]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
2 years agohier.7: Document /nonexistent
Mateusz Piotrowski [Thu, 30 Jun 2022 19:58:01 +0000 (21:58 +0200)]
hier.7: Document /nonexistent

MFC after: 1 week

(cherry picked from commit 606cdb2ef3a2574de0934f657eeb018ca6a470cc)

2 years agohier.7: Remove a.out lib directories
Mateusz Piotrowski [Thu, 30 Jun 2022 13:53:53 +0000 (15:53 +0200)]
hier.7: Remove a.out lib directories

Fixes: 76b4234537b2 BSD.usr.dist: remove aout
MFC after: 1 week

(cherry picked from commit d5a381f8c7bcf9ec391540158a3859d8bcd1d007)

2 years agopkg.7: Clean up mdoc
Mateusz Piotrowski [Thu, 30 Jun 2022 13:46:54 +0000 (15:46 +0200)]
pkg.7: Clean up mdoc

MFC after: 1 week

(cherry picked from commit 3beedc3840c29f3f728ed28cbc1450e1b9999f61)

2 years agosysctl.8: Reference sysctl(9)
Mateusz Piotrowski [Thu, 30 Jun 2022 13:34:58 +0000 (15:34 +0200)]
sysctl.8: Reference sysctl(9)

MFC after: 1 week

(cherry picked from commit 0438e9beb552b5e33c5d54748395429c6f9b8075)

2 years agobusdma_iommu: Fine-grained locking for the dmamap's map list
Alan Cox [Wed, 22 Jun 2022 21:51:47 +0000 (16:51 -0500)]
busdma_iommu: Fine-grained locking for the dmamap's map list

Introduce fine-grained locking on the dmamap's list of map entries,
replacing the use of the domain lock.  This is not the most significant
source of lock contention, but it is the easiest to address.

Reviewed by: kib
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D35557

(cherry picked from commit eeb46578c21ad37866f49f3bbb3ac738b44abbf6)

2 years agoiommu_gas: Eliminate unnecessary wrappers
Alan Cox [Tue, 14 Jun 2022 19:01:36 +0000 (14:01 -0500)]
iommu_gas: Eliminate unnecessary wrappers

Eliminate trivial wrappers for several iommu_gas functions that serve
no functional purpose.

Reviewed by: br, dougm, kib
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D35487

(cherry picked from commit fc2e4f15a9047bbf546cd675ed590b88e54362bd)

2 years agoiommu_gas: Fix a recent regression with IOMMU_MF_CANSPLIT
Alan Cox [Sun, 26 Jun 2022 16:48:12 +0000 (11:48 -0500)]
iommu_gas: Fix a recent regression with IOMMU_MF_CANSPLIT

As of 19bb5a7244ff, the IOMMU_MF_CANSPLIT case in iommu_gas_match_one()
must take into account the specified offset.  Otherwise, the recently
changed end calculation in iommu_gas_match_insert() could produce an
end address that crosses the specified boundary by one page.

Reviewed by: dougm
MFC with: 19bb5a7244ff

(cherry picked from commit 7b39a9bc1df37502e8186593f3427b7ff0e4cc71)

2 years agoiommu_gas: Tidy up
Alan Cox [Sun, 26 Jun 2022 05:44:47 +0000 (00:44 -0500)]
iommu_gas: Tidy up

Move a comment to the code that it describes.  Improve the wording.

Style fixes.

MFC after: 2 weeks

(cherry picked from commit da33f6d76b9d0c00649c692f6cb2b3b33bf5af30)

2 years agoiommu_gas: Add a missing error-case unlock
Alan Cox [Tue, 21 Jun 2022 04:48:31 +0000 (23:48 -0500)]
iommu_gas: Add a missing error-case unlock

Release the domain lock when iommu_gas_reserve_region_extend()'s call to
iommu_gas_reserve_region_locked() fails.

MFC after: 2 weeks

(cherry picked from commit 0ba1d8608234eee767b475627da6e5903ce7536a)

2 years agobusdma_iommu: Eliminate a redundant trunc_page()
Alan Cox [Mon, 20 Jun 2022 18:40:42 +0000 (13:40 -0500)]
busdma_iommu: Eliminate a redundant trunc_page()

Since OFF_TO_IDX() inherently truncates the given value, there is no
need to perform trunc_page() on it.

MFC after: 2 weeks

(cherry picked from commit 32e82bcc152783dfe0d03ffdd864cedfddbc80d7)

2 years agoiommu_gas: Eliminate a stale comment
Alan Cox [Mon, 20 Jun 2022 17:14:53 +0000 (12:14 -0500)]
iommu_gas: Eliminate a stale comment

As of 19bb5a7244ff, the size passed to iommu_gas_map is no longer
required to be a multiple of the CPU page size.

MFC after: 2 weeks

(cherry picked from commit 164491fb038724405b9e38355b569ef8dd4ad068)

2 years agoiommu_gas: Correct a broken KASSERT
Alan Cox [Fri, 17 Jun 2022 17:03:06 +0000 (12:03 -0500)]
iommu_gas: Correct a broken KASSERT

If iommu_gas_find_space() ever called iommu_gas_uppermatch(), and it
succeeded in allocating space, then the subsequent KASSERT would be
triggered.  Change that KASSERT to accept either success or ENOMEM.

MFC after: 1 week

(cherry picked from commit e6775534aee1963a39e5ee762b8eab1d7dfb1b6b)

2 years agoiommu_gas: make iommu_gas_lowermatch non-recursive
Doug Moore [Wed, 15 Jun 2022 16:32:56 +0000 (11:32 -0500)]
iommu_gas: make iommu_gas_lowermatch non-recursive

Change the recursive implementation to one that uses parent pointers
to walk back up the rb-tree, to slightly improve performance.

Reviewed by: alc, kib
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D35486

(cherry picked from commit f979ad00306508f0c9fc925ec05b2413b70ab5f1)

2 years agoiommu_gas: use to first-fit search for lowermatch
Doug Moore [Tue, 14 Jun 2022 05:47:22 +0000 (00:47 -0500)]
iommu_gas: use to first-fit search for lowermatch

Reverse the order of the search for a free space in lowermatch, to
make it a first-fit search. Iommu_gas_match_one always allocates from
the beginning of the free gap discovered from searching the tree, so
the current code isn't really allocating in a reverse first-fit
anyway, and making the search first-fit reduces the number of iommu
page table pages that are used.

