checkpoint-restore · osctobe · Jun 7, 2020 · Sep 28, 2020 · Aug 4, 2021 · Apr 7, 2022
diff --git a/.cirrus.yml b/.cirrus.yml
@@ -36,7 +36,10 @@ task:
     ln -sf /usr/include/google/protobuf/descriptor.proto images/google/protobuf/descriptor.proto
     dnf config-manager --set-enabled crb # Same as CentOS 8 powertools
     dnf -y install epel-release epel-next-release
-    dnf -y install --allowerasing asciidoc gcc git gnutls-devel libaio-devel libasan libcap-devel libnet-devel libnl3-devel libbsd-devel libselinux-devel make protobuf-c-devel protobuf-devel python-devel python-PyYAML python-future python-protobuf python-junit_xml python3-importlib-metadata python-flake8 xmlto
+    dnf -y install --allowerasing asciidoc gcc git gnutls-devel libaio-devel libasan libcap-devel libnet-devel libnl3-devel libbsd-devel libselinux-devel make protobuf-c-devel protobuf-devel python-devel python-PyYAML python-protobuf python-junit_xml python3-importlib-metadata python-flake8 xmlto libdrm-devel
+    # The image has a too old version of nettle which does not work with gnutls.
+    # Just upgrade to the latest to make the error go away.
+    dnf -y upgrade nettle nettle-devel
     systemctl stop sssd
     # Even with selinux in permissive mode the selinux tests will be executed.
     # The Cirrus CI user runs as a service from selinux point of view and is
@@ -108,7 +111,7 @@ task:
     yum install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm || :
     yum install -y dnf-plugins-core
     yum config-manager --set-enabled powertools
-    yum install -y --allowerasing asciidoc gcc git gnutls-devel libaio-devel libasan libcap-devel libnet-devel libnl3-devel libbsd-devel libselinux-devel make protobuf-c-devel protobuf-devel python3-devel python3-flake8 python3-PyYAML python3-future python3-protobuf python3-importlib-metadata python3-junit_xml xmlto
+    yum install -y --allowerasing asciidoc gcc git gnutls-devel libaio-devel libasan libcap-devel libnet-devel libnl3-devel libbsd-devel libselinux-devel make protobuf-c-devel protobuf-devel python3-devel python3-flake8 python3-PyYAML python3-protobuf python3-importlib-metadata python3-junit_xml xmlto libdrm-devel
     alternatives --set python /usr/bin/python3
     systemctl stop sssd
     # Even with selinux in permissive mode the selinux tests will be executed
@@ -120,38 +123,6 @@ task:
   build_script: |
     make -C scripts/ci local SKIP_CI_PREP=1 CC=gcc CD_TO_TOP=1 ZDTM_OPTS="-x zdtm/static/socket-raw"
 
-task:
-  name: CentOS 7 based test
-  environment:
-    HOME: "/root"
-    CIRRUS_WORKING_DIR: "/tmp/criu"
-
-  compute_engine_instance:
-    image_project: centos-cloud
-    image: family/centos-7
-    platform: linux
-    cpu: 4
-    memory: 8G
-
-  setup_script: |
-    # EPEL is needed for python2-future, python2-junit_xml, python-flake8 and libbsd-devel.
-    # Do not fail if latest epel repository definition is already installed
-    yum install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm || :
-    ln -sf /usr/include/google/protobuf/descriptor.proto images/google/protobuf/descriptor.proto
-    yum install -y findutils gcc git gnutls-devel iproute iptables libaio-devel libasan libcap-devel libnet-devel libnl3-devel libbsd-devel make procps-ng protobuf-c-devel protobuf-devel protobuf-python python python-flake8 python-ipaddress python2-future python2-junit_xml python-yaml python-six sudo tar which e2fsprogs python2-pip rubygem-asciidoctor libselinux-devel
-    # Even with selinux in permissive mode the selinux tests will be executed
-    # The Cirrus CI user runs as a service from selinux point of view and is
-    # much more restricted than a normal shell (system_u:system_r:unconfined_service_t:s0)
-    # The test case above (vagrant-fedora-no-vdso) should run selinux tests in enforcing mode
-    setenforce 0
-    # Enable user namespaces on CentOS 7
-    echo 10000 > /proc/sys/user/max_user_namespaces
-    # Adapt sudoers to our needs
-    echo 'root ALL=(ALL:ALL) ALL' | EDITOR='tee -a' visudo
-
-  build_script: |
-    make -C scripts/ci local SKIP_CI_PREP=1 CC=gcc CD_TO_TOP=1 ZDTM_IGNORE_TAINT=1 ZDTM_OPTS="-x zdtm/static/socket-raw -x zdtm/static/child_subreaper_existing_child -x zdtm/static/fifo_upon_unix_socket01 -x zdtm/static/overmount_sock -x zdtm/static/tempfs_overmounted"
-
 task:
   name: aarch64 build GCC (native)
   arm_container:

diff --git a/.clang-format b/.clang-format
@@ -53,7 +53,7 @@ BreakConstructorInitializersBeforeComma: false
 BreakConstructorInitializers: BeforeComma # Unknown to clang-format-4.0
 BreakAfterJavaFieldAnnotations: false
 BreakStringLiterals: false
-ColumnLimit: 120
+ColumnLimit: 0
 CommentPragmas: '^ IWYU pragma:'
 CompactNamespaces: false # Unknown to clang-format-4.0
 ConstructorInitializerAllOnOneLineOrOnePerLine: false

diff --git a/.codespellrc b/.codespellrc
@@ -1,3 +1,3 @@
 [codespell]
 skip = ./.git,./test/pki
-ignore-words-list = creat,fpr,fle,ue,bord,parms,nd,te,testng
+ignore-words-list = creat,fpr,fle,ue,bord,parms,nd,te,testng,inh,wronly,renderd,bui,clen
diff --git a/.github/workflows/check-commits.yml b/.github/workflows/check-commits.yml
@@ -0,0 +1,30 @@
+name: Verify self-contained commits
+
+on: pull_request
+
+# Cancel any preceding run on the pull request
+concurrency:
+  group: commit-test-${{ github.event.pull_request.number }}
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    # Check if pull request does not have label "not-selfcontained-ok"
+    if: "!contains(github.event.pull_request.labels.*.name, 'not-selfcontained-ok')"
+    steps:
+    - uses: actions/checkout@v3
+      with:
+        # Needed to rebase against the base branch
+        fetch-depth: 0
+        # Checkout pull request HEAD commit instead of merge commit
+        ref: ${{ github.event.pull_request.head.sha }}
+    - name: Install dependencies
+      run: sudo apt-get install -y libprotobuf-dev libprotobuf-c-dev protobuf-c-compiler protobuf-compiler python3-protobuf libnl-3-dev libnet-dev libcap-dev
+    - name: Configure git user details
+      run: |
+        git config --global user.email "[email protected]"
+        git config --global user.name "checkpoint-restore"
+    - name: Configure base branch without switching current branch
+      run: git fetch origin ${{ github.base_ref }}:${{ github.base_ref }}
+    - name: Build each commit
+      run: git rebase ${{ github.base_ref }} -x "make -C scripts/ci check-commit"
diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml
@@ -26,15 +26,15 @@ jobs:
       run: make lint
 
