Allow using virtio without hugepages

index a5847f0bd..6893448db 100644
@@ -61,6 +61,12 @@ eal_memalloc_set_seg_fd(int list_idx __rte_unused, int seg_idx __rte_unused,
 	return -ENOTSUP;
 }
 
+int
+eal_memalloc_set_seg_list_fd(int list_idx __rte_unused, int fd __rte_unused)
+{
+	return -ENOTSUP;
+}
+
 int
 eal_memalloc_get_seg_fd_offset(int list_idx __rte_unused,
 		int seg_idx __rte_unused, size_t *offset __rte_unused)

index af917c2f9..b96c9c512 100644
@@ -84,6 +84,10 @@ eal_memalloc_get_seg_fd(int list_idx, int seg_idx);
 int
 eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd);
 
+/* returns 0 or -errno */
+int
+eal_memalloc_set_seg_list_fd(int list_idx, int fd);
+
 int
 eal_memalloc_get_seg_fd_offset(int list_idx, int seg_idx, size_t *offset);
 

index 48b9c7360..5bda92717 100644
@@ -1529,6 +1529,10 @@ eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd)
 {
 	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
 
+	/* single file segments mode doesn't support individual segment fd's */
+	if (internal_config.single_file_segments)
+		return -ENOTSUP;
+
 	/* if list is not allocated, allocate it */
 	if (fd_list[list_idx].len == 0) {
 		int len = mcfg->memsegs[list_idx].memseg_arr.len;

@@ -1541,6 +1545,18 @@ eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd)
 	return 0;
 }
 
+int
+eal_memalloc_set_seg_list_fd(int list_idx, int fd)
+{
+	/* non-single file segment mode doesn't support segment list fd's */
+	if (!internal_config.single_file_segments)
+		return -ENOTSUP;
+
+	fd_list[list_idx].memseg_list_fd = fd;
+
+	return 0;
+}
+
 int
 eal_memalloc_get_seg_fd(int list_idx, int seg_idx)
 {

index 48b23ce19..8feac2c56 100644
@@ -25,6 +25,7 @@
 #include <sys/time.h>
 #include <signal.h>
 #include <setjmp.h>
+#include <linux/memfd.h>
 #ifdef RTE_EAL_NUMA_AWARE_HUGEPAGES
 #include <numa.h>
 #include <numaif.h>

@@ -1345,12 +1346,15 @@ eal_legacy_hugepage_init(void)
 	/* hugetlbfs can be disabled */
 	if (internal_config.no_hugetlbfs) {
 		struct rte_memseg_list *msl;
+		int n_segs, cur_seg, fd, memfd, flags;
 		uint64_t page_sz;
-		int n_segs, cur_seg;
 
 		/* nohuge mode is legacy mode */
 		internal_config.legacy_mem = 1;
 
+		/* nohuge mode is single-file segments mode */
+		internal_config.single_file_segments = 1;
+
 		/* create a memseg list */
 		msl = &mcfg->memsegs[0];
 

@@ -1363,8 +1367,36 @@ eal_legacy_hugepage_init(void)
 			return -1;
 		}
 
+		/* set up parameters for anonymous mmap */
+		fd = -1;
+		flags = MAP_PRIVATE | MAP_ANONYMOUS;
+
+		/* create a memfd and store it in the segment fd table */
+		memfd = memfd_create("nohuge", 0);
+		if (memfd < 0) {
+			RTE_LOG(ERR, EAL, "Cannot create memfd: %s\n",
+					strerror(errno));
+			RTE_LOG(ERR, EAL, "Falling back to anonymous map\n");
+		} else {
+			/* we got an fd - now resize it */
+			if (ftruncate(memfd, internal_config.memory) < 0) {
+				RTE_LOG(ERR, EAL, "Cannot resize memfd: %s\n",
+						strerror(errno));
+				RTE_LOG(ERR, EAL, "Falling back to anonymous map\n");
+				close(memfd);
+			} else {
+				/* creating memfd-backed file was successful.
+				 * we want changes to memfd to be visible to
+				 * other processes (such as vhost backend), so
+				 * map it as shared memory.
+				 */
+				RTE_LOG(DEBUG, EAL, "Using memfd for anonymous memory\n");
+				fd = memfd;
+				flags = MAP_SHARED;
+			}
+		}
 		addr = mmap(NULL, internal_config.memory, PROT_READ | PROT_WRITE,
-				MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+				flags, fd, 0);
 		if (addr == MAP_FAILED) {
 			RTE_LOG(ERR, EAL, "%s: mmap() failed: %s\n", __func__,
 					strerror(errno));

