WIP RDMA credits

.gitea/workflows/buildenv.Dockerfile

@@ -1,29 +1,28 @@
-FROM node:16-bookworm
+FROM node:16-bullseye
 
 WORKDIR /root
 
-ADD ./docker/etc/apt/trusted.gpg.d /etc/apt/trusted.gpg.d
+ADD ./docker/vitastor.gpg /etc/apt/trusted.gpg.d
 
-RUN echo 'deb http://deb.debian.org/debian bookworm-backports main' >> /etc/apt/sources.list; \
-    echo 'deb http://vitastor.io/debian bookworm main' >> /etc/apt/sources.list; \
+RUN echo 'deb http://deb.debian.org/debian bullseye-backports main' >> /etc/apt/sources.list; \
+    echo 'deb http://vitastor.io/debian bullseye main' >> /etc/apt/sources.list; \
     echo >> /etc/apt/preferences; \
     echo 'Package: *' >> /etc/apt/preferences; \
-    echo 'Pin: release n=bookworm-backports' >> /etc/apt/preferences; \
+    echo 'Pin: release a=bullseye-backports' >> /etc/apt/preferences; \
     echo 'Pin-Priority: 500' >> /etc/apt/preferences; \
     echo >> /etc/apt/preferences; \
     echo 'Package: *' >> /etc/apt/preferences; \
     echo 'Pin: origin "vitastor.io"' >> /etc/apt/preferences; \
     echo 'Pin-Priority: 1000' >> /etc/apt/preferences; \
-    perl -i -pe 's/Types: deb$/Types: deb deb-src/' /etc/apt/sources.list.d/debian.sources; \
     grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb/deb-src/' >> /etc/apt/sources.list; \
     echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf; \
     echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf
 
 RUN apt-get update
-RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan8 \
+RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan5 \
     libgoogle-perftools-dev devscripts libjerasure-dev cmake libibverbs-dev libisal-dev
 RUN apt-get -y build-dep fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
-RUN apt-get update && apt-get -y install jq lp-solve sudo nfs-common fdisk parted libc-ares-dev udev
+RUN apt-get update && apt-get -y install jq lp-solve sudo nfs-common fdisk parted
 RUN apt-get --download-only source fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
 
 RUN set -ex; \

.gitea/workflows/test.yml

@@ -63,7 +63,7 @@ jobs:
     container: ${{env.TEST_IMAGE}}:${{github.sha}}
     steps:
     # leak sanitizer sometimes crashes
-    - run: cd /root/vitastor/build && ASAN_OPTIONS=detect_leaks=0 make -j16 build_tests test
+    - run: cd /root/vitastor/build && ASAN_OPTIONS=detect_leaks=0 make -j16 test
 
   npm_lint:
     runs-on: ubuntu-latest
@@ -414,24 +414,6 @@ jobs:
           echo ""
         done
 
-  test_level_placement:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_level_placement.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
   test_snapshot:
     runs-on: ubuntu-latest
     needs: build
@@ -1206,78 +1188,6 @@ jobs:
           echo ""
         done
 
-  test_checksum:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_checksum.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
-  test_old_checksum:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: OLD=1 /root/vitastor/tests/test_checksum.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
-  test_corrupt_all:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_corrupt_all.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
-  test_old_corrupt_all:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: OLD=1 /root/vitastor/tests/test_corrupt_all.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
   test_reweight_half:
     runs-on: ubuntu-latest
     needs: build
@@ -1890,24 +1800,6 @@ jobs:
           echo ""
         done
 
-  test_old_partwr_csum:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: OLD=1 /root/vitastor/tests/test_partwr_csum.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-
   test_heal_old_csum_32k_dmj:
     runs-on: ubuntu-latest
     needs: build
@@ -2034,21 +1926,3 @@ jobs:
           echo ""
         done
 
-  test_nfs_unaligned_append:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_nfs_unaligned_append.sh
-    - name: Print logs
-      if: always() && steps.test.outcome == 'failure'
-      run: |
-        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
-          echo "-------- $i --------"
-          cat $i
-          echo ""
-        done
-

CMakeLists.txt

@@ -1,20 +1,20 @@
-cmake_minimum_required(VERSION 2.8...3.30)
+cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
-set(VITASTOR_VERSION "3.0.12")
+set(VITASTOR_VERSION "3.0.0")
 
 include(CTest)
 
 add_custom_target(build_tests)
-set_property(TEST PROPERTY ENVIRONMENT LSAN_OPTIONS=suppressions=${CMAKE_CURRENT_BINARY_DIR}/lsan-suppress.txt)
-add_test(gen_lsan_suppress
-    ${CMAKE_COMMAND} -E echo leak:tcmalloc > "${CMAKE_CURRENT_BINARY_DIR}/lsan-suppress.txt"
+add_custom_target(test
+	COMMAND
+	echo leak:tcmalloc > ${CMAKE_CURRENT_BINARY_DIR}/lsan-suppress.txt &&
+	env LSAN_OPTIONS=suppressions=${CMAKE_CURRENT_BINARY_DIR}/lsan-suppress.txt ${CMAKE_CTEST_COMMAND}
 )
-set_tests_properties(gen_lsan_suppress PROPERTIES FIXTURES_SETUP f_lsan_suppress)
-set_property(TEST PROPERTY FIXTURES_REQUIRED f_lsan_suppress)
 # make -j16 -C ../../build test_heap && ../../build/src/test/test_heap
 # make -j16 -C ../../build test_heap && rm -f $(find ../../build -name '*.gcda') && ctest -V -T test -T coverage -R heap --test-dir ../../build && (cd ../../build; gcovr -f ../src --html --html-nested -o coverage/index.html; cd ../src/test)
 # make -j16 -C ../../build test_blockstore && rm -f $(find ../../build -name '*.gcda') && ctest -V -T test -T coverage -R blockstore --test-dir ../../build && (cd ../../build; gcovr -f ../src --html --html-nested -o coverage/index.html; cd ../src/test)
 # kcov --include-path=../../../src ../../kcov ./test_blockstore
+add_dependencies(test build_tests)
 add_subdirectory(src)

csi/Dockerfile

@@ -1,5 +1,5 @@
 # Compile stage
-FROM golang:trixie AS build
+FROM golang:bookworm AS build
 
 ADD go.sum go.mod /app/
 RUN cd /app; CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go mod download -x
@@ -9,7 +9,7 @@ RUN perl -i -e '$/ = undef; while(<>) { s/\n\s*(\{\s*\n)/$1\n/g; s/\}(\s*\n\s*)e
     CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go build -o vitastor-csi
 
 # Final stage
-FROM debian:trixie
+FROM debian:bookworm
 
 LABEL maintainers="Vitaliy Filippov <vitalif@yourcmc.ru>"
 LABEL description="Vitastor CSI Driver"
@@ -25,20 +25,20 @@ RUN apt-get update && \
         # NFS mount dependencies
         nfs-common netbase \
         # dependencies of qemu-storage-daemon
-        libaio1t64 libc6 libfuse3-4 libglib2.0-0t64 libgmp10 libgnutls30t64 \
-        libhogweed6t64 libnettle8t64 libnuma1 libselinux1 liburing2 libzstd1 zlib1g && \
+        libnuma1 liburing2 libglib2.0-0 libfuse3-3 libaio1 libzstd1 libnettle8 \
+        libgmp10 libhogweed6 libp11-kit0 libidn2-0 libunistring2 libtasn1-6 libpcre2-8-0 libffi8 && \
     apt-get clean && \
     (echo options nbd nbds_max=128 > /etc/modprobe.d/nbd.conf)
 
 COPY --from=build /app/vitastor-csi /bin/
 
-RUN (echo deb http://vitastor.io/debian trixie main > /etc/apt/sources.list.d/vitastor.list) && \
+RUN (echo deb http://vitastor.io/debian bookworm main > /etc/apt/sources.list.d/vitastor.list) && \
     ((echo 'Package: *'; echo 'Pin: origin "vitastor.io"'; echo 'Pin-Priority: 1000') > /etc/apt/preferences.d/vitastor.pref) && \
     wget -q -O /etc/apt/trusted.gpg.d/vitastor.gpg https://vitastor.io/debian/pubkey.gpg && \
     apt-get update && \
     apt-get install -y vitastor-client ibverbs-providers && \
-    wget https://vitastor.io/archive/qemu/qemu-trixie-10.0.2%2Bds-2%2Bvitastor1/qemu-utils_10.0.2%2Bds-2%2Bvitastor1_amd64.deb && \
-    wget https://vitastor.io/archive/qemu/qemu-trixie-10.0.2%2Bds-2%2Bvitastor1/qemu-block-extra_10.0.2%2Bds-2%2Bvitastor1_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-utils_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-block-extra_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
     dpkg -x qemu-utils*.deb tmp1 && \
     dpkg -x qemu-block-extra*.deb tmp1 && \
     cp -a tmp1/usr/bin/qemu-storage-daemon /usr/bin/ && \

csi/Dockerfile.local

@@ -1,5 +1,5 @@
 # Compile stage
-FROM golang:trixie AS build
+FROM golang:bookworm AS build
 
 ADD go.sum go.mod /app/
 RUN cd /app; CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go mod download -x
@@ -9,7 +9,7 @@ RUN perl -i -e '$/ = undef; while(<>) { s/\n\s*(\{\s*\n)/$1\n/g; s/\}(\s*\n\s*)e
     CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go build -o vitastor-csi
 
 # Final stage
-FROM debian:trixie
+FROM debian:bookworm
 
 LABEL maintainers="Vitaliy Filippov <vitalif@yourcmc.ru>"
 LABEL description="Vitastor CSI Driver"
@@ -36,8 +36,8 @@ ADD deb /deb
 
 RUN apt-get update && \
     apt-get -y install /deb/vitastor-client_*.deb && \
-    wget https://vitastor.io/archive/qemu/qemu-trixie-9.2.2%2Bds-1%2Bvitastor4/qemu-utils_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
-    wget https://vitastor.io/archive/qemu/qemu-trixie-9.2.2%2Bds-1%2Bvitastor4/qemu-block-extra_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-utils_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-block-extra_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
     dpkg -x qemu-utils*.deb tmp1 && \
     dpkg -x qemu-block-extra*.deb tmp1 && \
     cp -a tmp1/usr/bin/qemu-storage-daemon /usr/bin/ && \

csi/Makefile

@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.12
+VITASTOR_VERSION ?= v3.0.0
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v3.0.12
+          image: vitalif/vitastor-csi:v3.0.0
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -121,7 +121,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v3.0.12
+          image: vitalif/vitastor-csi:v3.0.0
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "3.0.12"
+    vitastorCSIDriverVersion = "3.0.0"
 )
 
 // Config struct fills the parameters of request or user input

debian/build-vitastor-ubuntu-resolute.sh

@@ -1,5 +0,0 @@
-#!/bin/bash
-# 26.04 Resolute Raccoon
-
-docker build --build-arg DISTRO=ubuntu --build-arg REL=resolute -t vitastor-buildenv:resolute -f vitastor-buildenv.Dockerfile .
-docker run -it --rm -e REL=resolute -v `dirname $0`/../:/root/vitastor vitastor-buildenv:resolute /root/vitastor/debian/vitastor-build.sh

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (3.0.12-1) unstable; urgency=medium
+vitastor (3.0.0-1) unstable; urgency=medium
 
   * Bugfixes
 

debian/vitastor-build.sh

@@ -44,7 +44,7 @@ curl -s https://git.yourcmc.ru/vitalif/antietcd/archive/master.tar.gz | tar -zx
 curl -s https://git.yourcmc.ru/vitalif/tinyraft/archive/master.tar.gz | tar -zx
 
 cd /root/vitastor/packages/vitastor-$REL
-if [[ ( "$REL" = "trixie" || "$REL" = "resolute" ) && -e ../vitastor-bookworm/vitastor_$VER.orig.tar.xz ]]; then
+if [[ "$REL" = "trixie" && -e ../vitastor-bookworm/vitastor_$VER.orig.tar.xz ]]; then
     # Fucking shit, archives differ between bookworm (xz 5.4.1) and trixie (xz 5.8.1)
     cp ../vitastor-bookworm/vitastor_$VER.orig.tar.xz .
 else

docker/Dockerfile

@@ -1,6 +1,6 @@
 # Build Docker image with Vitastor packages
 
-FROM debian:trixie
+FROM debian:bookworm
 
 ADD etc/apt /etc/apt/
 RUN apt-get update && apt-get -y install vitastor ibverbs-providers udev systemd qemu-system-x86 qemu-system-common qemu-block-extra qemu-utils jq nfs-common && apt-get clean

docker/Makefile

@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.12
+VITASTOR_VERSION ?= v3.0.0
 
 all: build push
 

docker/etc/apt/preferences

@@ -1,3 +1,3 @@
 Package: *
-Pin: release n=trixie-backports
+Pin: release n=bookworm-backports
 Pin-Priority: 500

docker/etc/apt/sources.list.d/vitastor.list

@@ -1,2 +1,2 @@
-deb http://vitastor.io/debian trixie main
-#deb http://http.debian.net/debian/ trixie-backports main
+deb http://vitastor.io/debian bookworm main
+deb http://http.debian.net/debian/ bookworm-backports main

docker/etc/systemd/system/vitastor-host.service

@@ -7,7 +7,7 @@ PartOf=vitastor.target
 [Service]
 Restart=always
 EnvironmentFile=/etc/vitastor/docker.conf
-ExecStart=bash -c 'docker run --rm -i -v /etc/vitastor:/etc/vitastor -v /dev:/dev -v /run:/run -e SYSTEMD_IN_CHROOT=0 \
+ExecStart=bash -c 'docker run --rm -i -v /etc/vitastor:/etc/vitastor -v /dev:/dev -v /run:/run \
     --security-opt seccomp=unconfined --privileged --pid=host --log-driver none --network host --name vitastor vitastor:$VITASTOR_VERSION \
     sleep.sh'
 ExecStartPost=udevadm trigger

docker/etc/vitastor/docker.conf

@@ -4,7 +4,7 @@
 #
 
 # Desired Vitastor version
-VITASTOR_VERSION=v3.0.12
+VITASTOR_VERSION=v3.0.0
 
 # Additional arguments for all containers
 # For example, you may want to specify a custom logging driver here

docker/install.sh

@@ -2,7 +2,8 @@
 
 set -e
 
-cp -urv /etc/systemd/system/vitastor* /host-etc/systemd/system/
-cp -urv /etc/udev/rules.d /host-etc/udev/
+cp -urv /etc/default /host-etc/
+cp -urv /etc/systemd /host-etc/
+cp -urv /etc/udev /host-etc/
 cp -urnv /etc/vitastor /host-etc/
 cp -urnv /opt/scripts/* /host-bin/

docs/config/network.en.md

@@ -22,6 +22,7 @@ between clients, OSDs and etcd.
 - [rdma_max_msg](#rdma_max_msg)
 - [rdma_max_recv](#rdma_max_recv)
 - [rdma_max_send](#rdma_max_send)
+- [rdma_odp](#rdma_odp)
 - [peer_connect_interval](#peer_connect_interval)
 - [peer_connect_timeout](#peer_connect_timeout)
 - [osd_idle_timeout](#osd_idle_timeout)
@@ -101,6 +102,11 @@ found or if `osd_network` is not specified. Auto-selection is also
 unsupported with old libibverbs < v32, like in Debian 10 Buster or
 CentOS 7.
 
+Vitastor supports all adapters, even ones without ODP support, like
+Mellanox ConnectX-3 and non-Mellanox cards. Versions up to Vitastor
+1.2.0 required ODP which is only present in Mellanox ConnectX >= 4.
+See also [rdma_odp](#rdma_odp).
+
 Run `ibv_devinfo -v` as root to list available RDMA devices and their
 features.
 
@@ -110,23 +116,6 @@ the manual of your network vendor for details about setting up the switch
 for RoCEv2 correctly. Usually it means setting up Lossless Ethernet with
 PFC (Priority Flow Control) and ECN (Explicit Congestion Notification).
 
-Vitastor supports all adapters, even ones without ODP (On-Demand Paging)
-support, like Mellanox ConnectX-3 and non-Mellanox cards. ODP is only present
-in Mellanox ConnectX >= 4 adapters and allows to skip memory registration
-for RDMA and thus, in theory, avoid memory copying.
-
-Versions up to Vitastor 1.2.0 required ODP, then it was disabled by default,
-but it was still supported up to 3.0.3. Now ODP support is removed because it
-actually only hurts performance: an example 3-node cluster with 8 NVMe in each
-node and 2*25 GBit/s ConnectX-6 RDMA network pushed 3950000 read iops without
-ODP, but only 239000 iops with ODP.
-
-This happens because Mellanox ODP implementation seems to be based on
-message retransmissions when the adapter doesn't know about the buffer yet -
-it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
-which is generally slow in RDMA/RoCE networks. Here's a presentation about
-it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
-
 ## rdma_port_num
 
 - Type: integer
@@ -198,6 +187,28 @@ less than `rdma_max_recv` so the receiving side doesn't run out of buffers.
 Doesn't affect memory usage - additional memory isn't allocated for send
 operations.
 
+## rdma_odp
+
+- Type: boolean
+- Default: false
+
+Use RDMA with On-Demand Paging. ODP is currently only available on Mellanox
+ConnectX-4 and newer adapters. ODP allows to not register memory explicitly
+for RDMA adapter to be able to use it. This, in turn, allows to skip memory
+copying during sending. One would think this should improve performance, but
+**in reality** RDMA performance with ODP is **drastically** worse. Example
+3-node cluster with 8 NVMe in each node and 2*25 GBit/s ConnectX-6 RDMA network
+without ODP pushes 3950000 read iops, but only 239000 iops with ODP...
+
+This happens because Mellanox ODP implementation seems to be based on
+message retransmissions when the adapter doesn't know about the buffer yet -
+it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
+which is generally slow in RDMA/RoCE networks. Here's a presentation about
+it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
+
+ODP support is retained in the code just in case a good ODP implementation
+appears one day.
+
 ## peer_connect_interval
 
 - Type: seconds

docs/config/network.ru.md

@@ -22,6 +22,7 @@
 - [rdma_max_msg](#rdma_max_msg)
 - [rdma_max_recv](#rdma_max_recv)
 - [rdma_max_send](#rdma_max_send)
+- [rdma_odp](#rdma_odp)
 - [peer_connect_interval](#peer_connect_interval)
 - [peer_connect_timeout](#peer_connect_timeout)
 - [osd_idle_timeout](#osd_idle_timeout)
@@ -100,6 +101,12 @@ RoCEv1/RoCEv2, и даже позволяет полностью отключи
 не задана. Также автовыбор не поддерживается со старыми версиями библиотеки
 libibverbs < v32, например в Debian 10 Buster или CentOS 7.
 
+Vitastor поддерживает все модели адаптеров, включая те, у которых
+нет поддержки ODP, то есть вы можете использовать RDMA с ConnectX-3 и
+картами производства не Mellanox. Версии Vitastor до 1.2.0 включительно
+требовали ODP, который есть только на Mellanox ConnectX 4 и более новых.
+См. также [rdma_odp](#rdma_odp).
+
 Запустите `ibv_devinfo -v` от имени суперпользователя, чтобы посмотреть
 список доступных RDMA-устройств, их параметры и возможности.
 
@@ -110,24 +117,6 @@ libibverbs < v32, например в Debian 10 Buster или CentOS 7.
 подразумевает настройку сети без потерь на основе PFC (Priority Flow
 Control) и ECN (Explicit Congestion Notification).
 
-Vitastor поддерживает все модели адаптеров, включая те, у которых нет
-поддержки ODP (On-Demand Paging), например, ConnectX-3 и карты производства
-не Mellanox. Функция ODP доступна только на адаптерах Mellanox ConnectX-4 и
-более новых и позволяет не регистрировать память для её использования RDMA-картой,
-благодаря чему в теории можно избежать лишних копирований памяти.
-
-Версии Vitastor до 1.2.0 включительно требовали ODP, потом функция был отключена
-по умолчанию, но поддерживалась вплоть до версии 3.0.3. Сейчас поддержка ODP
-полностью удалена, так как на самом деле она только портит производительность:
-например, на 3-узловом кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на
-чтение с RDMA без ODP удаётся снять 3950000 iops, а с ODP - всего 239000 iops.
-
-Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
-основана на повторной передаче сообщений, когда карте не известен буфер -
-вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
-А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
-Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
-
 ## rdma_port_num
 
 - Тип: целое число
@@ -203,6 +192,29 @@ OSD в любом случае согласовывают реальное зн
 Не влияет на потребление памяти - дополнительная память на операции отправки
 не выделяется.
 
+## rdma_odp
+
+- Тип: булево (да/нет)
+- Значение по умолчанию: false
+
+Использовать RDMA с On-Demand Paging. ODP - функция, доступная пока что
+исключительно на адаптерах Mellanox ConnectX-4 и более новых. ODP позволяет
+не регистрировать память для её использования RDMA-картой. Благодаря этому
+можно не копировать данные при отправке их в сеть и, казалось бы, это должно
+улучшать производительность - но **по факту** получается так, что
+производительность только ухудшается, причём сильно. Пример - на 3-узловом
+кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на чтение с RDMA без ODP
+удаётся снять 3950000 iops, а с ODP - всего 239000 iops...
+
+Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
+основана на повторной передаче сообщений, когда карте не известен буфер -
+вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
+А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
+Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
+
+Возможность использования ODP сохранена в коде на случай, если вдруг в один
+прекрасный день появится хорошая реализация ODP.
+
 ## peer_connect_interval
 
 - Тип: секунды

docs/config/osd.en.md

@@ -38,7 +38,6 @@ with an OSD restart or, for some of them, even without restarting by updating co
 - [journal_io](#journal_io)
 - [journal_sector_buffer_count](#journal_sector_buffer_count)
 - [journal_no_same_sector_overwrites](#journal_no_same_sector_overwrites)
-- [skip_corrupted_meta_entries](#skip_corrupted_meta_entries)
 - [throttle_small_writes](#throttle_small_writes)
 - [throttle_target_iops](#throttle_target_iops)
 - [throttle_target_mbs](#throttle_target_mbs)
@@ -68,9 +67,6 @@ with an OSD restart or, for some of them, even without restarting by updating co
 - [pg_lock_retry_interval_ms](#pg_lock_retry_interval_ms)
 - [atomic_write_size](#atomic_write_size)
 - [use_atomic_flag](#use_atomic_flag)
-- [pg_reshard_chunk_size](#pg_reshard_chunk_size)
-- [pg_reshard_chunk_pause_ms](#pg_reshard_chunk_pause_ms)
-- [gc_on_start](#gc_on_start)
 
 ## bind_address
 
@@ -281,19 +277,13 @@ Maximum number of journal flushers (see above min_flusher_count).
 - Type: boolean
 - Default: true
 
-Only for the old store ([meta_format](layout-osd.en.md#meta_format) 2).
-
-This parameter makes Vitastor keep a copy of metadata area in memory as it is
-on disk, in addition to the metadata database. When the option is enabled, every
-metadata entry is effectively stored in RAM twice. It's required for good performance
-because it allows to avoid additional read-modify-write cycles during metadata
-modifications. Metadata area size with the old store is roughly 224 MB per 1 TB
-of data. You can turn the option off to reduce memory usage by this value, but
-it will reduce performance.
-
-For the new store ([meta_format](layout-osd.en.md#meta_format) 3), the option
-may be changed in the future to support operation without loading full metadata
-database in memory.
+This parameter makes Vitastor always keep metadata area of the block device
+in memory. It's required for good performance because it allows to avoid
+additional read-modify-write cycles during metadata modifications. Metadata
+area size is currently roughly 224 MB per 1 TB of data. You can turn it off
+to reduce memory usage by this value, but it will hurt performance. This
+restriction is likely to be removed in the future along with the upgrade
+of the metadata storage scheme.
 
 ## inmemory_journal
 
@@ -372,8 +362,6 @@ blocks. The only situation when you should increase it to a larger value
 is when you enable journal_no_same_sector_overwrites. In this case set
 it to, for example, 1024.
 
-Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
-
 ## journal_no_same_sector_overwrites
 
 - Type: boolean
@@ -387,17 +375,6 @@ journal after writing it instead of possibly overwriting it the second time.
 
 Most (99%) other SSDs don't need this option.
 
-Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
-
-## skip_corrupted_meta_entries
-
-- Type: boolean
-- Default: false
-
-Only for the new store ([meta_format](layout-osd.en.md#meta_format) 3).
-Allow OSD to start when some metadata entries or blocks are corrupted by
-skipping them. Should be only used as an emergency measure.
-
 ## throttle_small_writes
 
 - Type: boolean
@@ -705,10 +682,7 @@ with replicated pools and reach the best possible write performance.
 
 Default value is auto-detected during OSD initialization from
 `/sys/block/xx/queue/atomic_write_max_bytes` or assumed to be 4096 bytes
-because all known disks support 4 KB atomic writes. Auto-detection is only used for
-NVMe disks because SAS disks require the explicit WRITE ATOMIC command which requires
-RWF_ATOMIC (see below [#use_atomic_flag]) but that flag works incorrectly in current
-Linux versions.
+because all known disks support 4 KB atomic writes.
 
 You can also check if your NVMe drives support atomic writes by running
 the command `nvme id-ctrl /dev/nvme0n1 | grep awupf`. If the reported value,
@@ -723,40 +697,12 @@ reducing Write Amplification and improving write performance up to 2 times.
 
 - Type: boolean
 
-This option controls whether Vitastor OSDs use RWF_ATOMIC write flag with atomic writes.
-This flag is supported since Linux 6.11 and adds some safety to atomic writes - the kernel
-guarantees to not fragment write requests with it and also to check them against the actual
-device atomic write capabilities.
+This option controls whether the Vitastor OSD uses RWF_ATOMIC write flag with atomic
+writes. This flag is only supported on Linux kernel since 6.11. Atomic writes are
+generally only safe to use with this flag because it tells the kernel to never fragment
+write requests and also to check the write against the actual atomic write capabilities
+of the device.
 
-However, the option is disabled by default because the flag is currently UNUSABLE - Linux
-incorrectly requires writes with that flag to be of power-of-2 length and length-aligned.
-I.e., for example, 12 KB writes and not-8-KB aligned 8 KB writes are forbidden by the kernel,
-even though the NVMe specification allows them.
-
-For NVMe disks with `scheduler=none` writes aren't fragmented anyway so it's not a big deal.
-However, you can rebuild your kernel with [this patch](../../patches/linux-fix-atomic-write-checks.diff)
-and turn this option on. It will make your atomic writes a bit safer.
-
-## pg_reshard_chunk_size
-
-- Type: integer
-- Default: 100000
-
-Pool PG count change is a CPU-intensive operation because OSDs store the full object database
-in memory and have to move all entries between old and new PGs. Thus it's performed in chunks,
-with pauses between chunks to prevent blocking OSD's event loop and other clients' operations.
-This option sets the maximum number of object is a chunk. Moving 100k objects usually takes
-50-100ms. Chunk size equal to 0 means unlimited.
-
-## pg_reshard_chunk_pause_ms
-
-- Type: milliseconds
-- Default: 100
-
-This option sets the interval between handling two PG count change chunks.
-
-## gc_on_start
-
-- Type: boolean
-
-Forcibly clean all garbage entries in the new store on every OSD restart.
+This option is enabled by default when atomic_write_size is set to a value larger than 4 KB.
+You can disable it if you're sure that your disks support atomic writes and you want to
+bypass the Linux atomic write checks.

docs/config/osd.ru.md

@@ -39,7 +39,6 @@
 - [journal_io](#journal_io)
 - [journal_sector_buffer_count](#journal_sector_buffer_count)
 - [journal_no_same_sector_overwrites](#journal_no_same_sector_overwrites)
-- [skip_corrupted_meta_entries](#skip_corrupted_meta_entries)
 - [throttle_small_writes](#throttle_small_writes)
 - [throttle_target_iops](#throttle_target_iops)
 - [throttle_target_mbs](#throttle_target_mbs)
@@ -69,9 +68,6 @@
 - [pg_lock_retry_interval_ms](#pg_lock_retry_interval_ms)
 - [atomic_write_size](#atomic_write_size)
 - [use_atomic_flag](#use_atomic_flag)
-- [pg_reshard_chunk_size](#pg_reshard_chunk_size)
-- [pg_reshard_chunk_pause_ms](#pg_reshard_chunk_pause_ms)
-- [gc_on_start](#gc_on_start)
 
 ## bind_address
 
@@ -289,19 +285,13 @@ Flusher - это микро-поток (корутина), которая коп
 - Тип: булево (да/нет)
 - Значение по умолчанию: true
 
-Только для старого хранилища ([meta_format](layout-osd.en.md#meta_format) 2).
-
-Данный параметр заставляет Vitastor всегда держать копию области метаданных
-в памяти в том же виде, как она лежит на диске, в дополнение к БД метаданных.
-То есть, с включённой опцией каждая запись метаданных хранится в памяти дважды.
-Это нужно, чтобы избегать дополнительных операций чтения с диска при записи.
-Размер области метаданных в старом хранилище составляет примерно 224 МБ на
-1 ТБ данных. Вы можете отключить опцию, чтобы снизить потребление памяти
-примерно на эту величину, но при этом также снизится и производительность.
-
-Для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3) опция,
-возможно, будет переработана в будущем для поддержки работы без полной
-загрузки метаданных в памяти.
+Данный параметр заставляет Vitastor всегда держать область метаданных диска
+в памяти. Это нужно, чтобы избегать дополнительных операций чтения с диска
+при записи. Размер области метаданных на данный момент составляет примерно
+224 МБ на 1 ТБ данных. При включении потребление памяти снизится примерно
+на эту величину, но при этом также снизится и производительность. В будущем,
+после обновления схемы хранения метаданных, это ограничение, скорее всего,
+будет ликвидировано.
 
 ## inmemory_journal
 
@@ -384,8 +374,6 @@ fsync небезопасным даже с режимом "directsync".
 нужно менять - это если вы включаете journal_no_same_sector_overwrites. В
 этом случае установите данный параметр, например, в 1024.
 
-Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
-
 ## journal_no_same_sector_overwrites
 
 - Тип: булево (да/нет)
@@ -401,18 +389,6 @@ fsync небезопасным даже с режимом "directsync".
 
 Почти все другие SSD (99% моделей) не требуют данной опции.
 
-Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
-
-## skip_corrupted_meta_entries
-
-- Тип: булево (да/нет)
-- Значение по умолчанию: false
-
-Только для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3).
-Разрешить OSD запускаться, даже если часть блоков или записей метаданных
-повреждена, пропуская их. Опция предназначена для использования только в
-целях аварийного восстановления.
-
 ## throttle_small_writes
 
 - Тип: булево (да/нет)
@@ -741,9 +717,6 @@ pg_minsize OSD во время переключений, что может по
 Значение по умолчанию авто-определяется во время инициализации OSD из
 `/sys/block/xx/queue/atomic_write_max_bytes` либо принимается равным 4096,
 так как все известные диски поддерживают атомарную запись 4 КБ блоков.
-Автоопределение применяется только для NVMe-дисков, так как SAS диски требуют
-использования отдельной команды WRITE ATOMIC, а для неё нужен флаг RWF_ATOMIC
-(см. ниже [#use_atomic_flag]), а он в текущих версиях Linux работает некорректно.
 
 Вы также можете проверить, поддерживают ли ваши NVMe-диски атомарную запись,
 с помощью команды `nvme id-ctrl /dev/nvme0n1 | grep awupf`. Если значение awupf
@@ -762,41 +735,11 @@ pg_minsize OSD во время переключений, что может по
 - Тип: булево (да/нет)
 
 Данная опция контролирует использование Vitastor OSD флага RWF_ATOMIC при атомарной записи
-блоков. Этот флаг поддерживается, начиная с версии ядра Linux 6.11 и добавляет немного корректности
-атомарным записям - ядро гарантирует отсутствие фрагментации запросов записи с этим флагом и
-проверяет их на соответствие реальным возможностям устройства.
+блоков. Этот флаг поддерживается только в ядрах Linux начиная с 6.11. Атомарная запись
+является безопасной только при использовании этого флага, так как он сообщает ядру о том,
+что запрос записи нельзя фрагментировать и о том, что запрос нужно проверить на соответствие
+реальным возможностям атомарной записи устройства.
 
-Однако, данная опция по умолчанию отключена, так как флаг в текущих версиях Linux работает
-абсолютно НЕКОРРЕКТНО - при нём Linux требует, чтобы запросы записи имели длину, равную
-степени двойки и были выровнены на эту длину. То есть, например, 12 КБ запросы записи, а также
-8 КБ запросы записи по не-кратному 8 КБ смещению запрещаются ядром, хотя спецификация NVMe их
-разрешает.
-
-Для NVMe-дисков с `scheduler=none` запросы записи и так не фрагментируются, так что это не так
-уж и важно, однако вы можете пересобрать своё ядро с [этим патчем](../../patches/linux-fix-atomic-write-checks.diff)
-и включить данную опцию. Это сделает вашу атомарную запись капельку безопаснее.
-
-## pg_reshard_chunk_size
-
-- Тип: целое число
-- Значение по умолчанию: 100000
-
-Изменение числа PG в пуле заметно загружает процессор, так как OSD хранят полную базу данных
-объектов в памяти и им приходится перемещать все записи объектов между старыми и новыми PG.
-Поэтому изменение применяется порциями, с паузами между порциями, чтобы не блокировать обработку
-событий OSD и операции остальных клиентов. Данная опция задаёт максимальное число объектов
-в порции. Перемещение 100 тысяч объектов (значение по умолчанию) обычно занимает порядка
-50-100 миллисекунд. Значение опции 0 отключает лимит размера порции.
-
-## pg_reshard_chunk_pause_ms
-
-- Тип: миллисекунды
-- Значение по умолчанию: 100
-
-Данная опция задаёт интервал между обработкой двух порций изменения числа PG пулов.
-
-## gc_on_start
-
-- Тип: булево (да/нет)
-
-Принудительно очищать все мусорные записи в новом хранилище при каждом запуске OSD.
+Опция включается по умолчанию, когда atomic_write_size устанавливается в значение больше 4 КБ.
+Вы можете явно отключить её, если уверены, что ваши диски поддерживают атомарную запись и
+хотите обойти проверки уровня ядра.

docs/config/src/network.yml

@@ -84,6 +84,11 @@
     unsupported with old libibverbs < v32, like in Debian 10 Buster or
     CentOS 7.
 
+    Vitastor supports all adapters, even ones without ODP support, like
+    Mellanox ConnectX-3 and non-Mellanox cards. Versions up to Vitastor
+    1.2.0 required ODP which is only present in Mellanox ConnectX >= 4.
+    See also [rdma_odp](#rdma_odp).
+
     Run `ibv_devinfo -v` as root to list available RDMA devices and their
     features.
 
