Add pseudo-anonymous huge page mappings V2

PATCH In review Eric B Munson 3 participants 9 messages Aug 13, 2009 – Aug 14, 2009

Subsystems (6)

Memory management Documentation Generic include/asm header files Man-pages: manual pages for Linux -- Sections 2, 3, 4, 5, and 7 +2 more subsystems

lore

Share & tools

lore

Eric B Munson

author

2009-08-13 09:57 UTC #0 lore raw

cover Add pseudo-anonymous huge page mappings V2

This patch set adds a flag to mmap that allows the user to request a mapping to be backed with huge pages. This mapping will borrow functionality from the huge page shm code to create a file on the kernel internal mount and uses it to approximate an anonymous mapping. The MAP_HUGETLB flag is a modifier to MAP_ANONYMOUS and will not work without both flags being preset.

A new flag is necessary because there is no other way to hook into huge pages without creating a file on a hugetlbfs mount which wouldn't be MAP_ANONYMOUS.

To userspace, this mapping will behave just like an anonymous mapping because the file is not accessible outside of the kernel.

Eric B Munson (3):
hugetlbfs: Allow the creation of files suitable for MAP_PRIVATE on
the vfs internal mount
Add MAP_HUGETLB for mmaping pseudo-anonymous huge page regions Add MAP_HUGETLB example to vm/hugetlbpage.txt

 Documentation/vm/hugetlbpage.txt  |   80 +++++++++++++++++++++++++++++++++++++
 fs/hugetlbfs/inode.c              |   22 ++++++++--
 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |   17 +++++++-
 ipc/shm.c                         |    3 +-
 mm/mmap.c                         |   16 +++++++
 6 files changed, 133 insertions(+), 6 deletions(-)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Eric B Munson

author

2009-08-13 09:57 UTC #1 lore raw

PATCH 1/3 hugetlbfs: Allow the creation of files suitable for MAP_PRIVATE on the vfs internal mount V2

There are two means of creating mappings backed by huge pages:

1. mmap() a file created on hugetlbfs
2. Use shm which creates a file on an internal mount which essentially maps it MAP_SHARED

The internal mount is only used for shared mappings but there is very little that stops it being used for private mappings. This patch extends hugetlbfs_file_setup() to deal with the creation of files that will be mapped MAP_PRIVATE on the internal hugetlbfs mount. This extended API is used in a subsequent patch to implement the MAP_HUGETLB mmap() flag.

Signed-off-by Eric Munson <[email protected]>

---
Changes from V1:
Rebase to newest linux-2.6 tree

 fs/hugetlbfs/inode.c    |   22 ++++++++++++++++++----
 include/linux/hugetlb.h |   10 +++++++++-
 ipc/shm.c               |    3 ++-
 3 files changed, 29 insertions(+), 6 deletions(-)

fs/hugetlbfs/inode.c +18 -4

index 941c842..361f536 100644
@@ -506,6 +506,13 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb, uid_t uid,
 		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 		INIT_LIST_HEAD(&inode->i_mapping->private_list);
 		info = HUGETLBFS_I(inode);
+		/*
+		 * The policy is initialized here even if we are creating a
+		 * private inode because initialization simply creates an
+		 * an empty rb tree and calls spin_lock_init(), later when we
+		 * call mpol_free_shared_policy() it will just return because
+		 * the rb tree will still be empty.
+		 */
 		mpol_shared_policy_init(&info->policy, NULL);
 		switch (mode & S_IFMT) {
 		default:

@@ -930,12 +937,19 @@ static struct file_system_type hugetlbfs_fs_type = {
 
 static struct vfsmount *hugetlbfs_vfsmount;
 
-static int can_do_hugetlb_shm(void)
+static int can_do_hugetlb_shm(int creat_flags)
 {
-	return capable(CAP_IPC_LOCK) || in_group_p(sysctl_hugetlb_shm_group);
+	if (!(creat_flags & HUGETLB_SHMFS_INODE))
+		return 0;
+	if (capable(CAP_IPC_LOCK))
+		return 1;
+	if (in_group_p(sysctl_hugetlb_shm_group))
+		return 1;
+	return 0;
 }
 
-struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag)
+struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag,
+				int creat_flags)
 {
 	int error = -ENOMEM;
 	int unlock_shm = 0;

@@ -948,7 +962,7 @@ struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag)
 	if (!hugetlbfs_vfsmount)
 		return ERR_PTR(-ENOENT);
 
