rss_getkey function is now available even if kernel RSS option
is disabled. This makes ice_get_default_rss_key no longer needed.
Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
Reviewed by: gallatin
Approved by: kbowling (mentor)
Fixes: 012ecdde3a ("ice: use newly exposed RSS hash key API rather than ad-hoc hashing")
Sponsored by: Intel Corporation
Differential Revision: https://reviews.freebsd.org/D54281
If we get a random number, the NZCV is set to 0b0000. Then
"cset %w1, ne" will test whether Z == 0 and set %w1 to 1 if so.
More specifically, "cset %w1, ne" maps to "csinc %w1, wzr, wzr, eq",
which stores 0 in %w1 when NZCV == 0b0100 and 1 otherwise.
Thus, on a successful read we expect ret != 0, so the loop condition
needs to be fixed. In practice this means that we would end up trying
to fetch entropy up to ten times in a row. If all attempts are
successful, the last will be returned, otherwise no entropy will be
returned.
Reported by: Kevin Day <kevin@your.org>
Reviewed by: andrew
Fixes: 9eecef0521 ("Add an Armv8 rndr random number provider")
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D54259
This change retires two historic relics: the if_afdata[] array and the
dom_ifattach/dom_ifdetach methods.
The if_afdata[] array is a relic of the era, when there was expectation
that many transport protocols will coexist with IP, e.g. IPX or NetAtalk.
The array hasn't had any members except AF_INET and AF_INET6 for over a
decade already. This change removes the array and just leaves two pointer
fields: if_inet and if_inet6.
The dom_ifattach/dom_ifdetach predates the EVENTHANDLER(9) framework and
was a good enough method to initialize protocol contexts back then. Today
there is no good reason to treat IPv4 and IPv6 stacks differently to other
protocols/features that attach and detach from an interface.
The locking of if_afdata[] is a relic of SMPng times, when the system
startup and the interface attach was even more convoluted than before this
change, and we also had unloadable protocols that used a field in
if_afdata[]. Note that IPv4 and IPv6 are not unloadable.
Note that this change removes NET_EPOCH_WAIT() from the interface detach
sequence. This may surface several new races associated with interface
removal. I failed to hit any with consecutive test suite runs, though.
The expected general race scenario is that while struct ifnet is freed
with proper epoch_call(9) itself, some structures hanging off ifnet are
freed with direct free(9). The proper fix is either make if_foo point at
some static "dead" structure providing SMP visibility of this store, or
free those structure with epoch_call(9). All of these cases are planned
to be found and resolved during 16.0-CURRENT lifetime.
Reviewed by: zlei, gallatin, melifaro
Differential Revision: https://reviews.freebsd.org/D54089
Run the original ifnet_arrival_event before linking the interface.
Otherwise there is a race window when interface is already visible, but
not all of the protocols have completed their attach. Provide a new event
handler ifnet_attached_event, that is executed when the inteface is fully
visible. Use it in route(4) socket and netlink(4) to announce new
interface to the userland. Properly document the ifnet events in if_var.h.
Reviewed by: zlei, melifaro
Differential Revision: https://reviews.freebsd.org/D54085
Added libxo support to `gpart show`, also updated the man
pages for geom and gpart to show where you can expect
libxo formatted output.
PR: 290629
MFC after: 1 week
Sponsored by: ConnectWise
Reviewed by: asomers, mckusick, phil
Approved by: asomers (mentor)
Differential Revision: https://reviews.freebsd.org/D53950
Provide the flags used for a vtnet interface via the sysctl tree.
This is mostly used for debugging purposes.
Reviewed by: Timo Völker
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D54283
gcc complains when building libuvmem because the last two operands of ?:
in ORDER2SIZE and SIZE2ORDER have different signs. Add explicit casts
to address this.
Reported by: Jenkins
Reviewed by: alc, kib
MFC after: 1 week
Fixes: 1ecf01065b ("libuvmem: usermode port of vmem(9)")
Differential Revision: https://reviews.freebsd.org/D54268
Some refresh functions had two layers of 'do we need to refresh now?'
checks, leading to inconsistent refreshes.
Consolidate them.
PR: 291725
Sponsored by: Rubicon Communications, LLC ("Netgate")
This adds support for renaming a symbolic link found on the lower fs,
which necessitates copying it to the upper fs, as well as basic tests.
MFC after: 1 week
Sponsored by: Klara, Inc.
Sponsored by: NetApp, Inc.
Reviewed by: olce, siderop1_netapp.com, jah
Differential Revision: https://reviews.freebsd.org/D54229
When we are using Cirrus-CI and using a PR branch it creates a filename
in the form "pull/XXXX" which becomes a path seperator and the file
creation process fails.
Fails to complete the process with the following:
/bin/sh: cannot create
/tmp/meta.hUNGUq/ci-FreeBSD-16.0-pull/1932-amd64-nullhash-GENERIC.env:
No such file or directory
For future also apply the same regex for OSRELEASE and VOLUME_LABEL.
Coming vmap_pfn() implementation requires is_vmalloc_addr() to be in
the same file with other vmap-related code. Move code out from the
overcrowded file. No functional changes intended.
MFC after: 1 week
Reviewed by: bz
Differential Revision: https://reviews.freebsd.org/D54223
The ac6a7f6216 enabled execution of vlan_clone_dump_nl(), which
previously was effectively disabled. The function itself was added back
in 089104e0e0. This exposed a bug when Netlink dumps info on all
interfaces using a dangerous KPI if_foreach_sleep(), which may call its
callbacks on completely detached interfaces, hanging on the last
reference. The ifc_dump_ifp_nl_default() is able to digest such interface
without a panic, but vlan_clone_dump_nl() can't. Neither of the above
revisions is the actual culprit, rather it is design problem of detaching
interfaces and if_foreach_sleep().
