Priorities for bpf progs attached to the same tracepoint

.github/actions/vmtest/action.yml

@@ -0,0 +1,35 @@
+name: 'vmtest'
+description: 'Build + run vmtest'
+inputs:
+  arch:
+    description: 'what arch to test'
+    required: true
+    default: 'x86_64'
+runs:
+  using: "composite"
+  steps:
+    # 1. Setup environment
+    - name: Setup build environment
+      uses: libbpf/ci/setup-build-env@master
+    # 2. Build
+    - name: Build kernel image
+      shell: bash
+      run: ${GITHUB_ACTION_PATH}/build.sh ${{ inputs.arch }}
+    - name: Build selftests
+      shell: bash
+      run: ${GITHUB_ACTION_PATH}/build_selftests.sh
+      env:
+        VMLINUX_BTF: ${{ github.workspace }}/vmlinux
+    # 3. Test
+    - name: Prepare rootfs
+      uses: libbpf/ci/prepare-rootfs@master
+      with:
+        project-name: 'libbpf'
+        arch: ${{ inputs.arch }}
+        kernel-root: '.'
+    - name: Run selftests
+      uses: libbpf/ci/run-qemu@master
+      with:
+        arch: ${{ inputs.arch}}
+        img: '/tmp/root.img'
+        vmlinuz: '${{ github.workspace }}/vmlinuz'

.github/actions/vmtest/build.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+set -euo pipefail
+
+ARCH="$1"
+
+THISDIR="$(cd $(dirname $0) && pwd)"
+
+source "${THISDIR}"/helpers.sh
+
+travis_fold start build_kernel "Building kernel"
+
+cp ${GITHUB_WORKSPACE}/travis-ci/vmtest/configs/config-latest.${ARCH} .config
+
+make -j $((4*$(nproc))) olddefconfig all > /dev/null
+
+travis_fold end build_kernel

.github/actions/vmtest/build_selftests.sh

@@ -0,0 +1,42 @@
+#!/bin/bash
+
+set -euo pipefail
+
+THISDIR="$(cd $(dirname $0) && pwd)"
+
+source "${THISDIR}"/helpers.sh
+
+travis_fold start prepare_selftests "Building selftests"
+
+LLVM_VER=15
+LIBBPF_PATH="${REPO_ROOT}"
+
+PREPARE_SELFTESTS_SCRIPT=${THISDIR}/prepare_selftests-${KERNEL}.sh
+if [ -f "${PREPARE_SELFTESTS_SCRIPT}" ]; then
+	(cd "${REPO_ROOT}/${REPO_PATH}/tools/testing/selftests/bpf" && ${PREPARE_SELFTESTS_SCRIPT})
+fi
+
+if [[ "${KERNEL}" = 'LATEST' ]]; then
+	VMLINUX_H=
+else
+	VMLINUX_H=${THISDIR}/vmlinux.h
+fi
+
+cd ${REPO_ROOT}/${REPO_PATH}
+make \
+	CLANG=clang-${LLVM_VER} \
+	LLC=llc-${LLVM_VER} \
+	LLVM_STRIP=llvm-strip-${LLVM_VER} \
+	VMLINUX_BTF="${VMLINUX_BTF}" \
+	VMLINUX_H="${VMLINUX_H}" \
+	-C "${REPO_ROOT}/${REPO_PATH}/tools/testing/selftests/bpf" \
+	-j $((4*$(nproc))) > /dev/null
+cd -
+mkdir "${LIBBPF_PATH}"/selftests
+cp -R "${REPO_ROOT}/${REPO_PATH}/tools/testing/selftests/bpf" \
+	"${LIBBPF_PATH}"/selftests
+cd "${LIBBPF_PATH}"
+rm selftests/bpf/.gitignore
+git add selftests
+
+travis_fold end prepare_selftests

