#include "dma.h"
#include "dma_v2.h"
+extern struct kmem_cache *ioat3_sed_cache;
+
/* ioat hardware assumes at least two sources for raid operations */
#define src_cnt_to_sw(x) ((x) + 2)
#define src_cnt_to_hw(x) ((x) - 2)
static const u8 pq16_idx_to_field[] = { 1, 4, 1, 2, 3, 4, 5, 6, 7,
0, 1, 2, 3, 4, 5, 6 };
-/*
- * technically sources 1 and 2 do not require SED, but the op will have
- * at least 9 descriptors so that's irrelevant.
- */
-static const u8 pq16_idx_to_sed[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0,
- 1, 1, 1, 1, 1, 1, 1 };
-
static void ioat3_eh(struct ioat2_dma_chan *ioat);
-static dma_addr_t xor_get_src(struct ioat_raw_descriptor *descs[2], int idx)
-{
- struct ioat_raw_descriptor *raw = descs[xor_idx_to_desc >> idx & 1];
-
- return raw->field[xor_idx_to_field[idx]];
-}
-
static void xor_set_src(struct ioat_raw_descriptor *descs[2],
dma_addr_t addr, u32 offset, int idx)
{
pq->coef[idx] = coef;
}
-static int sed_get_pq16_pool_idx(int src_cnt)
-{
-
- return pq16_idx_to_sed[src_cnt];
-}
-
static bool is_jf_ioat(struct pci_dev *pdev)
{
switch (pdev->device) {
struct ioat_sed_ent *sed;
gfp_t flags = __GFP_ZERO | GFP_ATOMIC;
- sed = kmem_cache_alloc(device->sed_pool, flags);
+ sed = kmem_cache_alloc(ioat3_sed_cache, flags);
if (!sed)
return NULL;
sed->hw = dma_pool_alloc(device->sed_hw_pool[hw_pool],
flags, &sed->dma);
if (!sed->hw) {
- kmem_cache_free(device->sed_pool, sed);
+ kmem_cache_free(ioat3_sed_cache, sed);
return NULL;
}
return;
dma_pool_free(device->sed_hw_pool[sed->hw_pool], sed->hw, sed->dma);
- kmem_cache_free(device->sed_pool, sed);
-}
-
-static void ioat3_dma_unmap(struct ioat2_dma_chan *ioat,
- struct ioat_ring_ent *desc, int idx)
-{
- struct ioat_chan_common *chan = &ioat->base;
- struct pci_dev *pdev = chan->device->pdev;
- size_t len = desc->len;
- size_t offset = len - desc->hw->size;
- struct dma_async_tx_descriptor *tx = &desc->txd;
- enum dma_ctrl_flags flags = tx->flags;
-
- switch (desc->hw->ctl_f.op) {
- case IOAT_OP_COPY:
- if (!desc->hw->ctl_f.null) /* skip 'interrupt' ops */
- ioat_dma_unmap(chan, flags, len, desc->hw);
- break;
- case IOAT_OP_XOR_VAL:
- case IOAT_OP_XOR: {
- struct ioat_xor_descriptor *xor = desc->xor;
- struct ioat_ring_ent *ext;
- struct ioat_xor_ext_descriptor *xor_ex = NULL;
- int src_cnt = src_cnt_to_sw(xor->ctl_f.src_cnt);
- struct ioat_raw_descriptor *descs[2];
- int i;
-
- if (src_cnt > 5) {
- ext = ioat2_get_ring_ent(ioat, idx + 1);
- xor_ex = ext->xor_ex;
- }
-
- if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
- descs[0] = (struct ioat_raw_descriptor *) xor;
- descs[1] = (struct ioat_raw_descriptor *) xor_ex;
- for (i = 0; i < src_cnt; i++) {
- dma_addr_t src = xor_get_src(descs, i);
-
- ioat_unmap(pdev, src - offset, len,
- PCI_DMA_TODEVICE, flags, 0);
- }
-
- /* dest is a source in xor validate operations */
