jsg [Thu, 27 Jan 2022 23:05:39 +0000 (23:05 +0000)]
drm/amd/amdgpu: fix psp tmr bo pin count leak in SRIOV
From Jingwen Chen
8662d0c6a36807093aed34b17b930484fd4bf22f in linux 5.15.y/5.15.17
85dfc1d692c9434c37842e610be37cd4ae4e0081 in mainline linux
jsg [Thu, 27 Jan 2022 23:03:40 +0000 (23:03 +0000)]
drm/amdkfd: Fix error handling in svm_range_add
From Felix Kuehling
4a635b9d5ba5f691e378f7ec20e11b7b04f07a45 in linux 5.15.y/5.15.17
726be40607264b180a2b336c81e1dcff941de618 in mainline linux
jsg [Thu, 27 Jan 2022 23:01:23 +0000 (23:01 +0000)]
drm/amd/display: add else to avoid double destroy clk_mgr
From Martin Leung
58d33532664c1580beeb127ffe771a7739fb63d3 in linux 5.15.y/5.15.17
11dff0e871037a6ad978e52f826a2eb7f5fb274a in mainline linux
jsg [Thu, 27 Jan 2022 22:58:56 +0000 (22:58 +0000)]
drm/amdgpu/display: set vblank_disable_immediate for DC
From Alex Deucher
f94cf1cb17963905910c2adc332ebb9e44e7fa4b in linux 5.15.y/5.15.17
92020e81ddbeac351ea4a19bcf01743f32b9c800 in mainline linux
jsg [Thu, 27 Jan 2022 22:56:43 +0000 (22:56 +0000)]
drm/amd/display: check top_pipe_to_program pointer
From Yang Li
1fffa8ffd62fc511d4447b793641252c4743e81c in linux 5.15.y/5.15.17
a689e8d1f80012f90384ebac9dcfac4201f9f77e in mainline linux
jsg [Thu, 27 Jan 2022 22:52:54 +0000 (22:52 +0000)]
drm: panel-orientation-quirks: Add quirk for the Lenovo Yoga Book X91F/L
From Hans de Goede
c350fcc74035f48ee5a7226faee7079dd0ce9565 in linux 5.15.y/5.15.17
bc30c3b0c8a1904d83d5f0d60fb8650a334b207b in mainline linux
jsg [Thu, 27 Jan 2022 22:50:24 +0000 (22:50 +0000)]
drm/amd/display: Fix out of bounds access on DNC31 stream encoder regs
From Nicholas Kazlauskas
89166801f80dc64083212eef1308bfc62f6844f1 in linux 5.15.y/5.15.17
d374d3b493215d637b9e7be12a93f22caf4c1f97 in mainline linux
jsg [Thu, 27 Jan 2022 22:46:54 +0000 (22:46 +0000)]
drm/amd/display: Fix bug in debugfs crc_win_update entry
From Wayne Lin
6dcc6706dab10744d277bdc2613fb135c6a632dd in linux 5.15.y/5.15.17
4bef85d4c9491415b7931407b07f24841c1e0390 in mainline linux
jsg [Thu, 27 Jan 2022 22:44:38 +0000 (22:44 +0000)]
drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()
From Zhou Qingyang
e833ef0b545e1a5b3eaf00af391c4fd1fb47ffd8 in linux 5.15.y/5.15.17
ab50cb9df8896b39aae65c537a30de2c79c19735 in mainline linux
jsg [Thu, 27 Jan 2022 22:41:01 +0000 (22:41 +0000)]
drm/amdgpu: Fix a NULL pointer dereference in amdgpu_connector_lcd_native_mode()
From Zhou Qingyang
1e22b51876fa786c0b972e327ffb5cc686873cd1 in linux 5.15.y/5.15.17
b220110e4cd442156f36e1d9b4914bb9e87b0d00 in mainline linux
jsg [Thu, 27 Jan 2022 22:37:14 +0000 (22:37 +0000)]
drm/dp: Don't read back backlight mode in drm_edp_backlight_enable()
From Lyude Paul
56339a5546a7f66333f9d3c55e3fd1424caa4879 in linux 5.15.y/5.15.17
646596485e1ed2182adf293dfd5aec4a96c46330 in mainline linux
jsg [Thu, 27 Jan 2022 22:13:24 +0000 (22:13 +0000)]
drm: fix null-ptr-deref in drm_dev_init_release()
From Wang Hai
52e1bf49e3e25358519719284028334cb31ab543 in linux 5.15.y/5.15.17
acf20ed020ffa4d6cc8347e8d356509b95df3cbe in mainline linux
jsg [Thu, 27 Jan 2022 22:10:11 +0000 (22:10 +0000)]
drm/ttm: Put BO in its memory manager's lru list
From xinhui pan
0da09030aae4a3c9de4699e7ce71085639f8c378 in linux 5.15.y/5.15.17
781050b0a3164934857c300bb0bc291e38c26b6f in mainline linux
tb [Thu, 27 Jan 2022 20:31:21 +0000 (20:31 +0000)]
sort includes as usual
tb [Thu, 27 Jan 2022 20:30:29 +0000 (20:30 +0000)]
Do not depend on engine.h pulling in err.h and evp.h
bluhm [Thu, 27 Jan 2022 18:28:44 +0000 (18:28 +0000)]
Cast to unsigned integer before overflow can happen.