Reported by: alc
Reviewed by: alc, kib
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D35458

(cherry picked from commit 975715b78819c6de68df15a6dd78157c6dba0fcb)

2 years agoiommu_gas: pass size to iommu_map without rounding
Doug Moore [Fri, 10 Jun 2022 18:56:54 +0000 (13:56 -0500)]
iommu_gas: pass size to iommu_map without rounding

Let the caller to iommu_map pass the size parameter without rounding
it up to a multiple of page size.  Let iommu_map round it up when
necessary, which is not all of the time, so that in some cases less
space is reserved.

Reviewed by: alc, kib (previous version)
Tested by: pho, br
Discussed with: andrew
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D35424

(cherry picked from commit 19bb5a7244ff0374f44068a20df9393a1c88de5d)

2 years agoCheck if iommu ctx is valid before passing it to functions.
Ruslan Bukin [Sat, 7 May 2022 09:27:17 +0000 (10:27 +0100)]
Check if iommu ctx is valid before passing it to functions.

Differential Revision: https://reviews.freebsd.org/D35130
Sponsored by: UKRI

(cherry picked from commit 8b8e56dd5bc781b7d1c7185cd267b0b16a637aa2)

2 years agoiommu_gas: Change find_space lower search order
Doug Moore [Thu, 9 Jun 2022 04:14:28 +0000 (23:14 -0500)]
iommu_gas: Change find_space lower search order

iommu_gas_lowermatch looks right, then left, then right again in its
search for free space.  Change to a more straightforward last-fit
search that touches fewer tree nodes and improves performance.

Reported by: wxzhu@rice.edu
Reviewed by: alc, kib
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D35439

(cherry picked from commit 30031172534c22695ab7b26a9420bda7b20b0824)

2 years agovfs_mount.c: convert explicit panics and KASSERTs to MPASSERT/MPPASS
Konstantin Belousov [Wed, 29 Jun 2022 11:37:58 +0000 (14:37 +0300)]
vfs_mount.c: convert explicit panics and KASSERTs to MPASSERT/MPPASS

(cherry picked from commit ad175a107b4e0fdc0128c98c3ca9d6cf8006aba7)

2 years agoAdd MPASSERT() and MPPASS() macros
Konstantin Belousov [Wed, 29 Jun 2022 11:37:04 +0000 (14:37 +0300)]
Add MPASSERT() and MPPASS() macros

(cherry picked from commit 61ae7a355ce787aecd84dc35a047f458a5b5b1fb)

2 years agovfs_op_exit(): assert that mnt_vfs_ops stays non-zero for unmount or suspend
Konstantin Belousov [Mon, 20 Jun 2022 15:54:44 +0000 (18:54 +0300)]
vfs_op_exit(): assert that mnt_vfs_ops stays non-zero for unmount or suspend

(cherry picked from commit 1e54362824f62a9d43585e94422d1d3d5f0268f9)

2 years agolibsysdecode: Fix a typo in my email
Dmitry Chagin [Thu, 23 Jun 2022 09:29:29 +0000 (12:29 +0300)]
libsysdecode: Fix a typo in my email

Reported by: MikaĆ«l Urankar
MFC after: 2 weeks

(cherry picked from commit ed366c37d85a7ac7c58fcf3a0e8db980d099285c)

2 years agolinprocfs: Skip printing of the guard page in the /proc/self/maps
Dmitry Chagin [Wed, 22 Jun 2022 11:49:40 +0000 (14:49 +0300)]
linprocfs: Skip printing of the guard page in the /proc/self/maps

To calculate the base (lowest addressable) address of the stack of the
initial thread glibc parses /proc/self/maps.
In fact, the base address is calculated as 'to' value of stack entry of the
/proc/self/maps - stack size limit (if the stack grows down).
The base address should fit in between preceding entry and stack entry of
the /proc/self/maps.
In FreeBSD, since 19bd0d9 (Implement address space guards), we actually
have two mappings for the stack region. The first one is the no-access
mapping for the region the stack can grow into (guard page), and the
second - initial stack region with size sgrowsiz.
The first mapping confuses Glibc, in the end which is improperly
calculate stack size and the base address.

PR: 253337
Reviewed by: kib
Differential revision: https://reviews.freebsd.org/D35537
MFC after: 2 week

(cherry picked from commit ef1976ccf5420d0912afcb49733c7a88643069da)

2 years agokdump: Decode Linux *at() syscalls
Dmitry Chagin [Wed, 22 Jun 2022 11:29:38 +0000 (14:29 +0300)]
kdump: Decode Linux *at() syscalls

MFC after: 2 weeks

(cherry picked from commit 4d496ab44a09a3a20dbec8968d1b4ef837ab10bb)

2 years agolinux(4): To reuse MD linux.h hide kernel dependencies unde _KERNEL constraint
Dmitry Chagin [Wed, 22 Jun 2022 11:28:24 +0000 (14:28 +0300)]
linux(4): To reuse MD linux.h hide kernel dependencies unde _KERNEL constraint

MFC after: 2 weeks

(cherry picked from commit d416ee86c73bef55ba29c25cd594156441f1daa9)

2 years agokdump: For future use made decode_filemode() public
Dmitry Chagin [Wed, 22 Jun 2022 11:26:50 +0000 (14:26 +0300)]
kdump: For future use made decode_filemode() public

MFC after: 2 weeks

(cherry picked from commit ed2f123a217ff23800e97dde9801b6732310c57c)

2 years agolibsysdecode: Decode Linux open() flags
Dmitry Chagin [Wed, 22 Jun 2022 11:26:10 +0000 (14:26 +0300)]
libsysdecode: Decode Linux open() flags

MFC after: 2 weeks

(cherry picked from commit fde775bc53de5b5abecdf8679f47bb3aa236e546)

2 years agolibsysdecode: Decode Linux *at() flags
Dmitry Chagin [Wed, 22 Jun 2022 11:25:36 +0000 (14:25 +0300)]
libsysdecode: Decode Linux *at() flags

MFC after: 2 weeks

(cherry picked from commit d2b3237eebdfd46d54a9539acf13e3f24dc935b2)