     - name: Run make indent
-      run: >
+      continue-on-error: true
+      run: |
         if [ -z "${{github.base_ref}}" ]; then
-          git fetch --deepen=1 &&
-          if ! make indent OPTS=--diff; then
-            exit 1
-          fi
+          git fetch --deepen=1
+          make indent
         else
-          git fetch origin ${{github.base_ref}} &&
-          if ! make indent OPTS=--diff BASE=origin/${{github.base_ref}}; then
-            exit 1
-          fi
+          git fetch origin ${{github.base_ref}}
+          make indent BASE=origin/${{github.base_ref}}
         fi
+    - name: Raise in-line make indent warnings
+      run: |
+        git diff | ./scripts/github-indent-warnings.py
diff --git a/.github/workflows/loongarch64-qemu-test.yml b/.github/workflows/loongarch64-qemu-test.yml
@@ -0,0 +1,15 @@
+name: LoongArch64 Qemu Test
+
+on: [push, pull_request]
+
+# Cancel any preceding run on the pull request.
+concurrency:
+  group: loongarch64-qemu-test-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: ${{ github.ref != 'refs/heads/criu-dev' }}
+
+jobs:
+  build:
+    runs-on: ubuntu-22.04
+    steps:
+      - uses: actions/checkout@v2 
+      - run: sudo make -C scripts/ci loongarch64-qemu-test
diff --git a/.github/workflows/manage-labels.yml b/.github/workflows/manage-labels.yml
@@ -0,0 +1,14 @@
+name: Remove labels
+on: [issue_comment, pull_request_review_comment]
+jobs:
+  remove-labels-on-comments:
+    name: Remove labels on comments
+    if: github.event_name == 'issue_comment'
+    runs-on: ubuntu-latest
+    steps:
+      - uses: mondeja/remove-labels-gh-action@v1
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          labels: |
+            changes requested
+            awaiting reply
diff --git a/.gitignore b/.gitignore
@@ -25,13 +25,6 @@ images/google/protobuf/*.h
 .gitid
 criu/criu
 criu/unittest/unittest
-crit/crit
-criu/arch/*/sys-exec-tbl*.c
-# x86 syscalls-table is not generated
-!criu/arch/x86/sys-exec-tbl.c
-criu/arch/*/syscalls*.S
-criu/include/syscall-codes*.h
-criu/include/syscall*.h
 criu/include/version.h
 criu/pie/restorer-blob.h
 criu/pie/parasite-blob.h

diff --git a/.lgtm.yml b/.lgtm.yml
@@ -22,7 +22,6 @@ extraction:
       - "libbsd-dev"
       - "python3-yaml"
       - "libnl-route-3-dev"
-      - "python-future"
       - "gnutls-dev"
     configure:
       command:

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -46,19 +46,18 @@ This should create the `./criu/criu` executable.
 
 ## Edit the source code
 
-If you use ctags, you can generate the ctags file by running
-
-```
-        make tags
-```
-
 When you change the source code, please keep in mind the following code conventions:
 
+* code is written to be read, so the code readability is the most important thing you need to have in mind when preparing patches
 * we prefer tabs and indentations to be 8 characters width
-* CRIU mostly follows [Linux kernel coding style](https://www.kernel.org/doc/Documentation/process/coding-style.rst), but we are less strict than the kernel community.
+* we prefer line length of 80 characters or less, more is allowed if it helps with code readability
+* CRIU mostly follows [Linux kernel coding style](https://www.kernel.org/doc/Documentation/process/coding-style.rst), but we are less strict than the kernel community
 
-Other conventions can be learned from the source code itself. In short, make sure your new code
-looks similar to what is already there.
+Other conventions can be learned from the source code itself. In short, make sure your new code looks similar to what is already there.
+
+## Automatic tools to fix coding-style
+
+Important: These tools are there to advise you, but should not be considered as a "source of truth", as tools also make nasty mistakes from time to time which can completely break code readability.
 
 The following command can be used to automatically run a code linter for Python files (flake8), Shell scripts (shellcheck),
 text spelling (codespell), and a number of CRIU-specific checks (usage of print macros and EOL whitespace for C files).
@@ -90,6 +89,41 @@ to check the last *N* commits for formatting errors, without applying the change
 Note that for pull requests, the "Run code linter" workflow runs these checks for all commits. If a clang-format error is detected
 we need to review the suggested changes and decide if they should be fixed before merging.
 
+Here are some bad examples of clang-format-ing:
+
+* if clang-format tries to force 120 characters and breaks readability - it is wrong:
+
+```
+@@ -58,8 +59,7 @@ static int register_membarriers(void)
+         }
+
+         if (!all_ok) {
+-                fail("can't register membarrier()s - tried %#x, kernel %#x",
+-                     barriers_registered, barriers_supported);
++                fail("can't register membarrier()s - tried %#x, kernel %#x", barriers_registered, barriers_supported);
+                 return -1;
+         }
+```
+
+* if clang-format breaks your beautiful readability friendly alignment in structures, comments or defines - it is wrong:
+
+```
+--- a/test/zdtm/static/membarrier.c
++++ b/test/zdtm/static/membarrier.c
+@@ -27,9 +27,10 @@ static const struct {
+        int register_cmd;
+        int execute_cmd;
+ } membarrier_cmds[] = {
+-       { "",           MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED,           MEMBARRIER_CMD_PRIVATE_EXPEDITED },
+-       { "_SYNC_CORE", MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_SYNC_CORE, MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE },
+-       { "_RSEQ",      MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_RSEQ,      MEMBARRIER_CMD_PRIVATE_EXPEDITED_RSEQ },
++       { "", MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED, MEMBARRIER_CMD_PRIVATE_EXPEDITED },
++       { "_SYNC_CORE", MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_SYNC_CORE,
++         MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE },
++       { "_RSEQ", MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_RSEQ, MEMBARRIER_CMD_PRIVATE_EXPEDITED_RSEQ },
+ };
+```
+
 ## Test your changes
 