found by kubsan; joint work with tobhe@; OK miod@
visa [Thu, 27 Jan 2022 17:36:22 +0000 (17:36 +0000)]
cad(4): Mention PolarFire SoC.
visa [Thu, 27 Jan 2022 17:34:51 +0000 (17:34 +0000)]
cad(4): Disable unused priority queues
The GEMs on the PolarFire Soc implement four Rx queues and four
Tx queues for prioritizing traffic. All the queues are in a pseudo
active state after reset. The driver uses only queues with index 0.
Disable the rest to avoid DMA errors.
Tested on a SiFive FU740 by jca@, Microchip PolarFire SoC and
Xilinx Zynq-7000 by me.
millert [Thu, 27 Jan 2022 16:58:37 +0000 (16:58 +0000)]
Update awk to Dec 8, 2021 version.
Fixes error handling in closefile() and closeall(). Long standing
warnings had been made fatal and some fatal errors went undetected.
krw [Thu, 27 Jan 2022 16:26:32 +0000 (16:26 +0000)]
Nuke single use function PRT_fix_BN() and just use the ATA/ATAPI LBA -> CHS
conversion formula
LBA = (C × HPC + H) × SPT + (S − 1)
instead of the equivalent tortuous arithmetic used in PRT_fix_BN().
No intentional functional change.
jsg [Thu, 27 Jan 2022 12:25:22 +0000 (12:25 +0000)]
drop some local changes in an unused function
jsg [Thu, 27 Jan 2022 04:20:01 +0000 (04:20 +0000)]
raise sched_engine lock mtx ipl to IPL_TTY
used with spin_lock_irq() from
resubmit_virtual_request()
execlists_hold()
execlists_unhold()
execlists_capture()
remove_from_engine()
rcu_virtual_context_destroy()
virtual_submission_tasklet()
used with spin_lock_irqsave() from
execlists_submit_request()
execlists_reset_rewind()
execlists_reset_cancel()
virtual_submit_request()
intel_execlists_show_requests()
jsg [Thu, 27 Jan 2022 04:17:40 +0000 (04:17 +0000)]
raise mm obj_lock mtx ipl to IPL_TTY
used with spin_lock_irqsave() from
__i915_gem_object_set_pages()
i915_gem_suspend_late()
i915_gem_shrink()
i915_gem_shrinker_oom()
i915_gem_object_make_unshrinkable()
i915_gem_madvise_ioctl()
jsg [Thu, 27 Jan 2022 02:56:13 +0000 (02:56 +0000)]
use process name for both thread and process name
claudio [Wed, 26 Jan 2022 14:42:39 +0000 (14:42 +0000)]
Allow URI as file in -f mode. This makes it easier to explore
rpki repositories by following AIA and manifest URIs.
Also stop checking the the loaded file is not part of the auth tree,
it is possible that this file was loaded before as a dependency.
OK tb@
kettenis [Wed, 26 Jan 2022 14:39:07 +0000 (14:39 +0000)]
An ACPI device needs to be both present and enabled for it to function.
