platform_build_soong/android/packaging.go

272 lines
8.9 KiB
Go
Raw Normal View History

add PackagingSpec Currently, installation of a module is defined as an action of copying the built artifact of the module to an install path like out/soong/host (for host modules) and out/target/product/<device>/<partition> (for device modules). After the modules are installed, the installed files are further processed to create packages like system.img, vendor.img, cvd-host-package.tar.gz, etc. This notion of installation seems to have originated from the old time when system.img is the primary product of the entire build process (modulo a few more like root.img). Packaging the installed files as the filesystem image was considered as a post-build step then. However, this model doesn't seem to fit well to the current and future environment where we have a lot more filesystem images (system, vendor, system_ext, product, ...). The filesystem images themselves are even grouped together to form a higher-level filesystem image like super.img. Furthermore, things like cvd-host-package.tar.gz requires us to be able to group some of the host tools in a format that isn't filesystem image. Lastly, we are expected to have more filesystem images that are subsets of system.img (and their friends) for the Android-like mini OS that will be running on on-device virtual machines. These all imply that the packaging (which we call installation today) is not a global post-build step, but a part of the build rules for creating the package-like modules. A model better fits to the new sitatuation might be this; a module specifies its built artifact and the path where it should be placed. The latter path is not rooted at out/. It's a relative path to the root directory which will be determined by another module that implements the packaging. For example, cc_library will have ./lib (or ./lib64), not out/target/product/<device>/<partition>/lib as the path. Then packages like system.img, cvd-host-package.tar.gz, etc. are explicitly modeled as modules and they have deps to other modules. Then the modules are placed at the relative path under the package root, and the entire root directory finally is packaged as the output file (be it img, tar.gz, or whatever). PackagingSpec is the first step to implement the new model. It abstracts a request to place a built artifact at a certain path in a package. It has extra information about whether the path should be a symlink or not, and whether the path is for an executable. It currently is created when InstallFiles (and its friends) are called, and can be retrieved via the new method PackagingSpecs(). In this CL, no one is using PackagingSpec. The installation is still done by the existing rules created in InstallFiles, etc. and the structs are not used for the filesystem images like system.img. Bug: 159685774 Bug: 172414391 Test: m Change-Id: Ie1dec72d1ac14382fc3b74e5c850472e9320d6a3
2020-11-09 13:08:34 +08:00
// Copyright 2020 Google Inc. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License")
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package android
import (
"fmt"
"path/filepath"
"github.com/google/blueprint"
)
// PackagingSpec abstracts a request to place a built artifact at a certain path in a package. A
// package can be the traditional <partition>.img, but isn't limited to those. Other examples could
// be a new filesystem image that is a subset of system.img (e.g. for an Android-like mini OS
// running on a VM), or a zip archive for some of the host tools.
add PackagingSpec Currently, installation of a module is defined as an action of copying the built artifact of the module to an install path like out/soong/host (for host modules) and out/target/product/<device>/<partition> (for device modules). After the modules are installed, the installed files are further processed to create packages like system.img, vendor.img, cvd-host-package.tar.gz, etc. This notion of installation seems to have originated from the old time when system.img is the primary product of the entire build process (modulo a few more like root.img). Packaging the installed files as the filesystem image was considered as a post-build step then. However, this model doesn't seem to fit well to the current and future environment where we have a lot more filesystem images (system, vendor, system_ext, product, ...). The filesystem images themselves are even grouped together to form a higher-level filesystem image like super.img. Furthermore, things like cvd-host-package.tar.gz requires us to be able to group some of the host tools in a format that isn't filesystem image. Lastly, we are expected to have more filesystem images that are subsets of system.img (and their friends) for the Android-like mini OS that will be running on on-device virtual machines. These all imply that the packaging (which we call installation today) is not a global post-build step, but a part of the build rules for creating the package-like modules. A model better fits to the new sitatuation might be this; a module specifies its built artifact and the path where it should be placed. The latter path is not rooted at out/. It's a relative path to the root directory which will be determined by another module that implements the packaging. For example, cc_library will have ./lib (or ./lib64), not out/target/product/<device>/<partition>/lib as the path. Then packages like system.img, cvd-host-package.tar.gz, etc. are explicitly modeled as modules and they have deps to other modules. Then the modules are placed at the relative path under the package root, and the entire root directory finally is packaged as the output file (be it img, tar.gz, or whatever). PackagingSpec is the first step to implement the new model. It abstracts a request to place a built artifact at a certain path in a package. It has extra information about whether the path should be a symlink or not, and whether the path is for an executable. It currently is created when InstallFiles (and its friends) are called, and can be retrieved via the new method PackagingSpecs(). In this CL, no one is using PackagingSpec. The installation is still done by the existing rules created in InstallFiles, etc. and the structs are not used for the filesystem images like system.img. Bug: 159685774 Bug: 172414391 Test: m Change-Id: Ie1dec72d1ac14382fc3b74e5c850472e9320d6a3
2020-11-09 13:08:34 +08:00
type PackagingSpec struct {
// Path relative to the root of the package
relPathInPackage string
// The path to the built artifact
srcPath Path
// If this is not empty, then relPathInPackage should be a symlink to this target. (Then
// srcPath is of course ignored.)
