espie [Sun, 25 Jul 2021 11:29:42 +0000 (11:29 +0000)]
nobody noticed, but the variable is still called "MODULES" with an S
florian [Sun, 25 Jul 2021 08:36:06 +0000 (08:36 +0000)]
Do not doubt a secure (i.e. validated) NXDOMAIN response when we just
switched networks. We validated it, we can't do better than that!
While here reorder the long list of conditions to make it easier to
understand when we doubt a response because we might be behind a
captive portal. First list all conditions when we do not doubt the
response and then the two conditions when we do doubt the response.
OK benno
florian [Sun, 25 Jul 2021 08:34:43 +0000 (08:34 +0000)]
We store a list of resolver strategies in order of their preference in
the configuration struct. This is also an implicit list of enabled
resolver strategies. We have also stored an explict lookup array of
enabled strategies outside of the configuration to be able to quickly
answer "is this strategy enabled" without traversing the preferences
list.
Move this table into the configuration so that we don't need to
"repair" it on config reload.
This fixes a bug where on startup the preferences list and enabled
lookup table were not in sync. It didn't matter in practice since we
do a config reload and then pass in DNSSEC trustanchors on startup.
Both actions combined repaired things.
OK benno
aoyama [Sun, 25 Jul 2021 07:12:51 +0000 (07:12 +0000)]
Force to use serial console when no graphic board is found.
If no graphic board is found, LUNA's ROM monitor forces to use serial
console even though DIP switch setting is 'use graphic console'.
For the consistency, change the the kernel behavior the same as ROM
monitor does.
Tested on my LUNA-88K2.
mglocker [Sun, 25 Jul 2021 06:43:04 +0000 (06:43 +0000)]
Enable LEDs for the LAN7800 chip, as e.g. found on the
Raspberry Pi 3 Model B+.
ok kevlo@
jsg [Sun, 25 Jul 2021 06:11:48 +0000 (06:11 +0000)]
enable iwm(4)
from Ashton Fagg who tested on 9260
jsg [Sun, 25 Jul 2021 05:51:15 +0000 (05:51 +0000)]
enable ix(4)
from Ashton Fagg who tested on 82599
jca [Sat, 24 Jul 2021 22:41:09 +0000 (22:41 +0000)]
riscv64 userland timecounter support
ok kettenis@
deraadt [Sat, 24 Jul 2021 19:11:02 +0000 (19:11 +0000)]
sync
kettenis [Sat, 24 Jul 2021 18:15:13 +0000 (18:15 +0000)]
Implement a workaround for the SiFive FU740 CIP-1200 errata.
Remove the (incomplete) support for ASIDs. RISC-V hardware that implements
ASIDs doesn't exist at this moment and the current code interferes with
the errata workaround and other pmap improvements we're planning to make.
ok drahn@, jca@, deraadt@
jca [Sat, 24 Jul 2021 16:56:54 +0000 (16:56 +0000)]
Pretty print ELF machine name as "RISC-V"
Instead of "<unknown>: f3". ok kettenis@
mglocker [Sat, 24 Jul 2021 14:43:53 +0000 (14:43 +0000)]
The updated dwctwo(4) driver introduces two new parameters.
mglocker [Sat, 24 Jul 2021 14:41:41 +0000 (14:41 +0000)]
Fix octeon build after dwctwo(4) update.
From deraadt@
schwarze [Sat, 24 Jul 2021 14:33:14 +0000 (14:33 +0000)]
Two new manual pages X509_TRUST_set(3) and X509_check_trust(3)
documenting ten functions related to X509_TRUST objects,
trust identifiers, and trust indices.
kn [Sat, 24 Jul 2021 14:15:34 +0000 (14:15 +0000)]
Capitalise DNS response code, stub learns from SLAAC, clarify wording
OK florian
inoguchi [Sat, 24 Jul 2021 13:21:04 +0000 (13:21 +0000)]
Compare strcmp and strcasecmp return value with zero
patrick [Sat, 24 Jul 2021 10:52:07 +0000 (10:52 +0000)]
The I2C controller's clock- and bitrate properties used different naming
between ACPI and FDT. The ACPI tables have now been adjusted, so now we
need to accept the new scheme as well.
ok kettenis@
mpi [Sat, 24 Jul 2021 09:16:51 +0000 (09:16 +0000)]
Modifying a knote must be done with the corresponding lock held. Assert
that the KERNEL_LOCK() is held unless the filter is marked as MPSAFE.