So only attempt to attach hardware that has both bits enabled. This fixes
an issue where com(4) would attach for a disabled serial port leading to
misdetection of the hardware variant and a subsequent hang when /etc/rc
runs ttyflags -a.
ok anton@, deraadt@
claudio [Wed, 26 Jan 2022 13:57:56 +0000 (13:57 +0000)]
Change the repository layout by removing the valid directory. Everything
that was in that directory is now in the chachedir root. The rsync and
rrdp directories are now .rsync/ and .rrdp/. The ta/ directory still
remains because TAs are special.
Idea and most of the diff from job@
OK tb@ and benno@
uaa [Wed, 26 Jan 2022 12:05:33 +0000 (12:05 +0000)]
shrink uca.ibufsize from UCHCOMIBUFSIZE to wMaxPacketSize of bulk-in pipe
USB transaction is finished when whole requested data has transferred,
or short packet (the size is less than wMaxPacketSize) has sent.
UCHCOMIBUFSIZE(256) was multiply of wMaxPacketSize(32). When CH340 sends
exact wMaxPacketSize byte packet, this will cause Rx jam problem
due to transaction is not finished.
Now uca.ibufsize is same as wMaxPacketSize to avoid this problem.
ok kevlo@
tb [Wed, 26 Jan 2022 11:05:41 +0000 (11:05 +0000)]
whitespace
kn [Wed, 26 Jan 2022 06:31:31 +0000 (06:31 +0000)]
Make vmm(4/amd64) tracepoints amd64-only
One can use them on non-VMM architectures, but they obviously won't hit:
# arch -s ; btrace -l | grep vmm
sparc64
tracepoint:vmm:guest_enter
tracepoint:vmm:guest_exit
Move them under __amd64__ to avoid confusion and safe a few bytes.
OK dv
anton [Wed, 26 Jan 2022 06:05:59 +0000 (06:05 +0000)]
Rework initialization of sensors on device connect.
jsg [Wed, 26 Jan 2022 04:18:05 +0000 (04:18 +0000)]
implement might_alloc() using assertwaitok()
jsg [Wed, 26 Jan 2022 01:46:12 +0000 (01:46 +0000)]
reduce diff to linux ggtt_probe_common()
tb [Tue, 25 Jan 2022 21:55:13 +0000 (21:55 +0000)]
sync
eric [Tue, 25 Jan 2022 21:53:45 +0000 (21:53 +0000)]
minor bump after api additiom
eric [Tue, 25 Jan 2022 21:51:24 +0000 (21:51 +0000)]
Introduce a signer interface intented to make TLS privsep simpler
to implement.
Add a tls_config_set_sign_cb() function that allows to register
a callback for the signing operation on a tls_config. When used,
the context installs fake pivate keys internally, and the callback
receives the hash of the public key.
Add a tls_signer_*() set of functions to manage tls_signer objects.
A tls_signer is an opaque structure on which keys are added.
It is used to compute signatures with private keys identified by
their associated public key hash.
Discussed with and ok jsing@ tb@
jsg [Tue, 25 Jan 2022 21:31:26 +0000 (21:31 +0000)]
reduce diff to linux shmem_create_from_object()
includes linux
988d4ff6e3c2220d13d8dde22a98945b64fd7977
drm/i915: Fix ww locking in shmem_create_from_object
jsg [Tue, 25 Jan 2022 21:19:21 +0000 (21:19 +0000)]
reduce diff to linux vm_fault_cpu()
adds change made in linux
9fa1f4785f2a54286ccb8a850cda5661f0a3aaf9
drm/i915: Add object locking to vm_fault_cpu
tb [Tue, 25 Jan 2022 18:01:20 +0000 (18:01 +0000)]
Rewrite paragraph to refer to EVP_CIPHER_CTX_new() and HMAC_CTX_new()
to match reality.
spotted by/ok jmc
tb [Tue, 25 Jan 2022 17:55:39 +0000 (17:55 +0000)]
Remove some HMAC_CTX_init() remnants in HMAC(3).
spotted by/ok jmc
tb [Tue, 25 Jan 2022 15:00:09 +0000 (15:00 +0000)]
Fix another return 0 bug in SSL_shutdown()
If tls13_recod_layer_send_pending() returns TLS13_IO_EOF, we will
bubble this up to the caller via tls13_legacy_return_code(), which
translates TLS13_IO_EOF to 0. This can happen if we have pending
post handshake-handshake data and the peer closes the pipe.
