Add pseudo-anonymous huge page mappings V4

PATCH In review Eric B Munson 7 participants 20 messages Aug 25, 2009 – Sep 21, 2009

Subsystems (16)

Memory management Documentation Arm (32-bit) x86 +12 more subsystems

lore

Share & tools

lore

Eric B Munson

author

2009-08-25 11:14 UTC #0 lore raw

cover Add pseudo-anonymous huge page mappings V4

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 use 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.

This patch set is meant to simplify the programming model, presently there is a large chunk of boiler plate code, contained in libhugetlbfs, required to create private, hugepage backed mappings. This patch set would allow use of hugepages without linking to libhugetlbfs or having hugetblfs mounted.

Unification of the VM code would provide these same benefits, but it has been resisted each time that it has been suggested for several
reasons: it would break PAGE_SIZE assumptions across the kernel, it makes page-table abstractions really expensive, and it does not provide any benefit on architectures that do not support huge pages, incurring fast path penalties without providing any benefit on these architectures.

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

 Documentation/vm/00-INDEX         |    2 +
 Documentation/vm/hugetlbpage.txt  |   14 ++++---
 Documentation/vm/map_hugetlb.c    |   77 +++++++++++++++++++++++++++++++++++++
 fs/hugetlbfs/inode.c              |   21 ++++++++--
 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |   19 ++++++++-
 ipc/shm.c                         |    2 +-
 mm/mmap.c                         |   19 +++++++++
 8 files changed, 142 insertions(+), 13 deletions(-)
 create mode 100644 Documentation/vm/map_hugetlb.c

--
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

2009-08-25 11:14 UTC #1 lore raw

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

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]>

---
 fs/hugetlbfs/inode.c    |   21 +++++++++++++++++----
 include/linux/hugetlb.h |   12 ++++++++++--
 ipc/shm.c               |    2 +-
 3 files changed, 28 insertions(+), 7 deletions(-)

fs/hugetlbfs/inode.c +17 -4

index cb88dac..5584d55 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:

@@ -930,13 +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 user_struct **user)
+				struct user_struct **user, int creat_flags)
 {
 	int error = -ENOMEM;
 	struct file *file;

@@ -948,7 +961,7 @@ 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)) {
 		*user = current_user();
 		if (user_shm_lock(size, *user)) {
 			WARN_ONCE(1,

include/linux/hugetlb.h +10 -2

index 5cbc620..38bb552 100644
@@ -110,6 +110,14 @@ 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     = 1,
+};
+
 #ifdef CONFIG_HUGETLBFS
 struct hugetlbfs_config {
 	uid_t   uid;

@@ -148,7 +156,7 @@ 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 size, int acct,
-						struct user_struct **user);
+				struct user_struct **user, int creat_flags);
 int hugetlb_get_quota(struct address_space *mapping, long delta);
 void hugetlb_put_quota(struct address_space *mapping, long delta);

@@ -170,7 +178,7 @@ static inline void set_file_hugepages(struct file *file)
 
 #define is_file_hugepages(file)			0
 #define set_file_hugepages(file)		BUG()
-#define hugetlb_file_setup(name,size,acct,user)	ERR_PTR(-ENOSYS)
+#define hugetlb_file_setup(name,size,acct,user,creat)	ERR_PTR(-ENOSYS)
 
 #endif /* !CONFIG_HUGETLBFS */

ipc/shm.c +1 -1

index 1bc4701..5ba4962 100644
@@ -370,7 +370,7 @@ static int newseg(struct ipc_namespace *ns, struct ipc_params *params)
 		if (shmflg & SHM_NORESERVE)
 			acctflag = VM_NORESERVE;
 		file = hugetlb_file_setup(name, size, acctflag,
-							&shp->mlock_user);
+					&shp->mlock_user, HUGETLB_SHMFS_INODE);
 	} else {
 		/*
 		 * Do not allow no accounting for OVERCOMMIT_NEVER, even

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

2009-08-25 11:14 UTC #2 lore raw

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

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]>

---
 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |    7 +++++++
 mm/mmap.c                         |   19 +++++++++++++++++++
 3 files changed, 27 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 38bb552..b0bc0fd 100644
@@ -110,12 +110,19 @@ 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     = 1,
+	/*
+	 * The file is being created on the internal vfs mount and shmfs
+	 * accounting rules do not apply
+	 */
+	HUGETLB_ANONHUGE_INODE  = 2,
 };
 
