X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=lib%2Flibrte_eal%2Flinuxapp%2Feal%2Feal_memory.c;h=42a29fafd8d41b23df7ec544ba32f381c1fa85ce;hb=e33c4d97913e06b46573a9a43beaf979f8444790;hp=296f17272c522221ac3ed628f7125aad6d0b7201;hpb=1896b4ec5e7ad5089fa17120bebf17d5dea8f476;p=dpdk.git diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c index 296f17272c..42a29fafd8 100644 --- a/lib/librte_eal/linuxapp/eal/eal_memory.c +++ b/lib/librte_eal/linuxapp/eal/eal_memory.c @@ -1,13 +1,13 @@ /*- * 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 @@ -17,7 +17,7 @@ * * 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 @@ -80,12 +80,13 @@ #include #include #include +#include +#include #include #include #include #include -#include #include #include #include @@ -98,6 +99,13 @@ #include "eal_filesystem.h" #include "eal_hugepages.h" +#ifdef RTE_LIBRTE_XEN_DOM0 +int rte_xen_dom0_supported(void) +{ + return internal_config.xen_dom0_support; +} +#endif + /** * @file * Huge page mapping under linux @@ -112,8 +120,28 @@ static uint64_t baseaddr_offset; +static unsigned proc_pagemap_readable; + #define RANDOMIZE_VA_SPACE_FILE "/proc/sys/kernel/randomize_va_space" +static void +test_proc_pagemap_readable(void) +{ + int fd = open("/proc/self/pagemap", O_RDONLY); + + if (fd < 0) { + RTE_LOG(ERR, EAL, + "Cannot open /proc/self/pagemap: %s. " + "virt2phys address translation will not work\n", + strerror(errno)); + return; + } + + /* Is readable */ + close(fd); + proc_pagemap_readable = 1; +} + /* Lock page in physical memory and prevent from swapping. */ int rte_mem_lock_page(const void *virt) @@ -131,42 +159,69 @@ phys_addr_t rte_mem_virt2phy(const void *virtaddr) { int fd; - uint64_t page, physaddr, virtual; + uint64_t page, physaddr; unsigned long virt_pfn; int page_size; + off_t offset; + + /* when using dom0, /proc/self/pagemap always returns 0, check in + * dpdk memory by browsing the memsegs */ + if (rte_xen_dom0_supported()) { + struct rte_mem_config *mcfg; + struct rte_memseg *memseg; + unsigned i; + + mcfg = rte_eal_get_configuration()->mem_config; + for (i = 0; i < RTE_MAX_MEMSEG; i++) { + memseg = &mcfg->memseg[i]; + if (memseg->addr == NULL) + break; + if (virtaddr > memseg->addr && + virtaddr < RTE_PTR_ADD(memseg->addr, + memseg->len)) { + return memseg->phys_addr + + RTE_PTR_DIFF(virtaddr, memseg->addr); + } + } + + return RTE_BAD_PHYS_ADDR; + } + + /* Cannot parse /proc/self/pagemap, no need to log errors everywhere */ + if (!proc_pagemap_readable) + return RTE_BAD_PHYS_ADDR; /* standard page size */ page_size = getpagesize(); - virtual = (uint64_t) virtaddr; fd = open("/proc/self/pagemap", O_RDONLY); if (fd < 0) { RTE_LOG(ERR, EAL, "%s(): cannot open /proc/self/pagemap: %s\n", __func__, strerror(errno)); - return (uint64_t) -1; + return RTE_BAD_PHYS_ADDR; } - off_t offset; virt_pfn = (unsigned long)virtaddr / page_size; offset = sizeof(uint64_t) * virt_pfn; if (lseek(fd, offset, SEEK_SET) == (off_t) -1) { RTE_LOG(ERR, EAL, "%s(): seek error in /proc/self/pagemap: %s\n", __func__, strerror(errno)); close(fd); - return (uint64_t) -1; + return RTE_BAD_PHYS_ADDR; } if (read(fd, &page, sizeof(uint64_t)) < 0) { RTE_LOG(ERR, EAL, "%s(): cannot read /proc/self/pagemap: %s\n", __func__, strerror(errno)); close(fd); - return (uint64_t) -1; + return RTE_BAD_PHYS_ADDR; } /* * the pfn (page frame number) are bits 0-54 (see * pagemap.txt in linux Documentation) */ - physaddr = ((page & 0x7fffffffffffffULL) * page_size) + (virtual % page_size); + physaddr = ((page & 0x7fffffffffffffULL) * page_size) + + ((unsigned long)virtaddr % page_size); close(fd); return physaddr; } @@ -221,7 +276,7 @@ aslr_enabled(void) } /* - * Try to mmap *size bytes in /dev/zero. If it is succesful, return the + * Try to mmap *size bytes in /dev/zero. If it is successful, return the * pointer to the mmap'd area and keep *size unmodified. Else, retry * with a smaller zone: decrease *size by hugepage_sz until it reaches * 0. In this case, return NULL. Note: this function returns an address @@ -240,7 +295,7 @@ get_virtual_area(size_t *size, size_t hugepage_sz) } else addr = NULL; - RTE_LOG(INFO, EAL, "Ask a virtual area of 0x%zx bytes\n", *size); + RTE_LOG(DEBUG, EAL, "Ask a virtual area of 0x%zx bytes\n", *size); fd = open("/dev/zero", O_RDONLY); if (fd < 0){ @@ -256,7 +311,8 @@ get_virtual_area(size_t *size, size_t hugepage_sz) if (addr == MAP_FAILED) { close(fd); - RTE_LOG(INFO, EAL, "Cannot get a virtual area\n"); + RTE_LOG(ERR, EAL, "Cannot get a virtual area: %s\n", + strerror(errno)); return NULL; } @@ -269,7 +325,7 @@ get_virtual_area(size_t *size, size_t hugepage_sz) aligned_addr &= (~(hugepage_sz - 1)); addr = (void *)(aligned_addr); - RTE_LOG(INFO, EAL, "Virtual area found at %p (size = 0x%zx)\n", + RTE_LOG(DEBUG, EAL, "Virtual area found at %p (size = 0x%zx)\n", addr, *size); /* increment offset */ @@ -278,6 +334,22 @@ get_virtual_area(size_t *size, size_t hugepage_sz) return addr; } +static sigjmp_buf huge_jmpenv; + +static void huge_sigbus_handler(int signo __rte_unused) +{ + siglongjmp(huge_jmpenv, 1); +} + +/* Put setjmp into a wrap method to avoid compiling error. Any non-volatile, + * non-static local variable in the stack frame calling sigsetjmp might be + * clobbered by a call to longjmp. + */ +static int huge_wrap_sigsetjmp(void) +{ + return sigsetjmp(huge_jmpenv, 1); +} + /* * Mmap all hugepages of hugepage table: it first open a file in * hugetlbfs, then mmap() hugepage_sz data in it. If orig is set, the @@ -285,7 +357,7 @@ get_virtual_area(size_t *size, size_t hugepage_sz) * in hugepg_tbl[i].final_va. The second mapping (when orig is 0) tries to * map continguous physical blocks in contiguous virtual blocks. */ -static int +static unsigned map_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi, int orig) { @@ -300,7 +372,7 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, #endif for (i = 0; i < hpi->num_pages[0]; i++) { - size_t hugepage_sz = hpi->hugepage_sz; + uint64_t hugepage_sz = hpi->hugepage_sz; if (orig) { hugepg_tbl[i].file_id = i; @@ -316,11 +388,12 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, #endif hugepg_tbl[i].filepath[sizeof(hugepg_tbl[i].filepath) - 1] = '\0'; } -#ifndef RTE_ARCH_X86_64 - /* for 32-bit systems, don't remap 1G pages, just reuse original - * map address as final map address. +#ifndef RTE_ARCH_64 + /* for 32-bit systems, don't remap 1G and 16G pages, just reuse + * original map address as final map address. */ - else if (hugepage_sz == RTE_PGSIZE_1G){ + else 