From: Zhang Yi <yi.z.zhang@linux.intel.com>

When a file supporting DAX is used as vNVDIMM backend, mmap it with
MAP_SYNC flag in addition which can ensure file system metadata
synced in each guest writes to the backend file, without other QEMU
actions (e.g., periodic fsync() by QEMU).

Current, We have below different possible use cases:

1. pmem=on is set, shared=on is set, MAP_SYNC supported:
   a: backend is a dax supporting file.
    - MAP_SYNC will active.
   b: backend is not a dax supporting file.
    - mmap will trigger a warning. then MAP_SYNC flag will be ignored

2. The rest of cases:
   - we will never pass the MAP_SYNC to mmap2

Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
Signed-off-by: Zhang Yi <yi.z.zhang@linux.intel.com>
---
 include/qemu/osdep.h |  7 +++++++
 util/mmap-alloc.c    | 24 +++++++++++++++++++++++-
 2 files changed, 30 insertions(+), 1 deletion(-)

diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
index 457d24e..9a94cc3 100644
--- a/include/qemu/osdep.h
+++ b/include/qemu/osdep.h
@@ -115,6 +115,13 @@ extern int daemon(int, int);
 #include "sysemu/os-win32.h"
 #endif
 
+#ifdef CONFIG_LINUX
+#include <linux/mman.h>
+#else  /* !CONFIG_LINUX */
+#define MAP_SYNC              0x0
+#define MAP_SHARED_VALIDATE   0x0
+#endif /* CONFIG_LINUX */
+
 #ifdef CONFIG_POSIX
 #include "sysemu/os-posix.h"
 #endif
diff --git a/util/mmap-alloc.c b/util/mmap-alloc.c
index 97bbeed..e4e55fc 100644
--- a/util/mmap-alloc.c
+++ b/util/mmap-alloc.c
@@ -15,6 +15,7 @@
 #include "qemu/host-utils.h"
 
 #define HUGETLBFS_MAGIC       0x958458f6
+#define MAP_SYNC_FLAGS        (MAP_SYNC | MAP_SHARED_VALIDATE)
 
 #ifdef CONFIG_LINUX
 #include <sys/vfs.h>
@@ -101,6 +102,7 @@ void *qemu_ram_mmap(int fd,
 #else
     void *ptr = mmap(0, total, PROT_NONE, MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
 #endif
+    int mmap_flags;
     size_t offset;
     void *ptr1;
 
@@ -111,13 +113,33 @@ void *qemu_ram_mmap(int fd,
     assert(is_power_of_2(align));
     /* Always align to host page size */
     assert(align >= getpagesize());
+    mmap_flags = shared ? MAP_SHARED : MAP_PRIVATE;
+    if (shared && is_pmem) {
+        mmap_flags |= MAP_SYNC_FLAGS;
+    }
 
     offset = QEMU_ALIGN_UP((uintptr_t)ptr, align) - (uintptr_t)ptr;
     ptr1 = mmap(ptr + offset, size, PROT_READ | PROT_WRITE,
                 MAP_FIXED |
                 (fd == -1 ? MAP_ANONYMOUS : 0) |
-                (shared ? MAP_SHARED : MAP_PRIVATE),
+                mmap_flags,
                 fd, 0);
+
+
+    if (ptr1 == MAP_FAILED &&
+        (mmap_flags & MAP_SYNC_FLAGS) == MAP_SYNC_FLAGS) {
+        if (errno == ENOTSUP) {
+            perror("failed to validate with mapping flags");
+        }
+        /* if map failed with MAP_SHARED_VALIDATE | MAP_SYNC,
+         * we will remove these flags to handle compatibility.
+         */
+        ptr1 = mmap(ptr + offset, size, PROT_READ | PROT_WRITE,
+                    MAP_FIXED |
+                    (fd == -1 ? MAP_ANONYMOUS : 0) |
+                    MAP_SHARED,
+                    fd, 0);
+    }
     if (ptr1 == MAP_FAILED) {
         munmap(ptr, total);
         return MAP_FAILED;
-- 
2.7.4

