diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2013-05-09 09:46:45 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2013-05-09 09:46:45 -0700 |
commit | 1763e735b0a093a6747078b3bd101f079e576ab6 (patch) | |
tree | 75203a3229977d12dc5a5990d1122e7a3d5f30fc /drivers/dma/ioat | |
parent | b29bdba51924f6fd5971352ba111784dee3a5853 (diff) | |
parent | 3065c194670b61e213656ce25976d7c8a95e3c93 (diff) |
Merge branch 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma
Pull slave-dmaengine updates from Vinod Koul:
"This time we have dmatest improvements from Andy along with dw_dmac
fixes. He has also done support for acpi for dmanegine.
Also we have bunch of fixes going in DT support for dmanegine for
various folks. Then Haswell and other ioat changes from Dave and
SUDMAC support from Shimoda."
* 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma: (53 commits)
dma: tegra: implement suspend/resume callbacks
dma:of: Use a mutex to protect the of_dma_list
dma: of: Fix of_node reference leak
dmaengine: sirf: move driver init from module_init to subsys_initcall
sudmac: add support for SUDMAC
dma: sh: add Kconfig
at_hdmac: move to generic DMA binding
ioatdma: ioat3_alloc_sed can be static
ioatdma: Adding write back descriptor error status support for ioatdma 3.3
ioatdma: S1200 platforms ioatdma channel 2 and 3 falsely advertise RAID cap
ioatdma: Adding support for 16 src PQ ops and super extended descriptors
ioatdma: Removing hw bug workaround for CB3.x .2 and earlier
dw_dmac: add ACPI support
dmaengine: call acpi_dma_request_slave_channel as well
dma: acpi-dma: introduce ACPI DMA helpers
dma: of: Remove unnecessary list_empty check
DMA: OF: Check properties value before running be32_to_cpup() on it
DMA: of: Constant names
ioatdma: skip silicon bug workaround for pq_align for cb3.3
ioatdma: Removing PQ val disable for cb3.3
...
Diffstat (limited to 'drivers/dma/ioat')
-rw-r--r-- | drivers/dma/ioat/dma.c | 8 | ||||
-rw-r--r-- | drivers/dma/ioat/dma.h | 53 | ||||
-rw-r--r-- | drivers/dma/ioat/dma_v2.h | 2 | ||||
-rw-r--r-- | drivers/dma/ioat/dma_v3.c | 912 | ||||
-rw-r--r-- | drivers/dma/ioat/hw.h | 88 | ||||
-rw-r--r-- | drivers/dma/ioat/pci.c | 20 | ||||
-rw-r--r-- | drivers/dma/ioat/registers.h | 4 |
7 files changed, 950 insertions, 137 deletions
diff --git a/drivers/dma/ioat/dma.c b/drivers/dma/ioat/dma.c index 1879a5942bfc..17a2393b3e25 100644 --- a/drivers/dma/ioat/dma.c +++ b/drivers/dma/ioat/dma.c @@ -892,7 +892,7 @@ MODULE_PARM_DESC(ioat_interrupt_style, * ioat_dma_setup_interrupts - setup interrupt handler * @device: ioat device */ -static int ioat_dma_setup_interrupts(struct ioatdma_device *device) +int ioat_dma_setup_interrupts(struct ioatdma_device *device) { struct ioat_chan_common *chan; struct pci_dev *pdev = device->pdev; @@ -941,6 +941,7 @@ msix: } } intrctrl |= IOAT_INTRCTRL_MSIX_VECTOR_CONTROL; + device->irq_mode = IOAT_MSIX; goto done; msix_single_vector: @@ -956,6 +957,7 @@ msix_single_vector: pci_disable_msix(pdev); goto msi; } + device->irq_mode = IOAT_MSIX_SINGLE; goto done; msi: @@ -969,6 +971,7 @@ msi: pci_disable_msi(pdev); goto intx; } + device->irq_mode = IOAT_MSIX; goto done; intx: @@ -977,6 +980,7 @@ intx: if (err) goto err_no_irq; + device->irq_mode = IOAT_INTX; done: if (device->intr_quirk) device->intr_quirk(device); @@ -987,9 +991,11 @@ done: err_no_irq: /* Disable all interrupt generation */ writeb(0, device->reg_base + IOAT_INTRCTRL_OFFSET); + device->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 *device) { diff --git a/drivers/dma/ioat/dma.h b/drivers/dma/ioat/dma.h index 53a4cbb78f47..54fb7b9ff9aa 100644 --- a/drivers/dma/ioat/dma.h +++ b/drivers/dma/ioat/dma.h @@ -39,6 +39,7 @@ #define to_ioat_desc(lh) container_of(lh, struct ioat_desc_sw, node) #define tx_to_ioat_desc(tx) container_of(tx, struct ioat_desc_sw, txd) #define to_dev(ioat_chan) (&(ioat_chan)->device->pdev->dev) +#define to_pdev(ioat_chan) ((ioat_chan)->device->pdev) #define chan_num(ch) ((int)((ch)->reg_base - (ch)->device->reg_base) / 0x80) @@ -48,6 +49,14 @@ */ #define NULL_DESC_BUFFER_SIZE 1 +enum ioat_irq_mode { + IOAT_NOIRQ = 0, + IOAT_MSIX, + IOAT_MSIX_SINGLE, + IOAT_MSI, + IOAT_INTX +}; + /** * struct ioatdma_device - internal representation of a IOAT device * @pdev: PCI-Express device @@ -72,11 +81,16 @@ struct ioatdma_device { void __iomem *reg_base; struct pci_pool *dma_pool; struct pci_pool *completion_pool; +#define MAX_SED_POOLS 5 + struct dma_pool *sed_hw_pool[MAX_SED_POOLS]; + struct kmem_cache *sed_pool; struct dma_device common; u8 version; struct msix_entry msix_entries[4]; struct ioat_chan_common *idx[4]; struct dca_provider *dca; + enum ioat_irq_mode irq_mode; + u32 cap; void (*intr_quirk)(struct ioatdma_device *device); int (*enumerate_channels)(struct ioatdma_device *device); int (*reset_hw)(struct ioat_chan_common *chan); @@ -131,6 +145,20 @@ struct ioat_dma_chan { u16 active; }; +/** + * struct ioat_sed_ent - wrapper around super extended hardware descriptor + * @hw: hardware SED + * @sed_dma: dma address for the SED + * @list: list member + * @parent: point to the dma descriptor that's the parent + */ +struct ioat_sed_ent { + struct ioat_sed_raw_descriptor *hw; + dma_addr_t dma; + struct ioat_ring_ent *parent; + unsigned int hw_pool; +}; + static inline struct ioat_chan_common *to_chan_common(struct dma_chan *c) { return container_of(c, struct ioat_chan_common, common); @@ -179,7 +207,7 @@ __dump_desc_dbg(struct ioat_chan_common *chan, struct ioat_dma_descriptor *hw, struct device *dev = to_dev(chan); dev_dbg(dev, "desc[%d]: (%#llx->%#llx) cookie: %d flags: %#x" - " ctl: %#x (op: %d int_en: %d compl: %d)\n", id, + " ctl: %#10.8x (op: %#x int_en: %d compl: %d)\n", id, (unsigned long long) tx->phys, (unsigned long long) hw->next, tx->cookie, tx->flags, hw->ctl, hw->ctl_f.op, hw->ctl_f.int_en, hw->ctl_f.compl_write); @@ -201,7 +229,7 @@ ioat_chan_by_index(struct ioatdma_device *device, int index) return device->idx[index]; } -static inline u64 ioat_chansts(struct ioat_chan_common *chan) +static inline u64 ioat_chansts_32(struct ioat_chan_common *chan) { u8 ver = chan->device->version; u64 status; @@ -218,6 +246,26 @@ static inline u64 ioat_chansts(struct ioat_chan_common *chan) return status; } +#if BITS_PER_LONG == 64 + +static inline u64 ioat_chansts(struct ioat_chan_common *chan) +{ + u8 ver = chan->device->version; + u64 status; + + /* With IOAT v3.3 the status register is 64bit. */ + if (ver >= IOAT_VER_3_3) + status = readq(chan->reg_base + IOAT_CHANSTS_OFFSET(ver)); + else + status = ioat_chansts_32(chan); + + return status; +} + +#else +#define ioat_chansts ioat_chansts_32 +#endif + static inline void ioat_start(struct ioat_chan_common *chan) { u8 ver = chan->device->version; @@ -321,6 +369,7 @@ bool ioat_cleanup_preamble(struct ioat_chan_common *chan, dma_addr_t *phys_complete); void ioat_kobject_add(struct ioatdma_device *device, struct kobj_type *type); void ioat_kobject_del(struct ioatdma_device *device); +int ioat_dma_setup_interrupts(struct ioatdma_device *device); extern const struct sysfs_ops ioat_sysfs_ops; extern struct ioat_sysfs_entry ioat_version_attr; extern struct ioat_sysfs_entry ioat_cap_attr; diff --git a/drivers/dma/ioat/dma_v2.h b/drivers/dma/ioat/dma_v2.h index e100f644e344..29bf9448035d 100644 --- a/drivers/dma/ioat/dma_v2.h +++ b/drivers/dma/ioat/dma_v2.h @@ -137,6 +137,7 @@ struct ioat_ring_ent { #ifdef DEBUG int id; #endif + struct ioat_sed_ent *sed; }; static inline struct ioat_ring_ent * @@ -157,6 +158,7 @@ static inline void ioat2_set_chainaddr(struct ioat2_dma_chan *ioat, u64 addr) int ioat2_dma_probe(struct ioatdma_device *dev, int dca); int ioat3_dma_probe(struct ioatdma_device *dev, int dca); +void ioat3_dma_remove(struct ioatdma_device *dev); struct dca_provider *ioat2_dca_init(struct pci_dev *pdev, void __iomem *iobase); struct dca_provider *ioat3_dca_init(struct pci_dev *pdev, void __iomem *iobase); int ioat2_check_space_lock(struct ioat2_dma_chan *ioat, int num_descs); diff --git a/drivers/dma/ioat/dma_v3.c b/drivers/dma/ioat/dma_v3.c index e8336cce360b..ca6ea9b3551b 100644 --- a/drivers/dma/ioat/dma_v3.c +++ b/drivers/dma/ioat/dma_v3.c @@ -55,7 +55,7 @@ /* * Support routines for v3+ hardware */ - +#include <linux/module.h> #include <linux/pci.h> #include <linux/gfp.h> #include <linux/dmaengine.h> @@ -70,6 +70,10 @@ /* 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) +#define ndest_to_sw(x) ((x) + 1) +#define ndest_to_hw(x) ((x) - 1) +#define src16_cnt_to_sw(x) ((x) + 9) +#define src16_cnt_to_hw(x) ((x) - 9) /* provide a lookup table for setting the source address in the base or * extended descriptor of an xor or pq descriptor @@ -77,7 +81,20 @@ static const u8 xor_idx_to_desc = 0xe0; static const u8 xor_idx_to_field[] = { 1, 4, 5, 6, 7, 0, 1, 2 }; static const u8 pq_idx_to_desc = 0xf8; +static const u8 pq16_idx_to_desc[] = { 0, 0, 1, 1, 1, 1, 1, 1, 1, + 2, 2, 2, 2, 2, 2, 2 }; static const u8 pq_idx_to_field[] = { 1, 4, 5, 0, 1, 2, 4, 5 }; +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) { @@ -101,6 +118,13 @@ static dma_addr_t pq_get_src(struct ioat_raw_descriptor *descs[2], int idx) return raw->field[pq_idx_to_field[idx]]; } +static dma_addr_t pq16_get_src(struct ioat_raw_descriptor *desc[3], int idx) +{ + struct ioat_raw_descriptor *raw = desc[pq16_idx_to_desc[idx]]; + + return raw->field[pq16_idx_to_field[idx]]; +} + static void pq_set_src(struct ioat_raw_descriptor *descs[2], dma_addr_t addr, u32 offset, u8 coef, int idx) { @@ -111,6 +135,167 @@ static void pq_set_src(struct ioat_raw_descriptor *descs[2], 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) { + 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: + 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: + return true; + default: + return false; + } + +} + +static void pq16_set_src(struct