Release 0.9.6

- Fix vitastor-disk partition zeroing (sometimes it was writing garbage instead of zeroes)
- Fix incorrect EC space statistics in `vitastor-cli status`
- Several bug fixes for NFS:
  - Add . and .. in NFS directory listings
  - Return FILE_SYNC from NFS writes if immediate_commit is enabled
  - Return the same "verifier" in NFS COMMIT as in NFS WRITE
  - Make parallel NFS extending writes work correctly, without conflicts
  - Handle parallel NFS extending writes without imposing extra load on etcd
- Support UTF-8 in vitastor-cli table output
- Also allow "0" and "no" as false for inmemory_metadata and inmemory_journal
- Use HDD defaults for HDD-only in automatic `vitastor-disk prepare` mode

CMakeLists.txt

@@ -2,6 +2,6 @@ cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
-set(VERSION "0.9.4")
+set(VERSION "0.9.6")
 
 add_subdirectory(src)

csi/Makefile

@@ -1,4 +1,4 @@
-VERSION ?= v0.9.4
+VERSION ?= v0.9.6
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v0.9.4
+          image: vitalif/vitastor-csi:v0.9.6
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -116,7 +116,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v0.9.4
+          image: vitalif/vitastor-csi:v0.9.6
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "0.9.4"
+    vitastorCSIDriverVersion = "0.9.6"
 )
 
 // Config struct fills the parameters of request or user input

debian/changelog

@@ -1,10 +1,10 @@
-vitastor (0.9.4-1) unstable; urgency=medium
+vitastor (0.9.6-1) unstable; urgency=medium
 
   * Bugfixes
 
  -- Vitaliy Filippov <vitalif@yourcmc.ru>  Fri, 03 Jun 2022 02:09:44 +0300
 
-vitastor (0.9.4-1) unstable; urgency=medium
+vitastor (0.9.6-1) unstable; urgency=medium
 
   * Implement NFS proxy
   * Add documentation

debian/vitastor.Dockerfile

@@ -35,8 +35,8 @@ RUN set -e -x; \
     mkdir -p /root/packages/vitastor-$REL; \
     rm -rf /root/packages/vitastor-$REL/*; \
     cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-0.9.4; \
+    cp -r /root/vitastor vitastor-0.9.6; \
-    cd vitastor-0.9.4; \
+    cd vitastor-0.9.6; \
     ln -s /root/fio-build/fio-*/ ./fio; \
     FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
@@ -49,8 +49,8 @@ RUN set -e -x; \
     rm -rf a b; \
     echo "dep:fio=$FIO" > debian/fio_version; \
     cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.9.4.orig.tar.xz vitastor-0.9.4; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.9.6.orig.tar.xz vitastor-0.9.6; \
-    cd vitastor-0.9.4; \
+    cd vitastor-0.9.6; \
     V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     DEBFULLNAME="Vitaliy Filippov <vitalif@yourcmc.ru>" dch -D $REL -v "$V""$REL" "Rebuild for $REL"; \
     DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage --jobs=auto -sa; \

docs/usage/qemu.en.md

@@ -107,7 +107,8 @@ disabled by now, so if you want to try it on Debian, use a kernel from Ubuntu
 Commands to attach Vitastor image as a VDUSE device:
 
 ```
-modprobe vduse virtio-vdpa
+modprobe vduse
+modprobe virtio-vdpa
 qemu-storage-daemon --daemonize --blockdev '{"node-name":"test1","driver":"vitastor",\
   "etcd-host":"192.168.7.2:2379/v3","image":"testosd1","cache":{"direct":true,"no-flush":false},"discard":"unmap"}' \
   --export vduse-blk,id=test1,node-name=test1,name=test1,num-queues=16,queue-size=128,writable=true

docs/usage/qemu.ru.md

@@ -111,7 +111,8 @@ VDUSE (CONFIG_VIRTIO_VDPA=m и CONFIG_VDPA_USER=m). В ядрах в Debian Linu
 Команды для подключения виртуального диска через VDUSE:
 