symlinkTarget string
// Whether relPathInPackage should be marked as executable or not
executable bool
}
// Get file name of installed package
func (p *PackagingSpec) FileName() string {
if p.relPathInPackage != "" {
return filepath.Base(p.relPathInPackage)
}
return ""
}
// Path relative to the root of the package
func (p *PackagingSpec) RelPathInPackage() string {
return p.relPathInPackage
}
type PackageModule interface {
Module
packagingBase() *PackagingBase
// AddDeps adds dependencies to the `deps` modules. This should be called in DepsMutator.
// When adding the dependencies, depTag is used as the tag. If `deps` modules are meant to
// be copied to a zip in CopyDepsToZip, `depTag` should implement PackagingItem marker interface.
AddDeps(ctx BottomUpMutatorContext, depTag blueprint.DependencyTag)
// CopyDepsToZip zips the built artifacts of the dependencies into the given zip file and
// returns zip entries in it. This is expected to be called in GenerateAndroidBuildActions,
// followed by a build rule that unzips it and creates the final output (img, zip, tar.gz,
// etc.) from the extracted files
CopyDepsToZip(ctx ModuleContext, zipOut WritablePath) []string
}
// PackagingBase provides basic functionality for packaging dependencies. A module is expected to
// include this struct and call InitPackageModule.
type PackagingBase struct {
properties PackagingProperties
// Allows this module to skip missing dependencies. In most cases, this is not required, but
// for rare cases like when there's a dependency to a module which exists in certain repo
// checkouts, this is needed.
IgnoreMissingDependencies bool
}
type depsProperty struct {
// Modules to include in this package
Deps []string `android:"arch_variant"`
}
type packagingMultilibProperties struct {
First depsProperty `android:"arch_variant"`
Common depsProperty `android:"arch_variant"`
Lib32 depsProperty `android:"arch_variant"`
Lib64 depsProperty `android:"arch_variant"`
}
type packagingArchProperties struct {
Arm64 depsProperty
Arm depsProperty
X86_64 depsProperty
X86 depsProperty
}
type PackagingProperties struct {
Deps []string `android:"arch_variant"`
Multilib packagingMultilibProperties `android:"arch_variant"`
Arch packagingArchProperties
}
func InitPackageModule(p PackageModule) {
base := p.packagingBase()
p.AddProperties(&base.properties)
}
func (p *PackagingBase) packagingBase() *PackagingBase {
return p
}
// From deps and multilib.*.deps, select the dependencies that are for the given arch deps is for
// the current archicture when this module is not configured for multi target. When configured for
// multi target, deps is selected for each of the targets and is NOT selected for the current
// architecture which would be Common.
func (p *PackagingBase) getDepsForArch(ctx BaseModuleContext, arch ArchType) []string {
var ret []string
if arch == ctx.Target().Arch.ArchType && len(ctx.MultiTargets()) == 0 {
ret = append(ret, p.properties.Deps...)
} else if arch.Multilib == "lib32" {
ret = append(ret, p.properties.Multilib.Lib32.Deps...)
} else if arch.Multilib == "lib64" {
ret = append(ret, p.properties.Multilib.Lib64.Deps...)
} else if arch == Common {
ret = append(ret, p.properties.Multilib.Common.Deps...)
}
for i, t := range ctx.MultiTargets() {
if t.Arch.ArchType == arch {
ret = append(ret, p.properties.Deps...)
if i == 0 {
ret = append(ret, p.properties.Multilib.First.Deps...)
}
}
}
if ctx.Arch().ArchType == Common {
switch arch {
case Arm64:
ret = append(ret, p.properties.Arch.Arm64.Deps...)
case Arm:
ret = append(ret, p.properties.Arch.Arm.Deps...)
case X86_64:
ret = append(ret, p.properties.Arch.X86_64.Deps...)
case X86:
ret = append(ret, p.properties.Arch.X86.Deps...)
