[dpdk-dev] [PATCH 0/3] add unit tests for eal vfio library

[dpdk-dev] [PATCH 0/3] add unit tests for eal vfio library

[Chaitanya Babu Talluri]

1/3: fix vfio unmap that fails unexpectedly
2/3: fix vfio unmap that succeeds unexpectedly
3/3: add unit tests for eal vfio

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>

Chaitanya Babu Talluri (3):
  lib/eal: fix vfio unmap that fails unexpectedly
  lib/eal: fix vfio unmap that succeeds unexpectedly
  app/test: add unit tests for eal vfio

 app/test/Makefile                   |   1 +
 app/test/meson.build                |   2 +
 app/test/test_eal_vfio.c            | 728 ++++++++++++++++++++++++++++
 lib/librte_eal/linux/eal/eal_vfio.c |  59 ++-
 4 files changed, 783 insertions(+), 7 deletions(-)
 create mode 100644 app/test/test_eal_vfio.c

-- 
2.17.2

[dpdk-dev] [PATCH 1/3] lib/eal: fix vfio unmap that fails unexpectedly

[Chaitanya Babu Talluri]

Unmap of multiple pages fails after a sequence of partial map/unmaps.
The scenario is that multiple maps are created in user_mem_maps,
after multiple map/unmap/remap sequences.

For an example,
Steps:
1. Map 3 pages together
2. Un-map page1
3. Re-map page 1
4. Un-map page 2
5. Re-map page 2
6. Un-map page 3
7. Re-map page 3
8. Un-map all pages

Unmap fails when there are duplicate entries in user_mem_maps.

The fix is to validate if the input VA, IOVA exists in
user_mem_maps before creating map.

Fixes: 73a63908 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 46 +++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 501c74f23..104912077 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -212,6 +212,41 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 	return NULL;
 }
 
+static int
+find_user_mem_map_overlap(struct user_mem_maps *user_mem_maps, uint64_t addr,
+		uint64_t iova, uint64_t len)
+{
+	uint64_t va_end = addr + len;
+	uint64_t iova_end = iova + len;
+	int i;
+
+	for (i = 0; i < user_mem_maps->n_maps; i++) {
+		struct user_mem_map *map = &user_mem_maps->maps[i];
+		uint64_t map_va_end = map->addr + map->len;
+		uint64_t map_iova_end = map->iova + map->len;
+
+		bool no_lo_va_overlap = addr < map->addr && va_end <= map->addr;
+		bool no_hi_va_overlap = addr >= map_va_end &&
+			va_end > map_va_end;
+		bool no_lo_iova_overlap = iova < map->iova &&
+			iova_end <= map->iova;
+		bool no_hi_iova_overlap = iova >= map_iova_end &&
+			iova_end > map_iova_end;
+
+		/* check input VA and iova is not within the
+		 * existing map's range
+		 */
+		if ((no_lo_va_overlap || no_hi_va_overlap) &&
+				(no_lo_iova_overlap || no_hi_iova_overlap))
+			continue;
+		else
+			/* map overlaps */
+			return 1;
+	}
+	/* map doesn't overlap */
+	return 0;
+}
+
 /* this will sort all user maps, and merge/compact any adjacent maps */
 static void
 compact_user_maps(struct user_mem_maps *user_mem_maps)
@@ -1732,6 +1767,17 @@ container_dma_map(struct vfio_config *vfio_cfg, uint64_t vaddr, uint64_t iova,
 		ret = -1;
 		goto out;
 	}
+
+	/* check whether vaddr and iova exists in user_mem_maps */
+	ret = find_user_mem_map_overlap(user_mem_maps, vaddr, iova, len);
+	if (ret) {
+		RTE_LOG(ERR, EAL, "Mapping overlaps with a previously "
+				"existing mapping\n");
+		rte_errno = EEXIST;
+		ret = -1;
+		goto out;
+	}
+
 	/* map the entry */
 	if (vfio_dma_mem_map(vfio_cfg, vaddr, iova, len, 1)) {
 		/* technically, this will fail if there are currently no devices
-- 
2.17.2

Re: [dpdk-dev] [PATCH 1/3] lib/eal: fix vfio unmap that fails unexpectedly

[Burakov, Anatoly]

On 21-Aug-19 2:02 PM, Chaitanya Babu Talluri wrote:
> Unmap of multiple pages fails after a sequence of partial map/unmaps.
> The scenario is that multiple maps are created in user_mem_maps,
> after multiple map/unmap/remap sequences.
> 
> For an example,
> Steps:
> 1. Map 3 pages together
> 2. Un-map page1
> 3. Re-map page 1
> 4. Un-map page 2
> 5. Re-map page 2
> 6. Un-map page 3
> 7. Re-map page 3
> 8. Un-map all pages

I don't think this description is correct in relation to what is being 
fixed here. The code attempts to prevent overlaps, but there are no 
overlaps in the above example - none of the above operations would 
trigger the added code.

-- 
Thanks,
Anatoly

[dpdk-dev] [PATCH 2/3] lib/eal: fix vfio unmap that succeeds unexpectedly

[Chaitanya Babu Talluri]

Un-map of page with valid virtual address and
another page's IOVA succeeds unexpectedly.
An entry in user_mem_maps can refer multiple pages.
Currently in such case to unmap single page, VA
and IOVA related to entry in user_mem_maps is
checked but not based on page (based on the
page size), this is the cause.

The solution is that in find_user_mem_maps,
check whether user input iova is in relation with
input virtual address of the page which is to be
unmapped.

Fixes: 73a6390859 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 104912077..04c284cb2 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -184,13 +184,13 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		uint64_t iova, uint64_t len)
 {
 	uint64_t va_end = addr + len;
-	uint64_t iova_end = iova + len;
 	int i;
 
 	for (i = 0; i < user_mem_maps->n_maps; i++) {
 		struct user_mem_map *map = &user_mem_maps->maps[i];
 		uint64_t map_va_end = map->addr + map->len;
-		uint64_t map_iova_end = map->iova + map->len;
+		uint64_t diff_addr_len = addr - map->addr;
+		uint64_t expected_iova = map->iova + diff_addr_len;
 
 		/* check start VA */
 		if (addr < map->addr || addr >= map_va_end)
@@ -199,11 +199,10 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		if (va_end <= map->addr || va_end > map_va_end)
 			continue;
 
-		/* check start IOVA */
-		if (iova < map->iova || iova >= map_iova_end)
-			continue;
-		/* check if IOVA end is within boundaries */
-		if (iova_end <= map->iova || iova_end > map_iova_end)
+		/* check whether user input iova is in sync with
+		 * user_mem_map entry's iova
+		 */
+		if (expected_iova != iova)
 			continue;
 
 		/* we've found our map */
-- 
2.17.2

Re: [dpdk-dev] [PATCH 2/3] lib/eal: fix vfio unmap that succeeds unexpectedly

[Burakov, Anatoly]

On 21-Aug-19 2:02 PM, Chaitanya Babu Talluri wrote:
> Un-map of page with valid virtual address and
> another page's IOVA succeeds unexpectedly.
> An entry in user_mem_maps can refer multiple pages.
> Currently in such case to unmap single page, VA
> and IOVA related to entry in user_mem_maps is
> checked but not based on page (based on the
> page size), this is the cause.
> 
> The solution is that in find_user_mem_maps,
> check whether user input iova is in relation with
> input virtual address of the page which is to be
> unmapped.

The description could be clearer. Suggested rewording:

Unmapping page with a VA that is found in the list of current mappings 
will succeed even if the IOVA for the chunk that is being unmapped, is 
mismatched. Fix it by checking if IOVA address matches the expected IOVA 
address exactly.

-- 
Thanks,
Anatoly

[dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio

[Chaitanya Babu Talluri]

Unit test cases are added for eal vfio library.
eal_vfio_autotest added to meson build file.

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 app/test/Makefile        |   1 +
 app/test/meson.build     |   2 +
 app/test/test_eal_vfio.c | 728 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 731 insertions(+)
 create mode 100644 app/test/test_eal_vfio.c