 #ifdef CONFIG_HUGETLBFS

mm/mmap.c +19 -0

index 8101de4..9ca4f26 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>

@@ -951,6 +952,24 @@ 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) {
+		struct user_struct *user = NULL;
+		if (file)
+			return -EINVAL;
+
+		/*
+		 * VM_NORESERVE is used because the reservations will be
+		 * taken when vm_ops->mmap() is called
+		 * A dummy user value is used because we are not locking
+		 * memory so no accounting is necessary
+		 */
+		len = ALIGN(len, huge_page_size(&default_hstate));
+		file = hugetlb_file_setup(HUGETLB_ANON_FILE, len, VM_NORESERVE,
+						&user, HUGETLB_ANONHUGE_INODE);
+		if (IS_ERR(file))
+			return PTR_ERR(file);
+	}
+
 	/* 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

2009-08-25 11:14 UTC #3 lore raw

PATCH 3/3 Add MAP_HUGETLB example

This patch adds an example of how to use the MAP_HUGETLB flag to the vm documentation directory and a reference to the example in hugetlbpage.txt.

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

Acked-by David Rientjes <[email protected]>

---
 Documentation/vm/00-INDEX        |    2 +
 Documentation/vm/hugetlbpage.txt |   14 ++++---
 Documentation/vm/map_hugetlb.c   |   77 ++++++++++++++++++++++++++++++++++++++
 3 files changed, 87 insertions(+), 6 deletions(-)
 create mode 100644 Documentation/vm/map_hugetlb.c

Documentation/vm/00-INDEX +2 -0

index 2f77ced..aabd973 100644
@@ -20,3 +20,5 @@ slabinfo.c
 	- source code for a tool to get reports about slabs.
 slub.txt
 	- a short users guide for SLUB.
+map_hugetlb.c
+	- an example program that uses the MAP_HUGETLB mmap flag.

Documentation/vm/hugetlbpage.txt +8 -6

index ea8714f..6a8feab 100644
@@ -146,12 +146,14 @@ Regular chown, chgrp, and chmod commands (with right permissions) could be
 used to change the file attributes on hugetlbfs.
 
 Also, it is important to note that no such mount command is required if the
-applications are going to use only shmat/shmget system calls.  Users who
-wish to use hugetlb page via shared memory segment should be a member of
-a supplementary group and system admin needs to configure that gid into
-/proc/sys/vm/hugetlb_shm_group.  It is possible for same or different
-applications to use any combination of mmaps and shm* calls, though the
-mount of filesystem will be required for using mmap calls.
+applications are going to use only shmat/shmget system calls or mmap with
+MAP_HUGETLB.  Users who wish to use hugetlb page via shared memory segment
+should be a member of a supplementary group and system admin needs to
+configure that gid into /proc/sys/vm/hugetlb_shm_group.  It is possible for
+same or different applications to use any combination of mmaps and shm*
+calls, though the mount of filesystem will be required for using mmap calls
+without MAP_HUGETLB.  For an example of how to use mmap with MAP_HUGETLB see
+map_hugetlb.c.
 
 *******************************************************************

Documentation/vm/map_hugetlb.c new file +77 -0

77 lines not on this page · raw message

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>

Andrew Morton

maintainer

2009-09-17 22:44 UTC #4 lore raw

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

In reply to Eric B Munson

Quoted 11 lines · On Tue, 25 Aug 2009 12:14:53 +0100 Eric B Munson <[email protected]> wrote:

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]>
---
 include/asm-generic/mman-common.h |    1 +
 include/linux/hugetlb.h           |    7 +++++++
 mm/mmap.c                         |   19 +++++++++++++++++++

alpha fix:

From: Andrew Morton <[email protected]>

mm/mmap.c: In function 'do_mmap_pgoff':
mm/mmap.c:953: error: 'MAP_HUGETLB' undeclared (first use in this function) mm/mmap.c:953: error: (Each undeclared identifier is reported only once mm/mmap.c:953: error: for each function it appears in.)

Cc Adam Litke <[email protected]>

Cc David Gibson <[email protected]>

Cc David Rientjes <[email protected]>

Cc Eric B Munson <[email protected]>

Cc Hugh Dickins <[email protected]>

Cc Lee Schermerhorn <[email protected]>

Cc Mel Gorman <[email protected]>

Cc Nick Piggin <[email protected]>

Cc Ivan Kokshaysky <[email protected]>

Cc Richard Henderson <[email protected]>

Signed-off-by Andrew Morton <[email protected]>

---

 arch/alpha/include/asm/mman.h |    1 +
 1 file changed, 1 insertion(+)

arch/alpha/include/asm/mman.h renamed from arch/alpha/include/asm/mman.h~hugetlb-add-map_hugetlb-for-mmaping-pseudo-anonymous-huge-page-regions-alpha-fix +1 -0

diff -puN arch/alpha/include/asm/mman.h~hugetlb-add-map_hugetlb-for-mmaping-pseudo-anonymous-huge-page-regions-alpha-fix arch/alpha/include/asm/mman.h
@@ -28,6 +28,7 @@
 #define MAP_NORESERVE	0x10000		/* don't check for reservations */
 #define MAP_POPULATE	0x20000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x40000		/* do not block on IO */
+#define MAP_HUGETLB	0x80000		/* create a huge page mapping */
 
