HP van Braam [Tue, 23 Apr 2024 20:47:12 +0000 (14:47 -0600)]
ahc(4)/ahd(4): target mode: cancel outstanding AIOs and INOTs
When disabling a lun there can still be outstanding AIOs and INOTs, when
this happens previously the lun would just fail to disable and trying to
re-use the lun would break the card.
isp(4) in target mode does the same thing when disabling a lun, in
testing this allows re-starting of ctld(8) with connected initiators and
allows initiators to gracefully resume afterwards.
Signed-off-by: HP van Braam <hp@tmm.cx>
Reviewed by: imp, mav
Pull Request: https://github.com/freebsd/freebsd-src/pull/1190
Karl Tomlinson [Tue, 23 Apr 2024 18:18:25 +0000 (12:18 -0600)]
pow,powf(3),__ieee754_rem_pio2(f): Avoid negative integer left shift UB
A compiler clever enough to know that z is positive with a non-zero
biased exponent could, for example, optimize away the scalbnf(z,n) in
pow() because behavior for left shift of negative values is undefined.
`n` is negative when y*log2(|x|) < -0.5. i.e. |x^y| < sqrt(0.5)
The intended behavior for operator<< in this code is to shift the two's
complement representation of the first operand.
In the pow() functions, the result is added to the IEEE 754 exponent of
z = 2^y'. n may be negative enough to underflow the biased IEEE 754
exponent below zero, which is manifested in the sign bit of j
(which would correspond to the IEEE 754 sign bit).
The conversion from uint32_t to int32_t for out-of-int32_t-range values
is implementation defined. The assumed behavior of interpreting the
uint32_t value as a two's complement representation of a signed value
is already assumed in many parts of the code, such as uses of
GET_FLOAT_WORD() with signed integers.
This code passes all the current tests, and makes some out of tree
fuzzing tests pass again rather than hit UB (detailed in the commentary
of the pull request).
Signed-off-by: Karl Tomlinson <karlt+@karlt.net>
Reviewed by: imp, steve kargl, dim
Pull Request: https://github.com/freebsd/freebsd-src/pull/1137
__vprintf(): switch from strerror() to strerror_rl()
This eliminates the use of non-thread-safe function in printf*() family,
and make the call locale-aware. Also, it stops obliterating the
strerror() static buffer, which aligns with the POSIX requirement that
implementations must behave as if no standard-mandated functions call
strerror().
PR: 278556
Reported by: Jonathan Gruber <jonathan.gruber.jg@gmail.com>
Reviewed by: imp
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D44916
Mark Johnston [Tue, 23 Apr 2024 16:51:03 +0000 (12:51 -0400)]
nfsserver: Default to nfs_reserved_port_only="YES"
This setting causes the NFS server to check that all RPCs are sent from
a privileged (<= 1023) port, rejecting those that are not. This
slightly raises the bar for a user with network access to an
unauthenticated NFS server to access exported NFS filesystems.
Users that use traditional NFS clients (e.g., those provided by FreeBSD
or Linux) should not see any difference, assuming that unprivileged
filesystem mounting is disallowed.
Note that the setting is per-VNET, so may be overridden in VNET jails
without affecting the rest of the system.
Discussed with: freebsd-arch@
Reviewed by: rmacklem, bz, emaste
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D44906
sigqueue(2): add impl-specific flag __SIGQUEUE_TID
The flag allows the pid argument to designate a thread from the calling
process. The flag value is carved from the high bit of the signal
number, which slightly changes the ABI of syscall.
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D44867
This is used by 802.3 Ethernet. (Also be used by 802.4 Token Bus and
802.5 Token Ring, but we don't support those.)
This was accidentally removed along with FDDI support in commit 0437c8e3b198, presumably because comments implied it was used only by
FDDI or Token Ring.
Fixes: 0437c8e3b198 ("Remove support for FDDI networks.") Reviewed-by: emaste Signed-off-by: Denny Page <dennypage@me.com>
Pull-request: https://github.com/freebsd/freebsd-src/pull/1166
Update manual page references and macros to align to mandoc syntax
xlocale.3: Comment out reference to atof_l(3), atoi_l(3), atol_l(3), atoll_l(3)
These manual page references do not exist.
lagg.4: Change the reference for /etc/rc.conf from a reference link
.Xr -> .Pa based on the context within the manual page it is used.
buf.9: Remove .Xr entries from the file
The buf.9 manual page contains a commented out .Xr reference.