 ```
-modprobe vduse virtio-vdpa
+modprobe vduse
+modprobe virtio-vdpa
 qemu-storage-daemon --daemonize --blockdev '{"node-name":"test1","driver":"vitastor",\
   "etcd-host":"192.168.7.2:2379/v3","image":"testosd1","cache":{"direct":true,"no-flush":false},"discard":"unmap"}' \
   --export vduse-blk,id=test1,node-name=test1,name=test1,num-queues=16,queue-size=128,writable=true

mon/mon.js

@@ -1497,10 +1497,14 @@ class Mon
                     break;
                 }
             }
+            const pool_cfg = (this.state.config.pools[pool_id]||{});
             if (!object_size)
             {
-                object_size = (this.state.config.pools[pool_id]||{}).block_size ||
+                object_size = pool_cfg.block_size || this.config.block_size || 131072;
-                    this.config.block_size || 131072;
+            }
+            if (pool_cfg.scheme !== 'replicated')
+            {
+                object_size *= ((pool_cfg.pg_size||0) - (pool_cfg.parity_chunks||0));
             }
             object_size = BigInt(object_size);
             for (const pg_num in this.state.pg.stats[pool_id])

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VERSION = '0.9.4'
+VERSION = '0.9.6'
 
 LOG = logging.getLogger(__name__)
 

rpm/build-tarball.sh

@@ -24,4 +24,4 @@ rm fio
 mv fio-copy fio
 FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Version[\s:]+(\S+)/ && print $1; /^Release[\s:]+(\S+)/ && print "-$1"; }'`
 perl -i -pe 's/(Requires:\s*fio)([^\n]+)?/$1 = '$FIO'/' $VITASTOR/rpm/vitastor-el$EL.spec
-tar --transform 's#^#vitastor-0.9.4/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.9.4$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-0.9.6/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.9.6$(rpm --eval '%dist').tar.gz *

rpm/vitastor-el7.Dockerfile

@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.9.4.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.9.6.el7.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el7.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.9.4
+Version:        0.9.6
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-0.9.4.el7.tar.gz
+Source0:        vitastor-0.9.6.el7.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

rpm/vitastor-el8.Dockerfile

@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.9.4.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.9.6.el8.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el8.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.9.4
+Version:        0.9.6
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-0.9.4.el8.tar.gz
+Source0:        vitastor-0.9.6.el8.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

rpm/vitastor-el9.Dockerfile

@@ -18,7 +18,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.9.4.el9.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.9.6.el9.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el9.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.9.4
+Version:        0.9.6
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-0.9.4.el9.tar.gz
+Source0:        vitastor-0.9.6.el9.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

src/CMakeLists.txt

@@ -16,7 +16,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 	set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
 endif()
 
-add_definitions(-DVERSION="0.9.4")
+add_definitions(-DVERSION="0.9.6")
 add_definitions(-Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -I ${CMAKE_SOURCE_DIR}/src)
 if (${WITH_ASAN})
 	add_definitions(-fsanitize=address -fno-omit-frame-pointer)

src/blockstore_open.cpp

@@ -85,11 +85,13 @@ void blockstore_impl_t::parse_config(blockstore_config_t & config, bool init)
         immediate_commit = IMMEDIATE_SMALL;
     }
     metadata_buf_size = strtoull(config["meta_buf_size"].c_str(), NULL, 10);
-    inmemory_meta = config["inmemory_metadata"] != "false";
+    inmemory_meta = config["inmemory_metadata"] != "false" && config["inmemory_metadata"] != "0" &&
+        config["inmemory_metadata"] != "no";
     journal.sector_count = strtoull(config["journal_sector_buffer_count"].c_str(), NULL, 10);
     journal.no_same_sector_overwrites = config["journal_no_same_sector_overwrites"] == "true" ||
         config["journal_no_same_sector_overwrites"] == "1" || config["journal_no_same_sector_overwrites"] == "yes";
-    journal.inmemory = config["inmemory_journal"] != "false";
+    journal.inmemory = config["inmemory_journal"] != "false" && config["inmemory_journal"] != "0" &&
+        config["inmemory_journal"] != "no";
     // Validate
     if (journal.sector_count < 2)
     {

src/cli_ls.cpp

@@ -379,16 +379,18 @@ resume_1:
 
