Tracking user space vDSO remaping

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +1 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;

mm/mremap.c +1 -0

+#endif
 	}
 
 	/* Conceal VM_ACCOUNT so old reservation is not undone */

Signature · 7 lines

1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 301

Quoted 4 lines

mm/mremap.c +2 -0

+		arch_remap(mm, old_addr, old_addr+old_len,
+			   new_addr, new_addr+new_len);

Spaces around '+'?

Quoted 8 lines

mm/mremap.c +1 -0

+#endif
 	}
 
 	/* Conceal VM_ACCOUNT so old reservation is not undone */

Signature · 1 line

1.9.1

Quoted 5 lines

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 48 lines · Kirill A. Shutemov

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

Quoted 8 lines

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

The idea was to not impact all the architectures as arch_unmap(), arch_dup_mmap() or arch_exit_mmap() implies.

I look at the headers where such a dummy arch_remap could be put but I can't figure out one which will not impact all the architecture. What about defining a dummy service earlier in mm/remap.c in the case __HAVE_ARCH_REMAP is not defined ?
Something like :
#ifndef __HAVE_ARCH_REMAP
static inline void void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end)
{
}
#endif

Quoted 61 lines · Laurent Dufour

Quoted 48 lines · Kirill A. Shutemov

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

Quoted 8 lines

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

The idea was to not impact all the architectures as arch_unmap(), arch_dup_mmap() or arch_exit_mmap() implies.

Quoted 14 lines

I look at the headers where such a dummy arch_remap could be put but I can't figure out one which will not impact all the architecture. What about defining a dummy service earlier in mm/remap.c in the case __HAVE_ARCH_REMAP is not defined ?
Something like :
#ifndef __HAVE_ARCH_REMAP
static inline void void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end)
{
}
#endif

Or just #define arch_remap(...) do { } while (0)

Quoted 77 lines · Kirill A. Shutemov

Quoted 61 lines · Laurent Dufour

Quoted 48 lines · Kirill A. Shutemov

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

Quoted 8 lines

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

The idea was to not impact all the architectures as arch_unmap(), arch_dup_mmap() or arch_exit_mmap() implies.

Quoted 14 lines

I look at the headers where such a dummy arch_remap could be put but I can't figure out one which will not impact all the architecture. What about defining a dummy service earlier in mm/remap.c in the case __HAVE_ARCH_REMAP is not defined ?
Something like :
#ifndef __HAVE_ARCH_REMAP
static inline void void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end)
{
}
#endif

Quoted 2 lines

Or just #define arch_remap(...) do { } while (0)

I guessed you wanted the arch_remap() prototype to be exposed somewhere in the code.

To be honest, I can't find the benefit of defining a dummy arch_remap() in mm/remap.c if __HAVE_ARCH_REMAP is not defined instead of calling it in move_vma if __HAVE_ARCH_REMAP is defined.
Is it really what you want ?

Quoted 81 lines · Laurent Dufour

Quoted 77 lines · Kirill A. Shutemov

Quoted 61 lines · Laurent Dufour

Quoted 48 lines · Kirill A. Shutemov

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

Quoted 8 lines

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

The idea was to not impact all the architectures as arch_unmap(), arch_dup_mmap() or arch_exit_mmap() implies.

Quoted 14 lines

I look at the headers where such a dummy arch_remap could be put but I can't figure out one which will not impact all the architecture. What about defining a dummy service earlier in mm/remap.c in the case __HAVE_ARCH_REMAP is not defined ?
Something like :
#ifndef __HAVE_ARCH_REMAP
static inline void void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end)
{
}
#endif

Quoted 2 lines

Or just #define arch_remap(...) do { } while (0)

Quoted 7 lines

I guessed you wanted the arch_remap() prototype to be exposed somewhere in the code.

To be honest, I can't find the benefit of defining a dummy arch_remap() in mm/remap.c if __HAVE_ARCH_REMAP is not defined instead of calling it in move_vma if __HAVE_ARCH_REMAP is defined.
Is it really what you want ?