The <filmmein> 9 entry is a placeholder and has been removed for
clarity.
swab.c(libc): use a simplified version of byte swapping
This version of swab function simplifies the logic of swapping adjacent
bytes. Previous version of swab() used an arbitrary unrolling, which was
relevant back in the day but unnecessary for modern compilers, as if the
input size is known at compile time, they can do it automatically.
This version of swab() is inspired by musl.
A similar version can be found at: https://github.com/openbsd/src/blob/master/lib/libc/string/swab.c
Several users with alc(4)-based "Killer" Ethernet cards have reported
issues with this driver not passing traffic, which are solved by
disabling MSI-X using the provided tunable.
To work around this issue, disable MSI-X by default on this card.
This is done by having msix_disable default to 2, which means
"auto-detect". The user can still override this to either 0 or 1 as
desired.
Since these are slow (1Gbps) Ethernet ICs used in low-end systems, it's
unlikely this will cause any practical performance issues; on the other
hand, the card not working by default likely causes issues for many new
FreeBSD users who find their network port doesn't work and have no idea
why.
share/examples: move examples into appropriate packages
Allow a new variable SE_xxxPACKAGE to be used to place an example group
into a package.
Move the following examples into existing appropriate packages:
- bhyve examples into FreeBSD-bhyve
- bootforth examples into FreeBSD-bootloader
- csh examples into FreeBSD-csh
- ipfw examples into FreeBSD-ipfw
- jail examples into FreeBSD-jail
- pf examples into FreeBSD-pf
- ppp examples into FreeBSD-ppp
- printing examples into FreeBSD-lp
- uefisign examples into FreeBSD-efi-tools
- ypldap examples into FreeBSD-yp
- hast examples into FreeBSD-hast
Move all other examples into the new 'FreeBSD-examples' package.
This removes a significant number of files from the FreeBSD-utilities
package.
Reviewed by: imp, manu
Pull Request: https://github.com/freebsd/freebsd-src/pull/1176
Under Linux, the socket options IP_RECVERR and IPV6_RECVERR are used to
receive socket errors via a dedicated 'error queue' which can be
retrieved via recvmsg(). FreeBSD does not support this functionality.
For IPv4, the sysctl compat.linux.ignore_ip_recverr can be set to 1 to
silently ignore attempts to set IP_RECVERR and return success to the
application, which is wrong, but is required for (among other things)
a functional DNS client in recent versions of glibc.
Add support for ignoring IPV6_RECVERR, controlled by the same sysctl.
This fixes DNS in Linux when using IPv6 resolvers.
Reviewed by: imp, Jose Luis Duran
Pull Request: https://github.com/freebsd/freebsd-src/pull/1118
Brooks Davis [Mon, 22 Apr 2024 19:13:35 +0000 (20:13 +0100)]
libsys/arm: make vfork less of an outlier
Following the pattern of other architectures, make __sys_vfork the main
symbol and _vfork and vfork weak aliases. This avoids the need to
special case vfork in the generated list of symbols.
Brooks Davis [Mon, 22 Apr 2024 18:24:26 +0000 (19:24 +0100)]
libthr: always use __libc_interposing_slot()
Use __libc_interposing_slot() in favor of __libsys_interposing_slot() so
that the interposing interface is entierly between libc and libthr with
libsys only involved as an implementation detail.
Alan Somers [Mon, 22 Apr 2024 18:03:40 +0000 (12:03 -0600)]
geli: add a read-only kern.geom.eli.use_uma_bytes sysctl
It reports the value of the g_eli_alloc_sz variable. Allocations of
this size or less will use UMA. Larger allocations will use malloc.
Since malloc is slower, it is useful for users to know this variable so
they can avoid such allocations. For example, ZFS users can set
vfs.zfs.vdev.aggregation_limit to this value.
Joseph Mingrone [Mon, 22 Apr 2024 19:04:13 +0000 (16:04 -0300)]
adduser: Fix a syntax error
Reported by: Jose Luis Duran <jlduran@gmail.com>
Fixes: b37333899bd2
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D44871
Ed Maste [Mon, 22 Apr 2024 17:54:52 +0000 (13:54 -0400)]
mfc-candidates: use stable/14 as default MFC-to branch
This tool is typically invoked from within a working tree containing the
desired MFC target branch, which is detected automatically. It also has
a default for when it is invoked from elsewhere. Switch the default to
the most recent branch, stable/14.
