Summary:
the flags argument is unused.
Its initial design idea has been superceded by the addition of
virtio_setup_intr and related APIs.
Sponsored by: The FreeBSD Foundation
Reviewers: bryanv
Reviewed By: bryanv
Subscribers: cognet, imp
Differential Revision: https://reviews.freebsd.org/D41850
Summary:
these static functions:
- vq_ring_use_indirect
and these public functions:
- virtio_bus_is_modern
- virtio_with_feature
- virtqueue_empty
- virtqueue_full
exclusively return a boolean value.
Change their return values to bool.
Sponsored by: The FreeBSD Foundation
Reviewers: bryanv
Reviewed By: bryanv
Subscribers: cognet, imp
Differential Revision: https://reviews.freebsd.org/D41848
With bmake-20230909 we can use -j1.5C to use 1.5*ncpu
bmake will set .MAKE.JOBS.C=yes if this is supported,
and in that case jobs.mk will use JOB_MAX_C as default for JOB_MAX.
Remove logic from local.sys.mk which is now handled by jobs.mk
Also use latest dirdeps-targets.mk and meta.{autodep,subdir}.mk
If we've not (yet) resolved a neighbour nda_lladdr will be NULL, and
NLA_DATA_LEN(neigh->nda_lladdr) will dereference a NULL pointer.
Avoid that by checking nda_lladdr first, and only dereferencing if it's
not NULL.
Test case:
ping6 -c 1 <non-existant neighbour>
ndp -a
Reviewed by: melifaro
MFC after: 3 days
Sponsored by: Rubicon Communications, LLC ("Netgate")
Differential Revision: https://reviews.freebsd.org/D41903
- pclk and hclk are mandatory so always try to get them.
Don't make it fatal if it fails as some platform (like Zynq) don't
have a proper clock driver.
- Always use pclk for the reference clock.
- Try to get all the possible clocks and enable them.
Reviewed-by: mhorne
Tested-by: Milan Obuch <bsd@dino.sk>
Differential Revision: https://reviews.freebsd.org/D41857
Sponsored by: Beckhoff Automation GmbH & Co. KG
- Remove cdns,gem, it's the generic binding but for all platform that include
this one we need specific drivers setup so remove it.
- Remove cdns,macb, it's the generic binding for Atmel AT91 which we don't suport
- Remove cadence,gem, it's not an official binding and seems to be only used in some
obscure ARM11 SoC.
- Note that the cdns,zynq* are deprecated
- Add the new Xilinx compatible for zynq and zynqmp
Reviewed-by: mhorne
Tested-by: skibo, Milan Obuch <bsd@dino.sk>
Differential Revision: https://reviews.freebsd.org/D41856
Sponsored by: Beckhoff Automation GmbH & Co. KG
The binding says that we can have one or two clocks to export.
The first one is the actual sdclock while the second is the sample clock.
Both have the same parent, clk_xin.
Correctly export the clocks for RK3399 and ZynqMP.
No need to use a high ID as before, we have our own clock domain so use
ids starting at 1 as all exported clocks should be.
Reviewed-by: bz
Differential Revision: https://reviews.freebsd.org/D41810
Sponsored by: Beckhoff Automation GmbH & Co. KG
Per the bindings the phy and the syscon can always be present not just
for RK3399.
Reviewed-by: bz
Differiential Revision: https://reviews.freebsd.org/D41809
Sponsored by: Beckhoff Automation GmbH & Co. KG
Those two (in fact all of the supported one in this driver except RK3568) always
needs the clocks to be enabled.
Reviewed-by: bz
Differential Revision: https://reviews.freebsd.org/D41808
Sponsored by: Beckhoff Automation GmbH & Co. KG
This was used when we had our own DTS, it's not used anymore.
Reviewed-by: bz
Differential Revision: https://reviews.freebsd.org/D41807
Sponsored by: Beckhoff Automation GmbH & Co. KG
Add clock and reset drivers for the ZynqMP SoC.
The clocks are discovered by talking to the firmware as the topology isn't
fixed on this SoC.
Differential Revision: https://reviews.freebsd.org/D41812
Sponsored by: Beckhoff Automation GmbH & Co. KG
The ZynqMP SoC have a MCU running a firmware to control clocks, resets,
fpga loading etc ...
Add a driver that can be use to communicate with it.
For now only the clock and reset part are implemented.
Differential Revision: https://reviews.freebsd.org/D41811
Sponsored by: Beckhoff Automation GmbH & Co. KG
In 6cd1bc5316 for PR 263489 I changed freebsd-update to restart sshd
after upgrade, to avoid an upgrade-related incompatibility that made it
impossible to login.
This is intended to avoid losing access to remote hosts, and ought not
apply to upgrading jails (from outside).
PR: 263489, 272282
Reported by: otis
Reviewed by: otis, kevans
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Fixes: 6cd1bc5316 ("freebsd-update: restart sshd after upgrade")
Differential Revision: https://reviews.freebsd.org/D41890
Summary:
Provide an implementation of fpu_kern_enter/fpu_kern_leave for PPC to
enable FPU, VSX, and Altivec usage in-kernel. The functions currently
only support FPU_KERN_NOCTX, but this is sufficient for ossl(1) and many
other users of the API.
This patchset has been tested on powerpc64le using a modified version of
the in-tree tools/tools/crypto/cryptocheck.c tool to check for FPU/Vec
register clobbering along with a follow-up patch to enable ossl(4) on
powerpc64*.