 std::string print_table(json11::Json items, json11::Json header, bool use_esc)
 {
+    int header_sizes[header.array_items().size()];
     std::vector<int> sizes;
     for (int i = 0; i < header.array_items().size(); i++)
     {
-        sizes.push_back(header[i]["title"].string_value().length());
+        header_sizes[i] = utf8_length(header[i]["title"].string_value());
+        sizes.push_back(header_sizes[i]);
     }
     for (auto & item: items.array_items())
     {
         for (int i = 0; i < header.array_items().size(); i++)
         {
-            int l = item[header[i]["key"].string_value()].as_string().length();
+            int l = utf8_length(item[header[i]["key"].string_value()].as_string());
             sizes[i] = sizes[i] < l ? l : sizes[i];
         }
     }
@@ -400,7 +402,7 @@ std::string print_table(json11::Json items, json11::Json header, bool use_esc)
             // Separator
             str += "  ";
         }
-        int pad = sizes[i]-header[i]["title"].string_value().length();
+        int pad = sizes[i]-header_sizes[i];
         if (header[i]["right"].bool_value())
         {
             // Align right
@@ -428,7 +430,7 @@ std::string print_table(json11::Json items, json11::Json header, bool use_esc)
                 // Separator
                 str += "  ";
             }
-            int pad = sizes[i] - item[header[i]["key"].string_value()].as_string().length();
+            int pad = sizes[i] - utf8_length(item[header[i]["key"].string_value()].as_string());
             if (header[i]["right"].bool_value())
             {
                 // Align right

src/cli_modify.cpp

@@ -13,7 +13,7 @@ struct image_changer_t
     std::string image_name;
     std::string new_name;
     uint64_t new_size = 0;
-    bool force_size = false;
+    bool force_size = false, inc_size = false;
     bool set_readonly = false, set_readwrite = false, force = false;
     // interval between fsyncs
     int fsync_interval = 128;
@@ -81,14 +81,14 @@ struct image_changer_t
         }
         if ((!set_readwrite || !cfg.readonly) &&
             (!set_readonly || cfg.readonly) &&
-            (!new_size && !force_size || cfg.size == new_size) &&
+            (!new_size && !force_size || cfg.size == new_size || cfg.size >= new_size && inc_size) &&
             (new_name == "" || new_name == image_name))
         {
             result = (cli_result_t){ .text = "No change" };
             state = 100;
             return;
         }
-        if (new_size != 0 || force_size)
+        if ((new_size != 0 || force_size) && (cfg.size < new_size || !inc_size))
         {
             if (cfg.size >= new_size)
             {
@@ -233,6 +233,7 @@ std::function<bool(cli_result_t &)> cli_tool_t::start_modify(json11::Json cfg)
     changer->new_name = cfg["rename"].string_value();
     changer->new_size = parse_size(cfg["resize"].as_string());
     changer->force_size = cfg["force_size"].bool_value();
+    changer->inc_size = cfg["inc_size"].bool_value();
     changer->force = cfg["force"].bool_value();
     changer->set_readonly = cfg["readonly"].bool_value();
     changer->set_readwrite = cfg["readwrite"].bool_value();

src/disk_tool_prepare.cpp

@@ -147,7 +147,7 @@ int disk_tool_t::prepare_one(std::map<std::string, std::string> options, int is_
         for (int i = 0; i < sizeof(allow_additional_params)/sizeof(allow_additional_params[0]); i++)
         {
             auto it = options.find(allow_additional_params[i]);
-            if (it != options.end())
+            if (it != options.end() && it->second != "")
             {
                 sb[it->first] = it->second;
             }
@@ -621,7 +621,7 @@ int disk_tool_t::prepare(std::vector<std::string> devices)
                     }
                 }
                 // Treat all disks as SSDs if not in the hybrid mode
-                prepare_one(options, hybrid && dev.is_hdd ? 1 : 0);
+                prepare_one(options, dev.is_hdd ? 1 : 0);
                 if (hybrid)
                 {
                     options.erase("journal_device");

src/disk_tool_utils.cpp

@@ -264,6 +264,7 @@ int write_zero(int fd, uint64_t offset, uint64_t size)
 {
     uint64_t buf_len = 1024*1024;
     void *zero_buf = memalign_or_die(MEM_ALIGNMENT, buf_len);
+    memset(zero_buf, 0, buf_len);
     ssize_t r;
     while (size > 0)
     {

