2 The intent of this file is to give a brief summary of hugetlbpage support in
3 the Linux kernel. This support is built on top of multiple page size support
4 that is provided by most modern architectures. For example, i386
5 architecture supports 4K and 4M (2M in PAE mode) page sizes, ia64
6 architecture supports multiple page sizes 4K, 8K, 64K, 256K, 1M, 4M, 16M,
7 256M and ppc64 supports 4K and 16M. A TLB is a cache of virtual-to-physical
8 translations. Typically this is a very scarce resource on processor.
9 Operating systems try to make best use of limited number of TLB resources.
10 This optimization is more critical now as bigger and bigger physical memories
11 (several GBs) are more readily available.
13 Users can use the huge page support in Linux kernel by either using the mmap
14 system call or standard SYSv shared memory system calls (shmget, shmat).
16 First the Linux kernel needs to be built with the CONFIG_HUGETLBFS
17 (present under "File systems") and CONFIG_HUGETLB_PAGE (selected
18 automatically when CONFIG_HUGETLBFS is selected) configuration
21 The kernel built with hugepage support should show the number of configured
22 hugepages in the system by running the "cat /proc/meminfo" command.
24 /proc/meminfo also provides information about the total number of hugetlb
25 pages configured in the kernel. It also displays information about the
26 number of free hugetlb pages at any time. It also displays information about
27 the configured hugepage size - this is needed for generating the proper
28 alignment and size of the arguments to the above system calls.
30 The output of "cat /proc/meminfo" will have lines like:
40 HugePages_Total is the size of the pool of hugepages.
41 HugePages_Free is the number of hugepages in the pool that are not yet
43 HugePages_Rsvd is short for "reserved," and is the number of hugepages
44 for which a commitment to allocate from the pool has been made, but no
45 allocation has yet been made. It's vaguely analogous to overcommit.
46 HugePages_Surp is short for "surplus," and is the number of hugepages in
47 the pool above the value in /proc/sys/vm/nr_hugepages. The maximum
48 number of surplus hugepages is controlled by
49 /proc/sys/vm/nr_overcommit_hugepages.
51 /proc/filesystems should also show a filesystem of type "hugetlbfs" configured
54 /proc/sys/vm/nr_hugepages indicates the current number of configured hugetlb
55 pages in the kernel. Super user can dynamically request more (or free some
56 pre-configured) hugepages.
57 The allocation (or deallocation) of hugetlb pages is possible only if there are
58 enough physically contiguous free pages in system (freeing of hugepages is
59 possible only if there are enough hugetlb pages free that can be transferred
60 back to regular memory pool).
62 Pages that are used as hugetlb pages are reserved inside the kernel and cannot
63 be used for other purposes.
65 Once the kernel with Hugetlb page support is built and running, a user can
66 use either the mmap system call or shared memory system calls to start using
67 the huge pages. It is required that the system administrator preallocate
68 enough memory for huge page purposes.
70 Use the following command to dynamically allocate/deallocate hugepages:
72 echo 20 > /proc/sys/vm/nr_hugepages
74 This command will try to configure 20 hugepages in the system. The success
75 or failure of allocation depends on the amount of physically contiguous
76 memory that is preset in system at this time. System administrators may want
77 to put this command in one of the local rc init files. This will enable the
78 kernel to request huge pages early in the boot process (when the possibility
79 of getting physical contiguous pages is still very high). In either
80 case, administrators will want to verify the number of hugepages actually
81 allocated by checking the sysctl or meminfo.
83 /proc/sys/vm/nr_overcommit_hugepages indicates how large the pool of
84 hugepages can grow, if more hugepages than /proc/sys/vm/nr_hugepages are
85 requested by applications. echo'ing any non-zero value into this file
86 indicates that the hugetlb subsystem is allowed to try to obtain
87 hugepages from the buddy allocator, if the normal pool is exhausted. As
88 these surplus hugepages go out of use, they are freed back to the buddy
91 Caveat: Shrinking the pool via nr_hugepages such that it becomes less
92 than the number of hugepages in use will convert the balance to surplus
93 huge pages even if it would exceed the overcommit value. As long as
94 this condition holds, however, no more surplus huge pages will be
95 allowed on the system until one of the two sysctls are increased
96 sufficiently, or the surplus huge pages go out of use and are freed.
98 With support for multiple hugepage pools at run-time available, much of
99 the hugepage userspace interface has been duplicated in sysfs. The above
100 information applies to the default hugepage size (which will be
101 controlled by the proc interfaces for backwards compatibility). The root
102 hugepage control directory is
104 /sys/kernel/mm/hugepages
106 For each hugepage size supported by the running kernel, a subdirectory
107 will exist, of the form
111 Inside each of these directories, the same set of files will exist:
114 nr_overcommit_hugepages
119 which function as described above for the default hugepage-sized case.
121 If the user applications are going to request hugepages using mmap system
122 call, then it is required that system administrator mount a file system of
126 -o uid=<value>,gid=<value>,mode=<value>,size=<value>,nr_inodes=<value> \
129 This command mounts a (pseudo) filesystem of type hugetlbfs on the directory
130 /mnt/huge. Any files created on /mnt/huge uses hugepages. The uid and gid
131 options sets the owner and group of the root of the file system. By default
132 the uid and gid of the current process are taken. The mode option sets the
133 mode of root of file system to value & 0777. This value is given in octal.