ioat_raw_descriptor *desc[3], + dma_addr_t addr, u32 offset, u8 coef, int idx) +{ + struct ioat_pq_descriptor *pq = (struct ioat_pq_descriptor *)desc[0]; + struct ioat_pq16a_descriptor *pq16 = + (struct ioat_pq16a_descriptor *)desc[1]; + struct ioat_raw_descriptor *raw = desc[pq16_idx_to_desc[idx]]; + + raw->field[pq16_idx_to_field[idx]] = addr + offset; + + if (idx < 8) + pq->coef[idx] = coef; + else + pq16->coef[idx - 8] = coef; +} + +static struct ioat_sed_ent * +ioat3_alloc_sed(struct ioatdma_device *device, unsigned int hw_pool) +{ + struct ioat_sed_ent *sed; + gfp_t flags = __GFP_ZERO | GFP_ATOMIC; + + sed = kmem_cache_alloc(device->sed_pool, flags); + if (!sed) + return NULL; + + sed->hw_pool = hw_pool; + sed->hw = dma_pool_alloc(device->sed_hw_pool[hw_pool], + flags, &sed->dma); + if (!sed->hw) { + kmem_cache_free(device->sed_pool, sed); + return NULL; + } + + return sed; +} + +static void ioat3_free_sed(struct ioatdma_device *device, struct ioat_sed_ent *sed) +{ + if (!sed) + 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) { @@ -223,6 +408,54 @@ static void ioat3_dma_unmap(struct ioat2_dma_chan *ioat, } 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); @@ -250,6 +483,63 @@ static bool desc_has_ext(struct ioat_ring_ent *desc) return false; } +static u64 ioat3_get_current_completion(struct ioat_chan_common *chan) +{ + u64 phys_complete; + u64 completion; + + completion = *chan->completion; + phys_complete = ioat_chansts_to_addr(completion); + + dev_dbg(to_dev(chan), "%s: phys_complete: %#llx\n", __func__, + (unsigned long long) phys_complete); + + return phys_complete; +} + +static bool ioat3_cleanup_preamble(struct ioat_chan_common *chan, + u64 *phys_complete) +{ + *phys_complete = ioat3_get_current_completion(chan); + if (*phys_complete == chan->last_completion) + return false; + + clear_bit(IOAT_COMPLETION_ACK, &chan->state); + mod_timer(&chan->timer, jiffies + COMPLETION_TIMEOUT); + + return true; +} + +static void +desc_get_errstat(struct ioat2_dma_chan *ioat, struct ioat_ring_ent *desc) +{ + struct ioat_dma_descriptor *hw = desc->hw; + + switch (hw->ctl_f.op) { + case IOAT_OP_PQ_VAL: + case IOAT_OP_PQ_VAL_16S: + { + struct ioat_pq_descriptor *pq = desc->pq; + + /* check if there's error written */ + if (!pq->dwbes_f.wbes) + return; + + /* need to set a chanerr var for checking to clear later */ + + if (pq->dwbes_f.p_val_err) + *desc->result |= SUM_CHECK_P_RESULT; + + if (pq->dwbes_f.q_val_err) + *desc->result |= SUM_CHECK_Q_RESULT; + + return; + } + default: + return; + } +} + /** * __cleanup - reclaim used descriptors * @ioat: channel (ring) to clean @@ -260,6 +550,7 @@ static bool desc_has_ext(struct ioat_ring_ent *desc) static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete) { struct ioat_chan_common *chan = &ioat->base; + struct ioatdma_device *device = chan->device; struct ioat_ring_ent *desc; bool seen_current = false; int idx = ioat->tail, i; @@ -268,6 +559,16 @@ static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete) dev_dbg(to_dev(chan), "%s: head: %#x tail: %#x issued: %#x\n", __func__, ioat->head, ioat->tail, ioat->issued); + /* + * At restart of the channel, the completion address and the + * channel status will be 0 due to starting a new chain. Since + * it's