platform_system_core/init/mount_namespace.cpp

325 lines
12 KiB
C++
Raw Normal View History

Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
/*
* Copyright (C) 2019 The Android Open Source Project
*
* 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.
*/
#include "mount_namespace.h"
#include <sys/mount.h>
#include <string>
#include <vector>
#include <ApexProperties.sysprop.h>
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
#include <android-base/file.h>
#include <android-base/logging.h>
#include <android-base/properties.h>
#include <android-base/result.h>
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
#include <android-base/unique_fd.h>
#include <apex_manifest.pb.h>
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
#include "util.h"
namespace android {
namespace init {
namespace {
static bool BindMount(const std::string& source, const std::string& mount_point,
bool recursive = false) {
unsigned long mountflags = MS_BIND;
if (recursive) {
mountflags |= MS_REC;
}
if (mount(source.c_str(), mount_point.c_str(), nullptr, mountflags, nullptr) == -1) {
PLOG(ERROR) << "Failed to bind mount " << source;
return false;
}
return true;
}
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
static bool MakeShared(const std::string& mount_point, bool recursive = false) {
unsigned long mountflags = MS_SHARED;
if (recursive) {
mountflags |= MS_REC;
}
if (mount(nullptr, mount_point.c_str(), nullptr, mountflags, nullptr) == -1) {
PLOG(ERROR) << "Failed to change propagation type to shared";
return false;
}
return true;
}
static bool MakeSlave(const std::string& mount_point, bool recursive = false) {
unsigned long mountflags = MS_SLAVE;
if (recursive) {
mountflags |= MS_REC;
}
if (mount(nullptr, mount_point.c_str(), nullptr, mountflags, nullptr) == -1) {
PLOG(ERROR) << "Failed to change propagation type to slave";
return false;
}
return true;
}
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
static bool MakePrivate(const std::string& mount_point, bool recursive = false) {
unsigned long mountflags = MS_PRIVATE;
if (recursive) {
mountflags |= MS_REC;
}
if (mount(nullptr, mount_point.c_str(), nullptr, mountflags, nullptr) == -1) {
PLOG(ERROR) << "Failed to change propagation type to private";
return false;
}
return true;
}
static int OpenMountNamespace() {
int fd = open("/proc/self/ns/mnt", O_RDONLY | O_CLOEXEC);
if (fd < 0) {
PLOG(ERROR) << "Cannot open fd for current mount namespace";
}
return fd;
}
static std::string GetMountNamespaceId() {
std::string ret;
if (!android::base::Readlink("/proc/self/ns/mnt", &ret)) {
PLOG(ERROR) << "Failed to read namespace ID";
return "";
}
return ret;
}
static bool IsApexUpdatable() {
static bool updatable = android::sysprop::ApexProperties::updatable().value_or(false);
return updatable;
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
}
static Result<void> MountDir(const std::string& path, const std::string& mount_path) {
if (int ret = mkdir(mount_path.c_str(), 0755); ret != 0 && errno != EEXIST) {
return ErrnoError() << "Could not create mount point " << mount_path;
}
if (mount(path.c_str(), mount_path.c_str(), nullptr, MS_BIND, nullptr) != 0) {
return ErrnoError() << "Could not bind mount " << path << " to " << mount_path;
}
return {};
}
static Result<std::string> GetApexName(const std::string& apex_dir) {
const std::string manifest_path = apex_dir + "/apex_manifest.pb";
std::string content;
if (!android::base::ReadFileToString(manifest_path, &content)) {
return Error() << "Failed to read manifest file: " << manifest_path;
}
apex::proto::ApexManifest manifest;
if (!manifest.ParseFromString(content)) {
return Error() << "Can't parse manifest file: " << manifest_path;
}
return manifest.name();
}
static Result<void> ActivateFlattenedApexesFrom(const std::string& from_dir,
const std::string& to_dir) {
std::unique_ptr<DIR, decltype(&closedir)> dir(opendir(from_dir.c_str()), closedir);
if (!dir) {
return {};
}
dirent* entry;
while ((entry = readdir(dir.get())) != nullptr) {