-	if (!can_do_hugetlb_shm()) {
+	if (!can_do_hugetlb_shm(creat_flags)) {
 		if (user_shm_lock(size, user)) {
 			unlock_shm = 1;
 			WARN_ONCE(1,

include/linux/hugetlb.h +9 -1

index 2723513..78b6ddf 100644
@@ -109,6 +109,14 @@ static inline void hugetlb_report_meminfo(struct seq_file *m)
 
 #endif /* !CONFIG_HUGETLB_PAGE */
 
+enum {
+	/*
+	 * The file will be used as an shm file so shmfs accounting rules
+	 * apply
+	 */
+	HUGETLB_SHMFS_INODE     = 0x01,
+};
+
 #ifdef CONFIG_HUGETLBFS
 struct hugetlbfs_config {
 	uid_t   uid;

@@ -146,7 +154,7 @@ static inline struct hugetlbfs_sb_info *HUGETLBFS_SB(struct super_block *sb)
 
 extern const struct file_operations hugetlbfs_file_operations;
 extern struct vm_operations_struct hugetlb_vm_ops;
-struct file *hugetlb_file_setup(const char *name, size_t, int);
+struct file *hugetlb_file_setup(const char *name, size_t, int, int);
 int hugetlb_get_quota(struct address_space *mapping, long delta);
 void hugetlb_put_quota(struct address_space *mapping, long delta);

ipc/shm.c +2 -1

index 15dd238..801c68a 100644
@@ -369,7 +369,8 @@ static int newseg(struct ipc_namespace *ns, struct ipc_params *params)
 		/* hugetlb_file_setup applies strict accounting */
 		if (shmflg & SHM_NORESERVE)
 			acctflag = VM_NORESERVE;
-		file = hugetlb_file_setup(name, size, acctflag);
+		file = hugetlb_file_setup(name, size, acctflag,
+					HUGETLB_SHMFS_INODE);
 		shp->mlock_user = current_user();
 	} else {
 		/*

Signature · 7 lines

1.6.3.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Eric B Munson

author

2009-08-13 09:57 UTC #2 lore raw

PATCH 2/3 Add MAP_HUGETLB for mmaping pseudo-anonymous huge page regions V2

This patch adds a flag for mmap that will be used to request a huge page region that will look like anonymous memory to user space. This is accomplished by using a file on the internal vfsmount. MAP_HUGETLB is a modifier of MAP_ANONYMOUS and so must be specified with it. The region will behave the same as a MAP_ANONYMOUS region using small pages.

Signed-off-by Eric B Munson <[email protected]>

---
Changes from V1
Rebase to newest linux-2.6 tree
Rename MAP_LARGEPAGE to MAP_HUGETLB to match flag name for huge page shm

 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |    7 +++++++
 mm/mmap.c                         |   16 ++++++++++++++++
 3 files changed, 24 insertions(+), 0 deletions(-)

include/asm-generic/mman-common.h +1 -0

index 3b69ad3..12f5982 100644
@@ -19,6 +19,7 @@
 #define MAP_TYPE	0x0f		/* Mask for type of mapping */
 #define MAP_FIXED	0x10		/* Interpret addr exactly */
 #define MAP_ANONYMOUS	0x20		/* don't use a file */
+#define MAP_HUGETLB	0x40		/* create a huge page mapping */
 
 #define MS_ASYNC	1		/* sync memory asynchronously */
 #define MS_INVALIDATE	2		/* invalidate the caches */

include/linux/hugetlb.h +7 -0

index 78b6ddf..b84361c 100644
@@ -109,12 +109,19 @@ static inline void hugetlb_report_meminfo(struct seq_file *m)
 
 #endif /* !CONFIG_HUGETLB_PAGE */
 
+#define HUGETLB_ANON_FILE "anon_hugepage"
+
 enum {
 	/*
 	 * The file will be used as an shm file so shmfs accounting rules
 	 * apply
 	 */
 	HUGETLB_SHMFS_INODE     = 0x01,
+	/*
+	 * The file is being created on the internal vfs mount and shmfs
+	 * accounting rules do not apply
+	 */
+	HUGETLB_ANONHUGE_INODE  = 0x02,
 };
 
 #ifdef CONFIG_HUGETLBFS

mm/mmap.c +16 -0

index 34579b2..3612b20 100644
@@ -29,6 +29,7 @@
 #include <linux/rmap.h>
 #include <linux/mmu_notifier.h>
 #include <linux/perf_counter.h>
+#include <linux/hugetlb.h>
 
