stsp [Thu, 15 Feb 2024 11:57:38 +0000 (11:57 +0000)]
fix datapath Rx buffer management in qwx(4)
Fixes Tx/Rx stalls where the device ran out of free Rx buffers to use.
The device consumes buffers on the Rx refill ring out of order, which
the ring management code I wrote intially would not handle properly.
Instead of using an index into the ring where we would expect to see
a free slot which was in fact occupied, keep track of free buffers via
a bitmap.
mglocker [Thu, 15 Feb 2024 10:56:53 +0000 (10:56 +0000)]
Fix bogus packet length calculation in the RX/LRO path, which can lead to
TSO tagging forwarded packets which shouldn't. This will also fix the em(4)
watchdog timeouts seen after the em(4) TSO commit.
Thanks to Hrvoje Popovski for providing the infrastructure the reproduce
the issue, and test the fix.
Discussed with bluhm@. OK claudio@
tb [Thu, 15 Feb 2024 10:34:30 +0000 (10:34 +0000)]
BIO_dump*() avoid signed integer overflow
This API returns an int encoding the number of bytes printed. Thus, a dump
of a large enough byte string can make this overflow and rely on undefined
behavior. With an indent of 64, as little as 26 MB is enough to make this
happen.
ok jsing
jsg [Thu, 15 Feb 2024 09:48:03 +0000 (09:48 +0000)]
return non-zero if drm_fb_helper_alloc_info() fails in rkdrm_fb_probe()
found by smatch
tb [Thu, 15 Feb 2024 07:01:33 +0000 (07:01 +0000)]
Ensure that the FileAndHashes list in a mft has no duplicates
ok job
jsg [Thu, 15 Feb 2024 00:55:01 +0000 (00:55 +0000)]
fix fd leaks in error paths
ok miod@
bluhm [Wed, 14 Feb 2024 22:41:48 +0000 (22:41 +0000)]
Check IP length in ether_extract_headers().
For LRO with ix(4) it is necessary to detect ethernet padding.
Extract ip_len and ip6_plen from the mbuf and provide it to the
drivers.
Add extended sanitity checks, like IP packet is shorter than TCP
header. This prevents offloading to network hardware with bougus
packets.
Also iphlen of extracted headers contains header length for IPv4
and IPv6, to make code in drivers simpler.
OK mglocker@
miod [Wed, 14 Feb 2024 20:44:54 +0000 (20:44 +0000)]
Remove spurious GET_CURPROC in copyout() mistakenly introduced in r1.43.
claudio [Wed, 14 Feb 2024 13:18:21 +0000 (13:18 +0000)]
Hide struct ip6q, struct ip6asfrag, struct ip6_moptions,
struct ip6po_rhinfo and struct ip6_pktopts behind _KERNEL.
The only bit userland may want from netinet6/ip6_var.h is
struct ip6stat.
The recent change to struct ip6po_rhinfo to use struct route
resulted in various build failures in ports because code
included netinet6/ip6_var.h without net/route.h.
OK tb@ sthen@
miod [Wed, 14 Feb 2024 12:48:31 +0000 (12:48 +0000)]
Remove defines no longer needed in locore. NFC
miod [Wed, 14 Feb 2024 12:36:35 +0000 (12:36 +0000)]
Remove never ever used cpu_exec() macro, apparently already obsolete by the
time this file was introduced close to 30 years ago.
stsp [Wed, 14 Feb 2024 11:30:55 +0000 (11:30 +0000)]
make qwx(4) work with custom MAC addresses set via ifconfig(8)
ok phessler@
tb [Wed, 14 Feb 2024 10:49:00 +0000 (10:49 +0000)]
rpki-client: simplify x509_get_ski()
Use X509_public_digest() to calculate the SKI to get rid of a few dances
and weirdly named variables.
ok claudio
claudio [Wed, 14 Feb 2024 07:56:23 +0000 (07:56 +0000)]
Adjust test after the change to struct bgpd_addr.
