Skip to contentSkip to conversation

block: Gluster 6 compatibility

PATCH Superseded Niels de Vos 3 participants 7 messages Mar 4, 2019 – Mar 5, 2019 Block layer

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover block: Gluster 6 compatibility

Gluster 6 is currently available as release candidate. There have been a few changes to libgfapi.so that need to be adapted by consuming projects like QEMU. Fedora Rawhide already contains glusterfs-6.0-RC0, and this prevents rebuilds of QEMU there (https://bugzilla.redhat.com/1684298).

The following two changes should be sufficient to consume Gluster 6 once it is released. These have been tested on CentOS-7 with Gluster 5 and Gluster 6 (minimal manual qemu-img tests only).

Cheers,
Niels

Niels de Vos (2):
block/gluster: Handle changed glfs_ftruncate signature
gluster: the glfs_io_cbk callback function pointer adds pre/post stat
args

 block/gluster.c | 17 ++++++++++++++---
 configure       | 42 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 56 insertions(+), 3 deletions(-)
Signature · 1 line
2.20.1
#1 lore raw

PATCH 1/2 block/gluster: Handle changed glfs_ftruncate signature

From: Prasanna Kumar Kalever <[email protected]>

New versions of Glusters libgfapi.so have an updated glfs_ftruncate() function that returns additional 'struct stat' structures to enable advanced caching of attributes. This is useful for file servers, not so much for QEMU. Nevertheless, the API has changed and needs to be adopted.

---
v4: rebase to current master branch
v3: define old backwards compatible glfs_ftruncate() macro, from Eric Blake
v2: do a compile check as suggested by Eric Blake

---
 block/gluster.c | 11 +++++++++--
 configure       | 18 ++++++++++++++++++
 2 files changed, 27 insertions(+), 2 deletions(-)
block/gluster.c +9 -2
index af64330211..86e5278524 100644
@@ -20,6 +20,10 @@
 #include "qemu/option.h"
 #include "qemu/cutils.h"
 
+#ifdef CONFIG_GLUSTERFS_LEGACY_FTRUNCATE
+# define glfs_ftruncate(fd, offset, _u1, _u2) glfs_ftruncate(fd, offset)
+#endif
+
 #define GLUSTER_OPT_FILENAME        "filename"
 #define GLUSTER_OPT_VOLUME          "volume"
 #define GLUSTER_OPT_PATH            "path"
@@ -1005,6 +1009,7 @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset,
                                     PreallocMode prealloc, Error **errp)
 {
     int64_t current_length;
+    int ret;
 
     current_length = glfs_lseek(fd, 0, SEEK_END);
     if (current_length < 0) {
@@ -1032,7 +1037,8 @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset,
 #endif /* CONFIG_GLUSTERFS_FALLOCATE */
 #ifdef CONFIG_GLUSTERFS_ZEROFILL
     case PREALLOC_MODE_FULL:
-        if (glfs_ftruncate(fd, offset)) {
+        ret = glfs_ftruncate(fd, offset, NULL, NULL);
+        if (ret) {
             error_setg_errno(errp, errno, "Could not resize file");
             return -errno;
         }
@@ -1043,7 +1049,8 @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset,
         break;
 #endif /* CONFIG_GLUSTERFS_ZEROFILL */
     case PREALLOC_MODE_OFF:
-        if (glfs_ftruncate(fd, offset)) {
+        ret = glfs_ftruncate(fd, offset, NULL, NULL);
+        if (ret) {
             error_setg_errno(errp, errno, "Could not resize file");
             return -errno;
         }
configure +18 -0
index 540bee19ba..1d09bef1f9 100755
@@ -456,6 +456,7 @@ glusterfs_xlator_opt="no"
 glusterfs_discard="no"
 glusterfs_fallocate="no"
 glusterfs_zerofill="no"
+glusterfs_legacy_ftruncate="no"
 gtk=""
 gtk_gl="no"
 tls_priority="NORMAL"
@@ -4057,6 +4058,19 @@ if test "$glusterfs" != "no" ; then
       glusterfs_fallocate="yes"
       glusterfs_zerofill="yes"
     fi
+    cat > $TMPC << EOF
+#include <glusterfs/api/glfs.h>
+
+int
+main(void)
+{
+	/* new glfs_ftruncate() passes two additional args */
+	return glfs_ftruncate(NULL, 0 /*, NULL, NULL */);
+}
+EOF
+    if compile_prog "$glusterfs_cflags" "$glusterfs_libs" ; then
+      glusterfs_legacy_ftruncate="yes"
+    fi
   else
     if test "$glusterfs" = "yes" ; then
       feature_not_found "GlusterFS backend support" \
@@ -6853,6 +6867,10 @@ if test "$glusterfs_zerofill" = "yes" ; then
   echo "CONFIG_GLUSTERFS_ZEROFILL=y" >> $config_host_mak
 fi
 