src/nfs_conn.cpp

@@ -190,7 +190,15 @@ static int nfs3_setattr_proc(void *opaque, rpc_op_t *rop)
     {
         if (handle == "roothandle" || self->parent->dir_by_hash.find(handle) != self->parent->dir_by_hash.end())
         {
-            *reply = (SETATTR3res){ .status = NFS3ERR_ISDIR };
+            if (args->new_attributes.size.set_it)
+            {
+                *reply = (SETATTR3res){ .status = NFS3ERR_ISDIR };
+            }
+            else
+            {
+                // Silently ignore mode, uid, gid, atime, mtime changes
+                *reply = (SETATTR3res){ .status = NFS3_OK };
+            }
         }
         else
         {
@@ -358,7 +366,6 @@ static int nfs3_read_proc(void *opaque, rpc_op_t *rop)
 }
 
 static void nfs_resize_write(nfs_client_t *self, rpc_op_t *rop, uint64_t inode, uint64_t new_size, uint64_t offset, uint64_t count, void *buf);
-static void nfs_do_write(nfs_client_t *self, rpc_op_t *rop, uint64_t inode, uint64_t offset, uint64_t count, void *buf);
 
 static int nfs3_write_proc(void *opaque, rpc_op_t *rop)
 {
@@ -392,7 +399,6 @@ static int nfs3_write_proc(void *opaque, rpc_op_t *rop)
         .resok = (WRITE3resok){
             //.file_wcc = ...,
             .count = (unsigned)count,
-            .committed = args->stable,
         },
     };
     if ((args->offset % alignment) != 0 || (count % alignment) != 0)
@@ -436,42 +442,101 @@ static int nfs3_write_proc(void *opaque, rpc_op_t *rop)
     return 1;
 }
 
-static void nfs_resize_write(nfs_client_t *self, rpc_op_t *rop, uint64_t inode, uint64_t new_size, uint64_t offset, uint64_t count, void *buf)
+static void complete_extend_write(nfs_client_t *self, rpc_op_t *rop, inode_t inode, int res)
 {
-    // Check if we have to resize the inode before writing
+    WRITE3args *args = (WRITE3args*)rop->request;
+    WRITE3res *reply = (WRITE3res*)rop->reply;
+    if (res < 0)
+    {
+        *reply = (WRITE3res){ .status = vitastor_nfs_map_err(res) };
+        rpc_queue_reply(rop);
+        return;
+    }
+    bool imm = self->parent->cli->get_immediate_commit(inode);
+    reply->resok.committed = args->stable != UNSTABLE || imm ? FILE_SYNC : UNSTABLE;
+    *(uint64_t*)reply->resok.verf = self->parent->server_id;
+    if (args->stable != UNSTABLE && !imm)
+    {
+        // Client requested a stable write. Add an fsync
+        auto op = new cluster_op_t;
+        op->opcode = OSD_OP_SYNC;
+        op->callback = [rop](cluster_op_t *op)
+        {
+            if (op->retval != 0)
+            {
+                WRITE3res *reply = (WRITE3res*)rop->reply;
+                *reply = (WRITE3res){ .status = vitastor_nfs_map_err(-op->retval) };
+            }
+            delete op;
+            rpc_queue_reply(rop);
+        };
+        self->parent->cli->execute(op);
+    }
+    else
+    {
+        rpc_queue_reply(rop);
+    }
+}
+
+static void complete_extend_inode(nfs_client_t *self, uint64_t inode, uint64_t new_size, int err)
+{
+    auto ext_it = self->extend_writes.lower_bound((extend_size_t){ .inode = inode, .new_size = 0 });
+    while (ext_it != self->extend_writes.end() &&
+        ext_it->first.inode == inode &&
+        ext_it->first.new_size <= new_size)
+    {
+        ext_it->second.resize_res = err;
+        if (ext_it->second.write_res <= 0)
+        {
+            complete_extend_write(self, ext_it->second.rop, inode, ext_it->second.write_res < 0
+                ? ext_it->second.write_res : ext_it->second.resize_res);
+            self->extend_writes.erase(ext_it++);
+        }
+        else
+            ext_it++;
+    }
+}
+
+static void extend_inode(nfs_client_t *self, uint64_t inode, uint64_t new_size)
+{
+    // Send an extend request
+    auto & ext = self->extends[inode];
+    ext.cur_extend = new_size;
     auto inode_it = self->parent->cli->st_cli.inode_config.find(inode);
     if (inode_it != self->parent->cli->st_cli.inode_config.end() &&
         inode_it->second.size < new_size)
     {
         self->parent->cmd->loop_and_wait(self->parent->cmd->start_modify(json11::Json::object {
+            // FIXME: Resizing by ID is probably more correct
             { "image", inode_it->second.name },
             { "resize", new_size },
+            { "inc_size", true },
             { "force_size", true },
         }), [=](const cli_result_t & r)
         {
+            auto & ext = self->extends[inode];
             if (r.err)
             {
-                if (r.err == EAGAIN)
+                fprintf(stderr, "Error extending inode %lu to %lu bytes: %s\n", inode, new_size, r.text.c_str());
-                {
+            }
-                    // Multiple concurrent resize requests received, try to repeat
+            if (r.err == EAGAIN || ext.next_extend > ext.cur_extend)
-                    nfs_resize_write(self, rop, inode, new_size, offset, count, buf);
+            {
-                    return;
+                // Multiple concurrent resize requests received, try to repeat
-                }
+                extend_inode(self, inode, ext.next_extend > ext.cur_extend ? ext.next_extend : ext.cur_extend);
-                WRITE3res *reply = (WRITE3res*)rop->reply;
-                *reply = (WRITE3res){ .status = vitastor_nfs_map_err(r.err) };
-                rpc_queue_reply(rop);
                 return;
             }
-            nfs_do_write(self, rop, inode, offset, count, buf);
+            ext.cur_extend = ext.next_extend = 0;
+            complete_extend_inode(self, inode, new_size, r.err);
         });
     }
     else
     {
-        nfs_do_write(self, rop, inode, offset, count, buf);
+        complete_extend_inode(self, inode, new_size, 0);
     }
 }
 
