Page MenuHomeFreeBSD
Differential D29275 Diff 87200 website/content/en/releases/13.0R/relnotes.adoc

Changeset View

Standalone View

View Options

website/content/en/releases/13.0R/relnotes.adoc

Show First 20 Lines • Show All 417 Lines • ▼ Show 20 Lines
If the man:efibootmgr[8] utility is used to customize the boot environment, this file should be copied to the location set with the `-l` flag. If the man:efibootmgr[8] utility is used to customize the boot environment, this file should be copied to the location set with the `-l` flag.
[[network]] [[network]]
== Networking == Networking
This section describes changes that affect networking in FreeBSD. This section describes changes that affect networking in FreeBSD.
[[network-general]] [[network-general]]
debdrupUnsubmitted
Done
Inline Actions
This section describes changes that affect networking in FreeBSD.
- FreeBSD 13 features rewritten routing stack. It is based on the {{< rev-named "introduction" "a666325282eaed4b044459d121f339b2d6d0224b >}} of nexthops - objects holding all necessary state to pass a packet to the desired destination.
+ FreeBSD 13 features a rewritten routing stack. It is based on the {{< rev-named "introduction" "a666325282eaed4b044459d121f339b2d6d0224b >}} of nexthops - objects holding all necessary state to pass a packet to the desired destination.
Multipath routing support has been {{< rev-named "rewritten" "fedeb08b6a58e708e1153224d37ad26bdc1062a2" >}} in more scalable fashion, featuring 64-wide multipath routes with O(1) lookup time.
debdrup:
=== General Network === General Network
debdrupUnsubmitted
Done
Inline Actions
Multipath routing support has been {{< rev-named "rewritten" "fedeb08b6a58e708e1153224d37ad26bdc1062a2" >}} in more scalable fashion, featuring 64-wide multipath routes with O(1) lookup time.
- `RADIX_MPATH` kernel option got {{< rev-named replaced" "d1d941c5b910e075495cb06b92a99d3a3e7a3d6c >}} by the `ROUTE_MPATH`, which is turned {{< rev-named "on" "d5fe384b4d41a5c5adeaa6039d33d59d7bc33e76" }} by default. Additional `net.route.multipath` sysctl has been added to control the feature in runtime.
+ The `RADIX_MPATH` kernel option got {{< rev-named "replaced" "d1d941c5b910e075495cb06b92a99d3a3e7a3d6c >}} with `ROUTE_MPATH`, which is {{< rev-named "turned on" "d5fe384b4d41a5c5adeaa6039d33d59d7bc33e76" }} by default. Additionally, the `net.route.multipath` sysctl OID namespace has been added to control the feature at runtime.
Support for the custom route lookups algorithms has been {{< rev-named "added" "f5baf8bb12f39d0e8d64508c47eb6c4386ef716d" >}}. The framework decouples control plane and forwarding state, resulting in both
debdrup:
A new type of man:mbuf[9] (network data buffer) can represent multiple, A new type of man:mbuf[9] (network data buffer) can represent multiple,
debdrupUnsubmitted
Done
Inline Actions
`RADIX_MPATH` kernel option got {{< rev-named replaced" "d1d941c5b910e075495cb06b92a99d3a3e7a3d6c >}} by the `ROUTE_MPATH`, which is turned {{< rev-named "on" "d5fe384b4d41a5c5adeaa6039d33d59d7bc33e76" }} by default. Additional `net.route.multipath` sysctl has been added to control the feature in runtime.
- Support for the custom route lookups algorithms has been {{< rev-named "added" "f5baf8bb12f39d0e8d64508c47eb6c4386ef716d" >}}. The framework decouples control plane and forwarding state, resulting in both
+ Support for custom route lookup algorithms has been {{< rev-named "added" "f5baf8bb12f39d0e8d64508c47eb6c4386ef716d" >}}. The framework decouples control-plane and data-plane, resulting in both
faster lookups and much better convergence times for the large tables.
debdrup:
unmapped physical pages as a single buffer. unmapped physical pages as a single buffer.
This improves the performance of man:sendfile[2] by reducing the length of This improves the performance of man:sendfile[2] by reducing the length of
debdrupUnsubmitted
Done
Inline Actions
faster lookups and much better convergence times for the large tables.
- DPDK librte-based IPv4/IPv6 route lookup algorithms has been {{< rev-named "added" "537d134373141c2d25bfb24af6d661d0e6102927" >}}, optimising dataplane and control plane for the large routing tables.
+ DPDK librte-based IPv4/IPv6 route lookup algorithms has been {{< rev-named "added" "537d134373141c2d25bfb24af6d661d0e6102927" >}}, optimising control-plane and data-plane for large routing tables.
For {{< manpage "tcp" "4">}}, Proportional Rate Reduction, as described by RFC6937, to improve SACK loss recovery during burst loss and ACK thinning scenarios, was implemented and is enabled by default.
debdrup:
mbuf linked lists in socket buffers. gitref:82334850ea45[repository=src], gitref:cec06a3edc52[repository=src] {{< sponsored "Netflix" >}} mbuf linked lists in socket buffers. gitref:82334850ea45[repository=src], gitref:cec06a3edc52[repository=src] {{< sponsored "Netflix" >}}
The kernel now supports in-kernel framing and encryption of Transport The kernel now supports in-kernel framing and encryption of Transport
Layer Security (TLS) data on TCP sockets for TLS versions 1.0 through 1.3. Layer Security (TLS) data on TCP sockets for TLS versions 1.0 through 1.3.
Transmit offload via in-kernel crypto drivers is supported for Transmit offload via in-kernel crypto drivers is supported for
MtE cipher suites using AES-CBC as well as AEAD cipher suites using MtE cipher suites using AES-CBC as well as AEAD cipher suites using
AES-GCM. AES-GCM.
Receive offload via in-kernel crypto drivers is supported for AES-GCM Receive offload via in-kernel crypto drivers is supported for AES-GCM
cipher suites for TLS 1.2. cipher suites for TLS 1.2.
Using KTLS requires the use of a KTLS-aware userland SSL library. Using KTLS requires the use of a KTLS-aware userland SSL library.
The OpenSSL library included in the base system does not enable KTLS The OpenSSL library included in the base system does not enable KTLS
support by default, support by default,
but support can be enabled by building with the `WITH_OPENSSL_KTLS` option. but support can be enabled by building with the `WITH_OPENSSL_KTLS` option.
gitref:b2e60773c6b0[repository=src], gitref:6554362c6640[repository=src], gitref:f1f934754638[repository=src], gitref:3c0e56850511[repository=src], gitref:c1c52cd57e88[repository=src] {{< sponsored "Netflix" >}} {{< sponsored "Chelsio Communications" >}} gitref:b2e60773c6b0[repository=src], gitref:6554362c6640[repository=src], gitref:f1f934754638[repository=src], gitref:3c0e56850511[repository=src], gitref:c1c52cd57e88[repository=src] {{< sponsored "Netflix" >}} {{< sponsored "Chelsio Communications" >}}
man:tcp[4] now supports Proportional Rate Reduction (as described by RFC6937) to improve SACK loss recovery during burst loss and ACK thinning scenarios. man:tcp[4] now supports Proportional Rate Reduction (as described by RFC6937) to improve SACK loss recovery during burst loss and ACK thinning scenarios.
debdrupUnsubmitted
Done
Inline Actions
The {{< manpage "ping" "8" >}} and {{< manpage "ping6" "8" >}} utilities both support setting network QoS, with IP DSCP gitref:6034024daddb[repository=src] and Ethernet PCP gitref:81a6f4c7ae69[repository=src] {{< sponsored "NetApp" >}}
- Interface fib is now {{< rev-named "used" "66bc03d415666a5b0e8566a2d05649f13bf07ebe" >}} for the proxyarp checks.
+ Interface fib is now {{< rev-named "used" "66bc03d415666a5b0e8566a2d05649f13bf07ebe" >}} for proxyarp checks.
Loopback route installation has been {{< rev-named "fix" "9fdbf7eef5c006002769add15b1ebb8fa8d9e220" >}} for the interfaces in the different fibs using the same prefix.
debdrup:
This feature is enabled by default. This feature is enabled by default.
debdrupUnsubmitted
Done
Inline Actions
Interface fib is now {{< rev-named "used" "66bc03d415666a5b0e8566a2d05649f13bf07ebe" >}} for the proxyarp checks.
- Loopback route installation has been {{< rev-named "fix" "9fdbf7eef5c006002769add15b1ebb8fa8d9e220" >}} for the interfaces in the different fibs using the same prefix.
+ Loopback route installation has been {{< rev-named "fixed" "9fdbf7eef5c006002769add15b1ebb8fa8d9e220" >}} for the interfaces in different fibs using the same prefix.
Number of fibs can now be changed in {{< rev-named "runtime" "f5247a232a337c93c690068930f25503d4dab8e1" >>} by controlling `net.fibs` sysctl.
debdrup:
A new man:sysctl[8], `net.inet.tcp.do_prr`, can be set to `0` to restore the prior behavior. A new man:sysctl[8], `net.inet.tcp.do_prr`, can be set to `0` to restore the prior behavior.
debdrupUnsubmitted
Done
Inline Actions
Loopback route installation has been {{< rev-named "fix" "9fdbf7eef5c006002769add15b1ebb8fa8d9e220" >}} for the interfaces in the different fibs using the same prefix.
- Number of fibs can now be changed in {{< rev-named "runtime" "f5247a232a337c93c690068930f25503d4dab8e1" >>} by controlling `net.fibs` sysctl.
+ Number of fibs can now be changed at {{< rev-named "runtime" "f5247a232a337c93c690068930f25503d4dab8e1" >>} by controlling `net.fibs` sysctl.
`net.add_addr_allfibs` sysctl default has been {{< "changed" "2d39824195933c173bbfc9b31773070202d2e30e" >}} to 0.
debdrup:
PRR should generally help improve loss recovery performance and prevent numerous preventable retransmit timeout (RTO) stalls. PRR should generally help improve loss recovery performance and prevent numerous preventable retransmit timeout (RTO) stalls.
debdrupUnsubmitted
Not Done
Inline Actions

