T5 and above have extra bits for the optional filter fields. This is a
correctness issue and not just a waste because a filter mode valid on a
T4 (36b) may not be valid on a T5+ (40b).
MFC after: 2 weeks
Sponsored by: Chelsio Communications
Allow the filter mask (aka the hashfilter mode when hashfilters are
in use) to be set any time it is safe to do so. The requested mask
must be a subset of the filter mode already. The driver will not change
the mode or ingress config just to support a new mask.
MFC after: 2 weeks
Sponsored by: Chelsio Communications
1. Query the firmware for filter mode, mask, and related ingress config
instead of trying to figure them out from hardware registers. Read
configuration from the registers only when the firmware does not
support this query.
2. Use the firmware to set the filter mode. This is the correct way to
do it and is more flexible as well. The filter mode (and associated
ingress config) can now be changed any time it is safe to do so.
The user can specify a subset of a valid mode and the driver will
enable enough bits to make sure that the mode is maxed out -- that
is, it is not possible to set another bit without exceeding the
total width for optional filter fields. This is a hardware
requirement that was not enforced by the driver previously.
MFC after: 2 weeks
Sponsored by: Chelsio Communications
jail_attach(2) performs an internal chroot operation, leaving it up to
the calling process to assure the working directory is inside the jail.
Add a matching internal chdir operation to the jail's root. Also
ignore kern.chroot_allow_open_directories, and always disallow the
operation if there are any directory descriptors open.
Reported by: mjg
Approved by: markj, kib
MFC after: 3 days
In commit 38bfc6dee33b we added an IFDI_DETACH() call to
iflib_pseudo_deregister() since it looked like it was missing. One is
present in the error-handling path of iflib_pseudo_register(). However,
the detach actually comes from the DEVICE_DETACH() method for the
above-mentioned device_t, so now we're calling IFDI_DETACH() twice when
destroying a pseudo interface.
Fix the problem by not calling IFDI_DETACH() from the device detach
routine. This way we can ensure that iflib de-initialization always
happens in a consistent order. It also ensures that you can't do silly
things like "devctl detach <pseudo ifnet>", which would previously
detach the driver without tearing down the corresponding ifnet.
PR: 253541
Reviewed by: erj
MFC after: 1 week
Fixes: 38bfc6dee33b
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D28774
In 0ee0dbfb0d26cf4bc37f24f12e76c7f532b0f368 I imported a more recent
libcxxrt snapshot, which includes an upstream fix for the padding of
struct _Unwind_Exception:
e458560b7e
However, we also had a similar fix in our tree as:
https://cgit.freebsd.org/src/commit/?id=3c4fd2463bb29f65ef1404011fcb31e508cdf2e2
Since having both fixes makes the struct too large again, it leads to
SIGBUSes when throwing exceptions on amd64 (or other LP64 arches). This
is most easily tested by running kyua without any arguments.
It looks like our fix is no longer needed now, so revert it to reduce
diffs against upstream.
PR: 253226
Reviewed by: arichardson, kp
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D28799
Changes in the 2fe5a79425c7 moved dst sockaddr masking from the
routing control plane to the rtsock code.
It broke arp/ndp deletion.
It turns out, arp/ndp perform RTM_GET request first to get an
interface index necessary for the deletion.
Then they simply stamp the reply with RTF_LLDATA and set the
command to RTM_DELETE.
As a result, kernel receives request with non-empty RTA_NETMASK
and clears RTA_DST host bits before passing the message to the
lla code.
De facto, the only needed bits are RTA_DST, RTA_GATEWAY and the
subset of rtm_flags.
With that in mind, fix the interace by clearing RTA_NETMASK
for every messages with RTF_LLDATA.
While here, cleanup arp/ndp code a bit.
MFC after: 1 day
Reviewed by: gnn
Differential Revision: https://reviews.freebsd.org/D28804
Use appropriate types for the physical and virtual framebuffer address.
Fixes framebuffers mapped above 4G physical on 32-bit systems that
support physical address extensions like i386 and Book-E powerpc.
Patch developed by bdragon
Reviewed by: bdragon, luporl
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D28604
This fixes a warning from GCC for kernels without netmap since the
return value is never used.
Reviewed by: vmaffione, erj
Differential Revision: https://reviews.freebsd.org/D28598
Switch OOA queue from TAILQ to LIST and change its direction, so that
we traverse it forward, not backward. There is only one place where
we really need other direction, and it is not critical.