@@ -1375,6 +1407,16 @@ eal_legacy_hugepage_init(void)
 		msl->socket_id = 0;
 		msl->len = internal_config.memory;
 
+		/* we're in single-file segments mode, so only the segment list
+		 * fd needs to be set up.
+		 */
+		if (fd != -1) {
+			if (eal_memalloc_set_seg_list_fd(0, fd) < 0) {
+				RTE_LOG(ERR, EAL, "Cannot set up segment list fd\n");
+				/* not a serious error, proceed */
+			}
+		}
+
 		/* populate memsegs. each memseg is one page long */
 		for (cur_seg = 0; cur_seg < n_segs; cur_seg++) {
 			arr = &msl->memseg_arr;

index d47ea4938..999ba24b4 100644
@@ -704,6 +704,12 @@ rte_memseg_get_fd_thread_unsafe(const struct rte_memseg *ms)
 		return -1;
 	}
 
+	/* segment fd API is not supported for external segments */
+	if (msl->external) {
+		rte_errno = ENOTSUP;
+		return -1;
+	}
+
 	ret = eal_memalloc_get_seg_fd(msl_idx, seg_idx);
 	if (ret < 0) {
 		rte_errno = -ret;

@@ -754,6 +760,12 @@ rte_memseg_get_fd_offset_thread_unsafe(const struct rte_memseg *ms,
 		return -1;
 	}
 
+	/* segment fd API is not supported for external segments */
+	if (msl->external) {
+		rte_errno = ENOTSUP;
+		return -1;
+	}
+
 	ret = eal_memalloc_get_seg_fd_offset(msl_idx, seg_idx, offset);
 	if (ret < 0) {
 		rte_errno = -ret;

index 784939566..a93548b8c 100644
@@ -23,6 +23,10 @@
 #include <sys/time.h>
 #include <signal.h>
 #include <setjmp.h>
+#ifdef F_ADD_SEALS /* if file sealing is supported, so is memfd */
+#include <linux/memfd.h>
+#define MEMFD_SUPPORTED
+#endif
 #ifdef RTE_EAL_NUMA_AWARE_HUGEPAGES
 #include <numa.h>
 #include <numaif.h>

@@ -53,8 +57,8 @@ const int anonymous_hugepages_supported =
 #endif
 
 /*
- * we don't actually care if memfd itself is supported - we only need to check
- * if memfd supports hugetlbfs, as that already implies memfd support.
+ * we've already checked memfd support at compile-time, but we also need to
+ * check if we can create hugepage files with memfd.
  *
  * also, this is not a constant, because while we may be *compiled* with memfd
  * hugetlbfs support, we might not be *running* on a system that supports memfd

@@ -63,10 +67,11 @@ const int anonymous_hugepages_supported =
  */
 static int memfd_create_supported =
 #ifdef MFD_HUGETLB
-#define MEMFD_SUPPORTED
 		1;
+#define RTE_MFD_HUGETLB MFD_HUGETLB
 #else
 		0;
+#define RTE_MFD_HUGETLB 4U
 #endif
 
 /*

@@ -338,12 +343,12 @@ get_seg_memfd(struct hugepage_info *hi __rte_unused,
 	int fd;
 	char segname[250]; /* as per manpage, limit is 249 bytes plus null */
 
+	int flags = RTE_MFD_HUGETLB | pagesz_flags(hi->hugepage_sz);
+
 	if (internal_config.single_file_segments) {
 		fd = fd_list[list_idx].memseg_list_fd;
 
 		if (fd < 0) {
-			int flags = MFD_HUGETLB | pagesz_flags(hi->hugepage_sz);
-
 			snprintf(segname, sizeof(segname), "seg_%i", list_idx);
 			fd = memfd_create(segname, flags);
 			if (fd < 0) {

