Previously, we enforced the requirement that the DCIDs be the same for
all packets in a datagram by keeping a pointer to the first RXE
generated from a datagram. This is unsafe and could lead to a UAF if the
first packet is malformed, meaning that no RXE ended up being generated
from it. Keep track of the DCID directly instead, as we should enforce
this correctly even if the first packet in a datagram is malformed (but
has an intelligible header with a DCID and length).
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
Server mode not implemented yet.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
The documentation in the header file of the TXP stated that it is the
caller's responsibility to also notify the QTX of a discarded EL.
However, the implementation did not reflect this. Update the
implementation to reflect the intended design.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
Ordinarily we should not allow ELs to be rekeyed as it makes no sense to
do so. However the INITIAL EL can need to be rekeyed if a connection
retry occurs. Modify the QRL to allow this.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
This adds support for calculating and verifying retry integrity tags. In
order to support this, an 'unused' field is added to the QUIC packet
header structure so we can ensure that the serialization of the header
is bit-for-bit identical to what was decoded.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
While the QUIC RFCs state that the Initial EL should be auto-discarded
when successfully processing a packet at a higher EL, doing this inside
the QRX was not a good idea as this should be handled by the CSM.
We remove this functionality and adapt tests accordingly.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
Previously, the QRX filled in a OSSL_QRX_PKT structure provided by the
caller. This necessitated the caller managing reference counting itself
using a OSSL_QRX_PKT_WRAP structure. The need for this structure has
been eliminated by adding refcounting support to the QRX itself. The QRX
now outputs a pointer to an OSSL_QRX_PKT instead of filling in a
structure provided by the caller. The OSSL_QRX_PKT_WRAP structure has
been eliminated.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
This disables -Wtype-limits /
-Wtautological-constant-out-of-range-compare. Since it generates
warnings for valid and reasonable code, IMO this actually encourages
people to write worse code.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
This is required to support retries during connection establishment.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19703)
The internal error reason is confusing and indicating an error
in OpenSSL and not a configuration problem.
Fixes#19867
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19875)
TLS device offload allows to perform zerocopy sendfile transmissions.
FreeBSD provides this feature by default, and Linux 5.19 introduced it
as an opt-in. Zerocopy improves the TX rate significantly, but has a
side effect: if the underlying file is changed while being transmitted,
and a TCP retransmission happens, the receiver may get a TLS record
containing both new and old data, which leads to an authentication
failure and termination of connection. This effect is the reason Linux
makes a copy on sendfile by default.
This commit adds support for TLS zerocopy sendfile on Linux disabled by
default to avoid any unlikely backward compatibility issues on Linux,
although sacrificing consistency in OpenSSL's behavior on Linux and
FreeBSD. A new option called KTLSTxZerocopySendfile is added to enable
the new zerocopy behavior on Linux. This option should be used when the
the application guarantees that the file is not modified during
transmission, or it doesn't care about breaking the connection.
The related documentation is also added in this commit. The unit test
added doesn't test the actual functionality (it would require specific
hardware and a non-local peer), but solely checks that it's possible to
set the new option flag.
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Boris Pismenny <borisp@nvidia.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/18650)
