Sergey Kandaurov [Wed, 14 May 2025 19:33:00 +0000 (23:33 +0400)]
QUIC: reset qc->error to zero again.
Following the previous change that removed posting a close event
in OpenSSL compat layer, now ngx_quic_close_connection() is always
called on error path with either NGX_ERROR or qc->error set.
This allows to remove a special value -1 served as a missing error,
which simplifies the code. Partially reverts d3fb12d77.
Also, this improves handling of the draining connection state, which
consists of posting a close event with NGX_OK and no qc->error set,
where it was previously converted to NGX_QUIC_ERR_INTERNAL_ERROR.
Notably, this is rather a cosmetic fix, because drained connections
do not send any packets including CONNECTION_CLOSE, and qc->error
is not otherwise used.
Sergey Kandaurov [Tue, 13 May 2025 16:12:10 +0000 (20:12 +0400)]
QUIC: adjusted handling of callback errors.
Changed handshake callbacks to always return success. This allows to avoid
logging SSL_do_handshake() errors with empty or cryptic "internal error"
OpenSSL error messages at the inappropriate "crit" log level.
Further, connections with failed callbacks are closed now right away when
using OpenSSL compat layer. This change supersedes and reverts c37fdcdd1,
with the conditions to check callbacks invocation kept to slightly improve
code readability of control flow; they are optimized out in the resulting
assembly code.
Sergey Kandaurov [Wed, 21 May 2025 15:55:31 +0000 (19:55 +0400)]
QUIC: logging of SSL library errors.
Logging level for such errors, which should not normally happen,
is changed to NGX_LOG_ALERT, and ngx_log_error() is replaced with
ngx_ssl_error() for consistency with the rest of the code.
Previously, it was not possible to send acknowledgments if the
congestion window was limited or temporarily exceeded, such as
after sending a large response or MTU probe. If ACKs were not
received from the peer for some reason to update the in-flight
bytes counter below the congestion window, this might result in
a stalled connection.
The fix is to send ACKs regardless of congestion control. This
meets RFC 9002, Section 7:
: Similar to TCP, packets containing only ACK frames do not count
: toward bytes in flight and are not congestion controlled.
This is a simplified implementation to send ACK frames from the
head of the queue. This was made possible after 6f5f17358.
Reported in trac ticket #2621 and subsequently by Vladimir Homutov:
https://mailman.nginx.org/pipermail/nginx-devel/2025-April/ZKBAWRJVQXSZ2ISG3YJAF3EWMDRDHCMO.html
Aleksei Bavshin [Tue, 14 Jan 2025 19:11:28 +0000 (11:11 -0800)]
Win32: added detection of ARM64 target.
This extends the target selection implemented in dad6ec3aa63f to support
Windows ARM64 platforms. OpenSSL support for VC-WIN64-ARM target first
appeared in 1.1.1 and is present in all currently supported (3.x)
branches.
As a side effect, ARM64 Windows builds will get 16-byte alignment along
with the rest of non-x86 platforms. This is safe, as malloc on 64-bit
Windows guarantees the fundamental alignment of allocations, 16 bytes.
Aleksei Bavshin [Tue, 14 Jan 2025 18:32:24 +0000 (10:32 -0800)]
Core: improved NGX_ALIGNMENT detection on some x86_64 platforms.
Previously, the default pool alignment used sizeof(unsigned long), with
the expectation that this would match to a platform word size. Certain
64-bit platforms prove this assumption wrong by keeping the 32-bit long
type, which is fully compliant with the C standard.
This introduces a possibility of suboptimal misaligned access to the
data allocated with ngx_palloc() on the affected platforms, which is
addressed here by changing the default NGX_ALIGNMENT to a pointer size.
As we override the detection in auto/os/conf for all the machine types
except x86, and Unix-like 64-bit systems prefer the 64-bit long, the
impact of the change should be limited to Win64 x64.
Roman Arutyunyan [Fri, 18 Apr 2025 07:16:57 +0000 (11:16 +0400)]
HTTP/3: fixed NGX_HTTP_V3_VARLEN_INT_LEN value.