Presumably tls13_legacy_shutdown() should be rewritten yet again.
ok jsing
tb [Tue, 25 Jan 2022 14:51:54 +0000 (14:51 +0000)]
Avoid an infinite loop in SSL_shutdown()
If the peer closed the write side of the connection and we have not
yet received the close_notify, SSL_shutdown() makes an extra read to
try and read the peer's close_notify from the pipe. In that situation,
we receive EOF. The legacy stack will return -1 while the TLSv1.3
stack will end up returning 0.
Since the documentation is not super explicit about what should be
done if SSL_shutdown() returns 0, some applications will enter an
infinite loop. The code and documentation indicate that SSL_shutdown()
should only be called once more if it returned 0. Newer versions
of the OpenSSL documentation explicitly say that one should call
SSL_read() if SSL_shutdown() returns 0 in order to retrieve the
close_notify. Doing this would also have avoided this infinite loop.
Reported by Carsten Arzig and bluhm with a test case extracted from the
syslogd tests using IO::Socket::SSL, which has such an infinite loop.
ok bluhm jsing
visa [Tue, 25 Jan 2022 07:10:19 +0000 (07:10 +0000)]
Use unguarded loads in stack trace saving
The stack trace saver should see a system state that is not broken.
Therefore use unguarded memory accesses.
However, the unwinder is still haphazard. Terminate immediately if
the program counter or stack pointer look inconsistent.
visa [Tue, 25 Jan 2022 07:08:43 +0000 (07:08 +0000)]
Preserve pcb_onfault in kdbpeek() and kdbpoke()
This avoids misbehaviour if kdbpeek() or kdbpoke() is used in the middle
of a guarded copy operation.
gnezdo [Tue, 25 Jan 2022 04:04:40 +0000 (04:04 +0000)]
Capture a repeated pattern into sysctl_securelevel_int function
A few variables in the kernel are only writeable before securelevel is
raised. It makes sense to handle them with less code.
OK sthen@ bluhm@
bluhm [Mon, 24 Jan 2022 22:49:48 +0000 (22:49 +0000)]
An af-to pf rule must have an address family naf to use after
translation. Make stricter sanity checks in pf ioctl to avoid later
crashes during packet processing.
Reported-by: syzbot+0ef9190e7d0195496d0d@syzkaller.appspotmail.com
OK sashan@
tb [Mon, 24 Jan 2022 17:39:59 +0000 (17:39 +0000)]
Document X509_V_ERR_UNNESTED_RESOURCE. Previous version looked good
to claudio
claudio [Mon, 24 Jan 2022 17:29:37 +0000 (17:29 +0000)]
Adjust code to handle unsupported file types a bit more graceful.
The file still needs to match its hash to make the MFT valid but then
there will only be a warning printed. Parsing of other files from that
MFT are not influenced.
OK tb@
claudio [Mon, 24 Jan 2022 15:50:34 +0000 (15:50 +0000)]
When rename fails show the source filename and not the destination.
The error should be more helpful thisway.
OK tb@
tb [Mon, 24 Jan 2022 13:53:29 +0000 (13:53 +0000)]
Garbage collect unused ret in ssl3_get_new_session_ticket()
ret is used to bubble up an error from ssl3_get_message() and is unused
otherwise. Zap dead store and silly ret = 1; return ret;
ok inoguchi jsing
tb [Mon, 24 Jan 2022 13:51:48 +0000 (13:51 +0000)]
In ssl3_get_server_key_exchange() switch a 'goto err' to
'goto fatal_err' so that the illegal_parameter alert is
actually sent in case of X509_get0_pubkey() failure.
ok inoguchi jsing
tb [Mon, 24 Jan 2022 13:49:50 +0000 (13:49 +0000)]
Avoid use of uninitialized in tlsext_sni_server_parse()
If the hostname is too long, tlsext_sni_is_valid_hostname() will fail
without having initialized *is_ip. As a result, the garbage value could
lead to accepting (but otherwise ignoring) overlong and possibly invalid
hostnames without erroring in tlsext_sni_server_parse().
ok inoguchi jsing
tb [Mon, 24 Jan 2022 13:47:53 +0000 (13:47 +0000)]
Garbage collect the unused rv in tls1_check_ec_server_key() and
convert to usual form of error checking.