Clean this up by adding a couple of macros at the top of the file that
are conditionally defined based on whether INET and/or INET6 are enabled,
which makes the code more readable and easier to maintain.
Mark Johnston [Mon, 22 Apr 2024 15:48:00 +0000 (11:48 -0400)]
ng_hci: Add sockaddr validation to sendto()
ng_btsocket_hci_raw_send() wasn't verifying that the destination address
specified by sendto() is large enough to fill a struct sockaddr_hci.
Thus, when copying the socket address into an mbuf,
ng_btsocket_hci_raw_send() may read past the end of the input sockaddr
while copying.
In practice this is effectively harmless since
ng_btsocket_hci_raw_output() only uses the address to identify a
netgraph node.
Reported by: Oliver Sieber <oliver@secfault-security.com>
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Mark Johnston [Mon, 22 Apr 2024 15:43:17 +0000 (11:43 -0400)]
thread: Simplify sanitizer integration with thread creation
fork() may allocate a new thread in one of two ways: from UMA, or cached
in a freed proc that was just allocated from UMA. In either case, KASAN
and KMSAN need to initialize some state; in particular they need to
initialize the shadow mapping of the new thread's stack.
This is done differently between KASAN and KMSAN, which is confusing.
This patch improves things a bit:
- Add a new thread_recycle() function, which moves all kernel stack
handling out of kern_fork.c, since it doesn't really belong there.
- Then, thread_alloc_stack() has only one local caller, so just inline
it.
- Avoid redundant shadow stack initialization: thread_alloc()
initializes the KMSAN shadow stack (via kmsan_thread_alloc()) even
through vm_thread_new() already did that.
- Add kasan_thread_alloc(), for consistency with kmsan_thread_alloc().
Doug Moore [Sun, 21 Apr 2024 22:24:29 +0000 (17:24 -0500)]
riscv: create a convenience composite macro
Define PTE_TO_VM_PAGE to compose the PHYS_TO_VM_PAGE and PTE_TO_PHYS
macros. Use it where appropriate, and drop some variables that it
makes unnecessary.
HP van Braam [Tue, 16 Apr 2024 23:01:20 +0000 (01:01 +0200)]
Enable L1SS handling on RPI4 pcib
Thanks to @kevans91 for pointing me in the right direction. FreeBSD had
the same bug as Linux (see
https://bugzilla.kernel.org/show_bug.cgi?id=217276) where the ultimate
solution was to honor the brcm,enable-l1ss FDT property.
In current versions of the dtb files this property has been added by
default.
Without this on many, many pcie addin cards the pcib will Serror when
trying to assert the clreq# pin on the pcie bus. Many cards do not have
these hooked up.
PR: 260131, 277638, 277605
Reviewed-by: emaste Signed-off-by: HP van Braam <hp@tmm.cx>
Pull-request: https://github.com/freebsd/freebsd-src/pull/1179
Enji Cooper [Sat, 20 Apr 2024 19:12:50 +0000 (12:12 -0700)]
OpenSSL: use the upstream provided version.map files for the fips/legacy providers
This change introduces a static copy of the fips and legacy linker version maps
generated by the OpenSSL 3.0.13 build process.
This unbreaks the fips and legacy providers by not exposing unnecessary
symbols from the fips/legacy provider shared objects shared with other
providers (base, default) and libcrypto.
More discussion:
Prior to this change, loading the fips provider indirectly from a
FreeBSD 14.0-CURRENT and 15.0-CURRENT host would result in a
process-wide deadlock when invoking select OpenSSL APIs
(CONF_modules_load* in this particular example).
Speaking with the upstream maintainers [1], it became obvious that
the FreeBSD base system was incorrectly building/linking the fips
provider, resulting in a symbol collision at runtime, and thus a
process-wide deadlock in specific circumstances. The fips provider
would deadlock when trying to acquire a write lock on internal
structures which should have only been available to the base and
default providers, as certain preprocessor ifdefs only allow specific
internal calls to be made with the base and default providers.