 #define MS_ASYNC	1		/* sync memory asynchronously */
 #define MS_SYNC		2		/* synchronous memory sync */

Andrew Morton

maintainer

2009-09-18 00:46 UTC #5 lore raw

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

Quoted 2 lines · On Thu, 17 Sep 2009 15:44:04 -0700 Andrew Morton <[email protected]>…

mm/mmap.c: In function 'do_mmap_pgoff':
mm/mmap.c:953: error: 'MAP_HUGETLB' undeclared (first use in this function)

mips breaks as well.

I don't know how many other architectures broke. I disabled the patches.

Arnd Bergmann

maintainer

2009-09-18 15:19 UTC #6 lore raw

PATCH 1/1 Add MAP_HUGETLB for mmaping pseudo-anonymous huge page regions

The patch also adds the MAP_STACK flag, which was previously defined only on some architectures but not on others. Since MAP_STACK is meant to be a hint only, architectures can define it without assigning a specific meaning to it.

Signed-off-by Arnd Bergmann <[email protected]>

Cc Eric B Munson <[email protected]>

Cc Hugh Dickins <[email protected]>

Cc David Rientjes <[email protected]>

Cc Andrew Morton <[email protected]>

Cc [email protected]

---

Quoted 9 lines · Andrew Morton

Quoted 2 lines · On Thu, 17 Sep 2009 15:44:04 -0700 Andrew Morton <[email protected]>…
mm/mmap.c: In function 'do_mmap_pgoff':
mm/mmap.c:953: error: 'MAP_HUGETLB' undeclared (first use in this function)
mips breaks as well.
I don't know how many other architectures broke. I disabled the patches.

two: parisc and xtensa.

It's also done in a different way from the existing flags, which makes it more likely to break again if someone tries to add another flag.

This patch is more along the lines of how I explained it could be done, adding the MAP_HUGETLB (and MAP_STACK) flags to the existing lists in a consistent way.

A logical next step would be to replace the mman.h files that are basically copies of include/asm-generic/mman.h.

	Arnd <><
---
 arch/alpha/include/asm/mman.h   |    2 ++
 arch/arm/include/asm/mman.h     |    2 ++
 arch/avr32/include/asm/mman.h   |    2 ++
 arch/cris/include/asm/mman.h    |    2 ++
 arch/frv/include/asm/mman.h     |    2 ++
 arch/h8300/include/asm/mman.h   |    2 ++
 arch/ia64/include/asm/mman.h    |    2 ++
 arch/m32r/include/asm/mman.h    |    2 ++
 arch/m68k/include/asm/mman.h    |    2 ++
 arch/mips/include/asm/mman.h    |    2 ++
 arch/mn10300/include/asm/mman.h |    2 ++
 arch/parisc/include/asm/mman.h  |    2 ++
 arch/powerpc/include/asm/mman.h |    2 ++
 arch/s390/include/asm/mman.h    |    2 ++
 arch/sparc/include/asm/mman.h   |    2 ++
 arch/x86/include/asm/mman.h     |    1 +
 arch/xtensa/include/asm/mman.h  |    2 ++
 include/asm-generic/mman.h      |    1 +
 18 files changed, 34 insertions(+), 0 deletions(-)

arch/alpha/include/asm/mman.h +2 -0

index 90d7c35..6e4d854 100644
@@ -28,6 +28,8 @@
 #define MAP_NORESERVE	0x10000		/* don't check for reservations */
 #define MAP_POPULATE	0x20000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x40000		/* do not block on IO */
+#define MAP_STACK	0x80000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x100000	/* create a huge page mapping */
 