Is there a line too many here?

debdrup: Is there a line too many here?
This surpasses the prior behavior, but a strictly packet conserving variant can be enabled. This surpasses the prior behavior, but a strictly packet conserving variant can be enabled.
A misconfigured token bucket traffic policer can cause persistent loss even during loss recovery. A misconfigured token bucket traffic policer can cause persistent loss even during loss recovery.
debdrupUnsubmitted
Done
Inline Actions
`net.add_addr_allfibs` sysctl default has been {{< "changed" "2d39824195933c173bbfc9b31773070202d2e30e" >}} to 0.
- Tha amount of queued packets in for unresolved ARP/NDP entries has been {{< rev-named "increased" "0da3f8c98d17d9c6e67d0f6cf45085df6a5ef041" >}} to 16.
+ The amount of queued packets in for unresolved ARP/NDP entries has been {{< rev-named "increased" "0da3f8c98d17d9c6e67d0f6cf45085df6a5ef041" >}} to 16.
Temporal routes (routes with `-expire` time set) expiration has been {{< rev-named "fixed" "34a5582c47c7ec24a6bf7de8af32b1334a79f58e" >}} for both IPv4 and IPv6.
debdrup:
In that case, activating the conservative PRR variant may prevent some retransmission timeouts (RTO) and associated session stalls for a few milliseconds while behaving less optimal in the general case. In that case, activating the conservative PRR variant may prevent some retransmission timeouts (RTO) and associated session stalls for a few milliseconds while behaving less optimal in the general case.
debdrupUnsubmitted
Done
Inline Actions
Tha amount of queued packets in for unresolved ARP/NDP entries has been {{< rev-named "increased" "0da3f8c98d17d9c6e67d0f6cf45085df6a5ef041" >}} to 16.
- Temporal routes (routes with `-expire` time set) expiration has been {{< rev-named "fixed" "34a5582c47c7ec24a6bf7de8af32b1334a79f58e" >}} for both IPv4 and IPv6.
+ Temporal routes (routes with `-expire` time set) expiration have been {{< rev-named "fixed" "34a5582c47c7ec24a6bf7de8af32b1334a79f58e" >}} for both IPv4 and IPv6.
Duplicate routes installation issue for /32 or /128 interface aliases has been {{< rev-named "fixed" "81728a538d24f483d0986850fa3f51d5d84d8f26".
debdrup:
A new man:sysctl[8], `net.inet.tcp.do_prr_conservative`, A new man:sysctl[8], `net.inet.tcp.do_prr_conservative`,
debdrupUnsubmitted
Done
Inline Actions
Temporal routes (routes with `-expire` time set) expiration has been {{< rev-named "fixed" "34a5582c47c7ec24a6bf7de8af32b1334a79f58e" >}} for both IPv4 and IPv6.
- Duplicate routes installation issue for /32 or /128 interface aliases has been {{< rev-named "fixed" "81728a538d24f483d0986850fa3f51d5d84d8f26".
+ Duplicate routes installation issue for /32 or /128 interface aliases has been {{< rev-named "fixed" "81728a538d24f483d0986850fa3f51d5d84d8f26" >}}.
IPv6 interface routes are now {{< rev-named "marked" "81728a538d24f483d0986850fa3f51d5d84d8f26" >}} with RTF_PINNED as their IPv4 counterparts.
debdrup:
can be set to `1` to enable strictly packet conserving behavior (at most 1 segment for each ACK received), while the normal variant may send up to 2 segments per received ACK - helping in cases of ACK thinning or significant burst loss events. gitref:0e1d7c25c5ab[repository=src] {{< sponsored "NetApp" >}} can be set to `1` to enable strictly packet conserving behavior (at most 1 segment for each ACK received), while the normal variant may send up to 2 segments per received ACK - helping in cases of ACK thinning or significant burst loss events. gitref:0e1d7c25c5ab[repository=src] {{< sponsored "NetApp" >}}
