unmap_addr_set(&wq->rq, mapping, wq->rq.dma_addr);
wq->db = (void *)((unsigned long)rman_get_virtual(sc->regs_res) +
MYPF_REG(SGE_PF_KDOORBELL));
wq->gts = (void *)((unsigned long)rman_get_virtual(rdev->adap->regs_res)
+ MYPF_REG(SGE_PF_GTS));
if (user) {
wq->sq.udb = (u64)((char*)rman_get_virtual(rdev->adap->udbs_res) +
(wq->sq.qid << rdev->qpshift));
wq->sq.udb &= PAGE_MASK;
wq->rq.udb = (u64)((char*)rman_get_virtual(rdev->adap->udbs_res) +
(wq->rq.qid << rdev->qpshift));
wq->rq.udb &= PAGE_MASK;
}
wq->rdev = rdev;
wq->rq.msn = 1;
/* build fw_ri_res_wr */
wr_len = sizeof *res_wr + 2 * sizeof *res;
wr = alloc_wrqe(wr_len, &sc->sge.mgmtq);
if (wr == NULL)
return (0);
res_wr = wrtod(wr);
memset(res_wr, 0, wr_len);
res_wr->op_nres = cpu_to_be32(
V_FW_WR_OP(FW_RI_RES_WR) |
V_FW_RI_RES_WR_NRES(2) |
F_FW_WR_COMPL);
res_wr->len16_pkd = cpu_to_be32(DIV_ROUND_UP(wr_len, 16));
res_wr->cookie = (unsigned long) &wr_wait;
res = res_wr->res;
res->u.sqrq.restype = FW_RI_RES_TYPE_SQ;
res->u.sqrq.op = FW_RI_RES_OP_WRITE;
/* eqsize is the number of 64B entries plus the status page size. */
eqsize = wq->sq.size * T4_SQ_NUM_SLOTS + spg_creds;
res->u.sqrq.fetchszm_to_iqid = cpu_to_be32(
V_FW_RI_RES_WR_HOSTFCMODE(0) | /* no host cidx updates */
V_FW_RI_RES_WR_CPRIO(0) | /* don't keep in chip cache */
V_FW_RI_RES_WR_PCIECHN(0) | /* set by uP at ri_init time */
V_FW_RI_RES_WR_IQID(scq->cqid));
res->u.sqrq.dcaen_to_eqsize = cpu_to_be32(
V_FW_RI_RES_WR_DCAEN(0) |
V_FW_RI_RES_WR_DCACPU(0) |
V_FW_RI_RES_WR_FBMIN(2) |
V_FW_RI_RES_WR_FBMAX(2) |
V_FW_RI_RES_WR_CIDXFTHRESHO(0) |
V_FW_RI_RES_WR_CIDXFTHRESH(0) |
V_FW_RI_RES_WR_EQSIZE(eqsize));
res->u.sqrq.eqid = cpu_to_be32(wq->sq.qid);
res->u.sqrq.eqaddr = cpu_to_be64(wq->sq.dma_addr);
res++;
res->u.sqrq.restype = FW_RI_RES_TYPE_RQ;
res->u.sqrq.op = FW_RI_RES_OP_WRITE;
/* eqsize is the number of 64B entries plus the status page size. */
eqsize = wq->rq.size * T4_RQ_NUM_SLOTS + spg_creds ;
res->u.sqrq.fetchszm_to_iqid = cpu_to_be32(
V_FW_RI_RES_WR_HOSTFCMODE(0) | /* no host cidx updates */
V_FW_RI_RES_WR_CPRIO(0) | /* don't keep in chip cache */
V_FW_RI_RES_WR_PCIECHN(0) | /* set by uP at ri_init time */
V_FW_RI_RES_WR_IQID(rcq->cqid));
res->u.sqrq.dcaen_to_eqsize = cpu_to_be32(