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; continue; @@ -335,9 +408,17 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, * 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; @@ -354,34 +435,56 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, /* try to create hugepage file */ fd = open(hugepg_tbl[i].filepath, O_CREAT | O_RDWR, 0755); if (fd < 0) { - RTE_LOG(ERR, EAL, "%s(): open failed: %s\n", __func__, + RTE_LOG(DEBUG, EAL, "%s(): open failed: %s\n", __func__, strerror(errno)); - return -1; + return i; } + /* map the segment, and populate page tables, + * the kernel fills this segment with zeros */ virtaddr = mmap(vma_addr, hugepage_sz, PROT_READ | PROT_WRITE, - MAP_SHARED, fd, 0); + MAP_SHARED | MAP_POPULATE, fd, 0); if (virtaddr == MAP_FAILED) { - RTE_LOG(ERR, EAL, "%s(): mmap failed: %s\n", __func__, + RTE_LOG(DEBUG, EAL, "%s(): mmap failed: %s\n", __func__, strerror(errno)); close(fd); - return -1; + return i; } if (orig) { hugepg_tbl[i].orig_va = virtaddr; - memset(virtaddr, 0, hugepage_sz); } else { hugepg_tbl[i].final_va = virtaddr; } + if (orig) { + /* In linux, hugetlb limitations, like cgroup, are + * enforced at fault time instead of mmap(), even + * with the option of MAP_POPULATE. Kernel will send + * a SIGBUS signal. To avoid to be killed, save stack + * environment here, if SIGBUS happens, we can jump + * back here. + */ + if (huge_wrap_sigsetjmp()) { + RTE_LOG(DEBUG, EAL, "SIGBUS: Cannot mmap more " + "hugepages of size %u MB\n", + (unsigned)(hugepage_sz / 0x100000)); + munmap(virtaddr, hugepage_sz); + close(fd); + unlink(hugepg_tbl[i].filepath); + return i; + } + *(int *)virtaddr = 0; + } + + /* set shared flock on the file. */ if (flock(fd, LOCK_SH | LOCK_NB) == -1) { - RTE_LOG(ERR, EAL, "%s(): Locking file failed:%s \n", + RTE_LOG(DEBUG, EAL, "%s(): Locking file failed:%s \n", __func__, strerror(errno)); close(fd); - return -1; + return i; } close(fd); @@ -389,7 +492,8 @@ map_all_hugepages(struct hugepage_file *hugepg_tbl, vma_addr = (char *)vma_addr + hugepage_sz; vma_len -= hugepage_sz; } - return 0; + + return i; } #ifdef RTE_EAL_SINGLE_FILE_SEGMENTS @@ -412,11 +516,12 @@ remap_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) while (i < hpi->num_pages[0]) { -#ifndef RTE_ARCH_X86_64 - /* for 32-bit systems, don't remap 1G pages, just reuse original - * map address as final map address. +#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){ + 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++; @@ -428,8 +533,17 @@ remap_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) * physical block: count the number of * contiguous physical pages. */ for (j = i+1; j < hpi->num_pages[0] ; j++) { - if (hugepg_tbl[j].physaddr != hugepg_tbl[j-1].physaddr + hugepage_sz) +#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; @@ -479,22 +593,16 @@ remap_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) old_addr = vma_addr; - /* map new, bigger segment */ + /* 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, fd, 0); + 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; } - - /* touch the page. this is needed because kernel postpones mapping - * creation until the first page fault. with this, we pin down - * the page and it is marked as used and gets into process' pagemap. - */ - for (offset = 0; offset < total_size; offset += hugepage_sz) - *((volatile uint8_t*) RTE_PTR_ADD(vma_addr, offset)); } /* set shared flock on the file. */ @@ -505,7 +613,7 @@ remap_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) return -1; } - rte_snprintf(hugepg_tbl[page_idx].filepath, MAX_HUGEPAGE_PATH, "%s", + snprintf(hugepg_tbl[page_idx].filepath, MAX_HUGEPAGE_PATH, "%s", filepath); physaddr = rte_mem_virt2phy(vma_addr); @@ -542,9 +650,6 @@ remap_all_hugepages(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) } } - /* zero out the whole segment */ - memset(hugepg_tbl[page_idx].final_va, 0, total_size); - page_idx++; } @@ -586,13 +691,13 @@ find_numasocket(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) f = fopen("/proc/self/numa_maps", "r"); if (f == NULL) { - RTE_LOG(INFO, EAL, "cannot open /proc/self/numa_maps," + RTE_LOG(NOTICE, EAL, "cannot open /proc/self/numa_maps," " consider that all memory is in socket_id 0\n"); return 0; } - rte_snprintf(hugedir_str, sizeof(hugedir_str), - "%s/", hpi->hugedir); + snprintf(hugedir_str, sizeof(hugedir_str), + "%s/%s", hpi->hugedir, internal_config.hugefile_prefix); /* parse numa map */ while (fgets(buf, sizeof(buf), f) != NULL) { @@ -651,49 +756,23 @@ error: return -1; } -/* - * Sort the hugepg_tbl by physical address (lower addresses first). We - * use a slow algorithm, but we won't have millions of pages, and this - * is only done at init time. - */ static int -sort_by_physaddr(struct hugepage_file *hugepg_tbl, struct hugepage_info *hpi) +cmp_physaddr(const void *a, const void *b) { - unsigned i, j; - int smallest_idx; - uint64_t smallest_addr; - struct hugepage_file tmp; - - for (i = 0; i < hpi->num_pages[0]; i++) { - smallest_addr = 0; - smallest_idx = -1; - - /* - * browse all entries starting at 'i', and find the - * entry with the smallest addr - */ - for (j=i; j< hpi->num_pages[0]; j++) { - - if (smallest_addr == 0 || - hugepg_tbl[j].physaddr < smallest_addr) { - smallest_addr = hugepg_tbl[j].physaddr; - smallest_idx = j; - } - } - - /* should not happen */ - if (smallest_idx == -1) { - RTE_LOG(ERR, EAL, "%s(): error in physaddr sorting\n", __func__); - return -1; - } - - /* swap the 2 entries in the table */ - memcpy(&tmp, &hugepg_tbl[smallest_idx], sizeof(struct hugepage_file)); - memcpy(&hugepg_tbl[smallest_idx], &hugepg_tbl[i], - sizeof(struct hugepage_file)); - memcpy(&hugepg_tbl[i], &tmp, sizeof(struct hugepage_file)); - } - return 0; +#ifndef RTE_ARCH_PPC_64 + const struct hugepage_file *p1 = (const struct hugepage_file *)a; + const struct hugepage_file *p2 = (const struct hugepage_file *)b; +#else + /* PowerPC needs memory sorted in reverse order from x86 */ + const struct hugepage_file *p1 = (const struct hugepage_file *)b; + const struct hugepage_file *p2 = (const struct hugepage_file *)a; +#endif + if (p1->physaddr < p2->physaddr) + return -1; + else if (p1->physaddr > p2->physaddr) + return 1; + else + return 0; } /* @@ -738,6 +817,30 @@ copy_hugepages_to_shared_mem(struct hugepage_file * dst, int dest_size, return 0; } +static int +unlink_hugepage_files(struct hugepage_file *hugepg_tbl, + unsigned num_hp_info) +{ + unsigned socket, size; + int page, nrpages = 0; + + /* get total number of hugepages */ + for (size = 0; size < num_hp_info; size++) + for (socket = 0; socket < RTE_MAX_NUMA_NODES; socket++) + nrpages += + internal_config.hugepage_info[size].num_pages[socket]; + + for (page = 0; page < nrpages; page++) { + struct hugepage_file *hp = &hugepg_tbl[page]; + + if (hp->final_va != NULL && unlink(hp->filepath)) { + RTE_LOG(WARNING, EAL, "%s(): Removing %s failed: %s\n", + __func__, hp->filepath, strerror(errno)); + } + } + return 0; +} + /* * unmaps hugepages that are not going to be used. since we originally allocate * ALL hugepages (not just those we need), additional unmapping needs to be done. @@ -857,7 +960,7 @@ get_socket_mem_size(int socket) size += hpi->hugepage_sz * hpi->num_pages[socket]; } - return (size); + return size; } /* @@ -881,13 +984,53 @@ calc_num_pages_per_socket(uint64_t * memory, if (num_hp_info == 0) return -1; - for (socket = 0; socket < RTE_MAX_NUMA_NODES && total_mem != 0; socket++) { - /* if specific memory amounts per socket weren't requested */ - if (internal_config.force_sockets == 0) { + /* if specific memory amounts per socket weren't requested */ + if (internal_config.force_sockets == 0) { + int cpu_per_socket[RTE_MAX_NUMA_NODES]; + size_t default_size, total_size; + unsigned lcore_id; + + /* Compute number of cores per socket */ + memset(cpu_per_socket, 0, sizeof(cpu_per_socket)); + RTE_LCORE_FOREACH(lcore_id) { + cpu_per_socket[rte_lcore_to_socket_id(lcore_id)]++; + } + + /* + * Automatically spread requested memory amongst detected sockets according + * to number of cores from cpu mask present on each socket + */ + total_size = internal_config.memory; + for (socket = 0; socket < RTE_MAX_NUMA_NODES && total_size != 0; socket++) { + + /* Set memory amount per socket */ + default_size = (internal_config.memory * cpu_per_socket[socket]) + / rte_lcore_count(); + + /* Limit to maximum available memory on socket */ + default_size = RTE_MIN(default_size, get_socket_mem_size(socket)); + + /* Update sizes */ + memory[socket] = default_size; + total_size -= default_size; + } + + /* + * If some memory is remaining, try to allocate it by getting all + * available memory from sockets, one after the other + */ + for (socket = 0; socket < RTE_MAX_NUMA_NODES && total_size != 0; socket++) { /* take whatever is available */ - memory[socket] = RTE_MIN(get_socket_mem_size(socket), - total_mem); + default_size = RTE_MIN(get_socket_mem_size(socket) - memory[socket], + total_size); + + /* Update sizes */ + memory[socket] += default_size; + total_size -= default_size; } + } + + for (socket = 0; socket < RTE_MAX_NUMA_NODES && total_mem != 0; socket++) { /* skips if the memory on specific socket wasn't requested */ for (i = 0; i < num_hp_info && memory[socket] != 0; i++){ hp_used[i].hugedir = hp_info[i].hugedir; @@ -938,7 +1081,7 @@ calc_num_pages_per_socket(uint64_t * memory, 0x100000); available = requested - ((unsigned) (memory[socket] / 0x100000)); - RTE_LOG(INFO, EAL, "Not enough memory available on socket %u! " + RTE_LOG(ERR, EAL, "Not enough memory available on socket %u! " "Requested: %uMB, available: %uMB\n", socket, requested, available); return -1; @@ -949,13 +1092,58 @@ calc_num_pages_per_socket(uint64_t * memory, if (total_mem > 0) { requested = (unsigned) (internal_config.memory / 0x100000); available = requested - (unsigned) (total_mem / 0x100000); - RTE_LOG(INFO, EAL, "Not enough memory available! Requested: %uMB," + RTE_LOG(ERR, EAL, "Not enough memory available! Requested: %uMB," " available: %uMB\n", requested, available); return -1; } return total_num_pages; } +static inline size_t +eal_get_hugepage_mem_size(void) +{ + uint64_t size = 0; + unsigned i, j; + + for (i = 0; i < internal_config.num_hugepage_sizes; i++) { + struct hugepage_info *hpi = &internal_config.hugepage_info[i]; + if (hpi->hugedir != NULL) { + for (j = 0; j < RTE_MAX_NUMA_NODES; j++) { + size += hpi->hugepage_sz * hpi->num_pages[j]; + } + } + } + + return (size < SIZE_MAX) ? (size_t)(size) : SIZE_MAX; +} + +static struct sigaction huge_action_old; +static int huge_need_recover; + +static void +huge_register_sigbus(void) +{ + sigset_t mask; + struct sigaction action; + + sigemptyset(&mask); + sigaddset(&mask, SIGBUS); + action.sa_flags = 0; + action.sa_mask = mask; + action.sa_handler = huge_sigbus_handler; + + huge_need_recover = !sigaction(SIGBUS, &action, &huge_action_old); +} + +static void +huge_recover_sigbus(void) +{ + if (huge_need_recover) { + sigaction(SIGBUS, &huge_action_old, NULL); + huge_need_recover = 0; + } +} + /* * Prepare physical memory mapping: fill configuration structure with * these infos, return 0 on success. @@ -967,11 +1155,11 @@ calc_num_pages_per_socket(uint64_t * memory, * 6. unmap the first mapping * 7. fill memsegs in configuration with contiguous zones */ -static int +int rte_eal_hugepage_init(void) { struct rte_mem_config *mcfg; - struct hugepage_file *hugepage, *tmp_hp = NULL; + struct hugepage_file *hugepage = NULL, *tmp_hp = NULL; struct hugepage_info used_hp[MAX_HUGEPAGE_SIZES]; uint64_t memory[RTE_MAX_NUMA_NODES]; @@ -984,6 +1172,8 @@ rte_eal_hugepage_init(void) int new_pages_count[MAX_HUGEPAGE_SIZES]; #endif + test_proc_pagemap_readable(); + memset(used_hp, 0, sizeof(used_hp)); /* get pointer to global configuration */ @@ -991,11 +1181,18 @@ rte_eal_hugepage_init(void) /* hugetlbfs can be disabled */ if (internal_config.no_hugetlbfs) { - addr = malloc(internal_config.memory); + addr = mmap(NULL, internal_config.memory, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS, 0, 0); + if (addr == MAP_FAILED) { + RTE_LOG(ERR, EAL, "%s: mmap() failed: %s\n", __func__, + strerror(errno)); + return -1; + } mcfg->memseg[0].phys_addr = (phys_addr_t)(uintptr_t)addr; mcfg->memseg[0].addr = addr; + mcfg->memseg[0].hugepage_sz = RTE_PGSIZE_4K; mcfg->memseg[0].len = internal_config.memory; - mcfg->memseg[0].socket_id = SOCKET_ID_ANY; + mcfg->memseg[0].socket_id = 0; return 0; } @@ -1010,7 +1207,6 @@ rte_eal_hugepage_init(void) #endif } - /* calculate total number of hugepages available. at this point we haven't * yet started sorting them so they all are on socket 0 */ for (i = 0; i < (int) internal_config.num_hugepage_sizes; i++) { @@ -1034,8 +1230,11 @@ rte_eal_hugepage_init(void) hp_offset = 0; /* where we start the current page size entries */ + huge_register_sigbus(); + /* map all hugepages and sort them */ for (i = 0; i < (int)internal_config.num_hugepage_sizes; i ++){ + unsigned pages_old, pages_new; struct hugepage_info *hpi; /* @@ -1049,10 +1248,28 @@ rte_eal_hugepage_init(void) continue; /* map all hugepages