debdrupUnsubmitted
Done
Inline Actions
Duplicate routes installation issue for /32 or /128 interface aliases has been {{< rev-named "fixed" "81728a538d24f483d0986850fa3f51d5d84d8f26".
- IPv6 interface routes are now {{< rev-named "marked" "81728a538d24f483d0986850fa3f51d5d84d8f26" >}} with RTF_PINNED as their IPv4 counterparts.
+ IPv6 interface routes are now {{< rev-named "marked" "81728a538d24f483d0986850fa3f51d5d84d8f26" >}} with RTF_PINNED like their IPv4 counterparts.
The {{< manpage "route" "8">}} network auto-guessing has been {{< rev-named eliminated" "d28210b2c2aaf3200907ed30d296b0d4856dd03c" >}} by removing remnants of classful behavior.
debdrup:
The man:cc_cubic[4] man:tcp[4] congestion control algorithm aligns more closely with the standard in RFC8312. gitref:40f9078ff9d9[repository=src] {{< sponsored "NetApp" >}} The man:cc_cubic[4] man:tcp[4] congestion control algorithm aligns more closely with the standard in RFC8312. gitref:40f9078ff9d9[repository=src] {{< sponsored "NetApp" >}}
debdrupUnsubmitted
Done
Inline Actions
The {{< manpage "route" "8">}} network auto-guessing has been {{< rev-named eliminated" "d28210b2c2aaf3200907ed30d296b0d4856dd03c" >}} by removing remnants of classful behavior.
- Sysctl `net.inet6.ip6.deembed_scopeid` , allowing to disable IPv6 scope deembedding, has been {{< "removed" "bec053ffe01ddc571fb09d1758708c407c33ddfd" >}}.
+ Sysctl `net.inet6.ip6.deembed_scopeid`, making it possible to disable IPv6 scope de-embedding, has been {{< "removed" "bec053ffe01ddc571fb09d1758708c407c33ddfd" >}}.
Stacked vlans (802.1ad) support has been {{< "added" "c7cffd65c5d858425e90b847d2e8e583e3b13bf7" >}}.
debdrup:
The amount of queued packets in for unresolved ARP/NDP entries has been increased to 16. gitref:0da3f8c98d17d9[repository=src]
Stacked VLAN (802.1ad) support has been added: gitref:c7cffd65c5d8[repository=src].
melifaroAuthorUnsubmitted
Done
Inline Actions

Q: what is the proper way to reference the person who contributed this change?

melifaro: Q: what is the proper way to reference the person who contributed this change?
debdrupUnsubmitted
Not Done
Inline Actions

The authorship is noted in that commit, so unless they explicitly asked to be mentioned in the release notes, I'm not sure.

debdrup: The authorship is noted in that commit, so unless they explicitly asked to be mentioned in the…
debdrupUnsubmitted
Done
Inline Actions
Sysctl `net.inet6.ip6.deembed_scopeid` , allowing to disable IPv6 scope deembedding, has been {{< "removed" "bec053ffe01ddc571fb09d1758708c407c33ddfd" >}}.
- Stacked vlans (802.1ad) support has been {{< "added" "c7cffd65c5d858425e90b847d2e8e583e3b13bf7" >}}.
+ Stacked VLAN (802.1ad) support has been {{< "added" "c7cffd65c5d858425e90b847d2e8e583e3b13bf7" >}}.
[[network-general]]
debdrup:
debdrupUnsubmitted
Done
Inline Actions
The amount of queued packets in for unresolved ARP/NDP entries has been increased to 16. gitref:0da3f8c98d17d9[repository=src]
- Stacked VLAN (802.1ad) support has been added: gitref:c7cffd65c5d8[repository=src].
+ Stacked VLAN (802.1ad) support has been added. gitref:c7cffd65c5d8[repository=src].
The man:ping[8] utility now supports setting network QoS, with IP DSCP gitref:6034024daddb[repository=src] and Ethernet PCP gitref:81a6f4c7ae69[repository=src]. {{< sponsored "NetApp" >}}