.github/actions/vmtest/helpers.sh

@@ -0,0 +1,44 @@
+# $1 - start or end
+# $2 - fold identifier, no spaces
+# $3 - fold section description
+travis_fold() {
+  local YELLOW='\033[1;33m'
+  local NOCOLOR='\033[0m'
+  if [ -z ${GITHUB_WORKFLOW+x} ]; then
+    echo travis_fold:$1:$2
+    if [ ! -z "${3:-}" ]; then
+      echo -e "${YELLOW}$3${NOCOLOR}"
+    fi
+    echo
+  else
+    if [ $1 = "start" ]; then
+      line="::group::$2"
+      if [ ! -z "${3:-}" ]; then
+        line="$line - ${YELLOW}$3${NOCOLOR}"
+      fi
+    else
+      line="::endgroup::"
+    fi
+    echo -e "$line"
+  fi
+}
+
+__print() {
+  local TITLE=""
+  if [[ -n $2 ]]; then
+      TITLE=" title=$2"
+  fi
+  echo "::$1${TITLE}::$3"
+}
+
+# $1 - title
+# $2 - message
+print_error() {
+  __print error $1 $2
+}
+
+# $1 - title
+# $2 - message
+print_notice() {
+  __print notice $1 $2
+}

.github/workflows/test.yml

@@ -0,0 +1,50 @@
+name: bpf-ci
+
+on:
+  pull_request:
+
+concurrency: 
+  group: ci-test-${{ github.head_ref }}
+  cancel-in-progress: true
+
+jobs:
+  VM_Test:
+    runs-on: ${{ matrix.runs_on }}
+    name: Kernel ${{ matrix.kernel }} on ${{ matrix.runs_on }} + selftests
+    timeout-minutes: 100
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - kernel: 'LATEST'
+            runs_on: ubuntu-latest
+            arch: 'x86_64'
+          - kernel: 'LATEST'
+            runs_on: z15
+            arch: 's390x'
+    env:
+      AUTHOR_EMAIL: "$(git log -1 --pretty=\"%aE\")"
+      KERNEL: LATEST
+      REPO_ROOT: ${{ github.workspace }}
+      REPO_PATH: ""
+    steps:
+      - uses: actions/checkout@v2
+      - if: ${{ github.repository != 'kernel-patches/bpf' && github.repository != 'kernel-patches/bpf-rc' }}
+        name: Download bpf-next tree
+        uses: libbpf/ci/get-linux-source@master
+        with:
+          dest: '.kernel'
+      - if: ${{ github.repository != 'kernel-patches/bpf' && github.repository != 'kernel-patches/bpf-rc' }}
+        name: Move linux source in place
+        shell: bash
+        run: |
+          rm -rf .kernel/.git
+          cp -rf .kernel/. .
+          rm -rf .kernel
+      - uses: libbpf/ci/patch-kernel@master
+        with:
+          patches-root: '${{ github.workspace }}/travis-ci/diffs'
+          repo-root: '${{ github.workspace }}'
+      - uses: ./.github/actions/vmtest
+        with:
+          arch: ${{ matrix.arch }}

README

@@ -1,18 +0,0 @@
-Linux kernel
-============
-
-There are several guides for kernel developers and users. These guides can
-be rendered in a number of formats, like HTML and PDF. Please read
-Documentation/admin-guide/README.rst first.
-
-In order to build the documentation, use ``make htmldocs`` or
-``make pdfdocs``.  The formatted documentation can also be read online at:
-
-    https://www.kernel.org/doc/html/latest/
-
-There are various text files in the Documentation/ subdirectory,
-several of them using the Restructured Text markup notation.
-
-Please read the Documentation/process/changes.rst file, as it contains the
-requirements for building and running the kernel, and information about
-the problems which may result by upgrading your kernel.

drivers/media/rc/bpf-lirc.c

@@ -160,7 +160,7 @@ static int lirc_bpf_attach(struct rc_dev *rcdev, struct bpf_prog *prog)
 		goto unlock;
 	}
 