Noticed by anton@
miod [Wed, 14 Feb 2024 06:17:51 +0000 (06:17 +0000)]
Enable the pool gc thread on m88k MULTIPROCESSOR kernels now that
pmap_unmap_direct() has been fixed; also tested by aoyama@
miod [Wed, 14 Feb 2024 06:16:53 +0000 (06:16 +0000)]
Make sure pmap_unmap_direct flushes dirty cache lines.
jsg [Wed, 14 Feb 2024 03:07:58 +0000 (03:07 +0000)]
avoid use after free of frp and frp->tname
found by smatch, ok miod@ millert@
jsg [Wed, 14 Feb 2024 02:44:58 +0000 (02:44 +0000)]
avoid use after free of q
found by smatch, ok miod@ deraadt@
jsg [Wed, 14 Feb 2024 02:40:02 +0000 (02:40 +0000)]
avoid use after free of fcode
found by smatch, fix from deraadt@ ok miod@
job [Tue, 13 Feb 2024 22:44:21 +0000 (22:44 +0000)]
Add explicit ASN1_ITEM_EXP prototypes
In LibreSSL *_it are variables, in other implementations they might
be a function. This helps squash compiler warnings in -portable.
Related: https://github.com/openbsd/src/commit/
65af98848fc7a42e34d470d10fc1db8e23f9db93
OK tb@
job [Tue, 13 Feb 2024 21:18:55 +0000 (21:18 +0000)]
Refactor parse_takey()
Avoid i2d_RSAPublicKey() to help with future portability efforts.
Avoid a complication related to size_t/int for the return value
of i2d_X509_PUBKEY. While there, change the out label to 'err'.
OK tb@
job [Tue, 13 Feb 2024 20:41:22 +0000 (20:41 +0000)]
Remove the stalemanifests metrics (which are no longer in use)
OK tb@
job [Tue, 13 Feb 2024 20:40:17 +0000 (20:40 +0000)]
Improve printing of TALs extracted from .tak objects
This changeset makes the output align more with the TAL file syntax.
OK tb@
job [Tue, 13 Feb 2024 20:37:15 +0000 (20:37 +0000)]
Improve a comment about what exactly the SKI is
OK tb@
job [Tue, 13 Feb 2024 20:36:42 +0000 (20:36 +0000)]
Avoid using i2d_RSAPublicKey()
This should help with future portability efforts, and perhaps
makes the code a bit more readable.
OK tb@
miod [Tue, 13 Feb 2024 17:51:17 +0000 (17:51 +0000)]
Fix splbio usage in oosiop_scsicmd() - it does not need to be taken early
since the switch to iopool years ago, but on the other had the update of
the command TAILQ needs to be done at splbio, and it wasn't.
claudio [Tue, 13 Feb 2024 16:35:43 +0000 (16:35 +0000)]
Use an C99 anonymous union inside struct bgpd_addr
This kills the two far to generic v4 and v6 macros which is a big win.
OK tb@
claudio [Tue, 13 Feb 2024 16:12:37 +0000 (16:12 +0000)]
Remove IMSG_SESSION_DOWN handler in the session engine. The RDE does
not issue such imsgs.
OK tb@
claudio [Tue, 13 Feb 2024 14:00:24 +0000 (14:00 +0000)]
Stop logging misleading errors when custom generic error pages are in use.
Only call the open(2) log_warn for errnos that are not ENOENT. Since
that is an error worth logging.
Based on a diff from Carsten Reith (carsten.reith t-online.de)
OK florian@ deraadt@
bluhm [Tue, 13 Feb 2024 13:58:19 +0000 (13:58 +0000)]
Analyse header layout in ether_extract_headers().
Several drivers need IPv4 header length and TCP offset for checksum
offload, TSO and LRO. Accessing these fields directly caused crashes
on sparc64 due to misaligned access. It cannot be guaranteed that
IP and TCP header is 4 byte aligned in driver level. Also gcc 4.2.1
assumes that bit fields can be accessed with 32 bit load instructions.
Use memcpy() in ether_extract_headers() to get the bits from IPv4
and TCP header and store the header length in struct ether_extracted.