2 years agokdump: Decode Linux clock_nanosleep syscall
Dmitry Chagin [Wed, 22 Jun 2022 11:21:42 +0000 (14:21 +0300)]
kdump: Decode Linux clock_nanosleep syscall

MFC after: 2 weeks

(cherry picked from commit 39de84b619b0d8bb27aa77e49b6134ca06370e1a)

2 years agokdump: Decode Linux sigprocmask how argument
Dmitry Chagin [Wed, 22 Jun 2022 11:20:39 +0000 (14:20 +0300)]
kdump: Decode Linux sigprocmask how argument

MFC after: 2 weeks

(cherry picked from commit f587a2a765fb3bac352bedf6ef8b322eb3363339)

2 years agokdump: Decode getitimer, setitimer which argument
Dmitry Chagin [Wed, 22 Jun 2022 11:19:31 +0000 (14:19 +0300)]
kdump: Decode getitimer, setitimer which argument

Reviewed by: jhb (previous version, without truss)
Differential revision: https://reviews.freebsd.org/D35231
MFC after: 2 weeks

(cherry picked from commit b9b86b6742ad589e5ad1077c000bfdb2ce9ac198)

2 years agolibsysdecode: Decode setitimer, getitimer which argument
Dmitry Chagin [Wed, 22 Jun 2022 11:17:10 +0000 (14:17 +0300)]
libsysdecode: Decode setitimer, getitimer which argument

Reviewed by: jhb
Differential revision: https://reviews.freebsd.org/D35230
MFC after: 2 weeks

(cherry picked from commit 2fc3a51db14f2be7fd1f5515c6adfc60d4880207)

2 years agokdump: Decode Linux l_sigset_t.
Dmitry Chagin [Wed, 22 Jun 2022 11:15:20 +0000 (14:15 +0300)]
kdump: Decode Linux l_sigset_t.

Reviewed by: markj
Differential revision: https://reviews.freebsd.org/D35153
MFC after: 2 weeks

(cherry picked from commit 3606a213bfa42e1970197ba9b3494c88d5827259)

2 years agolinux(4): Trace Linux l_sigset_t.
Dmitry Chagin [Wed, 22 Jun 2022 11:09:54 +0000 (14:09 +0300)]
linux(4): Trace Linux l_sigset_t.

MFC after: 2 weeks

(cherry picked from commit 9310737333eb9e1a15340c10ea3fd68e0ccdcfa9)

2 years agokdump: Add preliminary support for decoding Linux syscalls
Dmitry Chagin [Wed, 22 Jun 2022 11:03:56 +0000 (14:03 +0300)]
kdump: Add preliminary support for decoding Linux syscalls

Reviewed by: jhb
Differential revision: https://reviews.freebsd.org/D35216
MFC after: 2 weeks

(cherry picked from commit 8a0f6d8cfdb75a9484c456984ff55192f8b16886)

2 years agokdump: For future use extract common code to a separate files
Dmitry Chagin [Wed, 22 Jun 2022 11:01:36 +0000 (14:01 +0300)]
kdump: For future use extract common code to a separate files

Reviewed by: jhb
Differential revision: https://reviews.freebsd.org/D35356
MFC after: 2 weeks

(cherry picked from commit 8b8e2e8f4154e02670123caf9a1dd1c5188463cc)

2 years agolibsysdecode: Add method for decoding Linux signal names
Dmitry Chagin [Wed, 22 Jun 2022 11:00:05 +0000 (14:00 +0300)]
libsysdecode: Add method for decoding Linux signal names

Reviewed by: jhb
Differential revision: https://reviews.freebsd.org/D35215
MFC after: 2 weeks

(cherry picked from commit f85848fc61752c3823825cdee34104a640d7f6e0)

2 years agolibsysdecode: Add preliminary support for decoding Linux syscalls
Dmitry Chagin [Wed, 22 Jun 2022 10:58:53 +0000 (13:58 +0300)]
libsysdecode: Add preliminary support for decoding Linux syscalls

Differential revision: https://reviews.freebsd.org/D35354
MFC after: 2 weeks

(cherry picked from commit b69ae1a34c6f918118693490f18a81ecd7163f83)

2 years agolibsysdecode: For future use extract common code to a separate files
Dmitry Chagin [Wed, 22 Jun 2022 10:12:21 +0000 (13:12 +0300)]
libsysdecode: For future use extract common code to a separate files

Reviewed by: jhb, emaste
Differential revision: https://reviews.freebsd.org/D35353
MFC after: 2 weeks

(cherry picked from commit 9dac609629d4e0ba813df6169f8bd8383fca024f)

2 years agowpa_supplicant: Resolve secondary VAP association issue
J.R. Oldroyd [Sat, 2 Jul 2022 18:15:31 +0000 (11:15 -0700)]
wpa_supplicant: Resolve secondary VAP association issue

Association will fail on a secondary open unprotected VAP when the
primary VAP is configured for WPA. Examples of secondary VAPs are,
hotels, universities, and commodity routers' guest networks.

A broadly similar bug was discussed on Red Hat's bugzilla affecting
association to a D-Link DIR-842.

This suggests that as IEs were added to the 802.11 protocol the old code
was increasingly inadaquate to handle the additional IEs, not only a
secondary VAP.

PR: 264238
Reported by: Jaskie <jiangjun12321@gmail.com>
"J.R. Oldroyd" <fbsd@opal.com>
Submitted by: "J.R. Oldroyd" <fbsd@opal.com>

(cherry picked from commit 775611ea11db0973fd8b7aef0f5eb527308efd05)

2 years agowpa: Enable WEP by default
Cy Schubert [Sun, 3 Jul 2022 18:41:00 +0000 (11:41 -0700)]
wpa: Enable WEP by default

As of hostap 2.10, WEP is disabled by default. This of course is not a
bad thing but requires some planning and an announcment to remove WEP
support by default. A possible src.conf knob or letting users know they
should use the port instead might different options.

(cherry picked from commit 7999a7f2a8679f73574b6704377436025601ee85)

2 years agoApply llvm fix for assertion/crash building archivers/c-blosc2
Dimitry Andric [Fri, 1 Jul 2022 23:13:46 +0000 (01:13 +0200)]
Apply llvm fix for assertion/crash building archivers/c-blosc2

Merge commit 88ce403c6aab from llvm git (by Florian Hahn):

  [LV] Add new block to place recurrence splice, if needed.