-	ret = bpf_prog_array_copy(old_array, NULL, prog, 0, &new_array);
+	ret = bpf_prog_array_copy(old_array, NULL, prog, 0, 0, &new_array);
 	if (ret < 0)
 		goto unlock;
 
@@ -193,7 +193,7 @@ static int lirc_bpf_detach(struct rc_dev *rcdev, struct bpf_prog *prog)
 	}
 
 	old_array = lirc_rcu_dereference(raw->progs);
-	ret = bpf_prog_array_copy(old_array, prog, NULL, 0, &new_array);
+	ret = bpf_prog_array_copy(old_array, prog, NULL, 0, 0, &new_array);
 	/*
 	 * Do not use bpf_prog_array_delete_safe() as we would end up
 	 * with a dummy entry in the array, and the we would free the

include/linux/bpf.h

@@ -1235,6 +1235,7 @@ struct bpf_prog_array_item {
 	union {
 		struct bpf_cgroup_storage *cgroup_storage[MAX_BPF_CGROUP_STORAGE_TYPE];
 		u64 bpf_cookie;
+		u32 prio;
 	};
 };
 
@@ -1274,7 +1275,7 @@ int bpf_prog_array_copy_info(struct bpf_prog_array *array,
 int bpf_prog_array_copy(struct bpf_prog_array *old_array,
 			struct bpf_prog *exclude_prog,
 			struct bpf_prog *include_prog,
-			u64 bpf_cookie,
+			u64 bpf_cookie, u32 prio,
 			struct bpf_prog_array **new_array);
 
 struct bpf_run_ctx {};

include/linux/trace_events.h

@@ -736,7 +736,7 @@ trace_trigger_soft_disabled(struct trace_event_file *file)
 
 #ifdef CONFIG_BPF_EVENTS
 unsigned int trace_call_bpf(struct trace_event_call *call, void *ctx);
-int perf_event_attach_bpf_prog(struct perf_event *event, struct bpf_prog *prog, u64 bpf_cookie);
+int perf_event_attach_bpf_prog(struct perf_event *event, struct bpf_prog *prog, u64 bpf_cookie, u32 prio);
 void perf_event_detach_bpf_prog(struct perf_event *event);
 int perf_event_query_prog_array(struct perf_event *event, void __user *info);
 int bpf_probe_register(struct bpf_raw_event_map *btp, struct bpf_prog *prog);
@@ -754,7 +754,7 @@ static inline unsigned int trace_call_bpf(struct trace_event_call *call, void *c
 }
 
 static inline int
-perf_event_attach_bpf_prog(struct perf_event *event, struct bpf_prog *prog, u64 bpf_cookie)
+perf_event_attach_bpf_prog(struct perf_event *event, struct bpf_prog *prog, u64 bpf_cookie, u32 prio)
 {
 	return -EOPNOTSUPP;
 }
@@ -871,7 +871,8 @@ extern void ftrace_profile_free_filter(struct perf_event *event);
 void perf_trace_buf_update(void *record, u16 type);
 void *perf_trace_buf_alloc(int size, struct pt_regs **regs, int *rctxp);
 
-int perf_event_set_bpf_prog(struct perf_event *event, struct bpf_prog *prog, u64 bpf_cookie);
+int perf_event_set_bpf_prog(struct perf_event *event, struct bpf_prog *prog,
+			u64 bpf_cookie, u32 prio);
 void perf_event_free_bpf_prog(struct perf_event *event);
 
 void bpf_trace_run1(struct bpf_prog *prog, u64 arg1);

include/uapi/linux/bpf.h

@@ -1481,6 +1481,7 @@ union bpf_attr {
 				 * accessible through bpf_get_attach_cookie() BPF helper
 				 */
 				__u64		bpf_cookie;
+				__u32		prio;
 			} perf_event;
 			struct {
 				__u32		flags;

