Send and receive backup over network using QR code

This adds functionality to send and receive a backup over the network
using a QR code.

The sender or provider prepares the backup, sets up a server that
waits for clients.  It provides a ticket in the form of a QR code
which contains connection and authentication information.

The receiver uses the QR code to connect to the provider and fetches
backup, restoring it locally.
This commit is contained in:
Floris Bruynooghe
2023-02-01 16:45:09 +01:00
parent e318f5c697
commit 6044e5961b
9 changed files with 1265 additions and 77 deletions

View File

@@ -160,9 +160,9 @@ impl<'a> BlobObject<'a> {
pub fn from_path(context: &'a Context, path: &Path) -> Result<BlobObject<'a>> {
let rel_path = path
.strip_prefix(context.get_blobdir())
.context("wrong blobdir")?;
.with_context(|| format!("wrong blobdir: {}", path.display()))?;
if !BlobObject::is_acceptible_blob_name(rel_path) {
return Err(format_err!("wrong name"));
return Err(format_err!("bad blob name: {}", rel_path.display()));
}
let name = rel_path.to_str().context("wrong name")?;
BlobObject::from_name(context, name.to_string())

View File

@@ -513,6 +513,13 @@ impl Context {
// Ongoing process allocation/free/check
/// Tries to acquire the global UI "ongoing" mutex.
///
/// This is for modal operations during which no other user actions are allowed. Only
/// one such operation is allowed at any given time.
///
/// The return value is a cancel token, which will release the ongoing mutex when
/// dropped.
pub(crate) async fn alloc_ongoing(&self) -> Result<Receiver<()>> {
let mut s = self.running_state.write().await;
ensure!(

