diff options
Diffstat (limited to 'lib/librte_eal/linuxapp/eal')
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/Makefile | 12 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/eal.c | 38 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/eal_ivshmem.c | 954 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/eal_log.c | 40 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/eal_memory.c | 327 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/eal_pci.c | 23 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/include/exec-env/rte_interrupts.h | 1 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/include/exec-env/rte_kni_common.h | 10 | ||||
| -rw-r--r-- | lib/librte_eal/linuxapp/eal/rte_eal_version.map | 12 |
9 files changed, 80 insertions, 1337 deletions
diff --git a/lib/librte_eal/linuxapp/eal/Makefile b/lib/librte_eal/linuxapp/eal/Makefile index 182729c0..4e206f09 100644 --- a/lib/librte_eal/linuxapp/eal/Makefile +++ b/lib/librte_eal/linuxapp/eal/Makefile @@ -37,19 +37,13 @@ ARCH_DIR ?= $(RTE_ARCH) EXPORT_MAP := rte_eal_version.map VPATH += $(RTE_SDK)/lib/librte_eal/common/arch/$(ARCH_DIR) -LIBABIVER := 2 +LIBABIVER := 3 VPATH += $(RTE_SDK)/lib/librte_eal/common CFLAGS += -I$(SRCDIR)/include CFLAGS += -I$(RTE_SDK)/lib/librte_eal/common CFLAGS += -I$(RTE_SDK)/lib/librte_eal/common/include -ifeq ($(CONFIG_RTE_LIBRTE_IVSHMEM),y) -# workaround for circular dependency eal -> ivshmem -> ring/mempool -> eal -CFLAGS += -I$(RTE_SDK)/lib/librte_ring -CFLAGS += -I$(RTE_SDK)/lib/librte_mempool -CFLAGS += -I$(RTE_SDK)/lib/librte_ivshmem -endif CFLAGS += $(WERROR_FLAGS) -O3 LDLIBS += -ldl @@ -76,9 +70,6 @@ SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_lcore.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_timer.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_interrupts.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_alarm.c -ifeq ($(CONFIG_RTE_LIBRTE_IVSHMEM),y) -SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_ivshmem.c -endif # from common dir SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_lcore.c @@ -86,6 +77,7 @@ SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_timer.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_memzone.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_log.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_launch.c +SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_vdev.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_pci.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_pci_uio.c SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_memory.c diff --git a/lib/librte_eal/linuxapp/eal/eal.c b/lib/librte_eal/linuxapp/eal/eal.c index 3fb2188f..2075282e 100644 --- a/lib/librte_eal/linuxapp/eal/eal.c +++ b/lib/librte_eal/linuxapp/eal/eal.c @@ -70,6 +70,7 @@ #include <rte_cpuflags.h> #include <rte_interrupts.h> #include <rte_pci.h> +#include <rte_dev.h> #include <rte_devargs.h> #include <rte_common.h> #include <rte_version.h> @@ -238,7 +239,8 @@ rte_eal_config_attach(void) mem_config = (struct rte_mem_config *) mmap(NULL, sizeof(*mem_config), PROT_READ, MAP_SHARED, mem_cfg_fd, 0); if (mem_config == MAP_FAILED) - rte_panic("Cannot mmap memory for rte_config\n"); + rte_panic("Cannot mmap memory for rte_config! error %i (%s)\n", + errno, strerror(errno)); rte_config.mem_config = mem_config; } @@ -263,9 +265,17 @@ rte_eal_config_reattach(void) mem_config = (struct rte_mem_config *) mmap(rte_mem_cfg_addr, sizeof(*mem_config), PROT_READ | PROT_WRITE, MAP_SHARED, mem_cfg_fd, 0); + if (mem_config == MAP_FAILED || mem_config != rte_mem_cfg_addr) { + if (mem_config != MAP_FAILED) + /* errno is stale, don't use */ + rte_panic("Cannot mmap memory for rte_config at [%p], got [%p]" + " - please use '--base-virtaddr' option\n", + rte_mem_cfg_addr, mem_config); + else + rte_panic("Cannot mmap memory for rte_config! error %i (%s)\n", + errno, strerror(errno)); + } close(mem_cfg_fd); - if (mem_config == MAP_FAILED || mem_config != rte_mem_cfg_addr) - rte_panic("Cannot mmap memory for rte_config\n"); rte_config.mem_config = mem_config; } @@ -740,6 +750,9 @@ rte_eal_init(int argc, char **argv) char cpuset[RTE_CPU_AFFINITY_STR_LEN]; char thread_name[RTE_MAX_THREAD_NAME_LEN]; + /* checks if the machine is adequate */ + rte_cpu_check_supported(); + if (!rte_atomic32_test_and_set(&run_once)) return -1; @@ -748,9 +761,6 @@ rte_eal_init(int argc, char **argv) thread_id = pthread_self(); - if (rte_eal_log_early_init() < 0) - rte_panic("Cannot init early logs\n"); - eal_log_level_parse(argc, argv); /* set log level as early as possible */ @@ -789,6 +799,9 @@ rte_eal_init(int argc, char **argv) rte_config_init(); + if (rte_eal_log_init(logid, internal_config.syslog_facility) < 0) + rte_panic("Cannot init logs\n"); + if (rte_eal_pci_init() < 0) rte_panic("Cannot init PCI\n"); @@ -797,11 +810,6 @@ rte_eal_init(int argc, char **argv) rte_panic("Cannot init VFIO\n"); #endif -#ifdef RTE_LIBRTE_IVSHMEM - if (rte_eal_ivshmem_init() < 0) - rte_panic("Cannot init IVSHMEM\n"); -#endif - if (rte_eal_memory_init() < 0) rte_panic("Cannot init memory\n"); @@ -814,14 +822,6 @@ rte_eal_init(int argc, char **argv) if (rte_eal_tailqs_init() < 0) rte_panic("Cannot init tail queues for objects\n"); -#ifdef RTE_LIBRTE_IVSHMEM - if (rte_eal_ivshmem_obj_init() < 0) - rte_panic("Cannot init IVSHMEM objects\n"); -#endif - - if (rte_eal_log_init(logid, internal_config.syslog_facility) < 0) - rte_panic("Cannot init logs\n"); - if (rte_eal_alarm_init() < 0) rte_panic("Cannot init interrupt-handling thread\n"); diff --git a/lib/librte_eal/linuxapp/eal/eal_ivshmem.c b/lib/librte_eal/linuxapp/eal/eal_ivshmem.c deleted file mode 100644 index 67b3caf2..00000000 --- a/lib/librte_eal/linuxapp/eal/eal_ivshmem.c +++ /dev/null @@ -1,954 +0,0 @@ -/*- - * BSD LICENSE - * - * Copyright(c) 2010-2014 Intel Corporation. All rights reserved. - * All rights reserved. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * - * * Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * * Redistributions in binary form must reproduce the above copyright - * notice, this list of conditions and the following disclaimer in - * the documentation and/or other materials provided with the - * distribution. - * * Neither the name of Intel Corporation nor the names of its - * contributors may be used to endorse or promote products derived - * from this software without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS - * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT - * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR - * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT - * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, - * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT - * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, - * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY - * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT - * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE - * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. - */ - -#ifdef RTE_LIBRTE_IVSHMEM /* hide it from coverage */ - -#include <stdint.h> -#include <unistd.h> -#include <inttypes.h> -#include <sys/mman.h> -#include <sys/file.h> -#include <string.h> -#include <sys/queue.h> - -#include <rte_log.h> -#include <rte_pci.h> -#include <rte_memory.h> -#include <rte_eal.h> -#include <rte_eal_memconfig.h> -#include <rte_string_fns.h> -#include <rte_errno.h> -#include <rte_ring.h> -#include <rte_malloc.h> -#include <rte_common.h> -#include <rte_ivshmem.h> - -#include "eal_internal_cfg.h" -#include "eal_private.h" - -#define PCI_VENDOR_ID_IVSHMEM 0x1Af4 -#define PCI_DEVICE_ID_IVSHMEM 0x1110 - -#define IVSHMEM_MAGIC 0x0BADC0DE - -#define IVSHMEM_RESOURCE_PATH "/sys/bus/pci/devices/%04x:%02x:%02x.%x/resource2" -#define IVSHMEM_CONFIG_PATH "/var/run/.%s_ivshmem_config" - -#define PHYS 0x1 -#define VIRT 0x2 -#define IOREMAP 0x4 -#define FULL (PHYS|VIRT|IOREMAP) - -#define METADATA_SIZE_ALIGNED \ - (RTE_ALIGN_CEIL(sizeof(struct rte_ivshmem_metadata),pagesz)) - -#define CONTAINS(x,y)\ - (((y).addr_64 >= (x).addr_64) && ((y).addr_64 < (x).addr_64 + (x).len)) - -#define DIM(x) (sizeof(x)/sizeof(x[0])) - -struct ivshmem_pci_device { - char path[PATH_MAX]; - phys_addr_t ioremap_addr; -}; - -/* data type to store in config */ -struct ivshmem_segment { - struct rte_ivshmem_metadata_entry entry; - uint64_t align; - char path[PATH_MAX]; -}; -struct ivshmem_shared_config { - struct ivshmem_segment segment[RTE_MAX_MEMSEG]; - uint32_t segment_idx; - struct ivshmem_pci_device pci_devs[RTE_LIBRTE_IVSHMEM_MAX_PCI_DEVS]; - uint32_t pci_devs_idx; -}; -static struct ivshmem_shared_config * ivshmem_config; -static int memseg_idx; -static int pagesz; - -/* Tailq heads to add rings to */ -TAILQ_HEAD(rte_ring_list, rte_tailq_entry); - -/* - * Utility functions - */ - -static int -is_ivshmem_device(struct rte_pci_device * dev) -{ - return dev->id.vendor_id == PCI_VENDOR_ID_IVSHMEM - && dev->id.device_id == PCI_DEVICE_ID_IVSHMEM; -} - -static void * -map_metadata(int fd, uint64_t len) -{ - size_t metadata_len = sizeof(struct rte_ivshmem_metadata); - size_t aligned_len = METADATA_SIZE_ALIGNED; - - return mmap(NULL, metadata_len, PROT_READ | PROT_WRITE, - MAP_SHARED, fd, len - aligned_len); -} - -static void -unmap_metadata(void * ptr) -{ - munmap(ptr, sizeof(struct rte_ivshmem_metadata)); -} - -static int -has_ivshmem_metadata(int fd, uint64_t len) -{ - struct rte_ivshmem_metadata metadata; - void * ptr; - - ptr = map_metadata(fd, len); - - if (ptr == MAP_FAILED) - return -1; - - metadata = *(struct rte_ivshmem_metadata*) (ptr); - - unmap_metadata(ptr); - - return metadata.magic_number == IVSHMEM_MAGIC; -} - -static void -remove_segment(struct ivshmem_segment * ms, int len, int idx) -{ - int i; - - for (i = idx; i < len - 1; i++) - memcpy(&ms[i], &ms[i+1], sizeof(struct ivshmem_segment)); - memset(&ms[len-1], 0, sizeof(struct ivshmem_segment)); -} - -static int -overlap(const struct rte_memzone * mz1, const struct rte_memzone * mz2) -{ - uint64_t start1, end1, start2, end2; - uint64_t p_start1, p_end1, p_start2, p_end2; - uint64_t i_start1, i_end1, i_start2, i_end2; - int result = 0; - - /* gather virtual addresses */ - start1 = mz1->addr_64; - end1 = mz1->addr_64 + mz1->len; - start2 = mz2->addr_64; - end2 = mz2->addr_64 + mz2->len; - - /* gather physical addresses */ - p_start1 = mz1->phys_addr; - p_end1 = mz1->phys_addr + mz1->len; - p_start2 = mz2->phys_addr; - p_end2 = mz2->phys_addr + mz2->len; - - /* gather ioremap addresses */ - i_start1 = mz1->ioremap_addr; - i_end1 = mz1->ioremap_addr + mz1->len; - i_start2 = mz2->ioremap_addr; - i_end2 = mz2->ioremap_addr + mz2->len; - - /* check for overlap in virtual addresses */ - if (start1 >= start2 && start1 < end2) - result |= VIRT; - if (start2 >= start1 && start2 < end1) - result |= VIRT; - - /* check for overlap in physical addresses */ - if (p_start1 >= p_start2 && p_start1 < p_end2) - result |= PHYS; - if (p_start2 >= p_start1 && p_start2 < p_end1) - result |= PHYS; - - /* check for overlap in ioremap addresses */ - if (i_start1 >= i_start2 && i_start1 < i_end2) - result |= IOREMAP; - if (i_start2 >= i_start1 && i_start2 < i_end1) - result |= IOREMAP; - - return result; -} - -static int -adjacent(const struct rte_memzone * mz1, const struct rte_memzone * mz2) -{ - uint64_t start1, end1, start2, end2; - uint64_t p_start1, p_end1, p_start2, p_end2; - uint64_t i_start1, i_end1, i_start2, i_end2; - int result = 0; - - /* gather virtual addresses */ - start1 = mz1->addr_64; - end1 = mz1->addr_64 + mz1->len; - start2 = mz2->addr_64; - end2 = mz2->addr_64 + mz2->len; - - /* gather physical addresses */ - p_start1 = mz1->phys_addr; - p_end1 = mz1->phys_addr + mz1->len; - p_start2 = mz2->phys_addr; - p_end2 = mz2->phys_addr + mz2->len; - - /* gather ioremap addresses */ - i_start1 = mz1->ioremap_addr; - i_end1 = mz1->ioremap_addr + mz1->len; - i_start2 = mz2->ioremap_addr; - i_end2 = mz2->ioremap_addr + mz2->len; - - /* check if segments are virtually adjacent */ - if (start1 == end2) - result |= VIRT; - if (start2 == end1) - result |= VIRT; - - /* check if segments are physically adjacent */ - if (p_start1 == p_end2) - result |= PHYS; - if (p_start2 == p_end1) - result |= PHYS; - - /* check if segments are ioremap-adjacent */ - if (i_start1 == i_end2) - result |= IOREMAP; - if (i_start2 == i_end1) - result |= IOREMAP; - - return result; -} - -static int -has_adjacent_segments(struct ivshmem_segment * ms, int len) -{ - int i, j; - - for (i = 0; i < len; i++) - for (j = i + 1; j < len; j++) { - /* we're only interested in fully adjacent segments; partially - * adjacent segments can coexist. - */ - if (adjacent(&ms[i].entry.mz, &ms[j].entry.mz) == FULL) - return 1; - } - return 0; -} - -static int -has_overlapping_segments(struct ivshmem_segment * ms, int len) -{ - int i, j; - - for (i = 0; i < len; i++) - for (j = i + 1; j < len; j++) - if (overlap(&ms[i].entry.mz, &ms[j].entry.mz)) - return 1; - return 0; -} - -static int -seg_compare(const void * a, const void * b) -{ - const struct ivshmem_segment * s1 = (const struct ivshmem_segment*) a; - const struct ivshmem_segment * s2 = (const struct ivshmem_segment*) b; - - /* move unallocated zones to the end */ - if (s1->entry.mz.addr == NULL && s2->entry.mz.addr == NULL) - return 0; - if (s1->entry.mz.addr == 0) - return 1; - if (s2->entry.mz.addr == 0) - return -1; - - return s1->entry.mz.phys_addr > s2->entry.mz.phys_addr; -} - -#ifdef RTE_LIBRTE_IVSHMEM_DEBUG -static void -entry_dump(struct rte_ivshmem_metadata_entry *e) -{ - RTE_LOG(DEBUG, EAL, "\tvirt: %p-%p\n", e->mz.addr, - RTE_PTR_ADD(e->mz.addr, e->mz.len)); - RTE_LOG(DEBUG, EAL, "\tphys: 0x%" PRIx64 "-0x%" PRIx64 "\n", - e->mz.phys_addr, - e->mz.phys_addr + e->mz.len); - RTE_LOG(DEBUG, EAL, "\tio: 0x%" PRIx64 "-0x%" PRIx64 "\n", - e->mz.ioremap_addr, - e->mz.ioremap_addr + e->mz.len); - RTE_LOG(DEBUG, EAL, "\tlen: 0x%" PRIx64 "\n", e->mz.len); - RTE_LOG(DEBUG, EAL, "\toff: 0x%" PRIx64 "\n", e->offset); -} -#endif - - - -/* - * Actual useful code - */ - -/* read through metadata mapped from the IVSHMEM device */ -static int -read_metadata(char * path, int path_len, int fd, uint64_t flen) -{ - struct rte_ivshmem_metadata metadata; - struct rte_ivshmem_metadata_entry * entry; - int idx, i; - void * ptr; - - ptr = map_metadata(fd, flen); - - if (ptr == MAP_FAILED) - return -1; - - metadata = *(struct rte_ivshmem_metadata*) (ptr); - - unmap_metadata(ptr); - - RTE_LOG(DEBUG, EAL, "Parsing metadata for \"%s\"\n", metadata.name); - - idx = ivshmem_config->segment_idx; - - for (i = 0; i < RTE_LIBRTE_IVSHMEM_MAX_ENTRIES && - idx <= RTE_MAX_MEMSEG; i++) { - - if (idx == RTE_MAX_MEMSEG) { - RTE_LOG(ERR, EAL, "Not enough memory segments!\n"); - return -1; - } - - entry = &metadata.entry[i]; - - /* stop on uninitialized memzone */ - if (entry->mz.len == 0) - break; - - /* copy metadata entry */ - memcpy(&ivshmem_config->segment[idx].entry, entry, - sizeof(struct rte_ivshmem_metadata_entry)); - - /* copy path */ - snprintf(ivshmem_config->segment[idx].path, path_len, "%s", path); - - idx++; - } - ivshmem_config->segment_idx = idx; - - return 0; -} - -/* check through each segment and look for adjacent or overlapping ones. */ -static int -cleanup_segments(struct ivshmem_segment * ms, int tbl_len) -{ - struct ivshmem_segment * s, * tmp; - int i, j, concat, seg_adjacent, seg_overlapping; - uint64_t start1, start2, end1, end2, p_start1, p_start2, i_start1, i_start2; - - qsort(ms, tbl_len, sizeof(struct ivshmem_segment), - seg_compare); - - while (has_overlapping_segments(ms, tbl_len) || - has_adjacent_segments(ms, tbl_len)) { - - for (i = 0; i < tbl_len; i++) { - s = &ms[i]; - - concat = 0; - - for (j = i + 1; j < tbl_len; j++) { - tmp = &ms[j]; - - /* check if this segment is overlapping with existing segment, - * or is adjacent to existing segment */ - seg_overlapping = overlap(&s->entry.mz, &tmp->entry.mz); - seg_adjacent = adjacent(&s->entry.mz, &tmp->entry.mz); - - /* check if segments fully overlap or are fully adjacent */ - if ((seg_adjacent == FULL) || (seg_overlapping == FULL)) { - -#ifdef RTE_LIBRTE_IVSHMEM_DEBUG - RTE_LOG(DEBUG, EAL, "Concatenating segments\n"); - RTE_LOG(DEBUG, EAL, "Segment %i:\n", i); - entry_dump(&s->entry); - RTE_LOG(DEBUG, EAL, "Segment %i:\n", j); - entry_dump(&tmp->entry); -#endif - - start1 = s->entry.mz.addr_64; - start2 = tmp->entry.mz.addr_64; - p_start1 = s->entry.mz.phys_addr; - p_start2 = tmp->entry.mz.phys_addr; - i_start1 = s->entry.mz.ioremap_addr; - i_start2 = tmp->entry.mz.ioremap_addr; - end1 = s->entry.mz.addr_64 + s->entry.mz.len; - end2 = tmp->entry.mz.addr_64 + tmp->entry.mz.len; - - /* settle for minimum start address and maximum length */ - s->entry.mz.addr_64 = RTE_MIN(start1, start2); - s->entry.mz.phys_addr = RTE_MIN(p_start1, p_start2); - s->entry.mz.ioremap_addr = RTE_MIN(i_start1, i_start2); - s->entry.offset = RTE_MIN(s->entry.offset, tmp->entry.offset); - s->entry.mz.len = RTE_MAX(end1, end2) - s->entry.mz.addr_64; - concat = 1; - -#ifdef RTE_LIBRTE_IVSHMEM_DEBUG - RTE_LOG(DEBUG, EAL, "Resulting segment:\n"); - entry_dump(&s->entry); - -#endif - } - /* if segments not fully overlap, we have an error condition. - * adjacent segments can coexist. - */ - else if (seg_overlapping > 0) { - RTE_LOG(ERR, EAL, "Segments %i and %i overlap!\n", i, j); -#ifdef RTE_LIBRTE_IVSHMEM_DEBUG - RTE_LOG(DEBUG, EAL, "Segment %i:\n", i); - entry_dump(&s->entry); - RTE_LOG(DEBUG, EAL, "Segment %i:\n", j); - entry_dump(&tmp->entry); -#endif - return -1; - } - if (concat) - break; - } - /* if we concatenated, remove segment at j */ - if (concat) { - remove_segment(ms, tbl_len, j); - tbl_len--; - break; - } - } - } - - return tbl_len; -} - -static int -create_shared_config(void) -{ - char path[PATH_MAX]; - int fd; - - /* build ivshmem config file path */ - snprintf(path, sizeof(path), IVSHMEM_CONFIG_PATH, - internal_config.hugefile_prefix); - - fd = open(path, O_CREAT | O_RDWR, 0600); - - if (fd < 0) { - RTE_LOG(ERR, EAL, "Could not open %s: %s\n", path, strerror(errno)); - return -1; - } - - /* try ex-locking first - if the file is locked, we have a problem */ - if (flock(fd, LOCK_EX | LOCK_NB) == -1) { - RTE_LOG(ERR, EAL, "Locking %s failed: %s\n", path, strerror(errno)); - close(fd); - return -1; - } - - if (ftruncate(fd, sizeof(struct ivshmem_shared_config)) < 0) { - RTE_LOG(ERR, EAL, "ftruncate failed: %s\n", strerror(errno)); - return -1; - } - - ivshmem_config = mmap(NULL, sizeof(struct ivshmem_shared_config), - PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); - - if (ivshmem_config == MAP_FAILED) - return -1; - - memset(ivshmem_config, 0, sizeof(struct ivshmem_shared_config)); - - /* change the exclusive lock we