 #define MS_ASYNC	1		/* sync memory asynchronously */
 #define MS_SYNC		2		/* synchronous memory sync */

arch/arm/include/asm/mman.h +2 -0

index fc26976..6464d47 100644
@@ -10,6 +10,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) page tables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/avr32/include/asm/mman.h +2 -0

index 9a92b15..38cea1b 100644
@@ -10,6 +10,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) page tables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/cris/include/asm/mman.h +2 -0

index b7f0afb..de6b903 100644
@@ -12,6 +12,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/frv/include/asm/mman.h +2 -0

index 58c1d11..1939343 100644
@@ -10,6 +10,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/h8300/include/asm/mman.h +2 -0

index cf35f0a..eacacd0 100644
@@ -10,6 +10,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/ia64/include/asm/mman.h +2 -0

index 48cf8b9..cf55884 100644
@@ -18,6 +18,8 @@
 #define MAP_NORESERVE	0x04000		/* don't check for reservations */
 #define MAP_POPULATE	0x08000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/m32r/include/asm/mman.h +2 -0

index 04a5f40..d191089 100644
@@ -10,6 +10,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/m68k/include/asm/mman.h +2 -0

index 9f5c4c4..c421fef 100644
@@ -10,6 +10,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/mips/include/asm/mman.h +2 -0

index e4d6f1f..1578166 100644
@@ -46,6 +46,8 @@
 #define MAP_LOCKED	0x8000		/* pages are locked */
 #define MAP_POPULATE	0x10000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x20000		/* do not block on IO */
+#define MAP_STACK	0x40000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x80000		/* create a huge page mapping */
 
 /*
  * Flags for msync

arch/mn10300/include/asm/mman.h +2 -0

index d04fac1..94611c3 100644
@@ -21,6 +21,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/parisc/include/asm/mman.h +2 -0

index defe752..ce4b265 100644
@@ -22,6 +22,8 @@
 #define MAP_GROWSDOWN	0x8000		/* stack-like segment */
 #define MAP_POPULATE	0x10000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x20000		/* do not block on IO */
+#define MAP_STACK	0x40000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x80000		/* create a huge page mapping */
 
 #define MS_SYNC		1		/* synchronous memory sync */
 #define MS_ASYNC	2		/* sync memory asynchronously */

arch/powerpc/include/asm/mman.h +2 -0

index 7b1c498..d4a7f64 100644
@@ -25,6 +25,8 @@
 
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #ifdef __KERNEL__
 #ifdef CONFIG_PPC64

arch/s390/include/asm/mman.h +2 -0

index f63fe7b..22714ca 100644
@@ -18,6 +18,8 @@
 #define MAP_NORESERVE	0x4000		/* don't check for reservations */
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/sparc/include/asm/mman.h +2 -0

index 988192e..c3029ad 100644
@@ -20,6 +20,8 @@
 
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
+#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #ifdef __KERNEL__
 #ifndef __ASSEMBLY__

arch/x86/include/asm/mman.h +1 -0

index 751af25..c719f36 100644
@@ -13,6 +13,7 @@
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
 #define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

arch/xtensa/include/asm/mman.h +2 -0

index 9b92620..f380d04 100644
@@ -53,6 +53,8 @@
 #define MAP_LOCKED	0x8000		/* pages are locked */
 #define MAP_POPULATE	0x10000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x20000		/* do not block on IO */
+#define MAP_STACK	0x40000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x80000		/* create a huge page mapping */
 
 /*
  * Flags for msync

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

index 7cab4de..32c8bd6 100644
@@ -11,6 +11,7 @@
 #define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
 #define MAP_NONBLOCK	0x10000		/* do not block on IO */
 #define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
+#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
 
 #define MCL_CURRENT	1		/* lock all current mappings */
 #define MCL_FUTURE	2		/* lock all future mappings */

Arnd Bergmann

maintainer

2009-09-18 16:48 UTC #7 lore raw

PATCH 1/1 remove duplicate asm/mman.h files

A number of architectures have identical asm/mman.h files, x86 differs only in a single line, so they can all be merged by using the new generic file.

The remaining asm/mman.h files are substantially different from each other.

Signed-off-by Arnd Bergmann <[email protected]>

---

Quoted 2 lines · Arnd Bergmann

A logical next step would be to replace the mman.h files that are basically copies of include/asm-generic/mman.h.