View File

@@ -4,14 +4,18 @@
use std::any::Any;
use std::ffi::OsStr;
use std::iter::FusedIterator;
use std::path::{Path, PathBuf};
use ::pgp::types::KeyTrait;
use anyhow::{bail, ensure, format_err, Context as _, Result};
use async_channel::Receiver;
use futures::StreamExt;
use futures_lite::FutureExt;
use rand::{thread_rng, Rng};
use tokio::fs::{self, File};
use tokio::task::JoinError;
use tokio_stream::wrappers::ReadDirStream;
use tokio_tar::Archive;
use crate::blob::BlobObject;
@@ -19,6 +23,7 @@ use crate::chat::{self, delete_and_reset_all_device_msgs, ChatId};
use crate::config::Config;
use crate::contact::ContactId;
use crate::context::Context;
use crate::e2ee;
use crate::events::EventType;
use crate::key::{self, DcKey, DcSecretKey, SignedPublicKey, SignedSecretKey};
use crate::log::LogExt;
@@ -26,13 +31,13 @@ use crate::message::{Message, MsgId, Viewtype};
use crate::mimeparser::SystemMessage;
use crate::param::Param;
use crate::pgp;
use crate::qr::Qr;
use crate::sql;
use crate::stock_str;
use crate::tools::{
create_folder, delete_file, get_filesuffix_lc, open_file_std, read_file, time, write_file,
EmailAddress,
};
use crate::{e2ee, tools};
// Name of the database file in the backup.
const DBFILE_BACKUP_NAME: &str = "dc_database_backup.sqlite";
@@ -511,23 +516,9 @@ async fn export_backup(context: &Context, dir: &Path, passphrase: String) -> Res
let _d1 = DeleteOnDrop(temp_db_path.clone());
let _d2 = DeleteOnDrop(temp_path.clone());
context
.sql
.set_raw_config_int("backup_time", now as i32)
.await?;
sql::housekeeping(context).await.ok_or_log(context);
context
.sql
.execute("VACUUM;", paramsv![])
export_database(context, &temp_db_path, passphrase)
.await
.map_err(|e| warn!(context, "Vacuum failed, exporting anyway {}", e))
.ok();
ensure!(
context.scheduler.read().await.is_none(),
"cannot export backup, IO is running"
);
.context("could not export database")?;
info!(
context,
@@ -536,12 +527,6 @@ async fn export_backup(context: &Context, dir: &Path, passphrase: String) -> Res
dest_path.display(),
);
context
.sql
.export(&temp_db_path, passphrase)
.await
.with_context(|| format!("failed to backup plaintext database to {temp_db_path:?}"))?;
let res = export_backup_inner(context, &temp_db_path, &temp_path).await;
match &res {
@@ -579,32 +564,11 @@ async fn export_backup_inner(
.append_path_with_name(temp_db_path, DBFILE_BACKUP_NAME)
.await?;
let read_dir = tools::read_dir(context.get_blobdir()).await?;
let count = read_dir.len();
let mut written_files = 0;
let mut last_progress = 0;
for entry in read_dir.into_iter() {
let name = entry.file_name();
if !entry.file_type().await?.is_file() {
warn!(
context,
"Export: Found dir entry {} that is not a file, ignoring",
name.to_string_lossy()
);
continue;
}
let mut file = File::open(entry.path()).await?;
let path_in_archive = PathBuf::from(BLOBS_BACKUP_NAME).join(name);
let blobdir = BlobDirContents::new(context).await?;
for blob in blobdir.iter() {
let mut file = File::open(blob.to_abs_path()).await?;
let path_in_archive = PathBuf::from(BLOBS_BACKUP_NAME).join(blob.as_name());
builder.append_file(path_in_archive, &mut file).await?;
written_files += 1;
let progress = 1000 * written_files / count;
if progress != last_progress && progress > 10 && progress < 1000 {
// We already emitted ImexProgress(10) above
context.emit_event(EventType::ImexProgress(progress));
last_progress = progress;
}
}
builder.finish().await?;
@@ -763,6 +727,242 @@ where
Ok(())
}
/// Exports the database to *file*, encrypted using *passphrase*.
///
/// The directory of *file* must already exist, if *file* itself exists it will be
/// overwritten.
///
/// This also verifies that IO is not running during the export.
async fn export_database(context: &Context, dest: &Path, passphrase: String) -> Result<()> {
ensure!(
context.scheduler.read().await.is_none(),
"cannot export backup, IO is running"
);
let now = time().try_into().context("32-bit UNIX time overflow")?;
context.sql.set_raw_config_int("backup_time", now).await?;
sql::housekeeping(context).await.ok_or_log(context);
context
.sql
.execute("VACUUM;", paramsv![])
.await
.map_err(|e| warn!(context, "Vacuum failed, exporting anyway {}", e))
.ok();
context
.sql
.export(dest, passphrase)
.await
.with_context(|| format!("failed to backup database to {}", dest.display()))?;
Ok(())
}
#[derive(Debug)]
pub struct BackupSender {
/// A handle to the running provider.
provider: sendme::provider::Provider,
/// The ticket to retrieve the backup collection.
ticket: sendme::provider::Ticket,
/// Token holding the "ongoing" mutex. When this completes the provider should shut
/// down.
cancel_token: Receiver<()>,
}
impl BackupSender {
// - [x] check i/o is not running
// - [x] check we have secret key
// - [x] alloc ongoing
// - [ ] correctly cancel Provider when cancelled
// - [x] create auth token
// - [x] export backup with generated token as password
// - needs a path to store the database
// - [x] create the sendme database
// - [x] start provider
// - [ ] wait for one successful connection