got earlier to a shared lock */ - if (flock(fd, LOCK_SH | LOCK_NB) == -1) { - RTE_LOG(ERR, EAL, "Locking %s failed: %s \n", path, strerror(errno)); - return -1; - } - - close(fd); - - return 0; -} - -/* open shared config file and, if present, map the config. - * having no config file is not an error condition, as we later check if - * ivshmem_config is NULL (if it is, that means nothing was mapped). */ -static int -open_shared_config(void) -{ - char path[PATH_MAX]; - int fd; - - /* build ivshmem config file path */ - snprintf(path, sizeof(path), IVSHMEM_CONFIG_PATH, - internal_config.hugefile_prefix); - - fd = open(path, O_RDONLY); - - /* if the file doesn't exist, just return success */ - if (fd < 0 && errno == ENOENT) - return 0; - /* else we have an error condition */ - else if (fd < 0) { - RTE_LOG(ERR, EAL, "Could not open %s: %s\n", - path, strerror(errno)); - return -1; - } - - /* try ex-locking first - if the lock *does* succeed, this means it's a - * stray config file, so it should be deleted. - */ - if (flock(fd, LOCK_EX | LOCK_NB) != -1) { - - /* if we can't remove the file, something is wrong */ - if (unlink(path) < 0) { - RTE_LOG(ERR, EAL, "Could not remove %s: %s\n", path, - strerror(errno)); - return -1; - } - - /* release the lock */ - flock(fd, LOCK_UN); - close(fd); - - /* return success as having a stray config file is equivalent to not - * having config file at all. - */ - return 0; - } - - ivshmem_config = mmap(NULL, sizeof(struct ivshmem_shared_config), - PROT_READ, MAP_SHARED, fd, 0); - - if (ivshmem_config == MAP_FAILED) - return -1; - - /* place a shared lock on config file */ - if (flock(fd, LOCK_SH | LOCK_NB) == -1) { - RTE_LOG(ERR, EAL, "Locking %s failed: %s \n", path, strerror(errno)); - return -1; - } - - close(fd); - - return 0; -} - -/* - * This function does the following: - * - * 1) Builds a table of ivshmem_segments with proper offset alignment - * 2) Cleans up that table so that we don't have any overlapping or adjacent - * memory segments - * 3) Creates memsegs from this table and maps them into memory. - */ -static inline int -map_all_segments(void) -{ - struct ivshmem_segment ms_tbl[RTE_MAX_MEMSEG]; - struct ivshmem_pci_device * pci_dev; - struct rte_mem_config * mcfg; - struct ivshmem_segment * seg; - int fd, fd_zero; - unsigned i, j; - struct rte_memzone mz; - struct rte_memseg ms; - void * base_addr; - uint64_t align, len; - phys_addr_t ioremap_addr; - - ioremap_addr = 0; - - memset(ms_tbl, 0, sizeof(ms_tbl)); - memset(&mz, 0, sizeof(struct rte_memzone)); - memset(&ms, 0, sizeof(struct rte_memseg)); - - /* first, build a table of memsegs to map, to avoid failed mmaps due to - * overlaps - */ - for (i = 0; i < ivshmem_config->segment_idx && i <= RTE_MAX_MEMSEG; i++) { - if (i == RTE_MAX_MEMSEG) { - RTE_LOG(ERR, EAL, "Too many segments requested!\n"); - return -1; - } - - seg = &ivshmem_config->segment[i]; - - /* copy segment to table */ - memcpy(&ms_tbl[i], seg, sizeof(struct ivshmem_segment)); - - /* find ioremap addr */ - for (j = 0; j < DIM(ivshmem_config->pci_devs); j++) { - pci_dev = &ivshmem_config->pci_devs[j]; - if (!strncmp(pci_dev->path, seg->path, sizeof(pci_dev->path))) { - ioremap_addr = pci_dev->ioremap_addr; - break; - } - } - if (ioremap_addr == 0) { - RTE_LOG(ERR, EAL, "Cannot find ioremap addr!\n"); - return -1; - } - - /* work out alignments */ - align = seg->entry.mz.addr_64 - - RTE_ALIGN_FLOOR(seg->entry.mz.addr_64, 0x1000); - len = RTE_ALIGN_CEIL(seg->entry.mz.len + align, 0x1000); - - /* save original alignments */ - ms_tbl[i].align = align; - - /* create a memory zone */ - mz.addr_64 = seg->entry.mz.addr_64 - align; - mz.len = len; - mz.hugepage_sz = seg->entry.mz.hugepage_sz; - mz.phys_addr = seg->entry.mz.phys_addr - align; - - /* find true physical address */ - mz.ioremap_addr = ioremap_addr + seg->entry.offset - align; - - ms_tbl[i].entry.offset = seg->entry.offset - align; - - memcpy(&ms_tbl[i].entry.mz, &mz, sizeof(struct rte_memzone)); - } - - /* clean up the segments */ - memseg_idx = cleanup_segments(ms_tbl, ivshmem_config->segment_idx); - - if (memseg_idx < 0) - return -1; - - mcfg = rte_eal_get_configuration()->mem_config; - - fd_zero = open("/dev/zero", O_RDWR); - - if (fd_zero < 0) { - RTE_LOG(ERR, EAL, "Cannot open /dev/zero: %s\n", strerror(errno)); - return -1; - } - - /* create memsegs and put them into DPDK memory */ - for (i = 0; i < (unsigned) memseg_idx; i++) { - - seg = &ms_tbl[i]; - - ms.addr_64 = seg->entry.mz.addr_64; - ms.hugepage_sz = seg->entry.mz.hugepage_sz; - ms.len = seg->entry.mz.len; - ms.nchannel = rte_memory_get_nchannel(); - ms.nrank = rte_memory_get_nrank(); - ms.phys_addr = seg->entry.mz.phys_addr; - ms.ioremap_addr = seg->entry.mz.ioremap_addr; - ms.socket_id = seg->entry.mz.socket_id; - - base_addr = mmap(ms.addr, ms.len, - PROT_READ | PROT_WRITE, MAP_PRIVATE, fd_zero, 0); - - if (base_addr == MAP_FAILED || base_addr != ms.addr) { - RTE_LOG(ERR, EAL, "Cannot map /dev/zero!\n"); - return -1; - } - - fd = open(seg->path, O_RDWR); - - if (fd < 0) { - RTE_LOG(ERR, EAL, "Cannot open %s: %s\n", seg->path, - strerror(errno)); - return -1; - } - - munmap(ms.addr, ms.len); - - base_addr = mmap(ms.addr, ms.len, - PROT_READ | PROT_WRITE, MAP_SHARED, fd, - seg->entry.offset); - - - if (base_addr == MAP_FAILED || base_addr != ms.addr) { - RTE_LOG(ERR, EAL, "Cannot map segment into memory: " - "expected %p got %p (%s)\n", ms.addr, base_addr, - strerror(errno)); - return -1; - } - - RTE_LOG(DEBUG, EAL, "Memory segment mapped: %p (len %" PRIx64 ") at " - "offset 0x%" PRIx64 "\n", - ms.addr, ms.len, seg->entry.offset); - - /* put the pointers back into their real positions using original - * alignment */ - ms.addr_64 += seg->align; - ms.phys_addr += seg->align; - ms.ioremap_addr += seg->align; - ms.len -= seg->align; - - /* at this point, the rest of DPDK memory is not initialized, so we - * expect memsegs to be empty */ - memcpy(&mcfg->memseg[i], &ms, - sizeof(struct rte_memseg)); - - close(fd); - - RTE_LOG(DEBUG, EAL, "IVSHMEM segment found, size: 0x%lx\n", - ms.len); - } - - return 0; -} - -/* this happens at a later stage, after general EAL memory initialization */ -int -rte_eal_ivshmem_obj_init(void) -{ - struct rte_ring_list* ring_list = NULL; - struct rte_mem_config * mcfg; - struct ivshmem_segment * seg; - struct rte_memzone * mz; - struct rte_ring * r; - struct rte_tailq_entry *te; - unsigned i, ms, idx; - uint64_t offset; - - /* secondary process would not need any object discovery - it'll all - * already be in shared config */ - if (rte_eal_process_type() != RTE_PROC_PRIMARY || ivshmem_config == NULL) - return 0; - - /* check that we have an initialised ring tail queue */ - ring_list = RTE_TAILQ_LOOKUP(RTE_TAILQ_RING_NAME, rte_ring_list); - if (ring_list == NULL) { - RTE_LOG(ERR, EAL, "No rte_ring tailq found!