After fixing ngx_http_v3_encode_varlen_int() in 400eb1b628,
NGX_HTTP_V3_VARLEN_INT_LEN retained the old value of 4, which is
insufficient for the values over 1073741823 (1G - 1).
The NGX_HTTP_V3_VARLEN_INT_LEN macro is used in ngx_http_v3_uni.c to
format stream and frame types. Old buffer size is enough for formatting
this data. Also, the macro is used in ngx_http_v3_filter_module.c to
format output chunks and trailers. Considering output_buffers and
proxy_buffer_size are below 1G in all realistic scenarios, the old buffer
size is enough here as well.
Roman Arutyunyan [Mon, 14 Apr 2025 13:16:47 +0000 (17:16 +0400)]
QUIC: dynamic packet threshold.
RFC 9002, Section 6.1.1 defines packet reordering threshold as 3. Testing
shows that such low value leads to spurious packet losses followed by
congestion window collapse. The change implements dynamic packet threshold
detection based on in-flight packet range. Packet threshold is defined
as half the number of in-flight packets, with mininum value of 3.
Also, renamed ngx_quic_lost_threshold() to ngx_quic_time_threshold()
for better compliance with RFC 9002 terms.
Previosly the threshold was hardcoded at 10000. This value is too low for
high BDP networks. For example, if all frames are STREAM frames, and MTU
is 1500, the upper limit for congestion window would be roughly 15M
(10000 * 1500). With 100ms RTT it's just a 1.2Gbps network (15M * 10 * 8).
In reality, the limit is even lower because of other frame types. Also,
the number of frames that could be used simultaneously depends on the total
amount of data buffered in all server streams, and client flow control.
The change sets frame threshold based on max concurrent streams and stream
buffer size, the product of which is the maximum number of in-flight stream
data in all server streams at any moment. The value is divided by 2000 to
account for a typical MTU 1500 and the fact that not all frames are STREAM
frames.
QUIC: ignore congestion control when sending MTU probes.
If connection is network-limited, MTU probes have little chance of being
sent since congestion window is almost always full. As a result, PMTUD
may not be able to reach the real MTU and the connection may operate with
a reduced MTU. The solution is to ignore the congestion window. This may
lead to a temporary increase in in-flight count beyond congestion window.
QUIC: do not shrink congestion window after losing an MTU probe.
As per RFC 9000, Section 14.4:
Loss of a QUIC packet that is carried in a PMTU probe is therefore
not a reliable indication of congestion and SHOULD NOT trigger a
congestion control reaction.
Previously, these functions operated on a per-level basis. This however
resulted in excessive logging of in_flight and will also led to extra
work detecting underutilized congestion window in the followup patches.
Roman Arutyunyan [Mon, 10 Mar 2025 08:19:25 +0000 (12:19 +0400)]
QUIC: ngx_msec_t overflow protection.
On some systems the value of ngx_current_msec is derived from monotonic
clock, for which the following is defined by POSIX:
For this clock, the value returned by clock_gettime() represents
the amount of time (in seconds and nanoseconds) since an unspecified
point in the past.
As as result, overflow protection is needed when comparing two ngx_msec_t.
The change adds such protection to the ngx_quic_detect_lost() function.
QUIC: prevent spurious congestion control recovery mode.
Since recovery_start field was initialized with ngx_current_msec, all
congestion events that happened within the same millisecond or cycle
iteration, were treated as in recovery mode.
Also, when handling persistent congestion, initializing recovery_start
with ngx_current_msec resulted in treating all sent packets as in recovery
mode, which violates RFC 9002, see example in Appendix B.8.
While here, also fixed recovery_start wrap protection. Previously it used
2 * max_idle_timeout time frame for all sent frames, which is not a
reliable protection since max_idle_timeout is unrelated to congestion
control. Now recovery_start <= now condition is enforced. Note that
recovery_start wrap is highly unlikely and can only occur on a
32-bit system if there are no congestion events for 24 days.
HTTP/3: graceful shutdown on keepalive timeout expiration.
Previously, the expiration caused QUIC connection finalization even if
there are application-terminated streams finishing sending data. Such
finalization terminated these streams.
An easy way to trigger this is to request a large file from HTTP/3 over
a small MTU. In this case keepalive timeout expiration may abruptly
terminate the request stream.