-static void nfs_do_write(nfs_client_t *self, rpc_op_t *rop, uint64_t inode, uint64_t offset, uint64_t count, void *buf)
+static void nfs_do_write(nfs_client_t *self, std::multimap<extend_size_t, extend_write_t>::iterator ewr_it,
+    rpc_op_t *rop, uint64_t inode, uint64_t offset, uint64_t count, void *buf)
 {
     cluster_op_t *op = new cluster_op_t;
     op->opcode = OSD_OP_WRITE;
@@ -479,48 +544,61 @@ static void nfs_do_write(nfs_client_t *self, rpc_op_t *rop, uint64_t inode, uint
     op->offset = offset;
     op->len = count;
     op->iov.push_back(buf, count);
-    op->callback = [self, rop](cluster_op_t *op)
+    op->callback = [self, ewr_it, rop](cluster_op_t *op)
     {
-        uint64_t inode = op->inode;
+        auto inode = op->inode;
-        WRITE3args *args = (WRITE3args*)rop->request;
+        int write_res = op->retval < 0 ? op->retval : (op->retval != op->len ? -ERANGE : 0);
-        WRITE3res *reply = (WRITE3res*)rop->reply;
+        if (ewr_it == self->extend_writes.end())
-        if (op->retval != op->len)
         {
-            *reply = (WRITE3res){ .status = vitastor_nfs_map_err(-op->retval) };
+            complete_extend_write(self, rop, inode, write_res);
-            delete op;
-            rpc_queue_reply(rop);
         }
         else
         {
-            *(uint64_t*)reply->resok.verf = self->parent->server_id;
+            ewr_it->second.write_res = write_res;
-            delete op;
+            if (ewr_it->second.resize_res <= 0)
-            if (args->stable != UNSTABLE &&
-                !self->parent->cli->get_immediate_commit(inode))
             {
-                // Client requested a stable write. Add an fsync
+                complete_extend_write(self, rop, inode, write_res < 0 ? write_res : ewr_it->second.resize_res);
-                op = new cluster_op_t;
+                self->extend_writes.erase(ewr_it);
-                op->opcode = OSD_OP_SYNC;
-                op->callback = [rop](cluster_op_t *op)
-                {
-                    if (op->retval != 0)
-                    {
-                        WRITE3res *reply = (WRITE3res*)rop->reply;
-                        *reply = (WRITE3res){ .status = vitastor_nfs_map_err(-op->retval) };
-                    }
-                    delete op;
-                    rpc_queue_reply(rop);
-                };
-                self->parent->cli->execute(op);
-            }
-            else
-            {
-                rpc_queue_reply(rop);
             }
         }
     };
     self->parent->cli->execute(op);
 }
 