  In some cases, a recurrence splice instructions needs to be inserted
  between to regions, for example if the regions get re-arranged during
  sinking.

  Fixes #56146.

PR: 264979
Reported by: Robert Clausecker <fuz@fuz.su>
MFC after: 3 days

(cherry picked from commit 836d47d38e0a586c9b307fd7fb46bf563acba151)

2 years agopmc: intr pmc.soft(3) update
Wojciech Macek [Wed, 22 Sep 2021 07:52:42 +0000 (09:52 +0200)]
pmc: intr pmc.soft(3) update

Obtained from: Semihalf
Sponsored by: Stormshield
Reviewed by: mhorne
Differential revision: https://reviews.freebsd.org/D32055

(cherry picked from commit 319b150003001fa09cb4a97452c57340ce42db35)

2 years agohwpmc: don't validate capabilities in allocation method
Mitchell Horne [Mon, 30 Aug 2021 17:03:18 +0000 (14:03 -0300)]
hwpmc: don't validate capabilities in allocation method

These checks were inconsistently applied across the various hwpmc
classes. The condition is already checked by the generic code in
hwpmc_mod.c, so remove them.

MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D31388

(cherry picked from commit 0e78510b75497d183ab1aafbd99ff1031798bd84)

2 years agohwpmc: consistently validate PMC class in allocation method
Mitchell Horne [Mon, 30 Aug 2021 17:02:23 +0000 (14:02 -0300)]
hwpmc: consistently validate PMC class in allocation method

It is always a good idea. In one case, attempting to allocate N+1 PMCs
from a class with N hardware counters would incorrectly attempt to
allocate from the next class in the list. Without this validation, this
can lead to all kinds of strange behaviour.

Since powerpc_allocate_pmc() is used by both the mpc7xxx and ppc970
classes, add a new global to track which is active (it will never be
both).

Reviewed by: luporl, ray
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D31387

(cherry picked from commit 315cd55dba61416495a847f0eed6e522421e2347)

2 years agoOpenSSL: Regen assembly file for OpenSSSL 1.1.1q
Jung-uk Kim [Tue, 5 Jul 2022 16:06:50 +0000 (12:06 -0400)]
OpenSSL: Regen assembly file for OpenSSSL 1.1.1q

(cherry picked from commit 9576bca5834b3ccfbf7ff7d3b49db9c05c51d44c)

2 years agoOpenSSL: Regen manual pages for OpenSSL 1.1.1q
Jung-uk Kim [Tue, 5 Jul 2022 16:01:07 +0000 (12:01 -0400)]
OpenSSL: Regen manual pages for OpenSSL 1.1.1q

(cherry picked from commit 25fb2515923796b329329b5c1c17d200ff416e84)

2 years agoOpenSSL: Merge OpenSSL 1.1.1q
Jung-uk Kim [Tue, 5 Jul 2022 15:47:01 +0000 (11:47 -0400)]
OpenSSL: Merge OpenSSL 1.1.1q

(cherry picked from commit 64cbf7cebc3b80a971e1d15124831d84604b9370)

2 years agovt: Improve multi lingual word separation.
Hans Petter Selasky [Sat, 25 Jun 2022 09:17:44 +0000 (11:17 +0200)]
vt: Improve multi lingual word separation.

Suggested by: Tomoaki AOKI <junchoon@dec.sakura.ne.jp>
Differential Revision: https://reviews.freebsd.org/D35552
PR: 263084
Sponsored by: NVIDIA Networking

(cherry picked from commit 9971e6aff1bef3d456172c41a3df3ce7266517cf)

2 years agovt: Fix contents of paste buffer for newcons.
Hans Petter Selasky [Wed, 22 Jun 2022 16:38:47 +0000 (18:38 +0200)]
vt: Fix contents of paste buffer for newcons.

Trim all word separators from end of line, except for last line and
only use '\r' to terminate the pasted lines as expected by TTY.

Submitted by: Ivan Quitschal <tezeka@hotmail.com>
Differential Revision: https://reviews.freebsd.org/D35552
PR: 263084
Sponsored by: NVIDIA Networking

(cherry picked from commit 5fe0a82501cb86278b9aa426c4bf11d992e95aaa)

2 years agorc.conf.5: Fix example of how rc.conf.d works
Mateusz Piotrowski [Tue, 28 Jun 2022 22:37:39 +0000 (00:37 +0200)]
rc.conf.5: Fix example of how rc.conf.d works

While here, fix a typo.

MFC after: 1 week

(cherry picked from commit 454adf0fa4d787bf6e6ff7be0aa6c27a91d2be8e)

2 years agohwpmc: Reenable PME before reenabling counters.
Alexander Motin [Wed, 15 Jun 2022 15:39:37 +0000 (11:39 -0400)]
hwpmc: Reenable PME before reenabling counters.

Doing otherwise may lead to lost interrupts, that in combination
with hardware PMCs freezing may leave them frozen forever.  It may
also slightly improve profiling accuracy.

MFC after: 2 weeks

(cherry picked from commit f9e62419f6daf226a7116be2601842f8f07b1928)

2 years agohwpmc: Correct sysctl description.
Alexander Motin [Wed, 15 Jun 2022 15:37:26 +0000 (11:37 -0400)]
hwpmc: Correct sysctl description.

MFC after: 2 weeks

(cherry picked from commit 918505d0a084d18ad8e440c12c24b58231b33626)

2 years agohwpmc: Increase thread priority while iterating CPUs.
Alexander Motin [Tue, 7 Jun 2022 02:36:16 +0000 (22:36 -0400)]
hwpmc: Increase thread priority while iterating CPUs.

This allows to profile already running high-priority threads, that
otherwise by blocking thread migration to respective CPUs blocked PMC
management, i.e. profiling could start only when workload completed.

While there, return the thread to its original CPU after iterating
the list.  Otherwise all threads using PMC end up on the last CPU.

MFC after: 1 month

(cherry picked from commit eff9ee7c0c8e1fe782d6c01a29bb23224b02a847)

2 years agohwpmc: Add IDs for few more Intel Atom CPUs.
Alexander Motin [Fri, 3 Jun 2022 02:55:42 +0000 (22:55 -0400)]
hwpmc: Add IDs for few more Intel Atom CPUs.