- if (xor->ctl_f.op == IOAT_OP_XOR_VAL) {
- ioat_unmap(pdev, xor->dst_addr - offset, len,
- PCI_DMA_TODEVICE, flags, 1);
- break;
- }
- }
-
- if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP))
- ioat_unmap(pdev, xor->dst_addr - offset, len,
- PCI_DMA_FROMDEVICE, flags, 1);
- break;
- }
- case IOAT_OP_PQ_VAL:
- case IOAT_OP_PQ: {
- struct ioat_pq_descriptor *pq = desc->pq;
- struct ioat_ring_ent *ext;
- struct ioat_pq_ext_descriptor *pq_ex = NULL;
- int src_cnt = src_cnt_to_sw(pq->ctl_f.src_cnt);
- struct ioat_raw_descriptor *descs[2];
- int i;
-
- if (src_cnt > 3) {
- ext = ioat2_get_ring_ent(ioat, idx + 1);
- pq_ex = ext->pq_ex;
- }
-
- /* in the 'continue' case don't unmap the dests as sources */
- if (dmaf_p_disabled_continue(flags))
- src_cnt--;
- else if (dmaf_continue(flags))
- src_cnt -= 3;
-
- if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
- descs[0] = (struct ioat_raw_descriptor *) pq;
- descs[1] = (struct ioat_raw_descriptor *) pq_ex;
- for (i = 0; i < src_cnt; i++) {
- dma_addr_t src = pq_get_src(descs, i);
-
- ioat_unmap(pdev, src - offset, len,
- PCI_DMA_TODEVICE, flags, 0);
- }
-
- /* the dests are sources in pq validate operations */
- if (pq->ctl_f.op == IOAT_OP_XOR_VAL) {
- if (!(flags & DMA_PREP_PQ_DISABLE_P))
- ioat_unmap(pdev, pq->p_addr - offset,
- len, PCI_DMA_TODEVICE, flags, 0);
- if (!(flags & DMA_PREP_PQ_DISABLE_Q))
- ioat_unmap(pdev, pq->q_addr - offset,
- len, PCI_DMA_TODEVICE, flags, 0);
- break;
- }
- }
-
- if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
- if (!(flags & DMA_PREP_PQ_DISABLE_P))
- ioat_unmap(pdev, pq->p_addr - offset, len,
- PCI_DMA_BIDIRECTIONAL, flags, 1);
- if (!(flags & DMA_PREP_PQ_DISABLE_Q))
- ioat_unmap(pdev, pq->q_addr - offset, len,
- PCI_DMA_BIDIRECTIONAL, flags, 1);
- }
- break;
- }
- case IOAT_OP_PQ_16S:
- case IOAT_OP_PQ_VAL_16S: {
- struct ioat_pq_descriptor *pq = desc->pq;
- int src_cnt = src16_cnt_to_sw(pq->ctl_f.src_cnt);
- struct ioat_raw_descriptor *descs[4];
- int i;
-
- /* in the 'continue' case don't unmap the dests as sources */
- if (dmaf_p_disabled_continue(flags))
- src_cnt--;
- else if (dmaf_continue(flags))
- src_cnt -= 3;
-
- if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
- descs[0] = (struct ioat_raw_descriptor *)pq;
- descs[1] = (struct ioat_raw_descriptor *)(desc->sed->hw);
- descs[2] = (struct ioat_raw_descriptor *)(&desc->sed->hw->b[0]);
- for (i = 0; i < src_cnt; i++) {
- dma_addr_t src = pq16_get_src(descs, i);
-
- ioat_unmap(pdev, src - offset, len,
- PCI_DMA_TODEVICE, flags, 0);
- }
-
- /* the dests are sources in pq validate operations */
- if (pq->ctl_f.op == IOAT_OP_XOR_VAL) {
- if (!(flags & DMA_PREP_PQ_DISABLE_P))
- ioat_unmap(pdev, pq->p_addr - offset,
- len, PCI_DMA_TODEVICE,
- flags, 0);
- if (!(flags & DMA_PREP_PQ_DISABLE_Q))