diff --git a/app/test/Makefile b/app/test/Makefile
index 26ba6fe2b..9b9c78b4e 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -137,6 +137,7 @@ SRCS-y += test_cpuflags.c
 SRCS-y += test_mp_secondary.c
 SRCS-y += test_eal_flags.c
 SRCS-y += test_eal_fs.c
+SRCS-y += test_eal_vfio.c
 SRCS-y += test_alarm.c
 SRCS-y += test_interrupts.c
 SRCS-y += test_version.c
diff --git a/app/test/meson.build b/app/test/meson.build
index ec40943bd..2ec9c863a 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -36,6 +36,7 @@ test_sources = files('commands.c',
 	'test_distributor_perf.c',
 	'test_eal_flags.c',
 	'test_eal_fs.c',
+	'test_eal_vfio.c',
 	'test_efd.c',
 	'test_efd_perf.c',
 	'test_errno.c',
@@ -175,6 +176,7 @@ fast_test_names = [
         'eal_flags_file_prefix_autotest',
         'eal_flags_misc_autotest',
         'eal_fs_autotest',
+	'eal_vfio_autotest',
         'errno_autotest',
         'event_ring_autotest',
         'func_reentrancy_autotest',
diff --git a/app/test/test_eal_vfio.c b/app/test/test_eal_vfio.c
new file mode 100644
index 000000000..8995573df
--- /dev/null
+++ b/app/test/test_eal_vfio.c
@@ -0,0 +1,728 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2019 Intel Corporation
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <stdint.h>
+#include <errno.h>
+#include <sys/mman.h>
+#include <unistd.h>
+#include <rte_vfio.h>
+#include <rte_malloc.h>
+#include <rte_eal_memconfig.h>
+
+#include "test.h"
+
+#if !defined(RTE_EXEC_ENV_LINUX) || !defined(RTE_EAL_VFIO)
+static int
+test_eal_vfio(void)
+{
+	printf("VFIO not supported, skipping test\n");
+	return TEST_SKIPPED;
+}
+
+#else
+
+#define PAGESIZE sysconf(_SC_PAGESIZE)
+#define INVALID_CONTAINER_FD -5
+#define THREE_PAGES 3
+#define UNMAPPED_ADDR 0x1500
+
+uint64_t virtaddr_64;
+const char *name = "heap";
+size_t map_length;
+int container_fds[RTE_MAX_VFIO_CONTAINERS];
+
+static int
+check_get_mem(void *addr, rte_iova_t *iova)
+{
+	const struct rte_memseg_list *msl;
+	const struct rte_memseg *ms;
+	rte_iova_t expected_iova;
+
+	msl = rte_mem_virt2memseg_list(addr);
+	if (!msl->external) {
+		printf("%s():%i: Memseg list is not marked as "
+				"external\n", __func__, __LINE__);
+		return -1;
+	}
+	ms = rte_mem_virt2memseg(addr, msl);
+	if (ms == NULL) {
+		printf("%s():%i: Failed to retrieve memseg for "
+				"external mem\n", __func__, __LINE__);
+		return -1;
+	}
+	if (ms->addr != addr) {
+		printf("%s():%i: VA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	expected_iova = (iova == NULL) ? RTE_BAD_IOVA : iova[0];
+	if (ms->iova != expected_iova) {
+		printf("%s():%i: IOVA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	return 0;
+}
+
+/* Initialize container fds */
+static int
+initialize_container_fds(void)
+{
+	int i = 0;
+
+	for (i = 0; i < RTE_MAX_VFIO_CONTAINERS; i++)
+		container_fds[i] = -1;
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container create */
+static int
+test_vfio_container_create(void)
+{
+	int ret = 0, i = 0;
+
+	/* check max containers limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		container_fds[i] = rte_vfio_container_create();
+		TEST_ASSERT(container_fds[i] >  0, "Test to check "
+				"rte_vfio_container_create with max "
+				"containers limit: Failed\n");
+	}
+
+	/* check rte_vfio_container_create when exceeds max containers limit */
+	ret = rte_vfio_container_create();
+	TEST_ASSERT(ret == -1, "Test to check "
+			"rte_vfio_container_create container "
+			"when exceeds limit: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container destroy */
+static int
+test_vfio_container_destroy(void)
+{
+	int i = 0, ret = 0;
+
+	/* check to destroy max container limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		ret = rte_vfio_container_destroy(container_fds[i]);
+		TEST_ASSERT(ret == 0, "Test to check "
+				"rte_vfio_container_destroy: Failed\n");
+		container_fds[i] = -1;
+	}
+
+	/* check rte_vfio_container_destroy with valid but non existing value */
+	ret = rte_vfio_container_destroy(0);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy with "
+			"valid but non existing value: Failed\n");
+
+	/* check rte_vfio_container_destroy with invalid value */
+	ret = rte_vfio_container_destroy(-5);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy "
+			"with invalid value: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to bind a IOMMU group to a container*/
+static int
+test_rte_vfio_container_group_bind(void)
+{
+	int ret = 0;
+
+	/* Test case to bind with invalid container fd */
+	ret = rte_vfio_container_group_bind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to bind with non-existing container fd */
+	ret = rte_vfio_container_group_bind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unbind a IOMMU group from a container*/
+static int
+test_rte_vfio_container_group_unbind(void)
+{
+	int ret = 0;
+
+	/* Test case to unbind container from invalid group*/
+	ret = rte_vfio_container_group_unbind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to unbind container from group*/
+	ret = rte_vfio_container_group_unbind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with  non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to get IOMMU group number for a device*/
+static int
+test_rte_vfio_get_group_num(void)
+{
+	int ret = 0, invalid_group_num = 0;
+
+	/* Test case to get IOMMU group num from invalid group */
+	ret = rte_vfio_get_group_num(NULL, NULL, &invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	/* Test case to get IOMMU group num from invalid device address and
+	 * valid sysfs_base
+	 */
+	ret = rte_vfio_get_group_num("/sys/bus/pci/devices/", NULL,
+			&invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA mapping for devices in a container */
+static int
+test_rte_vfio_container_dma_map(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to map device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with invalid "
+			"container: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to map device for existing fd with no device attached and
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check  map device for existing fd "
+			"with no device attached and non-zero "
+			"map_length: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA unmapping for devices in a container*/
+static int
+test_rte_vfio_container_dma_unmap(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*Function to setup external memory */
+static int
+test_heap_mem_setup(size_t map_length, int n_pages)
+{
+	rte_iova_t iova[map_length / PAGESIZE];
+	void *addr;
+
+	addr = mmap(NULL, map_length, PROT_WRITE | PROT_READ,
+			MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+	if (addr == MAP_FAILED) {
+		printf("%s():%i: Failed to create dummy memory area\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	rte_iova_t tmp = 0x100000000 + PAGESIZE;
+	iova[0] = tmp;
+
+	if (rte_malloc_heap_create(name) != 0) {
+		printf("%s():%i: Failed to Create heap with valid name\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (rte_malloc_heap_memory_add(name, addr, map_length, iova, n_pages,
+				PAGESIZE) != 0) {
+		printf("%s():%i: Failed to add memory to heap\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (check_get_mem(addr, iova) != 0) {
+		printf("%s():%i: Failed to verify memory\n",
+				__func__, __LINE__);
+
+		return -1;
+	}
+	virtaddr_64 = (uint64_t)(uintptr_t)addr;
+
+	return 0;
+}
+
+/* Function to free the external memory */
+static void
+test_heap_mem_free(void)
+{
+	if (rte_malloc_heap_memory_remove(name, (void *)virtaddr_64,
+				map_length) != 0) {
+		printf("%s():%i: Failed to remove memory\n",
+				__func__, __LINE__);
+		return;
+	}
+	rte_malloc_heap_destroy(name);
+
+	munmap((void *)virtaddr_64, map_length);
+}
+
+/* Test to map memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_map(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map memory for VFIO with zero  vaddr, iova addr
+	 * and map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with incorrect inputs: Failed\n");
+
+	/* Test case to map memory for VFIO with zero vaddr, iova addr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, unmapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(1000000, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"unmapped virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, mapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check already mapped virtual address */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check start virtual address + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + UNMAPPED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to check start virtual address before
+	 * existing map, overlaps
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNMAPPED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start virtual address "
+			"before existing map, overlaps: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNMAPPED_ADDR), 0, 500);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to check already mapped iova overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with already mapped iova overlaps: Failed\n");
+
+	/* Test case to check start iova + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNMAPPED_ADDR),
+			map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start iova + length range overlaps: Failed\n");
+
+	/* Test case to check invalid iova */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + 5000), map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid iova: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNMAPPED_ADDR), 100);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid vaddr, valid iova addr
+	 * and valid map_length
+	 */
+	uint64_t invalid_addr = virtaddr_64 + 1;
+	ret = rte_vfio_dma_map(invalid_addr, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with mapped invalid virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid iova addr, valid vaddr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, UNMAPPED_ADDR, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"invalid iova address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unmap memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_unmap(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with incorrect input: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero iova addr,
+	 * unmapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and unmapped addr: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with unmapped vaddr, iova
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"unmapped addr, iova: Failed\n");
+
+	/* Test case to map memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped invalid vaddr,
+	 * valid IOVA and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + 1), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"invalid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static int
+test_rte_vfio_dma_map_overlaps(void)
+{
+	int ret = 0;
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid address:Failed\n");
+
+	/* Test case to map same start virtual address and
+	 * extend beyond end virtual address
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and extend beyond end "
+			"virtual address: Failed\n");
+
+	/* Test case to map same start virtual address and same end address*/
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and "
+			"same end address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and "
+		"mapped valid virtual address: Failed\n");
+
+	/* Test case to map different virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			(PAGESIZE * 2));
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with different virtual address: Failed\n");
+
+	/* Test case to map different start virtual address and
+	 * ends with same address
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with different start virtual address and "
+			"ends with same address: Failed\n");
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address range: Failed\n");
+
+	/* Test case to map middle overlapping virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap 1st page: Failed\n");
+
+	/* Test case to map 1st and 2nd page overlaps */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st and 2nd page overlaps: Failed\n");
+
+	/* Test case to map 3rd and 4th pages */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 3rd and 4th pages: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 3rd page: Failed\n");
+
+	/* Test case to map 1st page with total length
+	 * that overlaps middle page
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st page with total length "
+			"that overlaps middle page: Failed\n");
+
+	/* Test case to unmap 2nd page  */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 2nd page: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*allocate three pages */
+static int
+test_rte_vfio_dma_map_threepages(void)
+{
+	int ret = 0;
+
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+	uint64_t page1_va, page2_va, page3_va;
+	rte_iova_t page1_iova, page2_iova, page3_iova;
+
+	page1_va = virtaddr_64;
+	page2_va = virtaddr_64 + PAGESIZE;
+	page3_va = virtaddr_64 + (PAGESIZE * 2);
+
+	page1_iova = 0;
+	page2_iova = 0 + PAGESIZE;
+	page3_iova = 0 + (PAGESIZE * 2);
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid 1st page map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 2nd page */
+	ret = rte_vfio_dma_unmap(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to map 2nd page */
+	ret = rte_vfio_dma_map(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 3rd page virtual address: Failed\n");
+
+	/* Test case to map 3rd page */
+	ret = rte_vfio_dma_map(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped 3rd page valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page, but used IOVA address of 2nd page */
+	ret = rte_vfio_dma_unmap(page1_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static struct
+unit_test_suite eal_vfio_testsuite  = {
+	.suite_name = "EAL VFIO Unit Test Suite",
+	.setup = initialize_container_fds,
+	.teardown = NULL,
+	.unit_test_cases = {
+		/* Test Case 1: To check vfio container create test cases */
+		TEST_CASE(test_vfio_container_create),
+
+		/* Test Case 2: To check vfio container destroy */
+		TEST_CASE(test_vfio_container_destroy),
+
+		/* Test Case 3: To  bind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_bind),
+
+		/* Test Case 4: To get IOMMU group number for a device*/
+		TEST_CASE(test_rte_vfio_get_group_num),
+
+		/* Test Case 5: To unbind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_unbind),
+
+		/* Test Case 6: To perform DMA mapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free, test_rte_vfio_dma_map),
+
+		/* Test Case 7: To perform DMA unmapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_unmap),
+
+		/* Test Case 8: To perform map devices in specific container */
+		TEST_CASE(test_rte_vfio_container_dma_map),
+
+		/* Test Case 9: To perform unmap devices in specific container
+		 */
+		TEST_CASE(test_rte_vfio_container_dma_unmap),
+
+		/* Test Case 10: To perform three pages */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_threepages),
+
+		/* Test Case 11: To check DMA overlaps */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_overlaps),
+
+		TEST_CASES_END()
+	}
+};
+
+static int
+test_eal_vfio(void)
+{
+	return unit_test_suite_runner(&eal_vfio_testsuite);
+}
+
+#endif
+
+REGISTER_TEST_COMMAND(eal_vfio_autotest, test_eal_vfio);
-- 
2.17.2

Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio

[Aaron Conole]

Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com> writes:

> Unit test cases are added for eal vfio library.
> eal_vfio_autotest added to meson build file.
>
> Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
> ---

Thanks for adding unit tests for the vfio library.

In this case, there seems to be some failures - can you help determine
the cause:

https://travis-ci.com/ovsrobot/dpdk/jobs/227066776

Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio

[Chaitanya Babu, TalluriX]

Hi Aaron,

>-----Original Message-----
>From: Aaron Conole [mailto:aconole@redhat.com]
>Sent: Wednesday, August 21, 2019 9:15 PM
>To: Chaitanya Babu, TalluriX <tallurix.chaitanya.babu@intel.com>
>Cc: dev@dpdk.org; Pattan, Reshma <reshma.pattan@intel.com>;
>Parthasarathy, JananeeX M <jananeex.m.parthasarathy@intel.com>;
>Burakov, Anatoly <anatoly.burakov@intel.com>
>Subject: Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio
>
>Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com> writes:
>
>> Unit test cases are added for eal vfio library.
>> eal_vfio_autotest added to meson build file.
>>
>> Signed-off-by: Chaitanya Babu Talluri
>> <tallurix.chaitanya.babu@intel.com>
>> ---
>
>Thanks for adding unit tests for the vfio library.
>
>In this case, there seems to be some failures - can you help determine the
>cause:
>
As per log we observed 
"EAL:   VFIO support not initialized"

Can you Please check if device is binded or not with vfio-pci.
If device is not binded we might get this error (EAL:   VFIO support not initialized).

We have binded the device and observed the tests are running fine as expected.

>https://travis-ci.com/ovsrobot/dpdk/jobs/227066776

Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio

[Burakov, Anatoly]

On 22-Aug-19 8:18 AM, Chaitanya Babu, TalluriX wrote:
> Hi Aaron,
> 
>> -----Original Message-----
>> From: Aaron Conole [mailto:aconole@redhat.com]
>> Sent: Wednesday, August 21, 2019 9:15 PM
>> To: Chaitanya Babu, TalluriX <tallurix.chaitanya.babu@intel.com>
>> Cc: dev@dpdk.org; Pattan, Reshma <reshma.pattan@intel.com>;
>> Parthasarathy, JananeeX M <jananeex.m.parthasarathy@intel.com>;
>> Burakov, Anatoly <anatoly.burakov@intel.com>
>> Subject: Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio
>>
>> Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com> writes:
>>
>>> Unit test cases are added for eal vfio library.
>>> eal_vfio_autotest added to meson build file.
>>>
>>> Signed-off-by: Chaitanya Babu Talluri
>>> <tallurix.chaitanya.babu@intel.com>
>>> ---
>>
>> Thanks for adding unit tests for the vfio library.
>>
>> In this case, there seems to be some failures - can you help determine the
>> cause:
>>
> As per log we observed
> "EAL:   VFIO support not initialized"
> 
> Can you Please check if device is binded or not with vfio-pci.
> If device is not binded we might get this error (EAL:   VFIO support not initialized).
> 
> We have binded the device and observed the tests are running fine as expected.
> 
>> https://travis-ci.com/ovsrobot/dpdk/jobs/227066776
> 

The test should be skipped if VFIO if VFIO is not initialized. I think 
we have an API for that, although it is not meant to be used by 
application code.

-- 
Thanks,
Anatoly

Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio

[Aaron Conole]

"Burakov, Anatoly" <anatoly.burakov@intel.com> writes:

> On 22-Aug-19 8:18 AM, Chaitanya Babu, TalluriX wrote:
>> Hi Aaron,
>>
>>> -----Original Message-----
>>> From: Aaron Conole [mailto:aconole@redhat.com]
>>> Sent: Wednesday, August 21, 2019 9:15 PM
>>> To: Chaitanya Babu, TalluriX <tallurix.chaitanya.babu@intel.com>
>>> Cc: dev@dpdk.org; Pattan, Reshma <reshma.pattan@intel.com>;
>>> Parthasarathy, JananeeX M <jananeex.m.parthasarathy@intel.com>;
>>> Burakov, Anatoly <anatoly.burakov@intel.com>
>>> Subject: Re: [dpdk-dev] [PATCH 3/3] app/test: add unit tests for eal vfio
>>>
>>> Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com> writes:
>>>
>>>> Unit test cases are added for eal vfio library.
>>>> eal_vfio_autotest added to meson build file.
>>>>
>>>> Signed-off-by: Chaitanya Babu Talluri
>>>> <tallurix.chaitanya.babu@intel.com>
>>>> ---
>>>
>>> Thanks for adding unit tests for the vfio library.
>>>
>>> In this case, there seems to be some failures - can you help determine the
>>> cause:
>>>
>> As per log we observed
>> "EAL:   VFIO support not initialized"
>>
>> Can you Please check if device is binded or not with vfio-pci.
>> If device is not binded we might get this error (EAL:   VFIO support not initialized).
>>
>> We have binded the device and observed the tests are running fine as expected.
>>
>>> https://travis-ci.com/ovsrobot/dpdk/jobs/227066776
>>
>
> The test should be skipped if VFIO if VFIO is not initialized.