Should help finding missing locks when unlocking various filters.
ok visa@
visa [Sat, 24 Jul 2021 08:21:13 +0000 (08:21 +0000)]
Replace cpus_running with CPU_IS_RUNNING().
mglocker [Sat, 24 Jul 2021 06:04:44 +0000 (06:04 +0000)]
The updated dwctwo(4) driver introduces two new parameters.
jmatthew [Sat, 24 Jul 2021 05:49:59 +0000 (05:49 +0000)]
Use the presence of an intrmap, rather than the number of queues, to
determine how admin interrupts are set up, so we don't get confused about
it when there's only one queue.
visa [Sat, 24 Jul 2021 05:45:49 +0000 (05:45 +0000)]
Add basic regression tests for strchr() and strrchr().
visa [Sat, 24 Jul 2021 05:35:56 +0000 (05:35 +0000)]
Fix strchr() and strrchr() on mips64
Truncate the character arguments of strchr() and strrchr() to eight bits
so that the implied char conversion would work correctly. Otherwise the
functions would always return NULL when the character argument is
negative.
OK miod@
dtucker [Sat, 24 Jul 2021 02:57:28 +0000 (02:57 +0000)]
Skip RFC4716 format import and export tests when built without OpenSSL.
dtucker [Sat, 24 Jul 2021 02:51:14 +0000 (02:51 +0000)]
Don't omit ssh-keygen -y from usage when built without OpenSSL. It is
actually available, albeit only for
ed25519 keys.
dtucker [Sat, 24 Jul 2021 02:08:13 +0000 (02:08 +0000)]
Exclude key conversion options from usage when built without OpenSSL
since those are not available, similar to what we currently do with
the moduli screening options. We can also use this to skip the
conversion regression tests in this case.
djm [Sat, 24 Jul 2021 01:55:19 +0000 (01:55 +0000)]
don't leak environment= variable when it is not the first match
djm [Sat, 24 Jul 2021 01:54:23 +0000 (01:54 +0000)]
test for first-match-wins in authorized_keys environment=
options
deraadt [Fri, 23 Jul 2021 23:06:39 +0000 (23:06 +0000)]
sync
mglocker [Fri, 23 Jul 2021 21:47:22 +0000 (21:47 +0000)]
Make GENERIC compile again.
schwarze [Fri, 23 Jul 2021 20:50:28 +0000 (20:50 +0000)]
Similar to x509/x509_purp.c rev. 1.5:
Delete some code from X509_TRUST_cleanup(3) that had no effect:
it called a function on static objects that returns right away
unless the argument is dynamically allocated.
Pointed out by tb@.
This commit is identical to:
OpenSSL commit
5e6e650d62af09f47d63bfdd6c92e3b16e9da644
Author: Kurt Cancemi <kurt at x64architecture dot com>
Date: Thu Jun 9 21:57:36 2016 -0400
schwarze [Fri, 23 Jul 2021 20:40:49 +0000 (20:40 +0000)]
Delete some code from X509_PURPOSE_cleanup(3) that had no effect:
it called a function on static objects that returns right away
unless the argument is dynamically allocated.
OK jsing@ tb@
The useless code was independently discovered while writing documentation.
This commit is identical to:
OpenSSL commit
fa3a0286d178eb3b87bf2eb5fd7af40f81453314
Author: Kurt Cancemi <kurt at x64architecture dot com>
Date: Wed Jun 8 19:15:38 2016 -0400
schwarze [Fri, 23 Jul 2021 20:25:36 +0000 (20:25 +0000)]
Add a roff(7) comment that X509_issuer_and_serial_hash() is
intentionally undocumented because it uses MD5 only and is
unused in real-world code according to codesearch.debian.net.
No objection from tb@.
otto [Fri, 23 Jul 2021 18:04:28 +0000 (18:04 +0000)]
Make MALLOC_STATS compile again; noted by Omar Polo and Joe Nelson
schwarze [Fri, 23 Jul 2021 16:43:56 +0000 (16:43 +0000)]
clarify the meaning of the argument of X509_VERIFY_PARAM_set_purpose(3)
deraadt [Fri, 23 Jul 2021 16:23:37 +0000 (16:23 +0000)]
compile in non-DIAGNOSTIC mode; ok mglocker
schwarze [Fri, 23 Jul 2021 16:22:59 +0000 (16:22 +0000)]
mention the possibility that user-defined purpose identifiers may have
been defined or user-supplied checking functions may have been installed
job [Fri, 23 Jul 2021 16:03:47 +0000 (16:03 +0000)]
Adjust HTTP/1.1 request string
* "Connection: keep-alive" isn't needed, as the HTTP 1.1 default is to
use persistent connections (RFC 7230, section 6.3).