\n"); - return -1; - } - - mcfg = rte_eal_get_configuration()->mem_config; - - /* create memzones */ - for (i = 0; i < ivshmem_config->segment_idx && i <= RTE_MAX_MEMZONE; i++) { - - seg = &ivshmem_config->segment[i]; - - /* add memzone */ - if (mcfg->memzone_cnt == RTE_MAX_MEMZONE) { - RTE_LOG(ERR, EAL, "No more memory zones available!\n"); - return -1; - } - - idx = mcfg->memzone_cnt; - - RTE_LOG(DEBUG, EAL, "Found memzone: '%s' at %p (len 0x%" PRIx64 ")\n", - seg->entry.mz.name, seg->entry.mz.addr, seg->entry.mz.len); - - memcpy(&mcfg->memzone[idx], &seg->entry.mz, - sizeof(struct rte_memzone)); - - /* find ioremap address */ - for (ms = 0; ms <= RTE_MAX_MEMSEG; ms++) { - if (ms == RTE_MAX_MEMSEG) { - RTE_LOG(ERR, EAL, "Physical address of segment not found!\n"); - return -1; - } - if (CONTAINS(mcfg->memseg[ms], mcfg->memzone[idx])) { - offset = mcfg->memzone[idx].addr_64 - - mcfg->memseg[ms].addr_64; - mcfg->memzone[idx].ioremap_addr = mcfg->memseg[ms].ioremap_addr + - offset; - break; - } - } - - mcfg->memzone_cnt++; - } - - rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK); - - /* find rings */ - for (i = 0; i < mcfg->memzone_cnt; i++) { - mz = &mcfg->memzone[i]; - - /* check if memzone has a ring prefix */ - if (strncmp(mz->name, RTE_RING_MZ_PREFIX, - sizeof(RTE_RING_MZ_PREFIX) - 1) != 0) - continue; - - r = (struct rte_ring*) (mz->addr_64); - - te = rte_zmalloc("RING_TAILQ_ENTRY", sizeof(*te), 0); - if (te == NULL) { - RTE_LOG(ERR, EAL, "Cannot allocate ring tailq entry!\n"); - return -1; - } - - te->data = (void *) r; - - TAILQ_INSERT_TAIL(ring_list, te, next); - - RTE_LOG(DEBUG, EAL, "Found ring: '%s' at %p\n", r->name, mz->addr); - } - rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK); - -#ifdef RTE_LIBRTE_IVSHMEM_DEBUG - rte_memzone_dump(stdout); - rte_ring_list_dump(stdout); -#endif - - return 0; -} - -/* initialize ivshmem structures */ -int rte_eal_ivshmem_init(void) -{ - struct rte_pci_device * dev; - struct rte_pci_resource * res; - int fd, ret; - char path[PATH_MAX]; - - /* initialize everything to 0 */ - memset(path, 0, sizeof(path)); - ivshmem_config = NULL; - - pagesz = getpagesize(); - - RTE_LOG(DEBUG, EAL, "Searching for IVSHMEM devices...\n"); - - if (rte_eal_process_type() == RTE_PROC_SECONDARY) { - - if (open_shared_config() < 0) { - RTE_LOG(ERR, EAL, "Could not open IVSHMEM config!\n"); - return -1; - } - } - else { - - TAILQ_FOREACH(dev, &pci_device_list, next) { - - if (is_ivshmem_device(dev)) { - - /* IVSHMEM memory is always on BAR2 */ - res = &dev->mem_resource[2]; - - /* if we don't have a BAR2 */ - if (res->len == 0) - continue; - - /* construct pci device path */ - snprintf(path, sizeof(path), IVSHMEM_RESOURCE_PATH, - dev->addr.domain, dev->addr.bus, dev->addr.devid, - dev->addr.function); - - /* try to find memseg */ - fd = open(path, O_RDWR); - if (fd < 0) { - RTE_LOG(ERR, EAL, "Could not open %s\n", path); - return -1; - } - - /* check if it's a DPDK IVSHMEM device */ - ret = has_ivshmem_metadata(fd, res->len); - - /* is DPDK device */ - if (ret == 1) { - - /* config file creation is deferred until the first - * DPDK device is found. then, it has to be created - * only once. */ - if (ivshmem_config == NULL && - create_shared_config() < 0) { - RTE_LOG(ERR, EAL, "Could not create IVSHMEM config!\n"); - close(fd); - return -1; - } - - if (read_metadata(path, sizeof(path), fd, res->len) < 0) { - RTE_LOG(ERR, EAL, "Could not read metadata from" - " device %02x:%02x.%x!\n", dev->addr.bus, - dev->addr.devid, dev->addr.function); - close(fd); - return -1; - } - - if (ivshmem_config->pci_devs_idx == RTE_LIBRTE_IVSHMEM_MAX_PCI_DEVS) { - RTE_LOG(WARNING, EAL, - "IVSHMEM PCI device limit exceeded. Increase " - "CONFIG_RTE_LIBRTE_IVSHMEM_MAX_PCI_DEVS in " - "your config file.\n"); - break; - } - - RTE_LOG(INFO, EAL, "Found IVSHMEM device %02x:%02x.%x\n", - dev->addr.bus, dev->addr.devid, dev->addr.function); - - ivshmem_config->pci_devs[ivshmem_config->pci_devs_idx].ioremap_addr = res->phys_addr; - snprintf(ivshmem_config->pci_devs[ivshmem_config->pci_devs_idx].path, - sizeof(ivshmem_config->pci_devs[ivshmem_config->pci_devs_idx].path), - "%s", path); - - ivshmem_config->pci_devs_idx++; - } - /* failed to read */ - else if (ret < 0) { - RTE_LOG(ERR, EAL, "Could not read IVSHMEM device: %s\n", - strerror(errno)); - close(fd); - return -1; - } - /* not a DPDK device */ - else - RTE_LOG(DEBUG, EAL, "Skipping non-DPDK IVSHMEM device\n"); - - /* close the BAR fd */ - close(fd); - } - } - } - - /* ivshmem_config is not NULL only if config was created and/or mapped */ - if (ivshmem_config) { - if (map_all_segments() < 0) { - RTE_LOG(ERR, EAL, "Mapping IVSHMEM segments failed!\n"); - return -1; - } - } - else { - RTE_LOG(DEBUG, EAL, "No IVSHMEM configuration found! \n"); - } - - return 0; -} - -#endif diff --git a/lib/librte_eal/linuxapp/eal/eal_log.c b/lib/librte_eal/linuxapp/eal/eal_log.c index d3911004..e3a50aa3 100644 --- a/lib/librte_eal/linuxapp/eal/eal_log.c +++ b/lib/librte_eal/linuxapp/eal/eal_log.c @@ -97,45 +97,7 @@ rte_eal_log_init(const char *id, int facility) openlog(id, LOG_NDELAY | LOG_PID, facility); - if (rte_eal_common_log_init(log_stream) < 0) - return -1; - - return 0; -} - -/* early logs */ - -/* - * early log function, used before rte_eal_log_init - */ -static ssize_t -early_log_write(__attribute__((unused)) void *c, const char *buf, size_t size) -{ - ssize_t ret; - ret = fwrite(buf, size, 1, stdout); - fflush(stdout); - if (ret == 0) - return -1; - return ret; -} - -static cookie_io_functions_t early_log_func = { - .write = early_log_write, -}; -static FILE *early_log_stream; + eal_log_set_default(log_stream); -/* - * init