 #include <asm/uaccess.h>
 #include <asm/cacheflush.h>

@@ -954,6 +955,21 @@ unsigned long do_mmap_pgoff(struct file *file, unsigned long addr,
 	if (mm->map_count > sysctl_max_map_count)
 		return -ENOMEM;
 
+	if (flags & MAP_HUGETLB) {
+		if (file)
+			return -EINVAL;
+
+		/*
+		 * VM_NORESERVE is used because the reservations will be
+		 * taken when vm_ops->mmap() is called
+		 */
+		len = ALIGN(len, huge_page_size(&default_hstate));
+		file = hugetlb_file_setup(HUGETLB_ANON_FILE, len, VM_NORESERVE,
+						HUGETLB_ANONHUGE_INODE);
+		if (IS_ERR(file))
+			return -ENOMEM;
+	}
+
 	/* Obtain the address to map to. we verify (or select) it and ensure
 	 * that it represents a valid section of the address space.
 	 */

Signature · 7 lines

1.6.3.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Eric B Munson

author

2009-08-13 09:57 UTC #3 lore raw

PATCH 3/3 Add MAP_HUGETLB example to vm/hugetlbpage.txt V2

This patch adds an example of how to use the MAP_HUGETLB flag to the vm documentation.

Signed-off-by Eric B Munson <[email protected]>

---
Changes from V1:
Rebase to newest linux-2.6 tree
Change MAP_LARGEPAGE to MAP_HUGETLB to match flag name in huge page shm

 Documentation/vm/hugetlbpage.txt |   80 ++++++++++++++++++++++++++++++++++++++
 1 files changed, 80 insertions(+), 0 deletions(-)

Documentation/vm/hugetlbpage.txt +80 -0

index ea8714f..d30fa1a 100644
@@ -337,3 +337,83 @@ int main(void)
 
 	return 0;
 }
+
+*******************************************************************
+
+/*
+ * Example of using hugepage memory in a user application using the mmap
+ * system call with MAP_LARGEPAGE flag.  Before running this program make
+ * sure the administrator has allocated enough default sized huge pages
+ * to cover the 256 MB allocation.
+ *
+ * For ia64 architecture, Linux kernel reserves Region number 4 for hugepages.
+ * That means the addresses starting with 0x800000... will need to be
+ * specified.  Specifying a fixed address is not required on ppc64, i386
+ * or x86_64.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+
+#define LENGTH (256UL*1024*1024)
+#define PROTECTION (PROT_READ | PROT_WRITE)
+
+#ifndef MAP_HUGETLB
+#define MAP_HUGETLB 0x40
+#endif
+
+/* Only ia64 requires this */
+#ifdef __ia64__
+#define ADDR (void *)(0x8000000000000000UL)
+#define FLAGS (MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB | MAP_FIXED)
+#else
+#define ADDR (void *)(0x0UL)
+#define FLAGS (MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB)
+#endif
+
+void check_bytes(char *addr)
+{
+	printf("First hex is %x\n", *((unsigned int *)addr));
+}
+
+void write_bytes(char *addr)
+{
+	unsigned long i;
+
+	for (i = 0; i < LENGTH; i++)
+		*(addr + i) = (char)i;
+}
+
+void read_bytes(char *addr)
+{
+	unsigned long i;
+
+	check_bytes(addr);
+	for (i = 0; i < LENGTH; i++)
+		if (*(addr + i) != (char)i) {
+			printf("Mismatch at %lu\n", i);
+			break;
+		}
+}
+
+int main(void)
+{
+	void *addr;
+
+	addr = mmap(ADDR, LENGTH, PROTECTION, FLAGS, 0, 0);
+	if (addr == MAP_FAILED) {
+		perror("mmap");
+		exit(1);
+	}
+
+	printf("Returned address is %p\n", addr);
+	check_bytes(addr);
+	write_bytes(addr);
+	read_bytes(addr);
+
+	munmap(addr, LENGTH);
+
+	return 0;
+}