+if test "$glusterfs_legacy_ftruncate" = "yes" ; then
+  echo "CONFIG_GLUSTERFS_LEGACY_FTRUNCATE=y" >> $config_host_mak
+fi
+
 if test "$libssh2" = "yes" ; then
   echo "CONFIG_LIBSSH2=m" >> $config_host_mak
   echo "LIBSSH2_CFLAGS=$libssh2_cflags" >> $config_host_mak
Signature · 1 line
2.20.1
#2 lore raw

Re: [Qemu-devel] [PATCH 1/2] block/gluster: Handle changed glfs_ftruncate signature

Quoted 25 lines · Niels de Vos

From: Prasanna Kumar Kalever <[email protected]>

New versions of Glusters libgfapi.so have an updated glfs_ftruncate() function that returns additional 'struct stat' structures to enable advanced caching of attributes. This is useful for file servers, not so much for QEMU. Nevertheless, the API has changed and needs to be adopted.

---
v4: rebase to current master branch
v3: define old backwards compatible glfs_ftruncate() macro, from Eric Blake
v2: do a compile check as suggested by Eric Blake

---
 block/gluster.c | 11 +++++++++--
 configure       | 18 ++++++++++++++++++
 2 files changed, 27 insertions(+), 2 deletions(-)
block/gluster.c +0 -0
index af64330211..86e5278524 100644

block/gluster.c line 25

Quoted 6 lines
block/gluster.c +3 -0
@@ -20,6 +20,10 @@
 #include "qemu/option.h"
 #include "qemu/cutils.h"
 
+#ifdef CONFIG_GLUSTERFS_LEGACY_FTRUNCATE
+# define glfs_ftruncate(fd, offset, _u1, _u2) glfs_ftruncate(fd, offset)
+#endif

I don't much like this approach as it sets up a trapdoor. If future QEMU passes a non-NULL value for the 3rd/4th argument it will be silently ignored depending on glfs version built against which can result in incorrect behaviour at runtime. If we reverse it:

#ifndef CONFIG_GLUSTERFS_LEGACY_FTRUNCATE

 # define glfs_ftruncate(fd, offset) glfs_ftruncate(fd, offset, NULL, NULL)
 #endif

then it ensures we can't silently do the wrong thing in future. Anyone who wants to use the 3rd/4th args will have to make an explicit effort to ensure it works correctly with old glfs APIs. An added benefit is that it avoids the following patch chunks.

Quoted 79 lines
block/gluster.c +6 -2
+
 #define GLUSTER_OPT_FILENAME        "filename"
 #define GLUSTER_OPT_VOLUME          "volume"
 #define GLUSTER_OPT_PATH            "path"
@@ -1005,6 +1009,7 @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset,
                                     PreallocMode prealloc, Error **errp)
 {
     int64_t current_length;
+    int ret;
 