if (entry->d_name[0] == '.') continue;
if (entry->d_type == DT_DIR) {
const std::string apex_path = from_dir + "/" + entry->d_name;
const auto apex_name = GetApexName(apex_path);
if (!apex_name.ok()) {
LOG(ERROR) << apex_path << " is not an APEX directory: " << apex_name.error();
continue;
}
const std::string mount_path = to_dir + "/" + (*apex_name);
if (auto result = MountDir(apex_path, mount_path); !result.ok()) {
return result;
}
}
}
return {};
}
static bool ActivateFlattenedApexesIfPossible() {
if (IsRecoveryMode() || IsApexUpdatable()) {
return true;
}
const std::string kApexTop = "/apex";
const std::vector<std::string> kBuiltinDirsForApexes = {
"/system/apex",
"/system_ext/apex",
"/product/apex",
"/vendor/apex",
};
for (const auto& dir : kBuiltinDirsForApexes) {
if (auto result = ActivateFlattenedApexesFrom(dir, kApexTop); !result.ok()) {
LOG(ERROR) << result.error();
return false;
}
}
return true;
}
static Result<void> MountLinkerConfigForDefaultNamespace() {
// No need to mount linkerconfig for default mount namespace if the path does not exist (which
// would mean it is already mounted)
if (access("/linkerconfig/default", 0) != 0) {
return {};
}
if (mount("/linkerconfig/default", "/linkerconfig", nullptr, MS_BIND | MS_REC, nullptr) != 0) {
return ErrnoError() << "Failed to mount linker configuration for default mount namespace.";
}
return {};
}
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
static android::base::unique_fd bootstrap_ns_fd;
static android::base::unique_fd default_ns_fd;
static std::string bootstrap_ns_id;
static std::string default_ns_id;
} // namespace
bool SetupMountNamespaces() {
// Set the propagation type of / as shared so that any mounting event (e.g.
// /data) is by default visible to all processes. When private mounting is
// needed for /foo/bar, then we will make /foo/bar as a mount point (by
// bind-mounting by to itself) and set the propagation type of the mount
// point to private.
if (!MakeShared("/", true /*recursive*/)) return false;
// /apex is a private mountpoint to give different sets of APEXes for
Activate system APEXes early Summary: Boot sequence around apexd is changed to make it possible for pre-apexd processes to use libraries from APEXes. They no longer need to wait for the apexd to finish activating APEXes, which again can be done only after /data/ is mounted. This improves overall boot performance. Detail: This change fixes the problem that processes that are started before apexd (so called pre-apexd processes) can't access libraries that are provided only by the APEXes but are not found in the system partition (e.g. libdexfile_external.so, etc.). Main idea is to activate system APEXes (/system/apex/*.apex) before /data is mounted and then activate the updated APEXes (/data/apex/*.apex) after the /data mount. Detailed boot sequence is as follows. 1) init prepares the bootstrap and default mount namespaces. A tmpfs is mounted on /apex and the propagation type of the mountpoint is set to private. 2) before any other process is started, apexd is started in bootstrap mode. When executed in the mode, apexd only activates APEXes under /system/apex. Note that APEXes activated in this phase are mounted in the bootstrap mount namespace only. 3) other pre-apexd processes are started. They are in the bootstrap mount namespace and thus are provided with the libraries from the system APEXes. 4) /data is mounted. init switches into the default mount namespace and starts apexd as a daemon as usual. 5) apexd scans both /data/apex and /system/apex, and activate latest APEXes from the directories. Note that APEXes activated in this phase are mounted in the default namespaces only and thus are not visible to the pre-apexd processes. Bug: 125549215 Test: m; device boots Change-Id: I21c60d0ebe188fa4f24d6e6861f85ca204843069
2019-02-22 14:15:25 +01:00
// the bootstrap and default mount namespaces. The processes running with
// the bootstrap namespace get APEXes from the read-only partition.