+1 - not every developer will have that set up for their systems.  And
for those developers building drive-by patches, we shouldn't make the
unit tests unusable.

> I think
> we have an API for that, although it is not meant to be used by
> application code.

[dpdk-dev] [PATCH v2 0/3] add unit tests for eal vfio library

[Chaitanya Babu Talluri]

1/3: fix vfio unmap that fails unexpectedly
2/3: fix vfio unmap that succeeds unexpectedly
3/3: add unit tests for eal vfio

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>

---
v2: Updated commit messages.
---
Chaitanya Babu Talluri (3):
  lib/eal: fix vfio unmap that fails unexpectedly
  lib/eal: fix vfio unmap that succeeds unexpectedly
  app/test: add unit tests for eal vfio

 app/test/Makefile                   |   1 +
 app/test/meson.build                |   2 +
 app/test/test_eal_vfio.c            | 728 ++++++++++++++++++++++++++++
 lib/librte_eal/linux/eal/eal_vfio.c |  59 ++-
 4 files changed, 783 insertions(+), 7 deletions(-)
 create mode 100644 app/test/test_eal_vfio.c

-- 
2.17.2

[dpdk-dev] [PATCH v3 0/4] add unit tests for eal vfio library

[Chaitanya Babu Talluri]

1/4: fix vfio unmap that fails unexpectedly
2/4: fix vfio unmap that succeeds unexpectedly
3/4: add API to check iommu type is set
4/4: add unit tests for eal vfio

Patch 4/4 depends on 1/4,2/4,3/4

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
v3: Added API to check iommu type.
v2: Updated commit messages.
---

Chaitanya Babu Talluri (4):
  lib/eal: fix vfio unmap that fails unexpectedly
  lib/eal: fix vfio unmap that succeeds unexpectedly
  lib/eal: add API to check iommu type is set
  app/test: add unit tests for eal vfio

 app/test/Makefile                        |   1 +
 app/test/meson.build                     |   2 +
 app/test/test_eal_vfio.c                 | 736 +++++++++++++++++++++++
 lib/librte_eal/common/include/rte_vfio.h |  10 +
 lib/librte_eal/linux/eal/eal_vfio.c      |  75 ++-
 5 files changed, 817 insertions(+), 7 deletions(-)
 create mode 100644 app/test/test_eal_vfio.c

-- 
2.17.2

[dpdk-dev] [PATCH v4 0/4] add unit tests for eal vfio library

[Chaitanya Babu Talluri]

1/4: fix vfio unmap that fails unexpectedly
2/4: fix vfio unmap that succeeds unexpectedly
3/4: add API to check iommu type is set
4/4: add unit tests for eal vfio

Patch 4/4 depends on 1/4,2/4,3/4

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
v4: Fixed compilation issues.
v3: Added API to check iommu type.
v2: Updated commit messages.
---

Chaitanya Babu Talluri (4):
  lib/eal: fix vfio unmap that fails unexpectedly
  lib/eal: fix vfio unmap that succeeds unexpectedly
  lib/eal: add API to check iommu type is set
  app/test: add unit tests for eal vfio

 app/test/Makefile                        |   1 +
 app/test/meson.build                     |   2 +
 app/test/test_eal_vfio.c                 | 737 +++++++++++++++++++++++
 lib/librte_eal/common/include/rte_vfio.h |  11 +
 lib/librte_eal/linux/eal/eal_vfio.c      |  75 ++-
 lib/librte_eal/rte_eal_version.map       |   3 +
 6 files changed, 822 insertions(+), 7 deletions(-)
 create mode 100644 app/test/test_eal_vfio.c

-- 
2.17.2

Re: [dpdk-dev] [PATCH v4 0/4] add unit tests for eal vfio library

[Burakov, Anatoly]

On 04-Sep-19 1:04 PM, Chaitanya Babu Talluri wrote:
> 1/4: fix vfio unmap that fails unexpectedly
> 2/4: fix vfio unmap that succeeds unexpectedly
> 3/4: add API to check iommu type is set
> 4/4: add unit tests for eal vfio
> 
> Patch 4/4 depends on 1/4,2/4,3/4
> 
> Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
> ---

Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

-- 
Thanks,
Anatoly

Re: [dpdk-dev] [PATCH v4 0/4] add unit tests for eal vfio library

[Stephen Hemminger]

On Wed,  4 Sep 2019 13:04:38 +0100
Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com> wrote:

> 1/4: fix vfio unmap that fails unexpectedly
> 2/4: fix vfio unmap that succeeds unexpectedly
> 3/4: add API to check iommu type is set
> 4/4: add unit tests for eal vfio
> 
> Patch 4/4 depends on 1/4,2/4,3/4
> 
> Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
> ---
> v4: Fixed compilation issues.
> v3: Added API to check iommu type.
> v2: Updated commit messages.
> ---
> 
> Chaitanya Babu Talluri (4):
>   lib/eal: fix vfio unmap that fails unexpectedly
>   lib/eal: fix vfio unmap that succeeds unexpectedly
>   lib/eal: add API to check iommu type is set
>   app/test: add unit tests for eal vfio
> 
>  app/test/Makefile                        |   1 +
>  app/test/meson.build                     |   2 +
>  app/test/test_eal_vfio.c                 | 737 +++++++++++++++++++++++
>  lib/librte_eal/common/include/rte_vfio.h |  11 +
>  lib/librte_eal/linux/eal/eal_vfio.c      |  75 ++-
>  lib/librte_eal/rte_eal_version.map       |   3 +
>  6 files changed, 822 insertions(+), 7 deletions(-)
>  create mode 100644 app/test/test_eal_vfio.c
> 


This patch series looks OK, but is so out of date that EAL layout
has changed. If still interested, then please rebase on current main branch;
retest and resubmit it.

Thanks

[dpdk-dev] [PATCH v4 1/4] lib/eal: fix vfio unmap that fails unexpectedly

[Chaitanya Babu Talluri]

Unmap fails when there are duplicate entries in user_mem_maps.

The fix is to validate if the input VA, IOVA exists or
overlaps in user_mem_maps before creating map.

Fixes: 73a63908 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 46 +++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 501c74f23..104912077 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -212,6 +212,41 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 	return NULL;
 }
 
+static int
+find_user_mem_map_overlap(struct user_mem_maps *user_mem_maps, uint64_t addr,
+		uint64_t iova, uint64_t len)
+{
+	uint64_t va_end = addr + len;
+	uint64_t iova_end = iova + len;
+	int i;
+
+	for (i = 0; i < user_mem_maps->n_maps; i++) {
+		struct user_mem_map *map = &user_mem_maps->maps[i];
+		uint64_t map_va_end = map->addr + map->len;
+		uint64_t map_iova_end = map->iova + map->len;
+
+		bool no_lo_va_overlap = addr < map->addr && va_end <= map->addr;
+		bool no_hi_va_overlap = addr >= map_va_end &&
+			va_end > map_va_end;
+		bool no_lo_iova_overlap = iova < map->iova &&
+			iova_end <= map->iova;
+		bool no_hi_iova_overlap = iova >= map_iova_end &&
+			iova_end > map_iova_end;
+
+		/* check input VA and iova is not within the
+		 * existing map's range
+		 */
+		if ((no_lo_va_overlap || no_hi_va_overlap) &&
+				(no_lo_iova_overlap || no_hi_iova_overlap))
+			continue;
+		else
+			/* map overlaps */
+			return 1;
+	}
+	/* map doesn't overlap */
+	return 0;
+}
+
 /* this will sort all user maps, and merge/compact any adjacent maps */
 static void
 compact_user_maps(struct user_mem_maps *user_mem_maps)
@@ -1732,6 +1767,17 @@ container_dma_map(struct vfio_config *vfio_cfg, uint64_t vaddr, uint64_t iova,
 		ret = -1;
 		goto out;
 	}
+
+	/* check whether vaddr and iova exists in user_mem_maps */
+	ret = find_user_mem_map_overlap(user_mem_maps, vaddr, iova, len);
+	if (ret) {
+		RTE_LOG(ERR, EAL, "Mapping overlaps with a previously "
+				"existing mapping\n");
+		rte_errno = EEXIST;
+		ret = -1;
+		goto out;
+	}
+
 	/* map the entry */
 	if (vfio_dma_mem_map(vfio_cfg, vaddr, iova, len, 1)) {
 		/* technically, this will fail if there are currently no devices
-- 
2.17.2

[dpdk-dev] [PATCH v4 2/4] lib/eal: fix vfio unmap that succeeds unexpectedly

[Chaitanya Babu Talluri]

Unmapping page with a VA that is found in the list
of current mappings will succeed even if the IOVA
for the chunk that is being unmapped,is mismatched.

Fix it by checking if IOVA address matches the
expected IOVA address exactly.

Fixes: 73a6390859 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 104912077..04c284cb2 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -184,13 +184,13 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		uint64_t iova, uint64_t len)
 {
 	uint64_t va_end = addr + len;
-	uint64_t iova_end = iova + len;
 	int i;
 
 	for (i = 0; i < user_mem_maps->n_maps; i++) {
 		struct user_mem_map *map = &user_mem_maps->maps[i];
 		uint64_t map_va_end = map->addr + map->len;
-		uint64_t map_iova_end = map->iova + map->len;
+		uint64_t diff_addr_len = addr - map->addr;
+		uint64_t expected_iova = map->iova + diff_addr_len;
 
 		/* check start VA */
 		if (addr < map->addr || addr >= map_va_end)
@@ -199,11 +199,10 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		if (va_end <= map->addr || va_end > map_va_end)
 			continue;
 
-		/* check start IOVA */
-		if (iova < map->iova || iova >= map_iova_end)
-			continue;
-		/* check if IOVA end is within boundaries */
-		if (iova_end <= map->iova || iova_end > map_iova_end)
+		/* check whether user input iova is in sync with
+		 * user_mem_map entry's iova
+		 */
+		if (expected_iova != iova)
 			continue;
 
 		/* we've found our map */
-- 
2.17.2

[dpdk-dev] [PATCH v4 3/4] lib/eal: add API to check iommu type is set

[Chaitanya Babu Talluri]

Add rte_vfio_iommu_type_is_set() to check IOMMU
type for default container.

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/common/include/rte_vfio.h | 11 +++++++++++
 lib/librte_eal/linux/eal/eal_vfio.c      | 16 ++++++++++++++++
 lib/librte_eal/rte_eal_version.map       |  3 +++
 3 files changed, 30 insertions(+)

diff --git a/lib/librte_eal/common/include/rte_vfio.h b/lib/librte_eal/common/include/rte_vfio.h
index b360485fa..e3f80f772 100644
--- a/lib/librte_eal/common/include/rte_vfio.h
+++ b/lib/librte_eal/common/include/rte_vfio.h
@@ -397,6 +397,17 @@ int
 rte_vfio_container_dma_unmap(int container_fd, uint64_t vaddr,
 		uint64_t iova, uint64_t len);
 
+/**
+ * Check VFIO IOMMU Type is set for default container.
+ *
+ * @return
+ *    0 if successful
+ *   <0 if failed
+ */
+__rte_experimental
+int
+rte_vfio_iommu_type_is_set(void);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 04c284cb2..a5bb1cff4 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -2071,6 +2071,17 @@ rte_vfio_container_dma_unmap(int container_fd, uint64_t vaddr, uint64_t iova,
 	return container_dma_unmap(vfio_cfg, vaddr, iova, len);
 }
 
+int
+rte_vfio_iommu_type_is_set(void)
+{
+	if (vfio_get_iommu_type() < 0) {
+		RTE_LOG(ERR, EAL, "VFIO IOMMU Type is not set\n");
+		return -1;
+	}
+
+	return 0;
+}
+
 #else
 
 int
@@ -2191,4 +2202,9 @@ rte_vfio_container_dma_unmap(__rte_unused int container_fd,
 	return -1;
 }
 
+int
+rte_vfio_iommu_type_is_set(void)
+{
+	return -1;
+}
 #endif /* VFIO_PRESENT */
diff --git a/lib/librte_eal/rte_eal_version.map b/lib/librte_eal/rte_eal_version.map
index 7cbf82d37..5b1ba6fbd 100644
--- a/lib/librte_eal/rte_eal_version.map
+++ b/lib/librte_eal/rte_eal_version.map
@@ -419,4 +419,7 @@ EXPERIMENTAL {
 	rte_mcfg_timer_lock;
 	rte_mcfg_timer_unlock;
 	rte_rand_max;
+
+	# added in 19.11
+	rte_vfio_iommu_type_is_set;
 };
-- 
2.17.2

[dpdk-dev] [PATCH v4 4/4] app/test: add unit tests for eal vfio

[Chaitanya Babu Talluri]

Unit test cases are added for eal vfio library.
eal_vfio_autotest added to meson build file.