- ioat_unmap(pdev, pq->q_addr - offset,
- len, PCI_DMA_TODEVICE,
- flags, 0);
- break;
- }
- }
-
- if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
- if (!(flags & DMA_PREP_PQ_DISABLE_P))
- ioat_unmap(pdev, pq->p_addr - offset, len,
- PCI_DMA_BIDIRECTIONAL, flags, 1);
- if (!(flags & DMA_PREP_PQ_DISABLE_Q))
- ioat_unmap(pdev, pq->q_addr - offset, len,
- PCI_DMA_BIDIRECTIONAL, flags, 1);
- }
- break;
- }
- default:
- dev_err(&pdev->dev, "%s: unknown op type: %#x\n",
- __func__, desc->hw->ctl_f.op);
- }
+ kmem_cache_free(ioat3_sed_cache, sed);
}
static bool desc_has_ext(struct ioat_ring_ent *desc)
tx = &desc->txd;
if (tx->cookie) {
dma_cookie_complete(tx);
- ioat3_dma_unmap(ioat, desc, idx + i);
+ dma_descriptor_unmap(tx);
if (tx->callback) {
tx->callback(tx->callback_param);
tx->callback = NULL;
enum dma_status ret;
ret = dma_cookie_status(c, cookie, txstate);
- if (ret == DMA_SUCCESS)
+ if (ret == DMA_COMPLETE)
return ret;
ioat3_cleanup(ioat);
u8 op;
int i, s, idx, num_descs;
- /* this function only handles src_cnt 9 - 16 */
- BUG_ON(src_cnt < 9);
-
/* this function is only called with 9-16 sources */
op = result ? IOAT_OP_PQ_VAL_16S : IOAT_OP_PQ_16S;
descs[0] = (struct ioat_raw_descriptor *) pq;
- desc->sed = ioat3_alloc_sed(device,
- sed_get_pq16_pool_idx(src_cnt));
+ desc->sed = ioat3_alloc_sed(device, (src_cnt-2) >> 3);
if (!desc->sed) {
dev_err(to_dev(chan),
"%s: no free sed entries\n", __func__);
return &desc->txd;
}
+static int src_cnt_flags(unsigned int src_cnt, unsigned long flags)
+{
+ if (dmaf_p_disabled_continue(flags))
+ return src_cnt + 1;
+ else if (dmaf_continue(flags))
+ return src_cnt + 3;
+ else
+ return src_cnt;
+}
+
static struct dma_async_tx_descriptor *
ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src,
unsigned int src_cnt, const unsigned char *scf, size_t len,
unsigned long flags)
{
- struct dma_device *dma = chan->device;
-
/* specify valid address for disabled result */
if (flags & DMA_PREP_PQ_DISABLE_P)
dst[0] = dst[1];
single_source_coef[0] = scf[0];
single_source_coef[1] = 0;
- return (src_cnt > 8) && (dma->max_pq > 8) ?
+ return src_cnt_flags(src_cnt, flags) > 8 ?
__ioat3_prep_pq16_lock(chan, NULL, dst, single_source,
2, single_source_coef, len,
flags) :
single_source_coef, len, flags);
} else {
- return (src_cnt > 8) && (dma->max_pq > 8) ?
+ return src_cnt_flags(src_cnt, flags) > 8 ?
__ioat3_prep_pq16_lock(chan, NULL, dst, src, src_cnt,
scf, len, flags) :
__ioat3_prep_pq_lock(chan, NULL, dst, src, src_cnt,
unsigned int src_cnt, const unsigned char *scf, size_t len,
enum sum_check_flags *pqres, unsigned long flags)
{
- struct dma_device *dma = chan->device;
-
/* specify valid address for disabled result */
if (flags & DMA_PREP_PQ_DISABLE_P)
pq[0] = pq[1];
*/
*pqres = 0;
- return (src_cnt > 8) && (dma->max_pq > 8) ?
+ return src_cnt_flags(src_cnt, flags) > 8 ?
