igc: Apply ASPM L1.2 workaround to all I226 devices
Classify I226_LMVP and I226_BLANK_NVM as I226 silicon so they
receive the I226-specific ASPM L1.2 workaround.
PR: 279245
MFC after: 1 week
Pull-Request: https://github.com/freebsd/freebsd-src/pull/2318
igc: Apply ASPM L1.2 workaround to all I226 devices
Classify I226_LMVP and I226_BLANK_NVM as I226 silicon so they
receive the I226-specific ASPM L1.2 workaround.
PR: 279245
MFC after: 1 week
Pull-Request: https://github.com/freebsd/freebsd-src/pull/2318
[libc] Move generated elf header to include (#211428)
The fact that `hdr/elf_proxy.h` was generated with headergen was causing
dependency issues. This shouldn't be necessary, and this PR moves it out
into `include` to avoid needing headergen for an overlay build.
Assisted-by: Automated tooling, human reviewed.
Key the SNMP sysObjectID off hardware rather than the license
This commit adds changes to derive the sysObjectID suffix from the new hardware identity package instead of system.is_enterprise. sysObjectID is meant to tell a monitoring station what kind of box it is talking to, so a licensing predicate was never the right input - the same physical chassis was advertising a different identity depending on whether a license happened to be installed.
The emitted value changes for some systems: a licensed Mini or whitebox now reports 1 instead of 2, and an unlicensed single-controller iX box reports 2 instead of 1. Physical iX hardware and HA VMs are unaffected. That is observable to anything polling us over SNMP, since a station keys its polling template off this value, so it wants a release note.
Worth noting this file has no test coverage and etc_files is excluded from mypy, so CI says nothing about it. The template was compiled and rendered locally against each hardware class to confirm the polarity and that the OID prefix is unchanged.
nvmecontrol.8: Explain non-operational power modes
`nvmecontrol power -l ...` lists the available power modes.
Non-operational modes are marked with an asterisk. While here,
add <device-id | namespace-id> to the "nvmecontrol power" synopsis.
MFC after: 3 days
Reviewed by: dab, imp, michaelo, ziaee
Differential Revision: https://reviews.freebsd.org/D58480
nvmecontrol.8: Explain non-operational power modes
`nvmecontrol power -l ...` lists the available power modes.
Non-operational modes are marked with an asterisk. While here,
add <device-id | namespace-id> to the "nvmecontrol power" synopsis.
MFC after: 3 days
Reviewed by: dab, imp, michaelo, ziaee
Differential Revision: https://reviews.freebsd.org/D58480
Add a hardware identity package and route entitlements through it
This commit adds changes to pull hardware detection out of the entitlement engine into utils/hardware, split into a pure classifier that takes DMI values as arguments and a probe that owns the one impure step. Detecting what a machine is was never an entitlement concern, and keeping the classifier pure means it is testable without mocking anything.
HA virtual machines are now recognised. IXKVM comes out of the QEMU manufacturer plus the node serial convention, and bhyve is only treated as HA when the TrueNAS_A/TrueNAS_B backplane device is actually attached, so ordinary development bhyve VMs are not swept in. Both map to TrueNAS hardware, which moves them from the community to the hardware side of the feature matrix: an unlicensed HA VM loses dedup and gains SED, and a licensed HA VM without the fibre channel key gains fibre channel. That is a deliberate call - an HA VM stands in for an appliance and should be entitled as one.
The whole answer is cached for the life of the process since nothing it reads can change without a reboot, and the backplane scan only runs when DMI already says the guest is bhyve. The backplane probe reads sysfs rather than going through udev; the two resolve to the same file and sysfs does not need udevd running. The mypy workflow picks up the new package, without which it would not be type checked at all.
[CIR][AMDGPU] Add support for AMDGCN ldexp builtins (#198087)
Adds support for the following AMDGCN ldexp builtins:
- __builtin_amdgcn_ldexp (double)
- __builtin_amdgcn_ldexpf (float)
- __builtin_amdgcn_ldexph (half)
These are lowered to the `llvm.ldexp` intrinsic.
all: expire direct consumers of www/node20 on 2026-09-30
This happens to be the date when www/qt5-webengine expires, which
is also a node20 consumer (on i386, same for qt6-webengine).
all: expire direct consumers of www/node20 on 2026-09-30
This happens to be the date when www/qt5-webengine expires, which
is also a node20 consumer (on i386, same for qt6-webengine).
textproc/sigil: Add patch to workaround crash on close
Add a patch to workaround for issue:
https://github.com/Sigil-Ebook/Sigil/issues/874
Bump port revision.
PR: 296967
Reported-by: Jonathan Chen <jonc at chen.org.nz> (maintainer)
textproc/sigil: Add patch to workaround crash on close
Add a patch to workaround for issue:
https://github.com/Sigil-Ebook/Sigil/issues/874
Bump port revision.
PR: 296967
Reported-by: Jonathan Chen <jonc at chen.org.nz> (maintainer)
[DropAssumes] Print drop-deref in pipeline where appropriate (#211905)
drop-deref was added as an option in #166947 and is parsed correctly,
but is not serialized. This can make it difficult to reduce test cases,
especially automatically with llvm/utils/reduce_pipeline.py.
This patch implements printPipeline() to serialize drop-deref where
appropriate.
news/eilmeldung: update to 1.7.1
1.7.1 - 2026-07-27
This is a minor release with mainly bugfixes. However, there is a kind of breaking change:
breaking change for scoop users: eilmeldung is now in scoop main.
bugfix in article content display
no content is shown if now article is selected
tags are now updated correctly
an error on command clear now leads to an error message instead of a crash
undo tagging now bevaves correctly if an article already had the tag before
bump dependencies
1.7.0 - 2026-07-20
🙈 Ever thought, it is a shame that life has no undo function? 🙈 At least, eilmeldung has one now!
[16 lines not shown]
devel/feluda: update to 1.15.0
What's Changed
build(deps): bump com.fasterxml.jackson.core:jackson-databind from 2.16.1 to 2.22.0 in /examples/java-example/maven-example in the maven group across 1 directory by @dependabot[bot] in #228
feat(licenses): detect licenses from SPDX source file headers by @anistark in #229
fix(licenses): offline license text and copyleft restrictiveness by @anistark in #230
refactor: tidy string-argument ergonomics with impl Into<String> by @anistark in #231
docs: surface the github action and open community channels by @anistark in #232
feat(scan): flag foreign license headers in own source files by @anistark in #233
fix(java): resolve maven pom property placeholders in parent coords by @anistark in #234
feat(go): migrate license resolution to the official pkg.go.dev API by @anistark in #235
feat(tui): overhaul the --gui table layout, theme, and interactions by @anistark in #236
feat(cli): update --help and --version output by @anistark in #237
build(deps): bump com.fasterxml.jackson.core:jackson-databind from 2.22.0 to 2.22.1 in /examples/java-example/maven-example in the maven group across 1 directory by @dependabot[bot] in #239
fix(node): strip YAML quotes from scoped dep names in pnpm-lock.yaml by @01luyicheng in #238
feat(scan): detect vendored and unmanaged dependencies by @anistark in #240
docs: redesign the site styling and add host serve by @anistark in #242
[OpenACC][NFC] Minor clean up in ACCRoutineLowering. (#213333)
Minor NFC clean up after recent changes to remove nohost handling from
ACCRoutineLowering.
Assisted-by: Codex