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 app/test/Makefile        |   1 +
 app/test/meson.build     |   2 +
 app/test/test_eal_vfio.c | 737 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 740 insertions(+)
 create mode 100644 app/test/test_eal_vfio.c

diff --git a/app/test/Makefile b/app/test/Makefile
index 26ba6fe2b..9b9c78b4e 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -137,6 +137,7 @@ SRCS-y += test_cpuflags.c
 SRCS-y += test_mp_secondary.c
 SRCS-y += test_eal_flags.c
 SRCS-y += test_eal_fs.c
+SRCS-y += test_eal_vfio.c
 SRCS-y += test_alarm.c
 SRCS-y += test_interrupts.c
 SRCS-y += test_version.c
diff --git a/app/test/meson.build b/app/test/meson.build
index ec40943bd..bd96ebb2b 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -36,6 +36,7 @@ test_sources = files('commands.c',
 	'test_distributor_perf.c',
 	'test_eal_flags.c',
 	'test_eal_fs.c',
+	'test_eal_vfio.c',
 	'test_efd.c',
 	'test_efd_perf.c',
 	'test_errno.c',
@@ -175,6 +176,7 @@ fast_test_names = [
         'eal_flags_file_prefix_autotest',
         'eal_flags_misc_autotest',
         'eal_fs_autotest',
+        'eal_vfio_autotest',
         'errno_autotest',
         'event_ring_autotest',
         'func_reentrancy_autotest',
diff --git a/app/test/test_eal_vfio.c b/app/test/test_eal_vfio.c
new file mode 100644
index 000000000..de186e4c9
--- /dev/null
+++ b/app/test/test_eal_vfio.c
@@ -0,0 +1,737 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2019 Intel Corporation
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <stdint.h>
+#include <errno.h>
+#include <sys/mman.h>
+#include <unistd.h>
+#include <rte_compat.h>
+#include <rte_vfio.h>
+#include <rte_malloc.h>
+#include <rte_eal_memconfig.h>
+
+#include "test.h"
+
+#if !defined(RTE_EXEC_ENV_LINUX) || !defined(RTE_EAL_VFIO)
+static int
+test_eal_vfio(void)
+{
+	printf("VFIO not supported, skipping test\n");
+	return TEST_SKIPPED;
+}
+
+#else
+
+#define PAGESIZE sysconf(_SC_PAGESIZE)
+#define INVALID_CONTAINER_FD -5
+#define THREE_PAGES 3
+#define UNALIGNED_ADDR 0x1500
+
+uint64_t virtaddr_64;
+const char *name = "heap";
+size_t map_length;
+int container_fds[RTE_MAX_VFIO_CONTAINERS];
+
+static int
+check_get_mem(void *addr, rte_iova_t *iova)
+{
+	const struct rte_memseg_list *msl;
+	const struct rte_memseg *ms;
+	rte_iova_t expected_iova;
+
+	msl = rte_mem_virt2memseg_list(addr);
+	if (!msl->external) {
+		printf("%s():%i: Memseg list is not marked as "
+				"external\n", __func__, __LINE__);
+		return -1;
+	}
+	ms = rte_mem_virt2memseg(addr, msl);
+	if (ms == NULL) {
+		printf("%s():%i: Failed to retrieve memseg for "
+				"external mem\n", __func__, __LINE__);
+		return -1;
+	}
+	if (ms->addr != addr) {
+		printf("%s():%i: VA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	expected_iova = (iova == NULL) ? RTE_BAD_IOVA : iova[0];
+	if (ms->iova != expected_iova) {
+		printf("%s():%i: IOVA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	return 0;
+}
+static int
+check_vfio_exist_and_initialize(void)
+{
+	int i = 0;
+
+	if (rte_vfio_is_enabled("vfio_pci") == 0) {
+		printf("VFIO is not enabled\n");
+		return TEST_SKIPPED;
+	}
+	if (rte_vfio_iommu_type_is_set() < 0) {
+		printf("VFIO IOMMU Type is not set\n");
+		return TEST_SKIPPED;
+	}
+
+	/* initialize_container_fds */;
+	for (i = 0; i < RTE_MAX_VFIO_CONTAINERS; i++)
+		container_fds[i] = -1;
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container create */
+static int
+test_vfio_container_create(void)
+{
+	int ret = 0, i = 0;
+
+	/* check max containers limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		container_fds[i] = rte_vfio_container_create();
+		TEST_ASSERT(container_fds[i] >  0, "Test to check "
+				"rte_vfio_container_create with max "
+				"containers limit: Failed\n");
+	}
+
+	/* check rte_vfio_container_create when exceeds max containers limit */
+	ret = rte_vfio_container_create();
+	TEST_ASSERT(ret == -1, "Test to check "
+			"rte_vfio_container_create container "
+			"when exceeds limit: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container destroy */
+static int
+test_vfio_container_destroy(void)
+{
+	int i = 0, ret = 0;
+
+	/* check to destroy max container limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		ret = rte_vfio_container_destroy(container_fds[i]);
+		TEST_ASSERT(ret == 0, "Test to check "
+				"rte_vfio_container_destroy: Failed\n");
+		container_fds[i] = -1;
+	}
+
+	/* check rte_vfio_container_destroy with valid but non existing value */
+	ret = rte_vfio_container_destroy(0);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy with "
+			"valid but non existing value: Failed\n");
+
+	/* check rte_vfio_container_destroy with invalid value */
+	ret = rte_vfio_container_destroy(-5);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy "
+			"with invalid value: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to bind a IOMMU group to a container*/
+static int
+test_rte_vfio_container_group_bind(void)
+{
+	int ret = 0;
+
+	/* Test case to bind with invalid container fd */
+	ret = rte_vfio_container_group_bind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to bind with non-existing container fd */
+	ret = rte_vfio_container_group_bind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unbind a IOMMU group from a container*/
+static int
+test_rte_vfio_container_group_unbind(void)
+{
+	int ret = 0;
+
+	/* Test case to unbind container from invalid group*/
+	ret = rte_vfio_container_group_unbind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to unbind container from group*/
+	ret = rte_vfio_container_group_unbind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with  non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to get IOMMU group number for a device*/
+static int
+test_rte_vfio_get_group_num(void)
+{
+	int ret = 0, invalid_group_num = 0;
+
+	/* Test case to get IOMMU group num from invalid group */
+	ret = rte_vfio_get_group_num(NULL, NULL, &invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	/* Test case to get IOMMU group num from invalid device address and
+	 * valid sysfs_base
+	 */
+	ret = rte_vfio_get_group_num("/sys/bus/pci/devices/", NULL,
+			&invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA mapping for devices in a container */
+static int
+test_rte_vfio_container_dma_map(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to map device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with invalid "
+			"container: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to map device for existing fd with no device attached and
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check  map device for existing fd "
+			"with no device attached and non-zero "
+			"map_length: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA unmapping for devices in a container*/
+static int
+test_rte_vfio_container_dma_unmap(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*Function to setup external memory */
+static int
+test_heap_mem_setup(size_t map_length, int n_pages)
+{
+	rte_iova_t iova[map_length / PAGESIZE];
+	void *addr;
+
+	addr = mmap(NULL, map_length, PROT_WRITE | PROT_READ,
+			MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+	if (addr == MAP_FAILED) {
+		printf("%s():%i: Failed to create dummy memory area\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	rte_iova_t tmp = 0x100000000 + PAGESIZE;
+	iova[0] = tmp;
+
+	if (rte_malloc_heap_create(name) != 0) {
+		printf("%s():%i: Failed to Create heap with valid name\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (rte_malloc_heap_memory_add(name, addr, map_length, iova, n_pages,
+				PAGESIZE) != 0) {
+		printf("%s():%i: Failed to add memory to heap\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (check_get_mem(addr, iova) != 0) {
+		printf("%s():%i: Failed to verify memory\n",
+				__func__, __LINE__);
+
+		return -1;
+	}
+	virtaddr_64 = (uint64_t)(uintptr_t)addr;
+
+	return 0;
+}
+
+/* Function to free the external memory */
+static void
+test_heap_mem_free(void)
+{
+	if (rte_malloc_heap_memory_remove(name, &virtaddr_64,
+				map_length) != 0) {
+		printf("%s():%i: Failed to remove memory\n",
+				__func__, __LINE__);
+		return;
+	}
+	rte_malloc_heap_destroy(name);
+
+	munmap(&virtaddr_64, map_length);
+}
+
+/* Test to map memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_map(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map memory for VFIO with zero  vaddr, iova addr
+	 * and map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with incorrect inputs: Failed\n");
+
+	/* Test case to map memory for VFIO with zero vaddr, iova addr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, unmapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(1000000, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"unmapped virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, mapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check already mapped virtual address */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check start virtual address + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + UNALIGNED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to check start virtual address before
+	 * existing map, overlaps
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNALIGNED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start virtual address "
+			"before existing map, overlaps: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNALIGNED_ADDR), 0, 500);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to check already mapped iova overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with already mapped iova overlaps: Failed\n");
+
+	/* Test case to check start iova + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNALIGNED_ADDR),
+			map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start iova + length range overlaps: Failed\n");
+
+	/* Test case to check invalid iova */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + 5000), map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid iova: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNALIGNED_ADDR), 100);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid vaddr, valid iova addr
+	 * and valid map_length
+	 */
+	uint64_t invalid_addr = virtaddr_64 + 1;
+	ret = rte_vfio_dma_map(invalid_addr, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with mapped invalid virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid iova addr, valid vaddr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, UNALIGNED_ADDR, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"invalid iova address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unmap memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_unmap(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with incorrect input: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero iova addr,
+	 * unmapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and unmapped addr: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with unmapped vaddr, iova
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"unmapped addr, iova: Failed\n");
+
+	/* Test case to map memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped invalid vaddr,
+	 * valid IOVA and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + 1), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"invalid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static int
+test_rte_vfio_dma_map_overlaps(void)
+{
+	int ret = 0;
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid address:Failed\n");
+
+	/* Test case to map same start virtual address and
+	 * extend beyond end virtual address
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and extend beyond end "
+			"virtual address: Failed\n");
+
+	/* Test case to map same start virtual address and same end address*/
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and "
+			"same end address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and "
+		"mapped valid virtual address: Failed\n");
+
+	/* Test case to map different virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			(PAGESIZE * 2));
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with different virtual address: Failed\n");
+
+	/* Test case to map different start virtual address and
+	 * ends with same address
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with different start virtual address and "
+			"ends with same address: Failed\n");
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address range: Failed\n");
+
+	/* Test case to map middle overlapping virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap 1st page: Failed\n");
+
+	/* Test case to map 1st and 2nd page overlaps */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st and 2nd page overlaps: Failed\n");
+
+	/* Test case to map 3rd and 4th pages */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 3rd and 4th pages: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 3rd page: Failed\n");
+
+	/* Test case to map 1st page with total length
+	 * that overlaps middle page
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st page with total length "
+			"that overlaps middle page: Failed\n");
+
+	/* Test case to unmap 2nd page  */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 2nd page: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*allocate three pages */
+static int
+test_rte_vfio_dma_map_threepages(void)
+{
+	int ret = 0;
+
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+	uint64_t page1_va, page2_va, page3_va;
+	rte_iova_t page1_iova, page2_iova, page3_iova;
+
+	page1_va = virtaddr_64;
+	page2_va = virtaddr_64 + PAGESIZE;
+	page3_va = virtaddr_64 + (PAGESIZE * 2);
+
+	page1_iova = 0;
+	page2_iova = 0 + PAGESIZE;
+	page3_iova = 0 + (PAGESIZE * 2);
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid 1st page map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 2nd page */
+	ret = rte_vfio_dma_unmap(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to map 2nd page */
+	ret = rte_vfio_dma_map(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 3rd page virtual address: Failed\n");
+
+	/* Test case to map 3rd page */
+	ret = rte_vfio_dma_map(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped 3rd page valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page, but used IOVA address of 2nd page */
+	ret = rte_vfio_dma_unmap(page1_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static struct
+unit_test_suite eal_vfio_testsuite  = {
+	.suite_name = "EAL VFIO Unit Test Suite",
+	.setup = check_vfio_exist_and_initialize,
+	.teardown = NULL,
+	.unit_test_cases = {
+		/* Test Case 1: To check vfio container create test cases */
+		TEST_CASE(test_vfio_container_create),
+
+		/* Test Case 2: To check vfio container destroy */
+		TEST_CASE(test_vfio_container_destroy),
+
+		/* Test Case 3: To  bind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_bind),
+
+		/* Test Case 4: To get IOMMU group number for a device*/
+		TEST_CASE(test_rte_vfio_get_group_num),
+
+		/* Test Case 5: To unbind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_unbind),
+
+		/* Test Case 6: To perform DMA mapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free, test_rte_vfio_dma_map),
+
+		/* Test Case 7: To perform DMA unmapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_unmap),
+
+		/* Test Case 8: To perform map devices in specific container */
+		TEST_CASE(test_rte_vfio_container_dma_map),
+
+		/* Test Case 9: To perform unmap devices in specific container
+		 */
+		TEST_CASE(test_rte_vfio_container_dma_unmap),
+
+		/* Test Case 10: To perform three pages */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_threepages),
+
+		/* Test Case 11: To check DMA overlaps */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_overlaps),
+
+		TEST_CASES_END()
+	}
+};
+
+static int
+test_eal_vfio(void)
+{
+	return unit_test_suite_runner(&eal_vfio_testsuite);
+}
+
+#endif
+
+REGISTER_TEST_COMMAND(eal_vfio_autotest, test_eal_vfio);
-- 
2.17.2

[dpdk-dev] [PATCH v3 1/4] lib/eal: fix vfio unmap that fails unexpectedly

[Chaitanya Babu Talluri]

Unmap fails when there are duplicate entries in user_mem_maps.