@@ -357,8 +362,6 @@ get_seg_memfd(struct hugepage_info *hi __rte_unused,
 		fd = fd_list[list_idx].fds[seg_idx];
 
 		if (fd < 0) {
-			int flags = MFD_HUGETLB | pagesz_flags(hi->hugepage_sz);
-
 			snprintf(segname, sizeof(segname), "seg_%i-%i",
 					list_idx, seg_idx);
 			fd = memfd_create(segname, flags);

@@ -1542,6 +1545,17 @@ int
 eal_memalloc_get_seg_fd(int list_idx, int seg_idx)
 {
 	int fd;
+
+	if (internal_config.in_memory || internal_config.no_hugetlbfs) {
+#ifndef MEMFD_SUPPORTED
+		/* in in-memory or no-huge mode, we rely on memfd support */
+		return -ENOTSUP;
+#endif
+		/* memfd supported, but hugetlbfs memfd may not be */
+		if (!internal_config.no_hugetlbfs && !memfd_create_supported)
+			return -ENOTSUP;
+	}
+
 	if (internal_config.single_file_segments) {
 		fd = fd_list[list_idx].memseg_list_fd;
 	} else if (fd_list[list_idx].len == 0) {

@@ -1565,7 +1579,7 @@ test_memfd_create(void)
 		int pagesz_flag = pagesz_flags(pagesz);
 		int flags;
 
-		flags = pagesz_flag | MFD_HUGETLB;
+		flags = pagesz_flag | RTE_MFD_HUGETLB;
 		int fd = memfd_create("test", flags);
 		if (fd < 0) {
 			/* we failed - let memalloc know this isn't working */

@@ -1589,6 +1603,16 @@ eal_memalloc_get_seg_fd_offset(int list_idx, int seg_idx, size_t *offset)
 {
 	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
 
+	if (internal_config.in_memory || internal_config.no_hugetlbfs) {
+#ifndef MEMFD_SUPPORTED
+		/* in in-memory or no-huge mode, we rely on memfd support */
+		return -ENOTSUP;
+#endif
+		/* memfd supported, but hugetlbfs memfd may not be */
+		if (!internal_config.no_hugetlbfs && !memfd_create_supported)
+			return -ENOTSUP;
+	}
+
 	/* fd_list not initialized? */
 	if (fd_list[list_idx].len == 0)
 		return -ENODEV;

index a5847f0bd..6893448db 100644
@@ -61,6 +61,12 @@ eal_memalloc_set_seg_fd(int list_idx __rte_unused, int seg_idx __rte_unused,
 	return -ENOTSUP;
 }
 
+int
+eal_memalloc_set_seg_list_fd(int list_idx __rte_unused, int fd __rte_unused)
+{
+	return -ENOTSUP;
+}
+
 int
 eal_memalloc_get_seg_fd_offset(int list_idx __rte_unused,
 		int seg_idx __rte_unused, size_t *offset __rte_unused)

index af917c2f9..b96c9c512 100644
@@ -84,6 +84,10 @@ eal_memalloc_get_seg_fd(int list_idx, int seg_idx);
 int
 eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd);
 
+/* returns 0 or -errno */
+int
+eal_memalloc_set_seg_list_fd(int list_idx, int fd);
+
 int
 eal_memalloc_get_seg_fd_offset(int list_idx, int seg_idx, size_t *offset);
 

index a93548b8c..eef140b33 100644
@@ -1529,6 +1529,10 @@ eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd)
 {
 	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
 
+	/* single file segments mode doesn't support individual segment fd's */
+	if (internal_config.single_file_segments)
+		return -ENOTSUP;
+
 	/* if list is not allocated, allocate it */
 	if (fd_list[list_idx].len == 0) {
 		int len = mcfg->memsegs[list_idx].memseg_arr.len;

@@ -1541,6 +1545,28 @@ eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd)
 	return 0;
 }
 