Improved logging for simpler data extraction for plotting congestion
window graphs. In particular, added current milliseconds number from
ngx_current_msec.
While here, simplified logging text and removed irrelevant data.
SSL: external groups support in $ssl_curve and $ssl_curves.
Starting with OpenSSL 3.0, groups may be added externally with pluggable
KEM providers. Using SSL_get_negotiated_group(), which makes lookup in a
static table with known groups, doesn't allow to list such groups by names
leaving them in hex. Adding X25519MLKEM768 to the default group list in
OpenSSL 3.5 made this problem more visible. SSL_get0_group_name() and,
apparently, SSL_group_to_name() allow to resolve such provider-implemented
groups, which is also "generally preferred" over SSL_get_negotiated_group()
as documented in OpenSSL git commit 93d4f6133f.
This change makes external groups listing by name using SSL_group_to_name()
available since OpenSSL 3.0. To preserve "prime256v1" naming for the group
0x0017, and to avoid breaking BoringSSL and older OpenSSL versions support,
it is used supplementary for a group that appears to be unknown.
See https://github.com/openssl/openssl/issues/27137 for related discussion.
The fix is to always preserve passwords, by copying to the configuration
pool, if dynamic certificates are used. This is done as part of merging
"ssl_passwords" configuration.
To minimize the number of copies, a preserved version is then used for
inheritance. A notable exception is inheritance of preserved empty
passwords to the context with statically configured certificates:
server {
proxy_ssl_certificate $ssl_server_name.crt;
proxy_ssl_certificate_key $ssl_server_name.key;
Sergey Kandaurov [Thu, 27 Feb 2025 14:42:06 +0000 (18:42 +0400)]
Charset filter: improved validation of charset_map with utf-8.
It was possible to write outside of the buffer used to keep UTF-8
decoded values when parsing conversion table configuration.
Since this happened before UTF-8 decoding, the fix is to check in
advance if character codes are of more than 3-byte sequence. Note
that this is already enforced by a later check for ngx_utf8_decode()
decoded values for 0xffff, which corresponds to the maximum value
encoded as a valid 3-byte sequence, so the fix does not affect the
valid values.
Found with AddressSanitizer.
Fixes GitHub issue #529.
Sergey Kandaurov [Thu, 27 Feb 2025 12:09:50 +0000 (16:09 +0400)]
Slice filter: improved memory allocation error handling.
As uncovered by recent addition in slice.t, a partially initialized
context, coupled with HTTP 206 response from stub backend, might be
accessed in the next slice subrequest.
Sergey Kandaurov [Tue, 25 Feb 2025 15:50:44 +0000 (19:50 +0400)]
SSL: raised limit for sessions stored in shared memory.
Upstream SSL sessions may be of a noticeably larger size with tickets
in TLSv1.2 and older versions, or with "stateless" tickets in TLSv1.3,
if a client certificate is saved into the session. Further, certain
stateless session resumption implemetations may store additional data.
Such one is JDK, known to also include server certificates in session
ticket data, which roughly doubles a decoded session size to slightly
beyond the previous limit. While it's believed to be an issue on the
JDK side, this change allows to save such sessions.
Another, innocent case is using RSA certificates with 8192 key size.
Sergey Kandaurov [Tue, 11 Feb 2025 18:54:04 +0000 (22:54 +0400)]
Improved ngx_http_subrequest() error handling.
Previously, request might be left in inconsistent state in case of error,
which manifested in "http request count is zero" alerts when used by SSI
filter.
The fix is to reshuffle initialization order to postpone committing state
changes until after any potentially failing parts.
Sergey Kandaurov [Wed, 22 Jan 2025 14:55:44 +0000 (18:55 +0400)]
SNI: added restriction for TLSv1.3 cross-SNI session resumption.
In OpenSSL, session resumption always happens in the default SSL context,
prior to invoking the SNI callback. Further, unlike in TLSv1.2 and older
protocols, SSL_get_servername() returns values received in the resumption
handshake, which may be different from the value in the initial handshake.
Notably, this makes the restriction added in b720f650b insufficient for
sessions resumed with different SNI server name.