* "Host" is recommended to be in the front.
* "Accept-Encoding: identity" makes it clear to the server compression
encodings are not supported.
Based on suggestions from Julian Reschke.
OK benno@ claudio@
jca [Fri, 23 Jul 2021 15:31:14 +0000 (15:31 +0000)]
Use 8/4/1 bytes loads/stores for copyin/copyout/kcopy
Only use multiple bytes operations on properly aligned addresses, as
I have observed a 40x penalty for unaligned 8 bytes operations compared
to equivalent 1-byte loops on this Sifive Unmatched. The speed gain is
small but significant.
Input & ok kettenis@
schwarze [Fri, 23 Jul 2021 14:27:32 +0000 (14:27 +0000)]
new manual page X509_PURPOSE_set(3) documenting 11 functions
related to X509_PURPOSE objects, purpose identifiers, and purpose indices
cheloha [Fri, 23 Jul 2021 12:47:21 +0000 (12:47 +0000)]
nanosleep.2: miscellaneous cleanup and rewrites
Clean up the nanosleep(2) manpage. Correct errors and rewrite the
clunky parts.
Lots of input and suggestions from schwarze@ and deraadt@.
Thread: https://marc.info/?l=openbsd-tech&m=
162681970507360&w=2
ok schwarze@
florian [Fri, 23 Jul 2021 11:56:01 +0000 (11:56 +0000)]
When dhcpleasectl asks to send a new request on an interface we are
probably stuck in some way and the user wants a mostly clean slate.
If we already have an IP address transition to state REBOOTING so that
we no longer unicast dhcp requests. We will then try to reacquire our
lease twice before giving up and transition to INIT and send dhcp
discover messages accepting any IP address.
jmc [Fri, 23 Jul 2021 06:02:39 +0000 (06:02 +0000)]
occured -> occurred;
jmc [Fri, 23 Jul 2021 06:01:17 +0000 (06:01 +0000)]
punctuation;
djm [Fri, 23 Jul 2021 05:56:47 +0000 (05:56 +0000)]
mention in comment that read_passphrase(..., RP_ALLOW_STDIN) will
try to use askpass first. bz3314
convert a couple of debug() -> debug_f() while here
dtucker [Fri, 23 Jul 2021 05:53:02 +0000 (05:53 +0000)]
Simplify keygen-convert by using $SSH_KEYTYPES directly.
djm [Fri, 23 Jul 2021 05:24:02 +0000 (05:24 +0000)]
note successful authentication method in final "Authenticated to ..."
message and partial auth success messages (all at LogLevel=verbose)
ok dtucker@
dtucker [Fri, 23 Jul 2021 05:07:16 +0000 (05:07 +0000)]
Test conversion of
ed25519 and ecdsa keys too.
dtucker [Fri, 23 Jul 2021 04:56:21 +0000 (04:56 +0000)]
Add test for exporting pubkey from a passphrase-protected private key.
djm [Fri, 23 Jul 2021 04:04:52 +0000 (04:04 +0000)]
Add a ForkAfterAuthentication ssh_config(5) counterpart to the
ssh(1) -f flag. Last part of GHPR231 from Volker Diels-Grabsch.
ok dtucker
djm [Fri, 23 Jul 2021 04:00:59 +0000 (04:00 +0000)]
Add a StdinNull directive to ssh_config(5) that allows the config
file to do the same thing as -n does on the ssh(1) commandline.