+int
+eal_memalloc_set_seg_list_fd(int list_idx, int fd)
+{
+	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
+
+	/* non-single file segment mode doesn't support segment list fd's */
+	if (!internal_config.single_file_segments)
+		return -ENOTSUP;
+
+	/* if list is not allocated, allocate it */
+	if (fd_list[list_idx].len == 0) {
+		int len = mcfg->memsegs[list_idx].memseg_arr.len;
+
+		if (alloc_list(list_idx, len) < 0)
+			return -ENOMEM;
+	}
+
+	fd_list[list_idx].memseg_list_fd = fd;
+
+	return 0;
+}
+
 int
 eal_memalloc_get_seg_fd(int list_idx, int seg_idx)
 {

84 lines not on this page · raw message

index b96bca771..3da803e4e 100644
@@ -37,10 +37,44 @@ check_mem(const struct rte_memseg_list *msl __rte_unused,
 	return 0;
 }
 
+static int
+check_seg_fds(const struct rte_memseg_list *msl, const struct rte_memseg *ms,
+		void *arg __rte_unused)
+{
+	size_t offset;
+	int ret;
+
+	/* skip external segments */
+	if (msl->external)
+		return 0;
+
+	/* try segment fd first. we're in a callback, so thread-unsafe */
+	ret = rte_memseg_get_fd_thread_unsafe(ms);
+	if (ret < 0) {
+		/* ENOTSUP means segment is valid, but there is not support for
+		 * segment fd API (e.g. on FreeBSD).
+		 */
+		if (errno == ENOTSUP)
+			return 1;
+		/* all other errors are treated as failures */
+		return -1;
+	}
+
+	/* we're able to get memseg fd - try getting its offset */
+	ret = rte_memseg_get_fd_offset_thread_unsafe(ms, &offset);
+	if (ret < 0) {
+		if (errno == ENOTSUP)
+			return 1;
+		return -1;
+	}
+	return 0;
+}
+
 static int
 test_memory(void)
 {
 	uint64_t s;
+	int ret;
 
 	/*
 	 * dump the mapped memory: the python-expect script checks

@@ -59,6 +93,15 @@ test_memory(void)
 	/* try to read memory (should not segfault) */
 	rte_memseg_walk(check_mem, NULL);
 
+	/* check segment fd support */
+	ret = rte_memseg_walk(check_seg_fds, NULL);
+	if (ret == 1) {
+		printf("Segment fd API is unsupported\n");
+	} else if (ret == -1) {
+		printf("Error getting segment fd's\n");
+		return -1;
+	}
+
 	return 0;
 }
 

index d47ea4938..999ba24b4 100644
@@ -704,6 +704,12 @@ rte_memseg_get_fd_thread_unsafe(const struct rte_memseg *ms)
 		return -1;
 	}
 
+	/* segment fd API is not supported for external segments */
+	if (msl->external) {
+		rte_errno = ENOTSUP;
+		return -1;
+	}
+
 	ret = eal_memalloc_get_seg_fd(msl_idx, seg_idx);
 	if (ret < 0) {
 		rte_errno = -ret;

@@ -754,6 +760,12 @@ rte_memseg_get_fd_offset_thread_unsafe(const struct rte_memseg *ms,
 		return -1;
 	}
 
+	/* segment fd API is not supported for external segments */
+	if (msl->external) {
+		rte_errno = ENOTSUP;
+		return -1;
+	}
+
 	ret = eal_memalloc_get_seg_fd_offset(msl_idx, seg_idx, offset);
 	if (ret < 0) {
 		rte_errno = -ret;

index 784939566..a93548b8c 100644
@@ -23,6 +23,10 @@
 #include <sys/time.h>
 #include <signal.h>
 #include <setjmp.h>
+#ifdef F_ADD_SEALS /* if file sealing is supported, so is memfd */
+#include <linux/memfd.h>
+#define MEMFD_SUPPORTED
+#endif
 #ifdef RTE_EAL_NUMA_AWARE_HUGEPAGES
 #include <numa.h>
 #include <numaif.h>