available */ - if (map_all_hugepages(&tmp_hp[hp_offset], hpi, 1) < 0){ - RTE_LOG(DEBUG, EAL, "Failed to mmap %u MB hugepages\n", - (unsigned)(hpi->hugepage_sz / 0x100000)); + 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, + (unsigned)(hpi->hugepage_sz / 0x100000)); + + int pages = pages_old - pages_new; + + nr_hugepages -= pages; + hpi->num_pages[0] = pages_new; + if (pages_new == 0) + continue; +#endif } /* find physical addresses and sockets for each hugepage */ @@ -1068,8 +1285,8 @@ rte_eal_hugepage_init(void) goto fail; } - if (sort_by_physaddr(&tmp_hp[hp_offset], hpi) < 0) - goto fail; + 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 */ @@ -1084,8 +1301,9 @@ rte_eal_hugepage_init(void) hp_offset += new_pages_count[i]; #else /* remap all hugepages */ - if (map_all_hugepages(&tmp_hp[hp_offset], hpi, 0) < 0){ - RTE_LOG(DEBUG, EAL, "Failed to remap %u MB pages\n", + if (map_all_hugepages(&tmp_hp[hp_offset], hpi, 0) != + hpi->num_pages[0]) { + RTE_LOG(ERR, EAL, "Failed to remap %u MB pages\n", (unsigned)(hpi->hugepage_sz / 0x100000)); goto fail; } @@ -1099,6 +1317,11 @@ rte_eal_hugepage_init(void) #endif } + huge_recover_sigbus(); + + 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++) { @@ -1119,7 +1342,9 @@ rte_eal_hugepage_init(void) int socket = tmp_hp[i].socket_id; /* find a hugepage info with right size and increment num_pages */ - for (j = 0; j < (int) internal_config.num_hugepage_sizes; j++) { + const int nb_hpsizes = RTE_MIN(MAX_HUGEPAGE_SIZES, + (int)internal_config.num_hugepage_sizes); + 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 @@ -1149,13 +1374,13 @@ rte_eal_hugepage_init(void) for (i = 0; i < (int) internal_config.num_hugepage_sizes; i++) { for (j = 0; j < RTE_MAX_NUMA_NODES; j++) { if (used_hp[i].num_pages[j] > 0) { - RTE_LOG(INFO, EAL, - "Requesting %u pages of size %uMB" - " from socket %i\n", - used_hp[i].num_pages[j], - (unsigned) - (used_hp[i].hugepage_sz / 0x100000), - j); + RTE_LOG(DEBUG, EAL, + "Requesting %u pages of size %uMB" + " from socket %i\n", + used_hp[i].num_pages[j], + (unsigned) + (used_hp[i].hugepage_sz / 0x100000), + j); } } } @@ -1191,6 +1416,13 @@ rte_eal_hugepage_init(void) goto fail; } + /* free the hugepage backing files */ + if (internal_config.hugepage_unlink && + unlink_hugepage_files(tmp_hp, internal_config.num_hugepage_sizes) < 0) { + RTE_LOG(ERR, EAL, "Unlinking hugepage files failed!\n"); + goto fail; + } + /* free the temporary hugepage table */ free(tmp_hp); tmp_hp = NULL; @@ -1214,12 +1446,25 @@ rte_eal_hugepage_init(void) new_memseg = 1; else if (hugepage[i].size != hugepage[i-1].size) new_memseg = 1; + +#ifdef RTE_ARCH_PPC_64 + /* On PPC64 architecture, the mmap always start from higher + * virtual address to lower address. Here, both the physical + * address and virtual address are in descending order */ + else if ((hugepage[i-1].physaddr - hugepage[i].physaddr) != + hugepage[i].size) + new_memseg = 1; + else if (((unsigned long)hugepage[i-1].final_va - + (unsigned long)hugepage[i].final_va) != hugepage[i].size) + new_memseg = 1; +#else else if ((hugepage[i].physaddr - hugepage[i-1].physaddr) != hugepage[i].size) new_memseg = 1; else if (((unsigned long)hugepage[i].final_va - (unsigned long)hugepage[i-1].final_va) != hugepage[i].size) new_memseg = 1; +#endif if (new_memseg) { j += 1; @@ -1238,6 +1483,12 @@ rte_eal_hugepage_init(void) } /* continuation of previous memseg */ else { +#ifdef RTE_ARCH_PPC_64 + /* Use the phy and virt address of the last page as segment + * address for IBM Power architecture */ + mcfg->memseg[j].phys_addr = hugepage[i].physaddr; + mcfg->memseg[j].addr = hugepage[i].final_va; +#endif mcfg->memseg[j].len += mcfg->memseg[j].hugepage_sz; } hugepage[i].memseg_id = j; @@ -1251,14 +1502,19 @@ rte_eal_hugepage_init(void) "of memory.\n", i, nr_hugefiles, RTE_STR(CONFIG_RTE_MAX_MEMSEG), RTE_MAX_MEMSEG); - return (-ENOMEM); + goto fail; } + munmap(hugepage, nr_hugefiles * sizeof(struct hugepage_file)); + return 0; fail: - if (tmp_hp) - free(tmp_hp); + huge_recover_sigbus(); + free(tmp_hp); + if (hugepage != NULL) + munmap(hugepage, nr_hugefiles * sizeof(struct hugepage_file)); + return -1; } @@ -1280,11 +1536,11 @@ getFileSize(int fd) * configuration and finds the hugepages which form that segment, mapping them * in order to form a contiguous block in the virtual memory space */ -static int +int rte_eal_hugepage_attach(void) { const struct rte_mem_config *mcfg = rte_eal_get_configuration()->mem_config; - const struct hugepage_file *hp = NULL; + struct hugepage_file *hp = NULL; unsigned num_hp = 0; unsigned i, s = 0; /* s used to track the segment number */ off_t size; @@ -1297,10 +1553,12 @@ rte_eal_hugepage_attach(void) "into secondary processes\n"); } + test_proc_pagemap_readable(); + if (internal_config.xen_dom0_support) { #ifdef RTE_LIBRTE_XEN_DOM0 if (rte_xen_dom0_memory_attach() < 0) { - RTE_LOG(ERR, EAL,"Failed to attach memory setments of primay " + RTE_LOG(ERR, EAL, "Failed to attach memory segments of primary " "process\n"); return -1; } @@ -1364,7 +1622,7 @@ rte_eal_hugepage_attach(void) size = getFileSize(fd_hugepage); hp = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd_hugepage, 0); - if (hp == NULL) { + if (hp == MAP_FAILED) { RTE_LOG(ERR, EAL, "Could not mmap %s\n", eal_hugepage_info_path()); goto error; } @@ -1428,48 +1686,22 @@ rte_eal_hugepage_attach(void) s++; } /* unmap the hugepage config file, since we are done using it */ - munmap((void *)(uintptr_t)hp, size); + munmap(hp, size); close(fd_zero); close(fd_hugepage); return 0; error: + s = 0; + while (s < RTE_MAX_MEMSEG && mcfg->memseg[s].len > 0) { + munmap(mcfg->memseg[s].addr, mcfg->memseg[s].len); + s++; + } + if (hp != NULL && hp != MAP_FAILED) + munmap(hp, size); if (fd_zero >= 0) close(fd_zero); if (fd_hugepage >= 0) close(fd_hugepage); return -1; } - -static int -rte_eal_memdevice_init(void) -{ - struct rte_config *config; - - if (rte_eal_process_type() == RTE_PROC_SECONDARY) - return 0; - - config = rte_eal_get_configuration(); - config->mem_config->nchannel = internal_config.force_nchannel; - config->mem_config->nrank = internal_config.force_nrank; - - return 0; -} - - -/* init memory subsystem */ -int -rte_eal_memory_init(void) -{ - RTE_LOG(INFO, EAL, "Setting up memory...\n"); - const int retval = rte_eal_process_type() == RTE_PROC_PRIMARY ? - rte_eal_hugepage_init() : - rte_eal_hugepage_attach(); - if (retval < 0) - return -1; - - if (internal_config.no_shconf == 0 && rte_eal_memdevice_init() < 0) - return -1; - - return 0; -}