MFC after: 1 month

(cherry picked from commit 1326017849ee0536e415d69f150514d410b588ff)

2 years agohwpmc: Skip GLOBAL_CTRL updates on stop_pmc().
Alexander Motin [Thu, 2 Jun 2022 22:20:33 +0000 (18:20 -0400)]
hwpmc: Skip GLOBAL_CTRL updates on stop_pmc().

After we wipe PMC configuration, including its local enable bit(s),
we don't really care about its global enable bit.  Global enable bits
now may only be cleared by interrupt handler in case of error (sample
buffer overflow).  Being set is actually a reset default for them.

This saves one WRMSR per process-scope PMC per context switch, that
is clearly visible in profiles.

MFC after: 1 month

(cherry picked from commit 326a8d3e085dda2b82c99cf1eb2997cb4dc99a71)

2 years agolibpmc: Prefer fixed counters in Intel event aliases.
Alexander Motin [Thu, 2 Jun 2022 19:09:38 +0000 (15:09 -0400)]
libpmc: Prefer fixed counters in Intel event aliases.

It is recommended by Intel to use fixed counters when possible to
leave programmable ones for different events.

This fixes `pmc stat`, which tries to count 6 events same time, while
many Intel CPUs have only 4 programmable counters.

MFC after: 1 month

(cherry picked from commit e144cd92bb150b5636f97af3c1c89da49edf8ac8)

2 years agohwpmc: Bump Intel's IA32_PERFEVTSELx width to 64 bits.
Alexander Motin [Thu, 2 Jun 2022 17:08:55 +0000 (13:08 -0400)]
hwpmc: Bump Intel's IA32_PERFEVTSELx width to 64 bits.

Haswell added there bits 32/33 for TSX, and AlderLake added bit 34
for Adaptive PEBS Record.

MFC after: 1 month

(cherry picked from commit 1a4614a51ee7e3d3cd4ef30b16e13f9ef1713a56)

2 years agohwpmc: Make powerpc memory allocation alike to x86.
Alexander Motin [Wed, 1 Jun 2022 18:50:27 +0000 (14:50 -0400)]
hwpmc: Make powerpc memory allocation alike to x86.

It should remove some unneeded pointer dereferences.

MFC after: 1 month

(cherry picked from commit 47699fc265b9d966b4eba6bc0c41739ac793f52b)

2 years agohwpmc: Remove always zero pc_resync field.
Alexander Motin [Wed, 1 Jun 2022 15:05:50 +0000 (11:05 -0400)]
hwpmc: Remove always zero pc_resync field.

MFC after: 1 month

(cherry picked from commit 5acb4c862b5b22c0556545a0e3a8dc2300238aca)

2 years agolibpmc: Fix INVERT flag mapping for both Intel and AMD.
Alexander Motin [Tue, 31 May 2022 04:02:06 +0000 (00:02 -0400)]
libpmc: Fix INVERT flag mapping for both Intel and AMD.

Looks like somebody's trivial copy/paste bug.

MFC after: 1 month

(cherry picked from commit 037dd0a9677144a870a381a1bbccbe61e34628dd)

2 years agohwpmc: Add basic Intel Alderlake CPUs support.
Alexander Motin [Tue, 31 May 2022 03:17:37 +0000 (23:17 -0400)]
hwpmc: Add basic Intel Alderlake CPUs support.

The PMC subsystem is not designed for non-uniform CPU capabilities
(P/E-cores are different), but at least several working architectural
events like cpu_clk_unhalted.thread_p should be better than nothing.

MFC after: 1 month

(cherry picked from commit fe109d3113166c8e3b8557f0569c4e5a3597ac93)

2 years agohwpmc: Update Intel's programmable counters restrictions.
Alexander Motin [Tue, 31 May 2022 02:51:00 +0000 (22:51 -0400)]
hwpmc: Update Intel's programmable counters restrictions.

Primarily remove unneeded restrictions from later CPUs.

MFC after: 1 month

(cherry picked from commit ae57fbc7ec65fcc75575a55c911e0da32ea9c20d)

2 years agolibpmc: Disable hardcoding of Intel fixed counters.
Alexander Motin [Mon, 30 May 2022 23:58:40 +0000 (19:58 -0400)]
libpmc: Disable hardcoding of Intel fixed counters.

Intel json's now have pretty reasonable fixed counters representation
via event=0 and umask.  Hardcoded values create more confusion than fix.

MFC after: 1 month

(cherry picked from commit 73b7b181e691ebbe8a0e411fd60f5003b51dc75d)

2 years agohwpmc: Correct selection of Intel fixed counters.
Alexander Motin [Mon, 30 May 2022 23:46:48 +0000 (19:46 -0400)]
hwpmc: Correct selection of Intel fixed counters.

Intel json's use event=0 to specify fixed counter number via umask.
Alternatively fixed counters have equivalent programmable event/umask.

MFC after: 1 month

(cherry picked from commit c1e813d1230915e19a236ec687cadc1051841e56)

2 years agolibpmc: jevents: Sync with the latest Linux kernel.
Alexander Motin [Mon, 30 May 2022 18:07:31 +0000 (14:07 -0400)]
libpmc: jevents: Sync with the latest Linux kernel.

commit f4df0dbbe62ee8e4405a57b27ccd54393971c773
Date:   Wed May 25 22:04:10 2022 +0800

MFC after: 1 month

(cherry picked from commit 62ff619dcc3540659a319be71c9a489f1659e14a)

2 years agoRevert "libpmc: jevents: Partial style(9) pass"
Alexander Motin [Mon, 30 May 2022 14:07:31 +0000 (10:07 -0400)]
Revert "libpmc: jevents: Partial style(9) pass"

This file is shared with Linux.  Having local style patches does not
help with updating it.

This reverts commit 3ea691c2612952dc5e6963649ad64312bd6dff4f.

(cherry picked from commit 51f329660f1c56f295fb13bb3a34fb0895f485a1)

2 years agohwpmc: Use hardware PMCs freezing on PMI on Intel v2+.
Alexander Motin [Mon, 30 May 2022 13:07:30 +0000 (09:07 -0400)]
hwpmc: Use hardware PMCs freezing on PMI on Intel v2+.

Since version 2 Intel CPUs can freeze PMCs when intering PMI to reduce
PMI effects on collected statistics.  Since version 4 hardware supports
"streamlined" mechanism, not requiring IA_GLOBAL_CTRL MSR access.