On Thu, Feb 07, 2019 at 01:27:19AM +0800, Zhang, Yi wrote:
> From: Zhang Yi <yi.z.zhang@linux.intel.com>
> 
> When a file supporting DAX is used as vNVDIMM backend, mmap it with
> MAP_SYNC flag in addition which can ensure file system metadata
> synced in each guest writes to the backend file, without other QEMU
> actions (e.g., periodic fsync() by QEMU).
> 
> Current, We have below different possible use cases:
> 
> 1. pmem=on is set, shared=on is set, MAP_SYNC supported:
>    a: backend is a dax supporting file.
>     - MAP_SYNC will active.
>    b: backend is not a dax supporting file.
>     - mmap will trigger a warning. then MAP_SYNC flag will be ignored
> 
> 2. The rest of cases:
>    - we will never pass the MAP_SYNC to mmap2
> 
> Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
> Signed-off-by: Zhang Yi <yi.z.zhang@linux.intel.com>
> ---
>  include/qemu/osdep.h |  7 +++++++
>  util/mmap-alloc.c    | 24 +++++++++++++++++++++++-
>  2 files changed, 30 insertions(+), 1 deletion(-)
> 
> diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
> index 457d24e..9a94cc3 100644
> --- a/include/qemu/osdep.h
> +++ b/include/qemu/osdep.h
> @@ -115,6 +115,13 @@ extern int daemon(int, int);
>  #include "sysemu/os-win32.h"
>  #endif
>  
> +#ifdef CONFIG_LINUX
> +#include <linux/mman.h>
> +#else  /* !CONFIG_LINUX */
> +#define MAP_SYNC              0x0
> +#define MAP_SHARED_VALIDATE   0x0
> +#endif /* CONFIG_LINUX */
> +
>  #ifdef CONFIG_POSIX
>  #include "sysemu/os-posix.h"
>  #endif

It's only used in one place. Maybe put this code in mmap-alloc.c ?

> diff --git a/util/mmap-alloc.c b/util/mmap-alloc.c
> index 97bbeed..e4e55fc 100644
> --- a/util/mmap-alloc.c
> +++ b/util/mmap-alloc.c
> @@ -15,6 +15,7 @@
>  #include "qemu/host-utils.h"
>  
>  #define HUGETLBFS_MAGIC       0x958458f6
> +#define MAP_SYNC_FLAGS        (MAP_SYNC | MAP_SHARED_VALIDATE)
>  

Pls don't do this, just put it in a local variable within qemu_ram_mmap.

>  #ifdef CONFIG_LINUX
>  #include <sys/vfs.h>
> @@ -101,6 +102,7 @@ void *qemu_ram_mmap(int fd,
>  #else
>      void *ptr = mmap(0, total, PROT_NONE, MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
>  #endif
> +    int mmap_flags;
>      size_t offset;
>      void *ptr1;
>  
> @@ -111,13 +113,33 @@ void *qemu_ram_mmap(int fd,
>      assert(is_power_of_2(align));
>      /* Always align to host page size */
>      assert(align >= getpagesize());
> +    mmap_flags = shared ? MAP_SHARED : MAP_PRIVATE;
> +    if (shared && is_pmem) {
> +        mmap_flags |= MAP_SYNC_FLAGS;
> +    }
>  
>      offset = QEMU_ALIGN_UP((uintptr_t)ptr, align) - (uintptr_t)ptr;
>      ptr1 = mmap(ptr + offset, size, PROT_READ | PROT_WRITE,
>                  MAP_FIXED |
>                  (fd == -1 ? MAP_ANONYMOUS : 0) |
> -                (shared ? MAP_SHARED : MAP_PRIVATE),
> +                mmap_flags,
>                  fd, 0);
> +
> +
> +    if (ptr1 == MAP_FAILED &&
> +        (mmap_flags & MAP_SYNC_FLAGS) == MAP_SYNC_FLAGS) {
> +        if (errno == ENOTSUP) {
> +            perror("failed to validate with mapping flags");

I don't think this warning message makes sense.
Are you trying to say:
                Warning: requesting persistence across crashes
		for file XYZ failed. Proceeding without persistence,
		data might become corrupted in case of host crash.
?

> +        }
> +        /* if map failed with MAP_SHARED_VALIDATE | MAP_SYNC,
> +         * we will remove these flags to handle compatibility.
> +         */
> +        ptr1 = mmap(ptr + offset, size, PROT_READ | PROT_WRITE,
> +                    MAP_FIXED |
> +                    (fd == -1 ? MAP_ANONYMOUS : 0) |
> +                    MAP_SHARED,
> +                    fd, 0);
> +    }
>      if (ptr1 == MAP_FAILED) {
>          munmap(ptr, total);
>          return MAP_FAILED;
> -- 
> 2.7.4