From: Dave Jiang Date: Tue, 11 Aug 2015 15:48:43 +0000 (-0700) Subject: dmaengine: ioatdma: move all the init routines X-Git-Tag: firefly_0821_release~176^2~1143^2~57 X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=c0f28ce66ecfd9fa0ae662a2c7f3e68e537e77f4;p=firefly-linux-kernel-4.4.55.git dmaengine: ioatdma: move all the init routines Moving all the init routines to init.c and fixup anything broken during the move. Signed-off-by: Dave Jiang Acked-by: Dan Williams Signed-off-by: Vinod Koul --- diff --git a/drivers/dma/ioat/Makefile b/drivers/dma/ioat/Makefile index d2555f4881d7..f785f8f42f7d 100644 --- a/drivers/dma/ioat/Makefile +++ b/drivers/dma/ioat/Makefile @@ -1,2 +1,2 @@ obj-$(CONFIG_INTEL_IOATDMA) += ioatdma.o -ioatdma-y := pci.o dma.o dma_v3.o dca.o sysfs.o +ioatdma-y := init.o dma.o dma_v3.o dca.o sysfs.o diff --git a/drivers/dma/ioat/dma.c b/drivers/dma/ioat/dma.c index 1746f7b4c3b4..5d78cafdd3f2 100644 --- a/drivers/dma/ioat/dma.c +++ b/drivers/dma/ioat/dma.c @@ -37,30 +37,12 @@ #include "../dmaengine.h" -int ioat_pending_level = 4; -module_param(ioat_pending_level, int, 0644); -MODULE_PARM_DESC(ioat_pending_level, - "high-water mark for pushing ioat descriptors (default: 4)"); -int ioat_ring_alloc_order = 8; -module_param(ioat_ring_alloc_order, int, 0644); -MODULE_PARM_DESC(ioat_ring_alloc_order, - "ioat+: allocate 2^n descriptors per channel (default: 8 max: 16)"); -static int ioat_ring_max_alloc_order = IOAT_MAX_ORDER; -module_param(ioat_ring_max_alloc_order, int, 0644); -MODULE_PARM_DESC(ioat_ring_max_alloc_order, - "ioat+: upper limit for ring size (default: 16)"); -static char ioat_interrupt_style[32] = "msix"; -module_param_string(ioat_interrupt_style, ioat_interrupt_style, - sizeof(ioat_interrupt_style), 0644); -MODULE_PARM_DESC(ioat_interrupt_style, - "set ioat interrupt style: msix (default), msi, intx"); - /** * ioat_dma_do_interrupt - handler used for single vector interrupt mode * @irq: interrupt id * @data: interrupt data */ -static irqreturn_t ioat_dma_do_interrupt(int irq, void *data) +irqreturn_t ioat_dma_do_interrupt(int irq, void *data) { struct ioatdma_device *instance = data; struct ioatdma_chan *ioat_chan; @@ -94,7 +76,7 @@ static irqreturn_t ioat_dma_do_interrupt(int irq, void *data) * @irq: interrupt id * @data: interrupt data */ -static irqreturn_t ioat_dma_do_interrupt_msix(int irq, void *data) +irqreturn_t ioat_dma_do_interrupt_msix(int irq, void *data) { struct ioatdma_chan *ioat_chan = data; @@ -104,28 +86,6 @@ static irqreturn_t ioat_dma_do_interrupt_msix(int irq, void *data) return IRQ_HANDLED; } -/* common channel initialization */ -void -ioat_init_channel(struct ioatdma_device *ioat_dma, - struct ioatdma_chan *ioat_chan, int idx) -{ - struct dma_device *dma = &ioat_dma->dma_dev; - struct dma_chan *c = &ioat_chan->dma_chan; - unsigned long data = (unsigned long) c; - - ioat_chan->ioat_dma = ioat_dma; - ioat_chan->reg_base = ioat_dma->reg_base + (0x80 * (idx + 1)); - spin_lock_init(&ioat_chan->cleanup_lock); - ioat_chan->dma_chan.device = dma; - dma_cookie_init(&ioat_chan->dma_chan); - list_add_tail(&ioat_chan->dma_chan.device_node, &dma->channels); - ioat_dma->idx[idx] = ioat_chan; - init_timer(&ioat_chan->timer); - ioat_chan->timer.function = ioat_dma->timer_fn; - ioat_chan->timer.data = data; - tasklet_init(&ioat_chan->cleanup_task, ioat_dma->cleanup_fn, data); -} - void ioat_stop(struct ioatdma_chan *ioat_chan) { struct ioatdma_device *ioat_dma = ioat_chan->ioat_dma; @@ -214,299 +174,6 @@ ioat_dma_tx_status(struct dma_chan *c, dma_cookie_t cookie, return dma_cookie_status(c, cookie, txstate); } -/* - * Perform a IOAT transaction to verify the HW works. - */ -#define IOAT_TEST_SIZE 2000 - -static void ioat_dma_test_callback(void *dma_async_param) -{ - struct completion *cmp = dma_async_param; - - complete(cmp); -} - -/** - * ioat_dma_self_test - Perform a IOAT transaction to verify the HW works. - * @ioat_dma: dma device to be tested - */ -int ioat_dma_self_test(struct ioatdma_device *ioat_dma) -{ - int i; - u8 *src; - u8 *dest; - struct dma_device *dma = &ioat_dma->dma_dev; - struct device *dev = &ioat_dma->pdev->dev; - struct dma_chan *dma_chan; - struct dma_async_tx_descriptor *tx; - dma_addr_t dma_dest, dma_src; - dma_cookie_t cookie; - int err = 0; - struct completion cmp; - unsigned long tmo; - unsigned long flags; - - src = kzalloc(sizeof(u8) * IOAT_TEST_SIZE, GFP_KERNEL); - if (!src) - return -ENOMEM; - dest = kzalloc(sizeof(u8) * IOAT_TEST_SIZE, GFP_KERNEL); - if (!dest) { - kfree(src); - return -ENOMEM; - } - - /* Fill in src buffer */ - for (i = 0; i < IOAT_TEST_SIZE; i++) - src[i] = (u8)i; - - /* Start copy, using first DMA channel */ - dma_chan = container_of(dma->channels.next, struct dma_chan, - device_node); - if (dma->device_alloc_chan_resources(dma_chan) < 1) { - dev_err(dev, "selftest cannot allocate chan resource\n"); - err = -ENODEV; - goto out; - } - - dma_src = dma_map_single(dev, src, IOAT_TEST_SIZE, DMA_TO_DEVICE); - if (dma_mapping_error(dev, dma_src)) { - dev_err(dev, "mapping src buffer failed\n"); - goto free_resources; - } - dma_dest = dma_map_single(dev, dest, IOAT_TEST_SIZE, DMA_FROM_DEVICE); - if (dma_mapping_error(dev, dma_dest)) { - dev_err(dev, "mapping dest buffer failed\n"); - goto unmap_src; - } - flags = DMA_PREP_INTERRUPT; - tx = ioat_dma->dma_dev.device_prep_dma_memcpy(dma_chan, dma_dest, - dma_src, IOAT_TEST_SIZE, - flags); - if (!tx) { - dev_err(dev, "Self-test prep failed, disabling\n"); - err = -ENODEV; - goto unmap_dma; - } - - async_tx_ack(tx); - init_completion(&cmp); - tx->callback = ioat_dma_test_callback; - tx->callback_param = &cmp; - cookie = tx->tx_submit(tx); - if (cookie < 0) { - dev_err(dev, "Self-test setup failed, disabling\n"); - err = -ENODEV; - goto unmap_dma; - } - dma->device_issue_pending(dma_chan); - - tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); - - if (tmo == 0 || - dma->device_tx_status(dma_chan, cookie, NULL) - != DMA_COMPLETE) { - dev_err(dev, "Self-test copy timed out, disabling\n"); - err = -ENODEV; - goto unmap_dma; - } - if (memcmp(src, dest, IOAT_TEST_SIZE)) { - dev_err(dev, "Self-test copy failed compare, disabling\n"); - err = -ENODEV; - goto free_resources; - } - -unmap_dma: - dma_unmap_single(dev, dma_dest, IOAT_TEST_SIZE, DMA_FROM_DEVICE); -unmap_src: - dma_unmap_single(dev, dma_src, IOAT_TEST_SIZE, DMA_TO_DEVICE); -free_resources: - dma->device_free_chan_resources(dma_chan); -out: - kfree(src); - kfree(dest); - return err; -} - -/** - * ioat_dma_setup_interrupts - setup interrupt handler - * @ioat_dma: ioat dma device - */ -int ioat_dma_setup_interrupts(struct ioatdma_device *ioat_dma) -{ - struct ioatdma_chan *ioat_chan; - struct pci_dev *pdev = ioat_dma->pdev; - struct device *dev = &pdev->dev; - struct msix_entry *msix; - int i, j, msixcnt; - int err = -EINVAL; - u8 intrctrl = 0; - - if (!strcmp(ioat_interrupt_style, "msix")) - goto msix; - if (!strcmp(ioat_interrupt_style, "msi")) - goto msi; - if (!strcmp(ioat_interrupt_style, "intx")) - goto intx; - dev_err(dev, "invalid ioat_interrupt_style %s\n", ioat_interrupt_style); - goto err_no_irq; - -msix: - /* The number of MSI-X vectors should equal the number of channels */ - msixcnt = ioat_dma->dma_dev.chancnt; - for (i = 0; i < msixcnt; i++) - ioat_dma->msix_entries[i].entry = i; - - err = pci_enable_msix_exact(pdev, ioat_dma->msix_entries, msixcnt); - if (err) - goto msi; - - for (i = 0; i < msixcnt; i++) { - msix = &ioat_dma->msix_entries[i]; - ioat_chan = ioat_chan_by_index(ioat_dma, i); - err = devm_request_irq(dev, msix->vector, - ioat_dma_do_interrupt_msix, 0, - "ioat-msix", ioat_chan); - if (err) { - for (j = 0; j < i; j++) { - msix = &ioat_dma->msix_entries[j]; - ioat_chan = ioat_chan_by_index(ioat_dma, j); - devm_free_irq(dev, msix->vector, ioat_chan); - } - goto msi; - } - } - intrctrl |= IOAT_INTRCTRL_MSIX_VECTOR_CONTROL; - ioat_dma->irq_mode = IOAT_MSIX; - goto done; - -msi: - err = pci_enable_msi(pdev); - if (err) - goto intx; - - err = devm_request_irq(dev, pdev->irq, ioat_dma_do_interrupt, 0, - "ioat-msi", ioat_dma); - if (err) { - pci_disable_msi(pdev); - goto intx; - } - ioat_dma->irq_mode = IOAT_MSI; - goto done; - -intx: - err = devm_request_irq(dev, pdev->irq, ioat_dma_do_interrupt, - IRQF_SHARED, "ioat-intx", ioat_dma); - if (err) - goto err_no_irq; - - ioat_dma->irq_mode = IOAT_INTX; -done: - if (ioat_dma->intr_quirk) - ioat_dma->intr_quirk(ioat_dma); - intrctrl |= IOAT_INTRCTRL_MASTER_INT_EN; - writeb(intrctrl, ioat_dma->reg_base + IOAT_INTRCTRL_OFFSET); - return 0; - -err_no_irq: - /* Disable all interrupt generation */ - writeb(0, ioat_dma->reg_base + IOAT_INTRCTRL_OFFSET); - ioat_dma->irq_mode = IOAT_NOIRQ; - dev_err(dev, "no usable interrupts\n"); - return err; -} -EXPORT_SYMBOL(ioat_dma_setup_interrupts); - -static void ioat_disable_interrupts(struct ioatdma_device *ioat_dma) -{ - /* Disable all interrupt generation */ - writeb(0, ioat_dma->reg_base + IOAT_INTRCTRL_OFFSET); -} - -int ioat_probe(struct ioatdma_device *ioat_dma) -{ - int err = -ENODEV; - struct dma_device *dma = &ioat_dma->dma_dev; - struct pci_dev *pdev = ioat_dma->pdev; - struct device *dev = &pdev->dev; - - /* DMA coherent memory pool for DMA descriptor allocations */ - ioat_dma->dma_pool = pci_pool_create("dma_desc_pool", pdev, - sizeof(struct ioat_dma_descriptor), - 64, 0); - if (!ioat_dma->dma_pool) { - err = -ENOMEM; - goto err_dma_pool; - } - - ioat_dma->completion_pool = pci_pool_create("completion_pool", pdev, - sizeof(u64), - SMP_CACHE_BYTES, - SMP_CACHE_BYTES); - - if (!ioat_dma->completion_pool) { - err = -ENOMEM; - goto err_completion_pool; - } - - ioat_dma->enumerate_channels(ioat_dma); - - dma_cap_set(DMA_MEMCPY, dma->cap_mask); - dma->dev = &pdev->dev; - - if (!dma->chancnt) { - dev_err(dev, "channel enumeration error\n"); - goto err_setup_interrupts; - } - - err = ioat_dma_setup_interrupts(ioat_dma); - if (err) - goto err_setup_interrupts; - - err = ioat_dma->self_test(ioat_dma); - if (err) - goto err_self_test; - - return 0; - -err_self_test: - ioat_disable_interrupts(ioat_dma); -err_setup_interrupts: - pci_pool_destroy(ioat_dma->completion_pool); -err_completion_pool: - pci_pool_destroy(ioat_dma->dma_pool); -err_dma_pool: - return err; -} - -int ioat_register(struct ioatdma_device *ioat_dma) -{ - int err = dma_async_device_register(&ioat_dma->dma_dev); - - if (err) { - ioat_disable_interrupts(ioat_dma); - pci_pool_destroy(ioat_dma->completion_pool); - pci_pool_destroy(ioat_dma->dma_pool); - } - - return err; -} - -void ioat_dma_remove(struct ioatdma_device *ioat_dma) -{ - struct dma_device *dma = &ioat_dma->dma_dev; - - ioat_disable_interrupts(ioat_dma); - - ioat_kobject_del(ioat_dma); - - dma_async_device_unregister(dma); - - pci_pool_destroy(ioat_dma->dma_pool); - pci_pool_destroy(ioat_dma->completion_pool); - - INIT_LIST_HEAD(&dma->channels); -} - void __ioat_issue_pending(struct ioatdma_chan *ioat_chan) { ioat_chan->dmacount += ioat_ring_pending(ioat_chan); @@ -577,7 +244,7 @@ static void __ioat_start_null_desc(struct ioatdma_chan *ioat_chan) __ioat_issue_pending(ioat_chan); } -static void ioat_start_null_desc(struct ioatdma_chan *ioat_chan) +void ioat_start_null_desc(struct ioatdma_chan *ioat_chan) { spin_lock_bh(&ioat_chan->prep_lock); __ioat_start_null_desc(ioat_chan); @@ -645,49 +312,6 @@ int ioat_reset_sync(struct ioatdma_chan *ioat_chan, unsigned long tmo) return err; } -/** - * ioat_enumerate_channels - find and initialize the device's channels - * @ioat_dma: the ioat dma device to be enumerated - */ -int ioat_enumerate_channels(struct ioatdma_device *ioat_dma) -{ - struct ioatdma_chan *ioat_chan; - struct device *dev = &ioat_dma->pdev->dev; - struct dma_device *dma = &ioat_dma->dma_dev; - u8 xfercap_log; - int i; - - INIT_LIST_HEAD(&dma->channels); - dma->chancnt = readb(ioat_dma->reg_base + IOAT_CHANCNT_OFFSET); - dma->chancnt &= 0x1f; /* bits [4:0] valid */ - if (dma->chancnt > ARRAY_SIZE(ioat_dma->idx)) { - dev_warn(dev, "(%d) exceeds max supported channels (%zu)\n", - dma->chancnt, ARRAY_SIZE(ioat_dma->idx)); - dma->chancnt = ARRAY_SIZE(ioat_dma->idx); - } - xfercap_log = readb(ioat_dma->reg_base + IOAT_XFERCAP_OFFSET); - xfercap_log &= 0x1f; /* bits [4:0] valid */ - if (xfercap_log == 0) - return 0; - dev_dbg(dev, "%s: xfercap = %d\n", __func__, 1 << xfercap_log); - - for (i = 0; i < dma->chancnt; i++) { - ioat_chan = devm_kzalloc(dev, sizeof(*ioat_chan), GFP_KERNEL); - if (!ioat_chan) - break; - - ioat_init_channel(ioat_dma, ioat_chan, i); - ioat_chan->xfercap_log = xfercap_log; - spin_lock_init(&ioat_chan->prep_lock); - if (ioat_dma->reset_hw(ioat_chan)) { - i = 0; - break; - } - } - dma->chancnt = i; - return i; -} - static dma_cookie_t ioat_tx_submit_unlock(struct dma_async_tx_descriptor *tx) { struct dma_chan *c = tx->chan; @@ -741,8 +365,7 @@ ioat_alloc_ring_ent(struct dma_chan *chan, gfp_t flags) return desc; } -static void -ioat_free_ring_ent(struct ioat_ring_ent *desc, struct dma_chan *chan) +void ioat_free_ring_ent(struct ioat_ring_ent *desc, struct dma_chan *chan) { struct ioatdma_device *ioat_dma; @@ -751,7 +374,7 @@ ioat_free_ring_ent(struct ioat_ring_ent *desc, struct dma_chan *chan) kmem_cache_free(ioat_cache, desc); } -static struct ioat_ring_ent ** +struct ioat_ring_ent ** ioat_alloc_ring(struct dma_chan *c, int order, gfp_t flags) { struct ioat_ring_ent **ring; @@ -788,128 +411,6 @@ ioat_alloc_ring(struct dma_chan *c, int order, gfp_t flags) return ring; } -/** - * ioat_free_chan_resources - release all the descriptors - * @chan: the channel to be cleaned - */ -void ioat_free_chan_resources(struct dma_chan *c) -{ - struct ioatdma_chan *ioat_chan = to_ioat_chan(c); - struct ioatdma_device *ioat_dma = ioat_chan->ioat_dma; - struct ioat_ring_ent *desc; - const int total_descs = 1 << ioat_chan->alloc_order; - int descs; - int i; - - /* Before freeing channel resources first check - * if they have been previously allocated for this channel. - */ - if (!ioat_chan->ring) - return; - - ioat_stop(ioat_chan); - ioat_dma->reset_hw(ioat_chan); - - spin_lock_bh(&ioat_chan->cleanup_lock); - spin_lock_bh(&ioat_chan->prep_lock); - descs = ioat_ring_space(ioat_chan); - dev_dbg(to_dev(ioat_chan), "freeing %d idle descriptors\n", descs); - for (i = 0; i < descs; i++) { - desc = ioat_get_ring_ent(ioat_chan, ioat_chan->head + i); - ioat_free_ring_ent(desc, c); - } - - if (descs < total_descs) - dev_err(to_dev(ioat_chan), "Freeing %d in use descriptors!\n", - total_descs - descs); - - for (i = 0; i < total_descs - descs; i++) { - desc = ioat_get_ring_ent(ioat_chan, ioat_chan->tail + i); - dump_desc_dbg(ioat_chan, desc); - ioat_free_ring_ent(desc, c); - } - - kfree(ioat_chan->ring); - ioat_chan->ring = NULL; - ioat_chan->alloc_order = 0; - pci_pool_free(ioat_dma->completion_pool, ioat_chan->completion, - ioat_chan->completion_dma); - spin_unlock_bh(&ioat_chan->prep_lock); - spin_unlock_bh(&ioat_chan->cleanup_lock); - - ioat_chan->last_completion = 0; - ioat_chan->completion_dma = 0; - ioat_chan->dmacount = 0; -} - -/* ioat_alloc_chan_resources - allocate/initialize ioat descriptor ring - * @chan: channel to be initialized - */ -int ioat_alloc_chan_resources(struct dma_chan *c) -{ - struct ioatdma_chan *ioat_chan = to_ioat_chan(c); - struct ioat_ring_ent **ring; - u64 status; - int order; - int i = 0; - u32 chanerr; - - /* have we already been set up? */ - if (ioat_chan->ring) - return 1 << ioat_chan->alloc_order; - - /* Setup register to interrupt and write completion status on error */ - writew(IOAT_CHANCTRL_RUN, ioat_chan->reg_base + IOAT_CHANCTRL_OFFSET); - - /* allocate a completion writeback area */ - /* doing 2 32bit writes to mmio since 1 64b write doesn't work */ - ioat_chan->completion = - pci_pool_alloc(ioat_chan->ioat_dma->completion_pool, - GFP_KERNEL, &ioat_chan->completion_dma); - if (!ioat_chan->completion) - return -ENOMEM; - - memset(ioat_chan->completion, 0, sizeof(*ioat_chan->completion)); - writel(((u64)ioat_chan->completion_dma) & 0x00000000FFFFFFFF, - ioat_chan->reg_base + IOAT_CHANCMP_OFFSET_LOW); - writel(((u64)ioat_chan->completion_dma) >> 32, - ioat_chan->reg_base + IOAT_CHANCMP_OFFSET_HIGH); - - order = ioat_get_alloc_order(); - ring = ioat_alloc_ring(c, order, GFP_KERNEL); - if (!ring) - return -ENOMEM; - - spin_lock_bh(&ioat_chan->cleanup_lock); - spin_lock_bh(&ioat_chan->prep_lock); - ioat_chan->ring = ring; - ioat_chan->head = 0; - ioat_chan->issued = 0; - ioat_chan->tail = 0; - ioat_chan->alloc_order = order; - set_bit(IOAT_RUN, &ioat_chan->state); - spin_unlock_bh(&ioat_chan->prep_lock); - spin_unlock_bh(&ioat_chan->cleanup_lock); - - ioat_start_null_desc(ioat_chan); - - /* check that we got off the ground */ - do { - udelay(1); - status = ioat_chansts(ioat_chan); - } while (i++ < 20 && !is_ioat_active(status) && !is_ioat_idle(status)); - - if (is_ioat_active(status) || is_ioat_idle(status)) - return 1 << ioat_chan->alloc_order; - - chanerr = readl(ioat_chan->reg_base + IOAT_CHANERR_OFFSET); - - dev_WARN(to_dev(ioat_chan), - "failed to start channel chanerr: %#x\n", chanerr); - ioat_free_chan_resources(c); - return -EFAULT; -} - bool reshape_ring(struct ioatdma_chan *ioat_chan, int order) { /* reshape differs from normal ring allocation in that we want diff --git a/drivers/dma/ioat/dma.h b/drivers/dma/ioat/dma.h index 2566ec6ae8a4..d2ffa5775d53 100644 --- a/drivers/dma/ioat/dma.h +++ b/drivers/dma/ioat/dma.h @@ -381,6 +381,43 @@ ioat_set_chainaddr(struct ioatdma_chan *ioat_chan, u64 addr) ioat_chan->reg_base + IOAT2_CHAINADDR_OFFSET_HIGH); } +irqreturn_t ioat_dma_do_interrupt(int irq, void *data); +irqreturn_t ioat_dma_do_interrupt_msix(int irq, void *data); +struct ioat_ring_ent ** +ioat_alloc_ring(struct dma_chan *c, int order, gfp_t flags); +void ioat_start_null_desc(struct ioatdma_chan *ioat_chan); +void ioat_free_ring_ent(struct ioat_ring_ent *desc, struct dma_chan *chan); +int ioat_reset_hw(struct ioatdma_chan *ioat_chan); +struct dma_async_tx_descriptor * +ioat_prep_interrupt_lock(struct dma_chan *c, unsigned long flags); +struct dma_async_tx_descriptor * +ioat_prep_xor(struct dma_chan *chan, dma_addr_t dest, dma_addr_t *src, + unsigned int src_cnt, size_t len, unsigned long flags); +struct dma_async_tx_descriptor * +ioat_prep_xor_val(struct dma_chan *chan, dma_addr_t *src, + unsigned int src_cnt, size_t len, + enum sum_check_flags *result, unsigned long flags); +struct dma_async_tx_descriptor * +ioat_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_async_tx_descriptor * +ioat_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, + unsigned int src_cnt, const unsigned char *scf, size_t len, + enum sum_check_flags *pqres, unsigned long flags); +struct dma_async_tx_descriptor * +ioat_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_async_tx_descriptor * +ioat_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, + unsigned int src_cnt, size_t len, + enum sum_check_flags *result, unsigned long flags); +enum dma_status +ioat_tx_status(struct dma_chan *c, dma_cookie_t cookie, + struct dma_tx_state *txstate); +void ioat_cleanup_event(unsigned long data); +void ioat_timer_event(unsigned long data); +bool is_bwd_ioat(struct pci_dev *pdev); int ioat_probe(struct ioatdma_device *ioat_dma); int ioat_register(struct ioatdma_device *ioat_dma); int ioat_dma_self_test(struct ioatdma_device *ioat_dma); @@ -421,5 +458,7 @@ extern int ioat_pending_level; extern int ioat_ring_alloc_order; extern struct kobj_type ioat_ktype; extern struct kmem_cache *ioat_cache; +extern int ioat_ring_max_alloc_order; +extern struct kmem_cache *ioat_sed_cache; #endif /* IOATDMA_H */ diff --git a/drivers/dma/ioat/dma_v3.c b/drivers/dma/ioat/dma_v3.c index 2a47f3968ce4..f6a194a3a463 100644 --- a/drivers/dma/ioat/dma_v3.c +++ b/drivers/dma/ioat/dma_v3.c @@ -62,8 +62,6 @@ #include "hw.h" #include "dma.