@@ -92,23 +97,6 @@
     the manual of your network vendor for details about setting up the switch
     for RoCEv2 correctly. Usually it means setting up Lossless Ethernet with
     PFC (Priority Flow Control) and ECN (Explicit Congestion Notification).
-
-    Vitastor supports all adapters, even ones without ODP (On-Demand Paging)
-    support, like Mellanox ConnectX-3 and non-Mellanox cards. ODP is only present
-    in Mellanox ConnectX >= 4 adapters and allows to skip memory registration
-    for RDMA and thus, in theory, avoid memory copying.
-
-    Versions up to Vitastor 1.2.0 required ODP, then it was disabled by default,
-    but it was still supported up to 3.0.3. Now ODP support is removed because it
-    actually only hurts performance: an example 3-node cluster with 8 NVMe in each
-    node and 2*25 GBit/s ConnectX-6 RDMA network pushed 3950000 read iops without
-    ODP, but only 239000 iops with ODP.
-
-    This happens because Mellanox ODP implementation seems to be based on
-    message retransmissions when the adapter doesn't know about the buffer yet -
-    it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
-    which is generally slow in RDMA/RoCE networks. Here's a presentation about
-    it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
   info_ru: |
     Название RDMA-устройства для связи с Vitastor OSD (например, "rocep5s0f0").
     Если не указано, Vitastor попробует найти RoCE-устройство, соответствующее
@@ -117,6 +105,12 @@
     не задана. Также автовыбор не поддерживается со старыми версиями библиотеки
     libibverbs < v32, например в Debian 10 Buster или CentOS 7.
 
+    Vitastor поддерживает все модели адаптеров, включая те, у которых
+    нет поддержки ODP, то есть вы можете использовать RDMA с ConnectX-3 и
+    картами производства не Mellanox. Версии Vitastor до 1.2.0 включительно
+    требовали ODP, который есть только на Mellanox ConnectX 4 и более новых.
+    См. также [rdma_odp](#rdma_odp).
+
     Запустите `ibv_devinfo -v` от имени суперпользователя, чтобы посмотреть
     список доступных RDMA-устройств, их параметры и возможности.
 
@@ -126,24 +120,6 @@
     коммутатора для RoCEv2 ищите в документации производителя. Обычно это
     подразумевает настройку сети без потерь на основе PFC (Priority Flow
     Control) и ECN (Explicit Congestion Notification).
-
-    Vitastor поддерживает все модели адаптеров, включая те, у которых нет
-    поддержки ODP (On-Demand Paging), например, ConnectX-3 и карты производства
-    не Mellanox. Функция ODP доступна только на адаптерах Mellanox ConnectX-4 и
-    более новых и позволяет не регистрировать память для её использования RDMA-картой,
-    благодаря чему в теории можно избежать лишних копирований памяти.
-
-    Версии Vitastor до 1.2.0 включительно требовали ODP, потом функция был отключена
-    по умолчанию, но поддерживалась вплоть до версии 3.0.3. Сейчас поддержка ODP
-    полностью удалена, так как на самом деле она только портит производительность:
-    например, на 3-узловом кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на
-    чтение с RDMA без ODP удаётся снять 3950000 iops, а с ODP - всего 239000 iops.
-
-    Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
-    основана на повторной передаче сообщений, когда карте не известен буфер -
-    вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
-    А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
-    Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
 - name: rdma_port_num
   type: int
   info: |
@@ -242,6 +218,45 @@
     у принимающей стороны в процессе работы не заканчивались буферы на приём.
     Не влияет на потребление памяти - дополнительная память на операции отправки
     не выделяется.
+- name: rdma_odp
+  type: bool
+  default: false
+  online: false
+  info: |
+    Use RDMA with On-Demand Paging. ODP is currently only available on Mellanox
+    ConnectX-4 and newer adapters. ODP allows to not register memory explicitly
+    for RDMA adapter to be able to use it. This, in turn, allows to skip memory
+    copying during sending. One would think this should improve performance, but
+    **in reality** RDMA performance with ODP is **drastically** worse. Example
+    3-node cluster with 8 NVMe in each node and 2*25 GBit/s ConnectX-6 RDMA network
+    without ODP pushes 3950000 read iops, but only 239000 iops with ODP...
+
+    This happens because Mellanox ODP implementation seems to be based on
+    message retransmissions when the adapter doesn't know about the buffer yet -
+    it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
+    which is generally slow in RDMA/RoCE networks. Here's a presentation about
+    it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
+
+    ODP support is retained in the code just in case a good ODP implementation
+    appears one day.
+  info_ru: |
+    Использовать RDMA с On-Demand Paging. ODP - функция, доступная пока что
+    исключительно на адаптерах Mellanox ConnectX-4 и более новых. ODP позволяет
+    не регистрировать память для её использования RDMA-картой. Благодаря этому
+    можно не копировать данные при отправке их в сеть и, казалось бы, это должно
+    улучшать производительность - но **по факту** получается так, что
+    производительность только ухудшается, причём сильно. Пример - на 3-узловом
+    кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на чтение с RDMA без ODP
+    удаётся снять 3950000 iops, а с ODP - всего 239000 iops...
+
+    Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
+    основана на повторной передаче сообщений, когда карте не известен буфер -
+    вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
+    А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
+    Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
+
+    Возможность использования ODP сохранена в коде на случай, если вдруг в один
+    прекрасный день появится хорошая реализация ODP.
 - name: peer_connect_interval
   type: sec
   min: 1

docs/config/src/osd.yml

@@ -253,33 +253,21 @@
   type: bool
   default: true
   info: |
-    Only for the old store ([meta_format](layout-osd.en.md#meta_format) 2).
-
-    This parameter makes Vitastor keep a copy of metadata area in memory as it is
-    on disk, in addition to the metadata database. When the option is enabled, every
-    metadata entry is effectively stored in RAM twice. It's required for good performance
-    because it allows to avoid additional read-modify-write cycles during metadata
-    modifications. Metadata area size with the old store is roughly 224 MB per 1 TB
-    of data. You can turn the option off to reduce memory usage by this value, but
-    it will reduce performance.
-
-    For the new store ([meta_format](layout-osd.en.md#meta_format) 3), the option
-    may be changed in the future to support operation without loading full metadata
-    database in memory.
+    This parameter makes Vitastor always keep metadata area of the block device
+    in memory. It's required for good performance because it allows to avoid
+    additional read-modify-write cycles during metadata modifications. Metadata
+    area size is currently roughly 224 MB per 1 TB of data. You can turn it off
+    to reduce memory usage by this value, but it will hurt performance. This
+    restriction is likely to be removed in the future along with the upgrade
+    of the metadata storage scheme.
   info_ru: |
-    Только для старого хранилища ([meta_format](layout-osd.en.md#meta_format) 2).
-
-    Данный параметр заставляет Vitastor всегда держать копию области метаданных
-    в памяти в том же виде, как она лежит на диске, в дополнение к БД метаданных.
-    То есть, с включённой опцией каждая запись метаданных хранится в памяти дважды.
-    Это нужно, чтобы избегать дополнительных операций чтения с диска при записи.
-    Размер области метаданных в старом хранилище составляет примерно 224 МБ на
-    1 ТБ данных. Вы можете отключить опцию, чтобы снизить потребление памяти
-    примерно на эту величину, но при этом также снизится и производительность.
-
-    Для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3) опция,
-    возможно, будет переработана в будущем для поддержки работы без полной
-    загрузки метаданных в памяти.
+    Данный параметр заставляет Vitastor всегда держать область метаданных диска
+    в памяти. Это нужно, чтобы избегать дополнительных операций чтения с диска
+    при записи. Размер области метаданных на данный момент составляет примерно
+    224 МБ на 1 ТБ данных. При включении потребление памяти снизится примерно
+    на эту величину, но при этом также снизится и производительность. В будущем,
+    после обновления схемы хранения метаданных, это ограничение, скорее всего,
+    будет ликвидировано.
 - name: inmemory_journal
   type: bool
   default: true
@@ -398,15 +386,11 @@
     blocks. The only situation when you should increase it to a larger value
     is when you enable journal_no_same_sector_overwrites. In this case set
     it to, for example, 1024.
-
-    Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
   info_ru: |
     Максимальное число буферов, разрешённых для использования под записываемые
     в журнал блоки метаданных. Единственная ситуация, в которой этот параметр
     нужно менять - это если вы включаете journal_no_same_sector_overwrites. В
     этом случае установите данный параметр, например, в 1024.
-
-    Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
 - name: journal_no_same_sector_overwrites
   type: bool
   default: false
@@ -418,8 +402,6 @@
     journal after writing it instead of possibly overwriting it the second time.
 
     Most (99%) other SSDs don't need this option.
-
-    Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
   info_ru: |
     Включайте данную опцию для SSD вроде Intel D3-S4510 и D3-S4610, которые
     ОЧЕНЬ не любят, когда ПО перезаписывает один и тот же сектор несколько раз
@@ -430,20 +412,6 @@
     самого сектора.
 
     Почти все другие SSD (99% моделей) не требуют данной опции.
-
-    Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
-- name: skip_corrupted_meta_entries
-  type: bool
-  default: false
-  info: |
-    Only for the new store ([meta_format](layout-osd.en.md#meta_format) 3).
-    Allow OSD to start when some metadata entries or blocks are corrupted by
-    skipping them. Should be only used as an emergency measure.
-  info_ru: |
-    Только для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3).
-    Разрешить OSD запускаться, даже если часть блоков или записей метаданных
-    повреждена, пропуская их. Опция предназначена для использования только в
-    целях аварийного восстановления.
 - name: throttle_small_writes
   type: bool
   default: false
@@ -845,10 +813,7 @@
 
     Default value is auto-detected during OSD initialization from
     `/sys/block/xx/queue/atomic_write_max_bytes` or assumed to be 4096 bytes
-    because all known disks support 4 KB atomic writes. Auto-detection is only used for
-    NVMe disks because SAS disks require the explicit WRITE ATOMIC command which requires
-    RWF_ATOMIC (see below [#use_atomic_flag]) but that flag works incorrectly in current
-    Linux versions.
+    because all known disks support 4 KB atomic writes.
 
     You can also check if your NVMe drives support atomic writes by running
     the command `nvme id-ctrl /dev/nvme0n1 | grep awupf`. If the reported value,
@@ -869,9 +834,6 @@
     Значение по умолчанию авто-определяется во время инициализации OSD из
     `/sys/block/xx/queue/atomic_write_max_bytes` либо принимается равным 4096,
     так как все известные диски поддерживают атомарную запись 4 КБ блоков.
-    Автоопределение применяется только для NVMe-дисков, так как SAS диски требуют
-    использования отдельной команды WRITE ATOMIC, а для неё нужен флаг RWF_ATOMIC
-    (см. ниже [#use_atomic_flag]), а он в текущих версиях Linux работает некорректно.
 
     Вы также можете проверить, поддерживают ли ваши NVMe-диски атомарную запись,
     с помощью команды `nvme id-ctrl /dev/nvme0n1 | grep awupf`. Если значение awupf
@@ -887,58 +849,22 @@
 - name: use_atomic_flag
   type: bool
   info: |
-    This option controls whether Vitastor OSDs use RWF_ATOMIC write flag with atomic writes.
-    This flag is supported since Linux 6.11 and adds some safety to atomic writes - the kernel
-    guarantees to not fragment write requests with it and also to check them against the actual
-    device atomic write capabilities.
+    This option controls whether the Vitastor OSD uses RWF_ATOMIC write flag with atomic
+    writes. This flag is only supported on Linux kernel since 6.11. Atomic writes are
+    generally only safe to use with this flag because it tells the kernel to never fragment
+    write requests and also to check the write against the actual atomic write capabilities
+    of the device.
 
-    However, the option is disabled by default because the flag is currently UNUSABLE - Linux
-    incorrectly requires writes with that flag to be of power-of-2 length and length-aligned.
-    I.e., for example, 12 KB writes and not-8-KB aligned 8 KB writes are forbidden by the kernel,
-    even though the NVMe specification allows them.
-
-    For NVMe disks with `scheduler=none` writes aren't fragmented anyway so it's not a big deal.
-    However, you can rebuild your kernel with [this patch](../../patches/linux-fix-atomic-write-checks.diff)
-    and turn this option on. It will make your atomic writes a bit safer.
+    This option is enabled by default when atomic_write_size is set to a value larger than 4 KB.
+    You can disable it if you're sure that your disks support atomic writes and you want to
+    bypass the Linux atomic write checks.
   info_ru: |
     Данная опция контролирует использование Vitastor OSD флага RWF_ATOMIC при атомарной записи
-    блоков. Этот флаг поддерживается, начиная с версии ядра Linux 6.11 и добавляет немного корректности
-    атомарным записям - ядро гарантирует отсутствие фрагментации запросов записи с этим флагом и
-    проверяет их на соответствие реальным возможностям устройства.
+    блоков. Этот флаг поддерживается только в ядрах Linux начиная с 6.11. Атомарная запись
+    является безопасной только при использовании этого флага, так как он сообщает ядру о том,
+    что запрос записи нельзя фрагментировать и о том, что запрос нужно проверить на соответствие
+    реальным возможностям атомарной записи устройства.
 
-    Однако, данная опция по умолчанию отключена, так как флаг в текущих версиях Linux работает
-    абсолютно НЕКОРРЕКТНО - при нём Linux требует, чтобы запросы записи имели длину, равную
-    степени двойки и были выровнены на эту длину. То есть, например, 12 КБ запросы записи, а также
-    8 КБ запросы записи по не-кратному 8 КБ смещению запрещаются ядром, хотя спецификация NVMe их
-    разрешает.
-
-    Для NVMe-дисков с `scheduler=none` запросы записи и так не фрагментируются, так что это не так
-    уж и важно, однако вы можете пересобрать своё ядро с [этим патчем](../../patches/linux-fix-atomic-write-checks.diff)
-    и включить данную опцию. Это сделает вашу атомарную запись капельку безопаснее.
-- name: pg_reshard_chunk_size
-  type: int
-  default: 100000
-  info: |
-    Pool PG count change is a CPU-intensive operation because OSDs store the full object database
-    in memory and have to move all entries between old and new PGs. Thus it's performed in chunks,
-    with pauses between chunks to prevent blocking OSD's event loop and other clients' operations.
-    This option sets the maximum number of object is a chunk. Moving 100k objects usually takes
-    50-100ms. Chunk size equal to 0 means unlimited.
-  info_ru: |
-    Изменение числа PG в пуле заметно загружает процессор, так как OSD хранят полную базу данных
-    объектов в памяти и им приходится перемещать все записи объектов между старыми и новыми PG.
-    Поэтому изменение применяется порциями, с паузами между порциями, чтобы не блокировать обработку
-    событий OSD и операции остальных клиентов. Данная опция задаёт максимальное число объектов
-    в порции. Перемещение 100 тысяч объектов (значение по умолчанию) обычно занимает порядка
-    50-100 миллисекунд. Значение опции 0 отключает лимит размера порции.
-- name: pg_reshard_chunk_pause_ms
-  type: ms
-  default: 100
-  info: |
-    This option sets the interval between handling two PG count change chunks.
-  info_ru: |
-    Данная опция задаёт интервал между обработкой двух порций изменения числа PG пулов.
-- name: gc_on_start
-  type: bool
-  info: Forcibly clean all garbage entries in the new store on every OSD restart.
-  info_ru: Принудительно очищать все мусорные записи в новом хранилище при каждом запуске OSD.
+    Опция включается по умолчанию, когда atomic_write_size устанавливается в значение больше 4 КБ.
+    Вы можете явно отключить её, если уверены, что ваши диски поддерживают атомарную запись и
+    хотите обойти проверки уровня ядра.

docs/installation/docker.en.md

@@ -26,37 +26,13 @@ at Vitastor Kubernetes operator: https://github.com/Antilles7227/vitastor-operat
 The instruction is very simple.
 
 1. Download a Docker image of the desired version: \
-   `docker pull vitalif/vitastor:v3.0.12`
+   `docker pull vitalif/vitastor:v3.0.0`
 2. Install scripts to the host system: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.12 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.0 install.sh`
 3. Reload udev rules: \
    `udevadm control --reload-rules`
-4. Enable the vitastor-host service: \
-   `systemctl enable --now vitastor-host`
 
-After these steps, you can return to [Quick Start](../intro/quickstart.en.md).
-
-## Podman
-
-If you use Podman, run the following commands as root before installing Vitastor containers:
-
-```
-ln -s podman /usr/bin/docker
-
-mkdir -p /etc/systemd/system/systemd-udevd.service.d
-
-cat >/etc/systemd/system/systemd-udevd.service.d/override.conf <<EOF
-[Service]
-CapabilityBoundingSet=~
-SystemCallFilter=@mount capset
-EOF
-
-systemctl daemon-reload
-
-systemctl restart systemd-udevd
-```
-
-Without it, udev fails to do calls into a Podman container and Vitastor disk detection doesn't work.
+And you can return to [Quick Start](../intro/quickstart.en.md).
 
 ## Upgrading Containers
 

docs/installation/docker.ru.md

@@ -25,39 +25,14 @@ Vitastor можно установить в Docker/Podman. При этом etcd,
 Инструкция по установке максимально простая.
 
 1. Скачайте Docker-образ желаемой версии: \
-   `docker pull vitalif/vitastor:v3.0.12`
+   `docker pull vitalif/vitastor:v3.0.0`
 2. Установите скрипты в хост-систему командой: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.12 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.0 install.sh`
 3. Перезагрузите правила udev: \
    `udevadm control --reload-rules`
-4. Включите сервис vitastor-host: \
-   `systemctl enable --now vitastor-host`
 
 После этого вы можете возвращаться к разделу [Быстрый старт](../intro/quickstart.ru.md).
 
-## Podman
-
-Если вы используете Podman, перед установкой контейнеров Vitastor выполните следующие
-команды от имени суперпользователя:
-
-```
-ln -s podman /usr/bin/docker
-
-mkdir -p /etc/systemd/system/systemd-udevd.service.d
-
-cat >/etc/systemd/system/systemd-udevd.service.d/override.conf <<EOF
-[Service]
-CapabilityBoundingSet=~
-SystemCallFilter=@mount capset
-EOF
-
-systemctl daemon-reload
-
-systemctl restart systemd-udevd
-```
-
-Без этих настроек udev не может делать вызовы внутрь Podman-контейнеров и определение дисков Vitastor не работает.
-
 ## Обновление контейнеров
 
 Сначала обязательно проверьте раздел [Обновление Vitastor](../usage/admin.ru.md#обновление-vitastor),

docs/installation/packages.en.md

@@ -17,7 +17,6 @@
   - Debian 10 (Buster): `deb https://vitastor.io/debian buster main`
   - Ubuntu 22.04 (Jammy): `deb https://vitastor.io/debian jammy main`
   - Ubuntu 24.04 (Noble): `deb https://vitastor.io/debian noble main`
-  - Ubuntu 26.04 (Resolute): `deb https://vitastor.io/debian resolute main`
   - Add `-oldstable` to bookworm/bullseye/buster in this line to install the last
     stable version from 0.9.x branch instead of 1.x
 - To always prefer vitastor-patched QEMU and Libvirt versions, add the following to `/etc/apt/preferences`:
@@ -34,17 +33,15 @@
   - CentOS 7: `yum install https://vitastor.io/rpms/centos/7/vitastor-release.rpm`
   - CentOS 8: `dnf install https://vitastor.io/rpms/centos/8/vitastor-release.rpm`
   - AlmaLinux 9 and other RHEL 9 clones (Rocky, Oracle...): `dnf install https://vitastor.io/rpms/centos/9/vitastor-release.rpm`
-  - AlmaLinux 10 and other RHEL 10 clones: `dnf install https://vitastor.io/rpms/centos/10/vitastor-release.rpm`
 - Enable EPEL: `yum/dnf install epel-release`
 - Enable additional CentOS repositories:
   - CentOS 7: `yum install centos-release-scl`
   - CentOS 8: `dnf install centos-release-advanced-virtualization`
-  - RHEL 9/10 clones: not required
+  - RHEL 9 clones: not required
 - Enable elrepo-kernel:
   - CentOS 7: `yum install https://www.elrepo.org/elrepo-release-7.el7.elrepo.noarch.rpm`
   - CentOS 8: `dnf install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm`
   - RHEL 9 clones: `dnf install https://www.elrepo.org/elrepo-release-9.el9.elrepo.noarch.rpm`
-  - RHEL 10 clones: not required
 - Install packages: `yum/dnf install vitastor lpsolve etcd kernel-ml qemu-kvm`
 
 ## Installation requirements

docs/installation/packages.ru.md

@@ -17,7 +17,6 @@
   - Debian 10 (Buster): `deb https://vitastor.io/debian buster main`
   - Ubuntu 22.04 (Jammy): `deb https://vitastor.io/debian jammy main`
   - Ubuntu 24.04 (Noble): `deb https://vitastor.io/debian noble main`
-  - Ubuntu 26.04 (Resolute): `deb https://vitastor.io/debian resolute main`
   - Добавьте `-oldstable` к слову bookworm/bullseye/buster в этой строке, чтобы
     установить последнюю стабильную версию из ветки 0.9.x вместо 1.x
 - Чтобы всегда предпочитались версии пакетов QEMU и Libvirt с патчами Vitastor, добавьте в `/etc/apt/preferences`:
@@ -34,17 +33,15 @@
   - CentOS 7: `yum install https://vitastor.io/rpms/centos/7/vitastor-release.rpm`
   - CentOS 8: `dnf install https://vitastor.io/rpms/centos/8/vitastor-release.rpm`
   - AlmaLinux 9 и другие клоны RHEL 9 (Rocky, Oracle...): `dnf install https://vitastor.io/rpms/centos/9/vitastor-release.rpm`
-  - AlmaLinux 10 и другие клоны RHEL 10: `dnf install https://vitastor.io/rpms/centos/10/vitastor-release.rpm`
 - Включите EPEL: `yum/dnf install epel-release`
 - Включите дополнительные репозитории CentOS:
   - CentOS 7: `yum install centos-release-scl`
   - CentOS 8: `dnf install centos-release-advanced-virtualization`
-  - Клоны RHEL 9/10: не нужно
+  - Клоны RHEL 9: не нужно
 - Включите elrepo-kernel:
   - CentOS 7: `yum install https://www.elrepo.org/elrepo-release-7.el7.elrepo.noarch.rpm`
   - CentOS 8: `dnf install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm`
   - Клоны RHEL 9: `dnf install https://www.elrepo.org/elrepo-release-9.el9.elrepo.noarch.rpm`
-  - Клоны RHEL 10: не нужно
 - Установите пакеты: `yum/dnf install vitastor lpsolve etcd kernel-ml qemu-kvm`
 
 ## Установочные требования

docs/installation/proxmox.en.md

@@ -6,7 +6,7 @@
 
 # Proxmox VE
 
-To enable Vitastor support in Proxmox Virtual Environment (6.4-9.x are supported):
+To enable Vitastor support in Proxmox Virtual Environment (6.4-8.x are supported):
 
 - Add the corresponding Vitastor Debian repository into sources.list on Proxmox hosts:
   trixie for 9.0+, bookworm for 8.1+, pve8.0 for 8.0, bullseye for 7.4, pve7.3 for 7.3, pve7.2 for 7.2, pve7.1 for 7.1, buster for 6.4

docs/installation/proxmox.ru.md

@@ -6,7 +6,7 @@
 
 # Proxmox VE
 
-Чтобы подключить Vitastor к Proxmox Virtual Environment (поддерживаются версии 6.4-9.x):
+Чтобы подключить Vitastor к Proxmox Virtual Environment (поддерживаются версии 6.4-8.x):
 
 - Добавьте соответствующий Debian-репозиторий Vitastor в sources.list на хостах Proxmox:
   trixie для 9.0+, bookworm для 8.1+, pve8.0 для 8.0, bullseye для 7.4, pve7.3 для 7.3, pve7.2 для 7.2, pve7.1 для 7.1, buster для 6.4

docs/installation/source.en.md

@@ -16,7 +16,8 @@
   designated initializers support from C++20
 - CMake
 - jerasure headers and libraries
-- ISA-L, libibverbs, librdmacm, libnl3 headers and libraries (optional)
+- ISA-L, libibverbs and librdmacm headers and libraries (optional)
+- tcmalloc (google-perftools-dev)
 
 ## Basic instructions
 

docs/installation/source.ru.md

@@ -16,7 +16,8 @@
   назначенных инициализаторов (designated initializers) из C++20
 - CMake
 - Заголовки и библиотеки jerasure
-- Опционально - заголовки и библиотеки ISA-L, libibverbs, librdmacm, libnl3
+- Опционально - заголовки и библиотеки ISA-L, libibverbs, librdmacm
+- tcmalloc (google-perftools-dev)
 
 ## Базовая инструкция
 

docs/usage/disk.en.md

@@ -95,8 +95,6 @@ Options (single-device mode):
 Options (both modes):
 
 ```
---tags tag1,tag2           Set new OSD tag(s)
---weight <number>          Set new OSD weight (between 0 to 1)
 --journal_size 1G/32M      Set journal size (area or partition size)
 --block_size 1M/128k       Set blockstore object size
 --bitmap_granularity 4k    Set bitmap granularity

docs/usage/disk.ru.md

@@ -96,8 +96,6 @@ vitastor-disk - инструмент командной строки для уп
 Опции для обоих режимов:
 
 ```
---tags tag1,tag2           Задать теги для новых OSD
---weight <number>          Задать вес для новых OSD (от 0 до 1)
 --journal_size 1G/32M      Задать размер журнала (области или раздела журнала)
 --block_size 1M/128k       Задать размер объекта хранилища
 --bitmap_granularity 4k    Задать гранулярность битовых карт

docs/usage/nfs.en.md

@@ -262,4 +262,3 @@ Options:
 | `--logfile <FILE>` | log to the specified file                                |
 | `--enforce 1`      | enforce permissions at the server side (no by default)   |
 | `--foreground 1`   | stay in foreground, do not daemonize                     |
-| `--trace`          | trace all NFS requests                                   |

docs/usage/nfs.ru.md

@@ -274,4 +274,3 @@ VitastorFS из GPUDirect.
 | `--logfile <FILE>` | записывать логи в заданный файл                         |
 | `--enforce 1`      | проверять права доступа на стороне сервера (по умолчанию нет) |
 | `--foreground 1`   | не уходить в фон после запуска                          |
-| `--trace`          | логгировать все запросы NFS                             |

mon/http_server.js

@@ -16,7 +16,7 @@ async function create_http_server(cfg, handler)
         };
         if (cfg.mon_https_ca)
         {
-            tls.ca = await fsp.readFile(cfg.mon_https_ca);
+            tls.mon_https_ca = await fsp.readFile(cfg.mon_https_ca);
         }
         if (cfg.mon_https_client_auth)
         {

mon/lp_optimizer/lp_optimizer.js

@@ -10,19 +10,16 @@ const NO_OSD = 'Z';
 async function lp_solve(text)
 {
     const cp = child_process.spawn('lp_solve');
-    let stdout = '', stderr = '', finish_cb, finished = 0;
+    let stdout = '', stderr = '', finish_cb;
     cp.stdout.on('data', buf => stdout += buf.toString());
     cp.stderr.on('data', buf => stderr += buf.toString());
-    cp.stdout.on('end', () => finish_cb());
-    cp.stderr.on('end', () => finish_cb());
+    cp.on('exit', () => finish_cb && finish_cb());
     cp.stdin.write(text);
     cp.stdin.end();
-    await new Promise(ok => (finish_cb = () =>
+    if (cp.exitCode == null)
     {
-        finished++;
-        if (finished == 2)
-            ok();
-    }));
+        await new Promise(ok => finish_cb = ok);
+    }
     if (!stdout.trim())
     {
         return null;

mon/mon.js

@@ -627,7 +627,7 @@ class Mon
                 if (this.state.pg.history[pool_id] &&
                     this.state.pg.history[pool_id][pg])
                 {
-                    pg_history[pg-1] = JSON.parse(JSON.stringify(this.state.pg.history[pool_id][pg]));
+                    pg_history[pg-1] = this.state.pg.history[pool_id][pg];
                 }
             }
             const real_prev_pgs = [];
@@ -719,7 +719,7 @@ class Mon
                 this.next_recheck_timer = null;
                 this.next_recheck_at = 0;
                 this.schedule_recheck();
-            }, (this.next_recheck_at-now)*1000);
+            }, now-this.next_recheck_at);
         }
     }
 

mon/osd_tree.js

@@ -87,7 +87,7 @@ function make_hier_tree(global_config, tree)
     tree[''] = { children: [] };
     for (const node_id in tree)
     {
-        if (node_id === '')
+        if (node_id === '' || !(tree[node_id].children||[]).length && (tree[node_id].size||0) <= 0)
         {
             continue;
         }

mon/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor-mon",
-  "version": "3.0.12",
+  "version": "3.0.0",
   "description": "Vitastor SDS monitor service",
   "main": "mon-main.js",
   "scripts": {
@@ -9,7 +9,7 @@
   "author": "Vitaliy Filippov",
   "license": "UNLICENSED",
   "dependencies": {
-    "antietcd": "^1.2.4",
+    "antietcd": "^1.1.3",
     "sprintf-js": "^1.1.2",
     "ws": "^7.2.5"
   },

mon/pg_gen.js

@@ -52,7 +52,6 @@ function recheck_primary(state, global_config, up_osds, osd_tree)
             continue;
         }
         const aff_osds = get_affinity_osds(pool_cfg, up_osds, osd_tree);
-        let paused = false;
         for (let pg_num = 1; pg_num <= pool_cfg.pg_count; pg_num++)
         {
             if (!state.pg.config.items[pool_id])
@@ -75,19 +74,6 @@ function recheck_primary(state, global_config, up_osds, osd_tree)
                     );
                     new_pg_config.items[pool_id][pg_num].primary = new_primary;
                 }
-                paused = paused || !!pg_cfg.pause;
-            }
-        }
-        if (paused)
-        {
-            if (!new_pg_config)
-            {
-                new_pg_config = JSON.parse(JSON.stringify(state.pg.config));
-            }
-            console.log(`Resuming paused pool ${pool_id}`);
-            for (const pg in new_pg_config.items[pool_id])
-            {
-                delete new_pg_config.items[pool_id][pg].pause;
             }
         }
     }
@@ -192,10 +178,10 @@ async function generate_pool_pgs(state, global_config, pool_id, osd_tree, levels
     const rules = use_rules ? get_pg_rules(pool_id, pool_cfg, global_config.placement_levels) : null;
     const folded = fold_failure_domains(Object.values(pool_tree), use_rules ? rules : [ [ [ pool_cfg.failure_domain ] ] ]);
     // FIXME: Remove/merge make_hier_tree() step somewhere, however it's needed to remove empty nodes
-    const folded_tree = make_hier_tree(global_config, folded.nodes.reduce((a, c) => { a[c.id] = c; return a; }, {}));
+    const folded_tree = make_hier_tree(global_config, folded.nodes);
     const old_pg_count = prev_pgs.length;
     const optimize_cfg = {
-        osd_weights: folded.nodes.reduce((a, c) => { if (/^\d+$/.exec(c.id) && c.size != null) { a[c.id] = c.size||0; } return a; }, {}),
+        osd_weights: folded.nodes.reduce((a, c) => { if (Number(c.id)) { a[c.id] = c.size; } return a; }, {}),
         combinator: use_rules
             // new algorithm:
             ? new RuleCombinator(folded_tree, rules, pool_cfg.max_osd_combinations)

mon/pg_utils.js

@@ -84,7 +84,7 @@ function scale_pg_history(prev_pg_history, prev_pgs, new_pgs)
         finish_pg_history(merged_history[1]);
         for (let i = 0; i < new_pg_count; i++)
         {
-            new_pg_history[i] = JSON.parse(JSON.stringify(merged_history[1]));
+            new_pg_history[i] = { ...merged_history[1] };
         }
     }
     // Mark history keys for removed PGs as removed
@@ -102,7 +102,7 @@ function scale_pg_count(prev_pgs, new_pg_count)
     {
         for (let i = prev_pgs.length; i < new_pg_count; i++)
         {
-            prev_pgs[i] = [ ...prev_pgs[i % prev_pgs.length] ];
+            prev_pgs[i] = prev_pgs[i % prev_pgs.length];
         }
     }
     else if (prev_pgs.length > new_pg_count)

mon/stats.js

@@ -37,7 +37,6 @@ function derive_osd_stats(st, prev, prev_diff)
         const n = c.count - BigInt(pr && pr.count||0);
         diff.recovery_stats[op] = { ...c, bps: n > 0 ? b*1000n/timediff : 0n, iops: n > 0 ? n*1000n/timediff : 0n };
     }
-    diff.inode_stats = {};
     for (const pool_id in st.inode_stats||{})
     {
         diff.inode_stats[pool_id] = {};

node-binding/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor",
-  "version": "3.0.12",
+  "version": "3.0.0",
   "description": "Low-level native bindings to Vitastor client library",
   "main": "index.js",
   "keywords": [

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VITASTOR_VERSION = '3.0.12'
+VITASTOR_VERSION = '3.0.0'
 
 LOG = logging.getLogger(__name__)
 
@@ -275,7 +275,7 @@ class VitastorDriver(driver.CloneableImageVD,
             LOG.exception('error getting vitastor pool stats: '+str(e))
 
         self._stats = stats
-
+        
     def get_volume_stats(self, refresh=False):
         """Get volume stats.
         If 'refresh' is True, run update the stats first.
@@ -291,14 +291,6 @@ class VitastorDriver(driver.CloneableImageVD,
         else:
             return (1 + resp['kvs'][0]['value'], resp['kvs'][0]['mod_revision'])
 
-    def _cli(self, descr, *args):
-        args = [ 'vitastor-cli', *args, *(self._vitastor_args()) ]
-        try:
-            self._execute(*args)
-        except processutils.ProcessExecutionError as exc:
-            LOG.error("Failed to "+descr+": "+exc)
-            raise exception.VolumeBackendAPIException(data = exc.stderr)
-
     def create_volume(self, volume):
         """Creates a logical volume."""
 
