| Age | Commit message (Collapse) | Author |
|---|
|
This is based on an initial implementation by michaelo in
https://reviews.freebsd.org/D54361.
PR: 270497
Reported by: michaelo
Reviewed by: michaelo
Tested by: michaelo
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D54606
|
|
The KERN_OPTS:MDEV_ACPI checks are fine for as long as we are building
modules along the kernel. If one wants to just build the module
standalone out of the module directory this would fail.
Add the missing include for kmod.opts.mk (as was done for tcp
in 1319a76179682).
Sponsored by: The FreeBSD Foundation
Reported by: Tassilo Philipp (tphilipp potion-studios.com)
Fixes: f5a77dc8f8df ("improve module Makefile dependency on ACPI")
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D54769
|
|
If APIC is left in somewhat bad state, with some source hot (not masked
and active, e.g. timers after kexec or due to BIOS bug), we get the
interrupt too early.
Reported by: jmg
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54543
|
|
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54543
|
|
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54543
|
|
MFC after: 3 days
|
|
No functional change intended.
Reviewed by: pouria, delphij, imp
Approved by: glebius (mentor)
Differential Revision: https://reviews.freebsd.org/D54752
|
|
While here, fix manlint warnings in rtnetlink(4).
Reviewed by: melifaro, ziaee, glebius
Approved by: glebius (mentor)
Differential Revision: https://reviews.freebsd.org/D53786
|
|
NB: Rest of ipfw(4) sources get sx.h via vnet.h, which isn't perfect.
|
|
Fix a typo in the rdr_action_head() test.
Fixes: 685fb4253819 ("pf: Log the intended action when a NAT rule matches a packet")
MFC after: 1 week
|
|
Fixes: 895eeb492 ("powerpc/loader: Add CAS support for older CPUs")
MFC after: 1 week
|
|
This patch adds support for receive checksum offload for TCP/IPv6
and UDP/IPv6. Since receive checksum offload can't be configured
separately for IPv4 and IPv6, IFCAP_RXCSUM and IFCAP_RXCSUM_IPV6
can't be changed independently.
Reviewed by: Timo Völker
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54756
|
|
This patch adds support for transmit checksum offload for TCP/IPv6
and UDP/IPv6.
Reviewed by: Timo Völker
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54754
|
|
QEMU creates a "ibm,arch-vec-5-platform-support" property for all
pseries emulations. Add POWER7 and POWER6 to the CAS list, more can be
added later as needed/desired.
MFC after: 1 week
|
|
- s/vaues/values/
MFC after: 5 days
|
|
- s/vaues/values/
MFC after: 5 days
|
|
MFC after: 1 week
|
|
Add a -U option to build NanoBSD images without root privileges. It
relies on makefs/mkimg and metalog (mtree) files, similar to what
release engineering uses to build images.
Keep the current way to build NanoBSD images untouched. Once this
method gets battle tested, it may be used to build images as root as
well.
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48793
|
|
Add a provisional _populate_data_part function. It populates the
optional /data partition, but using makefs(8), which is more in-line
with what release engineering uses to create images.
Reviewed by: emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48792
|
|
Add a provisional _populate_cfg_part function. It populates the /cfg
partition, but using makefs(8), which is more in-line with what release
engineering uses to create images.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48791
|
|
Add a _populate_part(ition) function that mimics the current
populate_slice. Note however, that this function is not
backward-compatible with populate_slice, hence the different name. A
"_" is prepended to signal that it still experimental.
It can be used to populate the /cfg and /data partition using makefs(8).
Initially not wired.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48790
|
|
Add a NANO_MAKEFS global variable with options equivalent to NANO_NEWFS
to be used with a nano_makefs function for creating images.
Also add a function that adjusts the code size calculation, so makefs -s
won't error about the minimum rounded size. Ideally this shim should be
removed, therefore the suspicious _xxx prefix.
Initially not wired.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48789
|
|
Introduce a function tgt_switch_root_fstab() that switches the root
partition in the target file system tab file. Initially not wired.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48788
|
|
Initially not wired, this variable will hold the time stamp for all the
files of the final image.
By default uses the last commit time stamp, if empty, it will use the
start time of the NanoBSD build.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48787
|
|
Add tgt_dir(), that creates and adds directory entries to the metalog
specification file. Initially not wired.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48786
|
|
Prefer tgt_touch() as it adds an entry to the metalog file.
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48785
|
|
The correct mode for /tmp is 1777, keep the same mode when symlinking it
to /var/tmp.
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48784
|
|
Prefer tgt_dir2symlink() to create symlinks, as it has the benefit of
appending the entry to the metalog file.
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48783
|
|
Make tgt_dir2symlink() take an optional third argument that sets the
mode. By default, assume an absolute mode of 0777, to maintain
backward compatibility.
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48782
|
|
Also include the default NanoBSD uname/gname in the specification file.
By default, NANO_DEF_UNAME and NANO_DEF_GNAME map to root and wheel
respectively.
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48781
|
|
nano_umount[^1] is a convenience routine used to override issues with
umount(8) without changing the nanobsd.sh source code. There were a few
places where it was not being used.
[^1]: Introduced in 6d12b61a88a6 ("Add a routine for easy workaround any
umount issues w/o hacking nanobsd.sh.")
Reviewed by: imp, emaste
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D48780
|
|
Bump the timeout value (for the third time), to avoid cutoff on emulated
architectures on ci.freebsd.org. One of the runners will take
considerably more time to complete this test.
Reported by: Jenkins
Fixes: a551b0524953 ("tarfs: tests: Increase timeout (again)")
MFC after: 3 days
|
|
No functional change intended.
