[llvm-cov] Add show-function-summary option in coverage report (#172980)
Adds a command line argument to llvm-cov to optionally disable the
function summary in the coverage report (but maintaining existing
behavior by keeping it enabled by default).
Add clustered SMB state
This commit adds working configuration for stateful SMB HA failover:
* reclock helper script - determines which node holds the cluster
mutex lock based on presence of data pool system dataset.
* ctdb-related etc files. We can hard-code the nodes config based
on our known HA nodes (this significantly eases past problems with
gluster that we saw with dynamic nodes and ctdb stability).
* smbd will remain stopped on standby controller until it becomes
standby.
* keepalived will continue to manage virtual IPs
NAS-139251 / 25.10.2 / Fix netdata plugin startup delay for disk temp and pool stats (by Qubad786) (#17994)
## Problem
Currently when netdata is restarted, for each plugin it gives data after
`2 * update every frequency` seconds. So for example, if a plugin data
is to be collected every 5 min, we get datapoints after 10 minutes which
is an unnecessary delay added by how netdata implements `SimpleService`.
## Solution
We have overridden default `SimpleService` to not sleep as soon as it
runs but rather collect data after initialization and then sleep. This
means that we now get data after every `update every frequency` which if
it is 5 min, we will get data after 5 min instead of 10.
The 5 min delay which we still see is because of how netdata dbengine
stores/aligns data and there is no good way around that as that logic
lives in core netdata.
[5 lines not shown]
NAS-139298 / 25.10.2 / Add 'limit' option to 'audit.query' call. (by mgrimesix) (#18019)
`audit.query` calls must include either a `count` or `limit` option,
else the test will fail.
Added the missing `limit` option for the `test_system_reboot` test in
`test_system_lifetime.py`
Original PR: https://github.com/truenas/middleware/pull/18018
Co-authored-by: Mark Grimes <mark.grimes at ixsystems.com>
[PatternMatch] Implement match_fn using bind_back (NFC) (#175811)
Use llvm::bind_back landed in d2a521750 ([ADT] Introduce
bind_{front,back}, [not_]equal_to, #175056) to simplify implementations
of match_fn in PatternMatch and VPlanPatternMatch.
[RISCV] AI Foundry ET extensions for RISC-V (#174571)
This patch is part of an ongoing effort for upstreaming the AI Foundry
ET
extensions for RISC-V, formerly known as the ET-SoC-1 Minion extensions
by
Esperanto Technologies.
Changes
=======
The changes in this patch are:
- Defining the XAIFET feature as an "umbrella" extension under which
all the vendor extensions from ET-SoC-1 are included.
- Defining Esperanto Technologies's ET-SoC-1 Minion processor (et-soc1).
- Defining Ainekko's Erbium Minion processor (an-erbium), using the same
[67 lines not shown]
rust: update to 1.91
Based on wip/rust191 by he@, jperkin@, and adam@
Changes:
Add lint against dangling raw pointers from local variables
Stabilized APIs
Path::file_prefix
AtomicPtr::fetch_ptr_add
AtomicPtr::fetch_ptr_sub
AtomicPtr::fetch_byte_add
AtomicPtr::fetch_byte_sub
AtomicPtr::fetch_or
AtomicPtr::fetch_and
AtomicPtr::fetch_xor
{integer}::strict_add
[61 lines not shown]
ADT: Add utility functions for comparing FPClassTest (#175380)
Add utility functions for checking if less and greater queries
are known to not evaluate to true. This will permit more precise
folding of min/max intrinsics. The test is kind of a mess.
[llvm] Fix comment references deprecated make_scope_exit (#175820)
After #173131 and #174030, make_scope_exit is no longer used in
ThreadPool. Fix comment that references old APIs and references the new
API instead.
interface: POC for multi-dhcp6c support
I'm not entirely sure why we settled for a single deamon of
dhcp6c back in the day, but there are certianly downsides to
it and I don't see something that wasn't fixed in the meantime
that makes this not work.
At the moment this splits off dhcp6c only but we need to
change the daemon's print a bit to avoid complaining about
"other" devices since the situation to ignore a non-listening
interface is normal and not "ignoring" something obvious as
the INFO log message suggests.
Remove the $id bootstrapping now that every service has its
own configuration and "0" cannot overlap and add some options
relavant to #7647. The assoc-id principle and some other options
are now implemented.
rtsold still needs to be split to allow for HUP reload of
[2 lines not shown]
[VPlan] Optimize BranchOnTwoConds to chain of 2 simple branches. (#174016)
This patch improves the lowering for BranchOnTwoConds added in
https://github.com/llvm/llvm-project/pull/172750 by replacing the branch
on OR with a chain of 2 branches.
On Apple M cores, the new lowering is ~8-10% faster for std::find-like
loops. It also makes it easier to determine the early exits in VPlan. I
am also planning on extensions to support loops with multiple early
exits and early-exits at different positions, which should also be
slightly easier to do with the new representation.
PR: https://github.com/llvm/llvm-project/pull/174016
[clang-tools-extra] Update Maintainers for Clang-Doc
Currently, Erick Velez has been doing the bulk of clang-doc development.
The maintainer being removed hasn't participated in almost a year, so it
would be good to have active maintainers listed in the file.
AMDGPU: Change ABI of 16-bit scalar values for gfx6/gfx7
Keep bf16/f16 values encoded as the low half of a 32-bit register,
instead of promoting to float. This avoids unwanted FP effects
from the fpext/fptrunc which should not be implied by just
passing an argument. This also fixes ABI divergence between
SelectionDAG and GlobalISel.
I've wanted to make this change for ages, and failed the last
few times. The main complication was the hack to return
shader integer types in SGPRs, which now needs to inspect
the underlying IR type.
AMDGPU: Change ABI of 16-bit element vectors on gfx6/7
Fix ABI on old subtargets so match new subtargets, packing
16-bit element subvectors into 32-bit registers. Previously
this would be scalarized and promoted to i32/float.
Note this only changes the vector cases. Scalar i16/half are
still promoted to i32/float for now. I've unsuccessfully tried
to make that switch in the past, so leave that for later.
This will help with removal of softPromoteHalfType.
17819 Update gitignore files for recently added artefacts
Reviewed by: Robert Mustacchi <rm+illumos at fingolfin.org>
Approved by: Dan McDonald <danmcd at edgecast.io>
InstCombine: Implement SimplifyDemandedFPClass for fma
This can't do much filtering on the sources, except for nans.
We can also attempt to introduce ninf/nnan.
