X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fvirtio%2Fvirtio_pci.c;h=1e1775720fbc73bc7cc5e3b1e29e8d18517fca96;hb=a60a0c15076224b664668c09a96922f9a2c6b142;hp=39610775d6aeb2cab12aa1a224536041981d88bc;hpb=3891f233f76270e00e45addf221bc280c585a1d9;p=dpdk.git

diff --git a/drivers/net/virtio/virtio_pci.c b/drivers/net/virtio/virtio_pci.c
index 39610775d6..1e1775720f 100644
--- a/drivers/net/virtio/virtio_pci.c
+++ b/drivers/net/virtio/virtio_pci.c
@@ -37,60 +37,134 @@
  #include <fcntl.h>
 #endif
 
+#include <rte_io.h>
+
 #include "virtio_pci.h"
 #include "virtio_logs.h"
 #include "virtqueue.h"
 
+/*
+ * Following macros are derived from linux/pci_regs.h, however,
+ * we can't simply include that header here, as there is no such
+ * file for non-Linux platform.
+ */
+#define PCI_CAPABILITY_LIST	0x34
+#define PCI_CAP_ID_VNDR		0x09
+#define PCI_CAP_ID_MSIX		0x11
+
+/*
+ * The remaining space is defined by each driver as the per-driver
+ * configuration space.
+ */
+#define VIRTIO_PCI_CONFIG(hw) (((hw)->use_msix) ? 24 : 20)
+
+static inline int
+check_vq_phys_addr_ok(struct virtqueue *vq)
+{
+	/* Virtio PCI device VIRTIO_PCI_QUEUE_PF register is 32bit,
+	 * and only accepts 32 bit page frame number.
+	 * Check if the allocated physical memory exceeds 16TB.
+	 */
+	if ((vq->vq_ring_mem + vq->vq_ring_size - 1) >>
+			(VIRTIO_PCI_QUEUE_ADDR_SHIFT + 32)) {
+		PMD_INIT_LOG(ERR, "vring address shouldn't be above 16TB!");
+		return 0;
+	}
+
+	return 1;
+}
+
+/*
+ * Since we are in legacy mode:
+ * http://ozlabs.org/~rusty/virtio-spec/virtio-0.9.5.pdf
+ *
+ * "Note that this is possible because while the virtio header is PCI (i.e.
+ * little) endian, the device-specific region is encoded in the native endian of
+ * the guest (where such distinction is applicable)."
+ *
+ * For powerpc which supports both, qemu supposes that cpu is big endian and
+ * enforces this for the virtio-net stuff.
+ */
 static void
 legacy_read_dev_config(struct virtio_hw *hw, size_t offset,
 		       void *dst, int length)
 {
-	uint64_t off;
-	uint8_t *d;
+#ifdef RTE_ARCH_PPC_64
 	int size;
 
-	off = VIRTIO_PCI_CONFIG(hw) + offset;
-	for (d = dst; length > 0; d += size, off += size, length -= size) {
+	while (length > 0) {
 		if (length >= 4) {
 			size = 4;
-			*(uint32_t *)d = VIRTIO_READ_REG_4(hw, off);
+			rte_eal_pci_ioport_read(VTPCI_IO(hw), dst, size,
+				VIRTIO_PCI_CONFIG(hw) + offset);
+			*(uint32_t *)dst = rte_be_to_cpu_32(*(uint32_t *)dst);
 		} else if (length >= 2) {
 			size = 2;
-			*(uint16_t *)d = VIRTIO_READ_REG_2(hw, off);
+			rte_eal_pci_ioport_read(VTPCI_IO(hw), dst, size,
+				VIRTIO_PCI_CONFIG(hw) + offset);
+			*(uint16_t *)dst = rte_be_to_cpu_16(*(uint16_t *)dst);
 		} else {
 			size = 1;
-			*d = VIRTIO_READ_REG_1(hw, off);
+			rte_eal_pci_ioport_read(VTPCI_IO(hw), dst, size,
+				VIRTIO_PCI_CONFIG(hw) + offset);
 		}
+
+		dst = (char *)dst + size;
+		offset += size;
+		length -= size;
 	}
+#else
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), dst, length,
+				VIRTIO_PCI_CONFIG(hw) + offset);
+#endif
 }
 
 static void
 legacy_write_dev_config(struct virtio_hw *hw, size_t offset,
 			const void *src, int length)
 {
-	uint64_t off;
-	const uint8_t *s;
+#ifdef RTE_ARCH_PPC_64
+	union {
+		uint32_t u32;
+		uint16_t u16;
+	} tmp;
 	int size;
 