@@ -53,8 +57,8 @@ const int anonymous_hugepages_supported =
 #endif
 
 /*
- * we don't actually care if memfd itself is supported - we only need to check
- * if memfd supports hugetlbfs, as that already implies memfd support.
+ * we've already checked memfd support at compile-time, but we also need to
+ * check if we can create hugepage files with memfd.
  *
  * also, this is not a constant, because while we may be *compiled* with memfd
  * hugetlbfs support, we might not be *running* on a system that supports memfd

@@ -63,10 +67,11 @@ const int anonymous_hugepages_supported =
  */
 static int memfd_create_supported =
 #ifdef MFD_HUGETLB
-#define MEMFD_SUPPORTED
 		1;
+#define RTE_MFD_HUGETLB MFD_HUGETLB
 #else
 		0;
+#define RTE_MFD_HUGETLB 4U
 #endif
 
 /*

@@ -338,12 +343,12 @@ get_seg_memfd(struct hugepage_info *hi __rte_unused,
 	int fd;
 	char segname[250]; /* as per manpage, limit is 249 bytes plus null */
 
+	int flags = RTE_MFD_HUGETLB | pagesz_flags(hi->hugepage_sz);
+
 	if (internal_config.single_file_segments) {
 		fd = fd_list[list_idx].memseg_list_fd;
 
 		if (fd < 0) {
-			int flags = MFD_HUGETLB | pagesz_flags(hi->hugepage_sz);
-
 			snprintf(segname, sizeof(segname), "seg_%i", list_idx);
 			fd = memfd_create(segname, flags);
 			if (fd < 0) {

@@ -357,8 +362,6 @@ get_seg_memfd(struct hugepage_info *hi __rte_unused,
 		fd = fd_list[list_idx].fds[seg_idx];
 
 		if (fd < 0) {
-			int flags = MFD_HUGETLB | pagesz_flags(hi->hugepage_sz);
-
 			snprintf(segname, sizeof(segname), "seg_%i-%i",
 					list_idx, seg_idx);
 			fd = memfd_create(segname, flags);

@@ -1542,6 +1545,17 @@ int
 eal_memalloc_get_seg_fd(int list_idx, int seg_idx)
 {
 	int fd;
+
+	if (internal_config.in_memory || internal_config.no_hugetlbfs) {
+#ifndef MEMFD_SUPPORTED
+		/* in in-memory or no-huge mode, we rely on memfd support */
+		return -ENOTSUP;
+#endif
+		/* memfd supported, but hugetlbfs memfd may not be */
+		if (!internal_config.no_hugetlbfs && !memfd_create_supported)
+			return -ENOTSUP;
+	}
+
 	if (internal_config.single_file_segments) {
 		fd = fd_list[list_idx].memseg_list_fd;
 	} else if (fd_list[list_idx].len == 0) {

@@ -1565,7 +1579,7 @@ test_memfd_create(void)
 		int pagesz_flag = pagesz_flags(pagesz);
 		int flags;
 
-		flags = pagesz_flag | MFD_HUGETLB;
+		flags = pagesz_flag | RTE_MFD_HUGETLB;
 		int fd = memfd_create("test", flags);
 		if (fd < 0) {
 			/* we failed - let memalloc know this isn't working */

@@ -1589,6 +1603,16 @@ eal_memalloc_get_seg_fd_offset(int list_idx, int seg_idx, size_t *offset)
 {
 	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
 
+	if (internal_config.in_memory || internal_config.no_hugetlbfs) {
+#ifndef MEMFD_SUPPORTED
+		/* in in-memory or no-huge mode, we rely on memfd support */
+		return -ENOTSUP;
+#endif
+		/* memfd supported, but hugetlbfs memfd may not be */
+		if (!internal_config.no_hugetlbfs && !memfd_create_supported)
+			return -ENOTSUP;
+	}
+
 	/* fd_list not initialized? */
 	if (fd_list[list_idx].len == 0)
 		return -ENODEV;

index a5847f0bd..6893448db 100644
@@ -61,6 +61,12 @@ eal_memalloc_set_seg_fd(int list_idx __rte_unused, int seg_idx __rte_unused,
 	return -ENOTSUP;
 }
 