The fix is to validate if the input VA, IOVA exists or
overlaps in user_mem_maps before creating map.

Fixes: 73a63908 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 46 +++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 501c74f23..104912077 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -212,6 +212,41 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 	return NULL;
 }
 
+static int
+find_user_mem_map_overlap(struct user_mem_maps *user_mem_maps, uint64_t addr,
+		uint64_t iova, uint64_t len)
+{
+	uint64_t va_end = addr + len;
+	uint64_t iova_end = iova + len;
+	int i;
+
+	for (i = 0; i < user_mem_maps->n_maps; i++) {
+		struct user_mem_map *map = &user_mem_maps->maps[i];
+		uint64_t map_va_end = map->addr + map->len;
+		uint64_t map_iova_end = map->iova + map->len;
+
+		bool no_lo_va_overlap = addr < map->addr && va_end <= map->addr;
+		bool no_hi_va_overlap = addr >= map_va_end &&
+			va_end > map_va_end;
+		bool no_lo_iova_overlap = iova < map->iova &&
+			iova_end <= map->iova;
+		bool no_hi_iova_overlap = iova >= map_iova_end &&
+			iova_end > map_iova_end;
+
+		/* check input VA and iova is not within the
+		 * existing map's range
+		 */
+		if ((no_lo_va_overlap || no_hi_va_overlap) &&
+				(no_lo_iova_overlap || no_hi_iova_overlap))
+			continue;
+		else
+			/* map overlaps */
+			return 1;
+	}
+	/* map doesn't overlap */
+	return 0;
+}
+
 /* this will sort all user maps, and merge/compact any adjacent maps */
 static void
 compact_user_maps(struct user_mem_maps *user_mem_maps)
@@ -1732,6 +1767,17 @@ container_dma_map(struct vfio_config *vfio_cfg, uint64_t vaddr, uint64_t iova,
 		ret = -1;
 		goto out;
 	}
+
+	/* check whether vaddr and iova exists in user_mem_maps */
+	ret = find_user_mem_map_overlap(user_mem_maps, vaddr, iova, len);
+	if (ret) {
+		RTE_LOG(ERR, EAL, "Mapping overlaps with a previously "
+				"existing mapping\n");
+		rte_errno = EEXIST;
+		ret = -1;
+		goto out;
+	}
+
 	/* map the entry */
 	if (vfio_dma_mem_map(vfio_cfg, vaddr, iova, len, 1)) {
 		/* technically, this will fail if there are currently no devices
-- 
2.17.2

[dpdk-dev] [PATCH v3 2/4] lib/eal: fix vfio unmap that succeeds unexpectedly

[Chaitanya Babu Talluri]

Unmapping page with a VA that is found in the list
of current mappings will succeed even if the IOVA
for the chunk that is being unmapped,is mismatched.

Fix it by checking if IOVA address matches the
expected IOVA address exactly.

Fixes: 73a6390859 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 104912077..04c284cb2 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -184,13 +184,13 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		uint64_t iova, uint64_t len)
 {
 	uint64_t va_end = addr + len;
-	uint64_t iova_end = iova + len;
 	int i;
 
 	for (i = 0; i < user_mem_maps->n_maps; i++) {
 		struct user_mem_map *map = &user_mem_maps->maps[i];
 		uint64_t map_va_end = map->addr + map->len;
-		uint64_t map_iova_end = map->iova + map->len;
+		uint64_t diff_addr_len = addr - map->addr;
+		uint64_t expected_iova = map->iova + diff_addr_len;
 
 		/* check start VA */
 		if (addr < map->addr || addr >= map_va_end)
@@ -199,11 +199,10 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		if (va_end <= map->addr || va_end > map_va_end)
 			continue;
 
-		/* check start IOVA */
-		if (iova < map->iova || iova >= map_iova_end)
-			continue;
-		/* check if IOVA end is within boundaries */
-		if (iova_end <= map->iova || iova_end > map_iova_end)
+		/* check whether user input iova is in sync with
+		 * user_mem_map entry's iova
+		 */
+		if (expected_iova != iova)
 			continue;
 
 		/* we've found our map */
-- 
2.17.2

[dpdk-dev] [PATCH v3 3/4] lib/eal: add API to check iommu type is set

[Chaitanya Babu Talluri]

Add rte_vfio_iommu_type_is_set() to check IOMMU
type for default container.

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/common/include/rte_vfio.h | 10 ++++++++++
 lib/librte_eal/linux/eal/eal_vfio.c      | 16 ++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/lib/librte_eal/common/include/rte_vfio.h b/lib/librte_eal/common/include/rte_vfio.h
index b360485fa..a62006e5a 100644
--- a/lib/librte_eal/common/include/rte_vfio.h
+++ b/lib/librte_eal/common/include/rte_vfio.h
@@ -397,6 +397,16 @@ int
 rte_vfio_container_dma_unmap(int container_fd, uint64_t vaddr,
 		uint64_t iova, uint64_t len);
 
+/**
+ * Check VFIO IOMMU Type is set for default container.
+ *
+ * @return
+ *    0 if successful
+ *   <0 if failed
+ */
+int
+rte_vfio_iommu_type_is_set(void);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 04c284cb2..a5bb1cff4 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -2071,6 +2071,17 @@ rte_vfio_container_dma_unmap(int container_fd, uint64_t vaddr, uint64_t iova,
 	return container_dma_unmap(vfio_cfg, vaddr, iova, len);
 }
 
+int
+rte_vfio_iommu_type_is_set(void)
+{
+	if (vfio_get_iommu_type() < 0) {
+		RTE_LOG(ERR, EAL, "VFIO IOMMU Type is not set\n");
+		return -1;
+	}
+
+	return 0;
+}
+
 #else
 
 int
@@ -2191,4 +2202,9 @@ rte_vfio_container_dma_unmap(__rte_unused int container_fd,
 	return -1;
 }
 
+int
+rte_vfio_iommu_type_is_set(void)
+{
+	return -1;
+}
 #endif /* VFIO_PRESENT */
-- 
2.17.2

[dpdk-dev] [PATCH v3 4/4] app/test: add unit tests for eal vfio

[Chaitanya Babu Talluri]

Unit test cases are added for eal vfio library.
eal_vfio_autotest added to meson build file.

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 app/test/Makefile        |   1 +
 app/test/meson.build     |   2 +
 app/test/test_eal_vfio.c | 736 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 739 insertions(+)
 create mode 100644 app/test/test_eal_vfio.c