-	off = VIRTIO_PCI_CONFIG(hw) + offset;
-	for (s = src; length > 0; s += size, off += size, length -= size) {
+	while (length > 0) {
 		if (length >= 4) {
 			size = 4;
-			VIRTIO_WRITE_REG_4(hw, off, *(const uint32_t *)s);
+			tmp.u32 = rte_cpu_to_be_32(*(const uint32_t *)src);
+			rte_eal_pci_ioport_write(VTPCI_IO(hw), &tmp.u32, size,
+				VIRTIO_PCI_CONFIG(hw) + offset);
 		} else if (length >= 2) {
 			size = 2;
-			VIRTIO_WRITE_REG_2(hw, off, *(const uint16_t *)s);
+			tmp.u16 = rte_cpu_to_be_16(*(const uint16_t *)src);
+			rte_eal_pci_ioport_write(VTPCI_IO(hw), &tmp.u16, size,
+				VIRTIO_PCI_CONFIG(hw) + offset);
 		} else {
 			size = 1;
-			VIRTIO_WRITE_REG_1(hw, off, *s);
+			rte_eal_pci_ioport_write(VTPCI_IO(hw), src, size,
+				VIRTIO_PCI_CONFIG(hw) + offset);
 		}
+
+		src = (const char *)src + size;
+		offset += size;
+		length -= size;
 	}
+#else
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), src, length,
+				 VIRTIO_PCI_CONFIG(hw) + offset);
+#endif
 }
 
 static uint64_t
 legacy_get_features(struct virtio_hw *hw)
 {
-	return VIRTIO_READ_REG_4(hw, VIRTIO_PCI_HOST_FEATURES);
+	uint32_t dst;
+
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), &dst, 4,
+				VIRTIO_PCI_HOST_FEATURES);
+	return dst;
 }
 
 static void
@@ -101,19 +175,23 @@ legacy_set_features(struct virtio_hw *hw, uint64_t features)
 			"only 32 bit features are allowed for legacy virtio!");
 		return;
 	}
-	VIRTIO_WRITE_REG_4(hw, VIRTIO_PCI_GUEST_FEATURES, features);
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &features, 4,
+				 VIRTIO_PCI_GUEST_FEATURES);
 }
 
 static uint8_t
 legacy_get_status(struct virtio_hw *hw)
 {
-	return VIRTIO_READ_REG_1(hw, VIRTIO_PCI_STATUS);
+	uint8_t dst;
+
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), &dst, 1, VIRTIO_PCI_STATUS);
+	return dst;
 }
 
 static void
 legacy_set_status(struct virtio_hw *hw, uint8_t status)
 {
-	VIRTIO_WRITE_REG_1(hw, VIRTIO_PCI_STATUS, status);
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &status, 1, VIRTIO_PCI_STATUS);
 }
 
 static void
@@ -125,322 +203,311 @@ legacy_reset(struct virtio_hw *hw)
 static uint8_t
 legacy_get_isr(struct virtio_hw *hw)
 {
-	return VIRTIO_READ_REG_1(hw, VIRTIO_PCI_ISR);
+	uint8_t dst;
+
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), &dst, 1, VIRTIO_PCI_ISR);
+	return dst;
 }
 
 /* Enable one vector (0) for Link State Intrerrupt */
 static uint16_t
 legacy_set_config_irq(struct virtio_hw *hw, uint16_t vec)
 {
-	VIRTIO_WRITE_REG_2(hw, VIRTIO_MSI_CONFIG_VECTOR, vec);
-	return VIRTIO_READ_REG_2(hw, VIRTIO_MSI_CONFIG_VECTOR);
+	uint16_t dst;
+
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &vec, 2,
+				 VIRTIO_MSI_CONFIG_VECTOR);
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), &dst, 2,
+				VIRTIO_MSI_CONFIG_VECTOR);
+	return dst;
+}
+
+static uint16_t
+legacy_set_queue_irq(struct virtio_hw *hw, struct virtqueue *vq, uint16_t vec)
+{
+	uint16_t dst;
+
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &vq->vq_queue_index, 2,
+				 VIRTIO_PCI_QUEUE_SEL);
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &vec, 2,
+				 VIRTIO_MSI_QUEUE_VECTOR);
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), &dst, 2, VIRTIO_MSI_QUEUE_VECTOR);
+	return dst;
 }
 
 static uint16_t
 legacy_get_queue_num(struct virtio_hw *hw, uint16_t queue_id)
 {
-	VIRTIO_WRITE_REG_2(hw, VIRTIO_PCI_QUEUE_SEL, queue_id);
-	return VIRTIO_READ_REG_2(hw, VIRTIO_PCI_QUEUE_NUM);
+	uint16_t dst;
+
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &queue_id, 2,
+				 VIRTIO_PCI_QUEUE_SEL);
+	rte_eal_pci_ioport_read(VTPCI_IO(hw), &dst, 2, VIRTIO_PCI_QUEUE_NUM);
+	return dst;
 }
 