new chain and the first descriptor "fails", there is + * nothing to clean up. We do not want to reap the entire submitted + * chain due to this 0 address value and then BUG. + */ + if (!phys_complete) + return; + active = ioat2_ring_active(ioat); for (i = 0; i < active && !seen_current; i++) { struct dma_async_tx_descriptor *tx; @@ -276,6 +577,11 @@ static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete) prefetch(ioat2_get_ring_ent(ioat, idx + i + 1)); desc = ioat2_get_ring_ent(ioat, idx + i); dump_desc_dbg(ioat, desc); + + /* set err stat if we are using dwbes */ + if (device->cap & IOAT_CAP_DWBES) + desc_get_errstat(ioat, desc); + tx = &desc->txd; if (tx->cookie) { dma_cookie_complete(tx); @@ -294,6 +600,12 @@ static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete) BUG_ON(i + 1 >= active); i++; } + + /* cleanup super extended descriptors */ + if (desc->sed) { + ioat3_free_sed(device, desc->sed); + desc->sed = NULL; + } } smp_mb(); /* finish all descriptor reads before incrementing tail */ ioat->tail = idx + i; @@ -314,11 +626,22 @@ static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete) static void ioat3_cleanup(struct ioat2_dma_chan *ioat) { struct ioat_chan_common *chan = &ioat->base; - dma_addr_t phys_complete; + u64 phys_complete; spin_lock_bh(&chan->cleanup_lock); - if (ioat_cleanup_preamble(chan, &phys_complete)) + + if (ioat3_cleanup_preamble(chan, &phys_complete)) __cleanup(ioat, phys_complete); + + if (is_ioat_halted(*chan->completion)) { + u32 chanerr = readl(chan->reg_base + IOAT_CHANERR_OFFSET); + + if (chanerr & IOAT_CHANERR_HANDLE_MASK) { + mod_timer(&chan->timer, jiffies + IDLE_TIMEOUT); + ioat3_eh(ioat); + } + } + spin_unlock_bh(&chan->cleanup_lock); } @@ -333,15 +656,78 @@ static void ioat3_cleanup_event(unsigned long data) static void ioat3_restart_channel(struct ioat2_dma_chan *ioat) { struct ioat_chan_common *chan = &ioat->base; - dma_addr_t phys_complete; + u64 phys_complete; ioat2_quiesce(chan, 0); - if (ioat_cleanup_preamble(chan, &phys_complete)) + if (ioat3_cleanup_preamble(chan, &phys_complete)) __cleanup(ioat, phys_complete); __ioat2_restart_chan(ioat); } +static void ioat3_eh(struct ioat2_dma_chan *ioat) +{ + struct ioat_chan_common *chan = &ioat->base; + struct pci_dev *pdev = to_pdev(chan); + struct ioat_dma_descriptor *hw; + u64 phys_complete; + struct ioat_ring_ent *desc; + u32 err_handled = 0; + u32 chanerr_int; + u32 chanerr; + + /* cleanup so tail points to descriptor that caused the error */ + if (ioat3_cleanup_preamble(chan, &phys_complete)) + __cleanup(ioat, phys_complete); + + chanerr = readl(chan->reg_base + IOAT_CHANERR_OFFSET); + pci_read_config_dword(pdev, IOAT_PCI_CHANERR_INT_OFFSET, &chanerr_int); + + dev_dbg(to_dev(chan), "%s: error = %x:%x\n", + __func__, chanerr, chanerr_int); + + desc = ioat2_get_ring_ent(ioat, ioat->tail); + hw = desc->hw; + dump_desc_dbg(ioat, desc); + + switch (hw->ctl_f.op) { + case IOAT_OP_XOR_VAL: + if (chanerr & IOAT_CHANERR_XOR_P_OR_CRC_ERR) { + *desc->result |= SUM_CHECK_P_RESULT; + err_handled |= IOAT_CHANERR_XOR_P_OR_CRC_ERR; + } + break; + case IOAT_OP_PQ_VAL: + case IOAT_OP_PQ_VAL_16S: + if (chanerr & IOAT_CHANERR_XOR_P_OR_CRC_ERR) { + *desc->result |= SUM_CHECK_P_RESULT; + err_handled |= IOAT_CHANERR_XOR_P_OR_CRC_ERR; + } + if (chanerr & IOAT_CHANERR_XOR_Q_ERR) { + *desc->result |= SUM_CHECK_Q_RESULT; + err_handled |= IOAT_CHANERR_XOR_Q_ERR; + } + break; + } + + /* fault on unhandled error or spurious halt */ + if (chanerr ^ err_handled || chanerr == 0) { + dev_err(to_dev(chan), "%s: fatal error (%x:%x)\n", + __func__, chanerr, err_handled); + BUG(); + } + + writel(chanerr, chan->reg_base + IOAT_CHANERR_OFFSET); + pci_write_config_dword(pdev, IOAT_PCI_CHANERR_INT_OFFSET, chanerr_int); + + /* mark faulting descriptor as complete */ + *chan->completion = desc->txd.phys; + + spin_lock_bh(&ioat->prep_lock); + ioat3_restart_channel(ioat); + spin_unlock_bh(&ioat->prep_lock); +} + static void check_active(struct ioat2_dma_chan *ioat) { struct ioat_chan_common *chan = &ioat->base; @@ -605,7 +991,8 @@ dump_pq_desc_dbg(struct ioat2_dma_chan *ioat, struct ioat_ring_ent *desc, struct int i; dev_dbg(dev, "desc[%d]: (%#llx->%#llx) flags: %#x" - " sz: %#x ctl: %#x (op: %d int: %d compl: %d pq: '%s%s' src_cnt: %d)\n", + " sz: %#10.8x ctl: %#x (op: %#x int: %d compl: %d pq: '%s%s'" + " src_cnt: %d)\n", desc_id(desc), (unsigned long long) desc->txd.phys, (unsigned long long) (pq_ex ? pq_ex->next : pq->next), desc->txd.flags, pq->size, pq->ctl, pq->ctl_f.op, pq->ctl_f.int_en, @@ -617,6 +1004,42 @@ dump_pq_desc_dbg(struct ioat2_dma_chan *ioat, struct ioat_ring_ent *desc, struct (unsigned long long) pq_get_src(descs, i), pq->coef[i]); dev_dbg(dev, "\tP: %#llx\n", pq->p_addr); dev_dbg(dev, "\tQ: %#llx\n", pq->q_addr); + dev_dbg(dev, "\tNEXT: %#llx\n", pq->next); +} + +static void dump_pq16_desc_dbg(struct ioat2_dma_chan *ioat, + struct ioat_ring_ent *desc) +{ + struct device *dev = to_dev(&ioat->base); + struct ioat_pq_descriptor *pq = desc->pq; + struct ioat_raw_descriptor *descs[] = { (void *)pq, + (void *)pq, + (void *)pq }; + int src_cnt = src16_cnt_to_sw(pq->ctl_f.src_cnt); + int i; + + if (desc->sed) { + descs[1] = (void *)desc->sed->hw; + descs[2] = (void *)desc->sed->hw + 64; + } + + dev_dbg(dev, "desc[%d]: (%#llx->%#llx) flags: %#x" + " sz: %#x ctl: %#x (op: %#x int: %d compl: %d pq: '%s%s'" + " src_cnt: %d)\n", + desc_id(desc), (unsigned long long) desc->txd.phys, + (unsigned long long) pq->next, + desc->txd.flags, pq->size, pq->ctl, + pq->ctl_f.op, pq->ctl_f.int_en, + pq->ctl_f.compl_write, + pq->ctl_f.p_disable ? "" : "p", pq->ctl_f.q_disable ? "" : "q", + pq->ctl_f.src_cnt); + for (i = 0; i < src_cnt; i++) { + dev_dbg(dev, "\tsrc[%d]: %#llx coef: %#x\n", i, + (unsigned long long) pq16_get_src(descs, i), + pq->coef[i]); + } + dev_dbg(dev, "\tP: %#llx\n", pq->p_addr); + dev_dbg(dev, "\tQ: %#llx\n", pq->q_addr); } static struct dma_async_tx_descriptor * @@ -627,6 +1050,7 @@ __ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, { struct ioat2_dma_chan *ioat = to_ioat2_chan(c); struct ioat_chan_common *chan = &ioat->base; + struct ioatdma_device *device = chan->device; struct ioat_ring_ent *compl_desc; struct ioat_ring_ent *desc; struct ioat_ring_ent *ext; @@ -637,6 +1061,7 @@ __ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, u32 offset = 0; u8 op = result ? IOAT_OP_PQ_VAL : IOAT_OP_PQ; int i, s, idx, with_ext, num_descs; + int cb32 = (device->version < IOAT_VER_3_3) ? 