We might as well do it now, since we're touching them anyway... Patch applies on top of the previous one.

 arch/arm/include/asm/mman.h     |   20 +-------------------
 arch/avr32/include/asm/mman.h   |   20 +-------------------
 arch/cris/include/asm/mman.h    |   22 +---------------------
 arch/frv/include/asm/mman.h     |   21 +--------------------
 arch/h8300/include/asm/mman.h   |   20 +-------------------
 arch/ia64/include/asm/mman.h    |   16 ++--------------
 arch/m32r/include/asm/mman.h    |   20 +-------------------
 arch/m68k/include/asm/mman.h    |   20 +-------------------
 arch/mn10300/include/asm/mman.h |   31 +------------------------------
 arch/s390/include/asm/mman.h    |   15 +--------------
 arch/x86/include/asm/mman.h     |   15 +--------------
 11 files changed, 12 insertions(+), 208 deletions(-)

arch/arm/include/asm/mman.h +1 ~~-19~~

index 6464d47..8eebf89 100644
@@ -1,19 +1 @@
-#ifndef __ARM_MMAN_H__
-#define __ARM_MMAN_H__
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) page tables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __ARM_MMAN_H__ */
+#include <asm-generic/mman.h>

arch/avr32/include/asm/mman.h +1 ~~-19~~

index 38cea1b..8eebf89 100644
@@ -1,19 +1 @@
-#ifndef __ASM_AVR32_MMAN_H__
-#define __ASM_AVR32_MMAN_H__
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) page tables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __ASM_AVR32_MMAN_H__ */
+#include <asm-generic/mman.h>

arch/cris/include/asm/mman.h +1 ~~-21~~

index de6b903..8eebf89 100644
@@ -1,21 +1 @@
-#ifndef __CRIS_MMAN_H__
-#define __CRIS_MMAN_H__
-
-/* verbatim copy of asm-i386/ version */
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __CRIS_MMAN_H__ */
+#include <asm-generic/mman.h>

arch/frv/include/asm/mman.h +1 ~~-20~~

index 1939343..8eebf89 100644
@@ -1,20 +1 @@
-#ifndef __ASM_MMAN_H__
-#define __ASM_MMAN_H__
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __ASM_MMAN_H__ */
-
+#include <asm-generic/mman.h>

arch/h8300/include/asm/mman.h +1 ~~-19~~

index eacacd0..8eebf89 100644
@@ -1,19 +1 @@
-#ifndef __H8300_MMAN_H__
-#define __H8300_MMAN_H__
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __H8300_MMAN_H__ */
+#include <asm-generic/mman.h>

arch/ia64/include/asm/mman.h +2 ~~-14~~

index cf55884..4459028 100644
@@ -8,21 +8,9 @@
  *	David Mosberger-Tang <[email protected]>, Hewlett-Packard Co
  */
 
-#include <asm-generic/mman-common.h>
+#include <asm-generic/mman.h>
 
-#define MAP_GROWSDOWN	0x00100		/* stack-like segment */
-#define MAP_GROWSUP	0x00200		/* register stack-like segment */
-#define MAP_DENYWRITE	0x00800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x01000		/* mark it as an executable */
-#define MAP_LOCKED	0x02000		/* pages are locked */
-#define MAP_NORESERVE	0x04000		/* don't check for reservations */
-#define MAP_POPULATE	0x08000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
+#define MAP_GROWSUP	0x0200		/* register stack-like segment */
 
 #ifdef __KERNEL__
 #ifndef __ASSEMBLY__

arch/m32r/include/asm/mman.h +1 ~~-19~~

index d191089..8eebf89 100644
@@ -1,19 +1 @@
-#ifndef __M32R_MMAN_H__
-#define __M32R_MMAN_H__
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __M32R_MMAN_H__ */
+#include <asm-generic/mman.h>

arch/m68k/include/asm/mman.h +1 ~~-19~~

index c421fef..8eebf89 100644
@@ -1,19 +1 @@
-#ifndef __M68K_MMAN_H__
-#define __M68K_MMAN_H__
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* __M68K_MMAN_H__ */
+#include <asm-generic/mman.h>

arch/mn10300/include/asm/mman.h +1 ~~-30~~

index 94611c3..8eebf89 100644
@@ -1,30 +1 @@
-/* MN10300 Constants for mmap and co.
- *
- * Copyright (C) 2007 Matsushita Electric Industrial Co., Ltd.
- * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
- * - Derived from asm-x86/mman.h
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public Licence
- * as published by the Free Software Foundation; either version
- * 2 of the Licence, or (at your option) any later version.
- */
-#ifndef _ASM_MMAN_H
-#define _ASM_MMAN_H
-
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
-
-#endif /* _ASM_MMAN_H */
+#include <asm-generic/mman.h>

arch/s390/include/asm/mman.h +1 ~~-14~~

index 22714ca..4e9c8ae 100644
@@ -9,20 +9,7 @@
 #ifndef __S390_MMAN_H__
 #define __S390_MMAN_H__
 
-#include <asm-generic/mman-common.h>
-
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
+#include <asm-generic/mman.h>
 