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) @@ -118,124 +116,6 @@ static void pq_set_src(struct ioat_raw_descriptor *descs[2], pq->coef[idx] = coef; } -static bool is_jf_ioat(struct pci_dev *pdev) -{ - switch (pdev->device) { - case PCI_DEVICE_ID_INTEL_IOAT_JSF0: - case PCI_DEVICE_ID_INTEL_IOAT_JSF1: - case PCI_DEVICE_ID_INTEL_IOAT_JSF2: - case PCI_DEVICE_ID_INTEL_IOAT_JSF3: - case PCI_DEVICE_ID_INTEL_IOAT_JSF4: - case PCI_DEVICE_ID_INTEL_IOAT_JSF5: - case PCI_DEVICE_ID_INTEL_IOAT_JSF6: - case PCI_DEVICE_ID_INTEL_IOAT_JSF7: - case PCI_DEVICE_ID_INTEL_IOAT_JSF8: - case PCI_DEVICE_ID_INTEL_IOAT_JSF9: - return true; - default: - return false; - } -} - -static bool is_snb_ioat(struct pci_dev *pdev) -{ - switch (pdev->device) { - case PCI_DEVICE_ID_INTEL_IOAT_SNB0: - case PCI_DEVICE_ID_INTEL_IOAT_SNB1: - case PCI_DEVICE_ID_INTEL_IOAT_SNB2: - case PCI_DEVICE_ID_INTEL_IOAT_SNB3: - case PCI_DEVICE_ID_INTEL_IOAT_SNB4: - case PCI_DEVICE_ID_INTEL_IOAT_SNB5: - case PCI_DEVICE_ID_INTEL_IOAT_SNB6: - case PCI_DEVICE_ID_INTEL_IOAT_SNB7: - case PCI_DEVICE_ID_INTEL_IOAT_SNB8: - case PCI_DEVICE_ID_INTEL_IOAT_SNB9: - return true; - default: - return false; - } -} - -static bool is_ivb_ioat(struct pci_dev *pdev) -{ - switch (pdev->device) { - case PCI_DEVICE_ID_INTEL_IOAT_IVB0: - case PCI_DEVICE_ID_INTEL_IOAT_IVB1: - case PCI_DEVICE_ID_INTEL_IOAT_IVB2: - case PCI_DEVICE_ID_INTEL_IOAT_IVB3: - case PCI_DEVICE_ID_INTEL_IOAT_IVB4: - case PCI_DEVICE_ID_INTEL_IOAT_IVB5: - case PCI_DEVICE_ID_INTEL_IOAT_IVB6: - case PCI_DEVICE_ID_INTEL_IOAT_IVB7: - case PCI_DEVICE_ID_INTEL_IOAT_IVB8: - case PCI_DEVICE_ID_INTEL_IOAT_IVB9: - return true; - default: - return false; - } - -} - -static bool is_hsw_ioat(struct pci_dev *pdev) -{ - switch (pdev->device) { - case PCI_DEVICE_ID_INTEL_IOAT_HSW0: - case PCI_DEVICE_ID_INTEL_IOAT_HSW1: - case PCI_DEVICE_ID_INTEL_IOAT_HSW2: - case PCI_DEVICE_ID_INTEL_IOAT_HSW3: - case PCI_DEVICE_ID_INTEL_IOAT_HSW4: - case PCI_DEVICE_ID_INTEL_IOAT_HSW5: - case PCI_DEVICE_ID_INTEL_IOAT_HSW6: - case PCI_DEVICE_ID_INTEL_IOAT_HSW7: - case PCI_DEVICE_ID_INTEL_IOAT_HSW8: - case PCI_DEVICE_ID_INTEL_IOAT_HSW9: - return true; - default: - return false; - } - -} - -static bool is_xeon_cb32(struct pci_dev *pdev) -{ - return is_jf_ioat(pdev) || is_snb_ioat(pdev) || is_ivb_ioat(pdev) || - is_hsw_ioat(pdev); -} - -static bool is_bwd_ioat(struct pci_dev *pdev) -{ - switch (pdev->device) { - case PCI_DEVICE_ID_INTEL_IOAT_BWD0: - case PCI_DEVICE_ID_INTEL_IOAT_BWD1: - case PCI_DEVICE_ID_INTEL_IOAT_BWD2: - case PCI_DEVICE_ID_INTEL_IOAT_BWD3: - /* even though not Atom, BDX-DE has same DMA silicon */ - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE0: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE1: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE2: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE3: - return true; - default: - return false; - } -} - -static bool is_bwd_noraid(struct pci_dev *pdev) -{ - switch (pdev->device) { - case PCI_DEVICE_ID_INTEL_IOAT_BWD2: - case PCI_DEVICE_ID_INTEL_IOAT_BWD3: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE0: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE1: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE2: - case PCI_DEVICE_ID_INTEL_IOAT_BDXDE3: - return true; - default: - return false; - } - -} - static void pq16_set_src(struct ioat_raw_descriptor *desc[3], dma_addr_t addr, u32 offset, u8 coef, unsigned idx) { @@ -258,7 +138,7 @@ ioat3_alloc_sed(struct ioatdma_device *ioat_dma, unsigned int hw_pool) struct ioat_sed_ent *sed; gfp_t flags = __GFP_ZERO | GFP_ATOMIC; - sed = kmem_cache_alloc(ioat3_sed_cache, flags); + sed = kmem_cache_alloc(ioat_sed_cache, flags); if (!sed) return NULL; @@ -266,7 +146,7 @@ ioat3_alloc_sed(struct ioatdma_device *ioat_dma, unsigned int hw_pool) sed->hw = dma_pool_alloc(ioat_dma->sed_hw_pool[hw_pool], flags, &sed->dma); if (!sed->hw) { - kmem_cache_free(ioat3_sed_cache, sed); + kmem_cache_free(ioat_sed_cache, sed); return NULL; } @@ -280,7 +160,7 @@ ioat3_free_sed(struct ioatdma_device *ioat_dma, struct ioat_sed_ent *sed) return; dma_pool_free(ioat_dma->sed_hw_pool[sed->hw_pool], sed->hw, sed->dma); - kmem_cache_free(ioat3_sed_cache, sed); + kmem_cache_free(ioat_sed_cache, sed); } static bool desc_has_ext(struct ioat_ring_ent *desc) @@ -464,7 +344,7 @@ static void ioat3_cleanup(struct ioatdma_chan *ioat_chan) spin_unlock_bh(&ioat_chan->cleanup_lock); } -static void ioat3_cleanup_event(unsigned long data) +void ioat_cleanup_event(unsigned long data) { struct ioatdma_chan *ioat_chan = to_ioat_chan((void *)data); @@ -582,7 +462,7 @@ static void check_active(struct ioatdma_chan *ioat_chan) } -static void ioat3_timer_event(unsigned long data) +void ioat_timer_event(unsigned long data) { struct ioatdma_chan *ioat_chan = to_ioat_chan((void *)data); dma_addr_t phys_complete; @@ -634,8 +514,8 @@ static void ioat3_timer_event(unsigned long data) spin_unlock_bh(&ioat_chan->cleanup_lock); } -static enum dma_status -ioat3_tx_status(struct dma_chan *c, dma_cookie_t cookie, +enum dma_status +ioat_tx_status(struct dma_chan *c, dma_cookie_t cookie, struct dma_tx_state *txstate) { struct ioatdma_chan *ioat_chan = to_ioat_chan(c); @@ -651,7 +531,7 @@ ioat3_tx_status(struct dma_chan *c, dma_cookie_t cookie, } static struct dma_async_tx_descriptor * -__ioat3_prep_xor_lock(struct dma_chan *c, enum sum_check_flags *result, +__ioat_prep_xor_lock(struct dma_chan *c, enum sum_check_flags *result, dma_addr_t dest, dma_addr_t *src, unsigned int src_cnt, size_t len, unsigned long flags) { @@ -743,15 +623,15 @@ __ioat3_prep_xor_lock(struct dma_chan *c, enum sum_check_flags *result, return &compl_desc->txd; } -static struct dma_async_tx_descriptor * -ioat3_prep_xor(struct dma_chan *chan, dma_addr_t dest, dma_addr_t *src, +struct dma_async_tx_descriptor * +ioat_prep_xor(struct dma_chan *chan, dma_addr_t dest, dma_addr_t *src, unsigned int src_cnt, size_t len, unsigned long flags) { - return __ioat3_prep_xor_lock(chan, NULL, dest, src, src_cnt, len, flags); + return __ioat_prep_xor_lock(chan, NULL, dest, src, src_cnt, len, flags); } -static struct dma_async_tx_descriptor * -ioat3_prep_xor_val(struct dma_chan *chan, dma_addr_t *src, +struct dma_async_tx_descriptor * +ioat_prep_xor_val(struct dma_chan *chan, dma_addr_t *src, unsigned int src_cnt, size_t len, enum sum_check_flags *result, unsigned long flags) { @@ -760,7 +640,7 @@ ioat3_prep_xor_val(struct dma_chan *chan, dma_addr_t *src, */ *result = 0; - return __ioat3_prep_xor_lock(chan, result, src[0], &src[1], + return __ioat_prep_xor_lock(chan, result, src[0], &src[1], src_cnt - 1, len, flags); } @@ -828,7 +708,7 @@ static void dump_pq16_desc_dbg(struct ioatdma_chan *ioat_chan, } static struct dma_async_tx_descriptor * -__ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, +__ioat_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, const dma_addr_t *dst, const dma_addr_t *src, unsigned int src_cnt, const unsigned char *scf, size_t len, unsigned long flags) @@ -952,7 +832,7 @@ __ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, } static struct dma_async_tx_descriptor * -__ioat3_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result, +__ioat_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result, const dma_addr_t *dst, const dma_addr_t *src, unsigned int src_cnt, const unsigned char *scf, size_t len, unsigned long flags) @@ -1062,8 +942,8 @@ static int src_cnt_flags(unsigned int src_cnt, unsigned long flags) return src_cnt; } -static struct dma_async_tx_descriptor * -ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src, +struct dma_async_tx_descriptor * +ioat_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) { @@ -1087,23 +967,23 @@ ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src, single_source_coef[1] = 0; return src_cnt_flags(src_cnt, flags) > 8 ? - __ioat3_prep_pq16_lock(chan, NULL, dst, single_source, + __ioat_prep_pq16_lock(chan, NULL, dst, single_source, 2, single_source_coef, len, flags) : - __ioat3_prep_pq_lock(chan, NULL, dst, single_source, 2, + __ioat_prep_pq_lock(chan, NULL, dst, single_source, 2, single_source_coef, len, flags); } else { return src_cnt_flags(src_cnt, flags) > 8 ? - __ioat3_prep_pq16_lock(chan, NULL, dst, src, src_cnt, + __ioat_prep_pq16_lock(chan, NULL, dst, src, src_cnt, scf, len, flags) : - __ioat3_prep_pq_lock(chan, NULL, dst, src, src_cnt, + __ioat_prep_pq_lock(chan, NULL, dst, src, src_cnt, scf, len, flags); } } -static struct dma_async_tx_descriptor * -ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, +struct dma_async_tx_descriptor * +ioat_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, unsigned int src_cnt, const unsigned char *scf, size_t len, enum sum_check_flags *pqres, unsigned long flags) { @@ -1119,14 +999,14 @@ ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, *pqres = 0; return src_cnt_flags(src_cnt, flags) > 8 ? - __ioat3_prep_pq16_lock(chan, pqres, pq, src, src_cnt, scf, len, + __ioat_prep_pq16_lock(chan, pqres, pq, src, src_cnt, scf, len, flags) : - __ioat3_prep_pq_lock(chan, pqres, pq, src, src_cnt, scf, len, + __ioat_prep_pq_lock(chan, pqres, pq, src, src_cnt, scf, len, flags); } -static struct dma_async_tx_descriptor * -ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src, +struct dma_async_tx_descriptor * +ioat_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src, unsigned int src_cnt, size_t len, unsigned long flags) { unsigned char scf[src_cnt]; @@ -1138,14 +1018,14 @@ ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src, pq[1] = dst; /* specify valid address for disabled result */ return src_cnt_flags(src_cnt, flags) > 8 ? - __ioat3_prep_pq16_lock(chan, NULL, pq, src, src_cnt, scf, len, + __ioat_prep_pq16_lock(chan, NULL, pq, src, src_cnt, scf, len, flags) : - __ioat3_prep_pq_lock(chan, NULL, pq, src, src_cnt, scf, len, + __ioat_prep_pq_lock(chan, NULL, pq, src, src_cnt, scf, len, flags); } -static struct dma_async_tx_descriptor * -ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, +struct dma_async_tx_descriptor * +ioat_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, unsigned int src_cnt, size_t len, enum sum_check_flags *result, unsigned long flags) { @@ -1163,14 +1043,14 @@ ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, pq[1] = pq[0]; /* specify valid address for disabled result */ return src_cnt_flags(src_cnt, flags) > 8 ? - __ioat3_prep_pq16_lock(chan, result, pq, &src[1], src_cnt - 1, + __ioat_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, + __ioat_prep_pq_lock(chan, result, pq, &src[1], src_cnt - 1, scf, len, flags); } -static struct dma_async_tx_descriptor * -ioat3_prep_interrupt_lock(struct dma_chan *c, unsigned long flags) +struct dma_async_tx_descriptor * +ioat_prep_interrupt_lock(struct dma_chan *c, unsigned long flags) { struct ioatdma_chan *ioat_chan = to_ioat_chan(c); struct ioat_ring_ent *desc; @@ -1200,293 +1080,6 @@ ioat3_prep_interrupt_lock(struct dma_chan *c, unsigned long flags) return &desc->txd; } -static void ioat3_dma_test_callback(void *dma_async_param) -{ - struct completion *cmp = dma_async_param; - - complete(cmp); -} - -#define IOAT_NUM_SRC_TEST 6 /* must be <= 8 */ -static int ioat_xor_val_self_test(struct ioatdma_device *ioat_dma) -{ - int i, src_idx; - struct page *dest; - struct page *xor_srcs[IOAT_NUM_SRC_TEST]; - struct page *xor_val_srcs[IOAT_NUM_SRC_TEST + 1]; - dma_addr_t dma_srcs[IOAT_NUM_SRC_TEST + 1]; - dma_addr_t dest_dma; - struct dma_async_tx_descriptor *tx; - struct dma_chan *dma_chan; - dma_cookie_t cookie; - u8 cmp_byte = 0; - u32 cmp_word; - u32 xor_val_result; - int err = 0; - struct completion cmp; - unsigned long tmo; - struct device *dev = &ioat_dma->pdev->dev; - struct dma_device *dma = &ioat_dma->dma_dev; - u8 op = 0; - - dev_dbg(dev, "%s\n", __func__); - - if (!dma_has_cap(DMA_XOR, dma->cap_mask)) - return 0; - - for (src_idx = 0; src_idx < IOAT_NUM_SRC_TEST; src_idx++) { - xor_srcs[src_idx] = alloc_page(GFP_KERNEL); - if (!xor_srcs[src_idx]) { - while (src_idx--) - __free_page(xor_srcs[src_idx]); - return -ENOMEM; - } - } - - dest = alloc_page(GFP_KERNEL); - if (!dest) { - while (src_idx--) - __free_page(xor_srcs[src_idx]); - return -ENOMEM; - } - - /* Fill in src buffers */ - for (src_idx = 0; src_idx < IOAT_NUM_SRC_TEST; src_idx++) { - u8 *ptr = page_address(xor_srcs[src_idx]); - for (i = 0; i < PAGE_SIZE; i++) - ptr[i] = (1 << src_idx); - } - - for (src_idx = 0; src_idx < IOAT_NUM_SRC_TEST; src_idx++) - cmp_byte ^= (u8) (1 << src_idx); - - cmp_word = (cmp_byte << 24) | (cmp_byte << 16) | - (cmp_byte << 8) | cmp_byte; - - memset(page_address(dest), 0, PAGE_SIZE); - - dma_chan = container_of(dma->channels.next, struct dma_chan, - device_node); - if (dma->device_alloc_chan_resources(dma_chan) < 1) { - err = -ENODEV; - goto out; - } - - /* test xor */ - op = IOAT_OP_XOR; - - dest_dma = dma_map_page(dev, dest, 0, PAGE_SIZE, DMA_FROM_DEVICE); - if (dma_mapping_error(dev, dest_dma)) - goto dma_unmap; - - for (i = 0; i < IOAT_NUM_SRC_TEST; i++) - dma_srcs[i] = DMA_ERROR_CODE; - for (i = 0; i < IOAT_NUM_SRC_TEST; i++) { - dma_srcs[i] = dma_map_page(dev, xor_srcs[i], 0, PAGE_SIZE, - DMA_TO_DEVICE); - if (dma_mapping_error(dev, dma_srcs[i])) - goto dma_unmap; - } - tx = dma->device_prep_dma_xor(dma_chan, dest_dma, dma_srcs, - IOAT_NUM_SRC_TEST, PAGE_SIZE, - DMA_PREP_INTERRUPT); - - if (!tx) { - dev_err(dev, "Self-test xor prep failed\n"); - err = -ENODEV; - goto dma_unmap; - } - - async_tx_ack(tx); - init_completion(&cmp); - tx->callback = ioat3_dma_test_callback; - tx->callback_param = &cmp; - cookie = tx->tx_submit(tx); - if (cookie < 0) { - dev_err(dev, "Self-test xor setup failed\n"); - err = -ENODEV; - goto dma_unmap; - } - dma->device_issue_pending(dma_chan); - - tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); - - if (tmo == 0 || - dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) { - dev_err(dev, "Self-test xor timed out\n"); - err = -ENODEV; - goto dma_unmap; - } - - for (i = 0; i < IOAT_NUM_SRC_TEST; i++) - dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, DMA_TO_DEVICE); - - dma_sync_single_for_cpu(dev, dest_dma, PAGE_SIZE, DMA_FROM_DEVICE); - for (i = 0; i < (PAGE_SIZE / sizeof(u32)); i++) { - u32 *ptr = page_address(dest); - if (ptr[i] != cmp_word) { - dev_err(dev, "Self-test xor failed compare\n"); - err = -ENODEV; - goto free_resources; - } - } - dma_sync_single_for_device(dev, dest_dma, PAGE_SIZE, DMA_FROM_DEVICE); - - dma_unmap_page(dev, dest_dma, PAGE_SIZE, DMA_FROM_DEVICE); - - /* skip validate if the capability is not present */ - if (!dma_has_cap(DMA_XOR_VAL, dma_chan->device->cap_mask)) - goto free_resources; - - op = IOAT_OP_XOR_VAL; - - /* validate the sources with the destintation page */ - for (i = 0; i < IOAT_NUM_SRC_TEST; i++) - xor_val_srcs[i] = xor_srcs[i]; - xor_val_srcs[i] = dest; - - xor_val_result = 1; - - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) - dma_srcs[i] = DMA_ERROR_CODE; - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) { - dma_srcs[i] = dma_map_page(dev, xor_val_srcs[i], 0, PAGE_SIZE, - DMA_TO_DEVICE); - if (dma_mapping_error(dev, dma_srcs[i])) - goto dma_unmap; - } - tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs, - IOAT_NUM_SRC_TEST + 1, PAGE_SIZE, - &xor_val_result, DMA_PREP_INTERRUPT); - if (!tx) { - dev_err(dev, "Self-test zero prep failed\n"); - err = -ENODEV; - goto dma_unmap; - } - - async_tx_ack(tx); - init_completion(&cmp); - tx->callback = ioat3_dma_test_callback; - tx->callback_param = &cmp; - cookie = tx->tx_submit(tx); - if (cookie < 0) { - dev_err(dev, "Self-test zero setup failed\n"); - err = -ENODEV; - goto dma_unmap; - } - dma->device_issue_pending(dma_chan); - - tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); - - if (tmo == 0 || - dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) { - dev_err(dev, "Self-test validate timed out\n"); - err = -ENODEV; - goto dma_unmap; - } - - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) - dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, DMA_TO_DEVICE); - - if (xor_val_result != 0) { - dev_err(dev, "Self-test validate failed compare\n"); - err = -ENODEV; - goto free_resources; - } - - memset(page_address(dest), 0, PAGE_SIZE); - - /* test for non-zero parity sum */ - op = IOAT_OP_XOR_VAL; - - xor_val_result = 0; - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) - dma_srcs[i] = DMA_ERROR_CODE; - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) { - dma_srcs[i] = dma_map_page(dev, xor_val_srcs[i], 0, PAGE_SIZE, - DMA_TO_DEVICE); - if (dma_mapping_error(dev, dma_srcs[i])) - goto dma_unmap; - } - tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs, - IOAT_NUM_SRC_TEST + 1, PAGE_SIZE, - &xor_val_result, DMA_PREP_INTERRUPT); - if (!tx) { - dev_err(dev, "Self-test 2nd zero prep failed\n"); - err = -ENODEV; - goto dma_unmap; - } - - async_tx_ack(tx); - init_completion(&cmp); - tx->callback = ioat3_dma_test_callback; - tx->callback_param = &cmp; - cookie = tx->tx_submit(tx); - if (cookie < 0) { - dev_err(dev, "Self-test 2nd zero setup failed\n"); - err = -ENODEV; - goto dma_unmap; - } - dma->device_issue_pending(dma_chan); - - tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); - - if (tmo == 0 || - 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; - } - - if (xor_val_result != SUM_CHECK_P_RESULT) { - dev_err(dev, "Self-test validate failed compare\n"); - err = -ENODEV; - goto dma_unmap; - } - - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) - dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, DMA_TO_DEVICE); - - goto free_resources; -dma_unmap: - if (op == IOAT_OP_XOR) { - if (dest_dma != DMA_ERROR_CODE) - dma_unmap_page(dev, dest_dma, PAGE_SIZE, - DMA_FROM_DEVICE); - for (i = 0; i < IOAT_NUM_SRC_TEST; i++) - if (dma_srcs[i] != DMA_ERROR_CODE) - dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, - DMA_TO_DEVICE); - } else if (op == IOAT_OP_XOR_VAL) { - for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) - if (dma_srcs[i] != DMA_ERROR_CODE) - dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, - DMA_TO_DEVICE); - } -free_resources: - dma->device_free_chan_resources(dma_chan); -out: - src_idx = IOAT_NUM_SRC_TEST; - while (src_idx--) - __free_page(xor_srcs[src_idx]); - __free_page(dest); - return err; -} - -static int ioat3_dma_self_test(struct ioatdma_device *ioat_dma) -{ - int rc = ioat_dma_self_test(ioat_dma); - - if (rc) - return rc; - - rc = ioat_xor_val_self_test(ioat_dma); - if (rc) - return rc; - - return 0; -} - static int ioat3_irq_reinit(struct ioatdma_device *ioat_dma) { struct pci_dev *pdev = ioat_dma->pdev; @@ -1521,7 +1114,7 @@ static int ioat3_irq_reinit(struct ioatdma_device *ioat_dma) return ioat_dma_setup_interrupts(ioat_dma); } -static int ioat3_reset_hw(struct ioatdma_chan *ioat_chan) +int ioat_reset_hw(struct ioatdma_chan *ioat_chan) { /* throw away whatever the channel was doing and get it * initialized, with ioat3 specific workarounds @@ -1569,148 +1162,3 @@ static int ioat3_reset_hw(struct ioatdma_chan *ioat_chan) return err; } - -static void ioat3_intr_quirk(struct ioatdma_device *ioat_dma) -{ - struct dma_device *dma; - struct dma_chan *c; - struct ioatdma_chan *ioat_chan; - u32 errmask; - - dma = &ioat_dma->dma_dev; - - /* - * if we have descriptor write back error status, we mask the - * error interrupts - */ - if (ioat_dma->cap & IOAT_CAP_DWBES) { - list_for_each_entry(c, &dma->channels, device_node) { - ioat_chan = to_ioat_chan(c); - errmask = readl(ioat_chan->reg_base + - IOAT_CHANERR_MASK_OFFSET); - errmask |= IOAT_CHANERR_XOR_P_OR_CRC_ERR | - IOAT_CHANERR_XOR_Q_ERR; - writel(errmask, ioat_chan->reg_base + - IOAT_CHANERR_MASK_OFFSET); - } - } -} - -int ioat3_dma_probe(struct ioatdma_device *ioat_dma, int dca) -{ - struct pci_dev *pdev = ioat_dma->pdev; - int dca_en = system_has_dca_enabled(pdev); - struct dma_device *dma; - struct dma_chan *c; - struct ioatdma_chan *ioat_chan; - bool is_raid_device = false; - int err; - - ioat_dma->enumerate_channels = ioat_enumerate_channels; - ioat_dma->reset_hw = ioat3_reset_hw; - ioat_dma->self_test = ioat3_dma_self_test; - ioat_dma->intr_quirk = ioat3_intr_quirk; - dma = &ioat_dma->dma_dev; - dma->device_prep_dma_memcpy = ioat_dma_prep_memcpy_lock; - dma->device_issue_pending = ioat_issue_pending; - dma->device_alloc_chan_resources = ioat_alloc_chan_resources; - dma->device_free_chan_resources = ioat_free_chan_resources; - - dma_cap_set(DMA_INTERRUPT, dma->cap_mask); - dma->device_prep_dma_interrupt = ioat3_prep_interrupt_lock; - - ioat_dma->cap = readl(ioat_dma->reg_base + IOAT_DMA_CAP_OFFSET); - - if (is_xeon_cb32(pdev) || is_bwd_noraid(pdev)) - ioat_dma->cap &= - ~(IOAT_CAP_XOR | IOAT_CAP_PQ | IOAT_CAP_RAID16SS); - - /* dca is incompatible with raid operations */ - if (dca_en && (ioat_dma->cap & (IOAT_CAP_XOR|IOAT_CAP_PQ))) - ioat_dma->cap &= ~(IOAT_CAP_XOR|IOAT_CAP_PQ); - - if (ioat_dma->cap & IOAT_CAP_XOR) { - is_raid_device = true; - dma->max_xor = 8; - - dma_cap_set(DMA_XOR, dma->cap_mask); - dma->device_prep_dma_xor = ioat3_prep_xor; - - dma_cap_set(DMA_XOR_VAL, dma->cap_mask); - dma->device_prep_dma_xor_val = ioat3_prep_xor_val; - } - - if (ioat_dma->cap & IOAT_CAP_PQ) { - is_raid_device = true; - - dma->device_prep_dma_pq = ioat3_prep_pq; - dma->device_prep_dma_pq_val = ioat3_prep_pq_val; - dma_cap_set(DMA_PQ, dma->cap_mask); - dma_cap_set(DMA_PQ_VAL, dma->cap_mask); - - if (ioat_dma->cap & IOAT_CAP_RAID16SS) { - dma_set_maxpq(dma, 16, 0); - } else { - dma_set_maxpq(dma, 8, 0); - } - - if (!(ioat_dma->cap & IOAT_CAP_XOR)) { - dma->device_prep_dma_xor = ioat3_prep_pqxor; - dma->device_prep_dma_xor_val = ioat3_prep_pqxor_val; - dma_cap_set(DMA_XOR, dma->cap_mask); - dma_cap_set(DMA_XOR_VAL, dma->cap_mask); - - if (ioat_dma->cap & IOAT_CAP_RAID16SS) { - dma->max_xor = 16; - } else { - dma->max_xor = 8; - } - } - } - - dma->device_tx_status = ioat3_tx_status; - ioat_dma->cleanup_fn = ioat3_cleanup_event; - ioat_dma->timer_fn = ioat3_timer_event; - - /* starting with CB3.3 super extended descriptors are supported */ - if (ioat_dma->cap & IOAT_CAP_RAID16SS) { - char pool_name[14]; - int i; - - for (i = 0; i < MAX_SED_POOLS; i++) { - snprintf(pool_name, 14, "ioat_hw%d_sed", i); - - /* allocate SED DMA pool */ - ioat_dma->sed_hw_pool[i] = dmam_pool_create(pool_name, - &pdev->dev, - SED_SIZE * (i + 1), 64, 0); - if (!ioat_dma->sed_hw_pool[i]) - return -ENOMEM; - - } - } - - if (!