1 : 0; dev_dbg(to_dev(chan), "%s\n", __func__); /* the engine requires at least two sources (we provide @@ -662,7 +1087,7 @@ __ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, * order. */ if (likely(num_descs) && - ioat2_check_space_lock(ioat, num_descs+1) == 0) + ioat2_check_space_lock(ioat, num_descs + cb32) == 0) idx = ioat->head; else return NULL; @@ -700,6 +1125,9 @@ __ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, pq->q_addr = dst[1] + offset; pq->ctl = 0; pq->ctl_f.op = op; + /* we turn on descriptor write back error status */ + if (device->cap & IOAT_CAP_DWBES) + pq->ctl_f.wb_en = result ? 1 : 0; pq->ctl_f.src_cnt = src_cnt_to_hw(s); pq->ctl_f.p_disable = !!(flags & DMA_PREP_PQ_DISABLE_P); pq->ctl_f.q_disable = !!(flags & DMA_PREP_PQ_DISABLE_Q); @@ -716,26 +1144,140 @@ __ioat3_prep_pq_lock(struct dma_chan *c, enum sum_check_flags *result, pq->ctl_f.fence = !!(flags & DMA_PREP_FENCE); dump_pq_desc_dbg(ioat, desc, ext); - /* completion descriptor carries interrupt bit */ - compl_desc = ioat2_get_ring_ent(ioat, idx + i); - compl_desc->txd.flags = flags & DMA_PREP_INTERRUPT; - hw = compl_desc->hw; - hw->ctl = 0; - hw->ctl_f.null = 1; - hw->ctl_f.int_en = !!(flags & DMA_PREP_INTERRUPT); - hw->ctl_f.compl_write = 1; - hw->size = NULL_DESC_BUFFER_SIZE; - dump_desc_dbg(ioat, compl_desc); + if (!cb32) { + pq->ctl_f.int_en = !!(flags & DMA_PREP_INTERRUPT); + pq->ctl_f.compl_write = 1; + compl_desc = desc; + } else { + /* completion descriptor carries interrupt bit */ + compl_desc = ioat2_get_ring_ent(ioat, idx + i); + compl_desc->txd.flags = flags & DMA_PREP_INTERRUPT; + hw = compl_desc->hw; + hw->ctl = 0; + hw->ctl_f.null = 1; + hw->ctl_f.int_en = !!(flags & DMA_PREP_INTERRUPT); + hw->ctl_f.compl_write = 1; + hw->size = NULL_DESC_BUFFER_SIZE; + dump_desc_dbg(ioat, compl_desc); + } + /* we leave the channel locked to ensure in order submission */ return &compl_desc->txd; } static struct dma_async_tx_descriptor * +__ioat3_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) +{ + struct ioat2_dma_chan *ioat = to_ioat2_chan(c); + struct ioat_chan_common *chan = &ioat->base; + struct ioatdma_device *device = chan->device; + struct ioat_ring_ent *desc; + size_t total_len = len; + struct ioat_pq_descriptor *pq; + u32 offset = 0; + 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; + + dev_dbg(to_dev(chan), "%s\n", __func__); + + num_descs = ioat2_xferlen_to_descs(ioat, len); + + /* + * 16 source pq is only available on cb3.3 and has no completion + * write hw bug. + */ + if (num_descs && ioat2_check_space_lock(ioat, num_descs) == 0) + idx = ioat->head; + else + return NULL; + + i = 0; + + do { + struct ioat_raw_descriptor *descs[4]; + size_t xfer_size = min_t(size_t, len, 1 << ioat->xfercap_log); + + desc = ioat2_get_ring_ent(ioat, idx + i); + pq = desc->pq; + + descs[0] = (struct ioat_raw_descriptor *) pq; + + desc->sed = ioat3_alloc_sed(device, + sed_get_pq16_pool_idx(src_cnt)); + if (!desc->sed) { + dev_err(to_dev(chan), + "%s: no free sed entries\n", __func__); + return NULL; + } + + pq->sed_addr = desc->sed->dma; + desc->sed->parent = desc; + + descs[1] = (struct ioat_raw_descriptor *)desc->sed->hw; + descs[2] = (void *)descs[1] + 64; + + for (s = 0; s < src_cnt; s++) + pq16_set_src(descs, src[s], offset, scf[s], s); + + /* see the comment for dma_maxpq in