 #if defined(__KERNEL__) && !defined(__ASSEMBLY__) && defined(CONFIG_64BIT)
 int s390_mmap_check(unsigned long addr, unsigned long len);

arch/x86/include/asm/mman.h +1 ~~-13~~

index c719f36..c582add 100644
@@ -1,21 +1,8 @@
 #ifndef _ASM_X86_MMAN_H
 #define _ASM_X86_MMAN_H
 
-#include <asm-generic/mman-common.h>
+#include <asm-generic/mman.h>
 
 #define MAP_32BIT	0x40		/* only give out 32bit addresses */
 
-#define MAP_GROWSDOWN	0x0100		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-#define MAP_LOCKED	0x2000		/* pages are locked */
-#define MAP_NORESERVE	0x4000		/* don't check for reservations */
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */

David Rientjes

2009-09-18 19:37 UTC #8 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

Quoted 18 lines · Arnd Bergmann

arch/ia64/include/asm/mman.h +1 -8

index cf55884..4459028 100644
@@ -8,21 +8,9 @@
  *	David Mosberger-Tang <[email protected]>, Hewlett-Packard Co
  */
 
-#include <asm-generic/mman-common.h>
+#include <asm-generic/mman.h>
 
-#define MAP_GROWSDOWN	0x00100		/* stack-like segment */
-#define MAP_GROWSUP	0x00200		/* register stack-like segment */
-#define MAP_DENYWRITE	0x00800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x01000		/* mark it as an executable */
-#define MAP_LOCKED	0x02000		/* pages are locked */
-#define MAP_NORESERVE	0x04000		/* don't check for reservations */
-#define MAP_POPULATE	0x08000		/* populate (prefault) pagetables */

arch/ia64/include/asm/mman.h line 16

Quoted 10 lines

arch/ia64/include/asm/mman.h +1 -6

-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-#define MAP_STACK	0x20000		/* give out an address that is best suited for process/thread stacks */
-#define MAP_HUGETLB	0x40000		/* create a huge page mapping */
-
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
+#define MAP_GROWSUP	0x0200		/* register stack-like segment */
 
 #ifdef __KERNEL__
 #ifndef __ASSEMBLY__

ia64 doesn't use MAP_GROWSUP, so it's probably not necessary to carry it along with your cleanup.
--
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

Arnd Bergmann

maintainer

2009-09-21 08:31 UTC #9 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

Quoted 11 lines · David Rientjes

arch/ia64/include/asm/mman.h line 16
Quoted 6 lines · Arnd Bergmann
arch/ia64/include/asm/mman.h +1 -2
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
+#define MAP_GROWSUP	0x0200		/* register stack-like segment */
 
 #ifdef __KERNEL__
 #ifndef __ASSEMBLY__
ia64 doesn't use MAP_GROWSUP, so it's probably not necessary to carry it along with your cleanup.

ia64 is the only architecture defining it, nobody uses it in the kernel. If the ia64 maintainers want to remove it in a separate patch, that would probably be a good idea.

I tried not to change the ABI in any way in my patch, and there is a theoretical possibility that some user space program on ia64 currently depends on that definition.

	Arnd <><

David Rientjes

2009-09-21 09:13 UTC #10 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

Quoted 10 lines · Arnd Bergmann

Quoted 9 lines
Review
Quoted 6 lines
(quoted diff) +1 -2
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
+#define MAP_GROWSUP	0x0200		/* register stack-like segment */
 
 #ifdef __KERNEL__
 #ifndef __ASSEMBLY__
ia64 doesn't use MAP_GROWSUP, so it's probably not necessary to carry it along with your cleanup.

Quoted 4 lines

ia64 is the only architecture defining it, nobody uses it in the kernel. If the ia64 maintainers want to remove it in a separate patch, that would probably be a good idea.

I'll do it then.

Quoted 4 lines

I tried not to change the ABI in any way in my patch, and there is a theoretical possibility that some user space program on ia64 currently depends on that definition.

I don't buy that as justification, if some userspace program uses it based on the false belief that it actually does what it says, it's probably better to break their build than perpetuating the lie that it's different than ~MAP_GROWSDOWN.

ia64: remove definition for MAP_GROWSUP

MAP_GROWSUP is unused.