@@ -310,7 +302,7 @@ class VitastorDriver(driver.CloneableImageVD,
 
         LOG.debug("creating volume '%s'", vol_name)
 
-        self._cli('create volume', 'create', vol_name, '--size', size)
+        self._create_image(vol_name, { 'size': size })
 
         if volume.encryption_key_id:
             self._create_encrypted_volume(volume, volume.obj_context)
@@ -354,7 +346,7 @@ class VitastorDriver(driver.CloneableImageVD,
         snap_name = utils.convert_str(snapshot.name)
         if snap_name.find('@') >= 0 or snap_name.find('/') >= 0:
             raise exception.VolumeBackendAPIException(data = '@ and / are forbidden in volume and snapshot names')
-        self._cli('create snapshot', 'snap-create', vol_name+'@'+snap_name)
+        self._create_snapshot(vol_name, vol_name+'@'+snap_name)
 
     def snapshot_revert_use_temp_snapshot(self):
         """Disable the use of a temporary snapshot on revert."""
@@ -367,8 +359,21 @@ class VitastorDriver(driver.CloneableImageVD,
         snap_name = utils.convert_str(snapshot.name)
 
         # Delete the image and recreate it from the snapshot
-        self._cli('delete image', 'rm', vol_name)
-        self._cli('recreate image', 'create', '--parent', vol_name+'@'+snap_name, vol_name)
+        args = [ 'vitastor-cli', 'rm', vol_name, *(self._vitastor_args()) ]
+        try:
+            self._execute(*args)
+        except processutils.ProcessExecutionError as exc:
+            LOG.error("Failed to delete image "+vol_name+": "+exc)
+            raise exception.VolumeBackendAPIException(data = exc.stderr)
+        args = [
+            'vitastor-cli', 'create', '--parent', vol_name+'@'+snap_name,
+            vol_name, *(self._vitastor_args())
+        ]
+        try:
+            self._execute(*args)
+        except processutils.ProcessExecutionError as exc:
+            LOG.error("Failed to recreate image "+vol_name+" from "+vol_name+"@"+snap_name+": "+exc)
+            raise exception.VolumeBackendAPIException(data = exc.stderr)
 
     def delete_snapshot(self, snapshot):
         """Deletes a snapshot."""
@@ -376,7 +381,15 @@ class VitastorDriver(driver.CloneableImageVD,
         vol_name = utils.convert_str(snapshot.volume_name)
         snap_name = utils.convert_str(snapshot.name)
 
-        self._cli('remove snapshot', 'rm', vol_name+'@'+snap_name)
+        args = [
+            'vitastor-cli', 'rm', vol_name+'@'+snap_name,
+            *(self._vitastor_args())
+        ]
+        try:
+            self._execute(*args)
+        except processutils.ProcessExecutionError as exc:
+            LOG.error("Failed to remove snapshot "+vol_name+'@'+snap_name+": "+exc)
+            raise exception.VolumeBackendAPIException(data = exc.stderr)
 
     def _child_count(self, parents):
         children = 0
@@ -414,7 +427,13 @@ class VitastorDriver(driver.CloneableImageVD,
         if src_vref.admin_metadata.get('readonly') == 'True':
             # source volume is a volume-image cache entry or other readonly volume
             # clone without intermediate snapshot
-            self._cli('create clone', 'create', '--parent', src_name, '--size', size, dest_name)
+            src = self._get_image(src_name)
+            LOG.debug("creating image '%s' from '%s'", dest_name, src_name)
+            new_cfg = self._create_image(dest_name, {
+                'size': size,
+                'parent_id': src['idx']['id'],
+                'parent_pool_id': src['idx']['pool_id'],
+            })
             return {}
 
         clone_snap = "%s@%s.clone_snap" % (src_name, dest_name)
@@ -427,12 +446,15 @@ class VitastorDriver(driver.CloneableImageVD,
             clone_snap = dest_name
             make_img = False
 
-        LOG.debug("creating snapshot '%s'", clone_snap)
-        self._cli('create base snapshot', 'snap-create', '--allow-existing', '1', clone_snap)
-
+        LOG.debug("creating layer '%s' under '%s'", clone_snap, src_name)
+        new_cfg = self._create_snapshot(src_name, clone_snap, True)
         if make_img:
             # Then create a clone from it
-            self._cli('create clone', 'create', '--parent', clone_snap, '--size', size, dest_name)
+            new_cfg = self._create_image(dest_name, {
+                'size': size,
+                'parent_id': new_cfg['parent_id'],
+                'parent_pool_id': new_cfg['parent_pool_id'],
+            })
 
         return {}
 
@@ -442,8 +464,7 @@ class VitastorDriver(driver.CloneableImageVD,
         vol_name = utils.convert_str(volume.name)
         snap_name = utils.convert_str(snapshot.name)
 
-        src_snap = 'volume-'+snapshot.volume_id+'@'+snap_name
-        snap = self._get_image(src_snap)
+        snap = self._get_image('volume-'+snapshot.volume_id+'@'+snap_name)
         if not snap:
             raise exception.SnapshotNotFound(snapshot_id = snap_name)
         snap_inode_id = int(resp['responses'][0]['kvs'][0]['value']['id'])
@@ -452,8 +473,12 @@ class VitastorDriver(driver.CloneableImageVD,
         size = snap['cfg']['size']
         if int(volume.size):
             size = int(volume.size) * units.Gi
+        new_cfg = self._create_image(vol_name, {
+            'size': size,
+            'parent_id': snap['idx']['id'],
+            'parent_pool_id': snap['idx']['pool_id'],
+        })
 
-        self._cli('create clone', 'create', vol_name, '--size', size, '--parent', src_snap)
         return {}
 
     def _vitastor_args(self):
@@ -480,7 +505,49 @@ class VitastorDriver(driver.CloneableImageVD,
         """Deletes a logical volume."""
 
         vol_name = utils.convert_str(volume.name)
-        self._cli('delete volume', 'rm', '--matching', vol_name, vol_name+'@*', '--progress', '0')
+
+        # Find the volume and all its snapshots
+        range_end = b'index/image/' + vol_name.encode('utf-8')
+        range_end = range_end[0 : len(range_end)-1] + six.int2byte(range_end[len(range_end)-1] + 1)
+        resp = self._etcd_txn({ 'success': [
+            { 'request_range': { 'key': 'index/image/'+vol_name, 'range_end': range_end } },
+        ] })
+        if len(resp['responses'][0]['kvs']) == 0:
+            # already deleted
+            LOG.info("volume %s no longer exists in backend", vol_name)
+            return
+        layers = resp['responses'][0]['kvs']
+        layer_ids = {}
+        for kv in layers:
+            inode_id = int(kv['value']['id'])
+            pool_id = int(kv['value']['pool_id'])
+            inode_pool_id = (pool_id << 48) | (inode_id & 0xffffffffffff)
+            layer_ids[inode_pool_id] = True
+
+        # Check if the volume has clones and raise 'busy' if so
+        children = self._child_count(layer_ids)
+        if children > 0:
+            raise exception.VolumeIsBusy(volume_name = vol_name)
+
+        # Clear data
+        for kv in layers:
+            args = [
+                'vitastor-cli', 'rm-data', '--pool', str(kv['value']['pool_id']),
+                '--inode', str(kv['value']['id']), '--progress', '0',
+                *(self._vitastor_args())
+            ]
+            try:
+                self._execute(*args)
+            except processutils.ProcessExecutionError as exc:
+                LOG.error("Failed to remove layer "+kv['key']+": "+exc)
+                raise exception.VolumeBackendAPIException(data = exc.stderr)
+
+        # Delete all layers from etcd
+        requests = []
+        for kv in layers:
+            requests.append({ 'request_delete_range': { 'key': kv['key'] } })
+            requests.append({ 'request_delete_range': { 'key': 'config/inode/'+str(kv['value']['pool_id'])+'/'+str(kv['value']['id']) } })
+        self._etcd_txn({ 'success': requests })
 
     def retype(self, context, volume, new_type, diff, host):
         """Change extra type specifications for a volume."""
@@ -500,6 +567,98 @@ class VitastorDriver(driver.CloneableImageVD,
         """Removes an export for a logical volume."""
         pass
 
+    def _create_image(self, vol_name, cfg):
+        pool_s = str(self.cfg['pool_id'])
+        image_id = 0
+        while image_id == 0:
+            # check if the image already exists and find a free ID
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'index/image/'+vol_name } },
+                { 'request_range': { 'key': 'index/maxid/'+pool_s } },
+            ] })
+            if len(resp['responses'][0]['kvs']) > 0:
+                # already exists
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' already exists')
+            image_id, id_mod = self._next_id(resp['responses'][1])
+            # try to create the image
+            resp = self._etcd_txn({ 'compare': [
+                { 'target': 'MOD', 'mod_revision': id_mod, 'key': 'index/maxid/'+pool_s },
+                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+vol_name },
+                { 'target': 'VERSION', 'version': 0, 'key': 'config/inode/'+pool_s+'/'+str(image_id) },
+            ], 'success': [
+                { 'request_put': { 'key': 'index/maxid/'+pool_s, 'value': image_id } },
+                { 'request_put': { 'key': 'index/image/'+vol_name, 'value': json.dumps({
+                    'id': image_id, 'pool_id': self.cfg['pool_id']
+                }) } },
+                { 'request_put': { 'key': 'config/inode/'+pool_s+'/'+str(image_id), 'value': json.dumps({
+                    **cfg, 'name': vol_name,
+                }) } },
+            ] })
+            if not resp.get('succeeded'):
+                # repeat
+                image_id = 0
+
+    def _create_snapshot(self, vol_name, snap_vol_name, allow_existing = False):
+        while True:
+            # check if the image already exists and snapshot doesn't
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'index/image/'+vol_name } },
+                { 'request_range': { 'key': 'index/image/'+snap_vol_name } },
+            ] })
+            if len(resp['responses'][0]['kvs']) == 0:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
+            if len(resp['responses'][1]['kvs']) > 0:
+                if allow_existing:
+                    snap_idx = resp['responses'][1]['kvs'][0]['value']
+                    resp = self._etcd_txn({ 'success': [
+                        { 'request_range': { 'key': 'config/inode/'+str(snap_idx['pool_id'])+'/'+str(snap_idx['id']) } },
+                    ] })
+                    if len(resp['responses'][0]['kvs']) == 0:
+                        raise exception.VolumeBackendAPIException(data =
+                            'Volume '+snap_vol_name+' is already indexed, but does not exist'
+                        )
+                    return resp['responses'][0]['kvs'][0]['value']
+                raise exception.VolumeBackendAPIException(
+                    data = 'Volume '+snap_vol_name+' already exists'
+                )
+            vol_idx = resp['responses'][0]['kvs'][0]['value']
+            vol_idx_mod = resp['responses'][0]['kvs'][0]['mod_revision']
+            # get image inode config and find a new ID
+            resp = self._etcd_txn({ 'success': [
+                { 'request_range': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) } },
+                { 'request_range': { 'key': 'index/maxid/'+str(self.cfg['pool_id']) } },
+            ] })
+            if len(resp['responses'][0]['kvs']) == 0:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
+            vol_cfg = resp['responses'][0]['kvs'][0]['value']
+            vol_mod = resp['responses'][0]['kvs'][0]['mod_revision']
+            new_id, id_mod = self._next_id(resp['responses'][1])
+            # try to redirect image to the new inode
+            new_cfg = {
+                **vol_cfg, 'name': vol_name, 'parent_id': vol_idx['id'], 'parent_pool_id': vol_idx['pool_id']
+            }
+            resp = self._etcd_txn({ 'compare': [
+                { 'target': 'MOD', 'mod_revision': vol_idx_mod, 'key': 'index/image/'+vol_name },
+                { 'target': 'MOD', 'mod_revision': vol_mod, 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) },
+                { 'target': 'MOD', 'mod_revision': id_mod, 'key': 'index/maxid/'+str(self.cfg['pool_id']) },
+                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+snap_vol_name },
+                { 'target': 'VERSION', 'version': 0, 'key': 'config/inode/'+str(self.cfg['pool_id'])+'/'+str(new_id) },
+            ], 'success': [
+                { 'request_put': { 'key': 'index/maxid/'+str(self.cfg['pool_id']), 'value': new_id } },
+                { 'request_put': { 'key': 'index/image/'+vol_name, 'value': json.dumps({
+                    'id': new_id, 'pool_id': self.cfg['pool_id']
+                }) } },
+                { 'request_put': { 'key': 'config/inode/'+str(self.cfg['pool_id'])+'/'+str(new_id), 'value': json.dumps(new_cfg) } },
+                { 'request_put': { 'key': 'index/image/'+snap_vol_name, 'value': json.dumps({
+                    'id': vol_idx['id'], 'pool_id': vol_idx['pool_id']
+                }) } },
+                { 'request_put': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']), 'value': json.dumps({
+                    **vol_cfg, 'name': snap_vol_name, 'readonly': True
+                }) } }
+            ] })
+            if resp.get('succeeded'):
+                return new_cfg
+
     def initialize_connection(self, volume, connector):
         data = {
             'driver_volume_type': 'vitastor',
@@ -538,9 +697,13 @@ class VitastorDriver(driver.CloneableImageVD,
                     size = int(volume.size) * units.Gi
                     dest_name = utils.convert_str(volume.name)
                     # Find or create the base snapshot
-                    self._cli('create base snapshot', 'create', '--allow-existing', '1', base_vol.name+'@.clone_snap')
+                    snap_cfg = self._create_snapshot(base_vol.name, base_vol.name+'@.clone_snap', True)
                     # Then create a clone from it
-                    self._cli('create clone', 'create', dest_name, '--size', size, '--parent', base_vol.name+'@.clone_snap')
+                    new_cfg = self._create_image(dest_name, {
+                        'size': size,
+                        'parent_id': snap_cfg['parent_id'],
+                        'parent_pool_id': snap_cfg['parent_pool_id'],
+                    })
                     return ({}, True)
         return ({}, False)
 
@@ -607,8 +770,26 @@ class VitastorDriver(driver.CloneableImageVD,
     def extend_volume(self, volume, new_size):
         """Extend an existing volume."""
         vol_name = utils.convert_str(volume.name)
-        size = int(new_size) * units.Gi
-        self._cli('extend volume', 'modify', vol_name, '--resize', new_size)
+        while True:
+            vol = self._get_image(vol_name)
+            if not vol:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
+            # change size
+            size = int(new_size) * units.Gi
+            if size == vol['cfg']['size']:
+                break
+            resp = self._etcd_txn({ 'compare': [ {
+                'target': 'MOD',
+                'mod_revision': vol['cfg_mod'],
+                'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
+            } ], 'success': [
+                { 'request_put': {
+                    'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
+                    'value': json.dumps({ **vol['cfg'], 'size': size }),
+                } },
+            ] })
+            if resp.get('succeeded'):
+                break
         LOG.debug(
             "Extend volume from %(old_size)s GB to %(new_size)s GB.",
             {'old_size': volume.size, 'new_size': new_size}
@@ -681,7 +862,28 @@ class VitastorDriver(driver.CloneableImageVD,
         """
         from_name = self._get_existing_name(existing_ref)
         to_name = utils.convert_str(volume.name)
-        self._cli('rename', 'modify', from_name, '--rename', to_name)
+        self._rename(from_name, to_name)
+
+    def _rename(self, from_name, to_name):
+        while True:
+            vol = self._get_image(from_name)
+            if not vol:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+from_name+' does not exist')
+            to = self._get_image(to_name)
+            if to:
+                raise exception.VolumeBackendAPIException(data = 'Volume '+to_name+' already exists')
+            resp = self._etcd_txn({ 'compare': [
+                { 'target': 'MOD', 'mod_revision': vol['idx_mod'], 'key': 'index/image/'+vol['cfg']['name'] },
+                { 'target': 'MOD', 'mod_revision': vol['cfg_mod'], 'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']) },
+                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+to_name },
+            ], 'success': [
+                { 'request_delete_range': { 'key': 'index/image/'+vol['cfg']['name'] } },
+                { 'request_put': { 'key': 'index/image/'+to_name, 'value': json.dumps(vol['idx']) } },
+                { 'request_put': { 'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
+                    'value': json.dumps({ **vol['cfg'], 'name': to_name }) } },
+            ] })
+            if resp.get('succeeded'):
+                break
 
     def unmanage(self, volume):
         pass
@@ -754,7 +956,7 @@ class VitastorDriver(driver.CloneableImageVD,
         snap_name = self._get_existing_name(existing_ref)
         from_name = vol_name+'@'+snap_name
         to_name = vol_name+'@'+utils.convert_str(snapshot.name)
-        self._cli('rename', 'modify', from_name, '--rename', to_name)
+        self._rename(from_name, to_name)
 
     def unmanage_snapshot(self, snapshot):
         """Removes the specified snapshot from Cinder management."""

patches/libvirt-12.0-vitastor.diff

@@ -1,637 +0,0 @@
-diff --git a/include/libvirt/libvirt-storage.h b/include/libvirt/libvirt-storage.h
-index aaad4a3da1..5f5daa8341 100644
---- a/include/libvirt/libvirt-storage.h
-+++ b/include/libvirt/libvirt-storage.h
-@@ -326,6 +326,7 @@ typedef enum {
-     VIR_CONNECT_LIST_STORAGE_POOLS_ZFS           = 1 << 17, /* (Since: 1.2.8) */
-     VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE      = 1 << 18, /* (Since: 3.1.0) */
-     VIR_CONNECT_LIST_STORAGE_POOLS_ISCSI_DIRECT  = 1 << 19, /* (Since: 5.6.0) */
-+    VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR      = 1 << 20, /* (Since: 5.0.0) */
- } virConnectListAllStoragePoolsFlags;
- 
- int                     virConnectListAllStoragePools(virConnectPtr conn,
-diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
-index 9ca5c2450c..cc52f00c0c 100644
---- a/src/conf/domain_conf.c
-+++ b/src/conf/domain_conf.c
-@@ -7453,7 +7453,8 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
-     src->configFile = virXPathString("string(./config/@file)", ctxt);
- 
-     if (src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP ||
--        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS)
-+        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS ||
-+        src->protocol == VIR_STORAGE_NET_PROTOCOL_VITASTOR)
-         src->query = virXMLPropString(node, "query");
- 
-     if (virDomainStorageNetworkParseHosts(node, ctxt, &src->hosts, &src->nhosts) < 0)
-@@ -32187,6 +32188,7 @@ virDomainStorageSourceTranslateSourcePool(virStorageSource *src,
- 
-     case VIR_STORAGE_POOL_MPATH:
-     case VIR_STORAGE_POOL_RBD:
-+    case VIR_STORAGE_POOL_VITASTOR:
-     case VIR_STORAGE_POOL_SHEEPDOG:
-     case VIR_STORAGE_POOL_GLUSTER:
-     case VIR_STORAGE_POOL_LAST:
-diff --git a/src/conf/domain_validate.c b/src/conf/domain_validate.c
-index 7346a61731..83e94d762e 100644
---- a/src/conf/domain_validate.c
-+++ b/src/conf/domain_validate.c
-@@ -520,6 +520,7 @@ virDomainDiskDefValidateSourceChainOne(const virStorageSource *src)
-         case VIR_STORAGE_NET_PROTOCOL_RBD:
-             break;
- 
-+        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-         case VIR_STORAGE_NET_PROTOCOL_NBD:
-         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
-         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-@@ -592,7 +593,7 @@ virDomainDiskDefValidateSourceChainOne(const virStorageSource *src)
-         }
-     }
- 
--    /* internal snapshots and config files are currently supported only with rbd: */
-+    /* internal snapshots are currently supported only with rbd: */
-     if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_NETWORK &&
-         src->protocol != VIR_STORAGE_NET_PROTOCOL_RBD) {
-         if (src->snapshot) {
-@@ -600,10 +601,14 @@ virDomainDiskDefValidateSourceChainOne(const virStorageSource *src)
-                            _("<snapshot> element is currently supported only with 'rbd' disks"));
-             return -1;
-         }
--
-+    }
-+    /* config files are currently supported only with rbd and vitastor: */
-+    if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_NETWORK &&
-+        src->protocol != VIR_STORAGE_NET_PROTOCOL_RBD &&
-+        src->protocol != VIR_STORAGE_NET_PROTOCOL_VITASTOR) {
-         if (src->configFile) {
-             virReportError(VIR_ERR_XML_ERROR, "%s",
--                           _("<config> element is currently supported only with 'rbd' disks"));
-+                           _("<config> element is currently supported only with 'rbd' and 'vitastor' disks"));
-             return -1;
-         }
-     }
-diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng
-index 114dd3f96f..c71f9a3277 100644
---- a/src/conf/schemas/domaincommon.rng
-+++ b/src/conf/schemas/domaincommon.rng
-@@ -2093,6 +2093,35 @@
-     </element>
-   </define>
- 
-+  <define name="diskSourceNetworkProtocolVitastor">
-+    <element name="source">
-+      <interleave>
-+        <attribute name="protocol">
-+          <value>vitastor</value>
-+        </attribute>
-+        <ref name="diskSourceCommon"/>
-+        <optional>
-+          <attribute name="name"/>
-+        </optional>
-+        <optional>
-+          <attribute name="query"/>
-+        </optional>
-+        <zeroOrMore>
-+          <ref name="diskSourceNetworkHost"/>
-+        </zeroOrMore>
-+        <optional>
-+          <element name="config">
-+            <attribute name="file">
-+              <ref name="absFilePath"/>
-+            </attribute>
-+            <empty/>
-+          </element>
-+        </optional>
-+        <empty/>
-+      </interleave>
-+    </element>
-+  </define>
-+
-   <define name="diskSourceNetworkProtocolISCSI">
-     <element name="source">
-       <attribute name="protocol">
-@@ -2443,6 +2472,7 @@
-       <ref name="diskSourceNetworkProtocolSimple"/>
-       <ref name="diskSourceNetworkProtocolVxHS"/>
-       <ref name="diskSourceNetworkProtocolNFS"/>
-+      <ref name="diskSourceNetworkProtocolVitastor"/>
-     </choice>
-   </define>
- 
-diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c
-index 1dc9365bf2..a8a736be81 100644
---- a/src/conf/storage_conf.c
-+++ b/src/conf/storage_conf.c
-@@ -56,7 +56,7 @@ VIR_ENUM_IMPL(virStoragePool,
-               "logical", "disk", "iscsi",
-               "iscsi-direct", "scsi", "mpath",
-               "rbd", "sheepdog", "gluster",
--              "zfs", "vstorage",
-+              "zfs", "vstorage", "vitastor",
- );
- 
- VIR_ENUM_IMPL(virStoragePoolFormatFileSystem,
-@@ -242,6 +242,18 @@ static virStoragePoolTypeInfo poolTypeInfo[] = {
-           .formatToString = virStorageFileFormatTypeToString,
-       }
-     },
-+    {.poolType = VIR_STORAGE_POOL_VITASTOR,
-+     .poolOptions = {
-+         .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
-+                   VIR_STORAGE_POOL_SOURCE_NETWORK |
-+                   VIR_STORAGE_POOL_SOURCE_NAME),
-+      },
-+      .volOptions = {
-+          .defaultFormat = VIR_STORAGE_FILE_RAW,
-+          .formatFromString = virStorageVolumeFormatFromString,
-+          .formatToString = virStorageFileFormatTypeToString,
-+      }
-+    },
-     {.poolType = VIR_STORAGE_POOL_SHEEPDOG,
-      .poolOptions = {
-          .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
-@@ -538,6 +550,11 @@ virStoragePoolDefParseSource(xmlXPathContextPtr ctxt,
-                        _("element 'name' is mandatory for RBD pool"));
-         return -1;
-     }
-+    if (pool_type == VIR_STORAGE_POOL_VITASTOR && source->name == NULL) {
-+        virReportError(VIR_ERR_XML_ERROR, "%s",
-+                       _("element 'name' is mandatory for Vitastor pool"));
-+        return -1;
-+    }
- 
-     if (options->formatFromString) {
-         g_autofree char *format = NULL;
-@@ -1127,6 +1144,7 @@ virStoragePoolDefFormatBuf(virBuffer *buf,
-     /* RBD, Sheepdog, Gluster and Iscsi-direct devices are not local block devs nor
-      * files, so they don't have a target */
-     if (def->type != VIR_STORAGE_POOL_RBD &&
-+        def->type != VIR_STORAGE_POOL_VITASTOR &&
-         def->type != VIR_STORAGE_POOL_SHEEPDOG &&
-         def->type != VIR_STORAGE_POOL_GLUSTER &&
-         def->type != VIR_STORAGE_POOL_ISCSI_DIRECT) {
-diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h
-index fc67957cfe..720c07ef74 100644
---- a/src/conf/storage_conf.h
-+++ b/src/conf/storage_conf.h
-@@ -103,6 +103,7 @@ typedef enum {
-     VIR_STORAGE_POOL_GLUSTER,  /* Gluster device */
-     VIR_STORAGE_POOL_ZFS,      /* ZFS */
-     VIR_STORAGE_POOL_VSTORAGE, /* Virtuozzo Storage */
-+    VIR_STORAGE_POOL_VITASTOR, /* Vitastor */
- 
-     VIR_STORAGE_POOL_LAST,
- } virStoragePoolType;
-@@ -454,6 +455,7 @@ VIR_ENUM_DECL(virStoragePartedFs);
-                  VIR_CONNECT_LIST_STORAGE_POOLS_SCSI     | \
-                  VIR_CONNECT_LIST_STORAGE_POOLS_MPATH    | \
-                  VIR_CONNECT_LIST_STORAGE_POOLS_RBD      | \
-+                 VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR | \
-                  VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG | \
-                  VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER  | \
-                  VIR_CONNECT_LIST_STORAGE_POOLS_ZFS      | \
-diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
-index d7b9bdfecb..38aefd0dd4 100644
---- a/src/conf/storage_source_conf.c
-+++ b/src/conf/storage_source_conf.c
-@@ -90,6 +90,7 @@ VIR_ENUM_IMPL(virStorageNetProtocol,
-               "ssh",
-               "vxhs",
-               "nfs",
-+              "vitastor",
- );
- 
- 
-@@ -1317,6 +1318,7 @@ virStorageSourceNetworkDefaultPort(virStorageNetProtocol protocol)
-         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-             return 24007;
- 
-+        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-         case VIR_STORAGE_NET_PROTOCOL_RBD:
-             /* we don't provide a default for RBD */
-             return 0;
-diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
-index 22c35d420d..f1e32ea83d 100644
---- a/src/conf/storage_source_conf.h
-+++ b/src/conf/storage_source_conf.h
-@@ -131,6 +131,7 @@ typedef enum {
-     VIR_STORAGE_NET_PROTOCOL_SSH,
-     VIR_STORAGE_NET_PROTOCOL_VXHS,
-     VIR_STORAGE_NET_PROTOCOL_NFS,
-+    VIR_STORAGE_NET_PROTOCOL_VITASTOR,
- 
-     VIR_STORAGE_NET_PROTOCOL_LAST
- } virStorageNetProtocol;
-diff --git a/src/conf/virstorageobj.c b/src/conf/virstorageobj.c
-index 59fa5da372..4739167f5f 100644
---- a/src/conf/virstorageobj.c
-+++ b/src/conf/virstorageobj.c
-@@ -1438,6 +1438,7 @@ virStoragePoolObjSourceFindDuplicateCb(const void *payload,
-             return 1;
-         break;
- 
-+    case VIR_STORAGE_POOL_VITASTOR:
-     case VIR_STORAGE_POOL_ISCSI_DIRECT:
-     case VIR_STORAGE_POOL_RBD:
-     case VIR_STORAGE_POOL_LAST:
-@@ -1921,6 +1922,8 @@ virStoragePoolObjMatch(virStoragePoolObj *obj,
-                (obj->def->type == VIR_STORAGE_POOL_MPATH))   ||
-               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_RBD) &&
-                (obj->def->type == VIR_STORAGE_POOL_RBD))     ||
-+              (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR) &&
-+               (obj->def->type == VIR_STORAGE_POOL_VITASTOR)) ||
-               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG) &&
-                (obj->def->type == VIR_STORAGE_POOL_SHEEPDOG)) ||
-               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER) &&
-diff --git a/src/libvirt-storage.c b/src/libvirt-storage.c
-index db7660aac4..561df34709 100644
---- a/src/libvirt-storage.c
-+++ b/src/libvirt-storage.c
-@@ -94,6 +94,7 @@ virStoragePoolGetConnect(virStoragePoolPtr pool)
-  * VIR_CONNECT_LIST_STORAGE_POOLS_SCSI
-  * VIR_CONNECT_LIST_STORAGE_POOLS_MPATH
-  * VIR_CONNECT_LIST_STORAGE_POOLS_RBD
-+ * VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR
-  * VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG
-  * VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER
-  * VIR_CONNECT_LIST_STORAGE_POOLS_ZFS
-diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
-index 2b988157fa..9d0eb47b25 100644
---- a/src/libxl/libxl_conf.c
-+++ b/src/libxl/libxl_conf.c
-@@ -1069,6 +1069,7 @@ libxlMakeNetworkDiskSrcStr(virStorageSource *src,
-     case VIR_STORAGE_NET_PROTOCOL_SSH:
-     case VIR_STORAGE_NET_PROTOCOL_VXHS:
-     case VIR_STORAGE_NET_PROTOCOL_NFS:
-+    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-     case VIR_STORAGE_NET_PROTOCOL_LAST:
-     case VIR_STORAGE_NET_PROTOCOL_NONE:
-         virReportError(VIR_ERR_NO_SUPPORT,
-diff --git a/src/libxl/xen_xl.c b/src/libxl/xen_xl.c
-index e72e7d7f44..8482c21805 100644
---- a/src/libxl/xen_xl.c
-+++ b/src/libxl/xen_xl.c
-@@ -1461,6 +1461,7 @@ xenFormatXLDiskSrcNet(virStorageSource *src)
-     case VIR_STORAGE_NET_PROTOCOL_SSH:
-     case VIR_STORAGE_NET_PROTOCOL_VXHS:
-     case VIR_STORAGE_NET_PROTOCOL_NFS:
-+    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-     case VIR_STORAGE_NET_PROTOCOL_LAST:
-     case VIR_STORAGE_NET_PROTOCOL_NONE:
-         virReportError(VIR_ERR_NO_SUPPORT,
-diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
-index 9b43279797..459d8e8a65 100644
---- a/src/qemu/qemu_block.c
-+++ b/src/qemu/qemu_block.c
-@@ -743,6 +743,38 @@ qemuBlockStorageSourceGetRBDProps(virStorageSource *src,
- }
- 
- 
-+static virJSONValue *
-+qemuBlockStorageSourceGetVitastorProps(virStorageSource *src)
-+{
-+    virJSONValue *ret = NULL;
-+    virStorageNetHostDef *host;
-+    size_t i;
-+    g_auto(virBuffer) buf = VIR_BUFFER_INITIALIZER;
-+    g_autofree char *etcd = NULL;
-+
-+    for (i = 0; i < src->nhosts; i++) {
-+        host = src->hosts + i;
-+        if ((virStorageNetHostTransport)host->transport != VIR_STORAGE_NET_HOST_TRANS_TCP) {
-+            return NULL;
-+        }
-+        virBufferAsprintf(&buf, i > 0 ? ",%s:%u" : "%s:%u", host->name, host->port);
-+    }
-+    if (src->nhosts > 0) {
-+        etcd = virBufferContentAndReset(&buf);
-+    }
-+
-+    if (virJSONValueObjectAdd(&ret,
-+                              "S:etcd-host", etcd,
-+                              "S:etcd-prefix", src->query,
-+                              "S:config-path", src->configFile,
-+                              "s:image", src->path,
-+                              NULL) < 0)
-+        return NULL;
-+
-+    return ret;
-+}
-+
-+
- static virJSONValue *
- qemuBlockStorageSourceGetSshProps(virStorageSource *src)
- {
-@@ -1094,6 +1126,12 @@ qemuBlockStorageSourceGetBackendProps(virStorageSource *src,
-                 return NULL;
-             break;
- 
-+        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-+            driver = "vitastor";
-+            if (!(fileprops = qemuBlockStorageSourceGetVitastorProps(src)))
-+                return NULL;
-+            break;
-+
-         case VIR_STORAGE_NET_PROTOCOL_SSH:
-             driver = "ssh";
-             if (!(fileprops = qemuBlockStorageSourceGetSshProps(src)))
-@@ -1997,6 +2035,7 @@ qemuBlockGetBackingStoreString(virStorageSource *src,
- 
-             case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
-             case VIR_STORAGE_NET_PROTOCOL_RBD:
-+            case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-             case VIR_STORAGE_NET_PROTOCOL_VXHS:
-             case VIR_STORAGE_NET_PROTOCOL_NFS:
-             case VIR_STORAGE_NET_PROTOCOL_SSH:
-@@ -2377,6 +2416,12 @@ qemuBlockStorageSourceCreateGetStorageProps(virStorageSource *src,
-                 return -1;
-             break;
- 
-+        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-+            driver = "vitastor";
-+            if (!(location = qemuBlockStorageSourceGetVitastorProps(src)))
-+                return -1;
-+            break;
-+
-         case VIR_STORAGE_NET_PROTOCOL_SSH:
-             if (srcPriv->nbdkitProcess) {
-                 /* disk creation not yet supported with nbdkit, and even if it
-diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
-index ac56fc7cb4..9e407b4aab 100644
---- a/src/qemu/qemu_domain.c
-+++ b/src/qemu/qemu_domain.c
-@@ -4677,7 +4677,8 @@ qemuDomainValidateStorageSource(virStorageSource *src,
-     if (src->query &&
-         (actualType != VIR_STORAGE_TYPE_NETWORK ||
-          (src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTPS &&
--          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP))) {
-+          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP &&
-+          src->protocol != VIR_STORAGE_NET_PROTOCOL_VITASTOR))) {
-         virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
-                        _("query is supported only with HTTP(S) protocols"));
-         return -1;
-@@ -9103,6 +9104,7 @@ qemuDomainPrepareStorageSourceTLS(virStorageSource *src,
-         break;
- 
-     case VIR_STORAGE_NET_PROTOCOL_RBD:
-+    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
-     case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-     case VIR_STORAGE_NET_PROTOCOL_ISCSI:
-diff --git a/src/qemu/qemu_snapshot.c b/src/qemu/qemu_snapshot.c
-index e738afffc3..37d64f469b 100644
---- a/src/qemu/qemu_snapshot.c
-+++ b/src/qemu/qemu_snapshot.c
-@@ -665,6 +665,7 @@ qemuSnapshotPrepareDiskExternalInactive(virDomainSnapshotDiskDef *snapdisk,
-         case VIR_STORAGE_NET_PROTOCOL_NONE:
-         case VIR_STORAGE_NET_PROTOCOL_NBD:
-         case VIR_STORAGE_NET_PROTOCOL_RBD:
-+        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
-         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
-@@ -893,6 +894,7 @@ qemuSnapshotPrepareDiskInternal(virDomainDiskDef *disk,
-         case VIR_STORAGE_NET_PROTOCOL_NONE:
-         case VIR_STORAGE_NET_PROTOCOL_NBD:
-         case VIR_STORAGE_NET_PROTOCOL_RBD:
-+        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
-         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
-         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
-diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c
-index e19e032427..59f91f4710 100644
---- a/src/storage/storage_driver.c
-+++ b/src/storage/storage_driver.c
-@@ -1626,6 +1626,7 @@ storageVolLookupByPathCallback(virStoragePoolObj *obj,
- 
-         case VIR_STORAGE_POOL_GLUSTER:
-         case VIR_STORAGE_POOL_RBD:
-+        case VIR_STORAGE_POOL_VITASTOR:
-         case VIR_STORAGE_POOL_SHEEPDOG:
-         case VIR_STORAGE_POOL_ZFS:
-         case VIR_STORAGE_POOL_LAST:
-diff --git a/src/storage_file/storage_source_backingstore.c b/src/storage_file/storage_source_backingstore.c
-index 821378883c..2211f6891b 100644
---- a/src/storage_file/storage_source_backingstore.c
-+++ b/src/storage_file/storage_source_backingstore.c
-@@ -264,6 +264,75 @@ virStorageSourceParseRBDColonString(const char *rbdstr,
- }
- 
- 
-+static int
-+virStorageSourceParseVitastorColonString(const char *colonstr,
-+                                         virStorageSource *src)
-+{
-+    char *p, *e, *next;
-+    g_autofree char *options = NULL;
-+
-+    /* optionally skip the "vitastor:" prefix if provided */
-+    if (STRPREFIX(colonstr, "vitastor:"))
-+        colonstr += strlen("vitastor:");
-+
-+    options = g_strdup(colonstr);
-+
-+    p = options;
-+    while (*p) {
-+        /* find : delimiter or end of string */
-+        for (e = p; *e && *e != ':'; ++e) {
-+            if (*e == '\\') {
-+                e++;
-+                if (*e == '\0')
-+                    break;
-+            }
-+        }
-+        if (*e == '\0') {
-+            next = e;    /* last kv pair */
-+        } else {
-+            next = e + 1;
-+            *e = '\0';
-+        }
-+
-+        if (STRPREFIX(p, "image=")) {
-+            src->path = g_strdup(p + strlen("image="));
-+        } else if (STRPREFIX(p, "etcd-prefix=")) {
-+            src->query = g_strdup(p + strlen("etcd-prefix="));
-+        } else if (STRPREFIX(p, "config-path=")) {
-+            src->configFile = g_strdup(p + strlen("config-path="));
-+        } else if (STRPREFIX(p, "etcd-host=")) {
-+            char *h, *sep;
-+
-+            h = p + strlen("etcd-host=");
-+            while (h < e) {
-+                for (sep = h; sep < e; ++sep) {
-+                    if (*sep == '\\' && (sep[1] == ',' ||
-+                                         sep[1] == ';' ||
-+                                         sep[1] == ' ')) {
-+                        *sep = '\0';
-+                        sep += 2;
-+                        break;
-+                    }
-+                }
-+
-+                if (virStorageSourceRBDAddHost(src, h) < 0)
-+                    return -1;
-+
-+                h = sep;
-+            }
-+        }
-+
-+        p = next;
-+    }
-+
-+    if (!src->path) {
-+        return -1;
-+    }
-+
-+    return 0;
-+}
-+
-+
- static int
- virStorageSourceParseNBDColonString(const char *nbdstr,
-                                     virStorageSource *src)
-@@ -379,6 +448,11 @@ virStorageSourceParseBackingColon(virStorageSource *src,
-             return -1;
-         break;
- 
-+    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
-+        if (virStorageSourceParseVitastorColonString(path, src) < 0)
-+            return -1;
-+        break;
-+
-     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
-     case VIR_STORAGE_NET_PROTOCOL_LAST:
-     case VIR_STORAGE_NET_PROTOCOL_NONE:
-@@ -953,6 +1027,54 @@ virStorageSourceParseBackingJSONRBD(virStorageSource *src,
-     return 0;
- }
- 
-+static int
-+virStorageSourceParseBackingJSONVitastor(virStorageSource *src,
-+                                         virJSONValue *json,
-+                                         const char *jsonstr G_GNUC_UNUSED,
-+                                         int opaque G_GNUC_UNUSED)
-+{
-+    const char *filename;
-+    const char *image = virJSONValueObjectGetString(json, "image");
-+    const char *conf = virJSONValueObjectGetString(json, "config-path");
-+    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd-prefix");
-+    virJSONValue *servers = virJSONValueObjectGetArray(json, "server");
-+    size_t nservers;
-+    size_t i;
-+
-+    src->type = VIR_STORAGE_TYPE_NETWORK;
-+    src->protocol = VIR_STORAGE_NET_PROTOCOL_VITASTOR;
-+
-+    /* legacy syntax passed via 'filename' option */
-+    if ((filename = virJSONValueObjectGetString(json, "filename")))
-+        return virStorageSourceParseVitastorColonString(filename, src);
-+
-+    if (!image) {
-+        virReportError(VIR_ERR_INVALID_ARG, "%s",
-+                       _("missing image name in Vitastor backing volume "
-+                         "JSON specification"));
-+        return -1;
-+    }
-+
-+    src->path = g_strdup(image);
-+    src->configFile = g_strdup(conf);
-+    src->query = g_strdup(etcd_prefix);
-+
-+    if (servers) {
-+        nservers = virJSONValueArraySize(servers);
-+
-+        src->hosts = g_new0(virStorageNetHostDef, nservers);
-+        src->nhosts = nservers;
-+
-+        for (i = 0; i < nservers; i++) {
-+            if (virStorageSourceParseBackingJSONInetSocketAddress(src->hosts + i,
-+                                                                  virJSONValueArrayGet(servers, i)) < 0)
-+                return -1;
-+        }
-+    }
-+
-+    return 0;
-+}
-+
- static int
- virStorageSourceParseBackingJSONRaw(virStorageSource *src,
-                                     virJSONValue *json,
-@@ -1130,6 +1252,7 @@ static const struct virStorageSourceJSONDriverParser jsonParsers[] = {
-     {"sheepdog", false, virStorageSourceParseBackingJSONSheepdog, 0},
-     {"ssh", false, virStorageSourceParseBackingJSONSSH, 0},
-     {"rbd", false, virStorageSourceParseBackingJSONRBD, 0},
-+    {"vitastor", false, virStorageSourceParseBackingJSONVitastor, 0},
-     {"raw", true, virStorageSourceParseBackingJSONRaw, 0},
-     {"nfs", false, virStorageSourceParseBackingJSONNFS, 0},
-     {"vxhs", false, virStorageSourceParseBackingJSONVxHS, 0},
-diff --git a/src/test/test_driver.c b/src/test/test_driver.c
-index 1165689de7..bba846351c 100644
---- a/src/test/test_driver.c
-+++ b/src/test/test_driver.c
-@@ -7345,6 +7345,7 @@ testStorageVolumeTypeForPool(int pooltype)
-     case VIR_STORAGE_POOL_ISCSI_DIRECT:
-     case VIR_STORAGE_POOL_GLUSTER:
-     case VIR_STORAGE_POOL_RBD:
-+    case VIR_STORAGE_POOL_VITASTOR:
-         return VIR_STORAGE_VOL_NETWORK;
-     case VIR_STORAGE_POOL_LOGICAL:
-     case VIR_STORAGE_POOL_DISK:
-diff --git a/tests/storagepoolcapsschemadata/poolcaps-fs.xml b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
-index eee75af746..8bd0a57bdd 100644
---- a/tests/storagepoolcapsschemadata/poolcaps-fs.xml
-+++ b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
-@@ -204,4 +204,11 @@
-       </enum>
-     </volOptions>
-   </pool>
-+  <pool type='vitastor' supported='no'>
-+    <volOptions>
-+      <defaultFormat type='raw'/>
-+      <enum name='targetFormatType'>
-+      </enum>
-+    </volOptions>
-+  </pool>
- </storagepoolCapabilities>
-diff --git a/tests/storagepoolcapsschemadata/poolcaps-full.xml b/tests/storagepoolcapsschemadata/poolcaps-full.xml
-index 805950a937..852df0de16 100644
---- a/tests/storagepoolcapsschemadata/poolcaps-full.xml
-+++ b/tests/storagepoolcapsschemadata/poolcaps-full.xml
-@@ -204,4 +204,11 @@
-       </enum>
-     </volOptions>
-   </pool>
-+  <pool type='vitastor' supported='yes'>
-+    <volOptions>
-+      <defaultFormat type='raw'/>
-+      <enum name='targetFormatType'>
-+      </enum>
-+    </volOptions>
-+  </pool>
- </storagepoolCapabilities>
-diff --git a/tests/storagepoolxml2argvtest.c b/tests/storagepoolxml2argvtest.c
-index d5c2531ab8..b19308ac38 100644
---- a/tests/storagepoolxml2argvtest.c
-+++ b/tests/storagepoolxml2argvtest.c
-@@ -57,6 +57,7 @@ testCompareXMLToArgvFiles(bool shouldFail,
-     case VIR_STORAGE_POOL_GLUSTER:
-     case VIR_STORAGE_POOL_ZFS:
-     case VIR_STORAGE_POOL_VSTORAGE:
-+    case VIR_STORAGE_POOL_VITASTOR:
-     case VIR_STORAGE_POOL_LAST:
-     default:
-         VIR_TEST_DEBUG("pool type '%s' has no xml2argv test", defTypeStr);
-diff --git a/tools/virsh-pool.c b/tools/virsh-pool.c
-index 2010ef1356..072e2ff9e8 100644
---- a/tools/virsh-pool.c
-+++ b/tools/virsh-pool.c
-@@ -1187,6 +1187,9 @@ cmdPoolList(vshControl *ctl, const vshCmd *cmd G_GNUC_UNUSED)
-             case VIR_STORAGE_POOL_VSTORAGE:
-                 flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE;
-                 break;
-+            case VIR_STORAGE_POOL_VITASTOR:
-+                flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR;
-+                break;
-             case VIR_STORAGE_POOL_LAST:
-                 break;
-             }

