Newer
Older
use std::{io::Cursor, net::IpAddr, sync::Arc, time::Duration};
use crate::{
service::media::{FileMeta, UrlPreviewData},
services, utils, Error, Result, Ruma,
};
use image::io::Reader as ImgReader;
use reqwest::Url;
use ruma::api::client::{
error::ErrorKind,
create_content, get_content, get_content_as_filename, get_content_thumbnail,
use tracing::{debug, error, info};
use webpage::HTML;
///
/// Returns max upload size.
pub async fn get_media_config_route(
_body: Ruma<get_media_config::v3::Request>,
) -> Result<get_media_config::v3::Response> {
Ok(get_media_config::v3::Response {
upload_size: services().globals.max_request_size().into(),
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
/// # `GET /_matrix/media/v3/preview_url`
///
/// Returns URL preview.
pub async fn get_media_preview_route(
body: Ruma<get_media_preview::v3::Request>,
) -> Result<get_media_preview::v3::Response> {
let url = &body.url;
if !url_preview_allowed(url) {
return Err(Error::BadRequest(
ErrorKind::Forbidden,
"URL is not allowed to be previewed",
));
}
if let Ok(preview) = get_url_preview(url).await {
let res = serde_json::value::to_raw_value(&preview).map_err(|e| {
error!(
"Failed to convert UrlPreviewData into a serde json value: {}",
e
);
Error::BadRequest(
ErrorKind::Unknown,
"Unknown error occurred parsing URL preview",
)
})?;
return Ok(get_media_preview::v3::Response::from_raw_value(res));
}
Err(Error::BadRequest(
ErrorKind::LimitExceeded {
retry_after_ms: Some(Duration::from_secs(5)),
},
"Retry later",
))
}
///
/// Permanently save media in the server.
///
/// - Some metadata will be saved in the database
/// - Media will be saved in the media/ directory
pub async fn create_content_route(
utils::random_string(MXC_LENGTH)
);
.as_ref()
.map(|filename| "inline; filename=".to_owned() + filename)
.as_deref(),
let content_uri = mxc.into();
/// helper method to fetch remote media from other servers over federation
pub async fn get_remote_content(
mxc: &str,
server_name: &ruma::ServerName,
allow_redirect: bool,
timeout_ms: Duration,
// we'll lie to the client and say the blocked server's media was not found and log.
// the client has no way of telling anyways so this is a security bonus.
if services()
.globals
.prevent_media_downloads_from()
.contains(&server_name.to_owned())
{
info!("Received request for remote media `{}` but server is in our media server blocklist. Returning 404.", mxc);
return Err(Error::BadRequest(ErrorKind::NotFound, "Media not found."));
}
.sending
.send_federation_request(
server_name,
},
)
.await?;
content_response.content_disposition.as_deref(),
content_response.content_type.as_deref(),
&content_response.file,
)
.await?;
Ok(content_response)
/// # `GET /_matrix/media/v3/download/{serverName}/{mediaId}`
///
/// Load media from our server or over federation.
///
/// - Only allows federation if `allow_remote` is true
/// - Only redirects if `allow_redirect` is true
/// - Uses client-provided `timeout_ms` if available, else defaults to 20 seconds
pub async fn get_content_route(
let mxc = format!("mxc://{}/{}", body.server_name, body.media_id);
}) = services().media.get(mxc.clone()).await?
content_type,
cross_origin_resource_policy: Some("cross-origin".to_owned()),
} else if &*body.server_name != services().globals.server_name() && body.allow_remote {
let remote_content_response = get_remote_content(
&mxc,
&body.server_name,
body.media_id.clone(),
body.allow_redirect,
body.timeout_ms,
)
.await?;
} else {
Err(Error::BadRequest(ErrorKind::NotFound, "Media not found."))
}
}
/// # `GET /_matrix/media/v3/download/{serverName}/{mediaId}/{fileName}`
///
/// Load media from our server or over federation, permitting desired filename.