Patch from Volker Diels-Grabsch via GHPR231; ok dtucker
djm [Fri, 23 Jul 2021 03:57:20 +0000 (03:57 +0000)]
make authorized_keys environment="..." directives first-match-wins
and more strictly limit their maximum number; prompted by OOM
reported by OSS-fuzz (35470).
feedback and ok dtucker@
djm [Fri, 23 Jul 2021 03:54:55 +0000 (03:54 +0000)]
regression test for time-limited signature keys
djm [Fri, 23 Jul 2021 03:37:52 +0000 (03:37 +0000)]
Let allowed signers files used by ssh-keygen(1) signatures support key
lifetimes, and allow the verification mode to specify a signature time
to check at. This is intended for use by git to support signing
objects using ssh keys. ok dtucker@
jmatthew [Fri, 23 Jul 2021 00:29:14 +0000 (00:29 +0000)]
pci_intr_msix_count() is the function that drivers using multiple MSI-X
vectors use to decide whether to use MSI-X, so make it return 0 if MSI
is not enabled for the device.
fixes problems with ix(4) on older amd64 hardware and current riscv64
ok kettenis@ dlg@
schwarze [Thu, 22 Jul 2021 19:44:30 +0000 (19:44 +0000)]
document X509_STORE_CTX_set_time(3) and X509_STORE_CTX_set_depth(3)
schwarze [Thu, 22 Jul 2021 19:09:26 +0000 (19:09 +0000)]
Major cleanup.
1. Fix the order of functions to match the order they occur in
application code, making the text significantly easier to follow.
2. Do not use the same argument placeholder *sk for several different
things; call the arguments *trusted, *untrusted, and *crls as
appropriate.
3. Avoid using the word "initialised" for two different concepts
in the same manual page; it was sometimes intended to mean "fill
with zeros" and sometimes "replace the zeros with useful data".
4. Generally, make the text more precise, more straightforward,
and shorter (-84 +65 lines of mdoc code).
krw [Thu, 22 Jul 2021 18:54:17 +0000 (18:54 +0000)]
mbrfile can be const.
mglocker [Thu, 22 Jul 2021 18:32:33 +0000 (18:32 +0000)]
Sync dwctwo(4) with the NetBSD-current code base.
On the Raspberry Pi 3 Model B+ this does as a benefit:
* Enable the USB on-board Ethernet controller through mue(4).
* Enable the two USB uhub2 ports for removable devices.
Feedback incorporated from kettenis@ and jsg@.
ok kettenis@
deraadt [Thu, 22 Jul 2021 18:29:47 +0000 (18:29 +0000)]
Only perform the default-route-pause if there are interfaces with the
AUTOCONF flag set. This removes the delay for even more (strange)
static configs.
ok sthen
kettenis [Thu, 22 Jul 2021 18:16:13 +0000 (18:16 +0000)]
Delete unused non-PIC code.
ok mlarkin@
sthen [Thu, 22 Jul 2021 18:09:04 +0000 (18:09 +0000)]
small tweak for default route checking, we don't care about the number of
routes, only if at least one exists, so can avoid the subshell and just use
grep -q. ok deraadt
stsp [Thu, 22 Jul 2021 18:01:29 +0000 (18:01 +0000)]
Make dhcpleased(8) always configure provided routes, regardless of
whether the address we received in our lease is already configured.
In the case I observed, no default route was added to the routing table
even though the server provided both an address and a route option.
As it happened the leased address was already configured on the interface.
This should not prevent routing table updates, but it did.
ok florian
schwarze [Thu, 22 Jul 2021 17:11:14 +0000 (17:11 +0000)]
Split the functions operating on the X509_VERIFY_PARAM object out
of X509_STORE_CTX_new(3) because i'm about to document five additional
functions of this kind and the page X509_STORE_CTX_new(3) is growing
unwieldy.
No text change yet, except that i added an introductory sentence
to the beginning of the DESCRIPTION of the new page.
tb [Thu, 22 Jul 2021 16:40:19 +0000 (16:40 +0000)]
Add sizes for free() in zlib
Rebased version of a diff from miod who described it as follows:
This tries to keep diffability against upstream, hence a questionable
choice of the size type for zcfree() - but all sizes should fit in 32
bits anyway.
Since all zcfree routines used in the tree cope with NULL arguments
(including the various alloc.c used by the boot blocks), I have
simplified TRY_FREE to compensate for the growth.
Reminded by and ok mpi
schwarze [Thu, 22 Jul 2021 15:35:50 +0000 (15:35 +0000)]
document X509_STORE_CTX_get0_current_issuer(3)
and X509_STORE_CTX_get0_current_crl(3)
kn [Thu, 22 Jul 2021 15:32:51 +0000 (15:32 +0000)]
Send rDNS proposals on ramdisks
Otherwise resolvd(8) will never learn nameservers and update
/etc/resolv.conf with IPv6 resolvers.
At the moment IPv6 only installations always prompt for nameservers
anyway, but that is its own bug and will be fixed soon.