patches/linux-fix-atomic-write-checks.diff

@@ -1,39 +0,0 @@
-From 98d3f68a40130c438854f61db6025f9e9b099cb6 Mon Sep 17 00:00:00 2001
-From: Vitaliy Filippov <vitalifster@gmail.com>
-Date: Sat, 20 Dec 2025 14:44:35 +0300
-Subject: [PATCH] Do not require atomic writes to be power of 2 sized and
- aligned on length boundary
-
-It contradicts NVMe specification where alignment is only required when atomic
-write boundary (NABSPF/NABO) is set and highly limits usage of NVMe atomic writes
-
-Signed-off-by: Vitaliy Filippov <vitalifster@gmail.com>
----
- fs/read_write.c | 8 --------
- 1 file changed, 8 deletions(-)
-
-diff --git a/fs/read_write.c b/fs/read_write.c
-index 833bae068770..5467d710108d 100644
---- a/fs/read_write.c
-+++ b/fs/read_write.c
-@@ -1802,17 +1802,9 @@ int generic_file_rw_checks(struct file *file_in, struct file *file_out)
- 
- int generic_atomic_write_valid(struct kiocb *iocb, struct iov_iter *iter)
- {
--	size_t len = iov_iter_count(iter);
--
- 	if (!iter_is_ubuf(iter))
- 		return -EINVAL;
- 
--	if (!is_power_of_2(len))
--		return -EINVAL;
--
--	if (!IS_ALIGNED(iocb->ki_pos, len))
--		return -EINVAL;
--
- 	if (!(iocb->ki_flags & IOCB_DIRECT))
- 		return -EOPNOTSUPP;
- 
--- 
-2.51.0
-

patches/qemu-10.1-vitastor.patch

@@ -1,172 +1,29 @@
-diff --git a/block/meson.build b/block/meson.build
-index 34b1b2a306..24ca0f1e52 100644
---- a/block/meson.build
-+++ b/block/meson.build
-@@ -114,6 +114,7 @@ foreach m : [
-   [libnfs, 'nfs', files('nfs.c')],
-   [libssh, 'ssh', files('ssh.c')],
-   [rbd, 'rbd', files('rbd.c')],
-+  [vitastor, 'vitastor', files('vitastor.c')],
- ]
-   if m[0].found()
-     module_ss = ss.source_set()
-diff --git a/meson.build b/meson.build
-index 50c774a195..e5c7a3a4b1 100644
---- a/meson.build
-+++ b/meson.build
-@@ -1652,6 +1652,26 @@ if not get_option('rbd').auto() or have_block
-   endif
- endif
+diff --git a/src/client/qemu_driver.c b/src/client/qemu_driver.c
+index d8356dab..5f4cd50d 100644
+--- a/src/client/qemu_driver.c
++++ b/src/client/qemu_driver.c
+@@ -974,14 +974,21 @@ static void vitastor_co_read_bitmap_cb(void *opaque, long retval, uint8_t *bitma
+ #endif
+ }
  
-+vitastor = not_found
-+if not get_option('vitastor').auto() or have_block
-+  libvitastor_client = cc.find_library('vitastor_client', has_headers: ['vitastor_c.h'],
-+    required: get_option('vitastor'))
-+  if libvitastor_client.found()
-+    if cc.links('''
-+      #include <vitastor_c.h>
-+      int main(void) {
-+        vitastor_c_create_qemu(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
-+        return 0;
-+      }''', dependencies: libvitastor_client)
-+      vitastor = declare_dependency(dependencies: libvitastor_client)
-+    elif get_option('vitastor').enabled()
-+      error('could not link libvitastor_client')
-+    else
-+      warning('could not link libvitastor_client, disabling')
-+    endif
-+  endif
-+endif
-+
- glusterfs = not_found
- glusterfs_ftruncate_has_stat = false
- glusterfs_iocb_has_stat = false
-@@ -2547,6 +2567,7 @@ endif
- config_host_data.set('CONFIG_OPENGL', opengl.found())
- config_host_data.set('CONFIG_PLUGIN', get_option('plugins'))
- config_host_data.set('CONFIG_RBD', rbd.found())
-+config_host_data.set('CONFIG_VITASTOR', vitastor.found())
- config_host_data.set('CONFIG_RDMA', rdma.found())
- config_host_data.set('CONFIG_RELOCATABLE', get_option('relocatable'))
- config_host_data.set('CONFIG_SAFESTACK', get_option('safe_stack'))
-@@ -4972,6 +4993,7 @@ summary_info += {'fdt support':       fdt_opt == 'internal' ? 'internal' : fdt}
- summary_info += {'libcap-ng support': libcap_ng}
- summary_info += {'bpf support':       libbpf}
- summary_info += {'rbd support':       rbd}
-+summary_info += {'vitastor support':  vitastor}
- summary_info += {'smartcard support': cacard}
- summary_info += {'U2F support':       u2f}
- summary_info += {'libusb':            libusb}
-diff --git a/meson_options.txt b/meson_options.txt
-index fff1521e58..f0844c0e00 100644
---- a/meson_options.txt
-+++ b/meson_options.txt
-@@ -202,6 +202,8 @@ option('pvg', type: 'feature', value: 'auto',
-        description: 'macOS paravirtualized graphics support')
- option('rbd', type : 'feature', value : 'auto',
-        description: 'Ceph block device driver')
-+option('vitastor', type : 'feature', value : 'auto',
-+       description: 'Vitastor block device driver')
- option('opengl', type : 'feature', value : 'auto',
-        description: 'OpenGL support')
- option('rdma', type : 'feature', value : 'auto',
-diff --git a/qapi/block-core.json b/qapi/block-core.json
-index dc6eb4ae23..d043f4340e 100644
---- a/qapi/block-core.json
-+++ b/qapi/block-core.json
-@@ -3280,7 +3280,7 @@
-             'parallels', 'preallocate', 'qcow', 'qcow2', 'qed', 'quorum',
-             'raw', 'rbd',
-             { 'name': 'replication', 'if': 'CONFIG_REPLICATION' },
--            'ssh', 'throttle', 'vdi', 'vhdx',
-+            'ssh', 'throttle', 'vdi', 'vhdx', 'vitastor',
-             { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
-             { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
-             { 'name': 'virtio-blk-vhost-vdpa', 'if': 'CONFIG_BLKIO' },
-@@ -4363,6 +4363,28 @@
-             '*key-secret': 'str',
-             '*server': ['InetSocketAddressBase'] } }
- 
-+##
-+# @BlockdevOptionsVitastor:
-+#
-+# Driver specific block device options for vitastor
-+#
-+# @image:       Image name
-+# @inode:       Inode number
-+# @pool:        Pool ID
-+# @size:        Desired image size in bytes
-+# @config-path: Path to Vitastor configuration
-+# @etcd-host:   etcd connection address(es)
-+# @etcd-prefix: etcd key/value prefix
-+##
-+{ 'struct': 'BlockdevOptionsVitastor',
-+  'data': { '*inode': 'uint64',
-+            '*pool': 'uint64',
-+            '*size': 'uint64',
-+            '*image': 'str',
-+            '*config-path': 'str',
-+            '*etcd-host': 'str',
-+            '*etcd-prefix': 'str' } }
-+
- ##
- # @ReplicationMode:
- #
-@@ -4831,6 +4853,7 @@
-       'throttle':   'BlockdevOptionsThrottle',
-       'vdi':        'BlockdevOptionsGenericFormat',
-       'vhdx':       'BlockdevOptionsGenericFormat',
-+      'vitastor':   'BlockdevOptionsVitastor',
-       'virtio-blk-vfio-pci':
-                     { 'type': 'BlockdevOptionsVirtioBlkVfioPci',
-                       'if': 'CONFIG_BLKIO' },
-@@ -5304,6 +5327,20 @@
-             '*cluster-size' :   'size',
-             '*encrypt' :        'RbdEncryptionCreateOptions' } }
- 
-+##
-+# @BlockdevCreateOptionsVitastor:
-+#
-+# Driver specific image creation options for Vitastor.
-+#
-+# @location: Where to store the new image file.  This location cannot
-+#     point to a snapshot.
-+#
-+# @size: Size of the virtual disk in bytes
-+##
-+{ 'struct': 'BlockdevCreateOptionsVitastor',
-+  'data': { 'location':         'BlockdevOptionsVitastor',
-+            'size':             'size' } }
-+
- ##
- # @BlockdevVmdkSubformat:
- #
-@@ -5526,6 +5563,7 @@
-       'ssh':            'BlockdevCreateOptionsSsh',
-       'vdi':            'BlockdevCreateOptionsVdi',
-       'vhdx':           'BlockdevCreateOptionsVhdx',
-+      'vitastor':       'BlockdevCreateOptionsVitastor',
-       'vmdk':           'BlockdevCreateOptionsVmdk',
-       'vpc':            'BlockdevCreateOptionsVpc'
-   } }
-diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh
-index 0ebe6bc52a..2c37ad3892 100644
---- a/scripts/meson-buildoptions.sh
-+++ b/scripts/meson-buildoptions.sh
-@@ -175,6 +175,7 @@ meson_options_help() {
-   printf "%s\n" '  qga-vss         build QGA VSS support (broken with MinGW)'
-   printf "%s\n" '  qpl             Query Processing Library support'
-   printf "%s\n" '  rbd             Ceph block device driver'
-+  printf "%s\n" '  vitastor        Vitastor block device driver'
-   printf "%s\n" '  rdma            Enable RDMA-based migration'
-   printf "%s\n" '  replication     replication support'
-   printf "%s\n" '  rust            Rust support'
-@@ -459,6 +460,8 @@ _meson_option_parse() {
-     --disable-qpl) printf "%s" -Dqpl=disabled ;;
-     --enable-rbd) printf "%s" -Drbd=enabled ;;
-     --disable-rbd) printf "%s" -Drbd=disabled ;;
-+    --enable-vitastor) printf "%s" -Dvitastor=enabled ;;
-+    --disable-vitastor) printf "%s" -Dvitastor=disabled ;;
-     --enable-rdma) printf "%s" -Drdma=enabled ;;
-     --disable-rdma) printf "%s" -Drdma=disabled ;;
-     --enable-relocatable) printf "%s" -Drelocatable=true ;;
+-static int coroutine_fn vitastor_co_block_status(
+-    BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes,
+-    int64_t *pnum, int64_t *map, BlockDriverState **file)
++static int coroutine_fn vitastor_co_block_status(BlockDriverState *bs,
++#if QEMU_VERSION_MAJOR > 10 || QEMU_VERSION_MAJOR == 10 && QEMU_VERSION_MINOR >= 1
++    unsigned int mode,
++#else
++    bool want_zero,
++#endif
++    int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file)
+ {
+     // Allocated => return BDRV_BLOCK_DATA|BDRV_BLOCK_OFFSET_VALID
+     // Not allocated => return 0
+     // Error => return -errno
+     // Set pnum to length of the extent, `*map` = `offset`, `*file` = `bs`
++#if QEMU_VERSION_MAJOR > 10 || QEMU_VERSION_MAJOR == 10 && QEMU_VERSION_MINOR >= 1
++    int want_zero = (mode == BDRV_WANT_PRECISE);
++#endif
+     VitastorRPC task;
+     VitastorClient *client = bs->opaque;
+     uint64_t inode = client->watch ? vitastor_c_inode_get_num(client->watch) : client->inode;

patches/qemu-10.2-vitastor.patch

@@ -1,172 +0,0 @@
-diff --git a/block/meson.build b/block/meson.build
-index 34b1b2a306..24ca0f1e52 100644
---- a/block/meson.build
-+++ b/block/meson.build
-@@ -114,6 +114,7 @@ foreach m : [
-   [libnfs, 'nfs', files('nfs.c')],
-   [libssh, 'ssh', files('ssh.c')],
-   [rbd, 'rbd', files('rbd.c')],
-+  [vitastor, 'vitastor', files('vitastor.c')],
- ]
-   if m[0].found()
-     module_ss = ss.source_set()
-diff --git a/meson.build b/meson.build
-index d9293294d8..776a5becc6 100644
---- a/meson.build
-+++ b/meson.build
-@@ -1665,6 +1665,26 @@ if not get_option('rbd').auto() or have_block
-   endif
- endif
- 
-+vitastor = not_found
-+if not get_option('vitastor').auto() or have_block
-+  libvitastor_client = cc.find_library('vitastor_client', has_headers: ['vitastor_c.h'],
-+    required: get_option('vitastor'))
-+  if libvitastor_client.found()
-+    if cc.links('''
-+      #include <vitastor_c.h>
-+      int main(void) {
-+        vitastor_c_create_qemu(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
-+        return 0;
-+      }''', dependencies: libvitastor_client)
-+      vitastor = declare_dependency(dependencies: libvitastor_client)
-+    elif get_option('vitastor').enabled()
-+      error('could not link libvitastor_client')
-+    else
-+      warning('could not link libvitastor_client, disabling')
-+    endif
-+  endif
-+endif
-+
- glusterfs = not_found
- glusterfs_ftruncate_has_stat = false
- glusterfs_iocb_has_stat = false
-@@ -2509,6 +2529,7 @@ endif
- config_host_data.set('CONFIG_OPENGL', opengl.found())
- config_host_data.set('CONFIG_PLUGIN', get_option('plugins'))
- config_host_data.set('CONFIG_RBD', rbd.found())
-+config_host_data.set('CONFIG_VITASTOR', vitastor.found())
- config_host_data.set('CONFIG_RDMA', rdma.found())
- config_host_data.set('CONFIG_RELOCATABLE', get_option('relocatable'))
- config_host_data.set('CONFIG_SAFESTACK', get_option('safe_stack'))
-@@ -4948,6 +4969,7 @@ summary_info += {'fdt support':       fdt_opt == 'internal' ? 'internal' : fdt}
- summary_info += {'libcap-ng support': libcap_ng}
- summary_info += {'bpf support':       libbpf}
- summary_info += {'rbd support':       rbd}
-+summary_info += {'vitastor support':  vitastor}
- summary_info += {'smartcard support': cacard}
- summary_info += {'U2F support':       u2f}
- summary_info += {'libusb':            libusb}
-diff --git a/meson_options.txt b/meson_options.txt
-index 2836156257..148086cc6f 100644
---- a/meson_options.txt
-+++ b/meson_options.txt
-@@ -206,6 +206,8 @@ option('pvg', type: 'feature', value: 'auto',
-        description: 'macOS paravirtualized graphics support')
- option('rbd', type : 'feature', value : 'auto',
-        description: 'Ceph block device driver')
-+option('vitastor', type : 'feature', value : 'auto',
-+       description: 'Vitastor block device driver')
- option('opengl', type : 'feature', value : 'auto',
-        description: 'OpenGL support')
- option('rdma', type : 'feature', value : 'auto',
-diff --git a/qapi/block-core.json b/qapi/block-core.json
-index b82af74256..f25a6f5ce8 100644
---- a/qapi/block-core.json
-+++ b/qapi/block-core.json
-@@ -3351,7 +3351,7 @@
-             'parallels', 'preallocate', 'qcow', 'qcow2', 'qed', 'quorum',
-             'raw', 'rbd',
-             { 'name': 'replication', 'if': 'CONFIG_REPLICATION' },
--            'ssh', 'throttle', 'vdi', 'vhdx',
-+            'ssh', 'throttle', 'vdi', 'vhdx', 'vitastor',
-             { 'name': 'virtio-blk-vfio-pci', 'if': 'CONFIG_BLKIO' },
-             { 'name': 'virtio-blk-vhost-user', 'if': 'CONFIG_BLKIO' },
-             { 'name': 'virtio-blk-vhost-vdpa', 'if': 'CONFIG_BLKIO' },
-@@ -4434,6 +4434,28 @@
-             '*key-secret': 'str',
-             '*server': ['InetSocketAddressBase'] } }
- 
-+##
-+# @BlockdevOptionsVitastor:
-+#
-+# Driver specific block device options for vitastor
-+#
-+# @image:       Image name
-+# @inode:       Inode number
-+# @pool:        Pool ID
-+# @size:        Desired image size in bytes
-+# @config-path: Path to Vitastor configuration
-+# @etcd-host:   etcd connection address(es)
-+# @etcd-prefix: etcd key/value prefix
-+##
-+{ 'struct': 'BlockdevOptionsVitastor',
-+  'data': { '*inode': 'uint64',
-+            '*pool': 'uint64',
-+            '*size': 'uint64',
-+            '*image': 'str',
-+            '*config-path': 'str',
-+            '*etcd-host': 'str',
-+            '*etcd-prefix': 'str' } }
-+
- ##
- # @ReplicationMode:
- #
-@@ -4902,6 +4924,7 @@
-       'throttle':   'BlockdevOptionsThrottle',
-       'vdi':        'BlockdevOptionsGenericFormat',
-       'vhdx':       'BlockdevOptionsGenericFormat',
-+      'vitastor':   'BlockdevOptionsVitastor',
-       'virtio-blk-vfio-pci':
-                     { 'type': 'BlockdevOptionsVirtioBlkVfioPci',
-                       'if': 'CONFIG_BLKIO' },
-@@ -5376,6 +5399,20 @@
-             '*cluster-size' :   'size',
-             '*encrypt' :        'RbdEncryptionCreateOptions' } }
- 
-+##
-+# @BlockdevCreateOptionsVitastor:
-+#
-+# Driver specific image creation options for Vitastor.
-+#
-+# @location: Where to store the new image file.  This location cannot
-+#     point to a snapshot.
-+#
-+# @size: Size of the virtual disk in bytes
-+##
-+{ 'struct': 'BlockdevCreateOptionsVitastor',
-+  'data': { 'location':         'BlockdevOptionsVitastor',
-+            'size':             'size' } }
-+
- ##
- # @BlockdevVmdkSubformat:
- #
-@@ -5598,6 +5635,7 @@
-       'ssh':            'BlockdevCreateOptionsSsh',
-       'vdi':            'BlockdevCreateOptionsVdi',
-       'vhdx':           'BlockdevCreateOptionsVhdx',
-+      'vitastor':       'BlockdevCreateOptionsVitastor',
-       'vmdk':           'BlockdevCreateOptionsVmdk',
-       'vpc':            'BlockdevCreateOptionsVpc'
-   } }
-diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh
-index 3d0d132344..65ee8c855e 100644
---- a/scripts/meson-buildoptions.sh
-+++ b/scripts/meson-buildoptions.sh
-@@ -177,6 +177,7 @@ meson_options_help() {
-   printf "%s\n" '  qga-vss         build QGA VSS support (broken with MinGW)'
-   printf "%s\n" '  qpl             Query Processing Library support'
-   printf "%s\n" '  rbd             Ceph block device driver'
-+  printf "%s\n" '  vitastor        Vitastor block device driver'
-   printf "%s\n" '  rdma            Enable RDMA-based migration'
-   printf "%s\n" '  replication     replication support'
-   printf "%s\n" '  rust            Rust support'
-@@ -464,6 +465,8 @@ _meson_option_parse() {
-     --disable-qpl) printf "%s" -Dqpl=disabled ;;
-     --enable-rbd) printf "%s" -Drbd=enabled ;;
-     --disable-rbd) printf "%s" -Drbd=disabled ;;
-+    --enable-vitastor) printf "%s" -Dvitastor=enabled ;;
-+    --disable-vitastor) printf "%s" -Dvitastor=disabled ;;
-     --enable-rdma) printf "%s" -Drdma=enabled ;;
-     --disable-rdma) printf "%s" -Drdma=disabled ;;
-     --enable-relocatable) printf "%s" -Drelocatable=true ;;

rpm/build-tarball.sh

@@ -21,7 +21,7 @@ rpmbuild -bp fio.spec
 cd $VITASTOR
 VER=$(grep ^Version: rpm/vitastor-$REL.spec | awk '{print $2}')
 rm -rf fio
-ln -s $(ls -d ~/rpmbuild/BUILD/fio*/ | grep -v SPECPARTS) fio
+ln -s ~/rpmbuild/BUILD/fio*/ fio
 sh copy-fio-includes.sh
 rm fio
 mv fio-copy fio

rpm/vitastor-el10.Dockerfile

@@ -1,17 +0,0 @@
-# Build packages for AlmaLinux 10 inside a container
-# cd ..
-# docker pull --platform=linux/amd64/v2 quay.io/almalinuxorg/almalinux:10
-# docker build -t vitastor-buildenv:el10 -f rpm/vitastor-el10.Dockerfile .
-# docker run -i --rm -v ./:/root/vitastor vitastor-buildenv:el10 /root/vitastor/rpm/vitastor-build.sh
-
-FROM quay.io/almalinuxorg/almalinux:10
-
-WORKDIR /root
-
-RUN sed -i 's/enabled=0/enabled=1/' /etc/yum.repos.d/*.repo
-RUN dnf -y install epel-release dnf-plugins-core
-RUN dnf -y install https://vitastor.io/rpms/centos/10/vitastor-release-1.0-1.el10.noarch.rpm
-RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel isa-l-devel gf-complete-devel rdma-core-devel cmake libnl3-devel
-RUN dnf download --source fio
-RUN rpm --nomd5 -i fio*.src.rpm
-RUN cd ~/rpmbuild/SPECS && dnf builddep -y --spec fio.spec

rpm/vitastor-el10.spec

@@ -1,198 +0,0 @@
-Name:           vitastor
-Version:        3.0.12
-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-3.0.12.el10.tar.gz
-
-BuildRequires:  gperftools-devel
-BuildRequires:  gcc-c++
-BuildRequires:  nodejs >= 10
-BuildRequires:  jerasure-devel
-BuildRequires:  isa-l-devel
-BuildRequires:  gf-complete-devel
-BuildRequires:  rdma-core-devel
-BuildRequires:  cmake
-BuildRequires:  libnl3-devel
-Requires:       vitastor-osd = %{version}-%{release}
-Requires:       vitastor-mon = %{version}-%{release}
-Requires:       vitastor-client = %{version}-%{release}
-Requires:       vitastor-client-devel = %{version}-%{release}
-Requires:       vitastor-fio = %{version}-%{release}
-
-%description
-Vitastor is a small, simple and fast clustered block storage (storage for VM drives),
-architecturally similar to Ceph which means strong consistency, primary-replication,
-symmetric clustering and automatic data distribution over any number of drives of any
-size with configurable redundancy (replication or erasure codes/XOR).
-
-
-%package -n vitastor-osd
-Summary:        Vitastor - OSD
-Requires:       vitastor-client = %{version}-%{release}
-Requires:       util-linux
-Requires:       parted
-
-
-%description -n vitastor-osd
-Vitastor object storage daemon, i.e. server program that stores data.
-
-
-%package -n vitastor-mon
-Summary:        Vitastor - monitor
-Requires:       nodejs >= 10
-Requires:       lpsolve
-
-
-%description -n vitastor-mon
-Vitastor monitor, i.e. server program responsible for watching cluster state and
-scheduling cluster-level operations.
-
-
-%package -n vitastor-client
-Summary:        Vitastor - client
-
-
-%description -n vitastor-client
-Vitastor client library and command-line interface.
-
-
-%package -n vitastor-client-devel
-Summary:        Vitastor - development files
-Group:          Development/Libraries
-Requires:       vitastor-client = %{version}-%{release}
-
-
-%description -n vitastor-client-devel
-Vitastor library headers for development.
-
-
-%package -n vitastor-fio
-Summary:        Vitastor - fio drivers
-Group:          Development/Libraries
-Requires:       vitastor-client = %{version}-%{release}
-Requires:       fio = 3.36-5.el10
-
-
-%description -n vitastor-fio
-Vitastor fio drivers for benchmarking.
-
-
-%package -n vitastor-opennebula
-Summary:        Vitastor for OpenNebula
-Group:          Development/Libraries
-Requires:       vitastor-client
-Requires:       jq
-Requires:       python3-lxml
-Requires:       patch
-Requires:       qemu-kvm-block-vitastor
-
-
-%description -n vitastor-opennebula
-Vitastor storage plugin for OpenNebula.
-
-
-%prep
-%setup -q
-
-
-%build
-%cmake
-%cmake_build
-
-
-%install
-rm -rf $RPM_BUILD_ROOT
-%cmake_install
-cd mon
-npm install --production
-cd ..
-mkdir -p %buildroot/usr/lib/vitastor
-cp -r mon %buildroot/usr/lib/vitastor
-mv %buildroot/usr/lib/vitastor/mon/scripts/make-etcd %buildroot/usr/lib/vitastor/mon/
-mkdir -p %buildroot/lib/systemd/system
-cp mon/scripts/vitastor.target mon/scripts/vitastor-mon.service mon/scripts/vitastor-osd@.service %buildroot/lib/systemd/system
-mkdir -p %buildroot/lib/udev/rules.d
-cp mon/scripts/90-vitastor.rules %buildroot/lib/udev/rules.d
-mkdir -p %buildroot/var/lib/one
-cp -r opennebula/remotes %buildroot/var/lib/one
-cp opennebula/install.sh %buildroot/var/lib/one/remotes/datastore/vitastor/
-mkdir -p %buildroot/etc/
-cp -r opennebula/sudoers.d %buildroot/etc/
-
-
-%files
-%doc GPL-2.0.txt VNPL-1.1.txt README.md README-ru.md
-
-
-%files -n vitastor-osd
-%_bindir/vitastor-osd
-%_bindir/vitastor-disk
-%_bindir/vitastor-dump-journal
-/lib/systemd/system/vitastor-osd@.service
-/lib/systemd/system/vitastor.target
-/lib/udev/rules.d/90-vitastor.rules
-
-
-%pre -n vitastor-osd
-groupadd -r -f vitastor 2>/dev/null ||:
-useradd -r -g vitastor -s /sbin/nologin -c "Vitastor daemons" -M -d /nonexistent vitastor 2>/dev/null ||:
-install -o vitastor -g vitastor -d /var/log/vitastor
-mkdir -p /etc/vitastor
-
-
-%files -n vitastor-mon
-/usr/lib/vitastor/mon
-/lib/systemd/system/vitastor-mon.service
-
-
-%pre -n vitastor-mon
-groupadd -r -f vitastor 2>/dev/null ||:
-useradd -r -g vitastor -s /sbin/nologin -c "Vitastor daemons" -M -d /nonexistent vitastor 2>/dev/null ||:
-mkdir -p /etc/vitastor
-mkdir -p /var/lib/vitastor
-chown vitastor:vitastor /var/lib/vitastor
-
-
-%files -n vitastor-client
-%_bindir/vitastor-nbd
-%_bindir/vitastor-ublk
-%_bindir/vitastor-nfs
-%_bindir/vitastor-cli
-%_bindir/vitastor-rm
-%_bindir/vitastor-kv
-%_bindir/vitastor-kv-stress
-%_bindir/vita
-%_libdir/libvitastor_client.so*
-%_libdir/libvitastor_kv.so*
-
-
-%files -n vitastor-client-devel
-%_includedir/vitastor_c.h
-%_includedir/vitastor_kv.h
-%_libdir/pkgconfig
-
-
-%files -n vitastor-fio
-%_libdir/libfio_vitastor.so
-%_libdir/libfio_vitastor_blk.so
-%_libdir/libfio_vitastor_sec.so
-
-
-%files -n vitastor-opennebula
-/var/lib/one
-/etc/sudoers.d/opennebula-vitastor
-
-
-%triggerin -n vitastor-opennebula -- opennebula
-[ $2 = 0 ] || exit 0
-/var/lib/one/remotes/datastore/vitastor/install.sh
-
-
-# Turn off the brp-python-bytecompile script
-%global __os_install_post %(echo '%{__os_install_post}' | sed -e 's!/usr/lib[^[:space:]]*/brp-python-bytecompile[[:space:]].*$!!g')
-
-
-%changelog