///
/// - Only allows federation if `allow_remote` is true
/// - Only redirects if `allow_redirect` is true
/// - Uses client-provided `timeout_ms` if available, else defaults to 20 seconds
pub async fn get_content_as_filename_route(
let mxc = format!("mxc://{}/{}", body.server_name, body.media_id);
if let Some(FileMeta {
}) = services().media.get(mxc.clone()).await?
file,
content_type,
content_disposition: Some(format!("inline; filename={}", body.filename)),
cross_origin_resource_policy: Some("cross-origin".to_owned()),
} else if &*body.server_name != services().globals.server_name() && body.allow_remote {
let remote_content_response = get_remote_content(
&mxc,
&body.server_name,
body.media_id.clone(),
body.allow_redirect,
body.timeout_ms,
)
.await?;
content_disposition: Some(format!("inline: filename={}", body.filename)),
content_type: remote_content_response.content_type,
cross_origin_resource_policy: Some("cross-origin".to_owned()),
} else {
Err(Error::BadRequest(ErrorKind::NotFound, "Media not found."))
}
}
/// # `GET /_matrix/media/v3/thumbnail/{serverName}/{mediaId}`
///
/// Load media thumbnail from our server or over federation.
///
/// - Only allows federation if `allow_remote` is true
/// - Only redirects if `allow_redirect` is true
/// - Uses client-provided `timeout_ms` if available, else defaults to 20 seconds
pub async fn get_content_thumbnail_route(
let mxc = format!("mxc://{}/{}", body.server_name, body.media_id);
if let Some(FileMeta {
content_type, file, ..
body.width
.try_into()
.map_err(|_| Error::BadRequest(ErrorKind::InvalidParam, "Width is invalid."))?,
body.height
.try_into()
.map_err(|_| Error::BadRequest(ErrorKind::InvalidParam, "Height is invalid."))?,
Ok(get_content_thumbnail::v3::Response {
file,
content_type,
cross_origin_resource_policy: Some("cross-origin".to_owned()),
})
} else if &*body.server_name != services().globals.server_name() && body.allow_remote {
// we'll lie to the client and say the blocked server's media was not found and log.
// the client has no way of telling anyways so this is a security bonus.
if services()
.globals
.prevent_media_downloads_from()
.contains(&body.server_name.to_owned())
{
info!("Received request for remote media `{}` but server is in our media server blocklist. Returning 404.", mxc);
return Err(Error::BadRequest(ErrorKind::NotFound, "Media not found."));
}
server_name: body.server_name.clone(),
media_id: body.media_id.clone(),
timeout_ms: body.timeout_ms,
allow_redirect: body.allow_redirect,
None,
get_thumbnail_response.content_type.as_deref(),
body.width.try_into().expect("all UInts are valid u32s"),
body.height.try_into().expect("all UInts are valid u32s"),
&get_thumbnail_response.file,
)
.await?;
} else {
Err(Error::BadRequest(ErrorKind::NotFound, "Media not found."))
}
}
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
async fn download_image(client: &reqwest::Client, url: &str) -> Result<UrlPreviewData> {
let image = client.get(url).send().await?.bytes().await?;
let mxc = format!(
"mxc://{}/{}",
services().globals.server_name(),
utils::random_string(MXC_LENGTH)
);
services()
.media
.create(mxc.clone(), None, None, &image)
.await?;
let (width, height) = match ImgReader::new(Cursor::new(&image)).with_guessed_format() {
Err(_) => (None, None),
Ok(reader) => match reader.into_dimensions() {
Err(_) => (None, None),
Ok((width, height)) => (Some(width), Some(height)),
},
};
Ok(UrlPreviewData {
image: Some(mxc),
image_size: Some(image.len()),
image_width: width,
image_height: height,
..Default::default()
})
}
async fn download_html(client: &reqwest::Client, url: &str) -> Result<UrlPreviewData> {
let max_download_size = 300_000; // TODO: is this bytes? kilobytes? megabytes?