(ioat_dma->cap & (IOAT_CAP_XOR | IOAT_CAP_PQ))) - dma_cap_set(DMA_PRIVATE, dma->cap_mask); - - err = ioat_probe(ioat_dma); - if (err) - return err; - - list_for_each_entry(c, &dma->channels, device_node) { - ioat_chan = to_ioat_chan(c); - writel(IOAT_DMA_DCA_ANY_CPU, - ioat_chan->reg_base + IOAT_DCACTRL_OFFSET); - } - - err = ioat_register(ioat_dma); - if (err) - return err; - - ioat_kobject_add(ioat_dma, &ioat_ktype); - - if (dca) - ioat_dma->dca = ioat3_dca_init(pdev, ioat_dma->reg_base); - - return 0; -} diff --git a/drivers/dma/ioat/init.c b/drivers/dma/ioat/init.c new file mode 100644 index 000000000000..de8141c7cd01 --- /dev/null +++ b/drivers/dma/ioat/init.c @@ -0,0 +1,1293 @@ +/* + * Intel I/OAT DMA Linux driver + * Copyright(c) 2004 - 2015 Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * The full GNU General Public License is included in this distribution in + * the file called "COPYING". + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "dma.h" +#include "registers.h" +#include "hw.h" + +#include "../dmaengine.h" + +MODULE_VERSION(IOAT_DMA_VERSION); +MODULE_LICENSE("Dual BSD/GPL"); +MODULE_AUTHOR("Intel Corporation"); + +static struct pci_device_id ioat_pci_tbl[] = { + /* I/OAT v3 platforms */ + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG3) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG4) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG5) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG6) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG7) }, + + /* I/OAT v3.2 platforms */ + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF3) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF4) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF5) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF6) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF7) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF8) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF9) }, + + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB3) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB4) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB5) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB6) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB7) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB8) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB9) }, + + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB3) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB4) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB5) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB6) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB7) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB8) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB9) }, + + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW3) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW4) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW5) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW6) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW7) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW8) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW9) }, + + /* I/OAT v3.3 platforms */ + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD3) }, + + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE0) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE1) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE2) }, + { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE3) }, + + { 0, } +}; +MODULE_DEVICE_TABLE(pci, ioat_pci_tbl); + +static int ioat_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id); +static void ioat_remove(struct pci_dev *pdev); + +static int ioat_dca_enabled = 1; +module_param(ioat_dca_enabled, int, 0644); +MODULE_PARM_DESC(ioat_dca_enabled, "control support of dca service (default: 1)"); +int ioat_pending_level = 4; +module_param(ioat_pending_level, int, 0644); +MODULE_PARM_DESC(ioat_pending_level, + "high-water mark for pushing ioat descriptors (default: 4)"); +int ioat_ring_alloc_order = 8; +module_param(ioat_ring_alloc_order, int, 0644); +MODULE_PARM_DESC(ioat_ring_alloc_order, + "ioat+: allocate 2^n descriptors per channel (default: 8 max: 16)"); +int ioat_ring_max_alloc_order = IOAT_MAX_ORDER; +module_param(ioat_ring_max_alloc_order, int, 0644); +MODULE_PARM_DESC(ioat_ring_max_alloc_order, + "ioat+: upper limit for ring size (default: 16)"); +static char ioat_interrupt_style[32] = "msix"; +module_param_string(ioat_interrupt_style, ioat_interrupt_style, + sizeof(ioat_interrupt_style), 0644); +MODULE_PARM_DESC(ioat_interrupt_style, + "set ioat interrupt style: msix (default), msi, intx"); + +struct kmem_cache *ioat_cache; +struct kmem_cache *ioat_sed_cache; + +static bool is_jf_ioat(struct pci_dev *pdev) +{ + switch (pdev->device) { + case PCI_DEVICE_ID_INTEL_IOAT_JSF0: + case PCI_DEVICE_ID_INTEL_IOAT_JSF1: + case PCI_DEVICE_ID_INTEL_IOAT_JSF2: + case PCI_DEVICE_ID_INTEL_IOAT_JSF3: + case PCI_DEVICE_ID_INTEL_IOAT_JSF4: + case PCI_DEVICE_ID_INTEL_IOAT_JSF5: + case PCI_DEVICE_ID_INTEL_IOAT_JSF6: + case PCI_DEVICE_ID_INTEL_IOAT_JSF7: + case PCI_DEVICE_ID_INTEL_IOAT_JSF8: + case PCI_DEVICE_ID_INTEL_IOAT_JSF9: + return true; + default: + return false; + } +} + +static bool is_snb_ioat(struct pci_dev *pdev) +{ + switch (pdev->device) { + case PCI_DEVICE_ID_INTEL_IOAT_SNB0: + case PCI_DEVICE_ID_INTEL_IOAT_SNB1: + case PCI_DEVICE_ID_INTEL_IOAT_SNB2: + case PCI_DEVICE_ID_INTEL_IOAT_SNB3: + case PCI_DEVICE_ID_INTEL_IOAT_SNB4: + case PCI_DEVICE_ID_INTEL_IOAT_SNB5: + case PCI_DEVICE_ID_INTEL_IOAT_SNB6: + case PCI_DEVICE_ID_INTEL_IOAT_SNB7: + case PCI_DEVICE_ID_INTEL_IOAT_SNB8: + case PCI_DEVICE_ID_INTEL_IOAT_SNB9: + return true; + default: + return false; + } +} + +static bool is_ivb_ioat(struct pci_dev *pdev) +{ + switch (pdev->device) { + case PCI_DEVICE_ID_INTEL_IOAT_IVB0: + case PCI_DEVICE_ID_INTEL_IOAT_IVB1: + case PCI_DEVICE_ID_INTEL_IOAT_IVB2: + case PCI_DEVICE_ID_INTEL_IOAT_IVB3: + case PCI_DEVICE_ID_INTEL_IOAT_IVB4: + case PCI_DEVICE_ID_INTEL_IOAT_IVB5: + case PCI_DEVICE_ID_INTEL_IOAT_IVB6: + case PCI_DEVICE_ID_INTEL_IOAT_IVB7: + case PCI_DEVICE_ID_INTEL_IOAT_IVB8: + case PCI_DEVICE_ID_INTEL_IOAT_IVB9: + return true; + default: + return false; + } + +} + +static bool is_hsw_ioat(struct pci_dev *pdev) +{ + switch (pdev->device) { + case PCI_DEVICE_ID_INTEL_IOAT_HSW0: + case PCI_DEVICE_ID_INTEL_IOAT_HSW1: + case PCI_DEVICE_ID_INTEL_IOAT_HSW2: + case PCI_DEVICE_ID_INTEL_IOAT_HSW3: + case PCI_DEVICE_ID_INTEL_IOAT_HSW4: + case PCI_DEVICE_ID_INTEL_IOAT_HSW5: + case PCI_DEVICE_ID_INTEL_IOAT_HSW6: + case PCI_DEVICE_ID_INTEL_IOAT_HSW7: + case PCI_DEVICE_ID_INTEL_IOAT_HSW8: + case PCI_DEVICE_ID_INTEL_IOAT_HSW9: + return true; + default: + return false; + } + +} + +static bool is_xeon_cb32(struct pci_dev *pdev) +{ + return is_jf_ioat(pdev) || is_snb_ioat(pdev) || is_ivb_ioat(pdev) || + is_hsw_ioat(pdev); +} + +bool is_bwd_ioat(struct pci_dev *pdev) +{ + switch (pdev->device) { + case PCI_DEVICE_ID_INTEL_IOAT_BWD0: + case PCI_DEVICE_ID_INTEL_IOAT_BWD1: + case PCI_DEVICE_ID_INTEL_IOAT_BWD2: + case PCI_DEVICE_ID_INTEL_IOAT_BWD3: + /* even though not Atom, BDX-DE has same DMA silicon */ + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE0: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE1: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE2: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE3: + return true; + default: + return false; + } +} + +static bool is_bwd_noraid(struct pci_dev *pdev) +{ + switch (pdev->device) { + case PCI_DEVICE_ID_INTEL_IOAT_BWD2: + case PCI_DEVICE_ID_INTEL_IOAT_BWD3: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE0: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE1: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE2: + case PCI_DEVICE_ID_INTEL_IOAT_BDXDE3: + return true; + default: + return false; + } + +} + +/* + * Perform a IOAT transaction to verify the HW works. + */ +#define IOAT_TEST_SIZE 2000 + +static void ioat_dma_test_callback(void *dma_async_param) +{ + struct completion *cmp = dma_async_param; + + complete(cmp); +} + +/** + * ioat_dma_self_test - Perform a IOAT transaction to verify the HW works. + * @ioat_dma: dma device to be tested + */ +int ioat_dma_self_test(struct ioatdma_device *ioat_dma) +{ + int i; + u8 *src; + u8 *dest; + struct dma_device *dma = &ioat_dma->dma_dev; + struct device *dev = &ioat_dma->pdev->dev; + struct dma_chan *dma_chan; + struct dma_async_tx_descriptor *tx; + dma_addr_t dma_dest, dma_src; + dma_cookie_t cookie; + int err = 0; + struct completion cmp; + unsigned long tmo; + unsigned long flags; + + src = kzalloc(sizeof(u8) * IOAT_TEST_SIZE, GFP_KERNEL); + if (!src) + return -ENOMEM; + dest = kzalloc(sizeof(u8) * IOAT_TEST_SIZE, GFP_KERNEL); + if (!dest) { + kfree(src); + return -ENOMEM; + } + + /* Fill in src buffer */ + for (i = 0; i < IOAT_TEST_SIZE; i++) + src[i] = (u8)i; + + /* Start copy, using first DMA channel */ + dma_chan = container_of(dma->channels.next, struct dma_chan, + device_node); + if (dma->device_alloc_chan_resources(dma_chan) < 1) { + dev_err(dev, "selftest cannot allocate chan resource\n"); + err = -ENODEV; + goto out; + } + + dma_src = dma_map_single(dev, src, IOAT_TEST_SIZE, DMA_TO_DEVICE); + if (dma_mapping_error(dev, dma_src)) { + dev_err(dev, "mapping src buffer failed\n"); + goto free_resources; + } + dma_dest = dma_map_single(dev, dest, IOAT_TEST_SIZE, DMA_FROM_DEVICE); + if (dma_mapping_error(dev, dma_dest)) { + dev_err(dev, "mapping dest buffer failed\n"); + goto unmap_src; + } + flags = DMA_PREP_INTERRUPT; + tx = ioat_dma->dma_dev.device_prep_dma_memcpy(dma_chan, dma_dest, + dma_src, IOAT_TEST_SIZE, + flags); + if (!tx) { + dev_err(dev, "Self-test prep failed, disabling\n"); + err = -ENODEV; + goto unmap_dma; + } + + async_tx_ack(tx); + init_completion(&cmp); + tx->callback = ioat_dma_test_callback; + tx->callback_param = &cmp; + cookie = tx->tx_submit(tx); + if (cookie < 0) { + dev_err(dev, "Self-test setup failed, disabling\n"); + err = -ENODEV; + goto unmap_dma; + } + dma->device_issue_pending(dma_chan); + + tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); + + if (tmo == 0 || + dma->device_tx_status(dma_chan, cookie, NULL) + != DMA_COMPLETE) { + dev_err(dev, "Self-test copy timed out, disabling\n"); + err = -ENODEV; + goto unmap_dma; + } + if (memcmp(src, dest, IOAT_TEST_SIZE)) { + dev_err(dev, "Self-test copy failed compare, disabling\n"); + err = -ENODEV; + goto free_resources; + } + +unmap_dma: + dma_unmap_single(dev, dma_dest, IOAT_TEST_SIZE, DMA_FROM_DEVICE); +unmap_src: + dma_unmap_single(dev, dma_src, IOAT_TEST_SIZE, DMA_TO_DEVICE); +free_resources: + dma->device_free_chan_resources(dma_chan); +out: + kfree(src); + kfree(dest); + return err; +} + +/** + * ioat_dma_setup_interrupts - setup interrupt handler + * @ioat_dma: ioat dma device + */ +int ioat_dma_setup_interrupts(struct ioatdma_device *ioat_dma) +{ + struct ioatdma_chan *ioat_chan; + struct