I think Kirill meant something like e.g. the arch_enter_lazy_mmu_mode() is implemented and called in mm/mremap.c -- the "generic" part is in the include/asm-generic/pgtable.h and those architectures willing to have their own implementation are in arch/$arch/...

Kirill, if I'm right with it, can you suggest the header where to put the "generic" mremap hook's (empty) body?

Quoted 90 lines · Pavel Emelyanov

Quoted 81 lines · Laurent Dufour

Quoted 77 lines · Kirill A. Shutemov

Quoted 61 lines · Laurent Dufour

Quoted 48 lines · Kirill A. Shutemov

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

Quoted 8 lines

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

The idea was to not impact all the architectures as arch_unmap(), arch_dup_mmap() or arch_exit_mmap() implies.

Quoted 14 lines

I look at the headers where such a dummy arch_remap could be put but I can't figure out one which will not impact all the architecture. What about defining a dummy service earlier in mm/remap.c in the case __HAVE_ARCH_REMAP is not defined ?
Something like :
#ifndef __HAVE_ARCH_REMAP
static inline void void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end)
{
}
#endif

Quoted 2 lines

Or just #define arch_remap(...) do { } while (0)

Quoted 7 lines

I guessed you wanted the arch_remap() prototype to be exposed somewhere in the code.

To be honest, I can't find the benefit of defining a dummy arch_remap() in mm/remap.c if __HAVE_ARCH_REMAP is not defined instead of calling it in move_vma if __HAVE_ARCH_REMAP is defined.
Is it really what you want ?

Quoted 7 lines

I think Kirill meant something like e.g. the arch_enter_lazy_mmu_mode() is implemented and called in mm/mremap.c -- the "generic" part is in the include/asm-generic/pgtable.h and those architectures willing to have their own implementation are in arch/$arch/...

Kirill, if I'm right with it, can you suggest the header where to put the "generic" mremap hook's (empty) body?

I initially thought it would be enough to put it into <asm-generic/mmu_context.h>, expecting it works as <asm-generic/pgtable.h>. But that's not the case.

It probably worth at some point rework all <asm/mmu_context.h> to include <asm-generic/mmu_context.h> at the end as we do for <asm/pgtable.h>. But that's outside the scope of the patchset, I guess.

I don't see any better candidate for such dummy header. :-/

Quoted 107 lines · Kirill A. Shutemov

Quoted 90 lines · Pavel Emelyanov

Quoted 81 lines · Laurent Dufour

Quoted 77 lines · Kirill A. Shutemov

Quoted 61 lines · Laurent Dufour

Quoted 48 lines · Kirill A. Shutemov

Quoted 47 lines · Laurent Dufour

Some architecture would like to be triggered when a memory area is moved through the mremap system call.

This patch is introducing a new arch_remap mm hook which is placed in the path of mremap, and is called before the old area is unmapped (and the arch_unmap hook is called).

The architectures which need to call this hook should define __HAVE_ARCH_REMAP in their asm/mmu_context.h and provide the arch_remap service with the following prototype:
void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end);

Signed-off-by Laurent Dufour <[email protected]>

Reviewed-by Ingo Molnar <[email protected]>

---
 mm/mremap.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

mm/mremap.c +6 -6

index 2dc44b1cb1df..009db5565893 100644
@@ -25,6 +25,7 @@
 
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
 
 #include "internal.h"
 

@@ -286,13 +287,19 @@ static unsigned long move_vma(struct vm_area_struct *vma,
 		old_len = new_len;
 		old_addr = new_addr;
 		new_addr = -ENOMEM;
-	} else if (vma->vm_file && vma->vm_file->f_op->mremap) {
-		err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
-		if (err < 0) {
-			move_page_tables(new_vma, new_addr, vma, old_addr,
-					 moved_len, true);
-			return err;
+	} else {
+		if (vma->vm_file && vma->vm_file->f_op->mremap) {
+			err = vma->vm_file->f_op->mremap(vma->vm_file, new_vma);
+			if (err < 0) {
+				move_page_tables(new_vma, new_addr, vma,

Quoted 8 lines

mm/mremap.c static unsigned long move_vma(struct vm_area_struct *vma, · line 299

Quoted 5 lines

mm/mremap.c +4 -0

+						  old_addr, moved_len, true);
+				return err;
+			}
 		}
+#ifdef __HAVE_ARCH_REMAP

It would be cleaner to provide dummy arch_remap() for !__HAVE_ARCH_REMAP in some generic header.