Considering the example from b720f650b, previously, a client was able to
request example.org by presenting a certificate for example.org, then to
resume and request example.com.
The fix is to reject handshakes resumed with a different server name, if
verification of client certificates is enabled in a corresponding server
configuration.
Roman Arutyunyan [Wed, 15 Jan 2025 08:42:39 +0000 (12:42 +0400)]
Added "keepalive_min_timeout" directive.
The directive sets a timeout during which a keepalive connection will
not be closed by nginx for connection reuse or graceful shutdown.
The change allows clients that send multiple requests over the same
connection without delay or with a small delay between them, to avoid
receiving a TCP RST in response to one of them. This excludes network
issues and non-graceful shutdown. As a side-effect, it also addresses
the TCP reset problem described in RFC 9112, Section 9.6, when the last
sent HTTP response could be damaged by a followup TCP RST. It is important
for non-idempotent requests, which cannot be retried by client.
It is not recommended to set keepalive_min_timeout to large values as
this can introduce an additional delay during graceful shutdown and may
restrict nginx from effective connection reuse.
Sergey Kandaurov [Fri, 17 Jan 2025 13:55:21 +0000 (17:55 +0400)]
Configure: fixed --with-libatomic=DIR with recent libatomic_ops.
The build location of the resulting libatomic_ops.a was changed in v7.4.0
after converting libatomic_ops to use libtool. The fix is to use library
from the install path, this allows building with both old and new versions.
SSL: avoid using mismatched certificate/key cached pairs.
This can happen with certificates and certificate keys specified
with variables due to partial cache update in various scenarios:
- cache expiration with only one element of pair evicted
- on-disk update with non-cacheable encrypted keys
- non-atomic on-disk update
The fix is to retry with fresh data on X509_R_KEY_VALUES_MISMATCH.
Sergey Kandaurov [Tue, 29 Oct 2024 12:25:11 +0000 (16:25 +0400)]
SSL: caching certificates and certificate keys with variables.
A new directive "ssl_certificate_cache max=N [valid=time] [inactive=time]"
enables caching of SSL certificate chain and secret key objects specified
by "ssl_certificate" and "ssl_certificate_key" directives with variables.
Co-authored-by: Aleksei Bavshin <a.bavshin@nginx.com>
Sergey Kandaurov [Wed, 18 Dec 2024 16:09:58 +0000 (20:09 +0400)]
SSL: encrypted certificate keys are exempt from object cache.
SSL object cache, as previously introduced in 1.27.2, did not take
into account encrypted certificate keys that might be unexpectedly
fetched from the cache regardless of the matching passphrase. To
avoid this, caching of encrypted certificate keys is now disabled
based on the passphrase callback invocation.
A notable exception is encrypted certificate keys configured without
ssl_password_file. They are loaded once resulting in the passphrase
prompt on startup and reused in other contexts as applicable.
Sergey Kandaurov [Wed, 18 Dec 2024 16:03:35 +0000 (20:03 +0400)]
SSL: object cache inheritance from the old configuration cycle.
Memory based objects are always inherited, engine based objects are
never inherited to adhere the volatile nature of engines, file based
objects are inherited subject to modification time and file index.
The previous behaviour to bypass cache from the old configuration cycle
is preserved with a new directive "ssl_object_cache_inheritable off;".
Sergey Kandaurov [Mon, 23 Dec 2024 13:57:45 +0000 (17:57 +0400)]
Gzip: compatibility with recent zlib-ng 2.2.x versions.
It now uses 5/4 times more memory for the pending buffer.
Further, a single allocation is now used, which takes additional 56 bytes
for deflate_allocs in 64-bit mode aligned to 16, to store sub-allocation
pointers, and the total allocation size now padded up to 128 bytes, which
takes theoretically 200 additional bytes in total. This fits though into
"4 * (64 + sizeof(void*))" additional space for ZALLOC used in zlib-ng
2.1.x versions. The comment was updated to reflect this.
Roman Arutyunyan [Tue, 10 Dec 2024 14:19:27 +0000 (18:19 +0400)]
QUIC: fixed accessing a released stream.