From there network drivers can esily use it without caring about
alignment and bit shift. Do some sanity checks with the length
values to prevent that invalid values from evil packets get stored
into hardware registers. If check fails, clear the pointer to the
header to hide it from the driver. Add debug prints that help to
figure out the reason for bad packets and provide information when
debugging drivers.
OK mglocker@
florian [Tue, 13 Feb 2024 12:58:05 +0000 (12:58 +0000)]
Fix for CVE-2023-50387 and CVE-2023-50868.
Both can be used to cause high CPU load and potentially denial of
service with specifically crafted DNSSEC responses.
This is a sync with unbound(8).
florian [Tue, 13 Feb 2024 12:57:11 +0000 (12:57 +0000)]
Fix for CVE-2023-50387 and CVE-2023-50868.
Both can be used to cause high CPU load and potentially denial of
service with specifically crafted DNSSEC responses.
florian [Tue, 13 Feb 2024 12:53:05 +0000 (12:53 +0000)]
Only generate a new xid at the start of getting a new lease.
"RFC 2131 4.1 Constructing and sending DHCP messages" has this:
| Selecting a new 'xid' for each retransmission is an implementation
| decision. A client may choose to reuse the same 'xid' or select a new
| 'xid' for each retransmitted message.
We used to change xid for each request / response cycle but this ran
into problems with slow dhcp servers where we would change the xid too
frequently and would ignore late coming replies from the server.
Andre S points out that table 5 in "4.4.1 Initialization and
allocation of network address" says for the xid field in "DHCPREQUEST"
messages:
| 'xid' from server DHCPOFFER message
This seems to suggest that we need to use the same xid for the whole
DHCPDISCOVER / DHCPOFFER / DHCPREQUEST / DHCPACK exchange of messages.
Nothing else in the RFC is saying this though.
But since there are DHCP servers out there that depend on this, we
only generate a new xid when entering the INIT, REBOOTING and RENEWING
state.
I do wonder if we should just go with a static value of 0x04, which
was chosen by a fair dice roll, so guaranteed to be random.
Issue reported, initial diff and fix tested by Andre S
deraadt likes this version
OK tb
job [Tue, 13 Feb 2024 12:38:43 +0000 (12:38 +0000)]
Document a portability caveat about GeneralizedTime and UTCTime
OK tb@
tobhe [Tue, 13 Feb 2024 12:25:11 +0000 (12:25 +0000)]
Control startup of PROC_CERT and PROC_IKEV2.
Currenly PROC_PARENT sends the configuration to both PROC_CERT and
PROC_IKEV2 and finishes by sending IMSG_CTL_ACTIVE to PROC_IKEV2.
However, when PROC_IKEV2 receives IMSG_CTL_ACTIVE it does not know
the state of PROC_CERT: PROC_CERT might not have processed the
initial configuration while PROC_IKEV2 already sends requests to
PROC_CERT, causing failed requests, or even crashes (NULL deref of
ca_certs).
In order to make sure that PROC_CERT is ready before
IMSG_CTL_ACTIVE is sent to PROC_IKEV2 that startup protocol
is changed as follows:
(1) PROC_PARENT sends configuration to both PROC_CERT and PROC_IKEV2
(2) PROC_PARENT sends IMSG_CTL_ACTIVE to PROC_CERT
(3) PROC_CERT acks IMSG_CTL_ACTIVE by sending it back to PROC_PARENT
(4) PROC_PARENT now knows that PROC_CERT is ready and has processed
all messages from step (1)
(5) PROC_PARENT sends IMSG_CTL_ACTIVE to PROC_IKEV2 and knows that
IMSG_CTL_ACTIVE will be processed by PROC_IKEV2 after all
messages from step (1)
(6) PROC_IKEV2 can now assume that PROC_CERT is ready because it has
already processed IMSG_CTL_ACTIVE
from markus@
bluhm [Tue, 13 Feb 2024 12:22:09 +0000 (12:22 +0000)]
Merge struct route and struct route_in6.