Use STAILQ_REMOVE_HEAD() instead of STAILQ_REMOVE() in backends.
Replace few impossible conditions with assertions.
MFC after: 1 month
high-resolution nanosecond timestamp used for the DTrace 'timestamp'
built-in variable. The new implementation uses the EL0 cycle
counter and frequency registers in ARMv8-A. This replaces a
previous implementation that relied on an instrumentation-safe
implementation of getnanotime(), which provided only timer
resolution.
MFC after: 3 days
Reviewed by: andrew, bsdimp (older version)
Useful comments appreciated: jrtc27, emaste
- Implements little-endian support (powerpc64le)
- Adds 'hw.ofwfb.physaddr' kernel parameter so user can manually
provide correct address if it's not detected correctly
- Adds 'hw.ofwfb.argb32_pixel' so user can set it manually if
colors are inverted due to incorrect pixel format (default = 1)
- Automatically selects RGBA32 pixel format if NVidia graphic adapter
is detected (sets hw.ofwfb.argb32_pixel=0)
Machines equipped with NVidia graphic adapters tend to use RGBA32
pixel format. By default ARGB32 pixel format is used, proved to work
on machines equipped with ATI graphic adapter and the onboard adapter
used on Talos II and Blackbird machines from Raptor Computing Systems.
Original patch developed by bdragon
Reviewed by: bdragon, luporl
MFC after: 3 days
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D28604
Under some circumstances, PRR may end up with a fully
collapsed cwnd when finalizing the loss recovery.
Reviewed By: #transport, kbowling
Reported by: Liang Tian
MFC after: 1 week
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D28780
Packages default to ending up in a different location compared to the
documentation, so catch up to the implementation by referring to the
location where packages can usually be found if no environment variables
have been set.
While here, also update the mention of the file extension to match the
txz format that packages use.
PR: 253179, 224370
Reported by: rwatson, jeromer at fastmail dotnet
TCP_LINGERTIME can be traced back to BSD 4.4 Lite and perhaps beyond, in
exactly the same form that it appears here modulo slightly different
context. It used to be the case that there was a single pr_usrreq
method with requests dispatched to it; these exact two lines appeared in
tcp_usrreq's PRU_ATTACH handling.
The only purpose of this that I can find is to cause surprising behavior
on accepted connections. Newly-created sockets will never hit these
paths as one cannot set SO_LINGER prior to socket(2). If SO_LINGER is
set on a listening socket and inherited, one would expect the timeout to
be inherited rather than changed arbitrarily like this -- noting that
SO_LINGER is nonsense on a listening socket beyond inheritance, since
they cannot be 'connected' by definition.
Neither Illumos nor Linux reset the timer like this based on testing and
inspection of Illumos, and testing of Linux.
Reviewed by: rscheff, tuexen
Differential Revision: https://reviews.freebsd.org/D28265
Introduce new CTL core KPI ctl_run(), preprocessing I/Os in the caller
context instead of scheduling another thread just for that. This call
may sleep, that is not acceptable for some frontends like the original
CAM/FC one, but iSCSI already has separate sleepable per-connection RX
threads, and another thread scheduling is mostly just a waste of time.
IOCTL frontend actually waits for the I/O completion in the caller
thread, so the use of another thread for this has even less sense.
With this change I can measure ~5% IOPS improvement on 4KB iSCSI I/Os
to ZFS.
MFC after: 1 month
df093aa9463b linked against libprivateauditd.a, but that is currently
(and incorrectly) built as position-dependent. For now just force PIE
off for this test to fix the WITH_PIE build.
Sponsored by: The FreeBSD Foundation
The ACPI NFIT specification defines a set of "NVDIMM State Flags". These
flags are already reported by `acpidump -t', but this change makes them
available on a per-device basis, in a format that is more easily parsed.
To simplify this, introduce acpi_nfit_get_memory_maps_by_dimm(), which
locates the (ACPI_NFIT_MEMORY_MAP)s associated with a given
(nfit_handle_t).
Reviewed by: mav, cem
Tested by: mav, rpokala (version for stable/12)
MFC after: 3 days
Sponsored by: Panasas
The kernel changes needed for nfs-over-tls have been committed to main.
However, nfs-over-tls requires user space daemons to handle the
TLS handshake and other non-application data TLS records.
There is one daemon (rpc.tlsclntd) for the client side and one daemon
(rpc.tlsservd) for the server side, although they share a fair amount
of code found in rpc.tlscommon.c and rpc.tlscommon.h.