-static void
+static int
 legacy_setup_queue(struct virtio_hw *hw, struct virtqueue *vq)
 {
-	VIRTIO_WRITE_REG_2(hw, VIRTIO_PCI_QUEUE_SEL, vq->vq_queue_index);
+	uint32_t src;
+
+	if (!check_vq_phys_addr_ok(vq))
+		return -1;
+
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &vq->vq_queue_index, 2,
+			 VIRTIO_PCI_QUEUE_SEL);
+	src = vq->vq_ring_mem >> VIRTIO_PCI_QUEUE_ADDR_SHIFT;
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &src, 4, VIRTIO_PCI_QUEUE_PFN);
 
-	VIRTIO_WRITE_REG_4(hw, VIRTIO_PCI_QUEUE_PFN,
-		vq->mz->phys_addr >> VIRTIO_PCI_QUEUE_ADDR_SHIFT);
+	return 0;
 }
 
 static void
 legacy_del_queue(struct virtio_hw *hw, struct virtqueue *vq)
 {
-	VIRTIO_WRITE_REG_2(hw, VIRTIO_PCI_QUEUE_SEL, vq->vq_queue_index);
+	uint32_t src = 0;
 
-	VIRTIO_WRITE_REG_4(hw, VIRTIO_PCI_QUEUE_PFN, 0);
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &vq->vq_queue_index, 2,
+			 VIRTIO_PCI_QUEUE_SEL);
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &src, 4, VIRTIO_PCI_QUEUE_PFN);
 }
 
 static void
 legacy_notify_queue(struct virtio_hw *hw, struct virtqueue *vq)
 {
-	VIRTIO_WRITE_REG_2(hw, VIRTIO_PCI_QUEUE_NOTIFY, vq->vq_queue_index);
+	rte_eal_pci_ioport_write(VTPCI_IO(hw), &vq->vq_queue_index, 2,
+			 VIRTIO_PCI_QUEUE_NOTIFY);
 }
 
 #ifdef RTE_EXEC_ENV_LINUXAPP
 static int
-parse_sysfs_value(const char *filename, unsigned long *val)
+legacy_virtio_has_msix(const struct rte_pci_addr *loc)
 {
-	FILE *f;
-	char buf[BUFSIZ];
-	char *end = NULL;
+	DIR *d;
+	char dirname[PATH_MAX];
 
-	f = fopen(filename, "r");
-	if (f == NULL) {
-		PMD_INIT_LOG(ERR, "%s(): cannot open sysfs value %s",
-			     __func__, filename);
-		return -1;
-	}
+	snprintf(dirname, sizeof(dirname),
+		     "%s/" PCI_PRI_FMT "/msi_irqs", pci_get_sysfs_path(),
+		     loc->domain, loc->bus, loc->devid, loc->function);
 
-	if (fgets(buf, sizeof(buf), f) == NULL) {
-		PMD_INIT_LOG(ERR, "%s(): cannot read sysfs value %s",
-			     __func__, filename);
-		fclose(f);
-		return -1;
-	}
-	*val = strtoul(buf, &end, 0);
-	if ((buf[0] == '\0') || (end == NULL) || (*end != '\n')) {
-		PMD_INIT_LOG(ERR, "%s(): cannot parse sysfs value %s",
-			     __func__, filename);
-		fclose(f);
-		return -1;
-	}
-	fclose(f);
+	d = opendir(dirname);
+	if (d)
+		closedir(d);
+
+	return d != NULL;
+}
+#else
+static int
+legacy_virtio_has_msix(const struct rte_pci_addr *loc __rte_unused)
+{
+	/* nic_uio does not enable interrupts, return 0 (false). */
 	return 0;
 }
+#endif
 
-static int
-get_uio_dev(struct rte_pci_addr *loc, char *buf, unsigned int buflen,
-			unsigned int *uio_num)
+const struct virtio_pci_ops legacy_ops = {
+	.read_dev_cfg	= legacy_read_dev_config,
+	.write_dev_cfg	= legacy_write_dev_config,
+	.reset		= legacy_reset,
+	.get_status	= legacy_get_status,
+	.set_status	= legacy_set_status,
+	.get_features	= legacy_get_features,
+	.set_features	= legacy_set_features,
+	.get_isr	= legacy_get_isr,
+	.set_config_irq	= legacy_set_config_irq,
+	.set_queue_irq  = legacy_set_queue_irq,
+	.get_queue_num	= legacy_get_queue_num,
+	.setup_queue	= legacy_setup_queue,
+	.del_queue	= legacy_del_queue,
+	.notify_queue	= legacy_notify_queue,
+};
+
+static inline void
+io_write64_twopart(uint64_t val, uint32_t *lo, uint32_t *hi)
 {
-	struct dirent *e;
-	DIR *dir;
-	char dirname[PATH_MAX];
+	rte_write32(val & ((1ULL << 32) - 1), lo);
+	rte_write32(val >> 32,		     hi);
+}
 