This enables the upcoming fix to detect learned nameservers in the
first place.
Feedback OK florian
schwarze [Thu, 22 Jul 2021 14:30:38 +0000 (14:30 +0000)]
Move X509_STORE_CTX_get0_cert(3) to the X509_STORE_CTX_new(3) manual.
OpenSSL documents it in X509_STORE_CTX_get_error(3), but it is
misplaced there. It has nothing to do with accessing status or
error information but merely retrieves a pointer to the certificate
that the users wants to validate. It is a companion function to
X509_STORE_CTX_init(3), X509_STORE_CTX_set_cert(3),
X509_STORE_CTX_get0_store(3), and X509_STORE_CTX_get0_untrusted(3).
While here:
1. Clarify how the new, init, verify, cleanup, and free calls interact,
and who owns the memory involved, because this is all really confusing
from the user perspective.
2. Clarify how X509_STORE_CTX_init(3), X509_STORE_CTX_set_cert(3), and
X509_STORE_CTX_set_chain(3) partially override each other.
3. Move X509_STORE_CTX_set0_untrusted(3) to the proper place because
it is the same as X509_STORE_CTX_set_chain(3).
4. Add a few missing words and improve some wordings.
krw [Thu, 22 Jul 2021 13:30:40 +0000 (13:30 +0000)]
Have PRT_protected_guid() return 0/-1 like everything else
reporting success/failure.
No functional change.
krw [Thu, 22 Jul 2021 13:17:59 +0000 (13:17 +0000)]
Tweak some err() verbiage.
mvs [Thu, 22 Jul 2021 11:07:17 +0000 (11:07 +0000)]
In revision 1.74 of netmpls/mpls_input.c we use rtisvalid(4) to check
the result returned by rtalloc(4) instead of a compare with NULL. Since
it's assumed this returned `rt' could be non NULL we should use
rtfree(4) in error path. rtfree(4) also allows NULL as input.
ok claudio@
visa [Thu, 22 Jul 2021 07:22:43 +0000 (07:22 +0000)]
Make kqpoll_dequeue() usable with lazy removal of knotes
Adjust kqpoll_dequeue() so that it will clear only badfd knotes when
called from kqpoll_init(). This is needed by kqpoll's lazy removal
of knotes. Eager removal in kqpoll_dequeue() would defeat kqpoll's
attempt to reuse previously established knotes under workloads where
knote activation tends to occur already before next kqpoll scan.
Prompted by mpi@
deraadt [Thu, 22 Jul 2021 01:53:45 +0000 (01:53 +0000)]
After netstart, dhcpleased, and resolved are running, spin up to 10
seconds waiting for a default route (v4 or v6) to exist, this
increases the chance of DNS lookups working earlier. This is done
before pf is configured, sorry we have good reasons. Static
configurations are unaffected. dhclient previously did this kind of
delay, and this is the lightest touch we can come up with which gives
the same effect.
While here, also start relinking earlier.
ok benno florian sthen
krw [Wed, 21 Jul 2021 20:26:30 +0000 (20:26 +0000)]
A couple of missed 'const'.
schwarze [Wed, 21 Jul 2021 15:03:45 +0000 (15:03 +0000)]
Document X509_STORE_CTX_get_chain(3).
It is deprecated, but it is still called by various application programs,
so let's better mention it.
bluhm [Wed, 21 Jul 2021 12:23:32 +0000 (12:23 +0000)]
Also count crypto errors in ipsec_input_cb() like IPsec output in
previous commit.
krw [Wed, 21 Jul 2021 12:22:54 +0000 (12:22 +0000)]
Use a dedicated struct mbr to hold the GPT protective MBR,
either the one read from disk or the default one created
when initializing GPT.
Simplifies logic, eliminates MBR_init_GPT() and demotes
MBR_protective_mbr() to a helper function called from
GPT_read().
Nuke an unused variable in passing.
No intentional functional change.
bluhm [Wed, 21 Jul 2021 11:11:41 +0000 (11:11 +0000)]
Propagate errors from crypto_invoke() and count them in IPsec. They
should not happen, but always check error conditions. tq is never
NULL, remove the check. tdb->tdb_odrops++ is not MP safe, but will
be addressed separately in ipsec_output_cb().
OK mvs@
sthen [Wed, 21 Jul 2021 09:18:32 +0000 (09:18 +0000)]
sync
jsing [Wed, 21 Jul 2021 08:42:14 +0000 (08:42 +0000)]
Remove DTLS processed_rcds queue.
