This fixes PCI devices not being found on QEMU ppce500. This
generic board used to have its first PCI slot at 0x11, like the
mpc8544dsi and some real HW. After commit [1], it was changed to
0x1 and our driver wasn't prepared for that.
Allan Jude [Thu, 4 Nov 2021 12:55:33 +0000 (12:55 +0000)]
Allow kern.ipc.maxsockets to be set to current value without error
Normally setting kern.ipc.maxsockets returns EINVAL if the new value
is not greater than the previous value. This can cause spurious
error messages when sysctl.conf is processed multiple times, or when
automation systems try to ensure the sysctl is set to the correct
value. If the value is unchanged, then just do nothing.
firk [Tue, 15 Mar 2022 22:22:21 +0000 (18:22 -0400)]
clock_gettime: Fix CLOCK_THREAD_CPUTIME_ID race
Use a spinlock section instead of a critical section to synchronize with
statclock(). Otherwise the CLOCK_THREAD_CPUTIME_ID clock can appear to
go backwards.
Mark Johnston [Wed, 16 Mar 2022 16:09:17 +0000 (12:09 -0400)]
hdac: Handle interrupts racing with device suspend
- Avoid looping forever if a concurrent reset causes a read of the
interrupt status register to return all ones.
- Lock the softc before reading the interrupt status, so as to avoid a
similar infinite loop in hdac_one_intr().
This fixes suspend-to-S3 on some laptops.
PR: 261207
Reviewed by: mav, imp
Tested by: uqs
Sponsored by: The FreeBSD Foundation
Ed Maste [Thu, 17 Mar 2022 19:47:55 +0000 (15:47 -0400)]
hda: add patch for Framework laptop headphone jack
For Framework laptops built after Oct 2021 (like mine) that have a Tempo
Semiconductor 92HD95B codec rather than Realtek ALC295 (see
https://frame.work/ca/en/blog/solving-for-silicon-shortages).
As we do for many other laptops, put the headphone jack and speakers in
the same association by default so that the generic sound device
automatically switches between them. Also specify the jack colour and
location.
Reviewed by: markj
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D34596
K Staring [Sat, 3 Jul 2021 06:15:49 +0000 (00:15 -0600)]
hdaa: update pin patch configurations
A number of structural changes:
- Use decimal nid numbers instead of hex
- updated the branch to incoorporate the suggestions made in the
ALC280 pull request github thread
- Convert magic pin values into strings.
- Also update hdaa_patches to use clearer enums..
- made pin patch type enum clearer, add macro for 'string' type
patches
- Added pin_patch structures to separate data from logic.
- Integrated Realtek patches into new structure.
These incorporate fixes for ALC255, ALC256, ALC260, ALC262, ALC268,
ALC269, ALC280, ALC282, ALC283, ALC286, ALC290, ALC293, ALC296, ALC2880
And have definitions for a number of Dell and HP laptops.
Much of this data has been mined fromt he tables in the Linux driver.
imp squashed these into one commit because the changes from the github
pull requests no longer cleanly apply individually and made light style
changes after feedback from jhb.
Dimitry Andric [Sat, 19 Mar 2022 23:12:58 +0000 (00:12 +0100)]
Fix llvm build after 1b3bef43e3cb, due to API change
After merging llvm commit b9ca73e1a8fd for PR 262608, it would fail to
compile with:
/usr/src/contrib/llvm-project/llvm/lib/IR/Operator.cpp:197:22: error: no member named 'isZero' in 'llvm::APInt'
if (!IndexedSize.isZero()) {
~~~~~~~~~~~ ^
Upstream refactored their APInt class, and isZero() was one of the newer
methods which did not yet exist in llvm 13.0.0. Fix this by using the
older but equivalent isNullValue() method instead.
Dimitry Andric [Sat, 19 Mar 2022 19:59:04 +0000 (20:59 +0100)]
Apply llvm fix for assertion compiling certain versions of Wine
Merge commit b9ca73e1a8fd from llvm git (by Stephen Tozer):
[DebugInfo] Correctly handle arrays with 0-width elements in GEP salvaging
Fixes an issue where GEP salvaging did not properly account for GEP
instructions which stepped over array elements of width 0 (effectively a
no-op). This unnecessarily produced long expressions by appending
`... + (x * 0)` and potentially extended the number of SSA values used
in the dbg.value. This also erroneously triggered an assert in the
salvage function that the element width would be strictly positive.
These issues are resolved by simply ignoring these useless operands.
There's too many broken hardware out there that wrongly has the
ACPI_FADT_NO_VGA bit set. Ignore it unless running as a virtualized
guest, as then the expectation would be that the hypervisor does
provide correct ACPI tables.
Reviewed by: emaste, 0mp, eugen
Sponsored by: Citrix Systems R&D
PR: 230172
/usr/freebsd-dist is used used by various programs as the location for
FreeBSD distribution files. In-tree programs following this convention
are bsdinstall(8) and release(7).