134 By default the value 0755 is picked. The size option sets the maximum value of
135 memory (huge pages) allowed for that filesystem (/mnt/huge). The size is
136 rounded down to HPAGE_SIZE. The option nr_inodes sets the maximum number of
137 inodes that /mnt/huge can use. If the size or nr_inodes option is not
138 provided on command line then no limits are set. For size and nr_inodes
139 options, you can use [G|g]/[M|m]/[K|k] to represent giga/mega/kilo. For
140 example, size=2K has the same meaning as size=2048.
142 While read system calls are supported on files that reside on hugetlb
143 file systems, write system calls are not.
145 Regular chown, chgrp, and chmod commands (with right permissions) could be
146 used to change the file attributes on hugetlbfs.
148 Also, it is important to note that no such mount command is required if the
149 applications are going to use only shmat/shmget system calls. Users who
150 wish to use hugetlb page via shared memory segment should be a member of
151 a supplementary group and system admin needs to configure that gid into
152 /proc/sys/vm/hugetlb_shm_group. It is possible for same or different
153 applications to use any combination of mmaps and shm* calls, though the
154 mount of filesystem will be required for using mmap calls.
156 *******************************************************************
159 * Example of using hugepage memory in a user application using Sys V shared
160 * memory system calls. In this example the app is requesting 256MB of
161 * memory that is backed by huge pages. The application uses the flag
162 * SHM_HUGETLB in the shmget system call to inform the kernel that it is
163 * requesting hugepages.
165 * For the ia64 architecture, the Linux kernel reserves Region number 4 for
166 * hugepages. That means the addresses starting with 0x800000... will need
167 * to be specified. Specifying a fixed address is not required on ppc64,
170 * Note: The default shared memory limit is quite low on many kernels,
171 * you may need to increase it via:
173 * echo 268435456 > /proc/sys/kernel/shmmax
175 * This will increase the maximum size per shared memory segment to 256MB.
176 * The other limit that you will hit eventually is shmall which is the
177 * total amount of shared memory in pages. To set it to 16GB on a system
178 * with a 4kB pagesize do:
180 * echo 4194304 > /proc/sys/kernel/shmall
184 #include <sys/types.h>
187 #include <sys/mman.h>
190 #define SHM_HUGETLB 04000
193 #define LENGTH (256UL*1024*1024)
195 #define dprintf(x) printf(x)
197 /* Only ia64 requires this */
199 #define ADDR (void *)(0x8000000000000000UL)
200 #define SHMAT_FLAGS (SHM_RND)
202 #define ADDR (void *)(0x0UL)
203 #define SHMAT_FLAGS (0)
212 if ((shmid = shmget(2, LENGTH,
213 SHM_HUGETLB | IPC_CREAT | SHM_R | SHM_W)) < 0) {
217 printf("shmid: 0x%x\n", shmid);
219 shmaddr = shmat(shmid, ADDR, SHMAT_FLAGS);
220 if (shmaddr == (char *)-1) {
221 perror("Shared memory attach failure");
222 shmctl(shmid, IPC_RMID, NULL);
225 printf("shmaddr: %p\n", shmaddr);
227 dprintf("Starting the writes:\n");
228 for (i = 0; i < LENGTH; i++) {
229 shmaddr[i] = (char)(i);
230 if (!(i % (1024 * 1024)))
235 dprintf("Starting the Check...");
236 for (i = 0; i < LENGTH; i++)
237 if (shmaddr[i] != (char)i)
238 printf("\nIndex %lu mismatched\n", i);
241 if (shmdt((const void *)shmaddr) != 0) {
242 perror("Detach failure");
243 shmctl(shmid, IPC_RMID, NULL);
247 shmctl(shmid, IPC_RMID, NULL);
252 *******************************************************************
255 * Example of using hugepage memory in a user application using the mmap
256 * system call. Before running this application, make sure that the
257 * administrator has mounted the hugetlbfs filesystem (on some directory
258 * like /mnt) using the command mount -t hugetlbfs nodev /mnt. In this
259 * example, the app is requesting memory of size 256MB that is backed by
262 * For ia64 architecture, Linux kernel reserves Region number 4 for hugepages.
263 * That means the addresses starting with 0x800000... will need to be
264 * specified. Specifying a fixed address is not required on ppc64, i386
270 #include <sys/mman.h>
273 #define FILE_NAME "/mnt/hugepagefile"
274 #define LENGTH (256UL*1024*1024)
275 #define PROTECTION (PROT_READ | PROT_WRITE)
277 /* Only ia64 requires this */
279 #define ADDR (void *)(0x8000000000000000UL)
280 #define FLAGS (MAP_SHARED | MAP_FIXED)
282 #define ADDR (void *)(0x0UL)
283 #define FLAGS (MAP_SHARED)
286 void check_bytes(char *addr)
288 printf("First hex is %x\n", *((unsigned int *)addr));
291 void write_bytes(char *addr)
295 for (i = 0; i < LENGTH; i++)
296 *(addr + i) = (char)i;
299 void read_bytes(char *addr)
304 for (i = 0; i < LENGTH; i++)
305 if (*(addr + i) != (char)i) {
306 printf("Mismatch at %lu\n", i);
316 fd = open(FILE_NAME, O_CREAT | O_RDWR, 0755);
318 perror("Open failed");
322 addr = mmap(ADDR, LENGTH, PROTECTION, FLAGS, fd, 0);
323 if (addr == MAP_FAILED) {
329 printf("Returned address is %p\n", addr);
334 munmap(addr, LENGTH);