dmaengine: idxd: Init and probe for Intel data accelerators

The idxd driver introduces the Intel Data Stream Accelerator [1] that will
be available on future Intel Xeon CPUs. One of the kernel access
point for the driver is through the dmaengine subsystem. It will initially
provide the DMA copy service to the kernel.

Some of the main functionality introduced with this accelerator
are: shared virtual memory (SVM) support, and descriptor submission using
Intel CPU instructions movdir64b and enqcmds. There will be additional
accelerator devices that share the same driver with variations to
capabilities.

This commit introduces the probe and initialization component of the
driver.

[1]: https://software.intel.com/en-us/download/intel-data-streaming-accelerator-preliminary-architecture-specification

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/157965023991.73301.6186843973135311580.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>

1 files changed, 468 insertions, 0 deletions

diff --git a/drivers/dma/idxd/init.c b/drivers/dma/idxd/init.c
new file mode 100644
index 000000000000..6e89a87d62b0
--- /dev/null
+++ b/drivers/dma/idxd/init.c
@@ -0,0 +1,468 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright(c) 2019 Intel Corporation. All rights rsvd. */
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/pci.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/workqueue.h>
+#include <linux/aer.h>
+#include <linux/fs.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
+#include <linux/device.h>
+#include <linux/idr.h>
+#include <uapi/linux/idxd.h>
+#include "registers.h"
+#include "idxd.h"
+
+MODULE_VERSION(IDXD_DRIVER_VERSION);
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Intel Corporation");
+
+#define DRV_NAME "idxd"
+
+static struct idr idxd_idrs[IDXD_TYPE_MAX];
+static struct mutex idxd_idr_lock;
+
+static struct pci_device_id idxd_pci_tbl[] = {
+	/* DSA ver 1.0 platforms */
+	{ PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_DSA_SPR0) },
+	{ 0, }
+};
+MODULE_DEVICE_TABLE(pci, idxd_pci_tbl);
+
+static char *idxd_name[] = {
+	"dsa",
+};
+
+const char *idxd_get_dev_name(struct idxd_device *idxd)
+{
+	return idxd_name[idxd->type];
+}
+
+static int idxd_setup_interrupts(struct idxd_device *idxd)
+{
+	struct pci_dev *pdev = idxd->pdev;
+	struct device *dev = &pdev->dev;
+	struct msix_entry *msix;
+	struct idxd_irq_entry *irq_entry;
+	int i, msixcnt;
+	int rc = 0;
+
+	msixcnt = pci_msix_vec_count(pdev);
+	if (msixcnt < 0) {
+		dev_err(dev, "Not MSI-X interrupt capable.\n");
+		goto err_no_irq;
+	}
+
+	idxd->msix_entries = devm_kzalloc(dev, sizeof(struct msix_entry) *
+			msixcnt, GFP_KERNEL);
+	if (!idxd->msix_entries) {
+		rc = -ENOMEM;
+		goto err_no_irq;
+	}
+
+	for (i = 0; i < msixcnt; i++)
+		idxd->msix_entries[i].entry = i;
+
+	rc = pci_enable_msix_exact(pdev, idxd->msix_entries, msixcnt);
+	if (rc) {
+		dev_err(dev, "Failed enabling %d MSIX entries.\n", msixcnt);
+		goto err_no_irq;
+	}
+	dev_dbg(dev, "Enabled %d msix vectors\n", msixcnt);
+
+	/*
+	 * We implement 1 completion list per MSI-X entry except for
+	 * entry 0, which is for errors and others.
+	 */
+	idxd->irq_entries = devm_kcalloc(dev, msixcnt,
+					 sizeof(struct idxd_irq_entry),
+					 GFP_KERNEL);
+	if (!idxd->irq_entries) {
+		rc = -ENOMEM;
+		goto err_no_irq;
+	}
+
+	for (i = 0; i < msixcnt; i++) {