With this merge, we return to our previous long-standing practice of
distributing the IANA Time Zone Database unmodified.
Releases of tzdb since 2021b have merged some time zones where clocks
have agreed since 1970. The overwhelming majority of users will not be
affected by this change. A port of the newly created global-tz fork of
the IANA Time Zone database (misc/global-tz) is available for users who
need more granular pre-1970 time zone history.
zfskeys: Support autoloading of keys stored on ZFS
The zfskeys service script starts before the zfs service script, so that
dataset decryption keys are available when `zfs mount -a` is run. One of
the potential edge cases of this design is that if a key is stored on
ZFS it won't be loaded until `zfs mount -a` is issued.
In order to address that let's try to load the additional keys and mount
related ZFS datasets after the zfs script finishes its standard mounting
procedure.
PR: 262468
Reported by: Graham Perrin <grahamperrin@gmail.com>
Reviewed by: allanjude
Approved by: allanjude (src)
Fixes: 33ff39796ffe Add zfskeys rc.d script for auto-loading encryption keys
MFC after: 3 days
Sponsored by: Modirum
Sponsored by: Klara Inc.
Differential Revision: https://reviews.freebsd.org/D34601
netmap: add a tunable for the maximum number of VALE switches
The new dev.netmap.max_bridges sysctl tunable can be set in
loader.conf(5) to change the default maximum number of VALE
switches that can be created. Current defaults is 8.
Kristof Provost [Thu, 17 Mar 2022 02:35:13 +0000 (03:35 +0100)]
if_epair: build fix
66acf7685b failed to build on riscv (and mips). This is because the
atomic_testandset_int() (and friends) functions do not exist there.
Happily those platforms do have the long variant, so switch to that.
The total size of the user-provided nmreq was first computed and then
trusted during the copyin. This might lead to kernel memory corruption
and escape from jails/containers.
Reported by: Lucas Leong (@_wmliang_) of Trend Micro Zero Day Initiative
Security: CVE-2022-23084
MFC after: 3 days
An unsanitized field in an option could be abused, causing an integer
overflow followed by kernel memory corruption. This might be used
to escape jails/containers.
Reported by: Reno Robert and Lucas Leong (@_wmliang_) of Trend Micro
Zero Day Initiative
Security: CVE-2022-23085
Warner Losh [Sun, 9 Jan 2022 17:13:58 +0000 (10:13 -0700)]
devd: remove the scsi regexp
We don't really use the scsi regexp for anything. The rescan was a
workaround that was fixed a long time ago and has been disabled for
ages. And the regexp was incomplete.
Eugene Grosbein [Wed, 16 Mar 2022 04:41:51 +0000 (11:41 +0700)]
virtio_random(8): MFC: avoid deadlock at shutdown time (regression fix)
FreeBSD 13+ running as virtual guest may load virtio_random(8) driver
by means of devd(8) unless the driver is blacklisted or disabled
via device.hints(5). Currently, the driver may prevent
the system from rebooting or shutting down correctly.
This change deactivates virtio_random at very late stage
during system shutdown sequence to avoid deadlock
that results in kernel hang.
Mark Johnston [Wed, 16 Feb 2022 02:50:41 +0000 (21:50 -0500)]
armv8crypto: Use cursors to access crypto buffer data
Currently armv8crypto copies the scheme used in aesni(9), where payload
data and output buffers are allocated on the fly if the crypto buffer is
not virtually contiguous. This scheme is simple but incurs a lot of
overhead: for an encryption request with a separate output buffer we
have to
- allocate a temporary buffer to hold the payload
- copy input data into the buffer
- copy the encrypted payload to the output buffer
- zero the temporary buffer before freeing it
We have a handy crypto buffer cursor abstraction now, so reimplement the
armv8crypto routines using that instead of temporary buffers. This
introduces some extra complexity, but gallatin@ reports a 10% throughput
improvement with a KTLS workload without additional CPU usage. The
driver still allocates an AAD buffer for AES-GCM if necessary.
Mark Johnston [Wed, 16 Feb 2022 02:45:32 +0000 (21:45 -0500)]
opencrypto: Add a routine to copy a crypto buffer cursor
This was useful in converting armv8crypto to use buffer cursors. There
are some cases where one wants to make two passes over data, and this
provides a way to "reset" a cursor.
Ed Maste [Fri, 11 Mar 2022 19:27:46 +0000 (14:27 -0500)]
loader: accept "yellow" as a named color
For historical reasons console color number 3 may be either yellow (most
consoles) or brown (VGA palette). The console escape code standard
uses "yellow", but teken color name constants appear to be based on the
VGA scheme and use TC_BROWN for color 3. Even so, the palette table
used 50,50,0 as the RGB percentage tuple, resulting in a dim yellow for
framebuffer consoles at the time teken was introduced.