The colon should be a period.

debdrup: The colon should be a period.
The man:ping[8] utility now supports setting network QoS, with IP DSCP gitref:6034024daddb[repository=src] and Ethernet PCP gitref:81a6f4c7ae69[repository=src]. {{< sponsored "NetApp" >}} The man:ping[8] utility now supports setting network QoS, with IP DSCP gitref:6034024daddb[repository=src] and Ethernet PCP gitref:81a6f4c7ae69[repository=src]. {{< sponsored "NetApp" >}}
Merged the man:ping[8] and man:ping6[8] utilities. Merged the man:ping[8] and man:ping6[8] utilities.
man:ping[8] supports both IPv4 and IPv6. man:ping[8] supports both IPv4 and IPv6.
A legacy man:ping6[8] is retained for backwards compatibility. gitref:3cde9171d2d5[repository=src] A legacy man:ping6[8] is retained for backwards compatibility. gitref:3cde9171d2d5[repository=src]
SCTP support is now available as a new [.filename]#sctp.ko# kernel module SCTP support is now available as a new [.filename]#sctp.ko# kernel module
and is no longer compiled into GENERIC by default. and is no longer compiled into GENERIC by default.
gitref:e64080e79c53[repository=src] {{< sponsored "The FreeBSD Foundation" >}} gitref:e64080e79c53[repository=src] {{< sponsored "The FreeBSD Foundation" >}}
[[[network-routing]]
=== Routing
FreeBSD 13 features a rewritten routing stack. It is based on the introduction of nexthops - objects holding all necessary state to pass a packet to the desired destination. gitref:a666325282ea[repository=src]
Multipath routing support has been rewritten in more scalable fashion, featuring 64-wide multipath routes with O(1) lookup time. gitref:fedeb08b6a58[repository=src].
The `RADIX_MPATH` kernel option got replaced with `ROUTE_MPATH`, which is turned on by default. Additionally, the `net.route.multipath` sysctl has been added to control the feature in runtime. gitref:d1d941c5b910[repository=src], gitref:d5fe384b4d41[repository=src]
Support for custom route lookup algorithms has been added. The framework decouples control-plane and data-plane, resulting in both faster lookups and better convergence times for large tables under load. gitref:f5baf8bb12f3[repository=src]
DPDK librte-based IPv4/IPv6 route lookup algorithms has been added, optimising control-plane and data-plane for large routing tables. gitref:537d13437314[repository=src]
Interface fib is now used for proxyarp checks. gitref:66bc03d41566[repository=src]
Loopback route installation has been fixed for the interfaces in different fibs using the same prefix. gitref:9fdbf7eef5c0[repository=src]
Number of fibs can now be changed at runtime by controlling `net.fibs` sysctl. gitref:f5247a232a33[repository=src]
`net.add_addr_allfibs` sysctl default has been changed to 0. gitref:2d3982419593[repository=src]
Temporal routes (routes with `-expire` time set) expiration have been for both IPv4 and IPv6. gitref:34a5582c47c7[repository=src]
Duplicate routes installation issue for /32 or /128 interface aliases has been fixed. gitref:81728a538d24[repository=src]
IPv6 interface routes are now marked with RTF_PINNED like their IPv4 counterparts. gitref:81728a538d24[repository=src]
The {{< manpage "route" "8">}} network auto-guessing has been eliminated by removing remnants of classful behavior. gitref:d28210b2c2aa[repository=src]
Sysctl `net.inet6.ip6.deembed_scopeid` , making it possible to disable IPv6 scope de-embedding, has been removed. gitref:bec053ffe01d[repository=src]
[[hardware]] [[hardware]]
== Hardware Support == Hardware Support
This section covers general hardware support for physical machines, hypervisors, and virtualization environments, as well as hardware changes and updates that do not otherwise fit in other sections of this document. This section covers general hardware support for physical machines, hypervisors, and virtualization environments, as well as hardware changes and updates that do not otherwise fit in other sections of this document.
=== AMD64 === AMD64
▲ Show 20 Lines • Show All 135 Lines • Show Last 20 Lines