// - [ ] shutdown when a connection is closed, successful or not.
//
// - [ ] provide progress report
/// Prepares for sending a backup to a second device.
///
/// Before calling this function all I/O must be stopped so that no changes to the blobs
/// or database are happening, this is done by calling the `dc_accounts_stop_io` or
/// `dc_stop_io` APIs first. TODO: Add the rust equivalents.
///
/// This will acquire the global "ongoing" mutex.
pub async fn perpare(context: &Context, dir: &Path) -> Result<Self> {
ensure!(
// TODO: Should we worry about path normalisation?
dir != context.get_blobdir(),
"Temporary database export directory should not be in blobdir"
);
e2ee::ensure_secret_key_exists(context)
.await
.context("Private key not available, aborting backup export")?;
// Acquire global "ongoing" mutex.
let cancel_token = context.alloc_ongoing().await?;
let res = tokio::select! {
biased;
res = Self::prepare_inner(context, dir) => {
match res {
Ok(slf) => {
// TODO: maybe this is the wrong place to log this
// TODO: Also needs to log progress somehow.
info!(context, "Waiting for remote to connect");
Ok(slf)
},
Err(err) => {
error!(context, "Failed to set up second device setup: {:#}", err);
Err(err)
},
}
},
_ = cancel_token.recv() => Err(format_err!("cancelled")),
};
// TODO: This is all wrong, too early to release
context.free_ongoing().await;
res
}
async fn prepare_inner(context: &Context, dir: &Path) -> Result<Self> {
// Generate the token up front: we also use it to encrypt the database.
let token = sendme::protocol::AuthToken::generate();
let dbfile = dir.join(DBFILE_BACKUP_NAME);
export_database(context, &dbfile, token.to_string())
.await
.context("Database export failed")?;
// Now we can be sure IO is not running.
let mut files = vec![sendme::provider::DataSource::from(dbfile)];
let blobdir = BlobDirContents::new(context).await?;
for blob in blobdir.iter() {
files.push(blob.to_abs_path().into());
}
// Start listening.
let (db, hash) = sendme::provider::create_collection(files).await?;
let provider = sendme::provider::Provider::builder(db)
.auth_token(token)
.spawn()?;
let ticket = provider.ticket(hash);
Ok(Self { provider, ticket })
}
pub fn qr(&self) -> Qr {
Qr::Backup {
ticket: self.ticket.clone(),
}
}
pub async fn join(self) -> Result<(), JoinError> {
// TODO: should wait for 1 transfer to complete or abort
self.provider.join().await
}
pub fn abort(&self) {
self.provider.abort()
}
}
/// All files in the blobdir.
///
/// This exists so we can have a [`BlobDirIter`] which needs something to own the data of
/// it's `&Path`. Use [`BlobDirContents::iter`] to create the iterator.
///
/// Additionally pre-allocating this means we get a length for progress report.
struct BlobDirContents<'a> {
inner: Vec<PathBuf>,
context: &'a Context,
}
impl<'a> BlobDirContents<'a> {
async fn new(context: &'a Context) -> Result<BlobDirContents<'a>> {
let readdir = fs::read_dir(context.get_blobdir()).await?;
let inner = ReadDirStream::new(readdir)
.filter_map(|entry| async move {
match entry {
Ok(entry) => Some(entry),
Err(err) => {
error!(context, "Failed to read blob file: {err}");
None
}
}
})
.filter_map(|entry| async move {
match entry.file_type().await.ok()?.is_file() {
true => Some(entry.path()),
false => {
warn!(
context,
"Export: Found blob dir entry {} that is not a file, ignoring",
entry.path().display()
);
None
}
}
})
.collect()
.await;
Ok(Self { inner, context })
}
fn iter(&self) -> BlobDirIter<'_> {
BlobDirIter::new(self.context, &self.inner)
}
}
/// A stream for [`Blob`]s.
///
/// The stream emits [`EventType::ImexProgress`] events as it being consumed.
///
/// Because we like to know the total number of blobs to emit progress all the blobs are
/// read up front and stored. Luckily this also makes our life easier, since we now only
/// need to implement `Iterator` and not `Stream`.
struct BlobDirIter<'a> {
paths: &'a [PathBuf],
offset: usize,
last_progress: usize,
context: &'a Context,
}
impl<'a> BlobDirIter<'a> {
fn new(context: &'a Context, paths: &'a [PathBuf]) -> BlobDirIter<'a> {
Self {
paths,
offset: 0,
last_progress: 0,
context,
}
}
}
impl<'a> Iterator for BlobDirIter<'a> {
type Item = BlobObject<'a>;
fn next(&mut self) -> Option<Self::Item> {
while let Some(path) = self.paths.get(self.offset) {
self.offset += 1;
// In theory this can error but we'd have corrupted filenames in the blobdir, so
// silently skipping them is fine.
match BlobObject::from_path(self.context, path) {
Ok(blob) => {
let progress = 1000 * self.offset / self.paths.len();
if progress != self.last_progress && progress > 10 && progress < 1000 {
self.context.emit_event(EventType::ImexProgress(progress));
self.last_progress = progress;
}
return Some(blob);
}
Err(err) => warn!(self.context, "{err}"),
}
}
None
}
}
impl FusedIterator for BlobDirIter<'_> {}
#[cfg(test)]
mod tests {
use ::pgp::armor::BlockType;

View File

@@ -64,6 +64,9 @@ pub enum Qr {
Account {
domain: String,
},
Backup {
ticket: sendme::provider::Ticket,
},
WebrtcInstance {
domain: String,
instance_pattern: String,