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.12
+Version:        3.0.0
 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-3.0.12.el7.tar.gz
+Source0:        vitastor-3.0.0.el7.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  devtoolset-9-gcc-c++

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.12
+Version:        3.0.0
 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-3.0.12.el8.tar.gz
+Source0:        vitastor-3.0.0.el8.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-toolset-9-gcc-c++

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.12
+Version:        3.0.0
 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-3.0.12.el9.tar.gz
+Source0:        vitastor-3.0.0.el9.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-c++

src/CMakeLists.txt

@@ -1,8 +1,9 @@
-cmake_minimum_required(VERSION 2.8...3.30)
+cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
 include(GNUInstallDirs)
+include(CTest)
 include(CheckIncludeFile)
 
 find_package(PkgConfig)
@@ -20,7 +21,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 endif()
 set(ENABLE_COVERAGE false CACHE BOOL "Enable code coverage")
 
-add_definitions(-DVITASTOR_VERSION="3.0.12")
+add_definitions(-DVITASTOR_VERSION="3.0.0")
 add_definitions(-D_GNU_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -fno-omit-frame-pointer -fvisibility=hidden -I ${CMAKE_SOURCE_DIR}/src)
 add_link_options(-fno-omit-frame-pointer)
 if (${WITH_ASAN})

src/blockstore/CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required(VERSION 2.8...3.30)
+cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 

src/blockstore/blockstore.h

@@ -183,11 +183,6 @@ public:
     // Update configuration
     virtual void parse_config(blockstore_config_t & config) = 0;
 
-    // Reshard database for a pool in chunks
-    // MUST be called only when nobody makes any modifications to the DB for this pool
-    virtual void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit) = 0;
-    virtual bool reshard_continue(void *reshard_state, uint64_t chunk_limit) = 0;
-
     // Event loop
     virtual void loop() = 0;
 
@@ -228,9 +223,4 @@ public:
     virtual uint64_t get_journal_size() = 0;
 
     virtual uint32_t get_bitmap_granularity() = 0;
-
-    virtual uint64_t get_live_entries() = 0;
-    virtual uint64_t get_live_memory() = 0;
-    virtual uint64_t get_garbage_entries() = 0;
-    virtual uint64_t get_garbage_memory() = 0;
 };

src/blockstore/blockstore_disk.cpp

@@ -94,9 +94,6 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
     csum_block_size = parse_size(config["csum_block_size"]);
     discard_on_start = config.find("discard_on_start") != config.end() &&
         (config["discard_on_start"] == "true" || config["discard_on_start"] == "1" || config["discard_on_start"] == "yes");
-    gc_on_start = config.find("gc_on_start") == config.end() ||
-        (config["gc_on_start"] == "true" || config["gc_on_start"] == "1" || config["gc_on_start"] == "yes");
-    skip_double_claim = (config["skip_double_claim"] == "true" || config["skip_double_claim"] == "1" || config["skip_double_claim"] == "yes");
     min_discard_size = parse_size(config["min_discard_size"]);
     if (!min_discard_size)
         min_discard_size = 1024*1024;
@@ -174,10 +171,6 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
     {
         throw std::runtime_error("Data block size must be a multiple of sparse write tracking granularity");
     }
-    if (data_block_size / bitmap_granularity < 8)
-    {
-        throw std::runtime_error("Data block size must be at least bitmap_granularity*8");
-    }
     if (!data_csum_type)
     {
         csum_block_size = 0;
@@ -266,7 +259,7 @@ void blockstore_disk_t::calc_lengths(bool skip_meta_check)
     }
     // required metadata size
     block_count = data_len / data_block_size;
-    clean_entry_bitmap_size = (data_block_size / bitmap_granularity + 7) / 8;
+    clean_entry_bitmap_size = data_block_size / bitmap_granularity / 8;
     clean_dyn_size = clean_entry_bitmap_size*2 + (csum_block_size
         ? data_block_size/csum_block_size*(data_csum_type & 0xFF) : 0);
 recalc:

src/blockstore/blockstore_disk.h

@@ -26,14 +26,14 @@ class allocator_t;
 struct blockstore_disk_t
 {
     std::string data_device, meta_device, journal_device;
-    uint64_t data_block_size;
+    uint32_t data_block_size;
     uint64_t cfg_journal_size, cfg_data_size;
     // Required write alignment and journal/metadata/data areas' location alignment
     uint32_t disk_alignment = 4096;
     // Journal block size - minimum_io_size of the journal device is the best choice
-    uint64_t journal_block_size = 4096;
+    uint32_t journal_block_size = 4096;
     // Metadata block size - minimum_io_size of the metadata device is the best choice
-    uint64_t meta_block_size = 4096;
+    uint32_t meta_block_size = 4096;
     // Atomic write size of the data block device
     uint32_t atomic_write_size = 4096;
     // Whether we should set RWF_ATOMIC on atomic writes
@@ -57,10 +57,6 @@ struct blockstore_disk_t
     bool inmemory_journal = true;
     // Data discard granularity and minimum size (for the sake of performance)
     bool discard_on_start = false;
-    // GC on start (new store)
-    bool gc_on_start = true;
-    // Skip double claim conflicts on start (new store, temporary until the bug is found)
-    bool skip_double_claim = false;
     uint64_t min_discard_size = 1024*1024;
     uint64_t discard_granularity = 0;
 

src/blockstore/blockstore_flush.cpp

@@ -174,18 +174,14 @@ bool journal_flusher_co::loop()
     else if (wait_state == 19) goto resume_19;
     else if (wait_state == 20) goto resume_20;
     else if (wait_state == 21) goto resume_21;
-    else if (wait_state == 22) goto resume_22;
-    else if (wait_state == 23) goto resume_23;
-    else if (wait_state == 24) goto resume_24;
-    else if (wait_state == 25) goto resume_25;
 resume_0:
     wait_state = 0;
     wait_count = 0;
     cur_oid = {};
     res = bs->heap->get_next_compact(cur_oid);
+    // Advance fsynced_lsn every <journal_trim_interval> intent writes
     if ((bs->intent_write_counter >= bs->journal_trim_interval) && co_id == 0)
     {
-        // Advance fsynced_lsn every <journal_trim_interval> intent writes
         bs->intent_write_counter = 0;
 resume_17:
 resume_18:
@@ -200,7 +196,6 @@ resume_21:
     if (res == ENOENT && flusher->force_start > 0 && co_id == 0 &&
         (!bs->dsk.disable_journal_fsync || !bs->dsk.disable_meta_fsync || !bs->dsk.disable_data_fsync))
     {
-        // When under pressure, do an additional fsync to force entries to be marked compactable
         flusher->active_flushers++;
 resume_14:
 resume_15:
@@ -264,9 +259,11 @@ resume_1:
         if (wr->type() == BS_HEAP_SMALL_WRITE ||
             wr->type() == BS_HEAP_INTENT_WRITE && bs->dsk.csum_block_size > bs->dsk.bitmap_granularity)
         {
-            bs->prepare_read(read_vec, cur_obj, wr, 0, bs->dsk.data_block_size,
+            auto res = bs->prepare_read(read_vec, cur_obj, wr, 0, bs->dsk.data_block_size,
                 wr->type() == BS_HEAP_INTENT_WRITE && bs->dsk.csum_block_size > bs->dsk.bitmap_granularity && !bs->perfect_csum_update
                 ? COPY_BUF_SKIP_CSUM : 0);
+            if (res > 0)
+                copy_count++;
         }
     });
     if (!compact_info.compact_lsn)
@@ -276,25 +273,6 @@ resume_1:
         bs->heap->unlock_entry(cur_oid);
         goto resume_0;
     }
-    flusher->active_flushers++;
-    for (i = 0; i < read_vec.size(); i++)
-    {
-        if ((read_vec[i].copy_flags & COPY_BUF_JOURNAL) &&
-            !(read_vec[i].copy_flags & COPY_BUF_COALESCED))
-        {
-            copy_count++;
-        }
-    }
-    if (copy_count > 0 && !bs->dsk.disable_data_fsync)
-    {
-        init_fsync_data();
-    }
-    if (bs->log_level > 10)
-    {
-        printf("Compacting %jx:%jx v%ju..v%ju / l%ju..l%ju (%d writes)\n", cur_oid.inode, cur_oid.stripe,
-            compact_info.clean_wr->version, compact_info.compact_version,
-            compact_info.clean_wr->lsn, compact_info.compact_lsn, copy_count);
-    }
     mem_or(new_bmp, compact_info.clean_wr->get_int_bitmap(bs->heap), bs->dsk.clean_entry_bitmap_size);
     if (!bitmap_copied)
     {
@@ -313,6 +291,13 @@ resume_1:
         csum_copy.clear();
     }
     clean_loc = compact_info.clean_wr->big_location(bs->heap);
+    flusher->active_flushers++;
+    if (bs->log_level > 10)
+    {
+        printf("Compacting %jx:%jx v%ju..v%ju / l%ju..l%ju (%d writes)\n", cur_oid.inode, cur_oid.stripe,
+            compact_info.clean_wr->version, compact_info.compact_version,
+            compact_info.clean_wr->lsn, compact_info.compact_lsn, copy_count);
+    }
     overwrite_start = overwrite_end = 0;
     if (read_vec.size() > 0)
     {
@@ -351,13 +336,6 @@ resume_3:
     if (res == ENOENT || res == EDOM)
     {
         // Abort compaction
-abort_compact:
-        if (copy_count > 0 && !bs->dsk.disable_data_fsync)
-        {
-            cur_sync->member_count--;
-            if (cur_sync->member_count > 0)
-                bs->ringloop->wakeup();
-        }
         flusher->flushing.erase(cur_oid);
         bs->heap->unlock_entry(cur_oid);
         flusher->active_flushers--;
@@ -371,7 +349,10 @@ resume_4:
         if (res == ENOENT)
         {
             // Abort compaction
-            goto abort_compact;
+            flusher->flushing.erase(cur_oid);
+            bs->heap->unlock_entry(cur_oid);
+            flusher->active_flushers--;
+            goto resume_0;
         }
         if (res == EAGAIN)
         {
@@ -400,14 +381,14 @@ resume_9:
     for (i = 0; i < read_vec.size(); i++)
     {
         if ((read_vec[i].copy_flags & COPY_BUF_JOURNAL) &&
-            !(read_vec[i].copy_flags & COPY_BUF_COALESCED))
+            !(read_vec[i].copy_flags & COPY_BUF_COALESCED) ||
+            (read_vec[i].copy_flags & COPY_BUF_PADDED)) // FIXME Shit, simplify these flags
         {
             assert(read_vec[i].buf);
             await_sqe(10);
             data->iov = (struct iovec){ read_vec[i].buf + (read_vec[i].copy_flags & COPY_BUF_PADDED
                 ? read_vec[i].offset - read_vec[i].disk_offset : 0), (size_t)read_vec[i].len };
             data->callback = simple_callback_w;
-            assert(clean_loc + read_vec[i].offset + data->iov.iov_len <= bs->dsk.block_count*bs->dsk.data_block_size);
             io_uring_prep_writev(sqe, bs->dsk.data_fd, &data->iov, 1, bs->dsk.data_offset + clean_loc + read_vec[i].offset);
             wait_count++;
         }
@@ -418,17 +399,6 @@ resume_11:
         wait_state = 11;
         return false;
     }
-    if (copy_count > 0 && !bs->dsk.disable_data_fsync)
-    {
-resume_22:
-resume_23:
-resume_24:
-resume_25:
-        if (!fsync_data(22))
-        {
-            return false;
-        }
-    }
     // Lock is only needed to prevent freeing the big_write because we overwrite it...
     bs->heap->unlock_entry(cur_oid);
     // Mark the object compacted, but don't free and remove small_writes
@@ -438,14 +408,12 @@ resume_25:
     if (!cur_obj)
     {
         // Abort compaction
-        flusher->active_flushers--;
         flusher->flushing.erase(cur_oid);
         goto resume_0;
     }
     if (!calc_block_checksums())
     {
         // Abort compaction
-        flusher->active_flushers--;
         flusher->flushing.erase(cur_oid);
         goto resume_0;
     }
@@ -454,7 +422,6 @@ resume_25:
     if (res == EBUSY)
     {
         // Abort compaction, object is already overwritten by something else
-        flusher->active_flushers--;
         flusher->flushing.erase(cur_oid);
         goto resume_0;
     }
@@ -619,7 +586,7 @@ int journal_flusher_co::check_and_punch_checksums()
             bs->heap->calc_block_checksums((uint32_t*)(new_csums+csum_off), vec.buf, punch_bmp, vec.offset, vec.offset+vec.len, true, NULL);
         }
     }
-    // Modified, we should punch_holes and then write the block to disk
+    // Modified, we should add_punch_holes and then write the block to disk
     return EBUSY;
 }
 
@@ -732,67 +699,6 @@ resume_1:
     return true;
 }
 
-void journal_flusher_co::init_fsync_data()
-{
-    cur_sync = flusher->data_syncs.begin();
-    if (cur_sync == flusher->data_syncs.end() || cur_sync->ready_count > 0)
-    {
-        cur_sync = flusher->data_syncs.emplace(cur_sync);
-    }
-    cur_sync->member_count++;
-}
-
-bool journal_flusher_co::fsync_data(int wait_base)
-{
-    if (wait_state == wait_base)
-        goto resume_0;
-    else if (wait_state == wait_base+1)
-        goto resume_1;
-    else if (wait_state == wait_base+2)
-        goto resume_2;
-    else if (wait_state == wait_base+3)
-        goto resume_3;
-    cur_sync->ready_count++;
-resume_0:
-    if (cur_sync->ready_count < cur_sync->member_count)
-    {
-        wait_state = wait_base;
-        return false;
-    }
-    if (!cur_sync->sent)
-    {
-        // Sync batch is ready. Do it.
-        await_sqe(1);
-        data->iov = { 0 };
-        data->callback = simple_callback_w;
-        io_uring_prep_fsync(sqe, bs->dsk.data_fd, IORING_FSYNC_DATASYNC);
-        cur_sync->sent = true;
-        wait_count++;
-resume_2:
-        if (wait_count > 0)
-        {
-            wait_state = wait_base+2;
-            return false;
-        }
-        cur_sync->done = true;
-        // Wake up other flushers
-        bs->ringloop->wakeup();
-    }
-resume_3:
-    if (!cur_sync->done)
-    {
-        wait_state = wait_base+3;
-        return false;
-    }
-    cur_sync->done_count++;
-    if (cur_sync->done_count >= cur_sync->member_count)
-    {
-        flusher->data_syncs.erase(cur_sync);
-        cur_sync = flusher->data_syncs.end();
-    }
-    return true;
-}
-
 bool journal_flusher_co::fsync_meta(int wait_base)
 {
     if (wait_state == wait_base)        goto resume_0;

src/blockstore/blockstore_flush.h

@@ -25,15 +25,6 @@ struct flusher_meta_write_t
     std::map<uint64_t, meta_sector_t>::iterator it;
 };
 
-struct flusher_data_sync_t
-{
-    int member_count = 0;
-    int ready_count = 0;
-    int done_count = 0;
-    bool sent = false;
-    bool done = false;
-};
-
 class journal_flusher_t;
 
 // Journal flusher coroutine
@@ -67,7 +58,7 @@ class journal_flusher_co
     int i, res;
     bool read_to_fill_incomplete;
     int copy_count;
-    std::list<flusher_data_sync_t>::iterator cur_sync;
+    bool do_repeat = false;
 
     friend class journal_flusher_t;
 
@@ -78,8 +69,6 @@ class journal_flusher_co
     bool calc_block_checksums();
     bool write_meta_block(int wait_base);
     bool read_buffered(int wait_base);
-    void init_fsync_data();
-    bool fsync_data(int wait_base);
     bool fsync_meta(int wait_base);
     bool fsync_buffer(int wait_base);
     bool trim_lsn(int wait_base);
@@ -100,7 +89,6 @@ class journal_flusher_t
 
     robin_hood::unordered_flat_set<object_id> flushing;
     int active_flushers = 0;
-    std::list<flusher_data_sync_t> data_syncs;
     int wanting_meta_fsync = 0;
     bool fsyncing_meta = false;
     int syncing_buffer = 0;

src/blockstore/blockstore_heap.h

@@ -57,11 +57,11 @@ struct __attribute__((__packed__)) heap_entry_t
     inline heap_small_write_t& small() { return *(heap_small_write_t*)this; }
     inline heap_big_write_t& big() { return *(heap_big_write_t*)this; }
     inline heap_big_intent_t& big_intent() { return *(heap_big_intent_t*)this; }
-    bool is_garbage() const;
+    bool is_garbage();
     void set_garbage();
-    bool is_overwrite() const;
-    bool is_compactable() const;
-    bool is_before(const heap_entry_t *other) const;
+    bool is_overwrite();
+    bool is_compactable();
+    bool is_before(heap_entry_t *other);
     uint32_t get_size(blockstore_heap_t *heap);
     uint8_t *get_ext_bitmap(blockstore_heap_t *heap);
     uint8_t *get_int_bitmap(blockstore_heap_t *heap);
@@ -117,13 +117,10 @@ struct heap_object_mvcc_t
 
 struct heap_block_info_t
 {
-    struct __attribute__((__packed__))
-    {
-        uint32_t used_space = 0;
-        uint32_t garbage_space = 0;
-    };
+    uint32_t used_space = 0;
     uint64_t mod_lsn = 0, mod_lsn_to = 0; // only 1 block write of LSN sequence is allowed at a moment
-    bool is_writing = false;
+    bool is_writing: 1;
+    bool has_garbage: 1;
     std::vector<heap_list_item_t*> entries;
 };
 
@@ -140,8 +137,6 @@ struct heap_compact_t
     bool do_delete;
 };
 
-struct heap_reshard_state_t;
-
 struct heap_li_hash
 {
     size_t operator()(const heap_list_item_t* li) const noexcept
@@ -158,16 +153,6 @@ struct heap_li_equal
     }
 };
 
-struct heap_recheck_state_t
-{
-    heap_entry_t *obj = NULL;
-    heap_entry_t *next_wr = NULL;
-    size_t total_reads = 0;
-    size_t sent_reads = 0;
-    size_t checked_reads = 0;
-    heap_entry_t *bad_wr = NULL;
-};
-
 using i64hash_t = robin_hood::hash<uint64_t>;
 using heap_inode_map_t = robin_hood::unordered_flat_set<heap_list_item_t*, heap_li_hash, heap_li_equal, 88>;
 using heap_block_index_t = robin_hood::unordered_flat_map<uint64_t,
@@ -176,7 +161,7 @@ using heap_mvcc_map_t = robin_hood::unordered_flat_map<object_id, heap_object_mv
 
 class blockstore_heap_t
 {
-    friend struct heap_entry_t;
+    friend class heap_entry_t;
 
     blockstore_disk_t *dsk = NULL;
     uint8_t* buffer_area = NULL;
@@ -197,11 +182,6 @@ class blockstore_heap_t
     uint64_t buffer_area_used_space = 0;
     uint64_t data_used_space = 0;
 
-    uint64_t live_entries = 0;
-    uint64_t live_memory = 0;
-    uint64_t garbage_entries = 0;
-    uint64_t garbage_memory = 0;
-
     uint64_t next_lsn = 0;
     uint32_t last_allocated_block = UINT32_MAX;
     heap_mvcc_map_t object_mvcc;
@@ -218,35 +198,29 @@ class blockstore_heap_t
 
     bool marked_used_blocks = false;
     bool recheck_queue_filled = false;
-    std::vector<heap_list_item_t*> postponed_items;
     std::set<uint32_t> recheck_modified_blocks;
     std::deque<heap_entry_t*> recheck_queue;
-    std::map<heap_entry_t*, heap_recheck_state_t> recheck_states;
-    size_t recheck_pending_reads = 0;
     int recheck_in_progress = 0;
     bool in_recheck = false;
     std::function<void(bool is_data, uint64_t offset, uint64_t len, uint8_t* buf, std::function<void()>)> recheck_cb;
     int recheck_queue_depth = 0;
 
+    void inode_map_put(void* & inode_idx, heap_list_item_t* li);
+    void inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_list_item_t* & li, uint64_t stripe);
+    void inode_map_free(void* inode_idx);
+    void inode_map_iterate(void* & inode_idx, std::function<void(heap_list_item_t*)> cb);
+    void inode_map_replace(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* new_li);
+    void inode_map_erase(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* li);
+
     uint64_t get_pg_id(inode_t inode, uint64_t stripe);
     bool validate_object(heap_entry_t *obj);
     void fill_recheck_queue();
-    void recheck_drop_entries(heap_entry_t *obj, heap_entry_t *bad_wr);
-    void recheck_start_reads(heap_recheck_state_t *st);
     int mark_used_blocks();
-    void init_free_bad_entry(heap_entry_t *wr);
-    void init_erase_bad_entry(heap_list_item_t *li);
-    bool init_erase_double_claim(heap_list_item_t *prev_li, heap_list_item_t *cur_li);
-    void recheck_full_gc();
     void recheck_buffer(heap_entry_t *cwr, uint8_t *buf);
     void defragment_block(uint32_t block_num);
-    void reshard_add(heap_reshard_state_t *st, heap_list_item_t *li);
 
-    void gc_block(heap_block_info_t & inf);
     int allocate_entry(uint32_t entry_size, uint32_t *block_num, bool allow_last_free);
-    void insert_list_items(heap_list_item_t** v, size_t count, bool postpone);
-    void remove_list_item(heap_list_item_t *li);
-    void unlink_list_item(heap_list_item_t *li);
+    void insert_list_item(heap_list_item_t *li);
     int add_entry(uint32_t wr_size, uint32_t *modified_block, bool allow_last_free,
         bool explicit_complete, std::function<void(heap_entry_t *wr)> fill_entry);
     int add_simple(heap_entry_t *obj, uint64_t version, uint32_t *modified_block, uint32_t entry_type);
@@ -267,29 +241,28 @@ public:
         std::function<void(uint32_t, uint32_t, uint8_t*)> handle_block);
     int load_blocks(uint64_t disk_offset, uint64_t size, uint8_t *buf,
         bool allow_corrupted, uint64_t &entries_loaded);
-    // finish loading - should be called after load_blocks
-    void finish_load();
+    // finish loading
+    int finish_load(bool allow_corrupted = false);
     // get blocks which are modified during loading and should be written to the disk
     // before finishing initialization if not R/O
     std::vector<uint32_t> get_recheck_modified_blocks();
     // recheck small write data after reading the database from disk
     bool recheck_small_writes(std::function<void(bool is_data, uint64_t offset, uint64_t len, uint8_t* buf, std::function<void()>)> read_buffer, int queue_depth);
-    int finish_recheck();
     // reshard database according to the pool's PG count
-    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit);
-    bool reshard_continue(void* reshard_state, uint64_t chunk_limit);
-    bool reshard_check(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size);
+    void reshard(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size);
     void set_no_inode_stats(const std::vector<uint64_t> & pool_ids);
     void recalc_inode_space_stats(uint64_t pool_id, bool per_inode);
     // read an object entry and lock it against removal
     // in the future, may become asynchronous
     heap_entry_t *lock_and_read_entry(object_id oid);
+    // re-read a locked object entry with the given lsn (pointer may be invalidated)
+    heap_entry_t *read_locked_entry(object_id oid, uint64_t lsn);
     // read an object entry without locking it
     heap_entry_t *read_entry(object_id oid);
     // unlock an entry
     bool unlock_entry(object_id oid);
     // set or verify checksums in a write request
-    bool calc_checksums(heap_entry_t *wr, uint8_t *data, bool set, uint32_t offset = UINT32_MAX, uint32_t len = UINT32_MAX);
+    bool calc_checksums(heap_entry_t *wr, uint8_t *data, bool set, uint32_t offset = 0, uint32_t len = 0);
     // set or verify raw block checksums
     bool calc_block_checksums(uint32_t *block_csums, uint8_t *data, uint8_t *bitmap, uint32_t start, uint32_t end,
         bool set, std::function<void(uint32_t, uint32_t, uint32_t)> bad_block_cb);
@@ -341,7 +314,6 @@ public:
     void start_block_write(uint32_t block_num);
     void complete_block_write(uint32_t block_num);
     void complete_lsn_write(uint64_t lsn);
-    bool is_lsn_completed(uint64_t lsn);
     uint64_t get_completed_lsn();
     uint64_t get_fsynced_lsn();
     void mark_lsn_fsynced(uint64_t lsn);
@@ -373,10 +345,6 @@ public:
     uint32_t get_compact_queue_size();
     uint32_t get_to_compact_count();
     uint64_t get_compacted_count();
-    uint64_t get_live_entries();
-    uint64_t get_live_memory();
-    uint64_t get_garbage_entries();
-    uint64_t get_garbage_memory();
 
     uint64_t entry_pos(uint32_t block_num, uint32_t offset);
     heap_entry_t *entry_from_pos(uint64_t entry_pos, bool allow_unallocated = false);

src/blockstore/blockstore_impl.cpp

@@ -23,7 +23,6 @@ blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *
         dsk.open_meta();
         dsk.open_journal();
         dsk.calc_lengths();
-        dsk.check_lengths();
     }
     catch (std::exception & e)
     {
@@ -32,13 +31,16 @@ blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *
     }
     meta_superblock = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, dsk.meta_block_size);
     memset(meta_superblock, 0, dsk.meta_block_size);
+}
+
+void blockstore_impl_t::init()
+{
     flusher = new journal_flusher_t(this);
     if (dsk.inmemory_journal)
     {
         buffer_area = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, dsk.journal_len);
     }
     heap = new blockstore_heap_t(&dsk, buffer_area, log_level);
-    ringloop->wakeup();
 }
 
 blockstore_impl_t::~blockstore_impl_t()
@@ -101,7 +103,6 @@ void blockstore_impl_t::loop()
         unsigned initial_ring_space = ringloop->space_left();
         int op_idx = 0, new_idx = 0;
         bool has_unfinished_writes = false;
-        bool has_unfinished_sync = false;
         for (; op_idx < submit_queue.size(); op_idx++, new_idx++)
         {
             auto op = submit_queue[op_idx];
@@ -139,13 +140,7 @@ void blockstore_impl_t::loop()
             else if (op->opcode == BS_OP_SYNC)
             {
                 // syncs only completed writes, so doesn't have to be blocked by anything
-                if (!has_unfinished_sync)
-                {
-                    wr_st = continue_sync(op);
-                    has_unfinished_sync = (wr_st != 2);
-                }
-                else
-                    wr_st = 0;
+                wr_st = continue_sync(op);
             }
             else if (op->opcode == BS_OP_STABLE || op->opcode == BS_OP_ROLLBACK)
             {
@@ -161,7 +156,9 @@ void blockstore_impl_t::loop()
                     wr_st = 2;
                 }
                 else
+                {
                     wr_st = 0;
+                }
             }
             if (wr_st == 2)
             {
@@ -198,12 +195,12 @@ void blockstore_impl_t::loop()
             heap->start_block_write(block_num);
             mb.sent = true;
         }
-        pending_modified_blocks.clear();
         int ret = ringloop->submit();
         if (ret < 0)
         {
             throw std::runtime_error(std::string("io_uring_submit: ") + strerror(-ret));
         }
+        pending_modified_blocks.clear();
         if ((initial_ring_space - ringloop->space_left()) > 0)
         {
             live = true;
@@ -328,13 +325,9 @@ void blockstore_impl_t::process_list(blockstore_op_t *op)
         FINISH_OP(op);
         return;
     }
-    // Check if the DB is sharded correctly
-    if (!heap->reshard_check(INODE_POOL(min_inode), pg_count, pg_stripe_size))
-    {
-        op->retval = -EAGAIN;
-        FINISH_OP(op);
-        return;
-    }
+    // Check if the DB needs resharding
+    // (we don't know about PGs from the beginning, we only create "shards" here)
+    heap->reshard(INODE_POOL(min_inode), pg_count, pg_stripe_size);
     obj_ver_id *result = NULL;
     size_t stable_count = 0, unstable_count = 0;
     int res = heap->list_objects(list_pg, op->min_oid, op->max_oid, &result, &stable_count, &unstable_count);
@@ -401,13 +394,3 @@ std::string blockstore_impl_t::get_op_diag(blockstore_op_t *op)
         snprintf(buf, sizeof(buf), "state=%d", priv->op_state);
     return std::string(buf);
 }
-
-void* blockstore_impl_t::reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit)
-{
-    return heap->reshard_start(pool, pg_count, pg_stripe_size, chunk_limit);
-}
-
-bool blockstore_impl_t::reshard_continue(void *reshard_state, uint64_t chunk_limit)
-{
-    return heap->reshard_continue(reshard_state, chunk_limit);
-}

src/blockstore/blockstore_impl.h

@@ -78,7 +78,6 @@ public:
     // Suitable only for server SSDs with capacitors, requires disabled data and journal fsyncs
     int immediate_commit = IMMEDIATE_NONE;
     bool inmemory_meta = false;
-    bool skip_corrupted_meta_entries = false;
     uint32_t meta_write_recheck_parallelism = 0;
     // Maximum and minimum flusher count
     unsigned max_flusher_count = 0, min_flusher_count = 0;
@@ -143,6 +142,7 @@ public:
     int metadata_buf_size;
     blockstore_init_meta* metadata_init_reader;
 
+    void init();
     void check_wait(blockstore_op_t *op);
     void init_op(blockstore_op_t *op);
 
@@ -190,9 +190,6 @@ public:
     void parse_config(blockstore_config_t & config);
     void parse_config(blockstore_config_t & config, bool init);
 
-    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit);
-    bool reshard_continue(void *reshard_state, uint64_t chunk_limit);
-
     // Event loop
     void loop();
 
@@ -229,9 +226,4 @@ public:
     uint64_t get_free_block_count();
     inline uint32_t get_bitmap_granularity() { return dsk.bitmap_granularity; }
     inline uint64_t get_journal_size() { return dsk.journal_len; }
-
-    inline uint64_t get_live_entries() { return heap->get_live_entries(); }
-    inline uint64_t get_live_memory() { return heap->get_live_memory(); }
-    inline uint64_t get_garbage_entries() { return heap->get_garbage_entries(); }
-    inline uint64_t get_garbage_memory() { return heap->get_garbage_memory(); }
 };

src/blockstore/blockstore_init.cpp

@@ -72,6 +72,7 @@ resume_1:
     }
     if (is_zero((uint64_t*)bs->meta_superblock, bs->dsk.meta_block_size))
     {
+        bs->dsk.check_lengths();
         {
             blockstore_meta_header_v3_t *hdr = (blockstore_meta_header_v3_t *)bs->meta_superblock;
             hdr->zero = 0;
@@ -140,12 +141,12 @@ resume_1:
             hdr->bitmap_granularity != bs->dsk.bitmap_granularity ||
             hdr->data_csum_type != bs->dsk.data_csum_type ||
             hdr->csum_block_size != bs->dsk.csum_block_size ||
-            hdr->meta_area_size != bs->dsk.meta_area_size)
+            hdr->meta_area_size > bs->dsk.meta_area_size)
         {
             printf(
                 "Configuration stored in metadata superblock"
                 " (meta_block_size=%u, data_block_size=%u, bitmap_granularity=%u, data_csum_type=%u, csum_block_size=%u, meta_area_size=%ju)"
-                " differs from OSD configuration (%ju/%ju/%u, %u/%u, %ju).\n",
+                " differs from OSD configuration (%u/%u/%u, %u/%u, %ju).\n",
                 hdr->meta_block_size, hdr->data_block_size, hdr->bitmap_granularity,
                 hdr->data_csum_type, hdr->csum_block_size, hdr->meta_area_size,
                 bs->dsk.meta_block_size, bs->dsk.data_block_size, bs->dsk.bitmap_granularity,
@@ -153,15 +154,15 @@ resume_1:
             );
             exit(1);
         }
-        uint32_t csum = hdr->header_csum;
-        hdr->header_csum = 0;
-        if (crc32c(0, hdr, sizeof(*hdr)) != csum)
+        bs->dsk.meta_area_size = hdr->meta_area_size;
+        if (bs->dsk.meta_format != hdr->version)
         {
-            printf("Metadata header is corrupt (checksum mismatch).\n");
-            exit(1);
+            bs->dsk.meta_format = hdr->version;
+            bs->dsk.calc_lengths();
         }
-        hdr->header_csum = csum;
+        bs->dsk.check_lengths();
     }
+    bs->init();
     bs->heap->start_load(((blockstore_meta_header_v3_t *)bs->meta_superblock)->completed_lsn);
     if (bs->dsk.inmemory_journal)
     {
@@ -233,7 +234,7 @@ resume_4:
         {
             // Handle result
             uint64_t loaded = 0;
-            int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, bs->skip_corrupted_meta_entries, loaded);
+            int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, false, loaded);
             if (r != 0)
                 exit(1);
             entries_loaded += loaded;
@@ -247,8 +248,23 @@ resume_4:
         return 1;
     }
     // metadata read finished
-    bs->heap->finish_load();
-    printf("Metadata entries loaded: %ju, rechecking unfinished writes and garbage entries\n", entries_loaded);
+    printf("Metadata entries loaded: %ju, used blocks: %ju / %ju\n", entries_loaded, bs->heap->get_data_used_space() / bs->dsk.data_block_size, bs->dsk.block_count);
+    if (zero_on_init && !bs->dsk.disable_meta_fsync)
+    {
+        GET_SQE();
+        io_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
+        last_read_offset = 0;
+        data->iov = { 0 };
+        data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
+        submitted++;
+        bs->ringloop->submit();
+    resume_5:
+        if (submitted > 0)
+        {
+            wait_state = 5;
+            return 1;
+        }
+    }
     // asynchronous recheck
 resume_6:
     wait_state = 6;
@@ -262,7 +278,7 @@ resume_6:
         }
         GET_SQE();
         data->iov = (iovec){ buf, len };
-        data->callback = [offset, cb](ring_data_t *data)
+        data->callback = [this, offset, cb](ring_data_t *data)
         {
             if (data->res < 0)
             {
@@ -277,7 +293,7 @@ resume_6:
     }, bs->meta_write_recheck_parallelism);
     return 1;
 resume_7:
-    if (bs->heap->finish_recheck() != 0)
+    if (bs->heap->finish_load() != 0)
     {
         exit(1);
     }
@@ -285,11 +301,6 @@ resume_7:
     if (bs->readonly)
     {
         recheck_mod.clear();
-        printf("Actual metadata entries: %ju\n", bs->heap->get_live_entries());
-    }
-    else
-    {
-        printf("Actual metadata entries: %ju, clearing garbage in %zu metadata blocks\n", bs->heap->get_live_entries(), recheck_mod.size());
     }
     for (i = 0; i < recheck_mod.size(); i++)
     {
@@ -303,7 +314,7 @@ resume_8:
         uint32_t block_num = recheck_mod[i];
         uint64_t block_offset = bs->dsk.meta_offset + (uint64_t)(block_num+1) * bs->dsk.meta_block_size;
         data = ((ring_data_t*)sqe->user_data);
-        uint8_t *buf = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, bs->dsk.meta_block_size);
+        uint8_t *buf = (uint8_t*)malloc_or_die(bs->dsk.meta_block_size);
         bs->heap->get_meta_block(block_num, buf);
         data->iov = { buf, bs->dsk.meta_block_size };
         data->callback = [this, buf, block_offset](ring_data_t *data)
@@ -329,25 +340,5 @@ resume_9:
     }
     free(metadata_buffer);
     metadata_buffer = NULL;
-    if (!bs->dsk.disable_meta_fsync && !bs->readonly)
-    {
-        GET_SQE();
-        io_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
-        last_read_offset = 0;
-        data->iov = { 0 };
-        data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
-        submitted++;
-        bs->ringloop->submit();
-    resume_5:
-        if (submitted > 0)
-        {
-            wait_state = 5;
-            return 1;
-        }
-    }
-    printf("Loading finished. Data used: %ju / %ju bytes (%s / %s)\n",
-        bs->heap->get_data_used_space(), bs->dsk.block_count * bs->dsk.data_block_size,
-        format_size(bs->heap->get_data_used_space()).c_str(),
-        format_size(bs->dsk.block_count * bs->dsk.data_block_size).c_str());
     return 0;
 }