kernel/bpf/core.c

@@ -2405,13 +2405,14 @@ int bpf_prog_array_update_at(struct bpf_prog_array *array, int index,
 int bpf_prog_array_copy(struct bpf_prog_array *old_array,
 			struct bpf_prog *exclude_prog,
 			struct bpf_prog *include_prog,
-			u64 bpf_cookie,
+			u64 bpf_cookie, u32 prio,
 			struct bpf_prog_array **new_array)
 {
 	int new_prog_cnt, carry_prog_cnt = 0;
 	struct bpf_prog_array_item *existing, *new;
 	struct bpf_prog_array *array;
 	bool found_exclude = false;
+	bool found_less_prio = false;
 
 	/* Figure out how many existing progs we need to carry over to
 	 * the new array.
@@ -2458,16 +2459,30 @@ int bpf_prog_array_copy(struct bpf_prog_array *old_array,
 			    existing->prog == &dummy_bpf_prog.prog)
 				continue;
 
+			if (include_prog && existing->prio <= prio) {
+				found_less_prio = true;
+
+				new->prog = include_prog;
+				new->prio = prio;
+				new->bpf_cookie = bpf_cookie;
+
+				new++;
+			}
+
 			new->prog = existing->prog;
 			new->bpf_cookie = existing->bpf_cookie;
+			new->prio = existing->prio;
 			new++;
 		}
 	}
-	if (include_prog) {
+
+	if (include_prog && !found_less_prio) {
 		new->prog = include_prog;
 		new->bpf_cookie = bpf_cookie;
+		new->prio = prio;
 		new++;
 	}
+
 	new->prog = NULL;
 	*new_array = array;
 	return 0;

kernel/bpf/syscall.c

@@ -3009,7 +3009,8 @@ static int bpf_perf_link_attach(const union bpf_attr *attr, struct bpf_prog *pro
 	}
 
 	event = perf_file->private_data;
-	err = perf_event_set_bpf_prog(event, prog, attr->link_create.perf_event.bpf_cookie);
+	err = perf_event_set_bpf_prog(event, prog, attr->link_create.perf_event.bpf_cookie,
+								  attr->link_create.perf_event.prio);
 	if (err) {
 		bpf_link_cleanup(&link_primer);
 		goto out_put_file;

kernel/events/core.c

@@ -5753,7 +5753,7 @@ static long _perf_ioctl(struct perf_event *event, unsigned int cmd, unsigned lon
 		if (IS_ERR(prog))
 			return PTR_ERR(prog);
 
-		err = perf_event_set_bpf_prog(event, prog, 0);
+		err = perf_event_set_bpf_prog(event, prog, 0, 0);
 		if (err) {
 			bpf_prog_put(prog);
 			return err;
@@ -10172,7 +10172,7 @@ static inline bool perf_event_is_tracing(struct perf_event *event)
 }
 
 int perf_event_set_bpf_prog(struct perf_event *event, struct bpf_prog *prog,
-			    u64 bpf_cookie)
+			    u64 bpf_cookie, u32 prio)
 {
 	bool is_kprobe, is_tracepoint, is_syscall_tp;
 
@@ -10203,7 +10203,7 @@ int perf_event_set_bpf_prog(struct perf_event *event, struct bpf_prog *prog,
 			return -EACCES;
 	}
 
-	return perf_event_attach_bpf_prog(event, prog, bpf_cookie);
+	return perf_event_attach_bpf_prog(event, prog, bpf_cookie, prio);
 }
 
 void perf_event_free_bpf_prog(struct perf_event *event)
@@ -10226,7 +10226,7 @@ static void perf_event_free_filter(struct perf_event *event)
 }
 
 int perf_event_set_bpf_prog(struct perf_event *event, struct bpf_prog *prog,
-			    u64 bpf_cookie)
+			    u64 bpf_cookie, u32 prio)
 {
 	return -ENOENT;
 }