While trying to close a stream in ngx_quic_close_streams() by calling its
read event handler, the next stream saved prior to that could be destroyed
recursively. This caused a segfault while trying to access the next stream.
The way the next stream could be destroyed in HTTP/3 is the following.
A request stream read event handler ngx_http_request_handler() could
end up calling ngx_http_v3_send_cancel_stream() to report a cancelled
request stream in the decoder stream. If sending stream cancellation
decoder instruction fails for any reason, and the decoder stream is the
next in order after the request stream, the issue is triggered.
The fix is to postpone calling read event handlers for all streams being
closed to avoid closing a released stream.
Roman Arutyunyan [Fri, 13 Dec 2024 09:25:26 +0000 (13:25 +0400)]
QUIC: ignore version negotiation packets.
Previously, such packets were treated as long header packets with unknown
version 0, and a version negotiation packet was sent in response. This
could be used to set up an infinite traffic reflect loop with another nginx
instance.
Now version negotiation packets are ignored. As per RFC 9000, Section 6.1:
An endpoint MUST NOT send a Version Negotiation packet in response to
receiving a Version Negotiation packet.
nandsky [Mon, 25 Nov 2024 07:26:29 +0000 (15:26 +0800)]
QUIC: fixed client request timeout in 0-RTT scenarios.
Since 0-RTT and 1-RTT data exist in the same packet number space,
ngx_quic_discard_ctx incorrectly discards 1-RTT packets when
0-RTT keys are discarded.
Sergey Kandaurov [Wed, 23 Oct 2024 20:52:21 +0000 (00:52 +0400)]
Mail: handling of LOGIN IMAP command untagged response.
In particular, an untagged CAPABILITY response as described in the
interim RFC 3501 internet drafts was seen in various IMAP servers.
Previously resulted in a broken connection, now an untagged response
is proxied to client.
Roman Arutyunyan [Mon, 11 Nov 2024 18:28:30 +0000 (22:28 +0400)]
Realip: allowed square brackets with portless IPv6 address.
When client address is received, IPv6 address could be specified without
square brackets and without port, as well as both with the brackets and
port. The change allows IPv6 in square brackets and no port, which was
previously considered an error. This format conforms to RFC 3986.
Sergey Kandaurov [Thu, 21 Nov 2024 08:35:50 +0000 (12:35 +0400)]
Upstream: disallow empty path in proxy_store and friends.
Renaming a temporary file to an empty path ("") returns NGX_ENOPATH
with a subsequent ngx_create_full_path() to create the full path.
This function skips initial bytes as part of path separator lookup,
which causes out of bounds access on short strings.
The fix is to avoid renaming a temporary file to an obviously invalid
path, as well as explicitly forbid such syntax for literal values.
Although Coverity reports about potential type underflow, it is not
actually possible because the terminating '\0' is always included.
Notably, the run-time check is sufficient enough for Win32 as well.
Other short invalid values result either in NGX_ENOENT or NGX_EEXIST
and "MoveFile() .. failed" critical log messages, which involves a
separate error handling.
Roman Arutyunyan [Tue, 22 Oct 2024 14:34:13 +0000 (18:34 +0400)]
Mp4: prevent chunk index underflow.
When cropping stsc atom, it's assumed that chunk index is never 0.
Based on this assumption, start_chunk and end_chunk are calculated
by subtracting 1 from it. If chunk index is zero, start_chunk or
end_chunk may underflow, which will later trigger
"start/end time is out mp4 stco chunks" error. The change adds an
explicit check for zero chunk index to avoid underflow and report
a proper error.
Zero chunk index is explicitly banned in ISO/IEC 14496-12, 8.7.4
Sample To Chunk Box. It's also implicitly banned in QuickTime File
Format specification. Description of chunk offset table references
"Chunk 1" as the first table element.
Mp4: unordered stsc chunks error for the final chunk.
Currently an error is triggered if any of the chunk runs in stsc are
unordered. This however does not include the final chunk run, which
ends with trak->chunks + 1. The previous chunk index can be larger
leading to a 32-bit overflow. This could allow to skip the validity
check "if (start_sample > n)". This could later lead to a large
trak->start_chunk/trak->end_chunk, which would be caught later in
ngx_http_mp4_update_stco_atom() or ngx_http_mp4_update_co64_atom().