pci_dev *pdev = ioat_dma->pdev; + struct device *dev = &pdev->dev; + struct msix_entry *msix; + int i, j, msixcnt; + int err = -EINVAL; + u8 intrctrl = 0; + + if (!strcmp(ioat_interrupt_style, "msix")) + goto msix; + if (!strcmp(ioat_interrupt_style, "msi")) + goto msi; + if (!strcmp(ioat_interrupt_style, "intx")) + goto intx; + dev_err(dev, "invalid ioat_interrupt_style %s\n", ioat_interrupt_style); + goto err_no_irq; + +msix: + /* The number of MSI-X vectors should equal the number of channels */ + msixcnt = ioat_dma->dma_dev.chancnt; + for (i = 0; i < msixcnt; i++) + ioat_dma->msix_entries[i].entry = i; + + err = pci_enable_msix_exact(pdev, ioat_dma->msix_entries, msixcnt); + if (err) + goto msi; + + for (i = 0; i < msixcnt; i++) { + msix = &ioat_dma->msix_entries[i]; + ioat_chan = ioat_chan_by_index(ioat_dma, i); + err = devm_request_irq(dev, msix->vector, + ioat_dma_do_interrupt_msix, 0, + "ioat-msix", ioat_chan); + if (err) { + for (j = 0; j < i; j++) { + msix = &ioat_dma->msix_entries[j]; + ioat_chan = ioat_chan_by_index(ioat_dma, j); + devm_free_irq(dev, msix->vector, ioat_chan); + } + goto msi; + } + } + intrctrl |= IOAT_INTRCTRL_MSIX_VECTOR_CONTROL; + ioat_dma->irq_mode = IOAT_MSIX; + goto done; + +msi: + err = pci_enable_msi(pdev); + if (err) + goto intx; + + err = devm_request_irq(dev, pdev->irq, ioat_dma_do_interrupt, 0, + "ioat-msi", ioat_dma); + if (err) { + pci_disable_msi(pdev); + goto intx; + } + ioat_dma->irq_mode = IOAT_MSI; + goto done; + +intx: + err = devm_request_irq(dev, pdev->irq, ioat_dma_do_interrupt, + IRQF_SHARED, "ioat-intx", ioat_dma); + if (err) + goto err_no_irq; + + ioat_dma->irq_mode = IOAT_INTX; +done: + if (ioat_dma->intr_quirk) + ioat_dma->intr_quirk(ioat_dma); + intrctrl |= IOAT_INTRCTRL_MASTER_INT_EN; + writeb(intrctrl, ioat_dma->reg_base + IOAT_INTRCTRL_OFFSET); + return 0; + +err_no_irq: + /* Disable all interrupt generation */ + writeb(0, ioat_dma->reg_base + IOAT_INTRCTRL_OFFSET); + ioat_dma->irq_mode = IOAT_NOIRQ; + dev_err(dev, "no usable interrupts\n"); + return err; +} +EXPORT_SYMBOL(ioat_dma_setup_interrupts); + +static void ioat_disable_interrupts(struct ioatdma_device *ioat_dma) +{ + /* Disable all interrupt generation */ + writeb(0, ioat_dma->reg_base + IOAT_INTRCTRL_OFFSET); +} + +int ioat_probe(struct ioatdma_device *ioat_dma) +{ + int err = -ENODEV; + struct dma_device *dma = &ioat_dma->dma_dev; + struct pci_dev *pdev = ioat_dma->pdev; + struct device *dev = &pdev->dev; + + /* DMA coherent memory pool for DMA descriptor allocations */ + ioat_dma->dma_pool = pci_pool_create("dma_desc_pool", pdev, + sizeof(struct ioat_dma_descriptor), + 64, 0); + if (!ioat_dma->dma_pool) { + err = -ENOMEM; + goto err_dma_pool; + } + + ioat_dma->completion_pool = pci_pool_create("completion_pool", pdev, + sizeof(u64), + SMP_CACHE_BYTES, + SMP_CACHE_BYTES); + + if (!ioat_dma->completion_pool) { + err = -ENOMEM; + goto err_completion_pool; + } + + ioat_dma->enumerate_channels(ioat_dma); + + dma_cap_set(DMA_MEMCPY, dma->cap_mask); + dma->dev = &pdev->dev; + + if (!dma->chancnt) { + dev_err(dev, "channel enumeration error\n"); + goto err_setup_interrupts; + } + + err = ioat_dma_setup_interrupts(ioat_dma); + if (err) + goto err_setup_interrupts; + + err = ioat_dma->self_test(ioat_dma); + if (err) + goto err_self_test; + + return 0; + +err_self_test: + ioat_disable_interrupts(ioat_dma); +err_setup_interrupts: + pci_pool_destroy(ioat_dma->completion_pool); +err_completion_pool: + pci_pool_destroy(ioat_dma->dma_pool); +err_dma_pool: + return err; +} + +int ioat_register(struct ioatdma_device *ioat_dma) +{ + int err = dma_async_device_register(&ioat_dma->dma_dev); + + if (err) { + ioat_disable_interrupts(ioat_dma); + pci_pool_destroy(ioat_dma->completion_pool); + pci_pool_destroy(ioat_dma->dma_pool); + } + + return err; +} + +void ioat_dma_remove(struct ioatdma_device *ioat_dma) +{ + struct dma_device *dma = &ioat_dma->dma_dev; + + ioat_disable_interrupts(ioat_dma); + + ioat_kobject_del(ioat_dma); + + dma_async_device_unregister(dma); + + pci_pool_destroy(ioat_dma->dma_pool); + pci_pool_destroy(ioat_dma->completion_pool); + + INIT_LIST_HEAD(&dma->channels); +} + +/** + * ioat_enumerate_channels - find and initialize the device's channels + * @ioat_dma: the ioat dma device to be enumerated + */ +int ioat_enumerate_channels(struct ioatdma_device *ioat_dma) +{ + struct ioatdma_chan *ioat_chan; + struct device *dev = &ioat_dma->pdev->dev; + struct dma_device *dma = &ioat_dma->dma_dev; + u8 xfercap_log; + int i; + + INIT_LIST_HEAD(&dma->channels); + dma->chancnt = readb(ioat_dma->reg_base + IOAT_CHANCNT_OFFSET); + dma->chancnt &= 0x1f; /* bits [4:0] valid */ + if (dma->chancnt > ARRAY_SIZE(ioat_dma->idx)) { + dev_warn(dev, "(%d) exceeds max supported channels (%zu)\n", + dma->chancnt, ARRAY_SIZE(ioat_dma->idx)); + dma->chancnt = ARRAY_SIZE(ioat_dma->idx); + } + xfercap_log = readb(ioat_dma->reg_base + IOAT_XFERCAP_OFFSET); + xfercap_log &= 0x1f; /* bits [4:0] valid */ + if (xfercap_log == 0) + return 0; + dev_dbg(dev, "%s: xfercap = %d\n", __func__, 1 << xfercap_log); + + for (i = 0; i < dma->chancnt; i++) { + ioat_chan = devm_kzalloc(dev, sizeof(*ioat_chan), GFP_KERNEL); + if (!ioat_chan) + break; + + ioat_init_channel(ioat_dma, ioat_chan, i); + ioat_chan->xfercap_log = xfercap_log; + spin_lock_init(&ioat_chan->prep_lock); + if (ioat_dma->reset_hw(ioat_chan)) { + i = 0; + break; + } + } + dma->chancnt = i; + return i; +} + +/** + * ioat_free_chan_resources - release all the descriptors + * @chan: the channel to be cleaned + */ +void ioat_free_chan_resources(struct dma_chan *c) +{ + struct ioatdma_chan *ioat_chan = to_ioat_chan(c); + struct ioatdma_device *ioat_dma = ioat_chan->ioat_dma; + struct ioat_ring_ent *desc; + const int total_descs = 1 << ioat_chan->alloc_order; + int descs; + int i; + + /* Before freeing channel resources first check + * if they have been previously allocated for this channel. + */ + if (!ioat_chan->ring) + return; + + ioat_stop(ioat_chan); + ioat_dma->reset_hw(ioat_chan); + + spin_lock_bh(&ioat_chan->cleanup_lock); + spin_lock_bh(&ioat_chan->prep_lock); + descs = ioat_ring_space(ioat_chan); + dev_dbg(to_dev(ioat_chan), "freeing %d idle descriptors\n", descs); + for (i = 0; i < descs; i++) { + desc = ioat_get_ring_ent(ioat_chan, ioat_chan->head + i); + ioat_free_ring_ent(desc, c); + } + + if (descs < total_descs) + dev_err(to_dev(ioat_chan), "Freeing %d in use descriptors!\n", + total_descs - descs); + + for (i = 0; i < total_descs - descs; i++) { + desc = ioat_get_ring_ent(ioat_chan, ioat_chan->tail + i); + dump_desc_dbg(ioat_chan, desc); + ioat_free_ring_ent(desc, c); + } + + kfree(ioat_chan->ring); + ioat_chan->ring = NULL; + ioat_chan->alloc_order = 0; + pci_pool_free(ioat_dma->completion_pool, ioat_chan->completion, + ioat_chan->completion_dma); + spin_unlock_bh(&ioat_chan->prep_lock); + spin_unlock_bh(&ioat_chan->cleanup_lock); + + ioat_chan->last_completion = 0; + ioat_chan->completion_dma = 0; + ioat_chan->dmacount = 0; +} + +/* ioat_alloc_chan_resources - allocate/initialize ioat descriptor ring + * @chan: channel to be initialized + */ +int ioat_alloc_chan_resources(struct dma_chan *c) +{ + struct ioatdma_chan *ioat_chan = to_ioat_chan(c); + struct ioat_ring_ent **ring; + u64 status; + int order; + int i = 0; + u32 chanerr; + + /* have we already been set up? */ + if (ioat_chan->ring) + return 1 << ioat_chan->alloc_order; + + /* Setup register to interrupt and write completion status on error */ + writew(IOAT_CHANCTRL_RUN, ioat_chan->reg_base + IOAT_CHANCTRL_OFFSET); + + /* allocate a completion writeback area */ + /* doing 2 32bit writes to mmio since 1 64b write doesn't work */ + ioat_chan->completion = + pci_pool_alloc(ioat_chan->ioat_dma->completion_pool, + GFP_KERNEL, &ioat_chan->completion_dma); + if (!ioat_chan->completion) + return -ENOMEM; + + memset(ioat_chan->completion, 0, sizeof(*ioat_chan->completion)); + writel(((u64)ioat_chan->completion_dma) & 0x00000000FFFFFFFF, + ioat_chan->reg_base + IOAT_CHANCMP_OFFSET_LOW); + writel(((u64)ioat_chan->completion_dma) >> 32, + ioat_chan->reg_base + IOAT_CHANCMP_OFFSET_HIGH); + + order = ioat_get_alloc_order(); + ring = ioat_alloc_ring(c, order, GFP_KERNEL); + if (!ring) + return -ENOMEM; + + spin_lock_bh(&ioat_chan->cleanup_lock); + spin_lock_bh(&ioat_chan->prep_lock); + ioat_chan->ring = ring; + ioat_chan->head = 0; + ioat_chan->issued = 0; + ioat_chan->tail = 0; + ioat_chan->alloc_order = order; + set_bit(IOAT_RUN, &ioat_chan->state); + spin_unlock_bh(&ioat_chan->prep_lock); + spin_unlock_bh(&ioat_chan->cleanup_lock); + + ioat_start_null_desc(ioat_chan); + + /* check that we got off the ground */ + do { + udelay(1); + status = ioat_chansts(ioat_chan); + } while (i++ < 20 && !is_ioat_active(status) && !is_ioat_idle(status)); + + if (is_ioat_active(status) || is_ioat_idle(status)) + return 1 << ioat_chan->alloc_order; + + chanerr = readl(ioat_chan->reg_base + IOAT_CHANERR_OFFSET); + + dev_WARN(to_dev(ioat_chan), + "failed to start channel chanerr: %#x\n", chanerr); + ioat_free_chan_resources(c); + return -EFAULT; +} + +/* common channel initialization */ +void +ioat_init_channel(struct ioatdma_device *ioat_dma, + struct ioatdma_chan *ioat_chan, int idx) +{ + struct dma_device *dma = &ioat_dma->dma_dev; + struct dma_chan *c = &ioat_chan->dma_chan; + unsigned long data = (unsigned long) c; + + ioat_chan->ioat_dma = ioat_dma; + ioat_chan->reg_base = ioat_dma->reg_base + (0x80 * (idx + 1)); + spin_lock_init(&ioat_chan->cleanup_lock); + ioat_chan->dma_chan.device = dma; + dma_cookie_init(&ioat_chan->dma_chan); + list_add_tail(&ioat_chan->dma_chan.device_node, &dma->channels); + ioat_dma->idx[idx] = ioat_chan; + init_timer(&ioat_chan->timer); + ioat_chan->timer.function = ioat_dma->timer_fn; + ioat_chan->timer.data = data; + tasklet_init(&ioat_chan->cleanup_task, ioat_dma->cleanup_fn, data); +} + +static void ioat3_dma_test_callback(void *dma_async_param) +{ + struct completion *cmp = dma_async_param; + + complete(cmp); +} + +#define IOAT_NUM_SRC_TEST 6 /* must be <= 8 */ +static int ioat_xor_val_self_test(struct ioatdma_device *ioat_dma) +{ + int i, src_idx; + struct page *dest; + struct page *xor_srcs[IOAT_NUM_SRC_TEST]; + struct page *xor_val_srcs[IOAT_NUM_SRC_TEST + 1]; + dma_addr_t dma_srcs[IOAT_NUM_SRC_TEST + 1]; + dma_addr_t dest_dma; + struct dma_async_tx_descriptor *tx; + struct dma_chan *dma_chan; + dma_cookie_t cookie; + u8 cmp_byte = 0; + u32 cmp_word; + u32 xor_val_result; + int err = 0; + struct completion cmp; + unsigned long tmo; + struct device *dev = &ioat_dma->pdev->dev; + struct dma_device *dma = &ioat_dma->dma_dev; + u8 op = 0; + + dev_dbg(dev, "%s\n", __func__); + + if (!dma_has_cap(DMA_XOR, dma->cap_mask)) + return 0; + + for (src_idx = 0; src_idx < IOAT_NUM_SRC_TEST; src_idx++) { + xor_srcs[src_idx] = alloc_page(GFP_KERNEL); + if (!xor_srcs[src_idx]) { + while (src_idx--) + __free_page(xor_srcs[src_idx]); + return -ENOMEM; + } + } + + dest = alloc_page(GFP_KERNEL); + if (!dest) { + while (src_idx--) + __free_page(xor_srcs[src_idx]); + return -ENOMEM; + } + + /* Fill in src buffers */ + for (src_idx = 0; src_idx < IOAT_NUM_SRC_TEST; src_idx++) { + u8 *ptr = page_address(xor_srcs[src_idx]); + + for (i = 0; i < PAGE_SIZE; i++) + ptr[i] = (1 << src_idx); + } + + for (src_idx = 0; src_idx < IOAT_NUM_SRC_TEST; src_idx++) + cmp_byte ^= (u8) (1 << src_idx); + + cmp_word = (cmp_byte << 24) | (cmp_byte << 16) | + (cmp_byte << 8) | cmp_byte; + + memset(page_address(dest), 0, PAGE_SIZE); + + dma_chan = container_of(dma->channels.next, struct dma_chan, + device_node); + if (dma->device_alloc_chan_resources(dma_chan) < 1) { + err = -ENODEV; + goto out; + } + + /* test xor */ + op = IOAT_OP_XOR; + + dest_dma = dma_map_page(dev, dest, 0, PAGE_SIZE, DMA_FROM_DEVICE); + if (dma_mapping_error(dev, dest_dma)) + goto dma_unmap; + + for (i = 0; i < IOAT_NUM_SRC_TEST; i++) + dma_srcs[i] = DMA_ERROR_CODE; + for (i = 0; i < IOAT_NUM_SRC_TEST; i++) { + dma_srcs[i] = dma_map_page(dev, xor_srcs[i], 0, PAGE_SIZE, + DMA_TO_DEVICE); + if (dma_mapping_error(dev, dma_srcs[i])) + goto dma_unmap; + } + tx = dma->device_prep_dma_xor(dma_chan, dest_dma, dma_srcs, + IOAT_NUM_SRC_TEST, PAGE_SIZE, + DMA_PREP_INTERRUPT); + + if (!tx) { + dev_err(dev, "Self-test xor prep failed\n"); + err = -ENODEV; + goto dma_unmap; + } + + async_tx_ack(tx); + init_completion(&cmp); + tx->callback = ioat3_dma_test_callback; + tx->callback_param = &cmp; + cookie = tx->tx_submit(tx); + if (cookie < 0) { + dev_err(dev, "Self-test xor setup failed\n"); + err = -ENODEV; + goto dma_unmap; + } + dma->device_issue_pending(dma_chan); + + tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); + + if (tmo == 0 || + dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) { + dev_err(dev, "Self-test xor timed out\n"); + err = -ENODEV; + goto dma_unmap; + } + + for (i = 0; i < IOAT_NUM_SRC_TEST; i++) + dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, DMA_TO_DEVICE); + + dma_sync_single_for_cpu(dev, dest_dma, PAGE_SIZE, DMA_FROM_DEVICE); + for (i = 0; i < (PAGE_SIZE / sizeof(u32)); i++) { + u32 *ptr = page_address(dest); + + if (ptr[i] != cmp_word) { + dev_err(dev, "Self-test xor failed compare\n"); + err = -ENODEV; + goto free_resources; + } + } + dma_sync_single_for_device(dev, dest_dma, PAGE_SIZE, DMA_FROM_DEVICE); + + dma_unmap_page(dev, dest_dma, PAGE_SIZE, DMA_FROM_DEVICE); + + /* skip validate if the capability is not present */ + if (!dma_has_cap(DMA_XOR_VAL, dma_chan->device->cap_mask)) + goto free_resources; + + op = IOAT_OP_XOR_VAL; + + /* validate the sources with the destintation page */ + for (i = 0; i < IOAT_NUM_SRC_TEST; i++) + xor_val_srcs[i] = xor_srcs[i]; + xor_val_srcs[i] = dest; + + xor_val_result = 1; + + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) + dma_srcs[i] = DMA_ERROR_CODE; + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) { + dma_srcs[i] = dma_map_page(dev, xor_val_srcs[i], 0, PAGE_SIZE, + DMA_TO_DEVICE); + if (dma_mapping_error(dev, dma_srcs[i])) + goto dma_unmap; + } + tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs, + IOAT_NUM_SRC_TEST + 1, PAGE_SIZE, + &xor_val_result, DMA_PREP_INTERRUPT); + if (!tx) { + dev_err(dev, "Self-test zero prep failed\n"); + err = -ENODEV; + goto dma_unmap; + } + + async_tx_ack(tx); + init_completion(&cmp); + tx->callback = ioat3_dma_test_callback; + tx->callback_param = &cmp; + cookie = tx->tx_submit(tx); + if (cookie < 0) { + dev_err(dev, "Self-test zero setup failed\n"); + err = -ENODEV; + goto dma_unmap; + } + dma->device_issue_pending(dma_chan); + + tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); + + if (tmo == 0 || + dma->device_tx_status(dma_chan, cookie, NULL) != DMA_COMPLETE) { + dev_err(dev, "Self-test validate timed out\n"); + err = -ENODEV; + goto dma_unmap; + } + + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) + dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, DMA_TO_DEVICE); + + if (xor_val_result != 0) { + dev_err(dev, "Self-test validate failed compare\n"); + err = -ENODEV; + goto free_resources; + } + + memset(page_address(dest), 0, PAGE_SIZE); + + /* test for non-zero parity sum */ + op = IOAT_OP_XOR_VAL; + + xor_val_result = 0; + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) + dma_srcs[i] = DMA_ERROR_CODE; + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) { + dma_srcs[i] = dma_map_page(dev, xor_val_srcs[i], 0, PAGE_SIZE, + DMA_TO_DEVICE); + if (dma_mapping_error(dev, dma_srcs[i])) + goto dma_unmap; + } + tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs, + IOAT_NUM_SRC_TEST + 1, PAGE_SIZE, + &xor_val_result, DMA_PREP_INTERRUPT); + if (!tx) { + dev_err(dev, "Self-test 2nd zero prep failed\n"); + err = -ENODEV; + goto dma_unmap; + } + + async_tx_ack(tx); + init_completion(&cmp); + tx->callback = ioat3_dma_test_callback; + tx->callback_param = &cmp; + cookie = tx->tx_submit(tx); + if (cookie < 0) { + dev_err(dev, "Self-test 2nd zero setup failed\n"); + err = -ENODEV; + goto dma_unmap; + } + dma->device_issue_pending(dma_chan); + + tmo = wait_for_completion_timeout(&cmp, msecs_to_jiffies(3000)); + + if (tmo == 0 || + 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; + } + + if (xor_val_result != SUM_CHECK_P_RESULT) { + dev_err(dev, "Self-test validate failed compare\n"); + err = -ENODEV; + goto dma_unmap; + } + + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) + dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, DMA_TO_DEVICE); + + goto free_resources; +dma_unmap: + if (op == IOAT_OP_XOR) { + if (dest_dma != DMA_ERROR_CODE) + dma_unmap_page(dev, dest_dma, PAGE_SIZE, + DMA_FROM_DEVICE); + for (i = 0; i < IOAT_NUM_SRC_TEST; i++) + if (dma_srcs[i] != DMA_ERROR_CODE) + dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, + DMA_TO_DEVICE); + } else if (op == IOAT_OP_XOR_VAL) { + for (i = 0; i < IOAT_NUM_SRC_TEST + 1; i++) + if (dma_srcs[i] != DMA_ERROR_CODE) + dma_unmap_page(dev, dma_srcs[i], PAGE_SIZE, + DMA_TO_DEVICE); + } +free_resources: + dma->device_free_chan_resources(dma_chan); +out: + src_idx = IOAT_NUM_SRC_TEST; + while (src_idx--) + __free_page(xor_srcs[src_idx]); + __free_page(dest); + return err; +} + +static int ioat3_dma_self_test(struct ioatdma_device *ioat_dma) +{ + int rc = ioat_dma_self_test(ioat_dma); + + if (rc) + return rc; + + rc = ioat_xor_val_self_test(ioat_dma); + if (rc) + return rc; + + return 0; +} + +static void ioat3_intr_quirk(struct ioatdma_device *ioat_dma) +{ + struct dma_device *dma; + struct dma_chan *c; + struct ioatdma_chan *ioat_chan; + u32 errmask; + + dma = &ioat_dma->dma_dev; + + /* + * if we have descriptor write back error status, we mask the + * error interrupts + */ + if (ioat_dma->cap & IOAT_CAP_DWBES) { + list_for_each_entry(c, &dma->channels, device_node) { + ioat_chan = to_ioat_chan(c); + errmask = readl(ioat_chan->reg_base + + IOAT_CHANERR_MASK_OFFSET); + errmask |= IOAT_CHANERR_XOR_P_OR_CRC_ERR | + IOAT_CHANERR_XOR_Q_ERR; + writel(errmask, ioat_chan->reg_base + + IOAT_CHANERR_MASK_OFFSET); + } + } +} + +int ioat3_dma_probe(struct ioatdma_device *ioat_dma, int dca) +{ + struct pci_dev *pdev = ioat_dma->pdev; + int dca_en = system_has_dca_enabled(pdev); + struct dma_device *dma; + struct dma_chan *c; + struct ioatdma_chan *ioat_chan; + bool is_raid_device = false; + int err; + + ioat_dma->enumerate_channels = ioat_enumerate_channels; + ioat_dma->reset_hw = ioat_reset_hw; + ioat_dma->self_test = ioat3_dma_self_test; + ioat_dma->intr_quirk = ioat3_intr_quirk; + dma = &ioat_dma->dma_dev; + dma->device_prep_dma_memcpy = ioat_dma_prep_memcpy_lock; + dma->device_issue_pending = ioat_issue_pending; + dma->device_alloc_chan_resources = ioat_alloc_chan_resources; + dma->device_free_chan_resources = ioat_free_chan_resources; + + dma_cap_set(DMA_INTERRUPT, dma->cap_mask); + dma->device_prep_dma_interrupt = ioat_prep_interrupt_lock; + + ioat_dma->cap = readl(ioat_dma->reg_base + IOAT_DMA_CAP_OFFSET); + + if (is_xeon_cb32(pdev) || is_bwd_noraid(pdev)) + ioat_dma->cap &= + ~(IOAT_CAP_XOR | IOAT_CAP_PQ | IOAT_CAP_RAID16SS); + + /* dca is incompatible with raid operations */ + if (dca_en && (ioat_dma->cap & (IOAT_CAP_XOR|IOAT_CAP_PQ))) + ioat_dma->cap &= ~(IOAT_CAP_XOR|IOAT_CAP_PQ); + + if (ioat_dma->cap & IOAT_CAP_XOR) { + is_raid_device = true; + dma->max_xor = 8; + + dma_cap_set(DMA_XOR, dma->cap_mask); + dma->device_prep_dma_xor = ioat_prep_xor; + + dma_cap_set(DMA_XOR_VAL, dma->cap_mask); + dma->device_prep_dma_xor_val = ioat_prep_xor_val; + } + + if (ioat_dma->cap & IOAT_CAP_PQ) { + is_raid_device = true; + + dma->device_prep_dma_pq = ioat_prep_pq; + dma->device_prep_dma_pq_val = ioat_prep_pq_val; + dma_cap_set(DMA_PQ, dma->cap_mask); + dma_cap_set(DMA_PQ_VAL, dma->cap_mask); + + if (ioat_dma->cap & IOAT_CAP_RAID16SS) + dma_set_maxpq(dma, 16, 0); + else + dma_set_maxpq(dma, 8, 0); + + if (!(ioat_dma->cap & IOAT_CAP_XOR)) { + dma->device_prep_dma_xor = ioat_prep_pqxor; + dma->device_prep_dma_xor_val = ioat_prep_pqxor_val; + dma_cap_set(DMA_XOR, dma->cap_mask); + dma_cap_set(DMA_XOR_VAL, dma->cap_mask); + + if (ioat_dma->cap & IOAT_CAP_RAID16SS) + dma->max_xor = 16; + else + dma->max_xor = 8; + } + } + + dma->device_tx_status = ioat_tx_status; + ioat_dma->cleanup_fn = ioat_cleanup_event; + ioat_dma->timer_fn = ioat_timer_event; + + /* starting with CB3.3 super extended descriptors are supported */ + if (ioat_dma->cap & IOAT_CAP_RAID16SS) { + char pool_name[14]; + int i; + + for (i = 0; i < MAX_SED_POOLS; i++) { + snprintf(pool_name, 14, "ioat_hw%d_sed", i); + + /* allocate SED DMA pool */ + ioat_dma->sed_hw_pool[i] = dmam_pool_create(pool_name, + &pdev->dev, + SED_SIZE * (i + 1), 64, 0); + if (!ioat_dma->sed_hw_pool[i]) + return -ENOMEM; + + } + } + + if (!