V_FW_RI_RES_WR_DCAEN(0) |
V_FW_RI_RES_WR_DCACPU(0) |
V_FW_RI_RES_WR_FBMIN(2) |
V_FW_RI_RES_WR_FBMAX(2) |
V_FW_RI_RES_WR_CIDXFTHRESHO(0) |
V_FW_RI_RES_WR_CIDXFTHRESH(0) |
V_FW_RI_RES_WR_EQSIZE(eqsize));
res->u.sqrq.eqid = cpu_to_be32(wq->rq.qid);
res->u.sqrq.eqaddr = cpu_to_be64(wq->rq.dma_addr);
c4iw_init_wr_wait(&wr_wait);
t4_wrq_tx(sc, wr);
ret = c4iw_wait_for_reply(rdev, &wr_wait, 0, wq->sq.qid, __func__);
if (ret)
goto err7;
CTR6(KTR_IW_CXGBE,
"%s sqid 0x%x rqid 0x%x kdb 0x%p squdb 0x%llx rqudb 0x%llx",
__func__, wq->sq.qid, wq->rq.qid, wq->db,
(unsigned long long)wq->sq.udb, (unsigned long long)wq->rq.udb);
return 0;
err7:
contigfree(wq->rq.queue, wq->rq.memsize, M_DEVBUF);
err6:
dealloc_sq(rdev, &wq->sq);
err5:
c4iw_rqtpool_free(rdev, wq->rq.rqt_hwaddr, wq->rq.rqt_size);
err4:
kfree(wq->rq.sw_rq);
err3:
kfree(wq->sq.sw_sq);
err2:
c4iw_put_qpid(rdev, wq->rq.qid, uctx);
err1:
c4iw_put_qpid(rdev, wq->sq.qid, uctx);
return -ENOMEM;
}
/*
* Called by c2_probe to initialize the RNIC. This principally
* involves initalizing the various limits and resouce pools that
* comprise the RNIC instance.
*/
int __devinit c2_rnic_init(struct c2_dev *c2dev)
{
int err;
u32 qsize, msgsize;
void *q1_pages;
void *q2_pages;
void __iomem *mmio_regs;
/* Device capabilities */
c2dev->device_cap_flags =
(IB_DEVICE_RESIZE_MAX_WR |
IB_DEVICE_CURR_QP_STATE_MOD |
IB_DEVICE_SYS_IMAGE_GUID |
IB_DEVICE_ZERO_STAG |
IB_DEVICE_MEM_WINDOW);
/* Allocate the qptr_array */
c2dev->qptr_array = vmalloc(C2_MAX_CQS * sizeof(void *));
if (!c2dev->qptr_array) {
return -ENOMEM;
}
/* Inialize the qptr_array */
memset(c2dev->qptr_array, 0, C2_MAX_CQS * sizeof(void *));
c2dev->qptr_array[0] = (void *) &c2dev->req_vq;
c2dev->qptr_array[1] = (void *) &c2dev->rep_vq;
c2dev->qptr_array[2] = (void *) &c2dev->aeq;
/* Initialize data structures */
init_waitqueue_head(&c2dev->req_vq_wo);
spin_lock_init(&c2dev->vqlock);
spin_lock_init(&c2dev->lock);
/* Allocate MQ shared pointer pool for kernel clients. User
* mode client pools are hung off the user context
*/
err = c2_init_mqsp_pool(c2dev, GFP_KERNEL, &c2dev->kern_mqsp_pool);
if (err) {
goto bail0;
}
/* Allocate shared pointers for Q0, Q1, and Q2 from
* the shared pointer pool.