They use a KTLS enabled OpenSSL to perform the actual work and, as such,
are only built when MK_OPENSSL_KTLS is set.
Communication with the kernel is done via upcall RPCs done on AF_LOCAL
sockets and the custom system call rpctls_syscall.
Reviewed by: gbe (man pages only), jhb (usr.sbin/Makefile only)
Comments by: jhb
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D28430
Relnotes: yes
BUS_DMA_NOCACHE should only be used when one needs to guarantee the
created mapping has uncached memory attributes, usually as a result
of buggy hardware. Normal use cases should pass BUS_DMA_COHERENT, to
create an appropriate mapping based on the flags passed to
bus_dma_tag_create().
This should have no functional change, since the DMA tags in this driver
are created without the BUS_DMA_COHERENT flag.
Reported by: mmel
Reviewed by: mmel, Thomas Skibo <thomas-bsd@skibo.net>
MFC after: 3 days
Add three examples showing the use of `-h`, `-l`, `-t`, `-w`
christos@netbsd.org to be notified.
Reviewed by: bcr@, gbe@, imp@
Approved by: bcr@, gbe@ (mentor), imp@
Differential Revision: https://reviews.freebsd.org/D25242
This supports Chacha20-Poly1305 for both send and receive for TLS 1.2
and for send in TLS 1.3.
Reviewed by: gallatin
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27841
Chacha20-Poly1305 for TLS is an AEAD cipher suite for both TLS 1.2 and
TLS 1.3 (RFCs 7905 and 8446). For both versions, Chacha20 uses the
server and client IVs as implicit nonces xored with the record
sequence number to generate the per-record nonce matching the
construction used with AES-GCM for TLS 1.3.
Reviewed by: gallatin
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27839
- Make openssl_gcm_encrypt generic to AEAD ciphers (aside from CCM)
and use it for Chacha20-Poly1305.
- Use generic AEAD control constants instead of GCM/CCM specific names.
Reviewed by: cem
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27838
This uses the chacha20 IETF and poly1305 implementations from
libsodium. A seperate auth_hash is created for the auth side whose
Setkey method derives the poly1305 key from the AEAD key and nonce as
described in RFC 8439.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27837
Note that this algorithm implements the mode defined in RFC 8439.
Reviewed by: cem
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27836
This script uses -e, so it's prone to exiting in awkward places. In
particular, if arc diff fails, the script just exits without restoring
the checkout.
Mitigate this for now by using a global variable to record the previous
checkout and use a trap handler to restore it in the face of errors. A
better solution might be to use arc diff's --head parameter but that
will require more testing.
Reported by: kevans
Reviewed by: jhb
Differential Revision: https://reviews.freebsd.org/D28631
The output has been broken since 1b6dd6d772ca. Casting to uintmax_t
before the call to printf is necessary to ensure that 32-bit addresses
are interpreted correctly.
PR: 243236
MFC after: 3 days
The scheduler uses this topology to try and preserve locality when
migrating threads between CPUs and when performing work stealing.
Ensure that on NUMA systems it will at least take the NUMA topology into
account.
Reviewed by: mmel
Submitted by: Klara, Inc.
Sponsored by: Ampere Computing
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D28579
Per the i2c spec, a slave device can stretch SCL idefinitely, so 25ms is
a bit arbitrary in general. smbus does specify an optional timeout
recovery mechanism to be done at about 25~35ms, but the IPMI SSIF spec
says that BMCs don't have any obligation to implement that.
The BMC on Altra seems to mostly respond within 25ms, but occasionally
will stretch SCL for ~300 msec.
Also, the count_us mechanism seems to actually timeout around 25%
earlier than it would claim (timeout really happening around 19ms
instead of 25ms).
Sponsored by: Ampere Computing LLC
Submitted by: Klara Inc.
Reviewed by: manu, imp
Differential Revision: https://reviews.freebsd.org/D28747
Some ibnd_* manpages source other manpages from a `man3/` directory when it
should reference the pages in the current directory.
Instead of modifying contributing sources and using `.so` (discouraged by
mandoc(1)) use MLINKS in the proper Makefile and do not install the affected
manpages.