-	/*
-	 * depending on kernel version, uio can be located in uio/uioX
-	 * or uio:uioX
-	 */
-	snprintf(dirname, sizeof(dirname),
-		     SYSFS_PCI_DEVICES "/" PCI_PRI_FMT "/uio",
-		     loc->domain, loc->bus, loc->devid, loc->function);
-	dir = opendir(dirname);
-	if (dir == NULL) {
-		/* retry with the parent directory */
-		snprintf(dirname, sizeof(dirname),
-			     SYSFS_PCI_DEVICES "/" PCI_PRI_FMT,
-			     loc->domain, loc->bus, loc->devid, loc->function);
-		dir = opendir(dirname);
-
-		if (dir == NULL) {
-			PMD_INIT_LOG(ERR, "Cannot opendir %s", dirname);
-			return -1;
-		}
-	}
+static void
+modern_read_dev_config(struct virtio_hw *hw, size_t offset,
+		       void *dst, int length)
+{
+	int i;
+	uint8_t *p;
+	uint8_t old_gen, new_gen;
 
-	/* take the first file starting with "uio" */
-	while ((e = readdir(dir)) != NULL) {
-		/* format could be uio%d ...*/
-		int shortprefix_len = sizeof("uio") - 1;
-		/* ... or uio:uio%d */
-		int longprefix_len = sizeof("uio:uio") - 1;
-		char *endptr;
-
-		if (strncmp(e->d_name, "uio", 3) != 0)
-			continue;
-
-		/* first try uio%d */
-		errno = 0;
-		*uio_num = strtoull(e->d_name + shortprefix_len, &endptr, 10);
-		if (errno == 0 && endptr != (e->d_name + shortprefix_len)) {
-			snprintf(buf, buflen, "%s/uio%u", dirname, *uio_num);
-			break;
-		}
+	do {
+		old_gen = rte_read8(&hw->common_cfg->config_generation);
 
-		/* then try uio:uio%d */
-		errno = 0;
-		*uio_num = strtoull(e->d_name + longprefix_len, &endptr, 10);
-		if (errno == 0 && endptr != (e->d_name + longprefix_len)) {
-			snprintf(buf, buflen, "%s/uio:uio%u", dirname,
-				     *uio_num);
-			break;
-		}
-	}
-	closedir(dir);
+		p = dst;
+		for (i = 0;  i < length; i++)
+			*p++ = rte_read8((uint8_t *)hw->dev_cfg + offset + i);
 
-	/* No uio resource found */
-	if (e == NULL) {
-		PMD_INIT_LOG(ERR, "Could not find uio resource");
-		return -1;
-	}
+		new_gen = rte_read8(&hw->common_cfg->config_generation);
+	} while (old_gen != new_gen);
+}
 
-	return 0;
+static void
+modern_write_dev_config(struct virtio_hw *hw, size_t offset,
+			const void *src, int length)
+{
+	int i;
+	const uint8_t *p = src;
+
+	for (i = 0;  i < length; i++)
+		rte_write8((*p++), (((uint8_t *)hw->dev_cfg) + offset + i));
 }
 
-static int
-legacy_virtio_has_msix(const struct rte_pci_addr *loc)
+static uint64_t
+modern_get_features(struct virtio_hw *hw)
 {
-	DIR *d;
-	char dirname[PATH_MAX];
+	uint32_t features_lo, features_hi;
 
-	snprintf(dirname, sizeof(dirname),
-		     SYSFS_PCI_DEVICES "/" PCI_PRI_FMT "/msi_irqs",
-		     loc->domain, loc->bus, loc->devid, loc->function);
+	rte_write32(0, &hw->common_cfg->device_feature_select);
+	features_lo = rte_read32(&hw->common_cfg->device_feature);
 
-	d = opendir(dirname);
-	if (d)
-		closedir(d);
+	rte_write32(1, &hw->common_cfg->device_feature_select);
+	features_hi = rte_read32(&hw->common_cfg->device_feature);
 
-	return (d != NULL);
+	return ((uint64_t)features_hi << 32) | features_lo;
 }
 
-/* Extract I/O port numbers from sysfs */
-static int
-virtio_resource_init_by_uio(struct rte_pci_device *pci_dev)
+static void
+modern_set_features(struct virtio_hw *hw, uint64_t features)
 {
-	char dirname[PATH_MAX];
-	char filename[PATH_MAX];
-	unsigned long start, size;
-	unsigned int uio_num;
-
-	if (get_uio_dev(&pci_dev->addr, dirname, sizeof(dirname), &uio_num) < 0)
-		return -1;
+	rte_write32(0, &hw->common_cfg->guest_feature_select);
+	rte_write32(features & ((1ULL << 32) - 1),
+		    &hw->common_cfg->guest_feature);
 