}
}
return FirstUniqueStrings(ret)
}
func (p *PackagingBase) getSupportedTargets(ctx BaseModuleContext) []Target {
var ret []Target
// The current and the common OS targets are always supported
ret = append(ret, ctx.Target())
if ctx.Arch().ArchType != Common {
ret = append(ret, Target{Os: ctx.Os(), Arch: Arch{ArchType: Common}})
}
// If this module is configured for multi targets, those should be supported as well
ret = append(ret, ctx.MultiTargets()...)
return ret
}
// PackagingItem is a marker interface for dependency tags.
// Direct dependencies with a tag implementing PackagingItem are packaged in CopyDepsToZip().
type PackagingItem interface {
// IsPackagingItem returns true if the dep is to be packaged
IsPackagingItem() bool
}
// DepTag provides default implementation of PackagingItem interface.
// PackagingBase-derived modules can define their own dependency tag by embedding this, which
// can be passed to AddDeps() or AddDependencies().
type PackagingItemAlwaysDepTag struct {
}
// IsPackagingItem returns true if the dep is to be packaged
func (PackagingItemAlwaysDepTag) IsPackagingItem() bool {
return true
}
// See PackageModule.AddDeps
func (p *PackagingBase) AddDeps(ctx BottomUpMutatorContext, depTag blueprint.DependencyTag) {
for _, t := range p.getSupportedTargets(ctx) {
for _, dep := range p.getDepsForArch(ctx, t.Arch.ArchType) {
if p.IgnoreMissingDependencies && !ctx.OtherModuleExists(dep) {
continue
}
ctx.AddFarVariationDependencies(t.Variations(), depTag, dep)
}
}
}
// See PackageModule.CopyDepsToZip
func (p *PackagingBase) CopyDepsToZip(ctx ModuleContext, zipOut WritablePath) (entries []string) {
m := make(map[string]PackagingSpec)
ctx.VisitDirectDeps(func(child Module) {
if pi, ok := ctx.OtherModuleDependencyTag(child).(PackagingItem); !ok || !pi.IsPackagingItem() {
return
}
for _, ps := range child.TransitivePackagingSpecs() {
if _, ok := m[ps.relPathInPackage]; !ok {
m[ps.relPathInPackage] = ps
}
}
})
builder := NewRuleBuilder(pctx, ctx)
dir := PathForModuleOut(ctx, ".zip")
builder.Command().Text("rm").Flag("-rf").Text(dir.String())
builder.Command().Text("mkdir").Flag("-p").Text(dir.String())
seenDir := make(map[string]bool)
for _, k := range SortedStringKeys(m) {
ps := m[k]
destPath := dir.Join(ctx, ps.relPathInPackage).String()
destDir := filepath.Dir(destPath)
entries = append(entries, ps.relPathInPackage)
if _, ok := seenDir[destDir]; !ok {
seenDir[destDir] = true
builder.Command().Text("mkdir").Flag("-p").Text(destDir)
}
if ps.symlinkTarget == "" {
builder.Command().Text("cp").Input(ps.srcPath).Text(destPath)
} else {
builder.Command().Text("ln").Flag("-sf").Text(ps.symlinkTarget).Text(destPath)
}
if ps.executable {
builder.Command().Text("chmod").Flag("a+x").Text(destPath)
}
}
builder.Command().
BuiltTool("soong_zip").
FlagWithOutput("-o ", zipOut).
FlagWithArg("-C ", dir.String()).
Flag("-L 0"). // no compression because this will be unzipped soon
FlagWithArg("-D ", dir.String())
builder.Command().Text("rm").Flag("-rf").Text(dir.String())
builder.Build("zip_deps", fmt.Sprintf("Zipping deps for %s", ctx.ModuleName()))
return entries
}
// packagingSpecsDepSet is a thin type-safe wrapper around the generic depSet. It always uses
// topological order.
type packagingSpecsDepSet struct {
depSet
}
// newPackagingSpecsDepSet returns an immutable packagingSpecsDepSet with the given direct and
// transitive contents.
func newPackagingSpecsDepSet(direct []PackagingSpec, transitive []*packagingSpecsDepSet) *packagingSpecsDepSet {
return &packagingSpecsDepSet{*newDepSet(TOPOLOGICAL, direct, transitive)}
}
// ToList returns the packagingSpecsDepSet flattened to a list in topological order.
func (d *packagingSpecsDepSet) ToList() []PackagingSpec {
if d == nil {
return nil
}
return d.depSet.ToList().([]PackagingSpec)
}