[PATCH v3 0/3] fetch the alignment of device dax

[PATCH v3 0/3] fetch the alignment of device dax

[Jingqi Liu]

This series adds libdaxctl support and fetchs the alignment of
device dax through libdaxctl [1] APIs.

QEMU uses mmap(2) to maps vNVDIMM backends and aligns the mapping
address to the page size (getpagesize(2)) by default. However, some
types of backends may require an alignment different than the page
size. The 'align' option is provided to memory-backend-file to allow
users to specify the proper alignment.

For device dax (e.g., /dev/dax0.0), the 'align' option needs to
match the alignment requirement of the device dax, which can be fetched
through the APIs of libdaxctl version 57 or up.

[1] Libdaxctl is a part of ndctl project.
The project's repository is: https://github.com/pmem/ndctl

Changelog:
  v3:
  - Per Joao's suggestion, require libdaxctl version 57 or up.
  - Per Joao's suggestion, suggest to query the @align with daxctl tool
    in docs/nvdimm.txt.

  v2:
  - Per Paolo and Dan's suggestions, fetch the alignment of device dax
    through libdaxctl APIs.

  v1:
  - The initial version.
    Fetch the alignment through "/sys/dev/char/%d:%d/device/align".

Jingqi Liu (3):
  exec: fetch the alignment of Linux devdax pmem character device nodes
  docs/nvdimm: add description of alignment requirement of device dax
  configure: add libdaxctl support

 configure       | 29 ++++++++++++++++++++++++++
 docs/nvdimm.txt | 10 +++++++++
 exec.c          | 54 ++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 92 insertions(+), 1 deletion(-)

-- 
2.17.1

[PATCH v3 1/3] exec: fetch the alignment of Linux devdax pmem character device nodes

[Jingqi Liu]

If the backend file is devdax pmem character device, the alignment
specified by the option 'align=NUM' in the '-object memory-backend-file'
needs to match the alignment requirement of the devdax pmem character device.

This patch uses the interfaces of libdaxctl to fetch the devdax pmem file
'align', so that we can compare it with the NUM of 'align=NUM'.
The NUM needs to be larger than or equal to the devdax pmem file 'align'.

It also fixes the problem that mmap() returns failure in qemu_ram_mmap()
when the NUM of 'align=NUM' is less than the devdax pmem file 'align'.

Suggested-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Jingqi Liu <jingqi.liu@intel.com>
---
 exec.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 53 insertions(+), 1 deletion(-)

diff --git a/exec.c b/exec.c
index de9d949902..2c3444e47e 100644
--- a/exec.c
+++ b/exec.c
@@ -77,6 +77,10 @@
 
 #include "monitor/monitor.h"
 
+#ifdef CONFIG_LIBDAXCTL
+#include <daxctl/libdaxctl.h>
+#endif
+
 //#define DEBUG_SUBPAGE
 
 #if !defined(CONFIG_USER_ONLY)
@@ -1736,6 +1740,46 @@ static int64_t get_file_size(int fd)
     return size;
 }
 
+static int64_t get_file_align(int fd)
+{
+    int64_t align = -1;
+#if defined(__linux__) && defined(CONFIG_LIBDAXCTL)
+    struct stat st;
+
+    if (fstat(fd, &st) < 0) {
+        return -errno;
+    }
+
+    /* Special handling for devdax character devices */
+    if (S_ISCHR(st.st_mode)) {
+        g_autofree char *path = NULL;
+        g_autofree char *rpath = NULL;
+        struct daxctl_ctx *ctx;
+        struct daxctl_region *region;
+        int rc = 0;
+
+        path = g_strdup_printf("/sys/dev/char/%d:%d",
+                    major(st.st_rdev), minor(st.st_rdev));
+        rpath = realpath(path, NULL);
+
+        rc = daxctl_new(&ctx);
+        if (rc) {
+            return -1;
+        }
+
+        daxctl_region_foreach(ctx, region) {
+            if (strstr(rpath, daxctl_region_get_path(region))) {
+                align = daxctl_region_get_align(region);
+                break;
+            }
+        }
+        daxctl_unref(ctx);
+    }
+#endif /* defined(__linux__) && defined(CONFIG_LIBDAXCTL) */
+
+    return align;
+}
+
 static int file_ram_open(const char *path,
                          const char *region_name,
                          bool *created,
@@ -2275,7 +2319,7 @@ RAMBlock *qemu_ram_alloc_from_fd(ram_addr_t size, MemoryRegion *mr,
 {
     RAMBlock *new_block;
     Error *local_err = NULL;
-    int64_t file_size;
+    int64_t file_size, file_align;
 
     /* Just support these ram flags by now. */
     assert((ram_flags & ~(RAM_SHARED | RAM_PMEM)) == 0);
@@ -2311,6 +2355,14 @@ RAMBlock *qemu_ram_alloc_from_fd(ram_addr_t size, MemoryRegion *mr,
         return NULL;
     }
 
+    file_align = get_file_align(fd);
+    if (file_align > 0 && mr && file_align > mr->align) {
+        error_setg(errp, "backing store align 0x%" PRIx64
+                   " is larger than 'align' option 0x" RAM_ADDR_FMT,
+                   file_align, mr->align);
+        return NULL;
+    }
+
     new_block = g_malloc0(sizeof(*new_block));
     new_block->mr = mr;
     new_block->used_length = size;
-- 
2.17.1

[PATCH v3 2/3] docs/nvdimm: add description of alignment requirement of device dax

[Jingqi Liu]

For device dax (e.g., /dev/dax0.0), the NUM of 'align=NUM' option
needs to match the alignment requirement of the device dax.
It must be larger than or equal to the 'align' of device dax.

Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Jingqi Liu <jingqi.liu@intel.com>
---
 docs/nvdimm.txt | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/docs/nvdimm.txt b/docs/nvdimm.txt
index 362e99109e..c2c6e441b3 100644
--- a/docs/nvdimm.txt
+++ b/docs/nvdimm.txt
@@ -132,6 +132,16 @@ address to the page size (getpagesize(2)) by default. However, some
 types of backends may require an alignment different than the page
 size. In that case, QEMU v2.12.0 and later provide 'align' option to
 memory-backend-file to allow users to specify the proper alignment.
+For device dax (e.g., /dev/dax0.0), this alignment needs to match the
+alignment requirement of the device dax. The NUM of 'align=NUM' option
+must be larger than or equal to the 'align' of device dax.
+We can use one of the following commands to show the 'align' of device dax.
+
+    ndctl list -X
+    daxctl list -R
+
+In order to get the proper 'align' of device dax, you need to install
+the library 'libdaxctl'.
 
 For example, device dax require the 2 MB alignment, so we can use
 following QEMU command line options to use it (/dev/dax0.0) as the
-- 
2.17.1

[PATCH v3 3/3] configure: add libdaxctl support

[Jingqi Liu]

Add a pair of configure options --{enable,disable}-libdaxctl to control
whether QEMU is compiled with libdaxctl [1]. Libdaxctl is a utility
library for managing the device dax subsystem.

QEMU uses mmap(2) to maps vNVDIMM backends and aligns the mapping
address to the page size (getpagesize(2)) by default. However, some
types of backends may require an alignment different than the page
size. The 'align' option is provided to memory-backend-file to allow
users to specify the proper alignment.

For device dax (e.g., /dev/dax0.0), the 'align' option needs to match
the alignment requirement of the device dax, which can be fetched
through the APIs of libdaxctl version 57 or up.

[1] Libdaxctl is a part of ndctl project.
The project's repository is: https://github.com/pmem/ndctl

For more information about libdaxctl APIs, you can refer to the
comments in source code of: pmem/ndctl/daxctl/lib/libdaxctl.c.

Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Jingqi Liu <jingqi.liu@intel.com>
---
 configure | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/configure b/configure
index e225a1e3ff..d2418084c1 100755
--- a/configure
+++ b/configure
@@ -509,6 +509,7 @@ libpmem=""
 default_devices="yes"
 plugins="no"
 fuzzing="no"
+libdaxctl=""
 
 supported_cpu="no"
 supported_os="no"
@@ -1601,6 +1602,10 @@ for opt do
   ;;
   --gdb=*) gdb_bin="$optarg"
   ;;
+  --enable-libdaxctl) libdaxctl=yes
+  ;;
+  --disable-libdaxctl) libdaxctl=no
+  ;;
   *)
       echo "ERROR: unknown option $opt"
       echo "Try '$0 --help' for more information"
@@ -1894,6 +1899,7 @@ disabled with --disable-FEATURE, default is enabled if available:
   debug-mutex     mutex debugging support
   libpmem         libpmem support
   xkbcommon       xkbcommon support
+  libdaxctl       libdaxctl support
 
 NOTE: The object files are built at the place where configure is launched
 EOF
@@ -6190,6 +6196,24 @@ if test "$libpmem" != "no"; then
 	fi
 fi
 
+##########################################
+# check for libdaxctl
+
+if test "$libdaxctl" != "no"; then
+	if $pkg_config --atleast-version=57 "libdaxctl"; then
+		libdaxctl="yes"
+		libdaxctl_libs=$($pkg_config --libs libdaxctl)
+		libdaxctl_cflags=$($pkg_config --cflags libdaxctl)
+		libs_softmmu="$libs_softmmu $libdaxctl_libs"
+		QEMU_CFLAGS="$QEMU_CFLAGS $libdaxctl_cflags"
+	else
+		if test "$libdaxctl" = "yes" ; then
+			feature_not_found "libdaxctl" "Install libdaxctl"
+		fi
+		libdaxctl="no"
+	fi
+fi
+
 ##########################################
 # check for slirp
 
@@ -6767,6 +6791,7 @@ echo "parallels support $parallels"
 echo "sheepdog support  $sheepdog"
 echo "capstone          $capstone"
 echo "libpmem support   $libpmem"
+echo "libdaxctl support $libdaxctl"
 echo "libudev           $libudev"
 echo "default devices   $default_devices"
 echo "plugin support    $plugins"
@@ -7590,6 +7615,10 @@ if test "$libpmem" = "yes" ; then
   echo "CONFIG_LIBPMEM=y" >> $config_host_mak
 fi
 
+if test "$libdaxctl" = "yes" ; then
+  echo "CONFIG_LIBDAXCTL=y" >> $config_host_mak
+fi
+
 if test "$bochs" = "yes" ; then
   echo "CONFIG_BOCHS=y" >> $config_host_mak
 fi
-- 
2.17.1