the log library, called by rte_eal_init() to enable early - * logs - */ -int -rte_eal_log_early_init(void) -{ - early_log_stream = fopencookie(NULL, "w+", early_log_func); - if (early_log_stream == NULL) { - printf("Cannot configure early_log_stream\n"); - return -1; - } - rte_openlog_stream(early_log_stream); return 0; } diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c index c04cff0c..a956bb22 100644 --- a/lib/librte_eal/linuxapp/eal/eal_memory.c +++ b/lib/librte_eal/linuxapp/eal/eal_memory.c @@ -376,25 +376,15 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, void *vma_addr = NULL; size_t vma_len = 0; -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - RTE_SET_USED(vma_len); -#endif - for (i = 0; i < hpi->num_pages[0]; i++) { uint64_t hugepage_sz = hpi->hugepage_sz; if (orig) { hugepg_tbl[i].file_id = i; hugepg_tbl[i].size = hugepage_sz; -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - eal_get_hugefile_temp_path(hugepg_tbl[i].filepath, - sizeof(hugepg_tbl[i].filepath), hpi->hugedir, - hugepg_tbl[i].file_id); -#else eal_get_hugefile_path(hugepg_tbl[i].filepath, sizeof(hugepg_tbl[i].filepath), hpi->hugedir, hugepg_tbl[i].file_id); -#endif hugepg_tbl[i].filepath[sizeof(hugepg_tbl[i].filepath) - 1] = '\0'; } #ifndef RTE_ARCH_64 @@ -408,8 +398,6 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, continue; } #endif - -#ifndef RTE_EAL_SINGLE_FILE_SEGMENTS else if (vma_len == 0) { unsigned j, num_pages; @@ -439,10 +427,9 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, if (vma_addr == NULL) vma_len = hugepage_sz; } -#endif /* try to create hugepage file */ - fd = open(hugepg_tbl[i].filepath, O_CREAT | O_RDWR, 0755); + fd = open(hugepg_tbl[i].filepath, O_CREAT | O_RDWR, 0600); if (fd < 0) { RTE_LOG(DEBUG, EAL, "%s(): open failed: %s\n", __func__, strerror(errno)); @@ -505,169 +492,6 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, return i; } -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - -/* - * Remaps all hugepages into single file segments - */ -static int -remap_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) -{ - int fd; - unsigned i = 0, j, num_pages, page_idx = 0; - void *vma_addr = NULL, *old_addr = NULL, *page_addr = NULL; - size_t vma_len = 0; - size_t hugepage_sz = hpi->hugepage_sz; - size_t total_size, offset; - char filepath[MAX_HUGEPAGE_PATH]; - phys_addr_t physaddr; - int socket; - - while (i < hpi->num_pages[0]) { - -#ifndef RTE_ARCH_64 - /* for 32-bit systems, don't remap 1G pages and 16G pages, - * just reuse original map address as final map address. - */ - if ((hugepage_sz == RTE_PGSIZE_1G) - || (hugepage_sz == RTE_PGSIZE_16G)) { - hugepg_tbl[i].final_va = hugepg_tbl[i].orig_va; - hugepg_tbl[i].orig_va = NULL; - i++; - continue; - } -#endif - - /* reserve a virtual area for next contiguous - * physical block: count the number of - * contiguous physical pages. */ - for (j = i+1; j < hpi->num_pages[0] ; j++) { -#ifdef RTE_ARCH_PPC_64 - /* The physical addresses are sorted in descending - * order on PPC64 */ - if (hugepg_tbl[j].physaddr != - hugepg_tbl[j-1].physaddr - hugepage_sz) - break; -#else - if (hugepg_tbl[j].physaddr != - hugepg_tbl[j-1].physaddr + hugepage_sz) - break; -#endif - } - num_pages = j - i; - vma_len = num_pages * hugepage_sz; - - socket = hugepg_tbl[i].socket_id; - - /* get the biggest virtual memory area up to - * vma_len. If it fails, vma_addr is NULL, so - * let the kernel provide the address. */ - vma_addr = get_virtual_area(&vma_len, hpi->hugepage_sz); - - /* If we can't find a big enough virtual area, work out how many pages - * we are going to get */ - if (vma_addr == NULL) - j = i + 1; - else if (vma_len != num_pages * hugepage_sz) { - num_pages = vma_len / hugepage_sz; - j = i + num_pages; - - } - - hugepg_tbl[page_idx].file_id = page_idx; - eal_get_hugefile_path(filepath, - sizeof(filepath), - hpi->hugedir, - hugepg_tbl[page_idx].file_id); - - /* try to create hugepage file */ - fd = open(filepath, O_CREAT | O_RDWR, 0755); - if (fd < 0) { - RTE_LOG(ERR, EAL, "%s(): open failed: %s\n", __func__, strerror(errno)); - return -1; - } - - total_size = 0; - for (;i < j; i++) { - - /* unmap current segment */ - if (total_size > 0) - munmap(vma_addr, total_size); - - /* unmap original page */ - munmap(hugepg_tbl[i].orig_va, hugepage_sz); - unlink(hugepg_tbl[i].filepath); - - total_size += hugepage_sz; - - old_addr = vma_addr; - - /* map new, bigger segment, and populate page tables, - * the kernel fills this segment with zeros */ - vma_addr = mmap(vma_addr, total_size, - PROT_READ | PROT_WRITE, MAP_SHARED | MAP_POPULATE, fd, 0); - - if (vma_addr == MAP_FAILED || vma_addr != old_addr) { - RTE_LOG(ERR, EAL, "%s(): mmap failed: %s\n", __func__, strerror(errno)); - close(fd); - return -1; - } - } - - /* set shared flock on the file. */ - if (flock(fd, LOCK_SH | LOCK_NB) == -1) { - RTE_LOG(ERR, EAL, "%s(): Locking file failed:%s \n", - __func__, strerror(errno)); - close(fd); - return -1; - } - - snprintf(hugepg_tbl[page_idx].filepath, MAX_HUGEPAGE_PATH, "%s", - filepath); - - physaddr = rte_mem_virt2phy(vma_addr); - - if (physaddr == RTE_BAD_PHYS_ADDR) - return -1; - - hugepg_tbl[page_idx].final_va = vma_addr; - - hugepg_tbl[page_idx].physaddr = physaddr; - - hugepg_tbl[page_idx].repeated = num_pages; - - hugepg_tbl[page_idx].socket_id = socket; - - close(fd); - - /* verify the memory segment - that is, check that every VA corresponds - * to the physical address we expect to see - */ - for (offset = 0; offset < vma_len; offset += hugepage_sz) { - uint64_t expected_physaddr; - - expected_physaddr = hugepg_tbl[page_idx].physaddr + offset; - page_addr = RTE_PTR_ADD(vma_addr, offset); - physaddr = rte_mem_virt2phy(page_addr); - - if (physaddr != expected_physaddr) { - RTE_LOG(ERR, EAL, "Segment sanity check failed: wrong physaddr " - "at %p (offset 0x%" PRIx64 ": 0x%" PRIx64 - " (expected 0x%" PRIx64 ")\n", - page_addr, offset, physaddr, expected_physaddr); - return -1; - } - } - - page_idx++; - } - - /* zero out the rest */ - memset(&hugepg_tbl[page_idx], 0, (hpi->num_pages[0] - page_idx) * sizeof(struct hugepage_file)); - return page_idx; -} -#else/* RTE_EAL_SINGLE_FILE_SEGMENTS=n */ - /* Unmap all hugepages from original mapping */ static int