Signature · 7 lines

1.6.3.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

David Rientjes

2009-08-13 21:49 UTC #4 lore raw

Re: [PATCH 3/3] Add MAP_HUGETLB example to vm/hugetlbpage.txt V2

Quoted 17 lines · Eric B Munson

This patch adds an example of how to use the MAP_HUGETLB flag to the vm documentation.
Signed-off-by Eric B Munson <[email protected]>
---
Changes from V1:
Rebase to newest linux-2.6 tree
Change MAP_LARGEPAGE to MAP_HUGETLB to match flag name in huge page shm
 Documentation/vm/hugetlbpage.txt |   80 ++++++++++++++++++++++++++++++++++++++
 1 files changed, 80 insertions(+), 0 deletions(-)
Documentation/vm/hugetlbpage.txt +0 -0
index ea8714f..d30fa1a 100644

Documentation/vm/hugetlbpage.txt int main(void) · line 345

Quoted 9 lines

Documentation/vm/hugetlbpage.txt +6 -0

@@ -337,3 +337,83 @@ int main(void)
 
 	return 0;
 }
+
+*******************************************************************
+
+/*
+ * Example of using hugepage memory in a user application using the mmap
+ * system call with MAP_LARGEPAGE flag.  Before running this program make

s/MAP_LARGEPAGE/MAP_HUGETLB/

Randy Dunlap

2009-08-14 00:46 UTC #5 lore raw

Re: [PATCH 3/3] Add MAP_HUGETLB example to vm/hugetlbpage.txt V2

Quoted 20 lines · David Rientjes

Quoted 18 lines · Eric B Munson
This patch adds an example of how to use the MAP_HUGETLB flag to the vm documentation.
Signed-off-by Eric B Munson <[email protected]>
---
Changes from V1:
Rebase to newest linux-2.6 tree
Change MAP_LARGEPAGE to MAP_HUGETLB to match flag name in huge page shm
 Documentation/vm/hugetlbpage.txt |   80 ++++++++++++++++++++++++++++++++++++++
 1 files changed, 80 insertions(+), 0 deletions(-)
Documentation/vm/hugetlbpage.txt +0 -0
index ea8714f..d30fa1a 100644
@@ -337,3 +337,83 @@ int main(void)
 

Quoted 10 lines

Documentation/vm/hugetlbpage.txt int main(void) · line 345
Quoted 8 lines
Documentation/vm/hugetlbpage.txt +6 -0
 	return 0;
 }
+
+*******************************************************************
+
+/*
+ * Example of using hugepage memory in a user application using the mmap
+ * system call with MAP_LARGEPAGE flag.  Before running this program make
s/MAP_LARGEPAGE/MAP_HUGETLB/

I'm (slowly) making source code examples in Documentation/ buildable, as this one should be, please.

I.e., put it in a separate source file (hugetlbpage.txt can refer to the source file if you want it to) and add a Makefile similar to other Makefiles in the Documentation/ tree.

~Randy

Eric B Munson

author

2009-08-14 12:39 UTC #6 lore raw

Re: [PATCH 3/3] Add MAP_HUGETLB example to vm/hugetlbpage.txt V2

Quoted 32 lines · Randy Dunlap

Quoted 20 lines · David Rientjes
Quoted 18 lines · Eric B Munson
This patch adds an example of how to use the MAP_HUGETLB flag to the vm documentation.
Signed-off-by Eric B Munson <[email protected]>
---
Changes from V1:
Rebase to newest linux-2.6 tree
Change MAP_LARGEPAGE to MAP_HUGETLB to match flag name in huge page shm
 Documentation/vm/hugetlbpage.txt |   80 ++++++++++++++++++++++++++++++++++++++
 1 files changed, 80 insertions(+), 0 deletions(-)
Documentation/vm/hugetlbpage.txt +0 -0
index ea8714f..d30fa1a 100644
@@ -337,3 +337,83 @@ int main(void)
 
Quoted 10 lines
Documentation/vm/hugetlbpage.txt int main(void) · line 345
Quoted 8 lines
Documentation/vm/hugetlbpage.txt +6 -0
 	return 0;
 }
+
+*******************************************************************
+
+/*
+ * Example of using hugepage memory in a user application using the mmap
+ * system call with MAP_LARGEPAGE flag.  Before running this program make
s/MAP_LARGEPAGE/MAP_HUGETLB/

Quoted 9 lines

I'm (slowly) making source code examples in Documentation/ buildable, as this one should be, please.
I.e., put it in a separate source file (hugetlbpage.txt can refer to the source file if you want it to) and add a Makefile similar to other Makefiles in the Documentation/ tree.
~Randy

I will make these changes for V3, thanks for the reviews.