diff --git a/app/test/Makefile b/app/test/Makefile
index 26ba6fe2b..9b9c78b4e 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -137,6 +137,7 @@ SRCS-y += test_cpuflags.c
 SRCS-y += test_mp_secondary.c
 SRCS-y += test_eal_flags.c
 SRCS-y += test_eal_fs.c
+SRCS-y += test_eal_vfio.c
 SRCS-y += test_alarm.c
 SRCS-y += test_interrupts.c
 SRCS-y += test_version.c
diff --git a/app/test/meson.build b/app/test/meson.build
index ec40943bd..bd96ebb2b 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -36,6 +36,7 @@ test_sources = files('commands.c',
 	'test_distributor_perf.c',
 	'test_eal_flags.c',
 	'test_eal_fs.c',
+	'test_eal_vfio.c',
 	'test_efd.c',
 	'test_efd_perf.c',
 	'test_errno.c',
@@ -175,6 +176,7 @@ fast_test_names = [
         'eal_flags_file_prefix_autotest',
         'eal_flags_misc_autotest',
         'eal_fs_autotest',
+        'eal_vfio_autotest',
         'errno_autotest',
         'event_ring_autotest',
         'func_reentrancy_autotest',
diff --git a/app/test/test_eal_vfio.c b/app/test/test_eal_vfio.c
new file mode 100644
index 000000000..ca3efb034
--- /dev/null
+++ b/app/test/test_eal_vfio.c
@@ -0,0 +1,736 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2019 Intel Corporation
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <stdint.h>
+#include <errno.h>
+#include <sys/mman.h>
+#include <unistd.h>
+#include <rte_vfio.h>
+#include <rte_malloc.h>
+#include <rte_eal_memconfig.h>
+
+#include "test.h"
+
+#if !defined(RTE_EXEC_ENV_LINUX) || !defined(RTE_EAL_VFIO)
+static int
+test_eal_vfio(void)
+{
+	printf("VFIO not supported, skipping test\n");
+	return TEST_SKIPPED;
+}
+
+#else
+
+#define PAGESIZE sysconf(_SC_PAGESIZE)
+#define INVALID_CONTAINER_FD -5
+#define THREE_PAGES 3
+#define UNALIGNED_ADDR 0x1500
+
+uint64_t virtaddr_64;
+const char *name = "heap";
+size_t map_length;
+int container_fds[RTE_MAX_VFIO_CONTAINERS];
+
+static int
+check_get_mem(void *addr, rte_iova_t *iova)
+{
+	const struct rte_memseg_list *msl;
+	const struct rte_memseg *ms;
+	rte_iova_t expected_iova;
+
+	msl = rte_mem_virt2memseg_list(addr);
+	if (!msl->external) {
+		printf("%s():%i: Memseg list is not marked as "
+				"external\n", __func__, __LINE__);
+		return -1;
+	}
+	ms = rte_mem_virt2memseg(addr, msl);
+	if (ms == NULL) {
+		printf("%s():%i: Failed to retrieve memseg for "
+				"external mem\n", __func__, __LINE__);
+		return -1;
+	}
+	if (ms->addr != addr) {
+		printf("%s():%i: VA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	expected_iova = (iova == NULL) ? RTE_BAD_IOVA : iova[0];
+	if (ms->iova != expected_iova) {
+		printf("%s():%i: IOVA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	return 0;
+}
+static int
+check_vfio_exist_and_initialize(void)
+{
+	int i = 0;
+
+	if (rte_vfio_is_enabled("vfio_pci") == 0) {
+		printf("VFIO is not enabled\n");
+		return TEST_SKIPPED;
+	}
+	if (rte_vfio_iommu_type_is_set() < 0) {
+		printf("VFIO IOMMU Type is not set\n");
+		return TEST_SKIPPED;
+	}
+
+	/* initialize_container_fds */;
+	for (i = 0; i < RTE_MAX_VFIO_CONTAINERS; i++)
+		container_fds[i] = -1;
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container create */
+static int
+test_vfio_container_create(void)
+{
+	int ret = 0, i = 0;
+
+	/* check max containers limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		container_fds[i] = rte_vfio_container_create();
+		TEST_ASSERT(container_fds[i] >  0, "Test to check "
+				"rte_vfio_container_create with max "
+				"containers limit: Failed\n");
+	}
+
+	/* check rte_vfio_container_create when exceeds max containers limit */
+	ret = rte_vfio_container_create();
+	TEST_ASSERT(ret == -1, "Test to check "
+			"rte_vfio_container_create container "
+			"when exceeds limit: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container destroy */
+static int
+test_vfio_container_destroy(void)
+{
+	int i = 0, ret = 0;
+
+	/* check to destroy max container limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		ret = rte_vfio_container_destroy(container_fds[i]);
+		TEST_ASSERT(ret == 0, "Test to check "
+				"rte_vfio_container_destroy: Failed\n");
+		container_fds[i] = -1;
+	}
+
+	/* check rte_vfio_container_destroy with valid but non existing value */
+	ret = rte_vfio_container_destroy(0);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy with "
+			"valid but non existing value: Failed\n");
+
+	/* check rte_vfio_container_destroy with invalid value */
+	ret = rte_vfio_container_destroy(-5);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy "
+			"with invalid value: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to bind a IOMMU group to a container*/
+static int
+test_rte_vfio_container_group_bind(void)
+{
+	int ret = 0;
+
+	/* Test case to bind with invalid container fd */
+	ret = rte_vfio_container_group_bind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to bind with non-existing container fd */
+	ret = rte_vfio_container_group_bind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unbind a IOMMU group from a container*/
+static int
+test_rte_vfio_container_group_unbind(void)
+{
+	int ret = 0;
+
+	/* Test case to unbind container from invalid group*/
+	ret = rte_vfio_container_group_unbind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to unbind container from group*/
+	ret = rte_vfio_container_group_unbind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with  non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to get IOMMU group number for a device*/
+static int
+test_rte_vfio_get_group_num(void)
+{
+	int ret = 0, invalid_group_num = 0;
+
+	/* Test case to get IOMMU group num from invalid group */
+	ret = rte_vfio_get_group_num(NULL, NULL, &invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	/* Test case to get IOMMU group num from invalid device address and
+	 * valid sysfs_base
+	 */
+	ret = rte_vfio_get_group_num("/sys/bus/pci/devices/", NULL,
+			&invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA mapping for devices in a container */
+static int
+test_rte_vfio_container_dma_map(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to map device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with invalid "
+			"container: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to map device for existing fd with no device attached and
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check  map device for existing fd "
+			"with no device attached and non-zero "
+			"map_length: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA unmapping for devices in a container*/
+static int
+test_rte_vfio_container_dma_unmap(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*Function to setup external memory */
+static int
+test_heap_mem_setup(size_t map_length, int n_pages)
+{
+	rte_iova_t iova[map_length / PAGESIZE];
+	void *addr;
+
+	addr = mmap(NULL, map_length, PROT_WRITE | PROT_READ,
+			MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+	if (addr == MAP_FAILED) {
+		printf("%s():%i: Failed to create dummy memory area\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	rte_iova_t tmp = 0x100000000 + PAGESIZE;
+	iova[0] = tmp;
+
+	if (rte_malloc_heap_create(name) != 0) {
+		printf("%s():%i: Failed to Create heap with valid name\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (rte_malloc_heap_memory_add(name, addr, map_length, iova, n_pages,
+				PAGESIZE) != 0) {
+		printf("%s():%i: Failed to add memory to heap\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (check_get_mem(addr, iova) != 0) {
+		printf("%s():%i: Failed to verify memory\n",
+				__func__, __LINE__);
+
+		return -1;
+	}
+	virtaddr_64 = (uint64_t)(uintptr_t)addr;
+
+	return 0;
+}
+
+/* Function to free the external memory */
+static void
+test_heap_mem_free(void)
+{
+	if (rte_malloc_heap_memory_remove(name, (void *)virtaddr_64,
+				map_length) != 0) {
+		printf("%s():%i: Failed to remove memory\n",
+				__func__, __LINE__);
+		return;
+	}
+	rte_malloc_heap_destroy(name);
+
+	munmap((void *)virtaddr_64, map_length);
+}
+
+/* Test to map memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_map(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map memory for VFIO with zero  vaddr, iova addr
+	 * and map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with incorrect inputs: Failed\n");
+
+	/* Test case to map memory for VFIO with zero vaddr, iova addr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, unmapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(1000000, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"unmapped virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, mapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check already mapped virtual address */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check start virtual address + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + UNALIGNED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to check start virtual address before
+	 * existing map, overlaps
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNALIGNED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start virtual address "
+			"before existing map, overlaps: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNALIGNED_ADDR), 0, 500);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to check already mapped iova overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with already mapped iova overlaps: Failed\n");
+
+	/* Test case to check start iova + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNALIGNED_ADDR),
+			map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start iova + length range overlaps: Failed\n");
+
+	/* Test case to check invalid iova */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + 5000), map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid iova: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNALIGNED_ADDR), 100);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid vaddr, valid iova addr
+	 * and valid map_length
+	 */
+	uint64_t invalid_addr = virtaddr_64 + 1;
+	ret = rte_vfio_dma_map(invalid_addr, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with mapped invalid virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid iova addr, valid vaddr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, UNALIGNED_ADDR, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"invalid iova address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unmap memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_unmap(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with incorrect input: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero iova addr,
+	 * unmapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and unmapped addr: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with unmapped vaddr, iova
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"unmapped addr, iova: Failed\n");
+
+	/* Test case to map memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped invalid vaddr,
+	 * valid IOVA and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + 1), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"invalid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static int
+test_rte_vfio_dma_map_overlaps(void)
+{
+	int ret = 0;
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid address:Failed\n");
+
+	/* Test case to map same start virtual address and
+	 * extend beyond end virtual address
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and extend beyond end "
+			"virtual address: Failed\n");
+
+	/* Test case to map same start virtual address and same end address*/
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and "
+			"same end address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and "
+		"mapped valid virtual address: Failed\n");
+
+	/* Test case to map different virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			(PAGESIZE * 2));
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with different virtual address: Failed\n");
+
+	/* Test case to map different start virtual address and
+	 * ends with same address
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with different start virtual address and "
+			"ends with same address: Failed\n");
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address range: Failed\n");
+
+	/* Test case to map middle overlapping virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap 1st page: Failed\n");
+
+	/* Test case to map 1st and 2nd page overlaps */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st and 2nd page overlaps: Failed\n");
+
+	/* Test case to map 3rd and 4th pages */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 3rd and 4th pages: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 3rd page: Failed\n");
+
+	/* Test case to map 1st page with total length
+	 * that overlaps middle page
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st page with total length "
+			"that overlaps middle page: Failed\n");
+
+	/* Test case to unmap 2nd page  */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 2nd page: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*allocate three pages */
+static int
+test_rte_vfio_dma_map_threepages(void)
+{
+	int ret = 0;
+
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+	uint64_t page1_va, page2_va, page3_va;
+	rte_iova_t page1_iova, page2_iova, page3_iova;
+
+	page1_va = virtaddr_64;
+	page2_va = virtaddr_64 + PAGESIZE;
+	page3_va = virtaddr_64 + (PAGESIZE * 2);
+
+	page1_iova = 0;
+	page2_iova = 0 + PAGESIZE;
+	page3_iova = 0 + (PAGESIZE * 2);
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid 1st page map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 2nd page */
+	ret = rte_vfio_dma_unmap(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to map 2nd page */
+	ret = rte_vfio_dma_map(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 3rd page virtual address: Failed\n");
+
+	/* Test case to map 3rd page */
+	ret = rte_vfio_dma_map(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped 3rd page valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page, but used IOVA address of 2nd page */
+	ret = rte_vfio_dma_unmap(page1_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static struct
+unit_test_suite eal_vfio_testsuite  = {
+	.suite_name = "EAL VFIO Unit Test Suite",
+	.setup = check_vfio_exist_and_initialize,
+	.teardown = NULL,
+	.unit_test_cases = {
+		/* Test Case 1: To check vfio container create test cases */
+		TEST_CASE(test_vfio_container_create),
+
+		/* Test Case 2: To check vfio container destroy */
+		TEST_CASE(test_vfio_container_destroy),
+
+		/* Test Case 3: To  bind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_bind),
+
+		/* Test Case 4: To get IOMMU group number for a device*/
+		TEST_CASE(test_rte_vfio_get_group_num),
+
+		/* Test Case 5: To unbind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_unbind),
+
+		/* Test Case 6: To perform DMA mapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free, test_rte_vfio_dma_map),
+
+		/* Test Case 7: To perform DMA unmapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_unmap),
+
+		/* Test Case 8: To perform map devices in specific container */
+		TEST_CASE(test_rte_vfio_container_dma_map),
+
+		/* Test Case 9: To perform unmap devices in specific container
+		 */
+		TEST_CASE(test_rte_vfio_container_dma_unmap),
+
+		/* Test Case 10: To perform three pages */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_threepages),
+
+		/* Test Case 11: To check DMA overlaps */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_overlaps),
+
+		TEST_CASES_END()
+	}
+};
+
+static int
+test_eal_vfio(void)
+{
+	return unit_test_suite_runner(&eal_vfio_testsuite);
+}
+
+#endif
+
+REGISTER_TEST_COMMAND(eal_vfio_autotest, test_eal_vfio);
-- 
2.17.2

[dpdk-dev] [PATCH v2 1/3] lib/eal: fix vfio unmap that fails unexpectedly

[Chaitanya Babu Talluri]

Unmap fails when there are duplicate entries in user_mem_maps.

The fix is to validate if the input VA, IOVA exists or
overlaps in user_mem_maps before creating map.

Fixes: 73a63908 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 46 +++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 501c74f23..104912077 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -212,6 +212,41 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 	return NULL;
 }
 
+static int
+find_user_mem_map_overlap(struct user_mem_maps *user_mem_maps, uint64_t addr,
+		uint64_t iova, uint64_t len)
+{
+	uint64_t va_end = addr + len;
+	uint64_t iova_end = iova + len;
+	int i;
+
+	for (i = 0; i < user_mem_maps->n_maps; i++) {
+		struct user_mem_map *map = &user_mem_maps->maps[i];
+		uint64_t map_va_end = map->addr + map->len;
+		uint64_t map_iova_end = map->iova + map->len;
+
+		bool no_lo_va_overlap = addr < map->addr && va_end <= map->addr;
+		bool no_hi_va_overlap = addr >= map_va_end &&
+			va_end > map_va_end;
+		bool no_lo_iova_overlap = iova < map->iova &&
+			iova_end <= map->iova;
+		bool no_hi_iova_overlap = iova >= map_iova_end &&
+			iova_end > map_iova_end;
+
+		/* check input VA and iova is not within the
+		 * existing map's range
+		 */
+		if ((no_lo_va_overlap || no_hi_va_overlap) &&
+				(no_lo_iova_overlap || no_hi_iova_overlap))
+			continue;
+		else
+			/* map overlaps */
+			return 1;
+	}
+	/* map doesn't overlap */
+	return 0;
+}
+
 /* this will sort all user maps, and merge/compact any adjacent maps */
 static void
 compact_user_maps(struct user_mem_maps *user_mem_maps)
@@ -1732,6 +1767,17 @@ container_dma_map(struct vfio_config *vfio_cfg, uint64_t vaddr, uint64_t iova,
 		ret = -1;
 		goto out;
 	}
+
+	/* check whether vaddr and iova exists in user_mem_maps */
+	ret = find_user_mem_map_overlap(user_mem_maps, vaddr, iova, len);
+	if (ret) {
+		RTE_LOG(ERR, EAL, "Mapping overlaps with a previously "
+				"existing mapping\n");
+		rte_errno = EEXIST;
+		ret = -1;
+		goto out;
+	}
+
 	/* map the entry */
 	if (vfio_dma_mem_map(vfio_cfg, vaddr, iova, len, 1)) {
 		/* technically, this will fail if there are currently no devices
-- 
2.17.2

[dpdk-dev] [PATCH v2 2/3] lib/eal: fix vfio unmap that succeeds unexpectedly

[Chaitanya Babu Talluri]

Unmapping page with a VA that is found in the list
of current mappings will succeed even if the IOVA
for the chunk that is being unmapped,is mismatched.

Fix it by checking if IOVA address matches the
expected IOVA address exactly.

Fixes: 73a6390859 ("vfio: allow to map other memory regions")
Cc: stable@dpdk.org

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 lib/librte_eal/linux/eal/eal_vfio.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
index 104912077..04c284cb2 100644
--- a/lib/librte_eal/linux/eal/eal_vfio.c
+++ b/lib/librte_eal/linux/eal/eal_vfio.c
@@ -184,13 +184,13 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		uint64_t iova, uint64_t len)
 {
 	uint64_t va_end = addr + len;
-	uint64_t iova_end = iova + len;
 	int i;
 
 	for (i = 0; i < user_mem_maps->n_maps; i++) {
 		struct user_mem_map *map = &user_mem_maps->maps[i];
 		uint64_t map_va_end = map->addr + map->len;
-		uint64_t map_iova_end = map->iova + map->len;
+		uint64_t diff_addr_len = addr - map->addr;
+		uint64_t expected_iova = map->iova + diff_addr_len;
 
 		/* check start VA */
 		if (addr < map->addr || addr >= map_va_end)
@@ -199,11 +199,10 @@ find_user_mem_map(struct user_mem_maps *user_mem_maps, uint64_t addr,
 		if (va_end <= map->addr || va_end > map_va_end)
 			continue;
 
-		/* check start IOVA */
-		if (iova < map->iova || iova >= map_iova_end)
-			continue;
-		/* check if IOVA end is within boundaries */
-		if (iova_end <= map->iova || iova_end > map_iova_end)
+		/* check whether user input iova is in sync with
+		 * user_mem_map entry's iova
+		 */
+		if (expected_iova != iova)
 			continue;
 
 		/* we've found our map */
-- 
2.17.2

[dpdk-dev] [PATCH v2 3/3] app/test: add unit tests for eal vfio

[Chaitanya Babu Talluri]

Unit test cases are added for eal vfio library.
eal_vfio_autotest added to meson build file.

Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
---
 app/test/Makefile        |   1 +
 app/test/meson.build     |   2 +
 app/test/test_eal_vfio.c | 728 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 731 insertions(+)
 create mode 100644 app/test/test_eal_vfio.c

diff --git a/app/test/Makefile b/app/test/Makefile
index 26ba6fe2b..9b9c78b4e 100644
--- a/app/test/Makefile
+++ b/app/test/Makefile
@@ -137,6 +137,7 @@ SRCS-y += test_cpuflags.c
 SRCS-y += test_mp_secondary.c
 SRCS-y += test_eal_flags.c
 SRCS-y += test_eal_fs.c
+SRCS-y += test_eal_vfio.c
 SRCS-y += test_alarm.c
 SRCS-y += test_interrupts.c
 SRCS-y += test_version.c
diff --git a/app/test/meson.build b/app/test/meson.build
index ec40943bd..bd96ebb2b 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -36,6 +36,7 @@ test_sources = files('commands.c',
 	'test_distributor_perf.c',
 	'test_eal_flags.c',
 	'test_eal_fs.c',
+	'test_eal_vfio.c',
 	'test_efd.c',
 	'test_efd_perf.c',
 	'test_errno.c',
@@ -175,6 +176,7 @@ fast_test_names = [
         'eal_flags_file_prefix_autotest',
         'eal_flags_misc_autotest',
         'eal_fs_autotest',
+        'eal_vfio_autotest',
         'errno_autotest',
         'event_ring_autotest',
         'func_reentrancy_autotest',
diff --git a/app/test/test_eal_vfio.c b/app/test/test_eal_vfio.c
new file mode 100644
index 000000000..8995573df
--- /dev/null
+++ b/app/test/test_eal_vfio.c
@@ -0,0 +1,728 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2019 Intel Corporation
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <stdint.h>
+#include <errno.h>
+#include <sys/mman.h>
+#include <unistd.h>
+#include <rte_vfio.h>
+#include <rte_malloc.h>
+#include <rte_eal_memconfig.h>
+
+#include "test.h"
+
+#if !defined(RTE_EXEC_ENV_LINUX) || !defined(RTE_EAL_VFIO)
+static int
+test_eal_vfio(void)
+{
+	printf("VFIO not supported, skipping test\n");
+	return TEST_SKIPPED;
+}
+
+#else
+
+#define PAGESIZE sysconf(_SC_PAGESIZE)
+#define INVALID_CONTAINER_FD -5
+#define THREE_PAGES 3
+#define UNMAPPED_ADDR 0x1500
+
+uint64_t virtaddr_64;
+const char *name = "heap";
+size_t map_length;
+int container_fds[RTE_MAX_VFIO_CONTAINERS];
+
+static int
+check_get_mem(void *addr, rte_iova_t *iova)
+{
+	const struct rte_memseg_list *msl;
+	const struct rte_memseg *ms;
+	rte_iova_t expected_iova;
+
+	msl = rte_mem_virt2memseg_list(addr);
+	if (!msl->external) {
+		printf("%s():%i: Memseg list is not marked as "
+				"external\n", __func__, __LINE__);
+		return -1;
+	}
+	ms = rte_mem_virt2memseg(addr, msl);
+	if (ms == NULL) {
+		printf("%s():%i: Failed to retrieve memseg for "
+				"external mem\n", __func__, __LINE__);
+		return -1;
+	}
+	if (ms->addr != addr) {
+		printf("%s():%i: VA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	expected_iova = (iova == NULL) ? RTE_BAD_IOVA : iova[0];
+	if (ms->iova != expected_iova) {
+		printf("%s():%i: IOVA mismatch\n", __func__, __LINE__);
+		return -1;
+	}
+	return 0;
+}
+
+/* Initialize container fds */
+static int
+initialize_container_fds(void)
+{
+	int i = 0;
+
+	for (i = 0; i < RTE_MAX_VFIO_CONTAINERS; i++)
+		container_fds[i] = -1;
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container create */
+static int
+test_vfio_container_create(void)
+{
+	int ret = 0, i = 0;
+
+	/* check max containers limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		container_fds[i] = rte_vfio_container_create();
+		TEST_ASSERT(container_fds[i] >  0, "Test to check "
+				"rte_vfio_container_create with max "
+				"containers limit: Failed\n");
+	}
+
+	/* check rte_vfio_container_create when exceeds max containers limit */
+	ret = rte_vfio_container_create();
+	TEST_ASSERT(ret == -1, "Test to check "
+			"rte_vfio_container_create container "
+			"when exceeds limit: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* To test vfio container destroy */
+static int
+test_vfio_container_destroy(void)
+{
+	int i = 0, ret = 0;
+
+	/* check to destroy max container limit */
+	for (i = 1; i < RTE_MAX_VFIO_CONTAINERS; i++) {
+		ret = rte_vfio_container_destroy(container_fds[i]);
+		TEST_ASSERT(ret == 0, "Test to check "
+				"rte_vfio_container_destroy: Failed\n");
+		container_fds[i] = -1;
+	}
+
+	/* check rte_vfio_container_destroy with valid but non existing value */
+	ret = rte_vfio_container_destroy(0);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy with "
+			"valid but non existing value: Failed\n");
+
+	/* check rte_vfio_container_destroy with invalid value */
+	ret = rte_vfio_container_destroy(-5);
+	TEST_ASSERT(ret == -1, "Test to check rte_vfio_container_destroy "
+			"with invalid value: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to bind a IOMMU group to a container*/
+static int
+test_rte_vfio_container_group_bind(void)
+{
+	int ret = 0;
+
+	/* Test case to bind with invalid container fd */
+	ret = rte_vfio_container_group_bind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to bind with non-existing container fd */
+	ret = rte_vfio_container_group_bind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to bind a IOMMU group to a container "
+			"with non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unbind a IOMMU group from a container*/
+static int
+test_rte_vfio_container_group_unbind(void)
+{
+	int ret = 0;
+
+	/* Test case to unbind container from invalid group*/
+	ret = rte_vfio_container_group_unbind(INVALID_CONTAINER_FD, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with invalid fd: Failed\n");
+
+	/* Test case to unbind container from group*/
+	ret = rte_vfio_container_group_unbind(0, 0);
+	TEST_ASSERT(ret == -1, "Test to unbind a IOMMU group to a container "
+			"with  non existing fd: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to get IOMMU group number for a device*/
+static int
+test_rte_vfio_get_group_num(void)
+{
+	int ret = 0, invalid_group_num = 0;
+
+	/* Test case to get IOMMU group num from invalid group */
+	ret = rte_vfio_get_group_num(NULL, NULL, &invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	/* Test case to get IOMMU group num from invalid device address and
+	 * valid sysfs_base
+	 */
+	ret = rte_vfio_get_group_num("/sys/bus/pci/devices/", NULL,
+			&invalid_group_num);
+	TEST_ASSERT(ret == 0, "Test to get IOMMU group num: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA mapping for devices in a container */
+static int
+test_rte_vfio_container_dma_map(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to map device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with invalid "
+			"container: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to map device for existing fd with no device attached and
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_map(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check  map device for existing fd "
+			"with no device attached and non-zero "
+			"map_length: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to perform DMA unmapping for devices in a container*/
+static int
+test_rte_vfio_container_dma_unmap(void)
+{
+	int ret = 0, container_fd;
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for non-existing container_fd, with
+	 * non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(0, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with non-existing "
+			"container fd: Failed\n");
+
+	container_fd = rte_vfio_container_create();
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with non-zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test case to unmap device for existing fd with no device attached
+	 * and with zero map_length
+	 */
+	ret = rte_vfio_container_dma_unmap(container_fd, 0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to check map device with unmapped "
+			"container fd: Failed\n");
+
+	/* Test to destroy for container fd */
+	ret = rte_vfio_container_destroy(container_fd);
+	TEST_ASSERT(ret == 0, "Container fd destroy failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*Function to setup external memory */
+static int
+test_heap_mem_setup(size_t map_length, int n_pages)
+{
+	rte_iova_t iova[map_length / PAGESIZE];
+	void *addr;
+
+	addr = mmap(NULL, map_length, PROT_WRITE | PROT_READ,
+			MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+	if (addr == MAP_FAILED) {
+		printf("%s():%i: Failed to create dummy memory area\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	rte_iova_t tmp = 0x100000000 + PAGESIZE;
+	iova[0] = tmp;
+
+	if (rte_malloc_heap_create(name) != 0) {
+		printf("%s():%i: Failed to Create heap with valid name\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (rte_malloc_heap_memory_add(name, addr, map_length, iova, n_pages,
+				PAGESIZE) != 0) {
+		printf("%s():%i: Failed to add memory to heap\n",
+				__func__, __LINE__);
+		return -1;
+	}
+	if (check_get_mem(addr, iova) != 0) {
+		printf("%s():%i: Failed to verify memory\n",
+				__func__, __LINE__);
+
+		return -1;
+	}
+	virtaddr_64 = (uint64_t)(uintptr_t)addr;
+
+	return 0;
+}
+
+/* Function to free the external memory */
+static void
+test_heap_mem_free(void)
+{
+	if (rte_malloc_heap_memory_remove(name, (void *)virtaddr_64,
+				map_length) != 0) {
+		printf("%s():%i: Failed to remove memory\n",
+				__func__, __LINE__);
+		return;
+	}
+	rte_malloc_heap_destroy(name);
+
+	munmap((void *)virtaddr_64, map_length);
+}
+
+/* Test to map memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_map(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map memory for VFIO with zero  vaddr, iova addr
+	 * and map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with incorrect inputs: Failed\n");
+
+	/* Test case to map memory for VFIO with zero vaddr, iova addr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, unmapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(1000000, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"unmapped virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with valid iova addr, mapped
+	 * vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check already mapped virtual address */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to check start virtual address + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + UNMAPPED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to check start virtual address before
+	 * existing map, overlaps
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNMAPPED_ADDR), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start virtual address "
+			"before existing map, overlaps: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 - UNMAPPED_ADDR), 0, 500);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to check already mapped iova overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with already mapped iova overlaps: Failed\n");
+
+	/* Test case to check start iova + length range overlaps */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNMAPPED_ADDR),
+			map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with start iova + length range overlaps: Failed\n");
+
+	/* Test case to check invalid iova */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + 5000), map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid iova: Failed\n");
+
+	/* Test case to check invalid map length */
+	ret = rte_vfio_dma_map((virtaddr_64 + 8192), (0 + UNMAPPED_ADDR), 100);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with invalid map length: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid vaddr, valid iova addr
+	 * and valid map_length
+	 */
+	uint64_t invalid_addr = virtaddr_64 + 1;
+	ret = rte_vfio_dma_map(invalid_addr, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with mapped invalid virtual address: Failed\n");
+
+	/* Test case to map memory for VFIO with invalid iova addr, valid vaddr
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, UNMAPPED_ADDR, map_length);
+	TEST_ASSERT(ret == -1, "Test to map devices within default container "
+			"with valid map_length and "
+			"invalid iova address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/* Test to unmap memory region for use with VFIO*/
+static int
+test_rte_vfio_dma_unmap(void)
+{
+	int ret = 0;
+
+	const int n_pages = 1;
+	map_length = PAGESIZE;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, 0);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with incorrect input: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero vaddr,
+	 * iova addr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(0, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with zero iova addr,
+	 * unmapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and unmapped addr: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with unmapped vaddr, iova
+	 * and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, virtaddr_64, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"unmapped addr, iova: Failed\n");
+
+	/* Test case to map memory region from VFIO with valid iova,
+	 * mapped vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped invalid vaddr,
+	 * valid IOVA and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + 1), 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"invalid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static int
+test_rte_vfio_dma_map_overlaps(void)
+{
+	int ret = 0;
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid address:Failed\n");
+
+	/* Test case to map same start virtual address and
+	 * extend beyond end virtual address
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and extend beyond end "
+			"virtual address: Failed\n");
+
+	/* Test case to map same start virtual address and same end address*/
+	ret = rte_vfio_dma_map(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with same start virtual address and "
+			"same end address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and "
+		"mapped valid virtual address: Failed\n");
+
+	/* Test case to map different virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			(PAGESIZE * 2));
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with different virtual address: Failed\n");
+
+	/* Test case to map different start virtual address and
+	 * ends with same address
+	 */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with different start virtual address and "
+			"ends with same address: Failed\n");
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address range: Failed\n");
+
+	/* Test case to map middle overlapping virtual address */
+	ret = rte_vfio_dma_map((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with overlapping virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(virtaddr_64, 0, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap 1st page: Failed\n");
+
+	/* Test case to map 1st and 2nd page overlaps */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st and 2nd page overlaps: Failed\n");
+
+	/* Test case to map 3rd and 4th pages */
+	ret = rte_vfio_dma_map((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), (PAGESIZE * 2));
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 3rd and 4th pages: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + (PAGESIZE * 2)),
+			(0 + (PAGESIZE * 2)), PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 3rd page: Failed\n");
+
+	/* Test case to map 1st page with total length
+	 * that overlaps middle page
+	 */
+	ret = rte_vfio_dma_map(virtaddr_64, 0, map_length);
+	TEST_ASSERT(ret == -1, "Test to map device in default container "
+			"with 1st page with total length "
+			"that overlaps middle page: Failed\n");
+
+	/* Test case to unmap 2nd page  */
+	ret = rte_vfio_dma_unmap((virtaddr_64 + PAGESIZE), (0 + PAGESIZE),
+			PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap 2nd page: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+/*allocate three pages */
+static int
+test_rte_vfio_dma_map_threepages(void)
+{
+	int ret = 0;
+
+	const int n_pages = THREE_PAGES;
+	map_length = PAGESIZE * THREE_PAGES;
+	uint64_t page1_va, page2_va, page3_va;
+	rte_iova_t page1_iova, page2_iova, page3_iova;
+
+	page1_va = virtaddr_64;
+	page2_va = virtaddr_64 + PAGESIZE;
+	page3_va = virtaddr_64 + (PAGESIZE * 2);
+
+	page1_iova = 0;
+	page2_iova = 0 + PAGESIZE;
+	page3_iova = 0 + (PAGESIZE * 2);
+
+	test_heap_mem_setup(map_length, n_pages);
+
+	/* Test case to map three pages */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid 1st page map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to map 1st page */
+	ret = rte_vfio_dma_map(page1_va, page1_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped valid virtual address: Failed\n");
+
+	/* Test case to unmap 2nd page */
+	ret = rte_vfio_dma_unmap(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to map 2nd page */
+	ret = rte_vfio_dma_map(page2_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and mapped "
+			"valid 2nd page virtual address: Failed\n");
+
+	/* Test case to unmap 3rd page */
+	ret = rte_vfio_dma_unmap(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to unmap device in default container "
+			"with valid map_length and mapped "
+			"valid 3rd page virtual address: Failed\n");
+
+	/* Test case to map 3rd page */
+	ret = rte_vfio_dma_map(page3_va, page3_iova, PAGESIZE);
+	TEST_ASSERT(ret == 0, "Test to map device in default container "
+			"with valid map_length and "
+			"mapped 3rd page valid virtual address: Failed\n");
+
+	/* Test case to unmap 1st page, but used IOVA address of 2nd page */
+	ret = rte_vfio_dma_unmap(page1_va, page2_iova, PAGESIZE);
+	TEST_ASSERT(ret == -1, "Test to unmap devices in default container "
+			"with valid map_length and mapped "
+			"valid virtual address: Failed\n");
+
+	/* Test case to unmap memory region from VFIO with mapped
+	 * valid iova addr, vaddr and valid map_length
+	 */
+	ret = rte_vfio_dma_unmap(page1_va, page1_iova, map_length);
+	TEST_ASSERT(ret == 0, "Test to unmap devices in default container "
+			 "with valid map_length and mapped "
+			 "valid virtual address: Failed\n");
+
+	return TEST_SUCCESS;
+}
+
+static struct
+unit_test_suite eal_vfio_testsuite  = {
+	.suite_name = "EAL VFIO Unit Test Suite",
+	.setup = initialize_container_fds,
+	.teardown = NULL,
+	.unit_test_cases = {
+		/* Test Case 1: To check vfio container create test cases */
+		TEST_CASE(test_vfio_container_create),
+
+		/* Test Case 2: To check vfio container destroy */
+		TEST_CASE(test_vfio_container_destroy),
+
+		/* Test Case 3: To  bind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_bind),
+
+		/* Test Case 4: To get IOMMU group number for a device*/
+		TEST_CASE(test_rte_vfio_get_group_num),
+
+		/* Test Case 5: To unbind a IOMMU group to a container.*/
+		TEST_CASE(test_rte_vfio_container_group_unbind),
+
+		/* Test Case 6: To perform DMA mapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free, test_rte_vfio_dma_map),
+
+		/* Test Case 7: To perform DMA unmapping for devices in default
+		 * container
+		 */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_unmap),
+
+		/* Test Case 8: To perform map devices in specific container */
+		TEST_CASE(test_rte_vfio_container_dma_map),
+
+		/* Test Case 9: To perform unmap devices in specific container
+		 */
+		TEST_CASE(test_rte_vfio_container_dma_unmap),
+
+		/* Test Case 10: To perform three pages */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_threepages),
+
+		/* Test Case 11: To check DMA overlaps */
+		TEST_CASE_ST(NULL, test_heap_mem_free,
+				test_rte_vfio_dma_map_overlaps),
+
+		TEST_CASES_END()
+	}
+};
+
+static int
+test_eal_vfio(void)
+{
+	return unit_test_suite_runner(&eal_vfio_testsuite);
+}
+
+#endif
+
+REGISTER_TEST_COMMAND(eal_vfio_autotest, test_eal_vfio);
-- 
2.17.2

