vmd(8): drop packets if tap(4) returns ENOMEM.
If there are no mbufs available to queue a packet on write(2), we
should drop the packet instead of toggling the device reset state.
The virtio network device was inspecting the wrong errno as well.
Also drop packets if EMSGSIZE is set on error. Only in unexpected
cases should we flag the device needs a reset.
Issue reported by Mischa Peters.
ok bluhm@, jan@, mlarkin@ and makes sense to deraadt@
test: improve output log and report format
Bring the test framework output closer to the output of Go's testing
package, in order to make it easier to trace issues in the output from
very large tests with many sub-tests.
Also add test_skipnow(), and make test_skip/test_skipf wrappers around
test_logf_internal so skip reasons get the same call location prefix as
other log output.
This also makes it easier for a followup to add timing for tests, which
will help determine which tests are slow.
ok tb
vmd(8): place acpi tables in upper reserved memory
The original decision to place ACPI tables at 0x90000 caused "legacy" linux
bootloaders like lilo to overwrite the tables. While this doesn't damage
the vm, it does cause linux to revert to uniprocessor mode. Place the tables
in a mem reserved region adjacent to the pci mmio hole instead.
tested on various guest OSes
ok dv
Add TLSv1.2 error handling
Provide error handling for use with the TLSv1.2 context - for now this
is a duplicate of the TLSv1.3 error handling, which we may deduplicate
in the future.
ok tb@
Add TLSv1.2 handshake message handling.
Provide struct tls12_handshake_msg and wire it to the new TLSv1.2
stack, allowing handshake messages to be built and parsed.
ok beck@ tb@
Add a TLSv1.2 state machine and skeleton client/server.
This provides the basis for a new TLSv1.2 stack and is heavily based on
our TLSv1.3 implementation. All TLSv1.2 handshake messages are included,
however the state table only includes a minimal TLS connection - additional
state transitions will be added in due course.
This code is not currently used and further development will occur in tree.
ok beck@ tb@
On some platforms, e.g. Linux Clang 22.1.8 / glibc 2.43, strchr(3)
now implements the C23 behaviour where passing a const pointer to
strchr(3) also returns a const pointer.
Solve this by disentangling the horribly complicated logic in
tblcalc_literal() that did multiple checks on "mw". Instead, handle
the case "maximum width -> parse into words and break lines" which
needs a mutable copy of the content separately from the case
"unlimited width" which works with the original const char*.
This makes the code much more readable. No functional change.
Small issue found by Lexi Winter <ivy at FreeBSD.org>,
reported via Alexander Ziaee <ziaee at freebsd.org>.
Get x and y the right way round, fixes crash when opening a mode prompt
when the client is offset into the window; may be the crash reported by
sthen@.
ixl(4): Fix hardmtu and rxmax
Table 1-3. Link layer Ethernet port features
Maximum Transmit Unit Size (MTU) is 9728 - Ethernet header/CRC =
9728 - 18 = 9710 bytes (jumbo frames)
MTU can be further reduced by additional header fields such as
Virtual Local Area Network (VLAN) tag(s), etc.
Table 8-12. LAN Rx Queue Context in the Private Host Memory
The "rxmax" parameter defines the whole packet size starting at
the L2 header up to including the Ethernet CRC.
...
The RXMAX must not be set to a larger value than 5 x DBUFF
(since receive packet must never span on more than 5 buffers).
Tested by bluhm@
ok jmatthew@
Some terminals have extended DECSCUSR to allow it to reset to the
default style and put that as the Se capability in their terminfo entry.
Change so that if Ss and Se exist, tmux does not overwrite them with its
own. GitHub issue 5476.
vmm(4): Skip VMCALL/VMMCALL at CPL 0; otherwise abort by signal
On Intel we have to use #GP(0), on AMD #UD. Adjust regression test.
With Miguel Landaeta (miguel at miguel cc), thanks!
go ahead mlarkin
drm/ttm: Fix bo resource use-after-free
From Thadeu Lima de Souza Cascardo
73867f2536704760dc42671eb508c1fdc7fd9b0b in linux-6.18.y/6.18.53
91d7e9df42598af28ca440b95b16a4e51a408771 in mainline linux
drm/ttm: Drop tt->restore after successful restore
From Matthew Brost
329ddc5d438f991f49e4fd24c704d1c7288d5e03 in linux-6.18.y/6.18.53
941ac10529b3be5965a88d432a161ab459672ba8 in mainline linux
Disable LZ77 dictionary coder to avoid a potential side-channel leak
A chosen-plaintext attack method exists which makes use of
dictionary-based compression to recover secrets from one channel by
interacting with the SSH session's shared compression dictionary through
another channel.
Attacker-controlled input can recognizably reflect into the total length
of transmitted ciphertexts by virtue of LZ77 replacing repeated strings
with back-references into the SSH session's encoder search buffer,
which is shared across all channels. For this reason, the documentation
already recommended against enabling compression for connections that
share trusted and untrusted traffic.
As an extra precaution, use only Huffman coding when compressing
plaintexts, as this algorithm does not use a dictionary. But, this
results in a reduction of compression effectiveness.
Inspired by "Crossing the Streams: SSH Plaintext Recovery via a Common
[4 lines not shown]
drm/ttm: Tidy usage of local variables a little bit
From Tvrtko Ursulin
b21c971e4cb11c0c7ea74261f966814280c02647 in linux-6.18.y/6.18.53
c06da4b3573a2d3c906c185450d0b1059d02820e in mainline linux
drm/amd/display: Honor Broadcast RGB for BT.2020 RGB output
From Satyajit Roy
037a00dfd5838d616bc04195b182055da1b33fe6 in linux-6.18.y/6.18.53
7fca7acd60a228b62b4e9efa5f184738041e9564 in mainline linux
drm/amd/display: Propagate HDMI RGB quantization selectability
From Satyajit Roy
6194e1b536dc624e05412507fb84c6c880e1f8b5 in linux-6.18.y/6.18.53
bdcd0411d7d186225a52458fd42bb70d54ca917a in mainline linux
drm/amd/display: Rebuild InfoFrames on output color space changes
From Satyajit Roy
925d9aa99a69ddf461f256c397590957d35dee8a in linux-6.18.y/6.18.53
8cfd9e22eb5c04b15b82985ff913944f84673d4f in mainline linux