src/blockstore/blockstore_open.cpp

@@ -28,7 +28,6 @@ void blockstore_impl_t::parse_config(blockstore_config_t & config, bool init)
     throttle_target_parallelism = strtoull(config["throttle_target_parallelism"].c_str(), NULL, 10);
     throttle_threshold_us = strtoull(config["throttle_threshold_us"].c_str(), NULL, 10);
     perfect_csum_update = config["perfect_csum_update"] == "true" || config["perfect_csum_update"] == "1" || config["perfect_csum_update"] == "yes";
-    skip_corrupted_meta_entries = config["skip_corrupted_meta_entries"] == "true" || config["skip_corrupted_meta_entries"] == "1" || config["skip_corrupted_meta_entries"] == "yes";
     if (config["autosync_writes"] != "")
     {
         autosync_writes = strtoull(config["autosync_writes"].c_str(), NULL, 10);

src/blockstore/blockstore_read.cpp

@@ -22,7 +22,7 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
     uint64_t result_version = 0;
     bool found = false;
     uint32_t skip_csum = 0;
-    uint32_t blk_start = op->offset, blk_end = op->offset+op->len;
+    uint32_t blk_start = 0, blk_end = 0;
     bool need_skip = dsk.csum_block_size > dsk.bitmap_granularity && !perfect_csum_update;
     if (need_skip)
     {
@@ -32,28 +32,12 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
         if (blk_end % dsk.csum_block_size)
             blk_end += dsk.csum_block_size - (blk_end % dsk.csum_block_size);
     }
-    bool need_wait = false;
     heap->iterate_with_stable(obj, obj->lsn, [&](heap_entry_t *wr, bool stable)
     {
         if (wr->type() == BS_HEAP_DELETE)
         {
             return false;
         }
-        if (!heap->is_lsn_completed(wr->lsn))
-        {
-            if (wr->type() == BS_HEAP_BIG_INTENT && wr->big_intent().offset < blk_end && wr->big_intent().offset+wr->big_intent().len > blk_start ||
-                wr->type() == BS_HEAP_INTENT_WRITE && wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start)
-            {
-                // Wait until intent write is completed
-                need_wait = true;
-                return false;
-            }
-            else if (wr->type() == BS_HEAP_SMALL_WRITE && wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start)
-            {
-                // Skip entry and read the previous one
-                return true;
-            }
-        }
         if (op->version >= wr->version && !found)
         {
             found = true;
@@ -63,6 +47,12 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
                 memcpy(op->bitmap, wr->get_ext_bitmap(heap), dsk.clean_entry_bitmap_size);
             }
         }
+        if (need_skip && wr->lsn < heap->get_completed_lsn() &&
+            (wr->type() == BS_HEAP_BIG_INTENT && wr->big_intent().offset < blk_end && wr->big_intent().offset+wr->big_intent().len > blk_start ||
+            wr->type() == BS_HEAP_INTENT_WRITE && wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start))
+        {
+            skip_csum = COPY_BUF_SKIP_CSUM;
+        }
         if (op->version >= wr->version)
         {
             fulfilled += prepare_read(PRIV(op)->read_vec, obj, wr, op->offset, op->offset+op->len,
@@ -75,23 +65,13 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
                 return false;
             }
         }
-        if (need_skip && wr->type() == BS_HEAP_SMALL_WRITE &&
+        if (need_skip && (wr->type() == BS_HEAP_SMALL_WRITE || wr->type() == BS_HEAP_INTENT_WRITE) &&
             wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start)
         {
-            // Small write may mutate big write checksums during flush
             skip_csum = COPY_BUF_SKIP_CSUM;
         }
         return true;
     });
-    if (need_wait)
-    {
-undo_wait:
-        // Need to wait. undo added requests, unlock lsn
-        heap->unlock_entry(op->oid);
-        free_read_buffers(rv);
-        rv.clear();
-        return 0;
-    }
     if (!found)
     {
         // May happen if there are entries but all of them are > requested version
@@ -104,7 +84,11 @@ undo_wait:
     assert(fulfilled == op->len);
     if (!fulfill_read(op))
     {
-        goto undo_wait;
+        // Need to wait. undo added requests, unlock lsn
+        heap->unlock_entry(op->oid);
+        free_read_buffers(rv);
+        rv.clear();
+        return 0;
     }
     op->version = result_version;
     if (!PRIV(op)->pending_ops)
@@ -462,10 +446,6 @@ int blockstore_impl_t::read_bitmap(object_id oid, uint64_t target_version, void
         {
             if (target_version >= wr->version)
             {
-                if (wr->type() == BS_HEAP_DELETE)
-                {
-                    return false;
-                }
                 found = true;
                 if (result_version)
                 {

src/blockstore/blockstore_stable.cpp

@@ -16,7 +16,6 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
     else if (priv->op_state == 5) goto resume_5;
     assert(!priv->op_state);
     op->retval = 0;
-    PRIV(op)->lsn = 0;
     priv->modified_block = priv->modified_block2 = UINT32_MAX;
     for (priv->stab_pos = 0; priv->stab_pos < op->len; priv->stab_pos++)
     {
@@ -37,12 +36,6 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
                 FINISH_OP(op);
                 return 2;
             }
-            if (res == ENOENT)
-            {
-                op->retval = -ENOENT;
-                FINISH_OP(op);
-                return 2;
-            }
             if (res == ENOSPC)
             {
                 if (!heap->get_to_compact_count())
@@ -64,9 +57,9 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
             }
             assert(res == 0);
         }
-resume_1:
         if (priv->modified_block != UINT32_MAX && priv->modified_block2 != priv->modified_block)
         {
+resume_1:
             BS_SUBMIT_CHECK_SQES(1);
             prepare_meta_block_write(priv->modified_block);
 resume_2:

src/blockstore/blockstore_sync.cpp

@@ -9,7 +9,6 @@ int blockstore_impl_t::continue_sync(blockstore_op_t *op)
     if (!PRIV(op)->op_state)
     {
         op->retval = 0;
-        PRIV(op)->lsn = 0;
     }
     int res = do_sync(op, 0);
     if (res == 2)
@@ -105,8 +104,7 @@ int blockstore_impl_t::do_sync(blockstore_op_t *op, int base_state)
         unsynced_data_write_count = unsynced_buffer_write_count = unsynced_meta_write_count = 0;
         return 2;
     }
-    assert(!PRIV(op)->lsn);
-    PRIV(op)->lsn = heap->get_completed_lsn();
+    PRIV(op)->modified_block = heap->get_completed_lsn();
     if (!submit_fsyncs(PRIV(op)->pending_ops))
     {
         PRIV(op)->wait_detail = 1;
@@ -120,6 +118,6 @@ resume_1:
         return 1;
     }
 resume_2:
-    heap->mark_lsn_fsynced(PRIV(op)->lsn);
+    heap->mark_lsn_fsynced(PRIV(op)->modified_block);
     return 2;
 }

src/blockstore/blockstore_write.cpp

@@ -22,24 +22,21 @@ void blockstore_impl_t::prepare_meta_block_write(uint32_t modified_block)
     ring_data_t *data = ((ring_data_t*)sqe->user_data);
     uint8_t *buf = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, dsk.meta_block_size);
     data->iov = (struct iovec){ buf, (size_t)dsk.meta_block_size };
-    data->callback = [this, modified_block](ring_data_t *data)
+    data->callback = [this, modified_block, buf](ring_data_t *data)
     {
+        free(buf);
         live = true;
         if (data->res != data->iov.iov_len)
         {
             // FIXME: our state becomes corrupted after a write error. maybe do something better than just die
             disk_error_abort("data write", data->res, data->iov.iov_len);
         }
-        auto it = modified_blocks.find(modified_block);
-        assert(it != modified_blocks.end());
-        free(it->second.buf);
-        modified_blocks.erase(it);
+        modified_blocks.erase(modified_block);
         heap->complete_block_write(modified_block);
         ringloop->wakeup();
     };
-    assert(((uint64_t)modified_block+2)*dsk.meta_block_size <= dsk.meta_area_size);
     io_uring_prep_writev(
-        sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + ((uint64_t)modified_block+1)*dsk.meta_block_size
+        sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + (modified_block+1)*dsk.meta_block_size
     );
     unsynced_meta_write_count++;
     pending_modified_blocks.push_back(modified_block);
@@ -178,7 +175,6 @@ enospc:
         ring_data_t *data = ((ring_data_t*)sqe->user_data);
         data->iov = (struct iovec){ op->buf, op->len };
         data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-        assert(loc+op->offset+op->len <= dsk.block_count*dsk.data_block_size);
         io_uring_prep_writev(sqe, dsk.data_fd, &data->iov, 1, dsk.data_offset + loc + op->offset);
         PRIV(op)->pending_ops++;
         write_iodepth++;
@@ -253,12 +249,13 @@ enospc:
             goto enospc;
         assert(res == 0);
         PRIV(op)->lsn = obj->lsn;
+        if (op->len)
+            heap->use_buffer_area(op->oid.inode, loc, op->len);
         prepare_meta_block_write(PRIV(op)->modified_block);
         PRIV(op)->pending_ops++;
         if (op->len > 0)
         {
             // Prepare buffered data write
-            heap->use_buffer_area(op->oid.inode, loc, op->len);
             if (dsk.inmemory_journal)
             {
                 memcpy((uint8_t*)buffer_area + loc, op->buf, op->len);
@@ -266,7 +263,6 @@ enospc:
             BS_SUBMIT_GET_SQE(sqe2, data2);
             data2->iov = (struct iovec){ op->buf, op->len };
             data2->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-            assert(loc+op->len <= dsk.journal_len);
             io_uring_prep_writev(sqe2, dsk.journal_fd, &data2->iov, 1, dsk.journal_offset + loc);
             PRIV(op)->pending_ops++;
         }
@@ -352,7 +348,6 @@ resume_12:
     }
 resume_4:
     {
-        BS_SUBMIT_CHECK_SQES(1);
         auto obj = heap->read_entry(op->oid);
         int res = 0;
         if (PRIV(op)->write_type == _REDIRECT_INTENT)
@@ -409,12 +404,11 @@ resume_6:
         if (ref_us > exec_us + throttle_threshold_us)
         {
             // Pause reply
-            PRIV(op)->pending_ops++;
             PRIV(op)->op_state = 7;
             // Remember that the timer can in theory be called right here
             tfd->set_timer_us(ref_us-exec_us, false, [this, op](int timer_id)
             {
-                PRIV(op)->pending_ops--;
+                PRIV(op)->op_state = 8;
                 ringloop->wakeup();
             });
             return 1;
@@ -456,7 +450,6 @@ resume_10:
     BS_SUBMIT_GET_SQE(sqe, data);
     data->iov = (struct iovec){ op->buf, op->len };
     data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-    assert(PRIV(op)->location + op->offset <= dsk.block_count*dsk.data_block_size);
     io_uring_prep_writev(sqe, dsk.data_fd, &data->iov, 1, dsk.data_offset + PRIV(op)->location + op->offset);
     if (dsk.use_atomic_flag)
         sqe->rw_flags = RWF_ATOMIC;

src/blockstore/ondisk_formats.h

@@ -141,7 +141,7 @@ struct __attribute__((__packed__)) journal_entry
 inline uint32_t je_crc32(journal_entry *je)
 {
     // 0x48674bc7 = crc32(4 zero bytes)
-    return je->size < 4 ? 0 : crc32c(0x48674bc7, ((uint8_t*)je)+4, je->size-4);
+    return crc32c(0x48674bc7, ((uint8_t*)je)+4, je->size-4);
 }
 
 // "VITAstor"

src/blockstore/v1/flush.cpp

@@ -520,7 +520,6 @@ resume_2:
                 await_sqe(15);
                 data->iov = (struct iovec){ it->buf, (size_t)it->len };
                 data->callback = simple_callback_w;
-                assert(clean_loc+it->offset+it->len <= bs->dsk.block_count*bs->dsk.data_block_size);
                 io_uring_prep_writev(
                     sqe, bs->dsk.data_fd, &data->iov, 1, bs->dsk.data_offset + clean_loc + it->offset
                 );
@@ -750,7 +749,6 @@ bool journal_flusher_co::write_meta_block(flusher_meta_write_t & meta_block, int
     await_sqe(0);
     data->iov = (struct iovec){ meta_block.buf, (size_t)bs->dsk.meta_block_size };
     data->callback = simple_callback_w;
-    assert(bs->dsk.meta_block_size + meta_block.sector + bs->dsk.meta_block_size <= bs->dsk.meta_area_size);
     io_uring_prep_writev(
         sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bs->dsk.meta_block_size + meta_block.sector
     );

src/blockstore/v1/impl.cpp

@@ -407,77 +407,32 @@ blockstore_clean_db_t& blockstore_impl_t::clean_db_shard(object_id oid)
     return clean_db_shards[(pool_id << (64-POOL_ID_BITS)) | pg_num];
 }
 
-struct bs_reshard_state_t
+void blockstore_impl_t::reshard_clean_db(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size)
 {
-    int state = 0;
-    uint64_t pool_id = 0;
-    uint32_t pg_count = 0;
-    uint32_t pg_stripe_size = 0;
-    uint64_t chunk_size = 0;
-    std::map<pool_pg_id_t, blockstore_clean_db_t> old_shards;
+    uint64_t pool_id = (uint64_t)pool;
     std::map<pool_pg_id_t, blockstore_clean_db_t> new_shards;
-    std::map<pool_pg_id_t, blockstore_clean_db_t>::iterator sh_it;
-    blockstore_clean_db_t::iterator obj_it;
-};
-
-void* blockstore_impl_t::reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit)
-{
-    auto & settings = clean_db_settings[pool];
-    if (settings.pg_count == pg_count && settings.pg_stripe_size == pg_stripe_size)
-    {
-        return NULL;
-    }
-    bs_reshard_state_t *st = new bs_reshard_state_t;
-    st->state = 0;
-    st->pool_id = pool;
-    st->pg_count = pg_count;
-    st->pg_stripe_size = pg_stripe_size;
-    auto sh_it = clean_db_shards.lower_bound((st->pool_id << (64-POOL_ID_BITS)));
+    auto sh_it = clean_db_shards.lower_bound((pool_id << (64-POOL_ID_BITS)));
     while (sh_it != clean_db_shards.end() &&
-        (sh_it->first >> (64-POOL_ID_BITS)) == st->pool_id)
+        (sh_it->first >> (64-POOL_ID_BITS)) == pool_id)
     {
-        st->old_shards[sh_it->first] = std::move(sh_it->second);
+        for (auto & pair: sh_it->second)
+        {
+            // like map_to_pg()
+            uint64_t pg_num = (pair.first.stripe / pg_stripe_size) % pg_count + 1;
+            uint64_t shard_id = (pool_id << (64-POOL_ID_BITS)) | pg_num;
+            new_shards[shard_id][pair.first] = pair.second;
+        }
         clean_db_shards.erase(sh_it++);
     }
-    bool finished = reshard_continue(st, chunk_limit);
-    return finished ? NULL : st;
-}
-
-bool blockstore_impl_t::reshard_continue(void *reshard_state, uint64_t chunk_limit)
-{
-    bs_reshard_state_t *st = (bs_reshard_state_t*)reshard_state;
-    uint64_t chunk_size = 0;
-    if (st->state == 1)
-        goto resume_1;
-    for (st->sh_it = st->old_shards.begin(); st->sh_it != st->old_shards.end(); )
-    {
-        for (st->obj_it = st->sh_it->second.begin(); st->obj_it != st->sh_it->second.end(); st->obj_it++)
-        {
-            if (chunk_limit > 0 && chunk_size >= chunk_limit)
-            {
-                st->state = 1;
-                return false;
-            }
-resume_1:
-            // like map_to_pg()
-            uint64_t pg_num = (st->obj_it->first.stripe / st->pg_stripe_size) % st->pg_count + 1;
-            uint64_t shard_id = (st->pool_id << (64-POOL_ID_BITS)) | pg_num;
-            st->new_shards[shard_id][st->obj_it->first] = st->obj_it->second;
-            chunk_size++;
-        }
-        st->old_shards.erase(st->sh_it++);
-    }
-    for (auto sh_it = st->new_shards.begin(); sh_it != st->new_shards.end(); sh_it++)
+    for (sh_it = new_shards.begin(); sh_it != new_shards.end(); sh_it++)
     {
         auto & to = clean_db_shards[sh_it->first];
         to.swap(sh_it->second);
     }
-    clean_db_settings[st->pool_id] = (pool_shard_settings_t){
-        .pg_count = st->pg_count,
-        .pg_stripe_size = st->pg_stripe_size,
+    clean_db_settings[pool_id] = (pool_shard_settings_t){
+        .pg_count = pg_count,
+        .pg_stripe_size = pg_stripe_size,
     };
-    delete st;
-    return true;
 }
 
 void blockstore_impl_t::process_list(blockstore_op_t *op)
@@ -510,10 +465,7 @@ void blockstore_impl_t::process_list(blockstore_op_t *op)
                 sh_it->second.pg_count != pg_count ||
                 sh_it->second.pg_stripe_size != pg_stripe_size)
             {
-                // Sharding mismatch
-                op->retval = -EAGAIN;
-                FINISH_OP(op);
-                return;
+                reshard_clean_db(pool_id, pg_count, pg_stripe_size);
             }
             first_shard = last_shard = ((uint64_t)pool_id << (64-POOL_ID_BITS)) | list_pg;
         }
@@ -855,29 +807,4 @@ std::string blockstore_impl_t::get_op_diag(blockstore_op_t *op)
     return std::string(buf);
 }
 
-uint64_t blockstore_impl_t::get_live_entries()
-{
-    return used_blocks;
-}
-
-uint64_t blockstore_impl_t::get_live_memory()
-{
-    uint64_t used = 0;
-    for (auto & kv: clean_db_shards)
-    {
-        used += kv.second.size() * sizeof(blockstore_clean_db_t::value_type);
-    }
-    return used;
-}
-
-uint64_t blockstore_impl_t::get_garbage_entries()
-{
-    return dirty_db.size();
-}
-
-uint64_t blockstore_impl_t::get_garbage_memory()
-{
-    return (sizeof(obj_ver_id) + sizeof(dirty_entry) + 32) * dirty_db.size();
-}
-
 } // namespace v1

src/blockstore/v1/impl.h

@@ -202,6 +202,7 @@ class blockstore_impl_t: public blockstore_i
     uint8_t* get_clean_entry_bitmap(uint64_t block_loc, int offset);
 
     blockstore_clean_db_t& clean_db_shard(object_id oid);
+    void reshard_clean_db(pool_id_t pool_id, uint32_t pg_count, uint32_t pg_stripe_size);
     void recalc_inode_space_stats(uint64_t pool_id, bool per_inode);
 
     // Journaling
@@ -287,10 +288,6 @@ public:
     void parse_config(blockstore_config_t & config);
     void parse_config(blockstore_config_t & config, bool init);
 
-    // Reshard database for a pool
-    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit);
-    bool reshard_continue(void *reshard_state, uint64_t chunk_limit);
-
     // Event loop
     void loop();
 
@@ -332,10 +329,6 @@ public:
     inline uint64_t get_free_block_count() { return dsk.block_count - used_blocks; }
     inline uint32_t get_bitmap_granularity() { return dsk.disk_alignment; }
     inline uint64_t get_journal_size() { return dsk.journal_len; }
-    uint64_t get_live_entries();
-    uint64_t get_live_memory();
-    uint64_t get_garbage_entries();
-    uint64_t get_garbage_memory();
 };
 
 } // namespace v1

src/blockstore/v1/init.cpp

@@ -189,7 +189,7 @@ resume_1:
             printf(
                 "Configuration stored in metadata superblock"
                 " (meta_block_size=%u, data_block_size=%u, bitmap_granularity=%u, data_csum_type=%u, csum_block_size=%u)"
-                " differs from OSD configuration (%ju/%ju/%u, %u/%u).\n",
+                " differs from OSD configuration (%u/%u/%u, %u/%u).\n",
                 hdr->meta_block_size, hdr->data_block_size, hdr->bitmap_granularity,
                 hdr->data_csum_type, hdr->csum_block_size,
                 bs->dsk.meta_block_size, bs->dsk.data_block_size, bs->dsk.bitmap_granularity,

src/blockstore/v1/journal.cpp

@@ -193,7 +193,6 @@ void blockstore_impl_t::prepare_journal_sector_write(int cur_sector, blockstore_
             (size_t)journal.block_size
         };
         data->callback = [this, flush_id = journal.submit_id](ring_data_t *data) { handle_journal_write(data, flush_id); };
-        assert(journal.sector_info[cur_sector].offset+journal.block_size <= dsk.journal_len);
         io_uring_prep_writev(
             sqe, dsk.journal_fd, &data->iov, 1, journal.offset + journal.sector_info[cur_sector].offset
         );

src/blockstore/v1/read.cpp

@@ -620,7 +620,8 @@ bool blockstore_impl_t::fulfill_clean_read(blockstore_op_t *read_op, uint64_t &
     else if (from_journal)
     {
         // Don't scan bitmap - journal writes don't have holes (internal bitmap)!
-        uint8_t *csum = !dsk.csum_block_size ? 0 : (clean_entry_bitmap + dsk.clean_entry_bitmap_size);
+        uint8_t *csum = !dsk.csum_block_size ? 0 : (clean_entry_bitmap + dsk.clean_entry_bitmap_size +
+            item_start/dsk.csum_block_size*(dsk.data_csum_type & 0xFF));
         if (!fulfill_read(read_op, fulfilled, item_start, item_end,
             (BS_ST_BIG_WRITE | BS_ST_STABLE), 0, clean_loc + item_start, 0, csum, dyn_data))
         {

src/blockstore/v1/write.cpp

@@ -183,7 +183,7 @@ bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
         uint32_t end = (op->offset+op->len-1) / dsk.csum_block_size;
         auto fn = state & BS_ST_BIG_WRITE ? crc32c_pad : crc32c_nopad;
         if (start == end)
-            data_csums[0] = fn(0, op->buf, op->len, op->offset - start*dsk.csum_block_size, (end+1)*dsk.csum_block_size - (op->offset+op->len));
+            data_csums[0] = fn(0, op->buf, op->len, op->offset - start*dsk.csum_block_size, end*dsk.csum_block_size - (op->offset+op->len));
         else
         {
             // First block
@@ -368,9 +368,9 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
         }
         data->iov.iov_len = op->len + stripe_offset + stripe_end; // to check it in the callback
         data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
-        const uint64_t write_offset = (loc * dsk.data_block_size) + op->offset - stripe_offset;
-        assert(write_offset+op->len+stripe_offset+stripe_end <= dsk.block_count*dsk.data_block_size);
-        io_uring_prep_writev(sqe, dsk.data_fd, PRIV(op)->iov_zerofill, vcnt, dsk.data_offset + write_offset);
+        io_uring_prep_writev(
+            sqe, dsk.data_fd, PRIV(op)->iov_zerofill, vcnt, dsk.data_offset + (loc * dsk.data_block_size) + op->offset - stripe_offset
+        );
         PRIV(op)->pending_ops = 1;
         if (!(dirty_it->second.state & BS_ST_INSTANT))
         {
@@ -495,8 +495,9 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
                 .op = op,
             });
             data2->callback = [this, flush_id = journal.submit_id](ring_data_t *data) { handle_journal_write(data, flush_id); };
-            assert(journal.next_free+op->len <= dsk.journal_len);
-            io_uring_prep_writev(sqe2, dsk.journal_fd, &data2->iov, 1, journal.offset + journal.next_free);
+            io_uring_prep_writev(
+                sqe2, dsk.journal_fd, &data2->iov, 1, journal.offset + journal.next_free
+            );
             PRIV(op)->pending_ops++;
         }
         else

src/client/CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required(VERSION 2.8...3.30)
+cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
@@ -12,7 +12,7 @@ if (RDMACM_LIBRARIES)
 	set(MSGR_RDMACM "msgr_rdmacm.cpp")
 endif (RDMACM_LIBRARIES)
 add_library(vitastor_common STATIC
-	../util/epoll_manager.cpp etcd_state_client.cpp messenger.cpp msgr_iothread.cpp ../util/addr_util.cpp
+	../util/epoll_manager.cpp etcd_state_client.cpp messenger.cpp ../util/addr_util.cpp
 	msgr_stop.cpp msgr_op.cpp msgr_send.cpp msgr_receive.cpp ../util/ringloop.cpp ../../json11/json11.cpp
 	http_client.cpp osd_ops.cpp pg_states.cpp ../util/timerfd_manager.cpp ../util/str_util.cpp ../util/json_util.cpp ${MSGR_RDMA} ${MSGR_RDMACM}
 )
@@ -52,6 +52,9 @@ if (${WITH_FIO})
 		../util/rw_blocking.cpp
 		../util/addr_util.cpp
 	)
+	target_link_libraries(fio_vitastor_sec
+		tcmalloc_minimal
+	)
 endif (${WITH_FIO})
 
 # vitastor-nbd
@@ -98,7 +101,6 @@ add_executable(test_cluster_client
 	pg_states.cpp osd_ops.cpp cluster_client.cpp cluster_client_list.cpp cluster_client_wb.cpp msgr_op.cpp ../test/mock/messenger.cpp msgr_stop.cpp
 	etcd_state_client.cpp ../util/timerfd_manager.cpp ../util/addr_util.cpp ../util/str_util.cpp ../util/json_util.cpp ../../json11/json11.cpp
 )
-target_link_libraries(test_cluster_client ${LIBURING_LIBRARIES})
 target_compile_definitions(test_cluster_client PUBLIC -D__MOCK__)
 target_include_directories(test_cluster_client BEFORE PUBLIC ${CMAKE_SOURCE_DIR}/src/test/mock)
 add_dependencies(build_tests test_cluster_client)

src/client/cluster_client.cpp

@@ -27,7 +27,7 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
     msgr.ringloop = ringloop;
     msgr.repeer_pgs = [this](osd_num_t peer_osd)
     {
-        if (msgr.osd_peers.find(peer_osd) != msgr.osd_peers.end())
+        if (msgr.osd_peer_fds.find(peer_osd) != msgr.osd_peer_fds.end())
         {
             // peer_osd just connected
             continue_ops();
@@ -47,8 +47,8 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
     msgr.exec_op = [this](osd_op_t *op)
     {
         // Garbage in
-        fprintf(stderr, "Can't handle incoming operation from client %lu\n", op->client_id);
-        msgr.stop_client(op->client_id);
+        fprintf(stderr, "Incoming garbage from peer %d\n", op->peer_fd);
+        msgr.stop_client(op->peer_fd);
         delete op;
     };
     msgr.parse_config(config);
@@ -156,7 +156,7 @@ void cluster_client_t::continue_raw_ops(osd_num_t peer_osd)
     {
         auto op = it->second;
         op->op_type = OSD_OP_OUT;
-        op->client_id = msgr.osd_peers.at(peer_osd)->client_id;
+        op->peer_fd = msgr.osd_peer_fds.at(peer_osd);
         msgr.outbox_push(op);
         raw_ops.erase(it++);
     }
@@ -590,7 +590,7 @@ void cluster_client_t::on_change_pool_config_hook()
         {
             if (log_level > 2 && pg_counts[pool_item.first])
             {
-                fprintf(stderr, "Pool %u (%s) PG count changed from %lu to %lu\n", pool_item.first, pool_item.second.name.c_str(),
+                printf("Pool %u (%s) PG count changed from %lu to %lu\n", pool_item.first, pool_item.second.name.c_str(),
                     pg_counts[pool_item.first], pool_item.second.real_pg_count);
             }
             // At this point, all pool operations should have been suspended
@@ -868,16 +868,15 @@ void cluster_client_t::execute_cas(cluster_op_t *op)
     {
         int expected = part->req.hdr.opcode == OSD_OP_DELETE ? 0 : part->req.rw.len;
         op->retval = part->reply.hdr.retval;
-        if (op->retval != expected && op->retval >= 0)
-            op->retval = -EIO;
+        op->retval = op->retval == expected ? 0 : (op->retval >= 0 ? -EIO : op->retval);
         op->retval = op->retval == -EPIPE ? -EINTR : op->retval;
-        auto peer_it = msgr.osd_peers.find(op->parts[0].osd_num);
+        auto peer_it = msgr.osd_peer_fds.find(op->parts[0].osd_num);
         if (op->retval != 0 || (op->flags & OP_IMMEDIATE_COMMIT))
         {
             auto cb = std::move(op->callback);
             cb(op);
         }
-        else if (peer_it == msgr.osd_peers.end())
+        else if (peer_it == msgr.osd_peer_fds.end())
         {
             // Care must be taken to make sure that the client doesn't reconnect to the OSD
             // before executing the previously completed operation callback (!)
@@ -888,23 +887,20 @@ void cluster_client_t::execute_cas(cluster_op_t *op)
         else
         {
             // CAS writes have a built-in sync
-            osd_client_t *cl = peer_it->second;
+            auto peer_fd = peer_it->second;
             *part = (osd_op_t){
                 .op_type = OSD_OP_OUT,
-                .client_id = cl->client_id,
+                .peer_fd = peer_fd,
                 .req = {
                     .hdr = {
                         .magic = SECONDARY_OSD_OP_MAGIC,
                         .opcode = OSD_OP_SYNC,
                     },
                 },
-                .callback = [op](osd_op_t *part)
+                .callback = [this, op](osd_op_t *part)
                 {
-                    if (part->reply.hdr.retval != 0)
-                    {
-                        op->retval = part->reply.hdr.retval;
-                        op->retval = op->retval == -EPIPE ? -EINTR : op->retval;
-                    }
+                    op->retval = part->reply.hdr.retval;
+                    op->retval = op->retval == -EPIPE ? -EINTR : op->retval;
                     auto cb = std::move(op->callback);
                     cb(op);
                 },
@@ -1004,11 +1000,11 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
 
 void cluster_client_t::execute_raw(osd_num_t osd_num, osd_op_t *op)
 {
-    auto peer_it = msgr.osd_peers.find(osd_num);
-    if (peer_it != msgr.osd_peers.end())
+    auto fd_it = msgr.osd_peer_fds.find(osd_num);
+    if (fd_it != msgr.osd_peer_fds.end())
     {
         op->op_type = OSD_OP_OUT;
-        op->client_id = peer_it->second->client_id;
+        op->peer_fd = fd_it->second;
         msgr.outbox_push(op);
     }
     else
@@ -1119,13 +1115,6 @@ resume_2:
         // Finished successfully
         // Even if the PG count has changed in meanwhile we treat it as success
         // because if some operations were invalid for the new PG count we'd get errors
-        if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
-        {
-            // Copy part bitmaps only after finishing all part reads
-            for (auto & part: op->parts)
-                if ((part.flags & (PART_SENT|PART_DONE|PART_VALID)) == (PART_SENT|PART_DONE|PART_VALID))
-                    copy_part_bitmap(op, &part);
-        }
         if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
         {
             // Check parent inode
@@ -1171,7 +1160,7 @@ resume_2:
         erase_op(op);
         return 1;
     }
-    else if (op->retval != 0 && op->opcode != OSD_OP_SYNC && !(op->flags & OP_FLUSH_BUFFER) &&
+    else if (op->retval != 0 && !(op->flags & OP_FLUSH_BUFFER) &&
         op->retval != -EPIPE && (op->retval != -EIO || !client_eio_retry_interval) && (op->retval != -ENOSPC || !client_retry_enospc))
     {
         // Fatal error (neither -EPIPE, -EIO nor -ENOSPC)
@@ -1264,9 +1253,9 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
         {
             op->bitmap_buf = realloc_or_die(op->bitmap_buf, bitmap_mem);
             op->part_bitmaps = (uint8_t*)op->bitmap_buf + object_bitmap_size;
-            memset((uint8_t*)op->bitmap_buf+op->bitmap_buf_size, 0, bitmap_mem-op->bitmap_buf_size);
             op->bitmap_buf_size = bitmap_mem;
         }
+        memset(op->bitmap_buf, 0, bitmap_mem);
     }
     int iov_idx = 0;
     size_t iov_pos = 0;
@@ -1408,11 +1397,11 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
                 primary_osd = nearest_osd;
         }
         part->osd_num = primary_osd;
-        auto peer_it = msgr.osd_peers.find(primary_osd);
-        if (peer_it != msgr.osd_peers.end())
+        auto peer_it = msgr.osd_peer_fds.find(primary_osd);
+        if (peer_it != msgr.osd_peer_fds.end())
         {
-            osd_client_t *cl = peer_it->second;
-            part->flags |= PART_SENT|PART_VALID;
+            int peer_fd = peer_it->second;
+            part->flags |= PART_SENT;
             op->inflight_count++;
             uint64_t pg_bitmap_size = (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8) * (
                 pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks
@@ -1426,7 +1415,7 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
             }
             part->op = (osd_op_t){
                 .op_type = OSD_OP_OUT,
-                .client_id = cl->client_id,
+                .peer_fd = peer_fd,
                 .req = { .rw = {
                     .header = {
                         .magic = SECONDARY_OSD_OP_MAGIC,
@@ -1475,8 +1464,8 @@ int cluster_client_t::continue_sync(cluster_op_t *op)
     for (auto do_it = dirty_osds.begin(); do_it != dirty_osds.end(); )
     {
         osd_num_t sync_osd = *do_it;
-        auto peer_it = msgr.osd_peers.find(sync_osd);
-        if (peer_it == msgr.osd_peers.end())
+        auto peer_it = msgr.osd_peer_fds.find(sync_osd);
+        if (peer_it == msgr.osd_peer_fds.end())
             dirty_osds.erase(do_it++);
         else
             do_it++;
@@ -1529,12 +1518,12 @@ resume_1:
 
 void cluster_client_t::send_sync(cluster_op_t *op, cluster_op_part_t *part)
 {
-    osd_client_t *cl = msgr.osd_peers.at(part->osd_num);
+    auto peer_fd = msgr.osd_peer_fds.at(part->osd_num);
     part->flags |= PART_SENT;
     op->inflight_count++;
     part->op = (osd_op_t){
         .op_type = OSD_OP_OUT,
-        .client_id = cl->client_id,
+        .peer_fd = peer_fd,
         .req = {
             .hdr = {
                 .magic = SECONDARY_OSD_OP_MAGIC,
@@ -1574,10 +1563,10 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             // Error priority: EIO > ENOSPC > ETIMEDOUT > EPIPE
             op->retval = part->op.reply.hdr.retval;
         }
-        uint64_t stop_client_id = 0;
+        int stop_fd = -1;
         if (op->retval != -EINTR && op->retval != -EIO && op->retval != -ENOSPC)
         {
-            stop_client_id = part->op.client_id;
+            stop_fd = part->op.peer_fd;
             if (op->retval != -EPIPE || log_level > 0)
             {
                 fprintf(
@@ -1604,9 +1593,9 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             op->retry_after = op->retval != -EPIPE ? client_eio_retry_interval : client_retry_interval;
         }
         reset_retry_timer(op->retry_after);
-        if (stop_client_id)
+        if (stop_fd >= 0)
         {
-            msgr.stop_client(stop_client_id);
+            msgr.stop_client(stop_fd);
         }
         op->inflight_count--;
         if (op->inflight_count == 0 && !op->retry_after)
@@ -1625,11 +1614,14 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             dirty_osds.insert(part->osd_num);
         part->flags |= PART_DONE;
         op->done_count++;
-        if ((op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
-            && op->inode == op->cur_inode)
+        if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
         {
-            // Read only returns the version of the uppermost layer
-            op->version = op->parts.size() == 1 ? part->op.reply.rw.version : 0;
+            copy_part_bitmap(op, part);
+            if (op->inode == op->cur_inode)
+            {
+                // Read only returns the version of the uppermost layer
+                op->version = op->parts.size() == 1 ? part->op.reply.rw.version : 0;
+            }
         }
         else if (op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE)
         {

src/client/cluster_client.h

@@ -83,6 +83,9 @@ class writeback_cache_t;
 // FIXME: Split into public and private interfaces
 class __attribute__((visibility("default"))) cluster_client_t
 {
+#ifdef __MOCK__
+public:
+#endif
     timerfd_manager_t *tfd = NULL;
     ring_loop_t *ringloop = NULL;
 