Reviewed by: jhibbits
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D41540
Update import.sh to generate below
ChangeLog since bmake-20230622
2023-09-09 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230909
Merge with NetBSD make, pick up
o main.c: allow -j to compute a multiple of ncpu
If _SC_NPROCESSORS_ONLN is supported; and -j arg is a floating
point number or ends in 'C' compute .MAKE.JOBS as a multiple of
_SC_NPROCESSORS_ONLN
.MAKE.JOBS.C will be "yes" if -jC is supported
2023-08-20 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230820
Merge with NetBSD make, pick up
o make.1: note that :localtime is better for %s
o parse.c: improve error messages for invalid input.
o var.c: fix for %s:L:gmtime - set TZ=UTC and use localtime to get
correct result, it is still better to use %s:L:localtime.
2023-08-18 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230818
Merge with NetBSD make, pick up
o meta.c: meta_ignore - check raw path against metaIgnorePaths
to potentially skip call to realpath.
o var.c: be strict when parsing the argument of the ':mtime' modifier
o unit-tests/varmod-mtime.mk: document why '${%s:L:localtime}'
should be used to get an equivalent value to time(3).
2023-08-16 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230816
Merge with NetBSD make, pick up
o cond.c: clean up multiple-inclusion guards
2023-07-25 Simon J Gerraty <sjg@beast.crufty.net>
* unit-tests/Makefile: addd varmod-localtime to BROKEN_TESTS
if configure cannot work out how to control TZ.
Remove varmod-localtime from BROKEN_TESTS for IRIX*
2023-07-24 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230723
* configure.in: fix the test for wether TZ=Europe/Berlin works.
Depending on the time of year, if run between 22:00 and 00:00 UTC
the check in configure would fail incorrectly.
Take the day into account as well.
2023-07-18 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230711
Merge with NetBSD make, pick up
o make.1: clean up wording, clarify scope of '!' in conditions
2023-07-15 Simon J Gerraty <sjg@beast.crufty.net>
* make-bootstrap.sh.in: set prefix
If configure is run using ksh we get unexpanded ${prefix} in
DEFAULT_SYS_PATH, by ensuring prefix is set we should still get
correct result.
2023-07-13 Simon J Gerraty <sjg@beast.crufty.net>
* VERSION (_MAKE_VERSION): 20230711
bump version for IRIX tweaks
* make.h: undef OP_NONE if defined
* unit-tests/Makefile: set BROKEN_TESTS for IRIX
* configure.in: override INSTALL on IRIX
2023-06-27 Simon J Gerraty <sjg@beast.crufty.net>
* boot-strap op_test: ensure we set TEST_MAKE as we want it.
mk/ChangeLog since bmake-20230622
2023-09-09 Simon J Gerraty <sjg@beast.crufty.net>
* jobs.mk (JOB_MAX): use -jC if we can
we actually use JOB_MAX_C which defaults to 1.33C
2023-08-18 Simon J Gerraty <sjg@beast.crufty.net>
* now_utc: %s only works with :localtime
2023-07-14 Simon J Gerraty <sjg@beast.crufty.net>
* install-sh: ignore -c as claimed and only insist on
a directory for destination when more than one file to copy.
* sys.mk: when looking for SYS_OS_MK try ${.MAKE.OS} and
${.MAKE.OS:S,64,,} early (so we find sys/IRIX.mk for IRIX64)
2023-07-13 Simon J Gerraty <sjg@beast.crufty.net>
* install-mk (MK_VERSION): 20230711
* sys.mk: set SYS_MK and INSTALL_SH for systems with incompatible
install(1)
* sys/IRIX.mk: when setting ROOT_GROUP only match the first :0:
set INSTALL to install-sh rather than pathname that may not exist
(yet).
2023-07-07 Simon J Gerraty <sjg@beast.crufty.net>
* dirdeps.mk: pass DIRDEP_TARGETS to DIRDEP_MAKE
normally this is empty - for the default target, but there are
use-cases where we might set it to something else.
2023-07-04 Simon J Gerraty <sjg@beast.crufty.net>
* install-mk (MK_VERSION): 20230704
* dirdeps.mk: apply DEBUG_DIRDEPS_LIST_FILTER to lists we output
when DEBUG_DIRDEPS is in effect.
Eg. DEBUG_DIRDEPS_LIST_FILTER=ts\n
can greatly improve readability.
Summary:
Provide an implementation of fpu_kern_enter/fpu_kern_leave for PPC to
enable FPU, VSX, and Altivec usage in-kernel. The functions currently
only support FPU_KERN_NOCTX, but this is sufficient for ossl(1) and many
other users of the API.
This patchset has been tested on powerpc64le using a modified version of
the in-tree tools/tools/crypto/cryptocheck.c tool to check for FPU/Vec
register clobbering along with a follow-up patch to enable ossl(4) on
powerpc64*.
Reviewed by: jhibbits
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D41540
Normally, we poll the device every so often to see if commands have
timed out. However, we'll go into the recovery state as part of failing
the drive. To account for all possibilties, if we're failed when we get
into the polling function, just stop polling: Party is over.
Sponsored by: Netflix
The pkg solver gets confused by the file moves and refuses to upgrade
without manually upgrading FreeBSD-utilities and FreeBSD-utilities-dev
first.
PR: 254173, 273859
MFC after: 1 week
Reviewed by: Mina Galić <freebsd@igalic.co>
Differential Revision: https://reviews.freebsd.org/D41892