MFC after: 1 month

(cherry picked from commit 81ffb45f02dac470cfeddb98f2e4af2d895fea4a)

2 years agohwpmc: Add Intel Core register defines up to version 5.
Alexander Motin [Sun, 29 May 2022 02:11:59 +0000 (22:11 -0400)]
hwpmc: Add Intel Core register defines up to version 5.

MFC after: 1 month

(cherry picked from commit 245b056556e61792ae3cb21fd61e9b2ce362f4b9)

2 years agolibpmc: Another update of x86 event definitions.
Alexander Motin [Fri, 27 May 2022 02:07:42 +0000 (22:07 -0400)]
libpmc: Another update of x86 event definitions.

MFC after: 1 month

(cherry picked from commit 18054d0220cfc8df9c9568c437bd6fbb59d53c3c)

2 years agohwpmc_intel: assert for correct nclasses value
Mitchell Horne [Wed, 4 Aug 2021 17:37:05 +0000 (14:37 -0300)]
hwpmc_intel: assert for correct nclasses value

This variable is set based on the exact CPU model detected. If this
value is set too small, it could lead to a NULL-dereference from an
improperly initialized pmc_rowindex_to_classdep array.

Though it has been fixed, this was previously the case for Broadwell.
Add two asserts to catch this in DEBUG kernels, as it represents a
configuration error that may be hard to uncover otherwise.

PR: 253687
Reported by: Zhenlei Huang <zlei.huang@gmail.com>
Sponsored by: The FreeBSD Foundation

(cherry picked from commit 8399d923a5697b7c194dbd44c33018c94ec42c4e)

2 years agoriscv timer: remove intermediate helper
Mitchell Horne [Tue, 21 Jun 2022 16:23:21 +0000 (13:23 -0300)]
riscv timer: remove intermediate helper

get_counts() doesn't do anything at the moment but return the result of
get_cycles(), so remove it.

For clarity, rename get_cycles() to get_timecount(); RISC-V defines
separate time and cyclecount CSRs, so let's avoid confusing the two.
They may be backed by the same underlying clock, but this is an
implementation detail.

MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35461

(cherry picked from commit b82f4170fca86e9b41b613715852d4f225b26e3a)

2 years agoset_cputicker: use a bool
Mitchell Horne [Tue, 21 Jun 2022 16:22:26 +0000 (13:22 -0300)]
set_cputicker: use a bool

The third argument to this function indicates whether the supplied
ticker is fixed or variable, i.e. requiring calibration. Give this
argument a type and name that better conveys this purpose.

Reviewed by: kib, markj
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35459

(cherry picked from commit 8701571df9fad787f5833310cd696fe51e0cde6d)

2 years agoriscv timer: cleanup
Mitchell Horne [Tue, 21 Jun 2022 16:22:40 +0000 (13:22 -0300)]
riscv timer: cleanup

- Prune unused definitions and includes
- Slight renaming of callback functions to indicate their usage
- Place vdso_fill_timehands callback logically in the file
- Small style nits

MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35460

(cherry picked from commit 715276a08b08c1f30ee3313758a936a417939ef4)

2 years agoriscv timer: implement riscv_timer_et_stop()
Mitchell Horne [Tue, 21 Jun 2022 16:23:40 +0000 (13:23 -0300)]
riscv timer: implement riscv_timer_et_stop()

Simply by masking timer interrupts.

MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35463

(cherry picked from commit 3a4256dd86f01a35b076892841042bac0054fd64)

2 years agoriscv timer: provide a function for cpu_ticks
Mitchell Horne [Tue, 21 Jun 2022 16:23:31 +0000 (13:23 -0300)]
riscv timer: provide a function for cpu_ticks

This is cheaper than the default of tc_cpu_ticks().

MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35462

(cherry picked from commit 33734a1f765f8d0ca6214918c3051d88b7b98764)

2 years agoif_ffec: free the dmamem allocation in detach
Mitchell Horne [Wed, 15 Jun 2022 15:53:25 +0000 (12:53 -0300)]
if_ffec: free the dmamem allocation in detach

Calling bus_dmamap_destroy() for a mapping which was allocated with
bus_dmamem_alloc() will result in a panic. This change is not run-time
tested, but I identified the issue while implementing the analogous
method in if_dwc(4), using this implementation as the template.

MFC after: 1 week
Sponsored by: The FreeBSD Foundation

(cherry picked from commit 8757d0fca9e6a9e174bc87728e296e5338d30eb7)

2 years agoif_ffec: fix some misleading indentation
Mitchell Horne [Sat, 7 May 2022 20:26:44 +0000 (17:26 -0300)]
if_ffec: fix some misleading indentation

(cherry picked from commit b93985c11a6e7c7dd5fcaca8db6f4847a0126991)

2 years agoif_dwc: avoid duplicate packet counts
Mitchell Horne [Tue, 21 Jun 2022 13:24:25 +0000 (10:24 -0300)]
if_dwc: avoid duplicate packet counts

We already increment the unicast IPACKETS and OPACKETS counters in the
rx/tx paths, respectively. Multicast packets are counted in the generic
ethernet code. Therefore, we shouldn't increment these counters in
dwc_harvest_stats().

Drop the early return from dwc_rxfinish_one() so that we still count
received packets with e.g. a checksum error.

PR: 263817
Reported by: Jiahao LI <jiahali@blackberry.com>
Reviewed by: manu
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35499

(cherry picked from commit 9718759043ec2ef36f12b15963194b866d731b5b)

2 years agoif_dwc: add detach method
Mitchell Horne [Tue, 21 Jun 2022 13:26:44 +0000 (10:26 -0300)]
if_dwc: add detach method

It can be useful for testing.

Reviewed by: manu
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35496

(cherry picked from commit 27b39e58b39edb76a4bffdfe730edfe6bc22f329)

2 years agoif_dwc: enable RX checksum offload feature
Jiahao Li [Tue, 21 Jun 2022 13:23:43 +0000 (10:23 -0300)]
if_dwc: enable RX checksum offload feature

We claim support in ifcaps, but don't actually enable it.