@@ -152,9 +155,15 @@ public:
     void list_inode(inode_t inode, uint64_t min_offset, uint64_t max_offset, int max_parallel_pgs, std::function<void(
         int status, int pgs_left, pg_num_t pg_num, std::set<object_id>&& objects)> pg_callback);
 
+    //inline uint32_t get_bs_bitmap_granularity() { return st_cli.global_bitmap_granularity; }
+    //inline uint64_t get_bs_block_size() { return st_cli.global_block_size; }
+
+#ifndef __MOCK__
 protected:
+#endif
     void continue_ops(int time_passed = 0);
 
+protected:
     bool affects_osd(uint64_t inode, uint64_t offset, uint64_t len, osd_num_t osd);
     bool affects_pg(uint64_t inode, uint64_t offset, uint64_t len, pool_id_t pool_id, pg_num_t pg_num);
 
@@ -195,5 +204,4 @@ protected:
     osd_num_t select_nearest_osd(const std::vector<osd_num_t> & osds);
 
     friend class writeback_cache_t;
-    friend class cluster_client_test_t;
 };

src/client/cluster_client_impl.h

@@ -10,7 +10,6 @@
 #define PART_DONE 2
 #define PART_ERROR 4
 #define PART_RETRY 8
-#define PART_VALID 16
 #define CACHE_DIRTY 1
 #define CACHE_WRITTEN 2
 #define CACHE_FLUSHING 3

src/client/cluster_client_list.cpp

@@ -295,7 +295,7 @@ int cluster_client_t::start_pg_listing(inode_list_pg_t *pg)
     bool conn = true;
     for (osd_num_t peer_osd: all_peers)
     {
-        if (msgr.osd_peers.find(peer_osd) == msgr.osd_peers.end())
+        if (msgr.osd_peer_fds.find(peer_osd) == msgr.osd_peer_fds.end())
         {
             // Initiate connection
             if (st_cli.peer_states[peer_osd].is_null())
@@ -340,7 +340,7 @@ void cluster_client_t::send_list(inode_list_osd_t *cur_list)
     osd_op_t *op = new osd_op_t();
     op->op_type = OSD_OP_OUT;
     // Already checked that it exists above, but anyway
-    op->client_id = msgr.osd_peers.at(cur_list->osd_num)->client_id;
+    op->peer_fd = msgr.osd_peer_fds.at(cur_list->osd_num);
     op->req = (osd_any_op_t){
         .sec_list = {
             .header = {

src/client/cluster_client_wb.cpp

@@ -88,11 +88,6 @@ void writeback_cache_t::copy_write(cluster_op_t *op, int state, uint64_t new_flu
     // ...or just save it for writeback if write buffering is enabled
     if (op->len == 0)
     {
-        // FIXME: OSD_OP_DELETEs are currently only sent by vitastor-cli rm/rm-data and
-        // actually have len=0, because delete is actually a delete of the full object
-        // containing the requested offset, not a "punch hole" operation. But here, writeback
-        // cache assumes it IS a "punch hole" operation. I should select one of these
-        // approaches and fix everything accordingly when I decide to implement TRIM.
         return;
     }
     auto dirty_it = find_dirty(op->inode, op->offset);
@@ -249,13 +244,12 @@ void writeback_cache_t::copy_write(cluster_op_t *op, int state, uint64_t new_flu
             writeback_queue_size--;
         }
     }
-    if (!is_del && op->len > 0)
+    if (!is_del)
     {
         uint64_t pos = 0, len = op->len, iov_idx = 0;
-        while (iov_idx < op->iov.count)
+        while (len > 0 && iov_idx < op->iov.count)
         {
             auto & iov = op->iov.buf[iov_idx];
-            assert(pos + iov.iov_len <= len);
             memcpy(buf + pos, iov.iov_base, iov.iov_len);
             pos += iov.iov_len;
             iov_idx++;
@@ -449,7 +443,7 @@ void writeback_cache_t::start_writebacks(cluster_client_t *cli, int count)
         started++;
         assert(writeback_queue_size > 0);
         writeback_queue_size--;
-        writeback_bytes -= (is_del ? 0 : off - from_it->first.stripe);
+        writeback_bytes -= off - from_it->first.stripe;
         assert(writeback_queue_size > 0 || !writeback_bytes);
         flush_buffers(cli, from_it, to_it);
     }

src/client/etcd_state_client.cpp

@@ -414,10 +414,7 @@ void etcd_state_client_t::start_etcd_watcher()
                 }
                 // Save revision only if it's present in the message - because sometimes etcd sends something without a header, like:
                 // {"error": {"grpc_code": 14, "http_code": 503, "http_status": "Service Unavailable", "message": "error reading from server: EOF"}}
-                // Also don't save revision from the initial created: true messages because they always contain the latest revision
-                if (etcd_watches_initialised == ETCD_TOTAL_WATCHES &&
-                    !data["result"]["header"]["revision"].is_null() &&
-                    !data["result"]["created"].bool_value())
+                if (etcd_watches_initialised == ETCD_TOTAL_WATCHES && !data["result"]["header"]["revision"].is_null())
                 {
                     // Restart watchers from the same revision number as in the last received message,
                     // not from the next one to protect against revision being split into multiple messages,
@@ -571,19 +568,9 @@ void etcd_state_client_t::start_ws_keepalive()
 
 void etcd_state_client_t::load_global_config()
 {
-    json11::Json::object req = { { "success", json11::Json::array {
-        json11::Json::object {
-            { "request_range", json11::Json::object {
-                { "key", base64_encode(etcd_prefix+"/config/global") },
-            } }
-        },
-        json11::Json::object {
-            { "request_range", json11::Json::object {
-                { "key", base64_encode(etcd_prefix+"/config/pools") },
-            } }
-        },
-    } } };
-    etcd_txn(req, etcd_quick_timeout, max_etcd_attempts, 0, [this](std::string err, json11::Json data)
+    etcd_call("/kv/range", json11::Json::object {
+        { "key", base64_encode(etcd_prefix+"/config/global") }
+    }, etcd_quick_timeout, max_etcd_attempts, 0, [this](std::string err, json11::Json data)
     {
         if (err != "")
         {
@@ -601,12 +588,10 @@ void etcd_state_client_t::load_global_config()
             }
             return;
         }
-        json11::Json config_kv = data["responses"][0]["response_range"]["kvs"][0];
-        json11::Json pools_kv = data["responses"][1]["response_range"]["kvs"][0];
         json11::Json::object global_config;
-        if (!config_kv.is_null())
+        if (data["kvs"].array_items().size() > 0)
         {
-            auto kv = parse_etcd_kv(config_kv);
+            auto kv = parse_etcd_kv(data["kvs"][0]);
             if (kv.value.is_object())
             {
                 global_config = kv.value.object_items();
@@ -623,11 +608,6 @@ void etcd_state_client_t::load_global_config()
             global_bitmap_granularity = DEFAULT_BITMAP_GRANULARITY;
         }
         global_immediate_commit = parse_immediate_commit(global_config["immediate_commit"].string_value(), IMMEDIATE_ALL);
-        if (!pools_kv.is_null())
-        {
-            auto kv = parse_etcd_kv(pools_kv);
-            parse_state(kv);
-        }
         on_load_config_hook(global_config);
     });
 }
@@ -965,12 +945,8 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
             if (pc.pg_stripe_size < min_stripe_size)
                 pc.pg_stripe_size = min_stripe_size;
             // Save
-            auto & old_pc = this->pool_config[pool_id];
-            pc.real_pg_count = old_pc.real_pg_count;
-            pc.applied_pg_count = old_pc.applied_pg_count;
-            pc.applied_pg_stripe_size = old_pc.applied_pg_stripe_size;
-            pc.reshard_state = old_pc.reshard_state;
-            std::swap(pc.pg_config, old_pc.pg_config);
+            pc.real_pg_count = this->pool_config[pool_id].real_pg_count;
+            std::swap(pc.pg_config, this->pool_config[pool_id].pg_config);
             std::swap(this->pool_config[pool_id], pc);
             auto & parsed_cfg = this->pool_config[pool_id];
             parsed_cfg.exists = true;
@@ -1185,6 +1161,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
                 if (i >= pg_state_bit_count)
                 {
                     fprintf(stderr, "Unexpected pool %u PG %u state keyword in etcd: %s\n", pool_id, pg_num, e.dump().c_str());
+                    return;
                 }
             }
             if (!cur_primary || !value["state"].is_array() || !state ||
@@ -1193,6 +1170,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
                 (state & PG_INCOMPLETE) && state != PG_INCOMPLETE && state != (PG_INCOMPLETE|PG_HAS_INVALID))
             {
                 fprintf(stderr, "Unexpected pool %u PG %u state in etcd: primary=%ju, state=%s\n", pool_id, pg_num, cur_primary, value["state"].dump().c_str());
+                return;
             }
             pg_cfg.cur_primary = cur_primary;
             pg_cfg.cur_state = state;

src/client/etcd_state_client.h

@@ -6,7 +6,7 @@
 #include <set>
 
 #include "json11/json11.hpp"
-#include "object_id.h"
+#include "osd_id.h"
 #include "timerfd_manager.h"
 
 #define ETCD_CONFIG_WATCH_ID 1
@@ -68,11 +68,6 @@ struct pool_config_t
     std::string used_for_app;
     int backfillfull = 0;
     int local_reads = 0;
-
-    // runtime data, used only by OSD:
-    uint64_t applied_pg_count = 0;
-    uint64_t applied_pg_stripe_size = 0;
-    void *reshard_state = NULL;
 };
 
 struct inode_config_t

src/client/messenger.cpp

@@ -15,6 +15,106 @@
 #include "msgr_rdma.h"
 #endif
 
+#include <sys/poll.h>
+
+msgr_iothread_t::msgr_iothread_t():
+    ring(RINGLOOP_DEFAULT_SIZE, true),
+    thread(&msgr_iothread_t::run, this)
+{
+    eventfd = ring.register_eventfd();
+    if (eventfd < 0)
+    {
+        throw std::runtime_error(std::string("failed to register eventfd: ") + strerror(-eventfd));
+    }
+}
+
+msgr_iothread_t::~msgr_iothread_t()
+{
+    stop();
+}
+
+void msgr_iothread_t::add_sqe(io_uring_sqe & sqe)
+{
+    mu.lock();
+    queue.push_back((iothread_sqe_t){ .sqe = sqe, .data = std::move(*(ring_data_t*)sqe.user_data) });
+    if (queue.size() == 1)
+    {
+        cond.notify_all();
+    }
+    mu.unlock();
+}
+
+void msgr_iothread_t::stop()
+{
+    mu.lock();
+    if (stopped)
+    {
+        mu.unlock();
+        return;
+    }
+    stopped = true;
+    if (outer_loop_data)
+    {
+        outer_loop_data->callback = [](ring_data_t*){};
+    }
+    cond.notify_all();
+    close(eventfd);
+    mu.unlock();
+    thread.join();
+}
+
+void msgr_iothread_t::add_to_ringloop(ring_loop_t *outer_loop)
+{
+    assert(!this->outer_loop || this->outer_loop == outer_loop);
+    io_uring_sqe *sqe = outer_loop->get_sqe();
+    assert(sqe != NULL);
+    this->outer_loop = outer_loop;
+    this->outer_loop_data = ((ring_data_t*)sqe->user_data);
+    io_uring_prep_poll_add(sqe, eventfd, POLLIN);
+    outer_loop_data->callback = [this](ring_data_t *data)
+    {
+        if (data->res < 0)
+        {
+            throw std::runtime_error(std::string("eventfd poll failed: ") + strerror(-data->res));
+        }
+        outer_loop_data = NULL;
+        if (stopped)
+        {
+            return;
+        }
+        add_to_ringloop(this->outer_loop);
+        ring.loop();
+    };
+}
+
+void msgr_iothread_t::run()
+{
+    while (true)
+    {
+        {
+            std::unique_lock<std::mutex> lk(mu);
+            while (!stopped && !queue.size())
+                cond.wait(lk);
+            if (stopped)
+                return;
+            int i = 0;
+            for (; i < queue.size(); i++)
+            {
+                io_uring_sqe *sqe = ring.get_sqe();
+                if (!sqe)
+                    break;
+                ring_data_t *data = ((ring_data_t*)sqe->user_data);
+                *data = std::move(queue[i].data);
+                *sqe = queue[i].sqe;
+                sqe->user_data = (uint64_t)data;
+            }
+            queue.erase(queue.begin(), queue.begin()+i);
+        }
+        // We only want to offload sendmsg/recvmsg. Callbacks will be called in main thread
+        ring.submit();
+    }
+}
+
 void osd_messenger_t::init()
 {
 #ifdef WITH_RDMACM
@@ -45,7 +145,7 @@ void osd_messenger_t::init()
         rdma_contexts = msgr_rdma_context_t::create_all(
             osd_num && osd_cluster_network_masks.size() ? osd_cluster_network_masks : osd_network_masks,
             rdma_device != "" ? rdma_device.c_str() : NULL,
-            rdma_port_num, rdma_gid_index, rdma_mtu, log_level
+            rdma_port_num, rdma_gid_index, rdma_mtu, rdma_odp, log_level
         );
         if (!rdma_contexts.size())
         {
@@ -73,17 +173,21 @@ void osd_messenger_t::init()
     }
     if (ringloop && iothread_count > 0)
     {
-        init_iothreads();
+        for (int i = 0; i < iothread_count; i++)
+        {
+            auto iot = new msgr_iothread_t();
+            iothreads.push_back(iot);
+            iot->add_to_ringloop(ringloop);
+        }
     }
     keepalive_timer_id = tfd->set_timer(1000, true, [this](int)
     {
-        std::vector<uint64_t> clients_to_stop;
-        std::vector<osd_op_t*> ops_to_send;
         auto cl_it = clients.begin();
         while (cl_it != clients.end())
         {
             auto cl = cl_it->second;
             cl_it++;
+            auto peer_fd = cl->peer_fd;
             if (!cl->osd_num && !cl->in_osd_num || cl->peer_state != PEER_CONNECTED && cl->peer_state != PEER_RDMA)
             {
                 // Do not run keepalive on regular clients
@@ -95,9 +199,10 @@ void osd_messenger_t::init()
                 if (!cl->ping_time_remaining)
                 {
                     // Ping timed out, stop the client
-                    fprintf(stderr, "Ping timed out for OSD %ju (client %ju), disconnecting peer\n",
-                        cl->in_osd_num ? cl->in_osd_num : cl->osd_num, cl->client_id);
-                    clients_to_stop.push_back(cl->client_id);
+                    fprintf(stderr, "Ping timed out for OSD %ju (client %d), disconnecting peer\n", cl->in_osd_num ? cl->in_osd_num : cl->osd_num, cl->peer_fd);
+                    stop_client(peer_fd, true);
+                    // Restart iterator because it may be invalidated
+                    cl_it = clients.upper_bound(peer_fd);
                 }
             }
             else if (cl->idle_time_remaining > 0)
@@ -108,36 +213,37 @@ void osd_messenger_t::init()
                     // Connection is idle for <osd_idle_time>, send ping
                     osd_op_t *op = new osd_op_t();
                     op->op_type = OSD_OP_OUT;
-                    op->client_id = cl->client_id;
+                    op->peer_fd = cl->peer_fd;
                     op->req = (osd_any_op_t){
                         .hdr = {
                             .magic = SECONDARY_OSD_OP_MAGIC,
                             .opcode = OSD_OP_PING,
                         },
                     };
-                    op->callback = [this](osd_op_t *op)
+                    op->callback = [this, cl](osd_op_t *op)
                     {
-                        auto cl_it = clients.find(op->client_id);
-                        if (cl_it == clients.end())
+                        auto cl_it = clients.find(op->peer_fd);
+                        if (cl_it == clients.end() || cl_it->second != cl)
                         {
                             // client is already dropped
                             delete op;
                             return;
                         }
-                        auto cl = cl_it->second;
-                        uint64_t fail_client_id = (op->reply.hdr.retval != 0 ? op->client_id : 0);
+                        int fail_fd = (op->reply.hdr.retval != 0 ? op->peer_fd : -1);
                         auto fail_osd_num = cl->in_osd_num ? cl->in_osd_num : cl->osd_num;
                         cl->ping_time_remaining = 0;
                         delete op;
-                        if (fail_client_id)
+                        if (fail_fd >= 0)
                         {
-                            fprintf(stderr, "Ping failed for OSD %ju (client %ju), disconnecting peer\n", fail_osd_num, fail_client_id);
-                            stop_client(fail_client_id);
+                            fprintf(stderr, "Ping failed for OSD %ju (client %d), disconnecting peer\n", fail_osd_num, fail_fd);
+                            stop_client(fail_fd, true);
                         }
                     };
                     cl->ping_time_remaining = osd_ping_timeout;
                     cl->idle_time_remaining = osd_idle_timeout;
-                    ops_to_send.push_back(op);
+                    outbox_push(op);
+                    // Restart iterator because it may be invalidated
+                    cl_it = clients.upper_bound(peer_fd);
                 }
             }
             else
@@ -145,14 +251,6 @@ void osd_messenger_t::init()
                 cl->idle_time_remaining = osd_idle_timeout;
             }
         }
-        for (uint64_t client_id: clients_to_stop)
-        {
-            stop_client(client_id);
-        }
-        for (osd_op_t *op: ops_to_send)
-        {
-            outbox_push(op);
-        }
     });
 }
 
@@ -165,9 +263,16 @@ osd_messenger_t::~osd_messenger_t()
     }
     while (clients.size() > 0)
     {
-        stop_client(clients.begin()->first, true);
+        stop_client(clients.begin()->first, true, true);
+    }
+    if (iothreads.size())
+    {
+        for (auto iot: iothreads)
+        {
+            delete iot;
+        }
+        iothreads.clear();
     }
-    destroy_iothreads();
 #ifdef WITH_RDMA
     for (auto rdma_context: rdma_contexts)
     {
@@ -217,6 +322,7 @@ void osd_messenger_t::parse_config(const json11::Json & config)
     this->rdma_max_msg = config["rdma_max_msg"].uint64_value();
     if (!this->rdma_max_msg || this->rdma_max_msg > 128*1024*1024)
         this->rdma_max_msg = 129*1024;
+    this->rdma_odp = config["rdma_odp"].bool_value();
 #endif
     if (!osd_num)
         this->iothread_count = (uint32_t)config["client_iothread_count"].uint64_value();
@@ -335,7 +441,7 @@ void osd_messenger_t::try_connect_peer(uint64_t peer_osd)
     {
         return;
     }
-    if (osd_peers.find(peer_osd) != osd_peers.end())
+    if (osd_peer_fds.find(peer_osd) != osd_peer_fds.end())
     {
         wanted_peers.erase(peer_osd);
         return;
@@ -362,20 +468,20 @@ void osd_messenger_t::try_connect_peer_tcp(osd_num_t peer_osd, const char *peer_
 #ifdef WITH_RDMACM
     if (disable_tcp)
     {
-        on_connect_peer(peer_osd, -EINVAL, 0);
+        on_connect_peer(peer_osd, -EINVAL);
         return;
     }
 #endif
     struct sockaddr_storage addr;
     if (!string_to_addr(peer_host, 0, peer_port, &addr))
     {
-        on_connect_peer(peer_osd, -EINVAL, 0);
+        on_connect_peer(peer_osd, -EINVAL);
         return;
     }
     int peer_fd = socket(addr.ss_family, SOCK_STREAM, 0);
     if (peer_fd < 0)
     {
-        on_connect_peer(peer_osd, -errno, 0);
+        on_connect_peer(peer_osd, -errno);
         return;
     }
     fcntl(peer_fd, F_SETFL, fcntl(peer_fd, F_GETFL, 0) | O_NONBLOCK);
@@ -383,25 +489,21 @@ void osd_messenger_t::try_connect_peer_tcp(osd_num_t peer_osd, const char *peer_
     if (r < 0 && errno != EINPROGRESS)
     {
         close(peer_fd);
-        on_connect_peer(peer_osd, -errno, 0);
+        on_connect_peer(peer_osd, -errno);
         return;
     }
-    const uint64_t client_id = next_client_id++;
-    osd_client_t *cl = new osd_client_t();
+    clients[peer_fd] = new osd_client_t();
     if (log_level > 0)
     {
-        fprintf(stderr, "Connecting to OSD %ju at %s:%d (client %ju, FD %d)\n", peer_osd, peer_host, peer_port, client_id, peer_fd);
+        fprintf(stderr, "Connecting to OSD %ju at %s:%d (client %d)\n", peer_osd, peer_host, peer_port, peer_fd);
     }
-    cl->client_id = client_id;
-    cl->peer_addr = addr;
-    cl->peer_port = peer_port;
-    cl->peer_fd = peer_fd;
-    cl->peer_state = PEER_CONNECTING;
-    cl->connect_timeout_id = -1;
-    cl->osd_num = peer_osd;
-    cl->in_buf = malloc_or_die(receive_buffer_size);
-    clients[client_id] = cl;
-    clients_by_fd[peer_fd] = cl;
+    clients[peer_fd]->peer_addr = addr;
+    clients[peer_fd]->peer_port = peer_port;
+    clients[peer_fd]->peer_fd = peer_fd;
+    clients[peer_fd]->peer_state = PEER_CONNECTING;
+    clients[peer_fd]->connect_timeout_id = -1;
+    clients[peer_fd]->osd_num = peer_osd;
+    clients[peer_fd]->in_buf = malloc_or_die(receive_buffer_size);
     tfd->set_fd_handler(peer_fd, true, [this](int peer_fd, int epoll_events)
     {
         // Either OUT (connected) or HUP
@@ -409,11 +511,11 @@ void osd_messenger_t::try_connect_peer_tcp(osd_num_t peer_osd, const char *peer_
     });
     if (peer_connect_timeout > 0)
     {
-        cl->connect_timeout_id = tfd->set_timer(1000*peer_connect_timeout, false, [this, client_id](int timer_id)
+        clients[peer_fd]->connect_timeout_id = tfd->set_timer(1000*peer_connect_timeout, false, [this, peer_fd](int timer_id)
         {
-            osd_num_t peer_osd = clients.at(client_id)->osd_num;
-            stop_client(client_id);
-            on_connect_peer(peer_osd, -EPIPE, 0);
+            osd_num_t peer_osd = clients.at(peer_fd)->osd_num;
+            stop_client(peer_fd, true);
+            on_connect_peer(peer_osd, -EPIPE);
             return;
         });
     }
@@ -421,7 +523,7 @@ void osd_messenger_t::try_connect_peer_tcp(osd_num_t peer_osd, const char *peer_
 
 void osd_messenger_t::handle_connect_epoll(int peer_fd)
 {
-    auto cl = clients_by_fd.at(peer_fd);
+    auto cl = clients[peer_fd];
     if (cl->connect_timeout_id >= 0)
     {
         tfd->clear_timer(cl->connect_timeout_id);
@@ -436,8 +538,8 @@ void osd_messenger_t::handle_connect_epoll(int peer_fd)
     }
     if (result != 0)
     {
-        stop_client(cl->client_id);
-        on_connect_peer(peer_osd, -result, 0);
+        stop_client(peer_fd, true);
+        on_connect_peer(peer_osd, -result);
         return;
     }
     int one = 1;
@@ -454,23 +556,23 @@ void osd_messenger_t::handle_connect_epoll(int peer_fd)
 void osd_messenger_t::handle_peer_epoll(int peer_fd, int epoll_events)
 {
     // Mark client as ready (i.e. some data is available)
-    auto cl = clients_by_fd.at(peer_fd);
     if (epoll_events & EPOLLRDHUP)
     {
         // Stop client
         if (log_level > 0)
         {
-            fprintf(stderr, "[OSD %ju] client %ju disconnected\n", this->osd_num, cl->client_id);
+            fprintf(stderr, "[OSD %ju] client %d disconnected\n", this->osd_num, peer_fd);
         }
-        stop_client(cl->client_id);
+        stop_client(peer_fd, true);
     }
     else if (epoll_events & EPOLLIN)
     {
         // Mark client as ready (i.e. some data is available)
+        auto cl = clients[peer_fd];
         cl->read_ready++;
         if (cl->read_ready == 1)
         {
-            read_ready_clients.push_back(cl->client_id);
+            read_ready_clients.push_back(cl->peer_fd);
             if (ringloop)
                 ringloop->wakeup();
             else
@@ -479,13 +581,13 @@ void osd_messenger_t::handle_peer_epoll(int peer_fd, int epoll_events)
     }
 }
 
-void osd_messenger_t::on_connect_peer(osd_num_t peer_osd, int errcode, uint64_t client_id)
+void osd_messenger_t::on_connect_peer(osd_num_t peer_osd, int peer_fd)
 {
     auto & wp = wanted_peers.at(peer_osd);
     wp.connecting = false;
-    if (errcode < 0)
+    if (peer_fd < 0)
     {
-        fprintf(stderr, "Failed to connect to peer OSD %ju address %s port %d: %s\n", peer_osd, wp.cur_addr.c_str(), wp.cur_port, strerror(-errcode));
+        fprintf(stderr, "Failed to connect to peer OSD %ju address %s port %d: %s\n", peer_osd, wp.cur_addr.c_str(), wp.cur_port, strerror(-peer_fd));
         if (wp.address_changed)
         {
             wp.address_changed = false;
@@ -512,7 +614,7 @@ void osd_messenger_t::on_connect_peer(osd_num_t peer_osd, int errcode, uint64_t
     }
     if (log_level > 0)
     {
-        fprintf(stderr, "[OSD %ju] Connected with peer OSD %ju (client %ju)\n", osd_num, peer_osd, client_id);
+        fprintf(stderr, "[OSD %ju] Connected with peer OSD %ju (client %d)\n", osd_num, peer_osd, peer_fd);
     }
     wanted_peers.erase(peer_osd);
     repeer_pgs(peer_osd);
@@ -522,7 +624,7 @@ void osd_messenger_t::check_peer_config(osd_client_t *cl)
 {
     osd_op_t *op = new osd_op_t();
     op->op_type = OSD_OP_OUT;
-    op->client_id = cl->client_id;
+    op->peer_fd = cl->peer_fd;
     op->req = (osd_any_op_t){
         .show_conf = {
             .header = {
@@ -546,7 +648,7 @@ void osd_messenger_t::check_peer_config(osd_client_t *cl)
         if (!selected_ctx)
         {
             if (log_level > 0)
-                fprintf(stderr, "No RDMA context for OSD %ju connection (client %ju), using only TCP\n", cl->osd_num, cl->client_id);
+                fprintf(stderr, "No RDMA context for OSD %ju connection (peer %d), using only TCP\n", cl->osd_num, cl->peer_fd);
         }
         else
         {
@@ -607,8 +709,8 @@ void osd_messenger_t::check_peer_config(osd_client_t *cl)
         if (err)
         {
             osd_num_t peer_osd = cl->osd_num;
-            stop_client(op->client_id);
-            on_connect_peer(peer_osd, -EINVAL, 0);
+            stop_client(op->peer_fd);
+            on_connect_peer(peer_osd, -EINVAL);
             delete op;
             return;
         }
@@ -638,13 +740,21 @@ void osd_messenger_t::check_peer_config(osd_client_t *cl)
                     fprintf(stderr, "Connected to OSD %ju using RDMA\n", cl->osd_num);
                 }
                 cl->peer_state = PEER_RDMA;
+                tfd->set_fd_handler(cl->peer_fd, false, [this](int peer_fd, int epoll_events)
+                {
+                    // Do not miss the disconnection!
+                    if (epoll_events & EPOLLRDHUP)
+                    {
+                        handle_peer_epoll(peer_fd, epoll_events);
+                    }
+                });
                 // Add the initial receive request
                 init_recv_rdma(cl);
             }
         }
 #endif
-        osd_peers[cl->osd_num] = cl;
-        on_connect_peer(cl->osd_num, 0, cl->client_id);
+        osd_peer_fds[cl->osd_num] = cl->peer_fd;
+        on_connect_peer(cl->osd_num, cl->peer_fd);
         delete op;
     };
     outbox_push(op);
@@ -659,16 +769,13 @@ void osd_messenger_t::accept_connections(int listen_fd)
     while ((peer_fd = accept(listen_fd, (sockaddr*)&addr, &peer_addr_size)) >= 0)
     {
         assert(peer_fd != 0);
-        const uint64_t client_id = next_client_id++;
-        fprintf(stderr, "[OSD %ju] new client %ju (FD %d): connection from %s\n", this->osd_num, client_id, peer_fd,
+        fprintf(stderr, "[OSD %ju] new client %d: connection from %s\n", this->osd_num, peer_fd,
             addr_to_string(addr).c_str());
         fcntl(peer_fd, F_SETFL, fcntl(peer_fd, F_GETFL, 0) | O_NONBLOCK);
         int one = 1;
         setsockopt(peer_fd, SOL_TCP, TCP_NODELAY, &one, sizeof(one));
         auto cl = new osd_client_t();
-        cl->client_id = client_id;
-        clients[cl->client_id] = cl;
-        clients_by_fd[peer_fd] = cl;
+        clients[peer_fd] = cl;
         cl->is_incoming = true;
         cl->peer_addr = addr;
         cl->peer_addr = addr;

src/client/messenger.h

@@ -12,7 +12,6 @@
 #include <deque>
 #include <vector>
 
-#include "../util/robin_hood.h"
 #include "malloc_or_die.h"
 #include "json11/json11.hpp"
 #include "msgr_op.h"
@@ -51,7 +50,6 @@ struct msgr_rdma_context_t;
 
 struct osd_client_t
 {
-    uint64_t client_id = 0;
     int refs = 0;
 
     sockaddr_storage peer_addr = {};
@@ -87,7 +85,7 @@ struct osd_client_t
     std::vector<osd_op_t*> received_ops;
 
     // Outbound operations
-    robin_hood::unordered_flat_map<uint64_t, osd_op_t*> sent_ops;
+    std::map<uint64_t, osd_op_t*> sent_ops;
     uint64_t send_op_id = 0;
 
     // PGs dirtied by this client's primary-writes
@@ -129,7 +127,43 @@ struct osd_op_stats_t
     uint64_t subop_stat_count[OSD_OP_MAX+1] = { 0 };
 };
 
+#include <mutex>
+#include <condition_variable>
+#include <thread>
+
+#ifdef __MOCK__
 class msgr_iothread_t;
+#else
+struct iothread_sqe_t
+{
+    io_uring_sqe sqe;
+    ring_data_t data;
+};
+
+class msgr_iothread_t
+{
+protected:
+    ring_loop_t ring;
+    ring_loop_t *outer_loop = NULL;
+    ring_data_t *outer_loop_data = NULL;
+    int eventfd = -1;
+    bool stopped = false;
+    std::mutex mu;
+    std::condition_variable cond;
+    std::vector<iothread_sqe_t> queue;
+    std::thread thread;
+
+    void run();
+public:
+
+    msgr_iothread_t();
+    ~msgr_iothread_t();
+
+    void add_sqe(io_uring_sqe & sqe);
+    void stop();
+    void add_to_ringloop(ring_loop_t *outer_loop);
+};
+#endif
 
 #ifdef WITH_RDMA
 struct rdma_event_channel;
@@ -166,28 +200,27 @@ protected:
     std::vector<msgr_rdma_context_t *> rdma_contexts;
     uint64_t rdma_max_sge = 0, rdma_max_send = 0, rdma_max_recv = 0;
     uint64_t rdma_max_msg = 0;
+    bool rdma_odp = false;
     rdma_event_channel *rdmacm_evch = NULL;
-    robin_hood::unordered_flat_map<rdma_cm_id*, osd_client_t*> rdmacm_connections;
-    robin_hood::unordered_flat_map<rdma_cm_id*, rdmacm_connecting_t*> rdmacm_connecting;
+    std::map<rdma_cm_id*, osd_client_t*> rdmacm_connections;
+    std::map<rdma_cm_id*, rdmacm_connecting_t*> rdmacm_connecting;
 #endif
 
     std::vector<msgr_iothread_t*> iothreads;
-    std::vector<uint64_t> read_ready_clients;
-    std::vector<uint64_t> write_ready_clients;
+    std::vector<int> read_ready_clients;
+    std::vector<int> write_ready_clients;
     // We don't use ringloop->set_immediate here because we may have no ringloop in client :)
-    std::deque<osd_op_t*> set_immediate_ops;
+    std::vector<osd_op_t*> set_immediate_ops;
 
 public:
     timerfd_manager_t *tfd = NULL;
-    ring_loop_i *ringloop = NULL;
+    ring_loop_t *ringloop = NULL;
     bool has_sendmsg_zc = false;
     // osd_num_t is only for logging and asserts
-    uint64_t next_client_id = 1;
     osd_num_t osd_num;
-    robin_hood::unordered_flat_map<uint64_t, osd_client_t*> clients;
-    robin_hood::unordered_flat_map<uint64_t, osd_client_t*> osd_peers;
-    robin_hood::unordered_flat_map<int, osd_client_t*> clients_by_fd;
-    robin_hood::unordered_flat_map<osd_num_t, osd_wanted_peer_t> wanted_peers;
+    std::map<int, osd_client_t*> clients;
+    std::map<osd_num_t, osd_wanted_peer_t> wanted_peers;
+    std::map<uint64_t, int> osd_peer_fds;
     std::vector<std::string> osd_networks;
     std::vector<addr_mask_t> osd_network_masks;
     std::vector<std::string> osd_cluster_networks;
@@ -198,11 +231,9 @@ public:
     osd_op_stats_t stats, recovery_stats;
 
     void init();
-    void init_iothreads();
     void parse_config(const json11::Json & config);
     void connect_peer(uint64_t osd_num, json11::Json peer_state);
-    void stop_client(uint64_t client_id, bool force_delete = false);
-    void destroy_client(osd_client_t *cl);
+    void stop_client(int peer_fd, bool force = false, bool force_delete = false);
     void outbox_push(osd_op_t *cur_op);
     std::function<void(osd_op_t*)> exec_op;
     std::function<void(osd_num_t)> repeer_pgs;
@@ -211,7 +242,6 @@ public:
     void read_requests();
     void send_replies();
     void accept_connections(int listen_fd);
-    void destroy_iothreads();
     ~osd_messenger_t();
 
     static json11::Json::object read_config(const json11::Json & config);
@@ -222,7 +252,7 @@ public:
 