__ioat3_prep_pq16_lock(chan, pqres, pq, src, src_cnt, scf, len,
flags) :
__ioat3_prep_pq_lock(chan, pqres, pq, src, src_cnt, scf, len,
ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src,
unsigned int src_cnt, size_t len, unsigned long flags)
{
- struct dma_device *dma = chan->device;
unsigned char scf[src_cnt];
dma_addr_t pq[2];
flags |= DMA_PREP_PQ_DISABLE_Q;
pq[1] = dst; /* specify valid address for disabled result */
- return (src_cnt > 8) && (dma->max_pq > 8) ?
+ return src_cnt_flags(src_cnt, flags) > 8 ?
__ioat3_prep_pq16_lock(chan, NULL, pq, src, src_cnt, scf, len,
flags) :
__ioat3_prep_pq_lock(chan, NULL, pq, src, src_cnt, scf, len,
unsigned int src_cnt, size_t len,
enum sum_check_flags *result, unsigned long flags)
{
- struct dma_device *dma = chan->device;
unsigned char scf[src_cnt];
dma_addr_t pq[2];
flags |= DMA_PREP_PQ_DISABLE_Q;
pq[1] = pq[0]; /* specify valid address for disabled result */
-
- return (src_cnt > 8) && (dma->max_pq > 8) ?
+ return src_cnt_flags(src_cnt, flags) > 8 ?
__ioat3_prep_pq16_lock(chan, result, pq, &src[1], src_cnt - 1,
scf, len, flags) :
__ioat3_prep_pq_lock(chan, result, pq, &src[1], src_cnt - 1,
DMA_TO_DEVICE);
tx = dma->device_prep_dma_xor(dma_chan, dest_dma, dma_srcs,
IOAT_NUM_SRC_TEST, PAGE_SIZE,
- DMA_PREP_INTERRUPT |
- DMA_COMPL_SKIP_SRC_UNMAP |
- DMA_COMPL_SKIP_DEST_UNMAP);
+ DMA_PREP_INTERRUPT);
if (!tx) {
dev_err(dev, "Self-test xor prep failed\n");
tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000));
- if (dma->device_tx_status(dma_chan, cookie, NULL) != DMA_SUCCESS) {
+ if (dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) {
dev_err(dev, "Self-test xor timed out\n");
err = -ENODEV;
goto dma_unmap;
DMA_TO_DEVICE);
tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs,
IOAT_NUM_SRC_TEST + 1, PAGE_SIZE,
- &xor_val_result, DMA_PREP_INTERRUPT |
- DMA_COMPL_SKIP_SRC_UNMAP |
- DMA_COMPL_SKIP_DEST_UNMAP);
+ &xor_val_result, DMA_PREP_INTERRUPT);
if (!tx) {
dev_err(dev, "Self-test zero prep failed\n");
err = -ENODEV;
tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000));
- if (dma->device_tx_status(dma_chan, cookie, NULL) != DMA_SUCCESS) {
+ if (dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) {
dev_err(dev, "Self-test validate timed out\n");
err = -ENODEV;
goto dma_unmap;
goto free_resources;
}
+ memset(page_address(dest), 0, PAGE_SIZE);
+
/* test for non-zero parity sum */
op = IOAT_OP_XOR_VAL;
DMA_TO_DEVICE);
tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs,
IOAT_NUM_SRC_TEST + 1, PAGE_SIZE,
- &xor_val_result, DMA_PREP_INTERRUPT |
- DMA_COMPL_SKIP_SRC_UNMAP |
- DMA_COMPL_SKIP_DEST_UNMAP);
+ &xor_val_result, DMA_PREP_INTERRUPT);
if (!tx) {
dev_err(dev, "Self-test 2nd zero prep failed\n");
err = -ENODEV;
tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000));
- if (dma->device_tx_status(dma_chan, cookie, NULL) != DMA_SUCCESS) {
+ if (dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) {