     current_length = glfs_lseek(fd, 0, SEEK_END);
     if (current_length < 0) {
@@ -1032,7 +1037,8 @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset,
 #endif /* CONFIG_GLUSTERFS_FALLOCATE */
 #ifdef CONFIG_GLUSTERFS_ZEROFILL
     case PREALLOC_MODE_FULL:
-        if (glfs_ftruncate(fd, offset)) {
+        ret = glfs_ftruncate(fd, offset, NULL, NULL);
+        if (ret) {
             error_setg_errno(errp, errno, "Could not resize file");
             return -errno;
         }
@@ -1043,7 +1049,8 @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset,
         break;
 #endif /* CONFIG_GLUSTERFS_ZEROFILL */
     case PREALLOC_MODE_OFF:
-        if (glfs_ftruncate(fd, offset)) {
+        ret = glfs_ftruncate(fd, offset, NULL, NULL);
+        if (ret) {
             error_setg_errno(errp, errno, "Could not resize file");
             return -errno;
         }
configure +18 -0
index 540bee19ba..1d09bef1f9 100755
@@ -456,6 +456,7 @@ glusterfs_xlator_opt="no"
 glusterfs_discard="no"
 glusterfs_fallocate="no"
 glusterfs_zerofill="no"
+glusterfs_legacy_ftruncate="no"
 gtk=""
 gtk_gl="no"
 tls_priority="NORMAL"
@@ -4057,6 +4058,19 @@ if test "$glusterfs" != "no" ; then
       glusterfs_fallocate="yes"
       glusterfs_zerofill="yes"
     fi
+    cat > $TMPC << EOF
+#include <glusterfs/api/glfs.h>
+
+int
+main(void)
+{
+	/* new glfs_ftruncate() passes two additional args */
+	return glfs_ftruncate(NULL, 0 /*, NULL, NULL */);
+}
+EOF
+    if compile_prog "$glusterfs_cflags" "$glusterfs_libs" ; then
+      glusterfs_legacy_ftruncate="yes"
+    fi
   else
     if test "$glusterfs" = "yes" ; then
       feature_not_found "GlusterFS backend support" \
@@ -6853,6 +6867,10 @@ if test "$glusterfs_zerofill" = "yes" ; then
   echo "CONFIG_GLUSTERFS_ZEROFILL=y" >> $config_host_mak
 fi
 
+if test "$glusterfs_legacy_ftruncate" = "yes" ; then
+  echo "CONFIG_GLUSTERFS_LEGACY_FTRUNCATE=y" >> $config_host_mak
+fi
+
 if test "$libssh2" = "yes" ; then
   echo "CONFIG_LIBSSH2=m" >> $config_host_mak
   echo "LIBSSH2_CFLAGS=$libssh2_cflags" >> $config_host_mak
Signature · 1 line
2.20.1

Regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#3 lore raw

Re: [Qemu-devel] [PATCH 1/2] block/gluster: Handle changed glfs_ftruncate signature

Quoted 38 lines · Daniel P. Berrangé
Quoted 25 lines · Niels de Vos

From: Prasanna Kumar Kalever <[email protected]>

New versions of Glusters libgfapi.so have an updated glfs_ftruncate() function that returns additional 'struct stat' structures to enable advanced caching of attributes. This is useful for file servers, not so much for QEMU. Nevertheless, the API has changed and needs to be adopted.

---
v4: rebase to current master branch
v3: define old backwards compatible glfs_ftruncate() macro, from Eric Blake
v2: do a compile check as suggested by Eric Blake

---
 block/gluster.c | 11 +++++++++--
 configure       | 18 ++++++++++++++++++
 2 files changed, 27 insertions(+), 2 deletions(-)
block/gluster.c +0 -0
index af64330211..86e5278524 100644

block/gluster.c line 25

Quoted 6 lines
block/gluster.c +3 -0
@@ -20,6 +20,10 @@
 #include "qemu/option.h"
 #include "qemu/cutils.h"
 
+#ifdef CONFIG_GLUSTERFS_LEGACY_FTRUNCATE
+# define glfs_ftruncate(fd, offset, _u1, _u2) glfs_ftruncate(fd, offset)
+#endif

I don't much like this approach as it sets up a trapdoor. If future QEMU passes a non-NULL value for the 3rd/4th argument it will be silently ignored depending on glfs version built against which can result in incorrect behaviour at runtime. If we reverse it:

Quoted 8 lines

#ifndef CONFIG_GLUSTERFS_LEGACY_FTRUNCATE

 # define glfs_ftruncate(fd, offset) glfs_ftruncate(fd, offset, NULL, NULL)
 #endif

then it ensures we can't silently do the wrong thing in future. Anyone who wants to use the 3rd/4th args will have to make an explicit effort to ensure it works correctly with old glfs APIs. An added benefit is that it avoids the following patch chunks.

Thanks for the suggestion. All makes sense and I'll update the patch accordingly.

Niels

#4 lore raw

PATCH 2/2 gluster: the glfs_io_cbk callback function pointer adds pre/post stat args

In reply to Niels de Vos

The glfs_*_async() functions do a callback once finished. This callback has changed its arguments, pre- and post-stat structures have been added. This makes it possible to improve cashing, which is useful for Samba and NFS-Ganesha, but not so much for QEMU. Gluster 6 is the first release that includes these new arguments.