-	/* get portio size */
-	snprintf(filename, sizeof(filename),
-		     "%s/portio/port0/size", dirname);
-	if (parse_sysfs_value(filename, &size) < 0) {
-		PMD_INIT_LOG(ERR, "%s(): cannot parse size",
-			     __func__);
-		return -1;
-	}
+	rte_write32(1, &hw->common_cfg->guest_feature_select);
+	rte_write32(features >> 32,
+		    &hw->common_cfg->guest_feature);
+}
 
-	/* get portio start */
-	snprintf(filename, sizeof(filename),
-		 "%s/portio/port0/start", dirname);
-	if (parse_sysfs_value(filename, &start) < 0) {
-		PMD_INIT_LOG(ERR, "%s(): cannot parse portio start",
-			     __func__);
-		return -1;
-	}
-	pci_dev->mem_resource[0].addr = (void *)(uintptr_t)start;
-	pci_dev->mem_resource[0].len =  (uint64_t)size;
-	PMD_INIT_LOG(DEBUG,
-		     "PCI Port IO found start=0x%lx with size=0x%lx",
-		     start, size);
-
-	/* save fd */
-	memset(dirname, 0, sizeof(dirname));
-	snprintf(dirname, sizeof(dirname), "/dev/uio%u", uio_num);
-	pci_dev->intr_handle.fd = open(dirname, O_RDWR);
-	if (pci_dev->intr_handle.fd < 0) {
-		PMD_INIT_LOG(ERR, "Cannot open %s: %s\n",
-			dirname, strerror(errno));
-		return -1;
-	}
+static uint8_t
+modern_get_status(struct virtio_hw *hw)
+{
+	return rte_read8(&hw->common_cfg->device_status);
+}
 
-	pci_dev->intr_handle.type = RTE_INTR_HANDLE_UIO;
-	pci_dev->driver->drv_flags |= RTE_PCI_DRV_INTR_LSC;
+static void
+modern_set_status(struct virtio_hw *hw, uint8_t status)
+{
+	rte_write8(status, &hw->common_cfg->device_status);
+}
 
-	return 0;
+static void
+modern_reset(struct virtio_hw *hw)
+{
+	modern_set_status(hw, VIRTIO_CONFIG_STATUS_RESET);
+	modern_get_status(hw);
 }
 
-/* Extract port I/O numbers from proc/ioports */
-static int
-virtio_resource_init_by_ioports(struct rte_pci_device *pci_dev)
+static uint8_t
+modern_get_isr(struct virtio_hw *hw)
 {
-	uint16_t start, end;
-	int size;
-	FILE *fp;
-	char *line = NULL;
-	char pci_id[16];
-	int found = 0;
-	size_t linesz;
-
-	snprintf(pci_id, sizeof(pci_id), PCI_PRI_FMT,
-		 pci_dev->addr.domain,
-		 pci_dev->addr.bus,
-		 pci_dev->addr.devid,
-		 pci_dev->addr.function);
-
-	fp = fopen("/proc/ioports", "r");
-	if (fp == NULL) {
-		PMD_INIT_LOG(ERR, "%s(): can't open ioports", __func__);
-		return -1;
-	}
+	return rte_read8(hw->isr);
+}
 
-	while (getdelim(&line, &linesz, '\n', fp) > 0) {
-		char *ptr = line;
-		char *left;
-		int n;
+static uint16_t
+modern_set_config_irq(struct virtio_hw *hw, uint16_t vec)
+{
+	rte_write16(vec, &hw->common_cfg->msix_config);
+	return rte_read16(&hw->common_cfg->msix_config);
+}
 
-		n = strcspn(ptr, ":");
-		ptr[n] = 0;
-		left = &ptr[n + 1];
+static uint16_t
+modern_set_queue_irq(struct virtio_hw *hw, struct virtqueue *vq, uint16_t vec)
+{
+	rte_write16(vq->vq_queue_index, &hw->common_cfg->queue_select);
+	rte_write16(vec, &hw->common_cfg->queue_msix_vector);
+	return rte_read16(&hw->common_cfg->queue_msix_vector);
+}
 
-		while (*left && isspace(*left))
-			left++;
+static uint16_t
+modern_get_queue_num(struct virtio_hw *hw, uint16_t queue_id)
+{
+	rte_write16(queue_id, &hw->common_cfg->queue_select);
+	return rte_read16(&hw->common_cfg->queue_size);
+}
 
-		if (!strncmp(left, pci_id, strlen(pci_id))) {
-			found = 1;
+static int
+modern_setup_queue(struct virtio_hw *hw, struct virtqueue *vq)
+{
+	uint64_t desc_addr, avail_addr, used_addr;
+	uint16_t notify_off;
 
-			while (*ptr && isspace(*ptr))
-				ptr++;
+	if (!check_vq_phys_addr_ok(vq))
+		return -1;
 