ok inoguchi jsing
jsg [Mon, 24 Jan 2022 08:55:58 +0000 (08:55 +0000)]
sync dma-resv with linux-5.15.y
revert linux
cd29f22019ec4ab998d2e1e8c831c7c42db4aa7d
dma-buf: Use sequence counter with associated wound/wait mutex
to not have to deal with seqcount_ww_mutex
jmc [Mon, 24 Jan 2022 06:54:15 +0000 (06:54 +0000)]
in the options list, show -f as taking "file ..." arguments;
tweak/ok claudio
afresh1 [Mon, 24 Jan 2022 00:47:05 +0000 (00:47 +0000)]
Redownload existing files with failed checksums
jsg [Sun, 23 Jan 2022 22:53:03 +0000 (22:53 +0000)]
move uao_reference() call before uvm_map()
other uses in the kernel do this as uvm_map() may sleep and the segment
may be deallocated while sleeping without a reference
kettenis notes that shouldn't happen here due to a obj reference from an
earlier i915_gem_object_lookup() call
ok visa@ kettenis@
bluhm [Sun, 23 Jan 2022 21:44:31 +0000 (21:44 +0000)]
Define all TCP TF_ flags as unsigned numbers. They are stored in
u_int t_flags. Shifting TF_TIMER with TCPT_DELACK can touch the
sign bit.
found by kubsan; suggested by deraadt@; OK miod@
jmc [Sun, 23 Jan 2022 18:40:55 +0000 (18:40 +0000)]
rearrange SYNOPSIS/usage to be a bit clearer;
discussed with and ok claudio
claudio [Sun, 23 Jan 2022 12:09:24 +0000 (12:09 +0000)]
Handle EINTR the same way in all poll loops. In all cases restart the
poll loop. In the main process move the timeout handling for repositories
into a single function that does the timeouts and the calculation of the
timeout in one go.
OK tb@
claudio [Sun, 23 Jan 2022 11:59:40 +0000 (11:59 +0000)]
On poll() failure we want to skip pollfd related action but the signal
delivery checks at the end still need to happen. So that on EINTR bgpd
processes reconfigure or mrt files ASAP.
Fix for mrt integration tests.
Reported by and ok anton@
tb [Sun, 23 Jan 2022 09:19:13 +0000 (09:19 +0000)]
zap extra blank line
claudio [Sun, 23 Jan 2022 07:21:12 +0000 (07:21 +0000)]
Allow rpki-client to display more than one file in -f mode.
Change -f to be a mode flag and pass one or multiple files as arguments
to rpki-client. Some extra checks need to be done to not load the same
certificate or CRL multiple times.
Input and OK tb@
claudio [Sun, 23 Jan 2022 05:59:35 +0000 (05:59 +0000)]
Simplify valid_cert() and valid_roa() by passing in struct auth instead
of looking it up again. For this valid_roa() needs to be moved up in
proc_parser_roa() also move out the assignment of the TAL id. Not the
right thing to alter an object in a validation function.
OK tb@
millert [Sat, 22 Jan 2022 23:22:11 +0000 (23:22 +0000)]
Flush all stdio streams before running a shell command.
Otherwise, if ed's output is not line buffered (e.g. if it is
redirected to a file or pipe) the shell command output may be
displayed before data buffered by ed itself is written.
From Soeren Tempel. OK deraadt@
krw [Sat, 22 Jan 2022 15:39:00 +0000 (15:39 +0000)]
Wrap some long lines.
tb [Sat, 22 Jan 2022 09:18:48 +0000 (09:18 +0000)]
Change valid_filename() ot return an enum rtype and rename it to
rtype_from_mftfile(). Move both rtype_from functions to mft.c.
ok beck claudio
afresh1 [Sat, 22 Jan 2022 05:03:47 +0000 (05:03 +0000)]
Support multiple -v as promised in the man page
There are now four levels of verbosity:
0. Prints only the summary
1. Prints a line when installing/removing
2. Uses the ftp(1) progress bar
3. Provides more details for debugging
With some excellent ksh knowledge provided by kn@
deraadt [Sat, 22 Jan 2022 00:49:39 +0000 (00:49 +0000)]
sync
djm [Sat, 22 Jan 2022 00:49:34 +0000 (00:49 +0000)]
add a ssh_packet_process_read() function that reads from a fd
directly into the transport input buffer.
Use this in the client and server mainloops to avoid unnecessary
copying. It also lets us use a more greedy read size without penalty.
