Merge pull request #38137 from tonistiigi/seccomp-ptrace

seccomp: allow ptrace(2) for 4.8+ kernels

api/types/seccomp.go

@@ -77,8 +77,9 @@ type Arg struct {
 
 // Filter is used to conditionally apply Seccomp rules
 type Filter struct {
-	Caps   []string `json:"caps,omitempty"`
-	Arches []string `json:"arches,omitempty"`
+	Caps      []string `json:"caps,omitempty"`
+	Arches    []string `json:"arches,omitempty"`
+	MinKernel string   `json:"minKernel,omitempty"`
 }
 
 // Syscall is used to match a group of syscalls in Seccomp

profiles/seccomp/default.json

@@ -366,6 +366,18 @@
 			"includes": {},
 			"excludes": {}
 		},
+		{
+			"names": [
+				"ptrace"
+			],
+			"action": "SCMP_ACT_ALLOW",
+			"args": null,
+			"comment": "",
+			"includes": {
+				"minKernel": "4.8.0"
+			},
+			"excludes": {}
+		},
 		{
 			"names": [
 				"personality"

profiles/seccomp/seccomp.go

@@ -8,7 +8,8 @@ import (
 	"fmt"
 
 	"github.com/docker/docker/api/types"
-	"github.com/opencontainers/runtime-spec/specs-go"
+	"github.com/docker/docker/pkg/parsers/kernel"
+	specs "github.com/opencontainers/runtime-spec/specs-go"
 	libseccomp "github.com/seccomp/libseccomp-golang"
 )
 
@@ -95,6 +96,21 @@ func setupSeccomp(config *types.Seccomp, rs *specs.Spec) (*specs.LinuxSeccomp, e
 
 	newConfig.DefaultAction = specs.LinuxSeccompAction(config.DefaultAction)
 
+	var currentKernelVersion *kernel.VersionInfo
+	kernelGreaterEqualThan := func(v string) (bool, error) {
+		version, err := kernel.ParseRelease(v)
+		if err != nil {
+			return false, err
+		}
+		if currentKernelVersion == nil {
+			currentKernelVersion, err = kernel.GetKernelVersion()
+			if err != nil {
+				return false, err
+			}
+		}
+		return kernel.CompareKernelVersion(*version, *currentKernelVersion) <= 0, nil
+	}
+
 Loop:
 	// Loop through all syscall blocks and convert them to libcontainer format after filtering them
 	for _, call := range config.Syscalls {
@@ -110,6 +126,13 @@ Loop:
 				}
 			}
 		}
+		if call.Excludes.MinKernel != "" {
+			if ok, err := kernelGreaterEqualThan(call.Excludes.MinKernel); err != nil {
+				return nil, err
+			} else if ok {
+				continue Loop
+			}
+		}
 		if len(call.Includes.Arches) > 0 {
 			if !inSlice(call.Includes.Arches, arch) {
 				continue Loop
@@ -122,6 +145,13 @@ Loop:
 				}
 			}
 		}
+		if call.Includes.MinKernel != "" {
+			if ok, err := kernelGreaterEqualThan(call.Includes.MinKernel); err != nil {
+				return nil, err
+			} else if !ok {
+				continue Loop
+			}
+		}
 
 		if call.Name != "" && len(call.Names) != 0 {
 			return nil, errors.New("'name' and 'names' were specified in the seccomp profile, use either 'name' or 'names'")

profiles/seccomp/seccomp_default.go

@@ -356,6 +356,13 @@ func DefaultProfile() *types.Seccomp {
 			Action: types.ActAllow,
 			Args:   []*types.Arg{},
 		},
+		{
+			Names:  []string{"ptrace"},
+			Action: types.ActAllow,
+			Includes: types.Filter{
+				MinKernel: "4.8.0",
+			},
+		},
 		{
 			Names:  []string{"personality"},
 			Action: types.ActAllow,