(ioat_dma->cap & (IOAT_CAP_XOR | IOAT_CAP_PQ))) + dma_cap_set(DMA_PRIVATE, dma->cap_mask); + + err = ioat_probe(ioat_dma); + if (err) + return err; + + list_for_each_entry(c, &dma->channels, device_node) { + ioat_chan = to_ioat_chan(c); + writel(IOAT_DMA_DCA_ANY_CPU, + ioat_chan->reg_base + IOAT_DCACTRL_OFFSET); + } + + err = ioat_register(ioat_dma); + if (err) + return err; + + ioat_kobject_add(ioat_dma, &ioat_ktype); + + if (dca) + ioat_dma->dca = ioat3_dca_init(pdev, ioat_dma->reg_base); + + return 0; +} + +#define DRV_NAME "ioatdma" + +static struct pci_driver ioat_pci_driver = { + .name = DRV_NAME, + .id_table = ioat_pci_tbl, + .probe = ioat_pci_probe, + .remove = ioat_remove, +}; + +static struct ioatdma_device * +alloc_ioatdma(struct pci_dev *pdev, void __iomem *iobase) +{ + struct device *dev = &pdev->dev; + struct ioatdma_device *d = devm_kzalloc(dev, sizeof(*d), GFP_KERNEL); + + if (!d) + return NULL; + d->pdev = pdev; + d->reg_base = iobase; + return d; +} + +static int ioat_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) +{ + void __iomem * const *iomap; + struct device *dev = &pdev->dev; + struct ioatdma_device *device; + int err; + + err = pcim_enable_device(pdev); + if (err) + return err; + + err = pcim_iomap_regions(pdev, 1 << IOAT_MMIO_BAR, DRV_NAME); + if (err) + return err; + iomap = pcim_iomap_table(pdev); + if (!iomap) + return -ENOMEM; + + err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64)); + if (err) + err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32)); + if (err) + return err; + + err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)); + if (err) + err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32)); + if (err) + return err; + + device = alloc_ioatdma(pdev, iomap[IOAT_MMIO_BAR]); + if (!device) + return -ENOMEM; + pci_set_master(pdev); + pci_set_drvdata(pdev, device); + + device->version = readb(device->reg_base + IOAT_VER_OFFSET); + if (device->version >= IOAT_VER_3_0) + err = ioat3_dma_probe(device, ioat_dca_enabled); + else + return -ENODEV; + + if (err) { + dev_err(dev, "Intel(R) I/OAT DMA Engine init failed\n"); + return -ENODEV; + } + + return 0; +} + +static void ioat_remove(struct pci_dev *pdev) +{ + struct ioatdma_device *device = pci_get_drvdata(pdev); + + if (!device) + return; + + dev_err(&pdev->dev, "Removing dma and dca services\n"); + if (device->dca) { + unregister_dca_provider(device->dca, &pdev->dev); + free_dca_provider(device->dca); + device->dca = NULL; + } + ioat_dma_remove(device); +} + +static int __init ioat_init_module(void) +{ + int err = -ENOMEM; + + pr_info("%s: Intel(R) QuickData Technology Driver %s\n", + DRV_NAME, IOAT_DMA_VERSION); + + ioat_cache = kmem_cache_create("ioat", sizeof(struct ioat_ring_ent), + 0, SLAB_HWCACHE_ALIGN, NULL); + if (!ioat_cache) + return -ENOMEM; + + ioat_sed_cache = KMEM_CACHE(ioat_sed_ent, 0); + if (!ioat_sed_cache) + goto err_ioat_cache; + + err = pci_register_driver(&ioat_pci_driver); + if (err) + goto err_ioat3_cache; + + return 0; + + err_ioat3_cache: + kmem_cache_destroy(ioat_sed_cache); + + err_ioat_cache: + kmem_cache_destroy(ioat_cache); + + return err; +} +module_init(ioat_init_module); + +static void __exit ioat_exit_module(void) +{ + pci_unregister_driver(&ioat_pci_driver); + kmem_cache_destroy(ioat_cache); +} +module_exit(ioat_exit_module); diff --git a/drivers/dma/ioat/pci.c b/drivers/dma/ioat/pci.c deleted file mode 100644 index 0ee610224ecd..000000000000 --- a/drivers/dma/ioat/pci.c +++ /dev/null @@ -1,278 +0,0 @@ -/* - * Intel I/OAT DMA Linux driver - * Copyright(c) 2007 - 2009 Intel Corporation. - * - * This program is free software; you can redistribute it and/or modify it - * under the terms and conditions of the GNU General Public License, - * version 2, as published by the Free Software Foundation. - * - * This program is distributed in the hope that it will be useful, but WITHOUT - * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or - * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for - * more details. - * - * The full GNU General Public License is included in this distribution in - * the file called "COPYING". - * - */ - -/* - * This driver supports an Intel I/OAT DMA engine, which does asynchronous - * copy operations. - */ - -#include -#include -#include -#include -#include -#include -#include -#include "dma.h" -#include "registers.h" -#include "hw.h" - -MODULE_VERSION(IOAT_DMA_VERSION); -MODULE_LICENSE("Dual BSD/GPL"); -MODULE_AUTHOR("Intel Corporation"); - -static struct pci_device_id ioat_pci_tbl[] = { - /* I/OAT v3 platforms */ - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG3) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG4) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG5) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG6) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_TBG7) }, - - /* I/OAT v3.2 platforms */ - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF3) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF4) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF5) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF6) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF7) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF8) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_JSF9) }, - - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB3) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB4) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB5) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB6) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB7) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB8) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_SNB9) }, - - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB3) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB4) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB5) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB6) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB7) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB8) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_IVB9) }, - - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW3) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW4) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW5) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW6) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW7) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW8) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_HSW9) }, - - /* I/OAT v3.3 platforms */ - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BWD3) }, - - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE0) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE1) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE2) }, - { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_IOAT_BDXDE3) }, - - { 0, } -}; -MODULE_DEVICE_TABLE(pci, ioat_pci_tbl); - -static int ioat_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id); -static void ioat_remove(struct pci_dev *pdev); - -static int ioat_dca_enabled = 1; -module_param(ioat_dca_enabled, int, 0644); -MODULE_PARM_DESC(ioat_dca_enabled, "control support of dca service (default: 1)"); - -struct kmem_cache *ioat_cache; -struct kmem_cache *ioat3_sed_cache; - -#define DRV_NAME "ioatdma" - -static struct pci_driver ioat_pci_driver = { - .name = DRV_NAME, - .id_table = ioat_pci_tbl, - .probe = ioat_pci_probe, - .remove = ioat_remove, -}; - -static struct ioatdma_device * -alloc_ioatdma(struct pci_dev *pdev, void __iomem *iobase) -{ - struct device *dev = &pdev->dev; - struct ioatdma_device *d = devm_kzalloc(dev, sizeof(*d), GFP_KERNEL); - - if (!d) - return NULL; - d->pdev = pdev; - d->reg_base = iobase; - return d; -} - -/* - * The dmaengine core assumes that async DMA devices will only be removed - * when they not used anymore, or it assumes dma_async_device_unregister() - * will only be called by dma driver exit routines. But this assumption is - * not true for the IOAT driver, which calls dma_async_device_unregister() - * from ioat_remove(). So current IOAT driver doesn't support device - * hot-removal because it may cause system crash to hot-remove inuse IOAT - * devices. - * - * This is a hack to disable IOAT devices under ejectable PCI host bridge - * so it won't break PCI host bridge hot-removal. - */ -static bool ioat_pci_has_ejectable_acpi_ancestor(struct pci_dev *pdev) -{ -#ifdef CONFIG_ACPI - struct pci_bus *bus = pdev->bus; - struct acpi_device *adev; - - while (bus->parent) - bus = bus->parent; - for (adev = ACPI_COMPANION(bus->bridge); adev; adev = adev->parent) - if (adev->flags.ejectable) - return true; -#endif - - return false; -} - -static int ioat_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) -{ - void __iomem * const *iomap; - struct device *dev = &pdev->dev; - struct ioatdma_device *device; - int err; - - if (ioat_pci_has_ejectable_acpi_ancestor(pdev)) { - dev_dbg(&pdev->dev, "ignore ejectable IOAT device.\n"); - return -ENODEV; - } - - err = pcim_enable_device(pdev); - if (err) - return err; - - err = pcim_iomap_regions(pdev, 1 << IOAT_MMIO_BAR, DRV_NAME); - if (err) - return err; - iomap = pcim_iomap_table(pdev); - if (!iomap) - return -ENOMEM; - - err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64)); - if (err) - err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32)); - if (err) - return err; - - err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)); - if (err) - err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32)); - if (err) - return err; - - device = alloc_ioatdma(pdev, iomap[IOAT_MMIO_BAR]); - if (!device) - return -ENOMEM; - pci_set_master(pdev); - pci_set_drvdata(pdev, device); - - device->version = readb(device->reg_base + IOAT_VER_OFFSET); - if (device->version >= IOAT_VER_3_0) - err = ioat3_dma_probe(device, ioat_dca_enabled); - else - return -ENODEV; - - if (err) { - dev_err(dev, "Intel(R) I/OAT DMA Engine init failed\n"); - return -ENODEV; - } - - return 0; -} - -static void ioat_remove(struct pci_dev *pdev) -{ - struct ioatdma_device *device = pci_get_drvdata(pdev); - - if (!device) - return; - - dev_err(&pdev->dev, "Removing dma and dca services\n"); - if (device->dca) { - unregister_dca_provider(device->dca, &pdev->dev); - free_dca_provider(device->dca); - device->dca = NULL; - } - ioat_dma_remove(device); -} - -static int __init ioat_init_module(void) -{ - int err = -ENOMEM; - - pr_info("%s: Intel(R) QuickData Technology Driver %s\n", - DRV_NAME, IOAT_DMA_VERSION); - - ioat_cache = kmem_cache_create("ioat", sizeof(struct ioat_ring_ent), - 0, SLAB_HWCACHE_ALIGN, NULL); - if (!ioat_cache) - return -ENOMEM; - - ioat3_sed_cache = KMEM_CACHE(ioat_sed_ent, 0); - if (!ioat3_sed_cache) - goto err_ioat_cache; - - err = pci_register_driver(&ioat_pci_driver); - if (err) - goto err_ioat3_cache; - - return 0; - - err_ioat3_cache: - kmem_cache_destroy(ioat3_sed_cache); - - err_ioat_cache: - kmem_cache_destroy(ioat_cache); - - return err; -} -module_init(ioat_init_module); - -static void __exit ioat_exit_module(void) -{ - pci_unregister_driver(&ioat_pci_driver); - kmem_cache_destroy(ioat_cache); -} -module_exit(ioat_exit_module);