+static void nfs_resize_write(nfs_client_t *self, rpc_op_t *rop, uint64_t inode, uint64_t new_size, uint64_t offset, uint64_t count, void *buf)
+{
+    // Check if we have to resize the inode during write
+    auto inode_it = self->parent->cli->st_cli.inode_config.find(inode);
+    if (inode_it != self->parent->cli->st_cli.inode_config.end() &&
+        inode_it->second.size < new_size)
+    {
+        auto ewr_it = self->extend_writes.emplace((extend_size_t){
+            .inode = inode,
+            .new_size = new_size,
+        }, (extend_write_t){
+            .rop = rop,
+            .resize_res = 1,
+            .write_res = 1,
+        });
+        auto & ext = self->extends[inode];
+        if (ext.cur_extend > 0)
+        {
+            // Already resizing, just wait
+            if (ext.next_extend < new_size)
+                ext.next_extend = new_size;
+        }
+        else
+        {
+            extend_inode(self, inode, new_size);
+        }
+        nfs_do_write(self, ewr_it, rop, inode, offset, count, buf);
+    }
+    else
+    {
+        nfs_do_write(self, self->extend_writes.end(), rop, inode, offset, count, buf);
+    }
+}
+
 static int nfs3_create_proc(void *opaque, rpc_op_t *rop)
 {
     nfs_client_t *self = (nfs_client_t*)opaque;
@@ -881,6 +959,27 @@ static int nfs3_link_proc(void *opaque, rpc_op_t *rop)
     return 0;
 }
 
+static void fill_dir_entry(nfs_client_t *self, rpc_op_t *rop,
+    std::map<std::string, nfs_dir_t>::iterator dir_id_it, struct entryplus3 *entry, bool is_plus)
+{
+    if (dir_id_it == self->parent->dir_info.end())
+    {
+        return;
+    }
+    entry->fileid = dir_id_it->second.id;
+    if (is_plus)
+    {
+        entry->name_attributes = (post_op_attr){
+            .attributes_follow = 1,
+            .attributes = get_dir_attributes(self, dir_id_it->first),
+        };
+        entry->name_handle = (post_op_fh3){
+            .handle_follows = 1,
+            .handle = xdr_copy_string(rop->xdrs, "S"+base64_encode(sha256(dir_id_it->first))),
+        };
+    }
+}
+
 static void nfs3_readdir_common(void *opaque, rpc_op_t *rop, bool is_plus)
 {
     nfs_client_t *self = (nfs_client_t*)opaque;
@@ -958,17 +1057,17 @@ static void nfs3_readdir_common(void *opaque, rpc_op_t *rop, bool is_plus)
             continue;
         std::string subname = dir_id_it->first.substr(prefix.size());
         // for directories, fileid changes when the user restarts proxy
-        entries[subname].fileid = dir_id_it->second.id;
+        fill_dir_entry(self, rop, dir_id_it, &entries[subname], is_plus);
-        if (is_plus)
+    }
+    // Add . and ..
+    {
+        auto dir_id_it = self->parent->dir_info.find(dir);
+        fill_dir_entry(self, rop, dir_id_it, &entries["."], is_plus);
+        auto sl = dir.rfind("/");
+        if (sl != std::string::npos)
         {
-            entries[subname].name_attributes = (post_op_attr){
+            auto dir_id_it = self->parent->dir_info.find(dir.substr(0, sl));
-                .attributes_follow = 1,
+            fill_dir_entry(self, rop, dir_id_it, &entries[".."], is_plus);
-                .attributes = get_dir_attributes(self, dir_id_it->first),
-            };
-            entries[subname].name_handle = (post_op_fh3){
-                .handle_follows = 1,
-                .handle = xdr_copy_string(rop->xdrs, "S"+base64_encode(sha256(dir_id_it->first))),
-            };
         }
     }
     // Offset results by the continuation cookie (equal to index in the listing)
@@ -1193,10 +1292,11 @@ static int nfs3_commit_proc(void *opaque, rpc_op_t *rop)
     cluster_op_t *op = new cluster_op_t;
     // fsync. we don't know how to fsync a single inode, so just fsync everything
     op->opcode = OSD_OP_SYNC;
-    op->callback = [rop](cluster_op_t *op)
+    op->callback = [self, rop](cluster_op_t *op)
     {
         COMMIT3res *reply = (COMMIT3res*)rop->reply;
         *reply = (COMMIT3res){ .status = vitastor_nfs_map_err(op->retval) };
+        *(uint64_t*)reply->resok.verf = self->parent->server_id;
         rpc_queue_reply(rop);
     };
     self->parent->cli->execute(op);

src/nfs_proxy.h

@@ -86,6 +86,28 @@ struct rpc_free_buffer_t
     unsigned size;
 };
 