unmap_all_hugepages_orig(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) @@ -681,7 +505,6 @@ unmap_all_hugepages_orig(struct hugepage_file *hugepg_tbl, struct hugepage_info } return 0; } -#endif /* RTE_EAL_SINGLE_FILE_SEGMENTS */ /* * Parse /proc/self/numa_maps to get the NUMA socket ID for each huge @@ -875,12 +698,6 @@ unmap_unneeded_hugepages(struct hugepage_file *hugepg_tbl, for (page = 0; page < nrpages; page++) { struct hugepage_file *hp = &hugepg_tbl[page]; -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - /* if this page was already cleared */ - if (hp->final_va == NULL) - continue; -#endif - /* find a page that matches the criteria */ if ((hp->size == hpi[size].hugepage_sz) && (hp->socket_id == (int) socket)) { @@ -889,11 +706,7 @@ unmap_unneeded_hugepages(struct hugepage_file *hugepg_tbl, if (pages_found == hpi[size].num_pages[socket]) { uint64_t unmap_len; -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - unmap_len = hp->size * hp->repeated; -#else unmap_len = hp->size; -#endif /* get start addr and len of the remaining segment */ munmap(hp->final_va, (size_t) unmap_len); @@ -904,50 +717,10 @@ unmap_unneeded_hugepages(struct hugepage_file *hugepg_tbl, __func__, hp->filepath, strerror(errno)); return -1; } - } -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - /* else, check how much do we need to map */ - else { - int nr_pg_left = - hpi[size].num_pages[socket] - pages_found; - - /* if we need enough memory to fit into the segment */ - if (hp->repeated <= nr_pg_left) { - pages_found += hp->repeated; - } - /* truncate the segment */ - else { - uint64_t final_size = nr_pg_left * hp->size; - uint64_t seg_size = hp->repeated * hp->size; - - void * unmap_va = RTE_PTR_ADD(hp->final_va, - final_size); - int fd; - - munmap(unmap_va, seg_size - final_size); - - fd = open(hp->filepath, O_RDWR); - if (fd < 0) { - RTE_LOG(ERR, EAL, "Cannot open %s: %s\n", - hp->filepath, strerror(errno)); - return -1; - } - if (ftruncate(fd, final_size) < 0) { - RTE_LOG(ERR, EAL, "Cannot truncate %s: %s\n", - hp->filepath, strerror(errno)); - return -1; - } - close(fd); - - pages_found += nr_pg_left; - hp->repeated = nr_pg_left; - } - } -#else - /* else, lock the page and skip */ - else + } else { + /* lock the page and skip */ pages_found++; -#endif + } } /* match page */ } /* foreach page */ @@ -1177,9 +950,6 @@ rte_eal_hugepage_init(void) int i, j, new_memseg; int nr_hugefiles, nr_hugepages = 0; void *addr; -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - int new_pages_count[MAX_HUGEPAGE_SIZES]; -#endif test_proc_pagemap_readable(); @@ -1260,13 +1030,6 @@ rte_eal_hugepage_init(void) pages_old = hpi->num_pages[0]; pages_new = map_all_hugepages(&tmp_hp[hp_offset], hpi, 1); if (pages_new < pages_old) { -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - RTE_LOG(ERR, EAL, - "%d not %d hugepages of size %u MB allocated\n", - pages_new, pages_old, - (unsigned)(hpi->hugepage_sz / 0x100000)); - goto fail; -#else RTE_LOG(DEBUG, EAL, "%d not %d hugepages of size %u MB allocated\n", pages_new, pages_old, @@ -1278,7 +1041,6 @@ rte_eal_hugepage_init(void) hpi->num_pages[0] = pages_new; if (pages_new == 0) continue; -#endif } /* find physical addresses and sockets for each hugepage */ @@ -1297,18 +1059,6 @@ rte_eal_hugepage_init(void) qsort(&tmp_hp[hp_offset], hpi->num_pages[0], sizeof(struct hugepage_file), cmp_physaddr); -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - /* remap all hugepages into single file segments */ - new_pages_count[i] = remap_all_hugepages(&tmp_hp[hp_offset], hpi); - if (new_pages_count[i] < 0){ - RTE_LOG(DEBUG, EAL, "Failed to remap %u MB pages\n", - (unsigned)(hpi->hugepage_sz / 0x100000)); - goto fail; - } - - /* we have processed a num of hugepages of this size, so inc offset */ - hp_offset += new_pages_count[i]; -#else /* remap all hugepages */ if (map_all_hugepages(&tmp_hp[hp_offset], hpi, 0) != hpi->num_pages[0]) { @@ -1323,7 +1073,6 @@ rte_eal_hugepage_init(void) /* we have processed a num of hugepages of this size, so inc offset */ hp_offset += hpi->num_pages[0]; -#endif } huge_recover_sigbus(); @@ -1331,14 +1080,7 @@ rte_eal_hugepage_init(void) if (internal_config.memory == 0 && internal_config.force_sockets == 0) internal_config.memory = eal_get_hugepage_mem_size(); -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - nr_hugefiles = 0; - for (i = 0; i < (int) internal_config.num_hugepage_sizes; i++) { - nr_hugefiles += new_pages_count[i]; - } -#else nr_hugefiles = nr_hugepages; -#endif /* clean out the numbers of pages */ @@ -1356,12 +1098,7 @@ rte_eal_hugepage_init(void) for (j = 0; j < nb_hpsizes; j++) { if (tmp_hp[i].size == internal_config.hugepage_info[j].hugepage_sz) { -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - internal_config.hugepage_info[j].num_pages[socket] += - tmp_hp[i].repeated; -#else internal_config.hugepage_info[j].num_pages[socket]++; -#endif } } } @@ -1436,15 +1173,8 @@ rte_eal_hugepage_init(void) free(tmp_hp); tmp_hp = NULL; - /* find earliest free memseg - this is needed because in case of IVSHMEM, - * segments might have already been initialized */ - for (j = 0; j < RTE_MAX_MEMSEG; j++) - if (mcfg->memseg[j].addr == NULL) { - /* move to previous segment and exit loop */ - j--; - break; - } - + /* first memseg index shall be 0 after incrementing it below */ + j = -1; for (i = 0; i < nr_hugefiles; i++) { new_memseg = 0; @@ -1482,11 +1212,7 @@ rte_eal_hugepage_init(void) mcfg->memseg[j].phys_addr = hugepage[i].physaddr; mcfg->memseg[j].addr = hugepage[i].final_va; -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - mcfg->memseg[j].len = hugepage[i].size * hugepage[i].repeated; -#else mcfg->memseg[j].len = hugepage[i].size; -#endif mcfg->memseg[j].socket_id = hugepage[i].socket_id; mcfg->memseg[j].hugepage_sz = hugepage[i].size; } @@ -1598,15 +1324,6 @@ rte_eal_hugepage_attach(void) if (mcfg->memseg[s].len == 0) break; -#ifdef RTE_LIBRTE_IVSHMEM - /* - * if segment has ioremap address set, it's an IVSHMEM segment and - * doesn't need mapping as it was already mapped earlier - */ - if (mcfg->memseg[s].ioremap_addr != 0) - continue; -#endif - /* * fdzero is mmapped to get a contiguous block of virtual * addresses of the appropriate memseg size. @@ -1616,13 +1333,21 @@ rte_eal_hugepage_attach(void) PROT_READ, MAP_PRIVATE, fd_zero, 0); if (base_addr == MAP_FAILED || base_addr != mcfg->memseg[s].addr) { - RTE_LOG(ERR, EAL, "Could not mmap %llu bytes " - "in /dev/zero to requested address [%p]: '%s'\n", - (unsigned long long)mcfg->memseg[s].len, - mcfg->memseg[s].addr, strerror(errno)); max_seg = s; - if (base_addr != MAP_FAILED) + if (base_addr != MAP_FAILED) { + /* errno is stale, don't use */ + RTE_LOG(ERR, EAL, "Could not mmap %llu bytes " + "in /dev/zero at [%p], got [%p] - " + "please use '--base-virtaddr' option\n", + (unsigned long long)mcfg->memseg[s].len, + mcfg->memseg[s].addr, base_addr); munmap(base_addr, mcfg->memseg[s].len); + } else { + RTE_LOG(ERR, EAL, "Could not mmap %llu bytes " + "in /dev/zero at [%p]: '%s'\n", + (unsigned long long)mcfg->memseg[s].len, + mcfg->memseg[s].addr, strerror(errno)); + } if (aslr_enabled() > 0) { RTE_LOG(ERR, EAL, "It is recommended to " "disable ASLR in the kernel " @@ -1648,16 +1373,6 @@ rte_eal_hugepage_attach(void) void *addr, *base_addr; uintptr_t offset = 0; size_t mapping_size; -#ifdef RTE_LIBRTE_IVSHMEM - /* - * if segment has ioremap address set, it's an IVSHMEM segment and - * doesn't need mapping as it was already mapped earlier - */ - if (mcfg->memseg[s].ioremap_addr != 0) { - s++; - continue; - } -#endif /* * free previously mapped memory so we can map the * hugepages into the space @@ -1676,11 +1391,7 @@ rte_eal_hugepage_attach(void) hp[i].filepath); goto error; } -#ifdef RTE_EAL_SINGLE_FILE_SEGMENTS - mapping_size = hp[i].size * hp[i].repeated; -#else mapping_size = hp[i].size; -#endif addr = mmap(RTE_PTR_ADD(base_addr, offset), mapping_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c b/lib/librte_eal/linuxapp/eal/eal_pci.c index cd9de7cc..876ba381 100644 --- a/lib/librte_eal/linuxapp/eal/eal_pci.c +++ b/lib/librte_eal/linuxapp/eal/eal_pci.c @@ -350,13 +350,13 @@ pci_scan_one(const char *dirname, uint16_t domain, uint8_t bus, dirname); if (access(filename, R_OK) != 0) { /* if no NUMA support, set default to 0 */ - dev->numa_node = 0; + dev->device.numa_node = 0; } else { if (eal_parse_sysfs_value(filename, &tmp) < 0) { free(dev); return -1; } - dev->numa_node = tmp; + dev->device.numa_node = tmp; } /* parse resources */ @@ -390,6 +390,7 @@ pci_scan_one(const char *dirname, uint16_t domain, uint8_t bus, /* device is valid, add in list (sorted) */ if (TAILQ_EMPTY(&pci_device_list)) { + rte_eal_device_insert(&dev->device); TAILQ_INSERT_TAIL(&pci_device_list, dev, next); } else { struct rte_pci_device *dev2; @@ -402,6 +403,7 @@ pci_scan_one(const char *dirname, uint16_t domain, uint8_t bus, if (ret < 0) { TAILQ_INSERT_BEFORE(dev2, dev, next); + rte_eal_device_insert(&dev->device); } else { /* already registered */ dev2->kdrv = dev->kdrv; dev2->max_vfs = dev->max_vfs; @@ -411,12 +413,26 @@ pci_scan_one(const char *dirname, uint16_t domain, uint8_t bus, } return 0; } + rte_eal_device_insert(&dev->device); TAILQ_INSERT_TAIL(&pci_device_list, dev, next); } return 0; } +int +pci_update_device(const struct rte_pci_addr *addr) +{ + char filename[PATH_MAX]; + + snprintf(filename, sizeof(filename), "%s/" PCI_PRI_FMT, + pci_get_sysfs_path(), addr->domain, addr->bus, addr->devid, + addr->function); + + return pci_scan_one(filename, addr->domain, addr->bus, addr->devid, + addr->function); +} + /* * split up a pci address into its constituent parts. */ @@ -743,9 +759,6 @@ rte_eal_pci_ioport_unmap(struct rte_pci_ioport *p) int rte_eal_pci_init(void) { - TAILQ_INIT(&pci_driver_list); - TAILQ_INIT(&pci_device_list); - /* for debug purposes, PCI can be disabled */ if (internal_config.no_pci) return 0; diff --git a/lib/librte_eal/linuxapp/eal/include/exec-env/rte_interrupts.h b/lib/librte_eal/linuxapp/eal/include/exec-env/rte_interrupts.h index 3dacbff8..d459bf48 100644 --- a/lib/librte_eal/linuxapp/eal/include/exec-env/rte_interrupts.h +++ b/lib/librte_eal/linuxapp/eal/include/exec-env/rte_interrupts.h @@ -82,6 +82,7 @@ struct rte_epoll_event { /** Handle for interrupts. */ struct rte_intr_handle { + RTE_STD_C11 union { int vfio_dev_fd; /**< VFIO device file descriptor */ int uio_cfg_fd; /**< UIO config file descriptor diff --git a/lib/librte_eal/linuxapp/eal/include/exec-env/rte_kni_common.h b/lib/librte_eal/linuxapp/eal/include/exec-env/rte_kni_common.h index 2acdfd9b..09713b0c 100644 --- a/lib/librte_eal/linuxapp/eal/include/exec-env/rte_kni_common.h +++ b/lib/librte_eal/linuxapp/eal/include/exec-env/rte_kni_common.h @@ -61,6 +61,9 @@ #ifdef __KERNEL__ #include <linux/if.h> +#define RTE_STD_C11 +#else +#include <rte_common.h> #endif /** @@ -85,6 +88,7 @@ enum rte_kni_req_id { */ struct rte_kni_request { uint32_t req_id; /**< Request id */ + RTE_STD_C11 union { uint32_t new_mtu; /**< New MTU */ uint8_t if_up; /**< 1: interface up, 0: interface down */ @@ -102,7 +106,7 @@ struct rte_kni_fifo { volatile unsigned read; /**< Next position to be read */ unsigned len; /**< Circular buffer length */ unsigned elem_size; /**< Pointer size - for 32/64 bit OS */ - void * volatile buffer[0]; /**< The buffer contains mbuf pointers */ + void *volatile buffer[]; /**< The buffer contains mbuf pointers */ }; /* @@ -111,7 +115,8 @@ struct rte_kni_fifo { */ struct rte_kni_mbuf { void *buf_addr __attribute__((__aligned__(RTE_CACHE_LINE_SIZE))); - char pad0[10]; + uint64_t buf_physaddr; + char pad0[2]; uint16_t data_off; /**< Start address of data in segment buffer. */ char pad1[2]; uint8_t nb_segs; /**< Number of segments. */ @@ -159,6 +164,7 @@ struct rte_kni_device_info { uint16_t group_id; /**< Group ID */ uint32_t core_id; /**< core ID to bind for kernel thread */ + __extension__ uint8_t force_bind : 1; /**< Flag for kernel thread binding */ /* mbuf size */ diff --git a/lib/librte_eal/linuxapp/eal/rte_eal_version.map b/lib/librte_eal/linuxapp/eal/rte_eal_version.map index db8c9845..83721ba5 100644 --- a/lib/librte_eal/linuxapp/eal/rte_eal_version.map +++ b/lib/librte_eal/linuxapp/eal/rte_eal_version.map @@ -166,3 +166,15 @@ DPDK_16.07 { rte_thread_setname; } DPDK_16.04; + +DPDK_16.11 { + global: + + rte_delay_us_block; + rte_delay_us_callback_register; + rte_eal_dev_attach; + rte_eal_dev_detach; + rte_eal_vdrv_register; + rte_eal_vdrv_unregister; + +} DPDK_16.07; |