While there are no implications of the validity check being avoided,
the change still adds a check to ensure the final chunk run is ordered,
to produce a meaningful error and avoid a potential integer overflow.
Roman Arutyunyan [Mon, 23 Sep 2024 11:51:30 +0000 (15:51 +0400)]
Mp4: fixed handling an empty run of chunks in stsc atom.
A specially crafted mp4 file with an empty run of chunks in the stsc atom
and a large value for samples per chunk for that run, combined with a
specially crafted request, allowed to store that large value in prev_samples
and later in trak->end_chunk_samples while in ngx_http_mp4_crop_stsc_data().
Later in ngx_http_mp4_update_stsz_atom() this could result in buffer
overread while calculating trak->end_chunk_samples_size.
Now the value of samples per chunk specified for an empty run is ignored.
Dan Callahan [Fri, 6 Sep 2024 15:57:50 +0000 (16:57 +0100)]
Fixed link to contributing guidelines.
Absolute paths in links end up being rooted at github.com.
The contributing guidelines link is broken unless we use the full URL.
Also, remove superfluous "monospace formatting" for the link.
Mini Hawthorne [Wed, 12 Jul 2023 19:20:45 +0000 (12:20 -0700)]
Upstream: copy upstream zone DNS valid time during config reload.
Previously, all upstream DNS entries would be immediately re-resolved
on config reload. With a large number of upstreams, this creates
a spike of DNS resolution requests. These spikes can overwhelm the
DNS server or cause drops on the network.
This patch retains the TTL of previous resolutions across reloads
by copying each upstream's name's expiry time across configuration
cycles. As a result, no additional resolutions are needed.
Ruslan Ermilov [Fri, 3 Nov 2017 19:22:23 +0000 (22:22 +0300)]
Upstream: pre-resolve servers on reload.
After configuration is reloaded, it may take some time for the
re-resolvable upstream servers to resolve and become available
as peers. During this time, client requests might get dropped.
Such servers are now pre-resolved using the "cache" of already
resolved peers from the old shared memory zone.
Ruslan Ermilov [Sat, 15 Feb 2014 11:12:34 +0000 (15:12 +0400)]
Upstream: re-resolvable servers.
Specifying the upstream server by a hostname together with the
"resolve" parameter will make the hostname to be periodically
resolved, and upstream servers added/removed as necessary.
This requires a "resolver" at the "http" configuration block.
The "resolver_timeout" parameter also affects when the failed
DNS requests will be attempted again. Responses with NXDOMAIN
will be attempted again in 10 seconds.
Upstream has a configuration generation number that is incremented each
time servers are added/removed to the primary/backup list. This number
is remembered by the peer.init method, and if peer.get detects a change
in configuration, it returns NGX_BUSY.
Each server has a reference counter. It is incremented by peer.get and
decremented by peer.free. When a server is removed, it is removed from
the list of servers and is marked as "zombie". The memory allocated by
a zombie peer is freed only when its reference count becomes zero.
Co-authored-by: Roman Arutyunyan <arut@nginx.com> Co-authored-by: Sergey Kandaurov <pluknet@nginx.com> Co-authored-by: Vladimir Homutov <vl@nginx.com>
TLSv1 and TLSv1.1 are formally deprecated and forbidden to negotiate due
to insufficient security reasons outlined in RFC 8996.
TLSv1 and TLSv1.1 are disabled in BoringSSL e95b0cad9 and LibreSSL 3.8.1
in the way they cannot be enabled in nginx configuration. In OpenSSL 3.0,
they are only permitted at security level 0 (disabled by default).
The support is dropped in Chrome 84, Firefox 78, and deprecated in Safari.
This change disables TLSv1 and TLSv1.1 by default for OpenSSL 1.0.1 and
newer, where TLSv1.2 support is available. For older library versions,
which do not have alternatives, these protocol versions remain enabled.
Since a2a513b93cae, stream frames no longer need to be retransmitted after it
was deleted. The frames which were retransmitted before, could be stream data
frames sent prior to a RESET_STREAM. Such retransmissions are explicitly
prohibited by RFC 9000, Section 19.4.