Yields a 2-3% performance gain on cipher-speed.sh (in a fairly
unscientific test tbf)
feedback dtucker@ ok markus@
djm [Sat, 22 Jan 2022 00:45:31 +0000 (00:45 +0000)]
Use sshbuf_read() to read directly into the channel input buffer
rather than into a stack buffer that needs to be copied again;
Improves performance by about 1% on cipher-speed.sh
feedback dtucker@ ok markus@
inoguchi [Sat, 22 Jan 2022 00:45:17 +0000 (00:45 +0000)]
Use memmove instead of memcpy for overlapping memory
CID 251047 251094
OK beck@ jsing@ millert@ tb@
djm [Sat, 22 Jan 2022 00:43:43 +0000 (00:43 +0000)]
Add a sshbuf_read() that attempts to read(2) directly in to a
sshbuf; ok markus@
inoguchi [Sat, 22 Jan 2022 00:43:41 +0000 (00:43 +0000)]
Use memmove instead of memcpy for overlapping memory
CID 250936 251103
OK beck@ jsing@ millert@ tb@
inoguchi [Sat, 22 Jan 2022 00:36:46 +0000 (00:36 +0000)]
X509_GET_PUBKEY(3) return value check in libcrypto
CID 345116
ok beck@ tb@
inoguchi [Sat, 22 Jan 2022 00:34:48 +0000 (00:34 +0000)]
X509_GET_PUBKEY(3) return value check in libcrypto
CID 25131
ok beck@ tb@
suggest using X509_REQ_get0_pubkey() and remove the EVP_PKEY_free() from tb@
inoguchi [Sat, 22 Jan 2022 00:33:02 +0000 (00:33 +0000)]
X509_GET_PUBKEY(3) return value check in libcrypto
ok beck@ tb@
suggest using X509_get0_pubkey() and remove EVP_PKEY_free() from tb@
inoguchi [Sat, 22 Jan 2022 00:31:23 +0000 (00:31 +0000)]
X509_GET_PUBKEY(3) return value check in libcrypto
ok beck@ tb@
inoguchi [Sat, 22 Jan 2022 00:29:59 +0000 (00:29 +0000)]
X509_GET_PUBKEY(3) return value check in libcrypto
ok beck@ tb@
jsg [Fri, 21 Jan 2022 23:49:36 +0000 (23:49 +0000)]
release mtx before pool_get() in __xa_alloc()/__xa_store() when
PR_WAITOK is used to not sleep with mtx held
millert [Fri, 21 Jan 2022 22:53:20 +0000 (22:53 +0000)]
Treat ppoll(2) errors as fatal. OK deraadt@
tb [Fri, 21 Jan 2022 18:49:44 +0000 (18:49 +0000)]
Add function to determine the file type from the file name extension
rpki-client uses the same idiom to determine the file type in too many
places. Use one function that determines the appropriate RTYPE from the
file name. Add that type to struct mftfile and use this new member to
simplify queue_add_from_mft*().
input/ok claudio
deraadt [Fri, 21 Jan 2022 18:46:21 +0000 (18:46 +0000)]
the wrapped cron line is going to lose people
espie [Fri, 21 Jan 2022 17:41:41 +0000 (17:41 +0000)]
"move" part of the OpenBSD::State code into its own (stateless) class
so that it can be reused light-weight by portgen and the likes.
(the actual move will happen once the base sets all have the new files)
krw [Fri, 21 Jan 2022 17:29:24 +0000 (17:29 +0000)]
Split PRT_print() into PRT_print_parthdr() and PRT_print_part()
to align with GPT functions. Simplifies logic and eliminates the
magic invocations to print the header.
No functional change.
visa [Fri, 21 Jan 2022 16:39:56 +0000 (16:39 +0000)]
Limit unwinder symbol search to FUNC symbols
The prolog scanning has to locate the start of the subroutine, not some
local symbol in the middle.
This should reduce cases where unwinding goes astray.
OK miod@
deraadt [Fri, 21 Jan 2022 16:18:16 +0000 (16:18 +0000)]
In 1999 fd_set overflowing beyond FD_SETSIZE became enough of a problem that I
changed the entire tree to use fd_set allocation, and this manpage documented
the "calloc(howmany(max+1, NFDBITS), sizeof(fd_mask))" idiom. Since then we
completed converting the entire tree to poll(2), for many reasons, even ssh/sshd.