Plug the problem with removing pointer to freed memory on detach and
making a NULL check later.
Reported by: pho
The reason the global list worked before 8774a990ee is that bpf_setif()
used if_unit(), which is a VNET-aware function, and then went through the
global list looking for bpf_if with matching pointer.
PR: 291735
Fixes: 8774a990ee
As we have received an external box sponsored by NetActuate we can now
enable the pre-commit tasks to be run in this box.
Once we have enabled this debug if there are any errors and after
confirmation we can enable this for other tasks too.
Sponsored by: NetActuate
Sponsored by: The FreeBSD Foundation
Approved by: emaste, lwhsu
Differential Revision: https://reviews.freebsd.org/D54220
The LIBRARY section has been deprecated by upstream.
Show the updated usage of the Lb macro in SYNOPSIS.
Fixes: 4c07abdbac (mandoc: Vendor import of upstream at 2025-06-13)
Modeled directly after the method used by the zfs/zpool commands: flag
commands with a "please log me" flag, and when there, reconstruct the
command line. On success, call the library function to add it to the
log.
(Majority of the change by Rob; minor edits by kevans@)
Signed-off-by: Rob Norris <rob.norris@klarasystems.com>
Co-authored-by: Kyle Evans <kevans@FreeBSD.org>
Sponsored by: Modirum MDPay
Sponsored by: Klara, Inc.
If the kernel panics while a thread is in the middle of an SDHCI
transaction, trying to dump to a dump target on the MMC would result in
a hang. Fix this by completing the transaction first.
Reviewed by: imp
Obtained from: Hewlett Packard Enterprise
Differential Revision: https://reviews.freebsd.org/D54255
This change introduces "monitor mode", a mechanism for automatically
releasing virtual machine resources when bhyve dies, bringing us
closer towards making non-root bhyve viable.
Under this regime bhyve will create a transient virtual machine using
`vmmctl`'s `VMMCTL_CREATE_DESTROY_ON_CLOSE` flag and automatically
reboot said virtual machine as long as it exits with "reboot" status.
This is done by splitting bhyve into two processes. The parent process
creates the virtual machine while the child process initializes and
runs the virtual machine. When the child exits the parent inspects
its exit status and either exits or forks again. `vmmctl` automatically
destroys the underlying virtual machine once the parent process dies.
Differential Revision: https://reviews.freebsd.org/D53731
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
MFC after: 3 months
bhyve's exit status codes indicate how the VM was terminated.
Unfortunately, their meaning within the source code is somewhat
unclear since they are only used as magic numbers.
Fix this by defining exit status macros and using them to replace
the magic numbers in exit(3) function calls.
Differential Revision: https://reviews.freebsd.org/D53730
Reviewed by: markj, corvink, emaste
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
MFC after: 3 months
This change adds the ability to tie a virtual machine's lifecycle to
a /dev/vmmctl file descriptor. A user can request `vmmctl` to destroy a
virtual machine on close using the `VMMCTL_CREATE_DESTROY_ON_CLOSE` flag
when creating the virtual machine. `vmmctl` tracks such virtual machines
in per-descriptor lists.
Differential Revision: https://reviews.freebsd.org/D53729
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
MFC after: 3 months
This change introduces a new per-UID limit for controlling the
number of vmm instances, in anticipation of unprivileged bhyve.
This allows ut to limit the amount of kernel memory allocated
by the vmm driver and prevent potential memory exhaustion attacks.
Differential Revision: https://reviews.freebsd.org/D53728
Reviewed by: markj, olce, corvink
MFC after: 3 months
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
These are the remaining trivial xfail cases where no other
test logic changes are necessary.
Signed-off-by: Siva Mahadevan <me@svmhdvn.name>
PR: 290099
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Pull Request: https://github.com/freebsd/freebsd-src/pull/1871
Replace ATF_CHECK_* with ATF_REQUIRE_* to fail fast and avoid
unexpected aborts.
Signed-off-by: Siva Mahadevan <me@svmhdvn.name>
PR: 290099
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Pull Request: https://github.com/freebsd/freebsd-src/pull/1871
Replace ATF_CHECK_* with ATF_REQUIRE_* to fail fast and avoid
unexpected abort.
PR: 290099
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Pull Request: https://github.com/freebsd/freebsd-src/pull/1871
Replace ATF_CHECK_* with ATF_REQUIRE_* to fail fast and
avoid unexpected aborts.
Signed-off-by: Siva Mahadevan <me@svmhdvn.name>
PR: 290099
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Pull Request: https://github.com/freebsd/freebsd-src/pull/1871
This fixes broken test reports with the following message:
lib/msun/fe_round_test:fe_round -> broken: Test result contains
multiple lines: expected_failure:
/usr/src/contrib/netbsd-tests/lib/libm/t_fe_round.c:95: Didn't get the
same rounding mode out!<<NEWLINE>>(index 0) fed in 64 rounding mode, got
0 out<<NEWLINE>> [0.079s]
Signed-off-by: Siva Mahadevan <me@svmhdvn.name>
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Pull Request: https://github.com/freebsd/freebsd-src/pull/1871
For reproducing errors or test results it is important to gather
environment information. These environments are divided into two parts.
One part is in which environment the artifacts were built into and the
second part is in which environment the tests were run.
This patch collects thesee information and saves into a .env file in
the metadir. After this patch lands we will also need to change our
jenkins job where we are uploading the artifact to a central location.
This environment file should also be stored along with the artifact.
For easier location the image basename and the environment basename are
kept same.
Approved by: lwhsu
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D54247
Konstantin Belousov