Signature · 3 lines

Eric B Munson
IBM Linux Technology Center
[email protected]

David Rientjes

2009-08-13 21:49 UTC #7 lore raw

Re: [PATCH 2/3] Add MAP_HUGETLB for mmaping pseudo-anonymous huge page regions V2

In reply to Eric B Munson

Quoted 43 lines · Eric B Munson

This patch adds a flag for mmap that will be used to request a huge page region that will look like anonymous memory to user space. This is accomplished by using a file on the internal vfsmount. MAP_HUGETLB is a modifier of MAP_ANONYMOUS and so must be specified with it. The region will behave the same as a MAP_ANONYMOUS region using small pages.
Signed-off-by Eric B Munson <ebmunson-r/[email protected]>
---
Changes from V1
Rebase to newest linux-2.6 tree
Rename MAP_LARGEPAGE to MAP_HUGETLB to match flag name for huge page shm
 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |    7 +++++++
 mm/mmap.c                         |   16 ++++++++++++++++
 3 files changed, 24 insertions(+), 0 deletions(-)
include/asm-generic/mman-common.h +1 -0
index 3b69ad3..12f5982 100644
@@ -19,6 +19,7 @@
 #define MAP_TYPE	0x0f		/* Mask for type of mapping */
 #define MAP_FIXED	0x10		/* Interpret addr exactly */
 #define MAP_ANONYMOUS	0x20		/* don't use a file */
+#define MAP_HUGETLB	0x40		/* create a huge page mapping */
 
 #define MS_ASYNC	1		/* sync memory asynchronously */
 #define MS_INVALIDATE	2		/* invalidate the caches */
include/linux/hugetlb.h +2 -0
index 78b6ddf..b84361c 100644
@@ -109,12 +109,19 @@ static inline void hugetlb_report_meminfo(struct seq_file *m)
 
 #endif /* !CONFIG_HUGETLB_PAGE */
 
+#define HUGETLB_ANON_FILE "anon_hugepage"
+
 enum {
 	/*
 	 * The file will be used as an shm file so shmfs accounting rules
 	 * apply

include/linux/hugetlb.h static inline void hugetlb_report_meminfo(struct seq_file *m) · line 127

Quoted 10 lines

include/linux/hugetlb.h +5 -0

 	 */
 	HUGETLB_SHMFS_INODE     = 0x01,
+	/*
+	 * The file is being created on the internal vfs mount and shmfs
+	 * accounting rules do not apply
+	 */
+	HUGETLB_ANONHUGE_INODE  = 0x02,
 };
 
 #ifdef CONFIG_HUGETLBFS

While I think it's appropriate to use an enum here, these two "flags" can't be used together so it would probably be better to avoid the hexadecimal.

If flags were ever needed in the future, you could reserve the upper eight bits of the int for such purposes similiar to mempolicy flags.

Quoted 24 lines

mm/mmap.c +9 -0

index 34579b2..3612b20 100644
@@ -29,6 +29,7 @@
 #include <linux/rmap.h>
 #include <linux/mmu_notifier.h>
 #include <linux/perf_counter.h>
+#include <linux/hugetlb.h>
 
 #include <asm/uaccess.h>
 #include <asm/cacheflush.h>

@@ -954,6 +955,21 @@ unsigned long do_mmap_pgoff(struct file *file, unsigned long addr,
 	if (mm->map_count > sysctl_max_map_count)
 		return -ENOMEM;
 
+	if (flags & MAP_HUGETLB) {
+		if (file)
+			return -EINVAL;
+
+		/*
+		 * VM_NORESERVE is used because the reservations will be
+		 * taken when vm_ops->mmap() is called
+		 */

mm/mmap.c unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, · line 975

Quoted 10 lines

mm/mmap.c +7 -0

+		len = ALIGN(len, huge_page_size(&default_hstate));
+		file = hugetlb_file_setup(HUGETLB_ANON_FILE, len, VM_NORESERVE,
+						HUGETLB_ANONHUGE_INODE);
+		if (IS_ERR(file))
+			return -ENOMEM;
+	}
+
 	/* Obtain the address to map to. we verify (or select) it and ensure
 	 * that it represents a valid section of the address space.
 	 */

hugetlb_file_setup() can fail for reasons other than failing to reserve pages, so maybe it would be better to return PTR_ERR(file) instead of hardcoding -ENOMEM?
--
To unsubscribe from this list: send the line "unsubscribe linux-man" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