if (!(MakePrivate("/apex"))) return false;
// /linkerconfig is a private mountpoint to give a different linker configuration
// based on the mount namespace. Subdirectory will be bind-mounted based on current mount
// namespace
if (!(MakePrivate("/linkerconfig"))) return false;
// The two mount namespaces present challenges for scoped storage, because
// vold, which is responsible for most of the mounting, lives in the
// bootstrap mount namespace, whereas most other daemons and all apps live
// in the default namespace. Scoped storage has a need for a
// /mnt/installer view that is a slave bind mount of /mnt/user - in other
// words, all mounts under /mnt/user should automatically show up under
// /mnt/installer. However, additional mounts done under /mnt/installer
// should not propagate back to /mnt/user. In a single mount namespace
// this is easy to achieve, by simply marking the /mnt/installer a slave
// bind mount. Unfortunately, if /mnt/installer is only created and
// bind mounted after the two namespaces are created below, we end up
// with the following situation:
// /mnt/user and /mnt/installer share the same peer group in both the
// bootstrap and default namespaces. Marking /mnt/installer slave in either
// namespace means that it won't propagate events to the /mnt/installer in
// the other namespace, which is still something we require - vold is the
// one doing the mounting under /mnt/installer, and those mounts should
// show up in the default namespace as well.
//
// The simplest solution is to do the bind mount before the two namespaces
// are created: the effect is that in both namespaces, /mnt/installer is a
// slave to the /mnt/user mount, and at the same time /mnt/installer in the
// bootstrap namespace shares a peer group with /mnt/installer in the
// default namespace.
if (!mkdir_recursive("/mnt/user", 0755)) return false;
if (!mkdir_recursive("/mnt/installer", 0755)) return false;
if (!(BindMount("/mnt/user", "/mnt/installer", true))) return false;
// First, make /mnt/installer a slave bind mount
if (!(MakeSlave("/mnt/installer"))) return false;
// Then, make it shared again - effectively creating a new peer group, that
// will be inherited by new mount namespaces.
if (!(MakeShared("/mnt/installer"))) return false;
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
bootstrap_ns_fd.reset(OpenMountNamespace());
bootstrap_ns_id = GetMountNamespaceId();
// When APEXes are updatable (e.g. not-flattened), we create separate mount
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
// namespaces for processes that are started before and after the APEX is
// activated by apexd. In the namespace for pre-apexd processes, small
// number of essential APEXes (e.g. com.android.runtime) are activated.
// In the namespace for post-apexd processes, all APEXes are activated.
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
bool success = true;
if (IsApexUpdatable() && !IsRecoveryMode()) {
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
// Creating a new namespace by cloning, saving, and switching back to
// the original namespace.
if (unshare(CLONE_NEWNS) == -1) {
PLOG(ERROR) << "Cannot create mount namespace";
return false;
}
default_ns_fd.reset(OpenMountNamespace());
default_ns_id = GetMountNamespaceId();
if (setns(bootstrap_ns_fd.get(), CLONE_NEWNS) == -1) {
PLOG(ERROR) << "Cannot switch back to bootstrap mount namespace";
return false;
}
} else {
// Otherwise, default == bootstrap
default_ns_fd.reset(OpenMountNamespace());
default_ns_id = GetMountNamespaceId();
}
success &= ActivateFlattenedApexesIfPossible();
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
LOG(INFO) << "SetupMountNamespaces done";
return success;
}
bool SwitchToDefaultMountNamespace() {
if (IsRecoveryMode()) {
// we don't have multiple namespaces in recovery mode
return true;
}
if (default_ns_id != GetMountNamespaceId()) {
if (setns(default_ns_fd.get(), CLONE_NEWNS) == -1) {
PLOG(ERROR) << "Failed to switch back to the default mount namespace.";
return false;
}
if (auto result = MountLinkerConfigForDefaultNamespace(); !result.ok()) {