let mut response = client.get(url).send().await?;
let mut bytes: Vec<u8> = Vec::new();
while let Some(chunk) = response.chunk().await? {
bytes.extend_from_slice(&chunk);
if bytes.len() > max_download_size {
break;
}
}
let body = String::from_utf8_lossy(&bytes);
let html = match HTML::from_string(body.to_string(), Some(url.to_owned())) {
Ok(html) => html,
Err(_) => {
return Err(Error::BadRequest(
ErrorKind::Unknown,
"Failed to parse HTML",
))
}
};
let mut data = match html.opengraph.images.first() {
None => UrlPreviewData::default(),
Some(obj) => download_image(client, &obj.url).await?,
};
let props = html.opengraph.properties;
/* use OpenGraph title/description, but fall back to HTML if not available */
data.title = props.get("title").cloned().or(html.title);
data.description = props.get("description").cloned().or(html.description);
Ok(data)
}
fn url_request_allowed(addr: &IpAddr) -> bool {
// TODO: make this check ip_range_denylist
// could be implemented with reqwest when it supports IP filtering:
// https://github.com/seanmonstar/reqwest/issues/1515
// These checks have been taken from the Rust core/net/ipaddr.rs crate,
// IpAddr::V4.is_global() and IpAddr::V6.is_global(), as .is_global is not
// yet stabilized. TODO: Once this is stable, this match can be simplified.
match addr {
IpAddr::V4(ip4) => {
!(ip4.octets()[0] == 0 // "This network"
|| ip4.is_private()
|| (ip4.octets()[0] == 100 && (ip4.octets()[1] & 0b1100_0000 == 0b0100_0000)) // is_shared()
|| ip4.is_loopback()
|| ip4.is_link_local()
// addresses reserved for future protocols (`192.0.0.0/24`)
|| (ip4.octets()[0] == 192 && ip4.octets()[1] == 0 && ip4.octets()[2] == 0)
|| ip4.is_documentation()
|| (ip4.octets()[0] == 198 && (ip4.octets()[1] & 0xfe) == 18) // is_benchmarking()
|| (ip4.octets()[0] & 240 == 240 && !ip4.is_broadcast()) // is_reserved()
|| ip4.is_broadcast())
}
IpAddr::V6(ip6) => {
!(ip6.is_unspecified()
|| ip6.is_loopback()
// IPv4-mapped Address (`::ffff:0:0/96`)
|| matches!(ip6.segments(), [0, 0, 0, 0, 0, 0xffff, _, _])
// IPv4-IPv6 Translat. (`64:ff9b:1::/48`)
|| matches!(ip6.segments(), [0x64, 0xff9b, 1, _, _, _, _, _])
// Discard-Only Address Block (`100::/64`)
|| matches!(ip6.segments(), [0x100, 0, 0, 0, _, _, _, _])
// IETF Protocol Assignments (`2001::/23`)
|| (matches!(ip6.segments(), [0x2001, b, _, _, _, _, _, _] if b < 0x200)
&& !(
// Port Control Protocol Anycast (`2001:1::1`)
u128::from_be_bytes(ip6.octets()) == 0x2001_0001_0000_0000_0000_0000_0000_0001
// Traversal Using Relays around NAT Anycast (`2001:1::2`)
|| u128::from_be_bytes(ip6.octets()) == 0x2001_0001_0000_0000_0000_0000_0000_0002
// AMT (`2001:3::/32`)
|| matches!(ip6.segments(), [0x2001, 3, _, _, _, _, _, _])
// AS112-v6 (`2001:4:112::/48`)
|| matches!(ip6.segments(), [0x2001, 4, 0x112, _, _, _, _, _])
// ORCHIDv2 (`2001:20::/28`)
|| matches!(ip6.segments(), [0x2001, b, _, _, _, _, _, _] if (0x20..=0x2F).contains(&b))
))
|| ((ip6.segments()[0] == 0x2001) && (ip6.segments()[1] == 0xdb8)) // is_documentation()
|| ((ip6.segments()[0] & 0xfe00) == 0xfc00) // is_unique_local()