-			sscanf(ptr, "%04hx-%04hx", &start, &end);
-			size = end - start + 1;
+	desc_addr = vq->vq_ring_mem;
+	avail_addr = desc_addr + vq->vq_nentries * sizeof(struct vring_desc);
+	used_addr = RTE_ALIGN_CEIL(avail_addr + offsetof(struct vring_avail,
+							 ring[vq->vq_nentries]),
+				   VIRTIO_PCI_VRING_ALIGN);
 
-			break;
-		}
-	}
+	rte_write16(vq->vq_queue_index, &hw->common_cfg->queue_select);
 
-	free(line);
-	fclose(fp);
+	io_write64_twopart(desc_addr, &hw->common_cfg->queue_desc_lo,
+				      &hw->common_cfg->queue_desc_hi);
+	io_write64_twopart(avail_addr, &hw->common_cfg->queue_avail_lo,
+				       &hw->common_cfg->queue_avail_hi);
+	io_write64_twopart(used_addr, &hw->common_cfg->queue_used_lo,
+				      &hw->common_cfg->queue_used_hi);
 
-	if (!found)
-		return -1;
+	notify_off = rte_read16(&hw->common_cfg->queue_notify_off);
+	vq->notify_addr = (void *)((uint8_t *)hw->notify_base +
+				notify_off * hw->notify_off_multiplier);
 
-	pci_dev->mem_resource[0].addr = (void *)(uintptr_t)(uint32_t)start;
-	pci_dev->mem_resource[0].len =  (uint64_t)size;
-	PMD_INIT_LOG(DEBUG,
-		"PCI Port IO found start=0x%x with size=0x%x",
-		start, size);
+	rte_write16(1, &hw->common_cfg->queue_enable);
 
-	/* can't support lsc interrupt without uio */
-	pci_dev->driver->drv_flags &= ~RTE_PCI_DRV_INTR_LSC;
+	PMD_INIT_LOG(DEBUG, "queue %u addresses:", vq->vq_queue_index);
+	PMD_INIT_LOG(DEBUG, "\t desc_addr: %" PRIx64, desc_addr);
+	PMD_INIT_LOG(DEBUG, "\t aval_addr: %" PRIx64, avail_addr);
+	PMD_INIT_LOG(DEBUG, "\t used_addr: %" PRIx64, used_addr);
+	PMD_INIT_LOG(DEBUG, "\t notify addr: %p (notify offset: %u)",
+		vq->notify_addr, notify_off);
 
 	return 0;
 }
 
-/* Extract I/O port numbers from sysfs */
-static int
-legacy_virtio_resource_init(struct rte_pci_device *pci_dev)
+static void
+modern_del_queue(struct virtio_hw *hw, struct virtqueue *vq)
 {
-	if (virtio_resource_init_by_uio(pci_dev) == 0)
-		return 0;
-	else
-		return virtio_resource_init_by_ioports(pci_dev);
-}
+	rte_write16(vq->vq_queue_index, &hw->common_cfg->queue_select);
 
-#else
-static int
-legayc_virtio_has_msix(const struct rte_pci_addr *loc __rte_unused)
-{
-	/* nic_uio does not enable interrupts, return 0 (false). */
-	return 0;
+	io_write64_twopart(0, &hw->common_cfg->queue_desc_lo,
+				  &hw->common_cfg->queue_desc_hi);
+	io_write64_twopart(0, &hw->common_cfg->queue_avail_lo,
+				  &hw->common_cfg->queue_avail_hi);
+	io_write64_twopart(0, &hw->common_cfg->queue_used_lo,
+				  &hw->common_cfg->queue_used_hi);
+
+	rte_write16(0, &hw->common_cfg->queue_enable);
 }
 
-static int
-legacy_virtio_resource_init(struct rte_pci_device *pci_dev __rte_unused)
+static void
+modern_notify_queue(struct virtio_hw *hw __rte_unused, struct virtqueue *vq)
 {
-	/* no setup required */
-	return 0;
+	rte_write16(vq->vq_queue_index, vq->notify_addr);
 }
-#endif
 
-static const struct virtio_pci_ops legacy_ops = {
-	.read_dev_cfg	= legacy_read_dev_config,
-	.write_dev_cfg	= legacy_write_dev_config,
-	.reset		= legacy_reset,
-	.get_status	= legacy_get_status,
-	.set_status	= legacy_set_status,
-	.get_features	= legacy_get_features,
-	.set_features	= legacy_set_features,
-	.get_isr	= legacy_get_isr,
-	.set_config_irq	= legacy_set_config_irq,
-	.get_queue_num	= legacy_get_queue_num,
-	.setup_queue	= legacy_setup_queue,
-	.del_queue	= legacy_del_queue,
-	.notify_queue	= legacy_notify_queue,
+const struct virtio_pci_ops modern_ops = {
+	.read_dev_cfg	= modern_read_dev_config,
+	.write_dev_cfg	= modern_write_dev_config,
+	.reset		= modern_reset,
+	.get_status	= modern_get_status,
+	.set_status	= modern_set_status,
+	.get_features	= modern_get_features,
+	.set_features	= modern_set_features,
+	.get_isr	= modern_get_isr,
+	.set_config_irq	= modern_set_config_irq,
+	.set_queue_irq  = modern_set_queue_irq,
+	.get_queue_num	= modern_get_queue_num,
+	.setup_queue	= modern_setup_queue,
+	.del_queue	= modern_del_queue,
+	.notify_queue	= modern_notify_queue,
 };
 