MFC after: 3 days
|
|
NO_FP_LIBC was added in 2004 to save space by disabling FP support in
*printf()/*scanf(). The size benefit is negligible on modern systems
and conflicts with assumptions made by current base utilities.
Remove the option and always build libc with floating-point support.
Reported by: Oskar Holmlund <eovholmlund at gmail com>
MFC after: 2 weeks
|
|
Unfortunately the aq driver is using readl/writel calls instead
of bus space routines.
This broke, well, everything else.
Fixes: c75eff16ef54aaae7b5dc52ed894cc73a855f469
|
|
Add atomic_set/clear_short/char for doing 8-bit and 16-bit operations
more efficiently on "newer" architectures (POWER7 and later).
Piggybacks on b31abc95eb.
|
|
|
|
Make the tap database belong to ip_fw_chain, but leave the default "ipfw0"
tap per-vnet. This is only slightly better than keeping the database per-
vnet, as the bpf name space is per-vnet. However, we yet have only single
ipfw chain. Whenever multiple chains will coexist, this needs to be
addressed.
Require the chain lock to make modifications to the database.
Move tap allocation to a later ruleset build stage, when all rule numbers
are known already. This fixes a panic introduced by 3daae1ac1d82.
Fixes: 3daae1ac1d82ecdcd855101bab5206e914b12350
|
|
The lock grab & drop predates epoch(9) introduction to the network
stack and it doesn't provide a true guarantee that all threads that
may use ipfw configuration have finished. Also the lock prevented
from sleepable operations when freeing the rules.
|
|
created to allow addressable leds to be driven by abusing spi bus as waveform generator. this might have other uses for similar "permanent" spi transfers
Differential Revision: https://reviews.freebsd.org/D54734
Reviewed by: adrian
|
|
Initialize pindex in swap_pager_getpages_locked() before the loop
actually calculating it by skipping bogus pages. Compiler is worried
that loop might never assign to it, which actually cannot happen.
Sponsored by: The FreeBSD Foundation
Fixes: d198ad51ea73bbb162336923a387f52b0b1c1f1d
MFC after: 1 week
|
|
Reviewed by: adrian
Differential Revision: https://reviews.freebsd.org/D54633
|
|
Same as vnode_pager_generic_getpages_async(), swap_pager_getpages() must
handle a possibility of the provided page run to include bogus_page on
some positions, when called from sendfile_swapin(). The swap pager is
used for tmpfs vnodes.
In particular, the bogus page must not be used for pindex calculation,
we better not update the flags on it or wait for the flag clearing, and
we must not call vm_page_valid() because the function expects busy page.
This was bisected down to 72ddb6de1028426 (unix: increase
net.local.(stream|seqpacket).(recv|send)space to 64 KiB),
which is somewhat surprising, but apparently reasonable because it
allowed the run of more than one page for page-in from the swap pager,
which now might include valid pages replaced by bogus one.
In collaboration with: pho
Reviewed by: glebius, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54713
|
|
Reviewed by: glebius, markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54713
|
|
Reviewed by: glebius, markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54713
|
|
Before the "upper half lock" became sleepable the table manipulation code
needed sophisticated workarounds to recover from races, where the lock is
temporarily dropped to do malloc(M_WAITOK). Remove all these workarounds
as they are no longer needed.
Differential Revision: https://reviews.freebsd.org/D54580
|
|
The so called upper half ipfw lock is not used in the forwarding path. It
is used only during configuration changes and servicing system events like
interface arrival/departure or vnet creation. The original code drops the
lock before malloc(M_WAITOK) and then goes into great efforts to recover
from possible races. But the races still exist, e.g. create_table() would
first check for table existence, but then drop the lock. The change also
fixes unlock leak in check_table_space() in a branch that apparently was
never entered.
Changing to a sleepable lock we can reduce a lot of existing complexity
associated with race recovery, and as use the lock to cover other
configuration time allocations, like recently added per-rule bpf(4) taps.
This change doesn't remove much of a race recovery code, to ease bisection
in case of a regression. This will be done in a separate commit. This
change just removes lock drops during configuration events. The only
reduction is removal of get_map(), which is a straightforward reduce to a
simple malloc(9).
The only sleepable context where the lock was acquired was dyn_tick(). The
comment said it is done to prevent parallel execution of
dyn_expire_states(). However, there is proper internal locking in there
and function should be safe to execute in parallel. The real problem is
dyn_expire_states() called via userland to race with dyn_grow_hashtable()
called via dyn_tick(). Protect against this condition with the main chain
lock.
Differential Revision: https://reviews.freebsd.org/D54535
|
|
Reviewed by: pouria, ae
Differential Revision: https://reviews.freebsd.org/D54579
|
|
Otherwise, a forwarding thread may use the interface being detached. This
is a regression from 0d469d23715d, which manifests itself as a reliably
reproducible panic in in6_selecthlim(). Note that there are old bug
reports about such a panic, and I believe this change will not fix them,
as their nature is not due to a screwed up detach sequence, but due to
lack of proper epoch(9) based synchronization between the detach and
forwarding.
Reviewed by: pouria
Reported & tested by: jhibbits
PR: 292162
Fixes: 0d469d23715d690b863787ebfa51529e1f6a9092
Differential Revision: https://reviews.freebsd.org/D54721
|
|
Building the LINT-NOIP kernel on amd64 with clang 21 results in a
-Werror warning similar to:
sys/dev/mxge/if_mxge.c:1846:44: error: variable 'sum' is uninitialized when passed as a const pointer argument here [-Werror,-Wuninitialized-const-pointer]
1846 | cksum_offset, sizeof(sum), (caddr_t)&sum);
| ^~~
Indeed, if both `INET` and `INET6` are undefined, `sum` is never
initialized. Initialize it to zero to silence the warning.
Reviewed by: jhibbits
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54730
|