The idea was to not impact all the architectures as arch_unmap(), arch_dup_mmap() or arch_exit_mmap() implies.

Quoted 14 lines

I look at the headers where such a dummy arch_remap could be put but I can't figure out one which will not impact all the architecture. What about defining a dummy service earlier in mm/remap.c in the case __HAVE_ARCH_REMAP is not defined ?
Something like :
#ifndef __HAVE_ARCH_REMAP
static inline void void arch_remap(struct mm_struct *mm,
unsigned long old_start, unsigned long old_end, unsigned long new_start, unsigned long new_end)
{
}
#endif

Quoted 2 lines

Or just #define arch_remap(...) do { } while (0)

Quoted 7 lines

I guessed you wanted the arch_remap() prototype to be exposed somewhere in the code.

To be honest, I can't find the benefit of defining a dummy arch_remap() in mm/remap.c if __HAVE_ARCH_REMAP is not defined instead of calling it in move_vma if __HAVE_ARCH_REMAP is defined.
Is it really what you want ?

Quoted 7 lines

I think Kirill meant something like e.g. the arch_enter_lazy_mmu_mode() is implemented and called in mm/mremap.c -- the "generic" part is in the include/asm-generic/pgtable.h and those architectures willing to have their own implementation are in arch/$arch/...

Kirill, if I'm right with it, can you suggest the header where to put the "generic" mremap hook's (empty) body?

I initially thought it would be enough to put it into <asm-generic/mmu_context.h>, expecting it works as <asm-generic/pgtable.h>. But that's not the case.

It probably worth at some point rework all <asm/mmu_context.h> to include <asm-generic/mmu_context.h> at the end as we do for <asm/pgtable.h>. But that's outside the scope of the patchset, I guess.

Quoted 1 line

I don't see any better candidate for such dummy header. :-/

Clearly, I'm not confortable with a rewrite of <asm/mmu_context.h> :(

So what about this patch, is this v3 acceptable ?

Quoted 8 lines

Quoted 8 lines

I initially thought it would be enough to put it into <asm-generic/mmu_context.h>, expecting it works as <asm-generic/pgtable.h>. But that's not the case.

It probably worth at some point rework all <asm/mmu_context.h> to include <asm-generic/mmu_context.h> at the end as we do for <asm/pgtable.h>. But that's outside the scope of the patchset, I guess.

Quoted 5 lines

Quoted 1 line

I don't see any better candidate for such dummy header. :-/

Clearly, I'm not confortable with a rewrite of <asm/mmu_context.h> :(

So what about this patch, is this v3 acceptable ?

Acked-by Kirill A. Shutemov <[email protected]>

Quoted 12 lines

Quoted 12 lines

Quoted 9 lines

I initially thought it would be enough to put it into <asm-generic/mmu_context.h>, expecting it works as <asm-generic/pgtable.h>. But that's not the case.

It probably worth at some point rework all <asm/mmu_context.h> to include <asm-generic/mmu_context.h> at the end as we do for <asm/pgtable.h>. But that's outside the scope of the patchset, I guess.

I don't see any better candidate for such dummy header. :-/

Clearly, I'm not confortable with a rewrite of <asm/mmu_context.h> :(

Quoted 3 lines

Quoted 1 line

So what about this patch, is this v3 acceptable ?