With an additional detection in ./configure, the new arguments can conditionally get included in the glfs_io_cbk handler.

---
 block/gluster.c |  6 +++++-
 configure       | 24 ++++++++++++++++++++++++
 2 files changed, 29 insertions(+), 1 deletion(-)
block/gluster.c +5 -1
index 86e5278524..7483c3b2aa 100644
@@ -729,7 +729,11 @@ static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf,
 /*
  * AIO callback routine called from GlusterFS thread.
  */
-static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
+static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret,
+#ifdef CONFIG_GLUSTERFS_IOCB_HAS_STAT
+                                 struct glfs_stat *pre, struct glfs_stat *post,
+#endif
+                                 void *arg)
 {
     GlusterAIOCB *acb = (GlusterAIOCB *)arg;
 
configure +24 -0
index 1d09bef1f9..d8ae8c2f50 100755
@@ -457,6 +457,7 @@ glusterfs_discard="no"
 glusterfs_fallocate="no"
 glusterfs_zerofill="no"
 glusterfs_legacy_ftruncate="no"
+glusterfs_iocb_has_stat="no"
 gtk=""
 gtk_gl="no"
 tls_priority="NORMAL"
@@ -4071,6 +4072,25 @@ EOF
     if compile_prog "$glusterfs_cflags" "$glusterfs_libs" ; then
       glusterfs_legacy_ftruncate="yes"
     fi
+    cat > $TMPC << EOF
+#include <glusterfs/api/glfs.h>
+
+/* new glfs_io_cbk() passes two additional glfs_stat structs */
+static void
+glusterfs_iocb(glfs_fd_t *fd, ssize_t ret, struct glfs_stat *prestat, struct glfs_stat *poststat, void *data)
+{}
+
+int
+main(void)
+{
+	glfs_io_cbk iocb = &glusterfs_iocb;
+	iocb(NULL, 0 , NULL, NULL, NULL);
+	return 0;
+}
+EOF
+    if compile_prog "$glusterfs_cflags" "$glusterfs_libs" ; then
+      glusterfs_iocb_has_stat="yes"
+    fi
   else
     if test "$glusterfs" = "yes" ; then
       feature_not_found "GlusterFS backend support" \
@@ -6871,6 +6891,10 @@ if test "$glusterfs_legacy_ftruncate" = "yes" ; then
   echo "CONFIG_GLUSTERFS_LEGACY_FTRUNCATE=y" >> $config_host_mak
 fi
 
+if test "$glusterfs_iocb_has_stat" = "yes" ; then
+  echo "CONFIG_GLUSTERFS_IOCB_HAS_STAT=y" >> $config_host_mak
+fi
+
 if test "$libssh2" = "yes" ; then
   echo "CONFIG_LIBSSH2=m" >> $config_host_mak
   echo "LIBSSH2_CFLAGS=$libssh2_cflags" >> $config_host_mak
Signature · 1 line
2.20.1
#5 lore raw

Re: [Qemu-devel] [Qemu-block] [PATCH 2/2] gluster: the glfs_io_cbk callback function pointer adds pre/post stat args

Am 04.03.2019 um 17:21 hat Niels de Vos geschrieben:

Quoted 3 lines

The glfs_*_async() functions do a callback once finished. This callback has changed its arguments, pre- and post-stat structures have been added. This makes it possible to improve cashing, which is useful for

Did you mean "caching"?

Quoted 7 lines

Samba and NFS-Ganesha, but not so much for QEMU. Gluster 6 is the first release that includes these new arguments.

With an additional detection in ./configure, the new arguments can conditionally get included in the glfs_io_cbk handler.

Kevin

#6 lore raw

Re: [Qemu-devel] [Qemu-block] [PATCH 2/2] gluster: the glfs_io_cbk callback function pointer adds pre/post stat args

Quoted 6 lines · Kevin Wolf

Am 04.03.2019 um 17:21 hat Niels de Vos geschrieben:

Quoted 3 lines

The glfs_*_async() functions do a callback once finished. This callback has changed its arguments, pre- and post-stat structures have been added. This makes it possible to improve cashing, which is useful for

Did you mean "caching"?

Yuck, yes, of course. I'll correct that when I send an updated version for patch 1/2.

Niels