 CRIU comes with an extensive test suite. To check whether your changes introduce any regressions, run

diff --git a/Documentation/Makefile b/Documentation/Makefile
@@ -12,11 +12,9 @@ endif
 
 FOOTER		:= footer.txt
 SRC1		+= crit.txt
-ifeq ($(PYTHON),python3)
 SRC1		+= criu-ns.txt
-endif
 SRC1		+= compel.txt
-SRC1            += amdgpu_plugin.txt
+SRC1		+= criu-amdgpu-plugin.txt
 SRC8		+= criu.txt
 SRC		:= $(SRC1) $(SRC8)
 XMLS		:= $(patsubst %.txt,%.xml,$(SRC))

diff --git a/Documentation/amdgpu_plugin.txt → Documentation/criu-amdgpu-plugin.txt b/Documentation/amdgpu_plugin.txt → Documentation/criu-amdgpu-plugin.txt
@@ -3,7 +3,7 @@ ROCM Support(1)
 
 NAME
 ----
-amdgpu_plugin - A plugin extension to CRIU to support checkpoint/restore in
+criu-amdgpu-plugin - A plugin extension to CRIU to support checkpoint/restore in
 userspace for AMD GPUs.
 
 
@@ -22,7 +22,7 @@ Though *criu* is a great tool for checkpointing and restoring running
 applications, it has certain limitations such as it cannot handle
 applications that have device files open. In order to support *ROCm* based
 workloads with *criu* we need to augment criu's core functionality with a
-plugin based extension mechanism. *amdgpu_plugin* provides the necessary support
+plugin based extension mechanism. *criu-amdgpu-plugin* provides the necessary support
 to criu to allow Checkpoint / Restore with ROCm.
 
 
@@ -97,6 +97,15 @@ executing criu command.
     E.g:
     KFD_CAPABILITY_CHECK=1
 
+*KFD_MAX_BUFFER_SIZE*::
+    On some systems, VRAM sizes may exceed RAM sizes, and so buffers for dumping
+    and restoring VRAM may be unable to fit. Set to a nonzero value (in bytes)
+    to set a limit on the plugin's memory usage.
+    Default:0 (Disabled)
+
+    E.g:
+    KFD_MAX_BUFFER_SIZE="2G"
+
 
 AUTHOR
 ------

diff --git a/Documentation/criu.txt b/Documentation/criu.txt
@@ -155,6 +155,11 @@ not compatible with *--external* *dev*.
             notification message contains a file descriptor for
             the master pty
 
+        *query-ext-files*:::
+            called after the process tree is stopped and network is locked.
+            This hook is used only in the RPC mode. The notification reply
+            contains file ids to be added to external file list (may be empty).
+
 *--unprivileged*::
     This option tells *criu* to accept the limitations when running
     as non-root. Running as non-root requires *criu* at least to have
@@ -208,6 +213,9 @@ In other words, do not use it unless really needed.
 *-s*, *--leave-stopped*::
     Leave tasks in stopped state after checkpoint, instead of killing.
 
+*--no-resume-on-error*::
+    Leave tasks in stopped state even if checkpoint completed unsuccessfully.
+
 *--external* __type__**[**__id__**]:**__value__::
     Dump an instance of an external resource. The generic syntax is
     'type' of resource, followed by resource 'id' (enclosed in literal
@@ -457,6 +465,9 @@ The 'mode' may be one of the following:
 
     *nftables*::: Use nftables rules to drop the packets.
 
+    *skip*::: Don't lock the network. If *--tcp-close* is not used, the network
+    must be locked externally to allow CRIU to dump TCP connections.
+
 *restore*
 ~~~~~~~~~
 Restores previously checkpointed processes.