PR: 263886
Reviewed by: manu
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35498

(cherry picked from commit 35c9edab4169f99de7e8fcada6d9b499c8405f87)

2 years agoif_dwc: recognize additional rgmii phy-modes
Mitchell Horne [Tue, 21 Jun 2022 13:22:53 +0000 (10:22 -0300)]
if_dwc: recognize additional rgmii phy-modes

Per the reports, some Allwinner device trees now list the desired
phy-mode as "rgmii-id". The manual string comparison fails to detect
this, and we end up falling back to MII mode. Instead, select the clock
name using the sc->phy_mode variable, which is set in the main attach
function.

The logic to actually handle rgmii-id mode delays will be added to the
relevant PHY driver.

PR: 261355, 264673
Reported by: Maren <marentoy@protonmail.com>
Reported by: Arie Bikker <src-2016@bikker.homeunix.net>
Reviewed by: manu
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35500

(cherry picked from commit 3428997cb398767655a5651b76eb177a56e4b00a)

2 years agoif_dwc: consistently use if.c helper methods
Mitchell Horne [Tue, 21 Jun 2022 13:28:06 +0000 (10:28 -0300)]
if_dwc: consistently use if.c helper methods

And if_t rather than struct ifnet *. No functional change intended.

Reviewed by: manu
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D35497

(cherry picked from commit ca01879004d4fe49c2617f15048a52281963d0ac)

2 years agovm: Fix racy checks for swap objects
Mark Johnston [Mon, 20 Jun 2022 16:18:15 +0000 (12:18 -0400)]
vm: Fix racy checks for swap objects

Commit 4b8365d752ef introduced the ability to dynamically register
VM object types, for use by tmpfs, which creates swap-backed objects.
As a part of this, checks for such objects changed from

  object->type == OBJT_DEFAULT || object->type == OBJT_SWAP

to

  object->type == OBJT_DEFAULT || (object->flags & OBJ_SWAP) != 0

In particular, objects of type OBJT_DEFAULT do not have OBJ_SWAP set;
the swap pager sets this flag when converting from OBJT_DEFAULT to
OBJT_SWAP.

A few of these checks are done without the object lock held.  It turns
out that this can result in false negatives since the swap pager
converts objects like so:

  object->type = OBJT_SWAP;
  object->flags |= OBJ_SWAP;

Fix the problem by adding explicit tests for OBJT_SWAP objects in
unlocked checks.

PR: 258932
Fixes: 4b8365d752ef ("Add OBJT_SWAP_TMPFS pager")
Reported by: bdrewery
Reviewed by: kib
Sponsored by: The FreeBSD Foundation

(cherry picked from commit e123264e4dc394602f9fed2f0376204b5998d815)

2 years agoaw_mmc: add crash dumping support in MMCCAM mode
Andriy Gapon [Wed, 16 Feb 2022 07:56:06 +0000 (09:56 +0200)]
aw_mmc: add crash dumping support in MMCCAM mode

(cherry picked from commit af8b51b0bac932fd0bb80cc1fe672b725bd5f0e8)

2 years agoliberate wdog_kern_pat call from SW_WATCHDOG in arm minimdump code
Andriy Gapon [Wed, 16 Feb 2022 07:53:42 +0000 (09:53 +0200)]
liberate wdog_kern_pat call from SW_WATCHDOG in arm minimdump code

Obviosuly, there are hardware watchdogs on arm.

(cherry picked from commit 1a93b2288cb1cf160214272d62dcda2ad76a1f95)

2 years agoloader: fix powerpc* ofw loader
Alfredo Dal'Ava Junior [Fri, 1 Jul 2022 15:54:01 +0000 (12:54 -0300)]
loader: fix powerpc* ofw loader

With the introduction of llvm14, the powerpc* loader used on ofw/pseries
for 32 and 64 bit architectures puts the .data.rel.ro section after
.data section. This caused a crash kernel didn't boot. Bisect pointed to
change https://reviews.llvm.org/D111717 but problem could be fixed by
adding a section description to make it appear in the expected order.

This patch is based on discussion at:
        https://github.com/llvm/llvm-project/issues/56306

MFC after: 1 day
Sponsored by: Instituto de Pesquisas Eldorado (eldorado.org.br)

(cherry picked from commit 5d48fb3b16c1496bf415fee620c61cc944b0326d)

2 years agoloader: GELI encrypted disk should still use device name disk
Toomas Soome [Mon, 20 Jun 2022 06:51:44 +0000 (09:51 +0300)]
loader: GELI encrypted disk should still use device name disk

geli_probe_and_attach() does pick geli_devsw structure for
encrypted disks, the implementation depends on device
name "disk" when device type is DEVT_DISK, but geli_devsw is
setting name field "gelidisk".

PR: 264282
Submitted by: yamagi@yamagi.org
Reported by: yamagi@yamagi.org

(cherry picked from commit e417249016efcca73c9edad21b94b1315bc44601)

2 years agoLink pwd.db.5 and spwd.db.5 to passwd.5
Mateusz Piotrowski [Thu, 30 Jun 2022 21:02:08 +0000 (23:02 +0200)]
Link pwd.db.5 and spwd.db.5 to passwd.5

Let's make it easier to find documentation for those databases.

MFC after: 3 days

(cherry picked from commit 87f49967d3249e0b536beaa9c7fb5604d5a3919a)

2 years agodevfs.5: Replace Nm with Xr devfs 8 where appropriate
Mateusz Piotrowski [Thu, 30 Jun 2022 14:45:23 +0000 (16:45 +0200)]
devfs.5: Replace Nm with Xr devfs 8 where appropriate

MFC after: 3 days

(cherry picked from commit 523477f8c5881b6ae88164ab0c93e88f8d01697c)

2 years agoMFC jail: Remove a prison's shared memory when it dies
Jamie Gritton [Wed, 29 Jun 2022 17:47:39 +0000 (10:47 -0700)]
MFC jail: Remove a prison's shared memory when it dies

Add shm_remove_prison(), that removes all POSIX shared memory segments
belonging to a prison.  Call it from prison_cleanup() so a prison
won't be stuck in a dying state due to the resources still held.

PR: 257555
Reported by: grembo

(cherry picked from commit 7060da62ff18e8e52c5e41f0794cc4f10dadfc6e)

2 years agoMFC jail: add prison_cleanup() to release resources held by a dying jail
Jamie Gritton [Wed, 29 Jun 2022 17:33:05 +0000 (10:33 -0700)]
MFC jail: add prison_cleanup() to release resources held by a dying jail

Currently, when a jail starts dying, either by losing its last user
reference or by being explicitly killed,
osd_jail_call(...PR_METHOD_REMOVE...) is called.  Encapsulate this
into a function prison_cleanup() that can then do other cleanup.