dev_err(dev, "Self-test 2nd validate timed out\n");
err = -ENODEV;
goto dma_unmap;
static int ioat3_irq_reinit(struct ioatdma_device *device)
{
- int msixcnt = device->common.chancnt;
struct pci_dev *pdev = device->pdev;
- int i;
- struct msix_entry *msix;
- struct ioat_chan_common *chan;
- int err = 0;
+ int irq = pdev->irq, i;
+
+ if (!is_bwd_ioat(pdev))
+ return 0;
switch (device->irq_mode) {
case IOAT_MSIX:
+ for (i = 0; i < device->common.chancnt; i++) {
+ struct msix_entry *msix = &device->msix_entries[i];
+ struct ioat_chan_common *chan;
- for (i = 0; i < msixcnt; i++) {
- msix = &device->msix_entries[i];
chan = ioat_chan_by_index(device, i);
devm_free_irq(&pdev->dev, msix->vector, chan);
}
pci_disable_msix(pdev);
break;
-
- case IOAT_MSIX_SINGLE:
- msix = &device->msix_entries[0];
- chan = ioat_chan_by_index(device, 0);
- devm_free_irq(&pdev->dev, msix->vector, chan);
- pci_disable_msix(pdev);
- break;
-
case IOAT_MSI:
- chan = ioat_chan_by_index(device, 0);
- devm_free_irq(&pdev->dev, pdev->irq, chan);
pci_disable_msi(pdev);
- break;
-
+ /* fall through */
case IOAT_INTX:
- chan = ioat_chan_by_index(device, 0);
- devm_free_irq(&pdev->dev, pdev->irq, chan);
+ devm_free_irq(&pdev->dev, irq, device);
break;
-
default:
return 0;
}
-
device->irq_mode = IOAT_NOIRQ;
- err = ioat_dma_setup_interrupts(device);
-
- return err;
+ return ioat_dma_setup_interrupts(device);
}
static int ioat3_reset_hw(struct ioat_chan_common *chan)
}
err = ioat2_reset_sync(chan, msecs_to_jiffies(200));
- if (err) {
- dev_err(&pdev->dev, "Failed to reset!\n");
- return err;
- }
-
- if (device->irq_mode != IOAT_NOIRQ && is_bwd_ioat(pdev))
+ if (!err)
err = ioat3_irq_reinit(device);
+ if (err)
+ dev_err(&pdev->dev, "Failed to reset: %d\n", err);
+
return err;
}
char pool_name[14];
int i;
- /* allocate sw descriptor pool for SED */
- device->sed_pool = kmem_cache_create("ioat_sed",
- sizeof(struct ioat_sed_ent), 0, 0, NULL);
- if (!device->sed_pool)
- return -ENOMEM;
-
for (i = 0; i < MAX_SED_POOLS; i++) {
snprintf(pool_name, 14, "ioat_hw%d_sed", i);
/* allocate SED DMA pool */
- device->sed_hw_pool[i] = dma_pool_create(pool_name,
+ device->sed_hw_pool[i] = dmam_pool_create(pool_name,
&pdev->dev,
SED_SIZE * (i + 1), 64, 0);
if (!device->sed_hw_pool[i])
- goto sed_pool_cleanup;
+ return -ENOMEM;
}
}
device->dca = ioat3_dca_init(pdev, device->reg_base);
return 0;
-
-sed_pool_cleanup:
- if (device->sed_pool) {
- int i;
- kmem_cache_destroy(device->sed_pool);
-
- for (i = 0; i < MAX_SED_POOLS; i++)
- if (device->sed_hw_pool[i])
- dma_pool_destroy(device->sed_hw_pool[i]);
- }
-
- return -ENOMEM;
-}
-
-void ioat3_dma_remove(struct ioatdma_device *device)
-{
- if (device->sed_pool) {
- int i;
- kmem_cache_destroy(device->sed_pool);
-
- for (i = 0; i < MAX_SED_POOLS; i++)
- if (device->sed_hw_pool[i])
- dma_pool_destroy(device->sed_hw_pool[i]);
- }
}
projects / linux-drm-fsl-dcu.git / blobdiff