Amusingly, in 19e2ce2d8367 the comment on the palette entry was changed
from "brown" to "dark yellow" but the colour itself was changed from
a pure yellow to being somewhat brown.
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
usbtest: Fix issue when multiple devices are sharing same USB vendor and product ID.
When there are multiple devices sharing the same USB vendor and product ID,
the wrong device may be selected. Fix this by also matching the bus and
device address, ugen<X>.<Y> .
Ed Maste [Mon, 7 Mar 2022 19:17:01 +0000 (14:17 -0500)]
dumpon: use underlying device if encrypted swap is in use
/etc/rc.d/dumpon runs before /etc/rc.d/swap. When encrypted swap is in
use the .eli or .bde device will not exist at the time dumpon runs.
Even if this is addressed it does not make sense to dump core to
encrypted swap, as the encryption key will not be available after
reboot rendering the dump useless. Thus, for the case that dumpdev=AUTO
and encrypted swap is in use, strip the extension and use the underlying
device.
Emit a warning if we are using the underlying device and the user has not
configured dump encryption, so that the user knows that the will not be
encrypted.
PR: 238301
Reported by: Ivan Rozhuk
Reviewed by: jilles
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D34474
Michael Gmelin [Wed, 16 Mar 2022 22:08:55 +0000 (23:08 +0100)]
if_epair: fix race condition on multi-core systems
As an unwanted side effect of the performance improvements in 24f0bfbad57b9, epair interfaces stop forwarding traffic on higher
load levels when running on multi-core systems.
This happens due to a race condition in the logic that decides when to
place work in the task queue(s) responsible for processing the content
of ring buffers.
In order to fix this, a field named state is added to the epair_queue
structure. This field is used by the affected functions to signal each
other that something happened in the underlying ring buffers that might
require work to be scheduled in task queue(s), replacing the existing
logic, which relied on checking if ring buffers are empty or not.
epair_menq() does:
- set BIT_MBUF_QUEUED
- queue mbuf
- if testandset BIT_QUEUE_TASK:
enqueue task
epair_tx_start_deferred() does:
- swap ring buffers
- process mbufs
- clear BIT_QUEUE_TASK
- if testandclear BIT_MBUF_QUEUED
enqueue task
PR: 262571
Approved by: re (gjb, early MFC)
Reported by: Johan Hendriks <joh.hendriks@gmail.com>
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D34569
Mark Johnston [Tue, 1 Mar 2022 16:53:42 +0000 (11:53 -0500)]
fasttrap: Avoid creating WX mappings
fasttrap instruments certain instructions by overwriting them and
copying the original instruction to some per-thread scratch space which
is executed after the probe fires. This trampoline jumps back to the
tracepoint after executing the original instruction.
The created mapping has both write and execute permissions, and so this
mechanism doesn't work when allow_wx is disabled. Work around the
restriction by using proc_rwmem() to write to the trampoline.
Reviewed by: vangyzen
Tested by: Amit <akamit91@hotmail.com>
Sponsored by: The FreeBSD Foundation
Mark Johnston [Tue, 1 Mar 2022 16:48:39 +0000 (11:48 -0500)]
proc: Relax proc_rwmem()'s assertion on the process hold count
This reference ensures that the process and its associated vmspace will
not be destroyed while proc_rwmem() is executing. If, however, the
calling thread belongs to the target process, then it is unnecessary to
hold the process. In particular, fasttrap - a module which enables
userspace dtrace - may frequently call proc_rwmem(), and we'd prefer to
avoid the overhead of locking and bumping the hold count when possible.
Thus, make the assertion conditional on "p != curproc". Also assert
that the process is not already exiting. No functional change intended.
Andrew Turner [Tue, 8 Mar 2022 11:38:51 +0000 (11:38 +0000)]
Make the arm64 get_pcpu a function again
We assume the pointer returned from get_pcpu will be consistent even
if the thread is moved to a new CPU. Fix this by partially reverting 63c858a04d565 to make get_pcpu a function again.
Mark Johnston [Tue, 1 Mar 2022 14:07:14 +0000 (09:07 -0500)]
riscv: Add support for enabling SV48 mode
This increases the size of the user map from 256GB to 128TB. The kernel
map is left unchanged for now.
For now SV48 mode is left disabled by default, but can be enabled with a
tunable. Note that extant hardware does not implement SV48, but QEMU
does.
- In pmap_bootstrap(), allocate a L0 page and attempt to enable SV48
mode. If the write to SATP doesn't take, the kernel continues to run
in SV39 mode.
- Define VM_MAX_USER_ADDRESS to refer to the SV48 limit. In SV39 mode,
the region [VM_MAX_USER_ADDRESS_SV39, VM_MAX_USER_ADDRESS_SV48] is not
mappable.
Reviewed by: jhb
Sponsored by: The FreeBSD Foundation