Signed-off-by David Rientjes <[email protected]>

---

arch/ia64/include/asm/mman.h +0 -1

@@ -11,7 +11,6 @@
 #include <asm-generic/mman-common.h>
 
 #define MAP_GROWSDOWN	0x00100		/* stack-like segment */
-#define MAP_GROWSUP	0x00200		/* register stack-like segment */
 #define MAP_DENYWRITE	0x00800		/* ETXTBSY */
 #define MAP_EXECUTABLE	0x01000		/* mark it as an executable */
 #define MAP_LOCKED	0x02000		/* pages are locked */

Arnd Bergmann

maintainer Acked-by

2009-09-21 09:30 UTC #11 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

Quoted 9 lines · David Rientjes

Quoted 4 lines
I tried not to change the ABI in any way in my patch, and there is a theoretical possibility that some user space program on ia64 currently depends on that definition.
I don't buy that as justification, if some userspace program uses it based on the false belief that it actually does what it says, it's probably better to break their build than perpetuating the lie that it's different than ~MAP_GROWSDOWN.

It's more a matter of principle of my patches. I try to strictly separate patches that move code around (like the one I sent) from those that change contents (like yours, or the one before that adds MAP_STACK and MAP_HUGETLB).

Removing a definition from an exported header file either requires specific knowledge about why it is there to start with, or more research on the topic than I wanted to do. For instance, a theoretical program might have a helper function correctly doing

void *xmmap(void *addr, size_t length, int prot, int flags,
int fd, off_t offset)

{
	if (flags & MAP_GROWSUP) { /* MAP_GROWSUP is not supported */
		errno = -EINVAL;
		return MAP_FAILED;
	}

	return mmap(addr, length, prot, flags, fd, offset);
}

Of course, such a program would only work on ia64 currently, so it should be safe to make ia64 behave like the other architectures in this regard.

Quoted 5 lines

ia64: remove definition for MAP_GROWSUP
MAP_GROWSUP is unused.
Signed-off-by David Rientjes <[email protected]>

Acked-by Arnd Bergmann <[email protected]>

Hugh Dickins

2009-09-21 12:02 UTC #12 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

In reply to David Rientjes

Quoted 30 lines · David Rientjes

Quoted 10 lines · Arnd Bergmann
Quoted 9 lines
Review
Quoted 6 lines
(quoted diff) +1 -2
-#define MCL_CURRENT	1		/* lock all current mappings */
-#define MCL_FUTURE	2		/* lock all future mappings */
+#define MAP_GROWSUP	0x0200		/* register stack-like segment */
 
 #ifdef __KERNEL__
 #ifndef __ASSEMBLY__
ia64 doesn't use MAP_GROWSUP, so it's probably not necessary to carry it along with your cleanup.
Quoted 4 lines
ia64 is the only architecture defining it, nobody uses it in the kernel. If the ia64 maintainers want to remove it in a separate patch, that would probably be a good idea.
I'll do it then.
Quoted 4 lines
I tried not to change the ABI in any way in my patch, and there is a theoretical possibility that some user space program on ia64 currently depends on that definition.
I don't buy that as justification, if some userspace program uses it based on the false belief that it actually does what it says, it's probably better to break their build than perpetuating the lie that it's different than ~MAP_GROWSDOWN.

Quoted 3 lines

ia64: remove definition for MAP_GROWSUP
MAP_GROWSUP is unused.

Is it perhaps the case that some UNIX on ia64 does implement MAP_GROWSUP, and these numbers in the Linux ia64 mman.h have been chosen to match that reference implementation? Tony will know. But I wonder if you'd do better at least to leave a MAP_GROWSUP comment on that line, so that somebody doesn't go and reuse the empty slot later on.

Hugh

Quoted 14 lines

Signed-off-by David Rientjes <[email protected]>

---

arch/ia64/include/asm/mman.h +0 -1

@@ -11,7 +11,6 @@
 #include <asm-generic/mman-common.h>
 
 #define MAP_GROWSDOWN	0x00100		/* stack-like segment */
-#define MAP_GROWSUP	0x00200		/* register stack-like segment */
 #define MAP_DENYWRITE	0x00800		/* ETXTBSY */
 #define MAP_EXECUTABLE	0x01000		/* mark it as an executable */
 #define MAP_LOCKED	0x02000		/* pages are locked */

David Rientjes

2009-09-21 22:55 UTC #13 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

Quoted 6 lines · Hugh Dickins

Is it perhaps the case that some UNIX on ia64 does implement MAP_GROWSUP, and these numbers in the Linux ia64 mman.h have been chosen to match that reference implementation? Tony will know. But I wonder if you'd do better at least to leave a MAP_GROWSUP comment on that line, so that somebody doesn't go and reuse the empty slot later on.

Reserving the bit from future use by adding a comment may be helpful, but then let's do it for MAP_GROWSDOWN too.