 
@@ -448,14 +515,14 @@ void
 vtpci_read_dev_config(struct virtio_hw *hw, size_t offset,
 		      void *dst, int length)
 {
-	hw->vtpci_ops->read_dev_cfg(hw, offset, dst, length);
+	VTPCI_OPS(hw)->read_dev_cfg(hw, offset, dst, length);
 }
 
 void
 vtpci_write_dev_config(struct virtio_hw *hw, size_t offset,
 		       const void *src, int length)
 {
-	hw->vtpci_ops->write_dev_cfg(hw, offset, src, length);
+	VTPCI_OPS(hw)->write_dev_cfg(hw, offset, src, length);
 }
 
 uint64_t
@@ -468,7 +535,7 @@ vtpci_negotiate_features(struct virtio_hw *hw, uint64_t host_features)
 	 * host all support.
 	 */
 	features = host_features & hw->guest_features;
-	hw->vtpci_ops->set_features(hw, features);
+	VTPCI_OPS(hw)->set_features(hw, features);
 
 	return features;
 }
@@ -476,9 +543,9 @@ vtpci_negotiate_features(struct virtio_hw *hw, uint64_t host_features)
 void
 vtpci_reset(struct virtio_hw *hw)
 {
-	hw->vtpci_ops->set_status(hw, VIRTIO_CONFIG_STATUS_RESET);
+	VTPCI_OPS(hw)->set_status(hw, VIRTIO_CONFIG_STATUS_RESET);
 	/* flush status write */
-	hw->vtpci_ops->get_status(hw);
+	VTPCI_OPS(hw)->get_status(hw);
 }
 
 void
@@ -491,34 +558,175 @@ void
 vtpci_set_status(struct virtio_hw *hw, uint8_t status)
 {
 	if (status != VIRTIO_CONFIG_STATUS_RESET)
-		status |= hw->vtpci_ops->get_status(hw);
+		status |= VTPCI_OPS(hw)->get_status(hw);
 
-	hw->vtpci_ops->set_status(hw, status);
+	VTPCI_OPS(hw)->set_status(hw, status);
+}
+
+uint8_t
+vtpci_get_status(struct virtio_hw *hw)
+{
+	return VTPCI_OPS(hw)->get_status(hw);
 }
 
 uint8_t
 vtpci_isr(struct virtio_hw *hw)
 {
-	return hw->vtpci_ops->get_isr(hw);
+	return VTPCI_OPS(hw)->get_isr(hw);
 }
 
+static void *
+get_cfg_addr(struct rte_pci_device *dev, struct virtio_pci_cap *cap)
+{
+	uint8_t  bar    = cap->bar;
+	uint32_t length = cap->length;
+	uint32_t offset = cap->offset;
+	uint8_t *base;
+
+	if (bar > 5) {
+		PMD_INIT_LOG(ERR, "invalid bar: %u", bar);
+		return NULL;
+	}
 