|| ((ip6.segments()[0] & 0xffc0) == 0xfe80)) // is_unicast_link_local
}
}
}
async fn request_url_preview(url: &str) -> Result<UrlPreviewData> {
let client = services().globals.url_preview_client();
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
let response = client.head(url).send().await?;
if !response
.remote_addr()
.map_or(false, |a| url_request_allowed(&a.ip()))
{
return Err(Error::BadRequest(
ErrorKind::Forbidden,
"Requesting from this address is forbidden",
));
}
let content_type = match response
.headers()
.get(reqwest::header::CONTENT_TYPE)
.and_then(|x| x.to_str().ok())
{
Some(ct) => ct,
None => {
return Err(Error::BadRequest(
ErrorKind::Unknown,
"Unknown Content-Type",
))
}
};
let data = match content_type {
html if html.starts_with("text/html") => download_html(&client, url).await?,
img if img.starts_with("image/") => download_image(&client, url).await?,
_ => {
return Err(Error::BadRequest(
ErrorKind::Unknown,
"Unsupported Content-Type",
))
}
};
services().media.set_url_preview(url, &data).await?;
Ok(data)
}
async fn get_url_preview(url: &str) -> Result<UrlPreviewData> {
if let Some(preview) = services().media.get_url_preview(url).await {
return Ok(preview);
}
// ensure that only one request is made per URL
let mutex_request = Arc::clone(
services()
.media
.url_preview_mutex
.write()
.unwrap()
.entry(url.to_owned())
.or_default(),
);
let _request_lock = mutex_request.lock().await;
match services().media.get_url_preview(url).await {
Some(preview) => Ok(preview),
None => request_url_preview(url).await,
}
}
fn url_preview_allowed(url_str: &str) -> bool {
let url: Url = match Url::parse(url_str) {
Ok(u) => u,
Err(_) => return false,
};
if ["http", "https"]
.iter()
.all(|&scheme| scheme != url.scheme().to_lowercase())
{
debug!("Ignoring non-HTTP/HTTPS URL to preview: {}", url);
return false;
}
let host = match url.host_str() {
None => {
debug!(
"Ignoring URL preview for a URL that does not have a host (?): {}",
url
);
return false;
}
Some(h) => h.to_owned(),
};
let allowlist_domain_contains = services().globals.url_preview_domain_contains_allowlist();
let allowlist_domain_explicit = services().globals.url_preview_domain_explicit_allowlist();
let allowlist_url_contains = services().globals.url_preview_url_contains_allowlist();
if allowlist_domain_contains.contains(&"*".to_owned())
|| allowlist_domain_explicit.contains(&"*".to_owned())
|| allowlist_url_contains.contains(&"*".to_owned())
{
debug!(
"Config key contains * which is allowing all URL previews. Allowing URL {}",
url
);
return true;
}
if !host.is_empty() {
if allowlist_domain_explicit.contains(&host) {
return true;
}
debug!(
"Host {} is allowed by url_preview_domain_explicit_allowlist (check 1/3)",
&host
);
if allowlist_domain_contains
.iter()
.any(|domain_s| domain_s.contains(&host.clone()))
{
return true;
}
debug!(
"Host {} is allowed by url_preview_domain_contains_allowlist (check 2/3)",
&host
);
if allowlist_url_contains
.iter()
.any(|url_s| url_s.contains(&url.to_string()))
{
return true;
}
debug!(
"URL {} is allowed by url_preview_url_contains_allowlist (check 3/3)",
&host
);
}
false
}