Acked-by Kirill A. Shutemov <[email protected]>

Other than the #ifdef thing, the same:

Acked-by Pavel Emelyanov <[email protected]>

Quoted 2 lines

Kirill, if I'm right with it, can you suggest the header where to put the "generic" mremap hook's (empty) body?

I initially thought it would be enough to put it into <asm-generic/mmu_context.h>, expecting it works as <asm-generic/pgtable.h>. But that's not the case.

It probably worth at some point rework all <asm/mmu_context.h> to include <asm-generic/mmu_context.h> at the end as we do for <asm/pgtable.h>. But that's outside the scope of the patchset, I guess.

I don't see any better candidate for such dummy header. :-/

Quoted 7 lines · On Mon, 13 Apr 2015 17:02:19 +0300 "Kirill A. Shutemov" <[email protected]>…

Quoted 2 lines

Kirill, if I'm right with it, can you suggest the header where to put the "generic" mremap hook's (empty) body?

I initially thought it would be enough to put it into <asm-generic/mmu_context.h>, expecting it works as <asm-generic/pgtable.h>. But that's not the case.

Quoted 5 lines

It probably worth at some point rework all <asm/mmu_context.h> to include <asm-generic/mmu_context.h> at the end as we do for <asm/pgtable.h>. But that's outside the scope of the patchset, I guess.

I don't see any better candidate for such dummy header. :-/

Do away with __HAVE_ARCH_REMAP and do it like this:

arch/x/include/asm/y.h:

	extern void arch_remap(...);
	#define arch_remap arch_remap

include/linux/z.h:

	#include <asm/y.h>

	#ifndef arch_remap
	static inline void arch_remap(...) { }
	#define arch_remap arch_remap
	#endif

Quoted 7 lines

Do away with __HAVE_ARCH_REMAP and do it like this:

arch/x/include/asm/y.h:

	extern void arch_remap(...);
	#define arch_remap arch_remap

Quoted 8 lines

include/linux/z.h:

	#include <asm/y.h>

	#ifndef arch_remap
	static inline void arch_remap(...) { }
	#define arch_remap arch_remap
	#endif

Hi Andrew,

I like your idea, but I can't find any good candidate for <asm/y.h> and <linux/z.h>.

I tried with <linux/mm.h> and <asm/mmu_context.h> but <asm/mmu_context.h> is already including <linux/mm.h>.

Do you have any suggestion ?

Another option could be to do it like the actual arch_unmap() in <asm-generic/mm_hooks.h> but this is the opposite of your idea, and Ingo was not comfortable with this idea due to the impact of the other architectures.

Quoted 24 lines · On Tue, 14 Apr 2015 11:26:13 +0200 Laurent Dufour <[email protected]>…

Quoted 7 lines

Do away with __HAVE_ARCH_REMAP and do it like this:

arch/x/include/asm/y.h:

	extern void arch_remap(...);
	#define arch_remap arch_remap

Quoted 8 lines

include/linux/z.h:

	#include <asm/y.h>

	#ifndef arch_remap
	static inline void arch_remap(...) { }
	#define arch_remap arch_remap
	#endif

Hi Andrew,

I like your idea, but I can't find any good candidate for <asm/y.h> and <linux/z.h>.

I tried with <linux/mm.h> and <asm/mmu_context.h> but <asm/mmu_context.h> is already including <linux/mm.h>.

Quoted 6 lines

Do you have any suggestion ?

Another option could be to do it like the actual arch_unmap() in <asm-generic/mm_hooks.h> but this is the opposite of your idea, and Ingo was not comfortable with this idea due to the impact of the other architectures.

I don't see any appropriate header files for this. mman.h is kinda close.

So we create new header files, that's not a problem. I'm torn between

a) include/linux/mm-arch-hooks.h (and 31 arch/X/include/asm/mm-arch-hooks.h). Mandate: mm stuff which can be overridded by arch

versus

b) include/linux/mremap.h (+31), with a narrower mandate.

This comes up fairly regularly so I suspect a) is better. We'll add things to it over time, and various bits of existing ad-hackery can be moved over as cleanups.