Luck, Tony

2009-09-21 23:24 UTC #14 lore raw

RE: [PATCH] remove duplicate asm/mman.h files

Quoted 9 lines

Quoted 6 lines
Is it perhaps the case that some UNIX on ia64 does implement MAP_GROWSUP, and these numbers in the Linux ia64 mman.h have been chosen to match that reference implementation? Tony will know. But I wonder if you'd do better at least to leave a MAP_GROWSUP comment on that line, so that somebody doesn't go and reuse the empty slot later on.
Reserving the bit from future use by adding a comment may be helpful, but then let's do it for MAP_GROWSDOWN too.

Tony can only speculate because this bit has been in asm/mman.h since before I started working on Linux (it is in the 2.4.0 version ... which is roughly when I started ... and long before I was responsible for it).

Perhaps it was assumed that it would be useful? Linux/ia64 does use upwardly growing memory areas (the h/w register stack engine saves "stack" registers to an area that grows upwards).

But since we have survived this long without it actually being implemented, it may be true that we don't really need it after all.

-Tony

David Rientjes

2009-09-21 23:46 UTC #15 lore raw

RE: [PATCH] remove duplicate asm/mman.h files

Quoted 15 lines · Tony

Quoted 9 lines
Quoted 6 lines
Is it perhaps the case that some UNIX on ia64 does implement MAP_GROWSUP, and these numbers in the Linux ia64 mman.h have been chosen to match that reference implementation? Tony will know. But I wonder if you'd do better at least to leave a MAP_GROWSUP comment on that line, so that somebody doesn't go and reuse the empty slot later on.
Reserving the bit from future use by adding a comment may be helpful, but then let's do it for MAP_GROWSDOWN too.
Tony can only speculate because this bit has been in asm/mman.h since before I started working on Linux (it is in the 2.4.0 version ... which is roughly when I started ... and long before I was responsible for it).

Quoted 8 lines

Perhaps it was assumed that it would be useful? Linux/ia64 does use upwardly growing memory areas (the h/w register stack engine saves "stack" registers to an area that grows upwards).
But since we have survived this long without it actually being implemented, it may be true that we don't really need it after all.

glibc notes that both MAP_GROWSUP and MAP_GROWSDOWN are specific to Linux, yet they don't functionally do anything. While it may be true that there's no cost associated with keeping them around, I also think exporting such flags to userspace may give developers the belief that the implementation actually respects them when they're passed.

Ulrich wanted to do this last year but it appears to have been dropped.

Unless there's a convincing argument in the other direction, I don't see why they both can't just be removed and their bits reserved.

Ulrich Drepper

2009-09-21 23:58 UTC #16 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 1 line · David Rientjes

Ulrich wanted to do this last year but it appears to have been dropped.

I've mentioned that at that time, these flags cannot be used at all for stacks. And I don't know for what else it is useful. They should really be removed.

- --
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖ -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAkq4Ez0ACgkQ2ijCOnn/RHQNtACfX+y5pIQhDusikKiQwQ8nvGRN cI8An0oThAXSwXRALt9598vbPbiVwEeJ
=sxfU
-----END PGP SIGNATURE-----

Eric B Munson

author Acked-by on v1 1/1

2009-09-21 12:27 UTC #17 lore raw

Re: [PATCH] remove duplicate asm/mman.h files

In reply to Arnd Bergmann

Quoted 8 lines · Arnd Bergmann

A number of architectures have identical asm/mman.h files, x86 differs only in a single line, so they can all be merged by using the new generic file.
The remaining asm/mman.h files are substantially different from each other.
Signed-off-by Arnd Bergmann <[email protected]>

Acked-by Eric B Munson <[email protected]>

Eric B Munson

author Acked-by on v1 1/1

2009-09-21 12:25 UTC #18 lore raw

Re: [PATCH] Add MAP_HUGETLB for mmaping pseudo-anonymous huge page regions

In reply to Arnd Bergmann

Quoted 6 lines · Arnd Bergmann

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.

Quoted 6 lines

The patch also adds the MAP_STACK flag, which was previously defined only on some architectures but not on others. Since MAP_STACK is meant to be a hint only, architectures can define it without assigning a specific meaning to it.
Signed-off-by Arnd Bergmann <[email protected]>

I think this is a more sane way of handling the mman flags.

Acked-by Eric B Munson <[email protected]>

David Rientjes

Acked-by on v1 1/3

2009-08-26 19:34 UTC #19 lore raw

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

In reply to Eric B Munson

Quoted 6 lines · Eric B Munson

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

Quoted 7 lines

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]>

Acked-by David Rientjes <[email protected]>