-/* Enable one vector (0) for Link State Intrerrupt */
-uint16_t
-vtpci_irq_config(struct virtio_hw *hw, uint16_t vec)
+	if (offset + length < offset) {
+		PMD_INIT_LOG(ERR, "offset(%u) + length(%u) overflows",
+			offset, length);
+		return NULL;
+	}
+
+	if (offset + length > dev->mem_resource[bar].len) {
+		PMD_INIT_LOG(ERR,
+			"invalid cap: overflows bar space: %u > %" PRIu64,
+			offset + length, dev->mem_resource[bar].len);
+		return NULL;
+	}
+
+	base = dev->mem_resource[bar].addr;
+	if (base == NULL) {
+		PMD_INIT_LOG(ERR, "bar %u base addr is NULL", bar);
+		return NULL;
+	}
+
+	return base + offset;
+}
+
+static int
+virtio_read_caps(struct rte_pci_device *dev, struct virtio_hw *hw)
 {
-	return hw->vtpci_ops->set_config_irq(hw, vec);
+	uint8_t pos;
+	struct virtio_pci_cap cap;
+	int ret;
+
+	if (rte_eal_pci_map_device(dev)) {
+		PMD_INIT_LOG(DEBUG, "failed to map pci device!");
+		return -1;
+	}
+
+	ret = rte_eal_pci_read_config(dev, &pos, 1, PCI_CAPABILITY_LIST);
+	if (ret < 0) {
+		PMD_INIT_LOG(DEBUG, "failed to read pci capability list");
+		return -1;
+	}
+
+	while (pos) {
+		ret = rte_eal_pci_read_config(dev, &cap, sizeof(cap), pos);
+		if (ret < 0) {
+			PMD_INIT_LOG(ERR,
+				"failed to read pci cap at pos: %x", pos);
+			break;
+		}
+
+		if (cap.cap_vndr == PCI_CAP_ID_MSIX)
+			hw->use_msix = 1;
+
+		if (cap.cap_vndr != PCI_CAP_ID_VNDR) {
+			PMD_INIT_LOG(DEBUG,
+				"[%2x] skipping non VNDR cap id: %02x",
+				pos, cap.cap_vndr);
+			goto next;
+		}
+
+		PMD_INIT_LOG(DEBUG,
+			"[%2x] cfg type: %u, bar: %u, offset: %04x, len: %u",
+			pos, cap.cfg_type, cap.bar, cap.offset, cap.length);
+
+		switch (cap.cfg_type) {
+		case VIRTIO_PCI_CAP_COMMON_CFG:
+			hw->common_cfg = get_cfg_addr(dev, &cap);
+			break;
+		case VIRTIO_PCI_CAP_NOTIFY_CFG:
+			rte_eal_pci_read_config(dev, &hw->notify_off_multiplier,
+						4, pos + sizeof(cap));
+			hw->notify_base = get_cfg_addr(dev, &cap);
+			break;
+		case VIRTIO_PCI_CAP_DEVICE_CFG:
+			hw->dev_cfg = get_cfg_addr(dev, &cap);
+			break;
+		case VIRTIO_PCI_CAP_ISR_CFG:
+			hw->isr = get_cfg_addr(dev, &cap);
+			break;
+		}
+
+next:
+		pos = cap.cap_next;
+	}
+
+	if (hw->common_cfg == NULL || hw->notify_base == NULL ||
+	    hw->dev_cfg == NULL    || hw->isr == NULL) {
+		PMD_INIT_LOG(INFO, "no modern virtio pci device found.");
+		return -1;
+	}
+
+	PMD_INIT_LOG(INFO, "found modern virtio pci device.");
+
+	PMD_INIT_LOG(DEBUG, "common cfg mapped at: %p", hw->common_cfg);
+	PMD_INIT_LOG(DEBUG, "device cfg mapped at: %p", hw->dev_cfg);
+	PMD_INIT_LOG(DEBUG, "isr cfg mapped at: %p", hw->isr);
+	PMD_INIT_LOG(DEBUG, "notify base: %p, notify off multiplier: %u",
+		hw->notify_base, hw->notify_off_multiplier);
+
+	return 0;
 }
 
+/*
+ * Return -1:
+ *   if there is error mapping with VFIO/UIO.
+ *   if port map error when driver type is KDRV_NONE.
+ *   if whitelisted but driver type is KDRV_UNKNOWN.
+ * Return 1 if kernel driver is managing the device.
+ * Return 0 on success.
+ */
 int
 vtpci_init(struct rte_pci_device *dev, struct virtio_hw *hw)
 {
-	hw->vtpci_ops = &legacy_ops;
+	/*
+	 * Try if we can succeed reading virtio pci caps, which exists
+	 * only on modern pci device. If failed, we fallback to legacy
+	 * virtio handling.
+	 */
+	if (virtio_read_caps(dev, hw) == 0) {
+		PMD_INIT_LOG(INFO, "modern virtio pci detected.");
+		virtio_hw_internal[hw->port_id].vtpci_ops = &modern_ops;
+		hw->modern = 1;
+		return 0;
+	}
 
-	if (legacy_virtio_resource_init(dev) < 0)
+	PMD_INIT_LOG(INFO, "trying with legacy virtio pci.");
+	if (rte_eal_pci_ioport_map(dev, 0, VTPCI_IO(hw)) < 0) {
+		if (dev->kdrv == RTE_KDRV_UNKNOWN &&
+		    (!dev->device.devargs ||
+		     dev->device.devargs->type !=
+			RTE_DEVTYPE_WHITELISTED_PCI)) {
+			PMD_INIT_LOG(INFO,
+				"skip kernel managed virtio device.");
+			return 1;
+		}
 		return -1;
+	}
+
+	virtio_hw_internal[hw->port_id].vtpci_ops = &legacy_ops;
 	hw->use_msix = legacy_virtio_has_msix(&dev->addr);
-	hw->io_base  = (uint32_t)(uintptr_t)dev->mem_resource[0].addr;
+	hw->modern   = 0;
 
 	return 0;
 }