+		idxd->irq_entries[i].id = i;
+		idxd->irq_entries[i].idxd = idxd;
+	}
+
+	msix = &idxd->msix_entries[0];
+	irq_entry = &idxd->irq_entries[0];
+	rc = devm_request_threaded_irq(dev, msix->vector, idxd_irq_handler,
+				       idxd_misc_thread, 0, "idxd-misc",
+				       irq_entry);
+	if (rc < 0) {
+		dev_err(dev, "Failed to allocate misc interrupt.\n");
+		goto err_no_irq;
+	}
+
+	dev_dbg(dev, "Allocated idxd-misc handler on msix vector %d\n",
+		msix->vector);
+
+	/* first MSI-X entry is not for wq interrupts */
+	idxd->num_wq_irqs = msixcnt - 1;
+
+	for (i = 1; i < msixcnt; i++) {
+		msix = &idxd->msix_entries[i];
+		irq_entry = &idxd->irq_entries[i];
+
+		init_llist_head(&idxd->irq_entries[i].pending_llist);
+		INIT_LIST_HEAD(&idxd->irq_entries[i].work_list);
+		rc = devm_request_threaded_irq(dev, msix->vector,
+					       idxd_irq_handler,
+					       idxd_wq_thread, 0,
+					       "idxd-portal", irq_entry);
+		if (rc < 0) {
+			dev_err(dev, "Failed to allocate irq %d.\n",
+				msix->vector);
+			goto err_no_irq;
+		}
+		dev_dbg(dev, "Allocated idxd-msix %d for vector %d\n",
+			i, msix->vector);
+	}
+
+	idxd_unmask_error_interrupts(idxd);
+
+	return 0;
+
+ err_no_irq:
+	/* Disable error interrupt generation */
+	idxd_mask_error_interrupts(idxd);
+	pci_disable_msix(pdev);
+	dev_err(dev, "No usable interrupts\n");
+	return rc;
+}
+
+static void idxd_wqs_free_lock(struct idxd_device *idxd)
+{
+	int i;
+
+	for (i = 0; i < idxd->max_wqs; i++) {
+		struct idxd_wq *wq = &idxd->wqs[i];
+
+		percpu_free_rwsem(&wq->submit_lock);
+	}
+}
+
+static int idxd_setup_internals(struct idxd_device *idxd)
+{
+	struct device *dev = &idxd->pdev->dev;
+	int i;
+
+	idxd->groups = devm_kcalloc(dev, idxd->max_groups,
+				    sizeof(struct idxd_group), GFP_KERNEL);
+	if (!idxd->groups)
+		return -ENOMEM;
+
+	for (i = 0; i < idxd->max_groups; i++) {
+		idxd->groups[i].idxd = idxd;
+		idxd->groups[i].id = i;
+		idxd->groups[i].tc_a = -1;
+		idxd->groups[i].tc_b = -1;
+	}
+
+	idxd->wqs = devm_kcalloc(dev, idxd->max_wqs, sizeof(struct idxd_wq),
+				 GFP_KERNEL);
+	if (!idxd->wqs)
+		return -ENOMEM;
+
+	idxd->engines = devm_kcalloc(dev, idxd->max_engines,
+				     sizeof(struct idxd_engine), GFP_KERNEL);
+	if (!idxd->engines)
+		return -ENOMEM;
+
+	for (i = 0; i < idxd->max_wqs; i++) {
+		struct idxd_wq *wq = &idxd->wqs[i];
+		int rc;
+
+		wq->id = i;
+		wq->idxd = idxd;
+		mutex_init(&wq->wq_lock);
+		atomic_set(&wq->dq_count, 0);
+		init_waitqueue_head(&wq->submit_waitq);
+		rc = percpu_init_rwsem(&wq->submit_lock);
+		if (rc < 0) {
+			idxd_wqs_free_lock(idxd);
+			return rc;
+		}
+	}
+
+	for (i = 0; i < idxd->max_engines; i++) {
+		idxd->engines[i].idxd = idxd;
+		idxd->engines[i].id = i;
+	}
+
+	return 0;
+}
+
+static void idxd_read_table_offsets(struct idxd_device *idxd)
+{
+	union offsets_reg offsets;
+	struct device *dev = &idxd->pdev->dev;
+
+	offsets.bits[0] = ioread64(idxd->reg_base + IDXD_TABLE_OFFSET);
+	offsets.bits[1] = ioread64(idxd->reg_base + IDXD_TABLE_OFFSET
+			+ sizeof(u64));
+	idxd->grpcfg_offset = offsets.grpcfg * 0x100;
+	dev_dbg(dev, "IDXD Group Config Offset: %#x\n", idxd->grpcfg_offset);
+	idxd->wqcfg_offset = offsets.wqcfg * 0x100;
+	dev_dbg(dev, "IDXD Work Queue Config Offset: %#x\n",
+		idxd->wqcfg_offset);
+	idxd->msix_perm_offset = offsets.msix_perm * 0x100;