 #ifdef WITH_RDMA
     bool is_rdma_enabled();
-    bool connect_rdma(uint64_t client_id, std::string rdma_address, uint64_t client_max_msg);
+    bool connect_rdma(int peer_fd, std::string rdma_address, uint64_t client_max_msg);
 #endif
 #ifdef WITH_RDMACM
     bool is_use_rdmacm();
@@ -238,7 +268,7 @@ protected:
     void try_connect_peer_tcp(osd_num_t peer_osd, const char *peer_host, int peer_port);
     void handle_peer_epoll(int peer_fd, int epoll_events);
     void handle_connect_epoll(int peer_fd);
-    void on_connect_peer(osd_num_t peer_osd, int errcode, uint64_t client_id);
+    void on_connect_peer(osd_num_t peer_osd, int peer_fd);
     void check_peer_config(osd_client_t *cl);
     void cancel_osd_ops(osd_client_t *cl);
     void cancel_op(osd_op_t *op);
@@ -253,13 +283,15 @@ protected:
     bool handle_reply_hdr(osd_client_t *cl);
     void handle_reply_ready(osd_op_t *op);
     void handle_immediate_ops();
+    void clear_immediate_ops(int peer_fd);
 
 #ifdef WITH_RDMA
     void try_send_rdma(osd_client_t *cl);
+    void try_send_rdma_odp(osd_client_t *cl);
+    void try_send_rdma_nodp(osd_client_t *cl);
     bool init_recv_rdma(osd_client_t *cl);
     void handle_rdma_events(msgr_rdma_context_t *rdma_context);
     msgr_rdma_context_t* choose_rdma_context(osd_client_t *cl);
-    void destroy_rdma_conn(msgr_rdma_connection_t *rdma_conn);
 #endif
 #ifdef WITH_RDMACM
     void handle_rdmacm_events();

src/client/msgr_iothread.cpp

@@ -1,129 +0,0 @@
-// Copyright (c) Vitaliy Filippov, 2019+
-// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
-
-#include <stdexcept>
-#include <sys/poll.h>
-#include <unistd.h>
-
-#include "messenger.h"
-#include "msgr_iothread.h"
-
-msgr_iothread_t::msgr_iothread_t():
-    ring(RINGLOOP_DEFAULT_SIZE, true),
-    thread(&msgr_iothread_t::run, this)
-{
-    eventfd = ring.register_eventfd();
-    if (eventfd < 0)
-    {
-        throw std::runtime_error(std::string("failed to register eventfd: ") + strerror(-eventfd));
-    }
-}
-
-msgr_iothread_t::~msgr_iothread_t()
-{
-    stop();
-}
-
-void msgr_iothread_t::add_sqe(io_uring_sqe & sqe)
-{
-    mu.lock();
-    queue.push_back((iothread_sqe_t){ .sqe = sqe, .data = std::move(*(ring_data_t*)sqe.user_data) });
-    if (queue.size() == 1)
-    {
-        cond.notify_all();
-    }
-    mu.unlock();
-}
-
-void msgr_iothread_t::stop()
-{
-    mu.lock();
-    if (stopped)
-    {
-        mu.unlock();
-        return;
-    }
-    stopped = true;
-    if (outer_loop_data)
-    {
-        outer_loop_data->callback = [](ring_data_t*){};
-    }
-    cond.notify_all();
-    close(eventfd);
-    mu.unlock();
-    thread.join();
-}
-
-void msgr_iothread_t::add_to_ringloop(ring_loop_i *outer_loop)
-{
-    assert(!this->outer_loop || this->outer_loop == outer_loop);
-    io_uring_sqe *sqe = outer_loop->get_sqe();
-    assert(sqe != NULL);
-    this->outer_loop = outer_loop;
-    this->outer_loop_data = ((ring_data_t*)sqe->user_data);
-    io_uring_prep_poll_add(sqe, eventfd, POLLIN);
-    outer_loop_data->callback = [this](ring_data_t *data)
-    {
-        if (data->res < 0)
-        {
-            throw std::runtime_error(std::string("eventfd poll failed: ") + strerror(-data->res));
-        }
-        outer_loop_data = NULL;
-        if (stopped)
-        {
-            return;
-        }
-        add_to_ringloop(this->outer_loop);
-        ring.loop();
-    };
-}
-
-void msgr_iothread_t::run()
-{
-    while (true)
-    {
-        {
-            std::unique_lock<std::mutex> lk(mu);
-            while (!stopped && !queue.size())
-                cond.wait(lk);
-            if (stopped)
-                return;
-            int i = 0;
-            for (; i < queue.size(); i++)
-            {
-                io_uring_sqe *sqe = ring.get_sqe();
-                if (!sqe)
-                    break;
-                ring_data_t *data = ((ring_data_t*)sqe->user_data);
-                *data = std::move(queue[i].data);
-                *sqe = queue[i].sqe;
-                sqe->user_data = (uint64_t)data;
-            }
-            queue.erase(queue.begin(), queue.begin()+i);
-        }
-        // We only want to offload sendmsg/recvmsg. Callbacks will be called in main thread
-        ring.submit();
-    }
-}
-
-void osd_messenger_t::init_iothreads()
-{
-    for (int i = 0; i < iothread_count; i++)
-    {
-        auto iot = new msgr_iothread_t();
-        iothreads.push_back(iot);
-        iot->add_to_ringloop(ringloop);
-    }
-}
-
-void osd_messenger_t::destroy_iothreads()
-{
-    if (iothreads.size())
-    {
-        for (auto iot: iothreads)
-        {
-            delete iot;
-        }
-        iothreads.clear();
-    }
-}

src/client/msgr_iothread.h

@@ -1,38 +0,0 @@
-// Copyright (c) Vitaliy Filippov, 2019+
-// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
-
-#include <mutex>
-#include <condition_variable>
-#include <thread>
-
-#include "ringloop.h"
-
-struct iothread_sqe_t
-{
-    io_uring_sqe sqe;
-    ring_data_t data;
-};
-
-class msgr_iothread_t
-{
-protected:
-    ring_loop_t ring;
-    ring_loop_i *outer_loop = NULL;
-    ring_data_t *outer_loop_data = NULL;
-    int eventfd = -1;
-    bool stopped = false;
-    std::mutex mu;
-    std::condition_variable cond;
-    std::vector<iothread_sqe_t> queue;
-    std::thread thread;
-
-    void run();
-public:
-
-    msgr_iothread_t();
-    ~msgr_iothread_t();
-
-    void add_sqe(io_uring_sqe & sqe);
-    void stop();
-    void add_to_ringloop(ring_loop_i *outer_loop);
-};

src/client/msgr_op.h

@@ -156,8 +156,7 @@ struct __attribute__((visibility("default"))) osd_op_t
 {
     timespec tv_begin = { 0 }, tv_end = { 0 };
     uint64_t op_type = OSD_OP_IN;
-    uint64_t client_id = 0;
-    osd_num_t osd_num = 0;
+    int peer_fd;
     osd_any_op_t req;
     osd_any_reply_t reply;
     blockstore_op_t *bs_op = NULL;

src/client/msgr_rdma.cpp

@@ -59,6 +59,8 @@ msgr_rdma_context_t::~msgr_rdma_context_t()
         ibv_destroy_cq(cq);
     if (channel)
         ibv_destroy_comp_channel(channel);
+    if (mr)
+        ibv_dereg_mr(mr);
     if (pd)
         ibv_dealloc_pd(pd);
     if (context && !is_cm)
@@ -180,15 +182,13 @@ static int match_port_gid(const std::vector<addr_mask_t> & osd_network_masks, ib
 #endif
 
 std::vector<msgr_rdma_context_t*> msgr_rdma_context_t::create_all(const std::vector<addr_mask_t> & osd_network_masks,
-    const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, int log_level)
+    const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, bool odp, int log_level)
 {
     int res;
     std::vector<msgr_rdma_context_t*> ret;
     ibv_device **raw_dev_list = NULL;
     ibv_device **dev_list = NULL;
     ibv_device *single_list[2] = {};
-    int up_ports = 0;
-    int single_port_num = 0;
 
     raw_dev_list = dev_list = ibv_get_device_list(NULL);
     if (!dev_list || !*dev_list)
@@ -223,7 +223,6 @@ std::vector<msgr_rdma_context_t*> msgr_rdma_context_t::create_all(const std::vec
         dev_list = single_list;
     }
 
-retry:
     for (int i = 0; dev_list[i]; ++i)
     {
         auto dev = dev_list[i];
@@ -261,9 +260,6 @@ retry:
                 fprintf(stderr, "RDMA device %s port %d GID %d does not exist\n", ibv_get_device_name(dev), port_num, sel_gid_index);
                 continue;
             }
-            up_ports++;
-            single_port_num = port_num;
-            single_list[0] = dev;
             uint32_t port_mtu = sel_mtu ? sel_mtu : ibv_mtu_to_bytes(portinfo.active_mtu);
 #ifdef IBV_ADVISE_MR_ADVICE_PREFETCH_NO_FAULT
             if (sel_gid_index < 0)
@@ -275,7 +271,7 @@ retry:
                 {
                     if (log_level > 0)
                         log_rdma_dev_port_gid(dev, port_num, best_gid_idx, port_mtu, best_gidx);
-                    auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, log_level);
+                    auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, odp, log_level);
                     if (ctx)
                     {
                         ctx->net_mask = osd_network_masks[net_num];
@@ -295,7 +291,7 @@ retry:
                     log_rdma_dev_port_gid(dev, port_num, best_gid_idx, port_mtu, gidx);
                 }
 #endif
-                auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, log_level);
+                auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, odp, log_level);
                 if (ctx)
                     ret.push_back(ctx);
             }
@@ -304,21 +300,13 @@ cleanup_dev:
         ibv_close_device(context);
     }
 
-    if (!ret.size() && up_ports == 1 && dev_list != single_list)
-    {
-        // Auto-select the only available device/port if there is only one
-        dev_list = single_list;
-        sel_port_num = single_port_num;
-        goto retry;
-    }
-
 cleanup:
     if (raw_dev_list)
         ibv_free_device_list(raw_dev_list);
     return ret;
 }
 
-msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr & portinfo, int ib_port, int gid_index, uint32_t mtu, int log_level)
+msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr & portinfo, int ib_port, int gid_index, uint32_t mtu, bool odp, int log_level)
 {
     msgr_rdma_context_t *ctx = new msgr_rdma_context_t();
     ibv_context *context = ibv_open_device(dev);
@@ -358,6 +346,30 @@ msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr
         goto cleanup;
     }
 
+    ctx->odp = odp;
+    if (ctx->odp)
+    {
+        if (!(ctx->attrx.odp_caps.general_caps & IBV_ODP_SUPPORT) ||
+            !(ctx->attrx.odp_caps.general_caps & IBV_ODP_SUPPORT_IMPLICIT) ||
+            !(ctx->attrx.odp_caps.per_transport_caps.rc_odp_caps & IBV_ODP_SUPPORT_SEND) ||
+            !(ctx->attrx.odp_caps.per_transport_caps.rc_odp_caps & IBV_ODP_SUPPORT_RECV))
+        {
+            ctx->odp = false;
+            if (log_level > 0)
+                fprintf(stderr, "The RDMA device isn't implicit ODP (On-Demand Paging) capable, disabling it\n");
+        }
+    }
+
+    if (ctx->odp)
+    {
+        ctx->mr = ibv_reg_mr(ctx->pd, NULL, SIZE_MAX, IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_ON_DEMAND);
+        if (!ctx->mr)
+        {
+            fprintf(stderr, "Couldn't register RDMA memory region\n");
+            goto cleanup;
+        }
+    }
+
     ctx->channel = ibv_create_comp_channel(ctx->context);
     if (!ctx->channel)
     {
@@ -376,6 +388,8 @@ msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr
     return ctx;
 
 cleanup:
+    if (context)
+        ibv_close_device(context);
     delete ctx;
     return NULL;
 }
@@ -417,6 +431,8 @@ msgr_rdma_connection_t *msgr_rdma_connection_t::create(msgr_rdma_context_t *ctx,
     conn->max_recv = max_recv;
     conn->max_sge = max_sge;
     conn->max_msg = max_msg;
+    conn->in_credit = max_recv;
+    conn->out_credit = max_recv;
 
     ibv_qp_init_attr init_attr = {
         .send_cq = ctx->cq,
@@ -507,7 +523,7 @@ int msgr_rdma_connection_t::connect(msgr_rdma_address_t *dest)
     return 0;
 }
 
-bool osd_messenger_t::connect_rdma(uint64_t client_id, std::string rdma_address, uint64_t client_max_msg)
+bool osd_messenger_t::connect_rdma(int peer_fd, std::string rdma_address, uint64_t client_max_msg)
 {
     // Try to connect to the peer using RDMA
     msgr_rdma_address_t addr;
@@ -517,12 +533,12 @@ bool osd_messenger_t::connect_rdma(uint64_t client_id, std::string rdma_address,
         {
             client_max_msg = rdma_max_msg;
         }
-        auto cl = clients.at(client_id);
+        auto cl = clients.at(peer_fd);
         msgr_rdma_context_t *selected_ctx = choose_rdma_context(cl);
         if (!selected_ctx)
         {
             if (log_level > 0)
-                fprintf(stderr, "No RDMA context for peer %ju, using only TCP\n", client_id);
+                fprintf(stderr, "No RDMA context for peer %d, using only TCP\n", cl->peer_fd);
             return false;
         }
         msgr_rdma_connection_t *rdma_conn = msgr_rdma_connection_t::create(selected_ctx, rdma_max_send, rdma_max_recv, rdma_max_sge, client_max_msg);
@@ -533,13 +549,14 @@ bool osd_messenger_t::connect_rdma(uint64_t client_id, std::string rdma_address,
             {
                 delete rdma_conn;
                 fprintf(
-                    stderr, "Failed to connect RDMA queue pair to %s (client %ju)\n",
-                    addr.to_string().c_str(), client_id
+                    stderr, "Failed to connect RDMA queue pair to %s (client %d)\n",
+                    addr.to_string().c_str(), peer_fd
                 );
             }
             else
             {
                 // Remember connection, but switch to RDMA only after sending the configuration response
+                auto cl = clients.at(peer_fd);
                 cl->rdma_conn = rdma_conn;
                 cl->peer_state = PEER_RDMA_CONNECTING;
                 return true;
@@ -553,12 +570,14 @@ static void try_send_rdma_wr(osd_client_t *cl, ibv_sge *sge, int op_sge)
 {
     ibv_send_wr *bad_wr = NULL;
     ibv_send_wr wr = {
-        .wr_id = cl->client_id,
+        .wr_id = (uint64_t)(cl->peer_fd*2+1),
         .sg_list = sge,
         .num_sge = op_sge,
-        .opcode = IBV_WR_SEND,
+        .opcode = IBV_WR_SEND_WITH_IMM,
         .send_flags = IBV_SEND_SIGNALED,
+        .imm_data = cl->rdma_conn->in_credit,
     };
+    cl->rdma_conn->in_credit = 0;
     int err = ibv_post_send(cl->rdma_conn->qp, &wr, &bad_wr);
     if (err || bad_wr)
     {
@@ -566,6 +585,7 @@ static void try_send_rdma_wr(osd_client_t *cl, ibv_sge *sge, int op_sge)
         exit(1);
     }
     cl->rdma_conn->cur_send++;
+    cl->rdma_conn->out_credit--;
 }
 
 static int try_send_rdma_copy(osd_client_t *cl, uint8_t *dst, int dst_len)
@@ -590,7 +610,52 @@ static int try_send_rdma_copy(osd_client_t *cl, uint8_t *dst, int dst_len)
     return total_dst_len-dst_len;
 }
 
-void osd_messenger_t::try_send_rdma(osd_client_t *cl)
+void osd_messenger_t::try_send_rdma_odp(osd_client_t *cl)
+{
+    auto rc = cl->rdma_conn;
+    if (!cl->send_list.size() || rc->cur_send >= rc->max_send || !rc->out_credit)
+    {
+        return;
+    }
+    uint64_t op_size = 0, op_sge = 0;
+    ibv_sge sge[rc->max_sge];
+    while (rc->send_pos < cl->send_list.size())
+    {
+        iovec & iov = cl->send_list[rc->send_pos];
+        if (op_size >= rc->max_msg || op_sge >= rc->max_sge)
+        {
+            rc->send_sizes.push_back(op_size);
+            try_send_rdma_wr(cl, sge, op_sge);
+            op_sge = 0;
+            op_size = 0;
+            if (rc->cur_send >= rc->max_send || !rc->out_credit)
+            {
+                break;
+            }
+        }
+        uint32_t len = (uint32_t)(op_size+iov.iov_len-rc->send_buf_pos < rc->max_msg
+            ? iov.iov_len-rc->send_buf_pos : rc->max_msg-op_size);
+        sge[op_sge++] = {
+            .addr = (uintptr_t)((uint8_t*)iov.iov_base+rc->send_buf_pos),
+            .length = len,
+            .lkey = rc->ctx->mr->lkey,
+        };
+        op_size += len;
+        rc->send_buf_pos += len;
+        if (rc->send_buf_pos >= iov.iov_len)
+        {
+            rc->send_pos++;
+            rc->send_buf_pos = 0;
+        }
+    }
+    if (op_sge > 0)
+    {
+        rc->send_sizes.push_back(op_size);
+        try_send_rdma_wr(cl, sge, op_sge);
+    }
+}
+
+void osd_messenger_t::try_send_rdma_nodp(osd_client_t *cl)
 {
     auto rc = cl->rdma_conn;
     if (!rc->send_out_size)
@@ -598,23 +663,24 @@ void osd_messenger_t::try_send_rdma(osd_client_t *cl)
         // Allocate send ring buffer, if not yet
         rc->send_out_size = rc->max_msg*rdma_max_send;
         rc->send_out.buf = (uint8_t*)malloc_or_die(rc->send_out_size);
-        rc->send_out.mr = ibv_reg_mr(rc->ctx->pd, rc->send_out.buf, rc->send_out_size, 0);
-        if (!rc->send_out.mr)
+        if (!rc->ctx->odp)
         {
-            fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
-            exit(1);
+            rc->send_out.mr = ibv_reg_mr(rc->ctx->pd, rc->send_out.buf, rc->send_out_size, 0);
+            if (!rc->send_out.mr)
+            {
+                fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
+                exit(1);
+            }
         }
     }
     // Copy data into the buffer and send it
     uint8_t *dst = NULL;
     int dst_len = 0;
     int copied = 1;
-    while (!rc->send_out_full && copied > 0 && rc->cur_send < rc->max_send)
+    while (!rc->send_out_full && copied > 0 && rc->cur_send < rc->max_send && rc->out_credit > 0)
     {
         dst = (uint8_t*)rc->send_out.buf + rc->send_out_pos;
-        dst_len = (rc->send_out_pos >= rc->send_done_pos
-            ? rc->send_out_size-rc->send_out_pos
-            : rc->send_done_pos-rc->send_out_pos);
+        dst_len = (rc->send_out_pos < rc->send_out_size ? rc->send_out_size-rc->send_out_pos : rc->send_done_pos-rc->send_out_pos);
         if (dst_len > rc->max_msg)
             dst_len = rc->max_msg;
         copied = try_send_rdma_copy(cl, dst, dst_len);
@@ -624,17 +690,35 @@ void osd_messenger_t::try_send_rdma(osd_client_t *cl)
             if (rc->send_out_pos == rc->send_out_size)
                 rc->send_out_pos = 0;
             assert(rc->send_out_pos < rc->send_out_size);
-            if (rc->send_out_pos == rc->send_done_pos)
+            if (rc->send_out_pos >= rc->send_done_pos)
                 rc->send_out_full = true;
             ibv_sge sge = {
                 .addr = (uintptr_t)dst,
                 .length = (uint32_t)copied,
-                .lkey = rc->send_out.mr->lkey,
+                .lkey = rc->ctx->odp ? rc->ctx->mr->lkey : rc->send_out.mr->lkey,
             };
             try_send_rdma_wr(cl, &sge, 1);
             rc->send_sizes.push_back(copied);
         }
     }
+    if (rc->cur_send < rc->max_send && rc->in_credit > 0)
+    {
+        ibv_sge sge = {
+            .addr = (uintptr_t)NULL,
+            .length = (uint32_t)0,
+            .lkey = 0,
+        };
+        try_send_rdma_wr(cl, &sge, 1);
+        rc->send_sizes.push_back(0);
+    }
+}
+
+void osd_messenger_t::try_send_rdma(osd_client_t *cl)
+{
+    if (cl->rdma_conn->ctx->odp)
+        try_send_rdma_odp(cl);
+    else
+        try_send_rdma_nodp(cl);
 }
 
 static void try_recv_rdma_wr(osd_client_t *cl, void *buf)
@@ -642,11 +726,11 @@ static void try_recv_rdma_wr(osd_client_t *cl, void *buf)
     ibv_sge sge = {
         .addr = (uintptr_t)buf,
         .length = (uint32_t)cl->rdma_conn->max_msg,
-        .lkey = cl->rdma_conn->recv_buf.mr->lkey,
+        .lkey = cl->rdma_conn->ctx->odp ? cl->rdma_conn->ctx->mr->lkey : cl->rdma_conn->recv_buf.mr->lkey,
     };
     ibv_recv_wr *bad_wr = NULL;
     ibv_recv_wr wr = {
-        .wr_id = cl->client_id,
+        .wr_id = (uint64_t)(cl->peer_fd*2),
         .sg_list = &sge,
         .num_sge = 1,
     };
@@ -657,6 +741,7 @@ static void try_recv_rdma_wr(osd_client_t *cl, void *buf)
         exit(1);
     }
     cl->rdma_conn->cur_recv++;
+    cl->rdma_conn->in_credit++;
 }
 
 bool osd_messenger_t::init_recv_rdma(osd_client_t *cl)
@@ -664,11 +749,14 @@ bool osd_messenger_t::init_recv_rdma(osd_client_t *cl)
     auto rc = cl->rdma_conn;
     assert(!rc->recv_buf.buf);
     rc->recv_buf.buf = (uint8_t*)malloc_or_die(rc->max_msg * rc->max_recv);
-    rc->recv_buf.mr = ibv_reg_mr(rc->ctx->pd, rc->recv_buf.buf, rc->max_msg * rc->max_recv, IBV_ACCESS_LOCAL_WRITE);
-    if (!rc->recv_buf.mr)
+    if (!rc->ctx->odp)
     {
-        fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
-        exit(1);
+        rc->recv_buf.mr = ibv_reg_mr(rc->ctx->pd, rc->recv_buf.buf, rc->max_msg * rc->max_recv, IBV_ACCESS_LOCAL_WRITE);
+        if (!rc->recv_buf.mr)
+        {
+            fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
+            exit(1);
+        }
     }
     for (uint32_t i = 0; i < rc->max_recv; i++)
     {
@@ -703,38 +791,41 @@ void osd_messenger_t::handle_rdma_events(msgr_rdma_context_t *rdma_context)
         event_count = ibv_poll_cq(rdma_context->cq, RDMA_EVENTS_AT_ONCE, wc);
         for (int i = 0; i < event_count; i++)
         {
-            uint64_t client_id = wc[i].wr_id;
-            bool is_send = wc[i].opcode == IBV_WC_SEND;
+            int client_id = wc[i].wr_id >> 1;
+            bool is_send = wc[i].wr_id & 1;
             auto cl_it = clients.find(client_id);
             if (cl_it == clients.end())
             {
                 continue;
             }
             osd_client_t *cl = cl_it->second;
-            if (cl->peer_state == PEER_STOPPED)
-            {
-                continue;
-            }
             auto rc = cl->rdma_conn;
             if (wc[i].status != IBV_WC_SUCCESS)
             {
-                fprintf(stderr, "RDMA work request failed for client %ju", client_id);
+                fprintf(stderr, "RDMA work request failed for client %d", client_id);
                 if (cl->osd_num)
                 {
                     fprintf(stderr, " (OSD %ju)", cl->osd_num);
                 }
                 fprintf(stderr, " with status: %s, stopping client\n", ibv_wc_status_str(wc[i].status));
                 stop_client(client_id);
+                clear_immediate_ops(client_id);
                 continue;
             }
             if (!is_send)
             {
                 // Reset OSD ping state - client is obviously alive
+                if (wc[i].wc_flags & IBV_WC_WITH_IMM)
+                {
+                    rc->out_credit += wc[i].imm_data;
+                }
                 cl->ping_time_remaining = 0;
                 cl->idle_time_remaining = osd_idle_timeout;
                 rc->cur_recv--;
                 if (!handle_read_buffer(cl, rc->recv_buffers[rc->next_recv_buf], wc[i].byte_len))
                 {
+                    // handle_read_buffer may stop the client
+                    clear_immediate_ops(client_id);
                     continue;
                 }
                 try_recv_rdma_wr(cl, rc->recv_buffers[rc->next_recv_buf]);
@@ -745,11 +836,14 @@ void osd_messenger_t::handle_rdma_events(msgr_rdma_context_t *rdma_context)
                 rc->cur_send--;
                 uint64_t sent_size = rc->send_sizes.at(0);
                 rc->send_sizes.erase(rc->send_sizes.begin(), rc->send_sizes.begin()+1);
-                rc->send_done_pos += sent_size;
-                rc->send_out_full = false;
-                if (rc->send_done_pos == rc->send_out_size)
-                    rc->send_done_pos = 0;
-                assert(rc->send_done_pos < rc->send_out_size);
+                if (!rdma_context->odp)
+                {
+                    rc->send_done_pos += sent_size;
+                    rc->send_out_full = false;
+                    if (rc->send_done_pos == rc->send_out_size)
+                        rc->send_done_pos = 0;
+                    assert(rc->send_done_pos < rc->send_out_size);
+                }
                 int send_pos = 0, send_buf_pos = 0;
                 while (sent_size > 0)
                 {
@@ -794,16 +888,3 @@ void osd_messenger_t::handle_rdma_events(msgr_rdma_context_t *rdma_context)
     } while (event_count > 0);
     handle_immediate_ops();
 }
-
-void osd_messenger_t::destroy_rdma_conn(msgr_rdma_connection_t *rdma_conn)
-{
-    if (rdma_conn->cmid)
-    {
-        auto rdma_it = rdmacm_connections.find(rdma_conn->cmid);
-        if (rdma_it != rdmacm_connections.end() && rdma_it->second->rdma_conn == rdma_conn)
-        {
-            rdmacm_connections.erase(rdma_it);
-        }
-    }
-    delete rdma_conn;
-}

src/client/msgr_rdma.h

@@ -26,6 +26,8 @@ struct msgr_rdma_context_t
     ibv_context *context = NULL;
     ibv_device_attr_ex attrx;
     ibv_pd *pd = NULL;
+    bool odp = false;
+    ibv_mr *mr = NULL;
     ibv_comp_channel *channel = NULL;
     ibv_cq *cq = NULL;
     ibv_port_attr portinfo;
@@ -41,9 +43,9 @@ struct msgr_rdma_context_t
     int cm_refs = 0;
 
     static std::vector<msgr_rdma_context_t*> create_all(const std::vector<addr_mask_t> & osd_network_masks,
-        const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, int log_level);
+        const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, bool odp, int log_level);
     static msgr_rdma_context_t *create(ibv_device *dev, ibv_port_attr & portinfo,
-        int ib_port, int gid_index, uint32_t mtu, int log_level);
+        int ib_port, int gid_index, uint32_t mtu, bool odp, int log_level);
     static msgr_rdma_context_t* create_cm(ibv_context *ctx);
     bool reserve_cqe(int n);
 
@@ -79,6 +81,8 @@ struct msgr_rdma_connection_t
     int send_out_pos = 0, send_done_pos = 0, send_out_size = 0;
     bool send_out_full = false;
 
+    uint32_t out_credit = 0, in_credit = 0;
+
     ~msgr_rdma_connection_t();
     static msgr_rdma_connection_t *create(msgr_rdma_context_t *ctx, uint32_t max_send, uint32_t max_recv, uint32_t max_sge, uint32_t max_msg);
     int connect(msgr_rdma_address_t *dest);

src/client/msgr_rdmacm.cpp

@@ -11,7 +11,7 @@
 struct rdmacm_connecting_t
 {
     rdma_cm_id *cmid = NULL;
-    uint64_t client_id = 0;
+    int peer_fd = -1;
     osd_num_t peer_osd = 0;
     std::string addr;
     sockaddr_storage parsed_addr = {};
@@ -117,9 +117,9 @@ void osd_messenger_t::handle_rdmacm_events()
             auto cli_it = rdmacm_connections.find(ev->id);
             if (cli_it != rdmacm_connections.end())
             {
-                fprintf(stderr, "Received %s event for client %ju, closing connection\n",
-                    event_type_name, cli_it->second->client_id);
-                stop_client(cli_it->second->client_id);
+                fprintf(stderr, "Received %s event for peer %d, closing connection\n",
+                    event_type_name, cli_it->second->peer_fd);
+                stop_client(cli_it->second->peer_fd);
             }
             else if (rdmacm_connecting.find(ev->id) != rdmacm_connecting.end())
             {
@@ -178,6 +178,7 @@ msgr_rdma_context_t* msgr_rdma_context_t::create_cm(ibv_context *ctx)
         delete rdma_context;
         return NULL;
     }
+    rdma_context->odp = false;
     rdma_context->channel = ibv_create_comp_channel(rdma_context->context);
     if (!rdma_context->channel)
     {
@@ -265,6 +266,14 @@ msgr_rdma_context_t* osd_messenger_t::rdmacm_create_qp(rdma_cm_id *cmid)
 
 void osd_messenger_t::rdmacm_accept(rdma_cm_event *ev)
 {
+    // Make a fake FD (FIXME: do not use FDs for identifying clients!)
+    int fake_fd = socket(AF_INET, SOCK_STREAM, 0);
+    if (fake_fd < 0)
+    {
+        fprintf(stderr, "Failed to allocate a fake socket for RDMA-CM client: %s (code %d)\n", strerror(errno), errno);
+        rdma_destroy_id(ev->id);
+        return;
+    }
     auto rdma_context = rdmacm_create_qp(ev->id);
     if (!rdma_context)
     {
@@ -289,12 +298,12 @@ void osd_messenger_t::rdmacm_accept(rdma_cm_event *ev)
     // Wait for RDMA_CM_ESTABLISHED, and enable the connection only after it
     auto conn = new rdmacm_connecting_t;
     conn->cmid = ev->id;
-    conn->client_id = next_client_id++;
+    conn->peer_fd = fake_fd;
     conn->parsed_addr = *(sockaddr_storage*)rdma_get_peer_addr(ev->id);
     conn->rdma_context = rdma_context;
     rdmacm_set_conn_timeout(conn);
     rdmacm_connecting[ev->id] = conn;
-    fprintf(stderr, "[OSD %ju] new client %ju: connection from %s via RDMA-CM\n", this->osd_num, conn->client_id,
+    fprintf(stderr, "[OSD %ju] new client %d: connection from %s via RDMA-CM\n", this->osd_num, conn->peer_fd,
         addr_to_string(conn->parsed_addr).c_str());
 }
 
@@ -324,6 +333,8 @@ void osd_messenger_t::rdmacm_on_connect_peer_error(rdma_cm_id *cmid, int res)
     auto peer_osd = conn->peer_osd;
     if (conn->timeout_id >= 0)
         tfd->clear_timer(conn->timeout_id);
+    if (conn->peer_fd >= 0)
+        close(conn->peer_fd);
     if (conn->rdma_context)
         conn->rdma_context->reserve_cqe(-rdma_max_send-rdma_max_recv);
     if (conn->cmid)
@@ -344,7 +355,7 @@ void osd_messenger_t::rdmacm_on_connect_peer_error(rdma_cm_id *cmid, int res)
         else
         {
             // TCP is disabled
-            on_connect_peer(peer_osd, res == 0 ? -EINVAL : (res > 0 ? -res : res), 0);
+            on_connect_peer(peer_osd, res == 0 ? -EINVAL : (res > 0 ? -res : res));
         }
     }
 }
@@ -355,7 +366,7 @@ void osd_messenger_t::rdmacm_try_connect_peer(uint64_t peer_osd, const std::stri
     if (!string_to_addr(addr, false, rdmacm_port, &sa))
     {
         fprintf(stderr, "Address %s is invalid\n", addr.c_str());
-        on_connect_peer(peer_osd, -EINVAL, 0);
+        on_connect_peer(peer_osd, -EINVAL);
         return;
     }
     rdma_cm_id *cmid = NULL;
@@ -366,7 +377,17 @@ void osd_messenger_t::rdmacm_try_connect_peer(uint64_t peer_osd, const std::stri
         if (!disable_tcp)
             try_connect_peer_tcp(peer_osd, addr.c_str(), fallback_tcp_port);
         else
-            on_connect_peer(peer_osd, res, 0);
+            on_connect_peer(peer_osd, res);
+        return;
+    }
+    // Make a fake FD (FIXME: do not use FDs for identifying clients!)
+    int fake_fd = socket(AF_INET, SOCK_STREAM, 0);
+    if (fake_fd < 0)
+    {
+        int res = -errno;
+        rdma_destroy_id(cmid);
+        // Can't create socket, pointless to try TCP
+        on_connect_peer(peer_osd, res);
         return;
     }
     if (log_level > 0)
@@ -374,7 +395,7 @@ void osd_messenger_t::rdmacm_try_connect_peer(uint64_t peer_osd, const std::stri
     auto conn = new rdmacm_connecting_t;
     rdmacm_connecting[cmid] = conn;
     conn->cmid = cmid;
-    conn->client_id = next_client_id++;
+    conn->peer_fd = fake_fd;
     conn->peer_osd = peer_osd;
     conn->addr = addr;
     conn->parsed_addr = sa;
@@ -491,13 +512,13 @@ void osd_messenger_t::rdmacm_established(rdma_cm_event *ev)
     auto cl = new osd_client_t();
     cl->peer_addr = conn->parsed_addr;
     cl->peer_port = conn->rdmacm_port;
-    cl->client_id = conn->client_id;
+    cl->peer_fd = conn->peer_fd;
     cl->peer_state = PEER_RDMA;
     cl->connect_timeout_id = -1;
     cl->osd_num = peer_osd;
     cl->in_buf = malloc_or_die(receive_buffer_size);
     cl->rdma_conn = rc;
-    clients[conn->client_id] = cl;
+    clients[conn->peer_fd] = cl;
     if (conn->timeout_id >= 0)
     {
         tfd->clear_timer(conn->timeout_id);