+struct extend_size_t
+{
+    inode_t inode;
+    uint64_t new_size;
+};
+
+inline bool operator < (const extend_size_t &a, const extend_size_t &b)
+{
+    return a.inode < b.inode || a.inode == b.inode && a.new_size < b.new_size;
+}
+
+struct extend_write_t
+{
+    rpc_op_t *rop;
+    int resize_res, write_res; // 1 = started, 0 = completed OK, -errno = completed with error
+};
+
+struct extend_inode_t
+{
+    uint64_t cur_extend = 0, next_extend = 0;
+};
+
 class nfs_client_t
 {
 public:
@@ -100,6 +122,8 @@ public:
     rpc_cur_buffer_t cur_buffer = { 0 };
     std::map<uint8_t*, rpc_used_buffer_t> used_buffers;
     std::vector<rpc_free_buffer_t> free_buffers;
+    std::map<inode_t, extend_inode_t> extends;
+    std::multimap<extend_size_t, extend_write_t> extend_writes;
 
     iovec read_iov;
     msghdr read_msg = { 0 };

src/qemu_driver.c

@@ -240,10 +240,7 @@ static void vitastor_uring_handler(void *opaque)
     VitastorClient *client = (VitastorClient*)opaque;
     qemu_mutex_lock(&client->mutex);
     client->bh_uring_scheduled = 0;
-    do
+    vitastor_c_uring_handle_events(client->proxy);
-    {
-        vitastor_c_uring_handle_events(client->proxy);
-    } while (vitastor_c_uring_has_work(client->proxy));
     qemu_mutex_unlock(&client->mutex);
 }
 

src/ringloop.cpp

@@ -66,6 +66,16 @@ void ring_loop_t::unregister_consumer(ring_consumer_t *consumer)
 
 void ring_loop_t::loop()
 {
+    if (ring_eventfd >= 0)
+    {
+        // Reset eventfd counter
+        uint64_t ctr = 0;
+        int r = read(ring_eventfd, &ctr, 8);
+        if (r < 0 && errno != EAGAIN && errno != EINTR)
+        {
+            fprintf(stderr, "Error resetting eventfd: %s\n", strerror(errno));
+        }
+    }
     struct io_uring_cqe *cqe;
     while (!io_uring_peek_cqe(&ring, &cqe))
     {
@@ -84,7 +94,7 @@ void ring_loop_t::loop()
         }
         else
         {
-            printf("Warning: empty callback in SQE\n");
+            fprintf(stderr, "Warning: empty callback in SQE\n");
             free_ring_data[free_ring_data_ptr++] = d - ring_datas;
         }
         io_uring_cqe_seen(&ring, cqe);

src/str_util.cpp

@@ -308,3 +308,19 @@ std::string str_repeat(const std::string & str, int times)
         r += str;
     return r;
 }
+
+size_t utf8_length(const std::string & s)
+{
+    size_t len = 0;
+    for (size_t i = 0; i < s.size(); i++)
+        len += (s[i] & 0xC0) != 0x80;
+    return len;
+}
+
+size_t utf8_length(const char *s)
+{
+    size_t len = 0;
+    for (; *s; s++)
+        len += (*s & 0xC0) != 0x80;
+    return len;
+}

src/str_util.h

@@ -18,3 +18,5 @@ void print_help(const char *help_text, std::string exe_name, std::string cmd, bo
 uint64_t parse_time(std::string time_str, bool *ok = NULL);
 std::string read_all_fd(int fd);
 std::string str_repeat(const std::string & str, int times);
+size_t utf8_length(const std::string & s);
+size_t utf8_length(const char *s);

src/vitastor.pc.in

@@ -6,7 +6,7 @@ includedir=${prefix}/@CMAKE_INSTALL_INCLUDEDIR@
 
 Name: Vitastor
 Description: Vitastor client library
-Version: 0.9.4
+Version: 0.9.6
 Libs: -L${libdir} -lvitastor_client
 Cflags: -I${includedir}