LOG(ERROR) << result.error();
return false;
}
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
}
Activate system APEXes early Summary: Boot sequence around apexd is changed to make it possible for pre-apexd processes to use libraries from APEXes. They no longer need to wait for the apexd to finish activating APEXes, which again can be done only after /data/ is mounted. This improves overall boot performance. Detail: This change fixes the problem that processes that are started before apexd (so called pre-apexd processes) can't access libraries that are provided only by the APEXes but are not found in the system partition (e.g. libdexfile_external.so, etc.). Main idea is to activate system APEXes (/system/apex/*.apex) before /data is mounted and then activate the updated APEXes (/data/apex/*.apex) after the /data mount. Detailed boot sequence is as follows. 1) init prepares the bootstrap and default mount namespaces. A tmpfs is mounted on /apex and the propagation type of the mountpoint is set to private. 2) before any other process is started, apexd is started in bootstrap mode. When executed in the mode, apexd only activates APEXes under /system/apex. Note that APEXes activated in this phase are mounted in the bootstrap mount namespace only. 3) other pre-apexd processes are started. They are in the bootstrap mount namespace and thus are provided with the libraries from the system APEXes. 4) /data is mounted. init switches into the default mount namespace and starts apexd as a daemon as usual. 5) apexd scans both /data/apex and /system/apex, and activate latest APEXes from the directories. Note that APEXes activated in this phase are mounted in the default namespaces only and thus are not visible to the pre-apexd processes. Bug: 125549215 Test: m; device boots Change-Id: I21c60d0ebe188fa4f24d6e6861f85ca204843069
2019-02-22 14:15:25 +01:00
LOG(INFO) << "Switched to default mount namespace";
return true;
}
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
bool SwitchToBootstrapMountNamespaceIfNeeded() {
if (IsRecoveryMode()) {
// we don't have multiple namespaces in recovery mode
return true;
}
if (bootstrap_ns_id != GetMountNamespaceId() && bootstrap_ns_fd.get() != -1 &&
IsApexUpdatable()) {
Proper mount namespace configuration for bionic This CL fixes the design problem of the previous mechanism for providing the bootstrap bionic and the runtime bionic to the same path. Previously, bootstrap bionic was self-bind-mounted; i.e. /system/bin/libc.so is bind-mounted to itself. And the runtime bionic was bind-mounted on top of the bootstrap bionic. This has not only caused problems like `adb sync` not working(b/122737045), but also is quite difficult to understand due to the double-and-self mounting. This is the new design: Most importantly, these four are all distinct: 1) bootstrap bionic (/system/lib/bootstrap/libc.so) 2) runtime bionic (/apex/com.android.runtime/lib/bionic/libc.so) 3) mount point for 1) and 2) (/bionic/lib/libc.so) 4) symlink for 3) (/system/lib/libc.so -> /bionic/lib/libc.so) Inside the mount namespace of the pre-apexd processes, 1) is bind-mounted to 3). Likewise, inside the mount namespace of the post-apexd processes, 2) is bind-mounted to 3). In other words, there is no self-mount, and no double-mount. Another change is that mount points are under /bionic and the legacy paths become symlinks to the mount points. This is to make sure that there is no bind mounts under /system, which is breaking some apps. Finally, code for creating mount namespaces, mounting bionic, etc are refactored to mount_namespace.cpp Bug: 120266448 Bug: 123275379 Test: m, device boots, adb sync/push/pull works, especially with following paths: /bionic/lib64/libc.so /bionic/bin/linker64 /system/lib64/bootstrap/libc.so /system/bin/bootstrap/linker64 Change-Id: Icdfbdcc1efca540ac854d4df79e07ee61fca559f
2019-01-16 15:00:59 +01:00
if (setns(bootstrap_ns_fd.get(), CLONE_NEWNS) == -1) {
PLOG(ERROR) << "Failed to switch to bootstrap mount namespace.";
return false;
}
}
return true;
}
} // namespace init
} // namespace android