Eric B Munson

author

2009-08-14 12:40 UTC #8 lore raw

Re: [PATCH 2/3] Add MAP_HUGETLB for mmaping pseudo-anonymous huge page regions V2

Quoted 93 lines · David Rientjes

Quoted 43 lines · Eric B Munson
This patch adds a flag for mmap that will be used to request a huge page region that will look like anonymous memory to user space. This is accomplished by using a file on the internal vfsmount. MAP_HUGETLB is a modifier of MAP_ANONYMOUS and so must be specified with it. The region will behave the same as a MAP_ANONYMOUS region using small pages.
Signed-off-by Eric B Munson <[email protected]>
---
Changes from V1
Rebase to newest linux-2.6 tree
Rename MAP_LARGEPAGE to MAP_HUGETLB to match flag name for huge page shm
 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |    7 +++++++
 mm/mmap.c                         |   16 ++++++++++++++++
 3 files changed, 24 insertions(+), 0 deletions(-)
include/asm-generic/mman-common.h +1 -0
index 3b69ad3..12f5982 100644
@@ -19,6 +19,7 @@
 #define MAP_TYPE	0x0f		/* Mask for type of mapping */
 #define MAP_FIXED	0x10		/* Interpret addr exactly */
 #define MAP_ANONYMOUS	0x20		/* don't use a file */
+#define MAP_HUGETLB	0x40		/* create a huge page mapping */
 
 #define MS_ASYNC	1		/* sync memory asynchronously */
 #define MS_INVALIDATE	2		/* invalidate the caches */
include/linux/hugetlb.h +2 -0
index 78b6ddf..b84361c 100644
@@ -109,12 +109,19 @@ static inline void hugetlb_report_meminfo(struct seq_file *m)
 
 #endif /* !CONFIG_HUGETLB_PAGE */
 
+#define HUGETLB_ANON_FILE "anon_hugepage"
+
 enum {
 	/*
 	 * The file will be used as an shm file so shmfs accounting rules
 	 * apply
include/linux/hugetlb.h static inline void hugetlb_report_meminfo(struct seq_file *m) · line 127
Quoted 10 lines
include/linux/hugetlb.h +5 -0
 	 */
 	HUGETLB_SHMFS_INODE     = 0x01,
+	/*
+	 * The file is being created on the internal vfs mount and shmfs
+	 * accounting rules do not apply
+	 */
+	HUGETLB_ANONHUGE_INODE  = 0x02,
 };
 
 #ifdef CONFIG_HUGETLBFS
While I think it's appropriate to use an enum here, these two "flags" can't be used together so it would probably be better to avoid the hexadecimal.
If flags were ever needed in the future, you could reserve the upper eight bits of the int for such purposes similiar to mempolicy flags.
Quoted 30 lines
mm/mmap.c +15 -0
index 34579b2..3612b20 100644
@@ -29,6 +29,7 @@
 #include <linux/rmap.h>
 #include <linux/mmu_notifier.h>
 #include <linux/perf_counter.h>
+#include <linux/hugetlb.h>
 
 #include <asm/uaccess.h>
 #include <asm/cacheflush.h>
@@ -954,6 +955,21 @@ unsigned long do_mmap_pgoff(struct file *file, unsigned long addr,
 	if (mm->map_count > sysctl_max_map_count)
 		return -ENOMEM;
 
+	if (flags & MAP_HUGETLB) {
+		if (file)
+			return -EINVAL;
+
+		/*
+		 * VM_NORESERVE is used because the reservations will be
+		 * taken when vm_ops->mmap() is called
+		 */
+		len = ALIGN(len, huge_page_size(&default_hstate));
+		file = hugetlb_file_setup(HUGETLB_ANON_FILE, len, VM_NORESERVE,
+						HUGETLB_ANONHUGE_INODE);
+		if (IS_ERR(file))
+			return -ENOMEM;
+	}

Quoted 9 lines

mm/mmap.c unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, · line 975
Quoted 4 lines
mm/mmap.c +1 -0
+
 	/* Obtain the address to map to. we verify (or select) it and ensure
 	 * that it represents a valid section of the address space.
 	 */
hugetlb_file_setup() can fail for reasons other than failing to reserve pages, so maybe it would be better to return PTR_ERR(file) instead of hardcoding -ENOMEM?

I will make these changes for V3, thanks for your review.

Signature · 3 lines

Eric B Munson
IBM Linux Technology Center
[email protected]