(cherry picked from commit a9f7455c38c19438d1061227b1fa11d40c5407a6)

2 years agokldload: Bring functionality in line with manual page
Juraj Lutter [Fri, 17 Jun 2022 13:36:53 +0000 (15:36 +0200)]
kldload: Bring functionality in line with manual page

Honor -q parameter and do not display any warning messages when -q is
specified.

Approved by: dfr
MFC after: 2 weeks
Sponsored by: Resulta, s.r.o.
Differential Revision: https://reviews.freebsd.org/D35511

(cherry picked from commit ad0a7ea65084e997f21fdb4724120e850b20659b)

2 years agoXHCI: clear warm and port reset
Bjoern A. Zeeb [Tue, 14 Jun 2022 16:52:45 +0000 (16:52 +0000)]
XHCI: clear warm and port reset

It seems we do not clear UPS_C_BH_PORT_RESET and UPS_C_PORT_RESET
conditions after warm or port reset.  Add that code.

Obtained from: an old patch mainly debugging other problems
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D35483

(cherry picked from commit 8f892e9bee33fda1e2a2efca4b3e8a1949264cac)

2 years agoUSB: add quirks to XHCI
Bjoern A. Zeeb [Tue, 14 Jun 2022 16:39:31 +0000 (16:39 +0000)]
USB: add quirks to XHCI

While XHCI is very generic some revisions of chipsets have problems.
On dwc3 <= 3.00a Port Disable does not seem to work so we need to not
enable it.
For that introduce quirks to xhci so that controllers can steer
certain features.  I would hope that this is and remains the only one.

Obtained from: an old patch mainly debugging other problems
Reviewed by: hselasky
Differential Revision: https://reviews.freebsd.org/D35482

(cherry picked from commit 447c418da03454a2a00bc115a69c62055a6d5272)

2 years agoLinuxKPI: 802.11: cleanup lsta better
Bjoern A. Zeeb [Sun, 26 Jun 2022 19:13:00 +0000 (19:13 +0000)]
LinuxKPI: 802.11: cleanup lsta better

This changes cleans up lsta from the VIF station list as well as
deals with freeing the lsta itself so it is not leaked.

lkpi_iv_update_bss() makes this more complicated than it should be
as we ties more sta state (incl. drv/fw) to the node that net80211
does not know about.  There is more work to be done detangling this
now that is better understood.

(cherry picked from commit e24e8103e0e2140751a1b2baf7b68725ca20aaa2)

2 years agoLinuxKPI: 802.11: remove an early bandaid to make sure queues are allocated
Bjoern A. Zeeb [Sun, 26 Jun 2022 19:17:04 +0000 (19:17 +0000)]
LinuxKPI: 802.11: remove an early bandaid to make sure queues are allocated

iwlwifi allocates queues on first wakeup.  This takes a lot longer on
FreeBSD's work implementation that it seems to on Linux based on some
discussion.  That meant that we couldn't get non-data frames out quickly
enough initially and failed to associate.
d0d2911035192473e8bd3f6b99ed5ca9b1b29e47 should have solved most of this
for us with iwlwifi.  None of the other drivers ported to LinuxKPI/802.11
up to today will call a dequeue so we get notified when the queus are
allocated or even need to do so.
Remove the bandaid initilly put in for iwlwifi now and speed up the
overall process of getting us associated.

(cherry picked from commit 841719c08fd3635a396f62a58aa017ae0bf563e0)

2 years agoLinuxKPI: 802.11: sync sta->addr in lkpi_iv_update_bss()
Bjoern A. Zeeb [Sun, 26 Jun 2022 19:04:16 +0000 (19:04 +0000)]
LinuxKPI: 802.11: sync sta->addr in lkpi_iv_update_bss()

In lkpi_iv_update_bss() introduced in d9f59799fc3e7 we swap lsta and
along with that sta and drv state if ni gets reused and swapped under
us by net80211.  What we did not do was to sync sta->addr which later
(usually in lkpi_sta_assoc_to_run) during a bss_info update cause
problems in drivers (or firmware) as the BSSID and the station address
were not aligned.

If this proves to hold up to fix iwlwifi issues seem on firmware
for older chipsets, multi-assoc runs, and rtw89 (which this fixes)
we should add asserts that lkpi_iv_update_bss() can only happen in
pre-auth stages and/or make sure we factor out synching more state
fields.

Found debugging: rtw89

(cherry picked from commit ed3ef56b29fd194a5ac0b820fd09bf01a4922bb7)

2 years agonet80211 / LinuxKPI: 802.11: add Control Trigger Subframe information
Bjoern A. Zeeb [Fri, 17 Jun 2022 22:55:13 +0000 (22:55 +0000)]
net80211 / LinuxKPI: 802.11: add Control Trigger Subframe information

Add definitions related to 802.11ax Control Trigger frame format
needed for rtw89.

(cherry picked from commit 4c3684ef5cf239efeae8675787ac30247e70be8a)

2 years agoLinuxKPI: 802.11: ieee80211_start_tx_ba_session()
Bjoern A. Zeeb [Wed, 15 Jun 2022 21:02:45 +0000 (21:02 +0000)]
LinuxKPI: 802.11: ieee80211_start_tx_ba_session()

For as long as we do not implement the compat code for tx aggregation
return -EINVAL in ieee80211_start_tx_ba_session() as both rtw88 and
rtw89 check for this value and only then disable further attempts.

(cherry picked from commit 799051e2caa8fff3359a3bb90e0896b71b6e25be)

2 years agortw88: update Realtek's rtw88 driver
Bjoern A. Zeeb [Mon, 13 Jun 2022 13:57:42 +0000 (13:57 +0000)]
rtw88: update Realtek's rtw88 driver

Update rtw88 based on wireless-testing at
4e051428044d5c47cd2c81c3b154788efe07ee11 (tag: wt-2022-06-10).

This is in preparation to apply USB changes to work on these and
LinuxKPI for them over the next weeks, as well to debug a
reported issue, and possibly extract and upstream some local fixes.

(cherry picked from commit 9c951734c28914f51b2fe2f2028272b572ade1ee)