Now the use of kernel-only sys/param.h-found howmany() and related macross grate
on me, so it is time to recommend use of poll(2) instead. [On a related note,
MacOS poll(2) is been dangerously broken for many years; that is their problem to
handle as the whole ecosystem joins us in pivoting select -> poll)
ok millert
stsp [Fri, 21 Jan 2022 15:51:02 +0000 (15:51 +0000)]
Fix and re-enable active scans on iwm(4) and iwx(4).
Ensure that we supply the access point's DTIM period to firmware after
an active scan, as soon as the next beacon arrives. This prevents the
problems which prompted us to keep active scans disabled in our drivers.
Problem debugged and patch by zxystd from the OpenIntelWireless project.
I made some tweaks regarding TIM parsing, which were reviewed by zxystd.
Johannes Berg from Intel has confirmed to me via IRC that firmware
will misbehave if running with a zero DTIM period.
Tested:
8265: jca, stsp
9260: kettenis (possible fallout observed here; will keep an eye on it)
9650: stsp
ax200: zxystd, kevlo, stsp
ax201: stsp
ok kevlo@ kettenis@
millert [Fri, 21 Jan 2022 15:23:36 +0000 (15:23 +0000)]
Document EAGAIN error return, as specified by POSIX.
Our poll does not use EAGAIN but code needs to handle it for portability.
OK deraadt@ visa@
kn [Fri, 21 Jan 2022 15:05:01 +0000 (15:05 +0000)]
Enable dt(4)
This produces helpful flamegraphs on macppc.
With Andrew Krasavin <noiseless-ak AT yandex DOT ru>.
OK mpi
tb [Fri, 21 Jan 2022 14:08:33 +0000 (14:08 +0000)]
Avoid out-of-bounds accesses in case the filename is too showrt to
contain an extension of length four.
ok claudio
tobhe [Fri, 21 Jan 2022 14:07:06 +0000 (14:07 +0000)]
Fix NULL dereference which allows user programs to read parts
of the kernel memory. Found with clang static analyzer.
Feedback and ok gkoehler@
ok bluhm@
jsg [Fri, 21 Jan 2022 11:59:40 +0000 (11:59 +0000)]
sync kfd_ioctl.h with linux 5.15
jsg [Fri, 21 Jan 2022 11:37:53 +0000 (11:37 +0000)]
move media-bus-format.h to match comment in drm_connector.c
espie [Fri, 21 Jan 2022 09:06:38 +0000 (09:06 +0000)]
document show-list
djm [Fri, 21 Jan 2022 07:04:19 +0000 (07:04 +0000)]
add a helper for writing an error message to the stderr_buf and setting
quit_pending; no functional change but saves a bunch of boilerplate
djm [Fri, 21 Jan 2022 06:58:06 +0000 (06:58 +0000)]
correct comment and use local variable instead of long indirection;
spotted by dtucker@
gnezdo [Fri, 21 Jan 2022 03:13:22 +0000 (03:13 +0000)]
Require trig_test-2 to pass on arm64 and powerpc64
Thanks to mbuhl@ for testing on a plethora of platforms.
OK mbuhl@, tested by mbuhl@ and phessler@
gnezdo [Fri, 21 Jan 2022 03:13:10 +0000 (03:13 +0000)]
Optimization for tiny x in cos and sin
Subset of FreeBSD commit
4339c67c485f0b1f7699863fc29f6c06862d1dde:
Moved the optimization for tiny x from __kernel_{cos,sin}[f](x) to
{cos_sin}[f](x) so that x doesn't need to be reclassified in the
"kernel" functions to determine if it is tiny (it still needs to be
reclassified in the cosine case for other reasons that will go away).
...
Parts of the original change were applied separately before.
Suggested by kettenis@
OK mbuhl@, tested by mbuhl@ and phessler@
gnezdo [Fri, 21 Jan 2022 03:12:55 +0000 (03:12 +0000)]
Unplug assembly implementations of trig functions on x86 platforms
The same change was done by NetBSD some time back as:
Disable x87 implementations of sin, cos, tan.
The x87 hardware uses a bad approximation to pi for argument
reduction, and consequently yields bad answers for inputs near pi or
pi/2.
OK mbuhl@, tested by mbuhl@ and phessler@