*/
c2dev->hint_count = c2_alloc_mqsp(c2dev, c2dev->kern_mqsp_pool,
&c2dev->hint_count_dma,
GFP_KERNEL);
c2dev->req_vq.shared = c2_alloc_mqsp(c2dev, c2dev->kern_mqsp_pool,
&c2dev->req_vq.shared_dma,
GFP_KERNEL);
c2dev->rep_vq.shared = c2_alloc_mqsp(c2dev, c2dev->kern_mqsp_pool,
&c2dev->rep_vq.shared_dma,
GFP_KERNEL);
c2dev->aeq.shared = c2_alloc_mqsp(c2dev, c2dev->kern_mqsp_pool,
&c2dev->aeq.shared_dma, GFP_KERNEL);
if (!c2dev->hint_count || !c2dev->req_vq.shared ||
!c2dev->rep_vq.shared || !c2dev->aeq.shared) {
err = -ENOMEM;
goto bail1;
}
mmio_regs = c2dev->kva;
/* Initialize the Verbs Request Queue */
c2_mq_req_init(&c2dev->req_vq, 0,
be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q0_QSIZE)),
be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q0_MSGSIZE)),
mmio_regs +
be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q0_POOLSTART)),
mmio_regs +
be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q0_SHARED)),
C2_MQ_ADAPTER_TARGET);
/* Initialize the Verbs Reply Queue */
qsize = be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q1_QSIZE));
msgsize = be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q1_MSGSIZE));
q1_pages = dma_alloc_coherent(&c2dev->pcidev->dev, qsize * msgsize,
&c2dev->rep_vq.host_dma, GFP_KERNEL);
if (!q1_pages) {
err = -ENOMEM;
goto bail1;
}
pci_unmap_addr_set(&c2dev->rep_vq, mapping, c2dev->rep_vq.host_dma);
pr_debug("%s rep_vq va %p dma %llx\n", __func__, q1_pages,
(unsigned long long) c2dev->rep_vq.host_dma);
c2_mq_rep_init(&c2dev->rep_vq,
1,
qsize,
msgsize,
q1_pages,
mmio_regs +
be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q1_SHARED)),
C2_MQ_HOST_TARGET);
/* Initialize the Asynchronus Event Queue */
qsize = be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q2_QSIZE));
msgsize = be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q2_MSGSIZE));
q2_pages = dma_alloc_coherent(&c2dev->pcidev->dev, qsize * msgsize,
&c2dev->aeq.host_dma, GFP_KERNEL);
if (!q2_pages) {
err = -ENOMEM;
goto bail2;
}
pci_unmap_addr_set(&c2dev->aeq, mapping, c2dev->aeq.host_dma);
pr_debug("%s aeq va %p dma %llx\n", __func__, q2_pages,
(unsigned long long) c2dev->aeq.host_dma);
c2_mq_rep_init(&c2dev->aeq,
2,
qsize,
msgsize,
q2_pages,
mmio_regs +
be32_to_cpu((__force __be32) readl(mmio_regs + C2_REGS_Q2_SHARED)),
C2_MQ_HOST_TARGET);
/* Initialize the verbs request allocator */
err = vq_init(c2dev);
if (err)
goto bail3;
/* Enable interrupts on the adapter */
writel(0, c2dev->regs + C2_IDIS);
/* create the WR init message */
err = c2_adapter_init(c2dev);
if (err)
goto bail4;
c2dev->init++;
/* open an adapter instance */
err = c2_rnic_open(c2dev);
if (err)
goto bail4;
/* Initialize cached the adapter limits */
if (c2_rnic_query(c2dev, &c2dev->props))
goto bail5;
/* Initialize the PD pool */
err = c2_init_pd_table(c2dev);
if (err)
goto bail5;
/* Initialize the QP pool */
c2_init_qp_table(c2dev);
return 0;
bail5:
c2_rnic_close(c2dev);
bail4:
vq_term(c2dev);
bail3:
dma_free_coherent(&c2dev->pcidev->dev,
c2dev->aeq.q_size * c2dev->aeq.msg_size,
q2_pages, pci_unmap_addr(&c2dev->aeq, mapping));
bail2:
dma_free_coherent(&c2dev->pcidev->dev,
c2dev->rep_vq.q_size * c2dev->rep_vq.msg_size,
q1_pages, pci_unmap_addr(&c2dev->rep_vq, mapping));
bail1:
c2_free_mqsp_pool(c2dev, c2dev->kern_mqsp_pool);
bail0:
vfree(c2dev->qptr_array);
return err;
}