include/linux/dmaengine.h */ + if (dmaf_p_disabled_continue(flags)) + pq16_set_src(descs, dst[1], offset, 1, s++); + else if (dmaf_continue(flags)) { + pq16_set_src(descs, dst[0], offset, 0, s++); + pq16_set_src(descs, dst[1], offset, 1, s++); + pq16_set_src(descs, dst[1], offset, 0, s++); + } + + pq->size = xfer_size; + pq->p_addr = dst[0] + offset; + pq->q_addr = dst[1] + offset; + pq->ctl = 0; + pq->ctl_f.op = op; + pq->ctl_f.src_cnt = src16_cnt_to_hw(s); + /* we turn on descriptor write back error status */ + if (device->cap & IOAT_CAP_DWBES) + pq->ctl_f.wb_en = result ? 1 : 0; + pq->ctl_f.p_disable = !!(flags & DMA_PREP_PQ_DISABLE_P); + pq->ctl_f.q_disable = !!(flags & DMA_PREP_PQ_DISABLE_Q); + + len -= xfer_size; + offset += xfer_size; + } while (++i < num_descs); + + /* last pq descriptor carries the unmap parameters and fence bit */ + desc->txd.flags = flags; + desc->len = total_len; + if (result) + desc->result = result; + pq->ctl_f.fence = !!(flags & DMA_PREP_FENCE); + + /* with cb3.3 we should be able to do completion w/o a null desc */ + pq->ctl_f.int_en = !!(flags & DMA_PREP_INTERRUPT); + pq->ctl_f.compl_write = 1; + + dump_pq16_desc_dbg(ioat, desc); + + /* we leave the channel locked to ensure in order submission */ + return &desc->txd; +} + +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]; @@ -755,11 +1297,20 @@ ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src, single_source_coef[0] = scf[0]; single_source_coef[1] = 0; - return __ioat3_prep_pq_lock(chan, NULL, dst, single_source, 2, - single_source_coef, len, flags); - } else - return __ioat3_prep_pq_lock(chan, NULL, dst, src, src_cnt, scf, - len, flags); + return (src_cnt > 8) && (dma->max_pq > 8) ? + __ioat3_prep_pq16_lock(chan, NULL, dst, single_source, + 2, single_source_coef, len, + flags) : + __ioat3_prep_pq_lock(chan, NULL, dst, single_source, 2, + single_source_coef, len, flags); + + } else { + return (src_cnt > 8) && (dma->max_pq > 8) ? + __ioat3_prep_pq16_lock(chan, NULL, dst, src, src_cnt, + scf, len, flags) : + __ioat3_prep_pq_lock(chan, NULL, dst, src, src_cnt, + scf, len, flags); + } } struct dma_async_tx_descriptor * @@ -767,6 +1318,8 @@ ioat3_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_device *dma = chan->device; + /* specify valid address for disabled result */ if (flags & DMA_PREP_PQ_DISABLE_P) pq[0] = pq[1]; @@ -778,14 +1331,18 @@ ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src, */ *pqres = 0; - return __ioat3_prep_pq_lock(chan, pqres, pq, src, src_cnt, scf, len, - flags); + return (src_cnt > 8) && (dma->max_pq > 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, + flags); } static struct dma_async_tx_descriptor * 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]; @@ -794,8 +1351,11 @@ ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src, flags |= DMA_PREP_PQ_DISABLE_Q; pq[1] = dst; /* specify valid address for disabled result */ - return __ioat3_prep_pq_lock(chan, NULL, pq, src, src_cnt, scf, len, - flags); + return (src_cnt > 8) && (dma->max_pq > 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, + flags); } struct dma_async_tx_descriptor * @@ -803,6 +1363,7 @@ ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src, unsigned int src_cnt, size_t len, enum sum_check_flags *result, uns |