When DTLS handshake records are received from the next epoch, we will
potentially queue them on the unprocessed_rcds queue - this is usually
a Finished message that has been received without the ChangeCipherSuite
(CCS) message (which may have been dropped or reordered).
After the epoch increments (due to the CCS being received), the current
code processes all records on the unprocessed queue and immediate queues
them on the processed queue, which dtls1_get_record() then pulls from.
This form of processing only adds more complexity and another queue.
Instead, once the epoch increments, pull a single record from the
unprocessed queue and process it, allowing the contents to be consumed
by the caller. We repeat this process until the unprocessed queue is
empty, at which point we go back to consuming messages from the wire.
ok inoguchi@ tb@
nicm [Wed, 21 Jul 2021 08:09:43 +0000 (08:09 +0000)]
Do not add height twice when calculating popup_mouse_top, from M Kelly.
nicm [Wed, 21 Jul 2021 08:06:36 +0000 (08:06 +0000)]
Do not close popups on resize, instead adjust them to fit, from Anindya
Mukherjee.
jsing [Wed, 21 Jul 2021 07:51:12 +0000 (07:51 +0000)]
Silently discard invalid DTLS records.
Per RFC 6347 section 4.1.2.1, DTLS should silently discard invalid records,
including those that have a bad MAC. When converting to the new record
layer, we inadvertantly switched to standard TLS behaviour, where an
invalid record is fatal. This restores the previous behaviour.
Issue noted by inoguchi@
ok inoguchi@
kn [Wed, 21 Jul 2021 03:53:50 +0000 (03:53 +0000)]
Use exclusive lock under /dev/, silence expected errors in installer
resolvd(8), slaacd(8) and dhcpleased(8) are different from other daemons
in that there must only be a single instance.
resolvd already does this, adjust slaacd and dhcpleased accordingly while
moving the lockfile paths under /dev/ such that they work early on boot and
don't run into races should /var be (un)mounted between daemon starts.
Locking is especially required in the installer where all three daemons are
started every time the "(I)nstall, (U)pgrade, (A)utoinstall or (S)hell? "
prompt is entered, i.e. restarting installation or dropping into a shell
and back into the prompt again would start multiple instances.
To avoid expected lockfile error messages in between installer prompts,
discard standard error when starting the autoconf daemons; none of them
has other potential failure cases in installer mode before daemon(3)izing.
Input sthen deraadt
OK deraadt
jsg [Wed, 21 Jul 2021 01:03:49 +0000 (01:03 +0000)]
drm/amdkfd: fix sysfs kobj leak
From Philip Yang
b3205768cd1a3a9869500ec27e129f8a71d2152a in linux 5.10.y/5.10.52
dcdb4d904b4bd3078fe8d4d24b1658560d6078ef in mainline linux
jsg [Wed, 21 Jul 2021 01:01:21 +0000 (01:01 +0000)]
drm/dp_mst: Add missing drm parameters to recently added call to drm_dbg_kms()
From Jose Souza
cae871baa4f30608dc2084ca8ddc8f7e49913721 in linux 5.10.y/5.10.52
24ff3dc18b99c4b912ab1746e803ddb3be5ced4c in mainline linux
jsg [Wed, 21 Jul 2021 00:59:32 +0000 (00:59 +0000)]
drm/dp_mst: Avoid to mess up payload table by ports in stale topology
From Wayne Lin
16fb4e9c39b9543efd0c3dbc3dbb8865d090646a in linux 5.10.y/5.10.52
3769e4c0af5b82c8ea21d037013cb9564dfaa51f in mainline linux
jsg [Wed, 21 Jul 2021 00:57:36 +0000 (00:57 +0000)]
drm/dp_mst: Do not set proposed vcpi directly
From Wayne Lin
3462bc8b1a1f1b507804d33d118402235e8a1fab in linux 5.10.y/5.10.52
35d3e8cb35e75450f87f87e3d314e2d418b6954b in mainline linux
jsg [Wed, 21 Jul 2021 00:53:30 +0000 (00:53 +0000)]
drm/i915/gt: Fix -EDEADLK handling regression
From Ville Syrjala
0728df8048060e9bdedb9dd38c62782ee97184ba in linux 5.10.y/5.10.52
2feeb52859fc1ab94cd35b61ada3a6ac4ff24243 in mainline linux
jsg [Wed, 21 Jul 2021 00:50:05 +0000 (00:50 +0000)]
drm/i915/gtt: drop the page table optimisation
From Matthew Auld
81dd2d60f677bbab622c52711a711f0f43d37458 in linux 5.10.y/5.10.52
0abb33bfca0fb74df76aac03e90ce685016ef7be in mainline linux
jcs [Tue, 20 Jul 2021 18:33:59 +0000 (18:33 +0000)]
add AMD 17h/6xh Root Complex
ok brynet
schwarze [Tue, 20 Jul 2021 17:31:32 +0000 (17:31 +0000)]
Split X509_NAME_hash(3) out of d2i_X509_NAME(3) and document
X509_issuer_name_hash(3), X509_subject_name_hash(3), and the _old variants.