Re: [dpdk-dev] [PATCH v2 3/3] app/test: add unit tests for eal vfio

[Burakov, Anatoly]

On 22-Aug-19 12:53 PM, Chaitanya Babu Talluri wrote:
> Unit test cases are added for eal vfio library.
> eal_vfio_autotest added to meson build file.
> 
> Signed-off-by: Chaitanya Babu Talluri <tallurix.chaitanya.babu@intel.com>
> ---

I think the suggestion to check if VFIO is enabled was a good one. There 
is an API to do that - i think you can do

rte_vfio_is_enabled("vfio_pci")

and if returns 0, you can return TEST_SKIPPED straight away.

>   app/test/Makefile        |   1 +
>   app/test/meson.build     |   2 +
>   app/test/test_eal_vfio.c | 728 +++++++++++++++++++++++++++++++++++++++
>   3 files changed, 731 insertions(+)
>   create mode 100644 app/test/test_eal_vfio.c
> 
> diff --git a/app/test/Makefile b/app/test/Makefile

<snip>

> +static int
> +test_eal_vfio(void)
> +{
> +	printf("VFIO not supported, skipping test\n");
> +	return TEST_SKIPPED;
> +}
> +
> +#else
> +
> +#define PAGESIZE sysconf(_SC_PAGESIZE)
> +#define INVALID_CONTAINER_FD -5
> +#define THREE_PAGES 3
> +#define UNMAPPED_ADDR 0x1500

UNMAPPED_ADDR is not 4K-aligned and will probably trigger other errors 
regarding page alignment checks. I suggest making it either 4K-aligned 
(e.g. 0x2000), or explicitly align it (e.g. RTE_ALIGN(0x1500, PAGESIZE)).


-- 
Thanks,
Anatoly

Re: [dpdk-dev] [PATCH v2 3/3] app/test: add unit tests for eal vfio

[Chaitanya Babu, TalluriX]

Hi Anatoly,

>-----Original Message-----
>From: Burakov, Anatoly
>Sent: Thursday, August 22, 2019 5:45 PM
>To: Chaitanya Babu, TalluriX <tallurix.chaitanya.babu@intel.com>;
>dev@dpdk.org
>Cc: Pattan, Reshma <reshma.pattan@intel.com>; Parthasarathy, JananeeX M
><jananeex.m.parthasarathy@intel.com>
>Subject: Re: [PATCH v2 3/3] app/test: add unit tests for eal vfio
>
>On 22-Aug-19 12:53 PM, Chaitanya Babu Talluri wrote:
>> Unit test cases are added for eal vfio library.
>> eal_vfio_autotest added to meson build file.
>>
>> Signed-off-by: Chaitanya Babu Talluri
>> <tallurix.chaitanya.babu@intel.com>
>> ---
>
>I think the suggestion to check if VFIO is enabled was a good one. There is an
>API to do that - i think you can do
>
>rte_vfio_is_enabled("vfio_pci")
>
>and if returns 0, you can return TEST_SKIPPED straight away.

rte_vfio_is_enabled() checks whether vfio module is loaded  and container fd is created.

The below error occurs even when the iommu type of vfio_cfg is not set.
Ideally iommu _type is set only during device probe.
Since device is not binded, iommu type is not set.

In order to skip the test in such scenario (vfio module is loaded but device not binded), 
we need to check the device count/groups of the default container.

Currently there is no RTE APIs to fetch the default vfio_cfg structure in order to check the active vfio_cfgs.vfio_active_groups / vfio_cfgs.vfio_iommu_type of the default container.

Please let us know whether we can enhance any API in this regard. 
>
>>   app/test/Makefile        |   1 +
>>   app/test/meson.build     |   2 +
>>   app/test/test_eal_vfio.c | 728
>+++++++++++++++++++++++++++++++++++++++
>>   3 files changed, 731 insertions(+)
>>   create mode 100644 app/test/test_eal_vfio.c
>>
>> diff --git a/app/test/Makefile b/app/test/Makefile
>
><snip>
>
>> +static int
>> +test_eal_vfio(void)
>> +{
>> +	printf("VFIO not supported, skipping test\n");
>> +	return TEST_SKIPPED;
>> +}
>> +
>> +#else
>> +
>> +#define PAGESIZE sysconf(_SC_PAGESIZE) #define
>INVALID_CONTAINER_FD
>> +-5 #define THREE_PAGES 3 #define UNMAPPED_ADDR 0x1500
>
>UNMAPPED_ADDR is not 4K-aligned and will probably trigger other errors
>regarding page alignment checks. I suggest making it either 4K-aligned (e.g.
>0x2000), or explicitly align it (e.g. RTE_ALIGN(0x1500, PAGESIZE)).

Tests were added to check page un-aligned address.
The IOCTL fails as expected.

Please let us know whether we need to remove this.
>
>
>--
>Thanks,
>Anatoly

Re: [dpdk-dev] [PATCH v2 3/3] app/test: add unit tests for eal vfio

[Burakov, Anatoly]

On 23-Aug-19 8:57 AM, Chaitanya Babu, TalluriX wrote:
> Hi Anatoly,
> 
>> -----Original Message-----
>> From: Burakov, Anatoly
>> Sent: Thursday, August 22, 2019 5:45 PM
>> To: Chaitanya Babu, TalluriX <tallurix.chaitanya.babu@intel.com>;
>> dev@dpdk.org
>> Cc: Pattan, Reshma <reshma.pattan@intel.com>; Parthasarathy, JananeeX M
>> <jananeex.m.parthasarathy@intel.com>
>> Subject: Re: [PATCH v2 3/3] app/test: add unit tests for eal vfio
>>
>> On 22-Aug-19 12:53 PM, Chaitanya Babu Talluri wrote:
>>> Unit test cases are added for eal vfio library.
>>> eal_vfio_autotest added to meson build file.
>>>
>>> Signed-off-by: Chaitanya Babu Talluri
>>> <tallurix.chaitanya.babu@intel.com>
>>> ---
>>
>> I think the suggestion to check if VFIO is enabled was a good one. There is an
>> API to do that - i think you can do
>>
>> rte_vfio_is_enabled("vfio_pci")
>>
>> and if returns 0, you can return TEST_SKIPPED straight away.
> 
> rte_vfio_is_enabled() checks whether vfio module is loaded  and container fd is created.
> 
> The below error occurs even when the iommu type of vfio_cfg is not set.
> Ideally iommu _type is set only during device probe.
> Since device is not binded, iommu type is not set.
> 
> In order to skip the test in such scenario (vfio module is loaded but device not binded),
> we need to check the device count/groups of the default container.
> 
> Currently there is no RTE APIs to fetch the default vfio_cfg structure in order to check the active vfio_cfgs.vfio_active_groups / vfio_cfgs.vfio_iommu_type of the default container.
> 
> Please let us know whether we can enhance any API in this regard.

I don't quite like the idea of adding a new API just to enable pass/skip 
of unit tests. However, i don't think we have a choice here - i can't 
see any other way than to check if IOMMU type is set. So perhaps

int rte_vfio_iommu_type_is_set()

API could be added to enable graceful skipping for this test.

>>
>>>    app/test/Makefile        |   1 +
>>>    app/test/meson.build     |   2 +
>>>    app/test/test_eal_vfio.c | 728
>> +++++++++++++++++++++++++++++++++++++++
>>>    3 files changed, 731 insertions(+)
>>>    create mode 100644 app/test/test_eal_vfio.c
>>>
>>> diff --git a/app/test/Makefile b/app/test/Makefile
>>
>> <snip>
>>
>>> +static int
>>> +test_eal_vfio(void)
>>> +{
>>> +	printf("VFIO not supported, skipping test\n");
>>> +	return TEST_SKIPPED;
>>> +}
>>> +
>>> +#else
>>> +
>>> +#define PAGESIZE sysconf(_SC_PAGESIZE) #define
>> INVALID_CONTAINER_FD
>>> +-5 #define THREE_PAGES 3 #define UNMAPPED_ADDR 0x1500
>>
>> UNMAPPED_ADDR is not 4K-aligned and will probably trigger other errors
>> regarding page alignment checks. I suggest making it either 4K-aligned (e.g.
>> 0x2000), or explicitly align it (e.g. RTE_ALIGN(0x1500, PAGESIZE)).
> 
> Tests were added to check page un-aligned address.
> The IOCTL fails as expected.
> 
> Please let us know whether we need to remove this.

Ah, so it's *intentionally* unaligned! In that case, maybe 
"UNALIGNED_ADDR" is a better name?

>>
>>
>> --
>> Thanks,
>> Anatoly


-- 
Thanks,
Anatoly