+int
+eal_memalloc_set_seg_list_fd(int list_idx __rte_unused, int fd __rte_unused)
+{
+	return -ENOTSUP;
+}
+
 int
 eal_memalloc_get_seg_fd_offset(int list_idx __rte_unused,
 		int seg_idx __rte_unused, size_t *offset __rte_unused)

index af917c2f9..b96c9c512 100644
@@ -84,6 +84,10 @@ eal_memalloc_get_seg_fd(int list_idx, int seg_idx);
 int
 eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd);
 
+/* returns 0 or -errno */
+int
+eal_memalloc_set_seg_list_fd(int list_idx, int fd);
+
 int
 eal_memalloc_get_seg_fd_offset(int list_idx, int seg_idx, size_t *offset);
 

index a93548b8c..eef140b33 100644
@@ -1529,6 +1529,10 @@ eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd)
 {
 	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
 
+	/* single file segments mode doesn't support individual segment fd's */
+	if (internal_config.single_file_segments)
+		return -ENOTSUP;
+
 	/* if list is not allocated, allocate it */
 	if (fd_list[list_idx].len == 0) {
 		int len = mcfg->memsegs[list_idx].memseg_arr.len;

@@ -1541,6 +1545,28 @@ eal_memalloc_set_seg_fd(int list_idx, int seg_idx, int fd)
 	return 0;
 }
 
+int
+eal_memalloc_set_seg_list_fd(int list_idx, int fd)
+{
+	struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config;
+
+	/* non-single file segment mode doesn't support segment list fd's */
+	if (!internal_config.single_file_segments)
+		return -ENOTSUP;
+
+	/* if list is not allocated, allocate it */
+	if (fd_list[list_idx].len == 0) {
+		int len = mcfg->memsegs[list_idx].memseg_arr.len;
+
+		if (alloc_list(list_idx, len) < 0)
+			return -ENOMEM;
+	}
+
+	fd_list[list_idx].memseg_list_fd = fd;
+
+	return 0;
+}
+
 int
 eal_memalloc_get_seg_fd(int list_idx, int seg_idx)
 {

84 lines not on this page · raw message

index b96bca771..3da803e4e 100644
@@ -37,10 +37,44 @@ check_mem(const struct rte_memseg_list *msl __rte_unused,
 	return 0;
 }
 
+static int
+check_seg_fds(const struct rte_memseg_list *msl, const struct rte_memseg *ms,
+		void *arg __rte_unused)
+{
+	size_t offset;
+	int ret;
+
+	/* skip external segments */
+	if (msl->external)
+		return 0;
+
+	/* try segment fd first. we're in a callback, so thread-unsafe */
+	ret = rte_memseg_get_fd_thread_unsafe(ms);
+	if (ret < 0) {
+		/* ENOTSUP means segment is valid, but there is not support for
+		 * segment fd API (e.g. on FreeBSD).
+		 */
+		if (errno == ENOTSUP)
+			return 1;
+		/* all other errors are treated as failures */
+		return -1;
+	}
+
+	/* we're able to get memseg fd - try getting its offset */
+	ret = rte_memseg_get_fd_offset_thread_unsafe(ms, &offset);
+	if (ret < 0) {
+		if (errno == ENOTSUP)
+			return 1;
+		return -1;
+	}
+	return 0;
+}
+
 static int
 test_memory(void)
 {
 	uint64_t s;
+	int ret;
 
 	/*
 	 * dump the mapped memory: the python-expect script checks

@@ -59,6 +93,15 @@ test_memory(void)
 	/* try to read memory (should not segfault) */
 	rte_memseg_walk(check_mem, NULL);
 
+	/* check segment fd support */
+	ret = rte_memseg_walk(check_seg_fds, NULL);
+	if (ret == 1) {
+		printf("Segment fd API is unsupported\n");
+	} else if (ret == -1) {
+		printf("Error getting segment fd's\n");
+		return -1;
+	}
+
 	return 0;
 }