Even though this is only tangentially related to decoding and encoding,
including a single function in d2i_X509_NAME(3) was probably OK,
but let's not bog down that page with six functions that are likely
to become obsolete at some point - even though right now, they are
still being used both internally and by external software.
mvs [Tue, 20 Jul 2021 16:44:55 +0000 (16:44 +0000)]
Turn pipex(4) session statistics to per-CPU counters. This makes pipex(4)
more compliant to bluhm@'s work on traffic forwarding parallelization.
ok yasuoka@ bluhm@
bluhm [Tue, 20 Jul 2021 16:32:28 +0000 (16:32 +0000)]
The current workaround to disable parallel IPsec did not work.
Variable nettaskqs must not change at runtime. Interface input
queues choose the thread during init with ifiq_softnet = net_tq().
So it cannot be modified after pfkeyv2_send() sets the first SA in
kernel. Also changing the calculation in net_tq() may call task_del()
with a different taskq than task_add().
Instead of restricting the index to the first softnet task, use an
exclusive lock. For now just move the comment. We can later decide
if a write net lock or kernel lock is better.
OK mvs@
stsp [Tue, 20 Jul 2021 16:01:52 +0000 (16:01 +0000)]
Add the 7265-17 image back into the list of firmware used by iwm(4).
stsp [Tue, 20 Jul 2021 16:00:47 +0000 (16:00 +0000)]
Do not attempt to load 7265D iwm(4) firmware on all 7265 devices.
There are several variants of 7265 devices, all of which share a PCI ID.
7265D devices can be told apart by looking at the hardware revision number.
Non-D 7265 devices won't load 7265D firmware. Load the 7265-17 image instead.
Problem reported and fix tested on a non-D 7265 by Tilo Stritzky.
Tested on a 7265D by myself.
kettenis [Tue, 20 Jul 2021 15:25:48 +0000 (15:25 +0000)]
Use installboot -p such that the right partition will be newfs'ed.
ok krw@, deraadt@
kettenis [Tue, 20 Jul 2021 14:51:56 +0000 (14:51 +0000)]
Add -p option to "prepare" (newfs) a filesystem that will be used for
the bootloader. This is a no-op on architectures where such a filesystem
isn't needed.
ok krw@, deraadt@
stsp [Tue, 20 Jul 2021 14:44:37 +0000 (14:44 +0000)]
Make iwm_init() call iwm_stop() if we fail to move into SCAN state.
Generally, iwm_init() must either succeed or reset everything. In the case
I observed, the missing call to iwm_stop() left us with a non-zero refcount
for the newstate task which failed to perform the INIT -> SCAN transition.
The next ioctl request from userspace would then trigger a KASSERT in
iwm_init() which checks for old tasks that haven't run to completion.
This problem won't happen under normal conditions. It will happen if
the firmware crashes in response to a bad scan command, for instance.
I ran into this while working on support for newer iwx(4) firmware.
stsp [Tue, 20 Jul 2021 14:44:09 +0000 (14:44 +0000)]
Make iwx_init() call iwx_stop() if we fail to move into SCAN state.
Generally, iwx_init() must either succeed or reset everything. In the case
I observed, the missing call to iwx_stop() left us with a non-zero refcount
for the newstate task which failed to perform the INIT -> SCAN transition.
The next ioctl request from userspace would then trigger a KASSERT in
iwx_init() which checks for old tasks that haven't run to completion.
This problem won't happen under normal conditions. It will happen if
the firmware crashes in response to a bad scan command, for instance.
I ran into this while working on support for newer iwx(4) firmware.