+	dev_dbg(dev, "IDXD MSIX Permission Offset: %#x\n",
+		idxd->msix_perm_offset);
+	idxd->perfmon_offset = offsets.perfmon * 0x100;
+	dev_dbg(dev, "IDXD Perfmon Offset: %#x\n", idxd->perfmon_offset);
+}
+
+static void idxd_read_caps(struct idxd_device *idxd)
+{
+	struct device *dev = &idxd->pdev->dev;
+	int i;
+
+	/* reading generic capabilities */
+	idxd->hw.gen_cap.bits = ioread64(idxd->reg_base + IDXD_GENCAP_OFFSET);
+	dev_dbg(dev, "gen_cap: %#llx\n", idxd->hw.gen_cap.bits);
+	idxd->max_xfer_bytes = 1ULL << idxd->hw.gen_cap.max_xfer_shift;
+	dev_dbg(dev, "max xfer size: %llu bytes\n", idxd->max_xfer_bytes);
+	idxd->max_batch_size = 1U << idxd->hw.gen_cap.max_batch_shift;
+	dev_dbg(dev, "max batch size: %u\n", idxd->max_batch_size);
+	if (idxd->hw.gen_cap.config_en)
+		set_bit(IDXD_FLAG_CONFIGURABLE, &idxd->flags);
+
+	/* reading group capabilities */
+	idxd->hw.group_cap.bits =
+		ioread64(idxd->reg_base + IDXD_GRPCAP_OFFSET);
+	dev_dbg(dev, "group_cap: %#llx\n", idxd->hw.group_cap.bits);
+	idxd->max_groups = idxd->hw.group_cap.num_groups;
+	dev_dbg(dev, "max groups: %u\n", idxd->max_groups);
+	idxd->max_tokens = idxd->hw.group_cap.total_tokens;
+	dev_dbg(dev, "max tokens: %u\n", idxd->max_tokens);
+
+	/* read engine capabilities */
+	idxd->hw.engine_cap.bits =
+		ioread64(idxd->reg_base + IDXD_ENGCAP_OFFSET);
+	dev_dbg(dev, "engine_cap: %#llx\n", idxd->hw.engine_cap.bits);
+	idxd->max_engines = idxd->hw.engine_cap.num_engines;
+	dev_dbg(dev, "max engines: %u\n", idxd->max_engines);
+
+	/* read workqueue capabilities */
+	idxd->hw.wq_cap.bits = ioread64(idxd->reg_base + IDXD_WQCAP_OFFSET);
+	dev_dbg(dev, "wq_cap: %#llx\n", idxd->hw.wq_cap.bits);
+	idxd->max_wq_size = idxd->hw.wq_cap.total_wq_size;
+	dev_dbg(dev, "total workqueue size: %u\n", idxd->max_wq_size);
+	idxd->max_wqs = idxd->hw.wq_cap.num_wqs;
+	dev_dbg(dev, "max workqueues: %u\n", idxd->max_wqs);
+
+	/* reading operation capabilities */
+	for (i = 0; i < 4; i++) {
+		idxd->hw.opcap.bits[i] = ioread64(idxd->reg_base +
+				IDXD_OPCAP_OFFSET + i * sizeof(u64));
+		dev_dbg(dev, "opcap[%d]: %#llx\n", i, idxd->hw.opcap.bits[i]);
+	}
+}
+
+static struct idxd_device *idxd_alloc(struct pci_dev *pdev,
+				      void __iomem * const *iomap)
+{
+	struct device *dev = &pdev->dev;
+	struct idxd_device *idxd;
+
+	idxd = devm_kzalloc(dev, sizeof(struct idxd_device), GFP_KERNEL);
+	if (!idxd)
+		return NULL;
+
+	idxd->pdev = pdev;
+	idxd->reg_base = iomap[IDXD_MMIO_BAR];
+	spin_lock_init(&idxd->dev_lock);
+
+	return idxd;
+}
+
+static int idxd_probe(struct idxd_device *idxd)
+{
+	struct pci_dev *pdev = idxd->pdev;
+	struct device *dev = &pdev->dev;
+	int rc;
+
+	dev_dbg(dev, "%s entered and resetting device\n", __func__);
+	rc = idxd_device_reset(idxd);
+	if (rc < 0)
+		return rc;
+	dev_dbg(dev, "IDXD reset complete\n");
+
+	idxd_read_caps(idxd);
+	idxd_read_table_offsets(idxd);
+
+	rc = idxd_setup_internals(idxd);
+	if (rc)
+		goto err_setup;
+
+	rc = idxd_setup_interrupts(idxd);
+	if (rc)
+		goto err_setup;
+
+	dev_dbg(dev, "IDXD interrupt setup complete.\n");
+
+	mutex_lock(&idxd_idr_lock);
+	idxd->id = idr_alloc(&idxd_idrs[idxd->type], idxd, 0, 0, GFP_KERNEL);
+	mutex_unlock(&idxd_idr_lock);
+	if (idxd->id < 0) {
+		rc = -ENOMEM;
+		goto err_idr_fail;
+	}
+
+	dev_dbg(dev, "IDXD device %d probed successfully\n", idxd->id);
+	return 0;
+
+ err_idr_fail:
+	idxd_mask_error_interrupts(idxd);
+	idxd_mask_msix_vectors(idxd);
+ err_setup:
+	return rc;
+}
+
+static int idxd_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
+{
+	void __iomem * const *iomap;
+	struct device *dev = &pdev->dev;
+	struct idxd_device *idxd;
+	int rc;
+	unsigned int mask;
+
+	rc = pcim_enable_device(pdev);
+	if (rc)
+		return rc;
+
+	dev_dbg(dev, "Mapping BARs\n");
+	mask = (1 << IDXD_MMIO_BAR);
+	rc = pcim_iomap_regions(pdev, mask, DRV_NAME);
+	if (rc)
+		return rc;
+
+	iomap = pcim_iomap_table(pdev);
+	if (!iomap)
+		return -ENOMEM;
+
+	dev_dbg(dev, "Set DMA masks\n");
+	rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
+	if (rc)
+		rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (rc)
+		return rc;
+
+	rc = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+	if (rc)
+		rc = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (rc)
+		return rc;
+
+	dev_dbg(dev, "Alloc IDXD context\n");
+	idxd = idxd_alloc(pdev, iomap);
+	if (!idxd)
+		return -ENOMEM;
+
+	idxd_set_type(idxd);
+
+	dev_dbg(dev, "Set PCI master\n");
+	pci_set_master(pdev);
+	pci_set_drvdata(pdev, idxd);
+
+	idxd->hw.version = ioread32(idxd->reg_base + IDXD_VER_OFFSET);
+	rc = idxd_probe(idxd);
+	if (rc) {
+		dev_err(dev, "Intel(R) IDXD DMA Engine init failed\n");
+		return -ENODEV;
+	}
+
+	dev_info(&pdev->dev, "Intel(R) Accelerator Device (v%x)\n",
+		 idxd->hw.version);
+
+	return 0;
+}
+
+static void idxd_shutdown(struct pci_dev *pdev)
+{
+	struct idxd_device *idxd = pci_get_drvdata(pdev);
+	int rc, i;
+	struct idxd_irq_entry *irq_entry;
+	int msixcnt = pci_msix_vec_count(pdev);
+	unsigned long flags;
+
+	spin_lock_irqsave(&idxd->dev_lock, flags);
+	rc = idxd_device_disable(idxd);
+	spin_unlock_irqrestore(&idxd->dev_lock, flags);
+	if (rc)
+		dev_err(&pdev->dev, "Disabling device failed\n");
+
+	dev_dbg(&pdev->dev, "%s called\n", __func__);
+	idxd_mask_msix_vectors(idxd);
+	idxd_mask_error_interrupts(idxd);
+
+	for (i = 0; i < msixcnt; i++) {
+		irq_entry = &idxd->irq_entries[i];
+		synchronize_irq(idxd->msix_entries[i].vector);
+		if (i == 0)
+			continue;
+	}
+}
+
+static void idxd_remove(struct pci_dev *pdev)
+{
+	struct idxd_device *idxd = pci_get_drvdata(pdev);
+
+	dev_dbg(&pdev->dev, "%s called\n", __func__);
+	idxd_shutdown(pdev);
+	idxd_wqs_free_lock(idxd);
+	mutex_lock(&idxd_idr_lock);
+	idr_remove(&idxd_idrs[idxd->type], idxd->id);
+	mutex_unlock(&idxd_idr_lock);
+}
+
+static struct pci_driver idxd_pci_driver = {
+	.name		= DRV_NAME,
+	.id_table	= idxd_pci_tbl,
+	.probe		= idxd_pci_probe,
+	.remove		= idxd_remove,
+	.shutdown	= idxd_shutdown,
+};
+
+static int __init idxd_init_module(void)
+{
+	int err, i;
+
+	/*
+	 * If the CPU does not support write512, there's no point in
+	 * enumerating the device. We can not utilize it.
+	 */
+	if (!boot_cpu_has(X86_FEATURE_MOVDIR64B)) {
+		pr_warn("idxd driver failed to load without MOVDIR64B.\n");
+		return -ENODEV;
+	}
+
+	pr_info("%s: Intel(R) Accelerator Devices Driver %s\n",
+		DRV_NAME, IDXD_DRIVER_VERSION);
+
+	mutex_init(&idxd_idr_lock);
+	for (i = 0; i < IDXD_TYPE_MAX; i++)
+		idr_init(&idxd_idrs[i]);
+
+	err = pci_register_driver(&idxd_pci_driver);
+	if (err)
+		return err;
+
+	return 0;
+}
+module_init(idxd_init_module);
+
+static void __exit idxd_exit_module(void)
+{
+	pci_unregister_driver(&idxd_pci_driver);
+}
+module_exit(idxd_exit_module);