PR: 237693
Reported by: wosch@FreeBSD.org
Reviewed by: gbe@ (mentor) yuripv@
Approved by: gbe@ (mentor) yuripv@
Differential Revision: https://reviews.freebsd.org/D28727
Notable upstream changes:
bf156c966 Remove unused abd_alloc_scatter_offset_chunkcnt
658fb8020 Add "compatibility" property for zpool feature sets
This update introduces a new pool property called "compatibility"
that can be used to enable a limited set of pool features on pool
creation and "stick" to it, so the "zpool upgrade" does not
accidentally enable features that are not desired. The value of
this property may then be changed later.
See zpool-features(5) for more information about the "compatibility"
pool property.
Obtained from: OpenZFS
MFC after: 2 weeks
KCSAN complains about racy accesses in the locking code. Those races are
fine since they are inside a TD_SET_RUNNING() loop that expects the value
to be changed by another CPU.
Use relaxed atomic stores/loads to indicate that this variable can be
written/read by multiple CPUs at the same time. This will also prevent
the compiler from doing unexpected re-ordering.
Reported by: GENERIC-KCSAN
Test Plan: KCSAN no longer complains, kernel still runs fine.
Reviewed By: markj, mjg (earlier version)
Differential Revision: https://reviews.freebsd.org/D28569
In the CheriBSD CI we reproducibly see the first test in sys/audit
(administrative:acct_failure) fail due to a missing startup message.
It appears this is caused by a race condition when starting auditd:
`service auditd onestart` returns as soon as the initial auditd() parent
exits (after the daemon(3) call).
We can avoid this problem by setting up the auditd infrastructure
in-process: libauditd contains audit_quick_{start,stop}() functions that
look like they are ideally suited to this task.
This patch also avoids forking lots of shell processes for each of the 418
tests by using `auditon(A_SENDTRIGGER, &trigger, sizeof(trigger))` to check
for a running auditd(8) instead of using `service auditd onestatus`.
With these two changes (and D28388 to fix the XFAIL'd test) I can now
boot and run `cd /usr/tests/sys/audit && kyua test` without any failures
in a single-core QEMU instance. Before there would always be at least one
failed test.
Besides making the tests more reliable in CI, a nice side-effect of this
change is that it also significantly speeds up running them by avoiding
lots of fork()/execve() caused by shell scripts:
Running kyua test on an AArch64 QEMU took 315s before and now takes 68s,
so it's roughly 3.5 times faster. This effect is even larger when running
on a CHERI-RISC-V QEMU since emulating CHERI instructions on an x86 host
is noticeably slower than emulating AArch64.
Test Plan: aarch64+amd64 QEMU no longer fail.
Reviewed By: asomers
Differential Revision: https://reviews.freebsd.org/D28451
This causes problems when using ASAN with a runtime older than 12.0 since
the intercept does not expect qsort() to call itself using an interposable
function call. This results in infinite recursion and stack exhaustion
when a binary compiled with -fsanitize=address calls qsort.
See also https://bugs.llvm.org/show_bug.cgi?id=46832 and
https://reviews.llvm.org/D84509 (ASAN runtime patch).
To prevent this problem, this patch uses a static helper function
for the actual qsort() implementation. This prevents interposition and
allows for direct calls. As a nice side-effect, we can also move the
qsort_s checks to the top-level function and out of the recursive calls.
Reviewed By: kib
Differential Revision: https://reviews.freebsd.org/D28133
Since 4581cefc1e
ATF opens the results file on startup. This fixes problems like
capsicumized tests not being able to open the file on exit.
However, this test closes all file descriptors above 3 to get a
deterministic fd table allocation for the child. Instead of using closefrom
(which will close the ATF output file FD) I've changed this test use
the lowest available fd and pass that to the helper program as a string.
We could also try to re-open the results file in ATF if we get a EBADF
error, but that will fail when running under Capsicum.
Reviewed By: cem
Differential Revision: https://reviews.freebsd.org/D28684
In the new ENA-based instances like c6gn, the vector table moved to a
new PCIe bar - BAR1. Previously it was always located on the BAR0, so
the resources were already allocated together with the registers.
As the FreeBSD isn't doing any resource allocation behind the scenes,
the driver is responsible to allocate them explicitly, before other
parts of the OS (like the PCI code allocating MSIx) will be able to
access them.
To determine dynamically BAR on which the MSIx vector table is present
the pci_msix_table_bar() is being used and the new BAR is allocated if
needed.
Submitted by: Michal Krawczyk <mk@semihalf.com>
Obtained from: Semihalf
Sponsored by: Amazon, Inc
MFC after: 3 days