Enji Cooper [Sun, 21 Apr 2024 16:27:48 +0000 (09:27 -0700)]
bsd.subdir.mk: improve `SUBDIR.${MK_FOO}` advice
- Remove superfluous whitespace by removing trailing whitespace
before `\` (line continuation character)
- Quote `SUBDIR.` to clarify the fact that this is a variable
reference--not the end of a sentence.
Ed Maste [Tue, 16 Apr 2024 16:49:05 +0000 (12:49 -0400)]
mfc-candidates: move pretty printing into lua
d51c59002367 moved the MFC hash matching logic into a lua utility
script but left the output formatting in the shell script. Simplify this
slightly by just printing the formatted output from lua.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D44836
The general idea looked good to me. In particular, it allowed to save
some memory and avoid memory allocation failures when a large buffer
size was requested along with ring and fill policies.
But I didn't take into account that the second, supposedly unused
buffer, was actually used as the scratch buffer. The scratch buffer is
used as a temporary space for DTrace subroutines like copyin, copyinstr,
and alloca.
I think that the change can be fixed by allocating a separate smaller
buffer for the scratch buffer, but that fix would require more work than
I am able to do now. Hence the revert.
One of the comments in ahc_execute_scb() notes that the CAM direction is
actually w.r.t. the initiator. As a consequence, all of our sync ops
end up being wrong because the direction is flipped from that of the
transfer. Fix it to do proper invalidation and avoid spewing random
garbage out on the SCSI bus.
Reported and tested by: HP van Braam <hp@tmm.cx>
Reviewed by: imp, mav
Differential Revision: https://reviews.freebsd.org/D44860
In ahc_init(), qoutfifo is already assigned to effectively the same
value a couple lines up, except in the first assignment it uses the
proper definition; keep the more descriptive assignment.
ahc_targetcmd_offset() gets the offset wrong entirely; as per the
area of ahc_init() this diff also touches, targetcmds is laid out first
in the shared map and it's followed by the qoutfifo. As a result, we'd
generally be getting negative offsets here. We can't actually do a
partial sync anyways, so there was no consequence to getting this wrong.
Mark Johnston [Sat, 20 Apr 2024 16:02:06 +0000 (12:02 -0400)]
graid3: Fix teardown in g_raid3_try_destroy()
Commit 33cb9b3c3a22 replaced a g_raid3_destroy_device() call with a
g_raid3_free_device() call, which was incorrect and could lead to a
panic if a RAID3 GEOM failed to start (e.g., due to missing disks).
Reported by: graid3 tests
Fixes: 33cb9b3c3a22 ("graid3: Fix teardown races")
MFC after: 3 days
Sponsored by: Klara, Inc.
Mark Johnston [Sat, 20 Apr 2024 16:01:28 +0000 (12:01 -0400)]
wg: Add netmap support
When in netmap (emulated) mode, wireguard interfaces prepend or strip a
dummy ethernet header when interfacing with netmap. The netmap
application thus sees unencrypted, de-encapsulated frames with a fixed
header.
In this mode, netmap hooks the if_input and if_transmit routines of the
ifnet. Packets from the host TX ring are handled by wg_if_input(),
which simply hands them to the netisr layer; packets which would
otherwise be tunneled are intercepted in wg_output() and placed in the
host RX ring.
The "physical" TX ring is processed by wg_transmit(), which behaves
identically to wg_output() when netmap is not enabled, and packets
appear in the "physical" RX ring by hooking wg_deliver_in().
Stefan Eßer [Sat, 20 Apr 2024 06:30:35 +0000 (08:30 +0200)]
ldconfig: document changed default byte-order
Document that the hints file is created in little-endian byte-order
by default now. The -B option can be used to create a big-endian hints
file. Currently, both formats are accepted by the run-time linker on
all architectures.
- if the kernel is configured with a packet filter which drops packets
by default, pfctl or ipfw will be required in the rescue environment
to make the network functional.
- rescue's stated purpose is to be useful for small/embedded systems
(and is also quite useful for small jails); a rescue-based router
would want these binaries.
On amd64, this increases the size of rescue from 17346200 to 17907248,
or 561048 bytes (3.2%).
usr.bin/units: use else..if to avoid calling the next branch
Even if the first branch succeeds, next time it will still check for
the second branch (which will be false) as the first one was true.
Add an else..if statement to address this.
projects / FreeBSD / FreeBSD.git / log