Use a common struct route for both inet and inet6. Unfortunately
struct sockaddr is shorter than sockaddr_in6, so netinet/in.h has
to be exposed from net/route.h. Struct route has to be bsd visible
for userland as netstat kvm code inspects inp_route. Internet PCB
and TCP SYN cache can use a plain struct route now. All specific
sockaddr types for inet and inet6 are embeded there.
OK claudio@
miod [Tue, 13 Feb 2024 10:16:28 +0000 (10:16 +0000)]
Remove sanity checks from uvm_pagefree(). The first thing this function does
is invoke uvm_pageclean(), which performs the exact same sanity check, so
one set of checks is enough.
ok mpi@
claudio [Tue, 13 Feb 2024 09:29:39 +0000 (09:29 +0000)]
There is no way to implement the INKERNEL() macro on sparc64.
It was set to just 1 since the dawn of sparc64 support. It is
time to accept the fact and remove all INKERNEL() checks.
OK miod@ mpi@
tb [Tue, 13 Feb 2024 08:26:56 +0000 (08:26 +0000)]
zlib: sync with base
tb [Tue, 13 Feb 2024 08:26:26 +0000 (08:26 +0000)]
Pull in z_off64_t handling from upstream
Since we don't define Z_LARGE64, we continue to define z_off64_t to z_off_t
and all the other changes are no-ops for OpenBSD.
ok kettenis millert
nicm [Tue, 13 Feb 2024 08:10:23 +0000 (08:10 +0000)]
Add two new values for the destroy-unattached option to destroy sessions
only if they are not members of sessions groups, from Mark Huang, GitHub
issue 3806.
nicm [Tue, 13 Feb 2024 08:03:50 +0000 (08:03 +0000)]
Do not allow paste into panes which have exited, from Romain Francoise
in GitHub issue 3830.
jsg [Tue, 13 Feb 2024 02:14:25 +0000 (02:14 +0000)]
fix off-by-one in bounds test
found by "buffer overflow 'mvacc_cpu_freqs' 13 <= 13" smatch error
ok miod@
mvs [Mon, 12 Feb 2024 22:48:27 +0000 (22:48 +0000)]
Pass protosw instead of domain structure to soalloc() to get real
`pr_type'. The corresponding domain is referenced as `pr_domain'.
Otherwise dp->dom_protosw->pr_type of inet sockets always points
to inetsw[0].
ok bluhm
cheloha [Mon, 12 Feb 2024 22:07:33 +0000 (22:07 +0000)]
kernel: disable hardclock() on secondary CPUs
There is no useful work left for secondary CPUs to do in hardclock().
Disable cq_hardclock on secondary CPUs and remove the now-unnecessary
early-return from hardclock().
This change reduces every system's normal clock interrupt rate by
(HZ - HZ/10) per secondary CPU. For example, an 8-core machine
with a HZ=100 kernel should see its clock interrupt rate drop from
~1600 to ~960.
Thread: https://marc.info/?l=openbsd-tech&m=
170750140915898&w=2
ok kettenis@
uaa [Mon, 12 Feb 2024 21:37:25 +0000 (21:37 +0000)]
add Allwinner H616 support (and errata fix)
ok kettenis@
job [Mon, 12 Feb 2024 16:42:42 +0000 (16:42 +0000)]
Add showfilename set option
Pressing control-G all the time to understand 'what file is in what
window' might be tedious. Instead, offer a configurable option
(default off) to display the file name in the lower left corner.
OK millert@ otto@
mpi [Mon, 12 Feb 2024 15:12:09 +0000 (15:12 +0000)]
Fix tuple & string comparisons in maps.
mpi [Mon, 12 Feb 2024 15:11:06 +0000 (15:11 +0000)]
Trailing spaces
kettenis [Mon, 12 Feb 2024 10:46:10 +0000 (10:46 +0000)]
Revert the change that enables retpoline PLTs by default. While these
provide a mitigation against branch speculation attacks, they also make
IBT control flow integrity less effective. Our kernel now uses IBPB to
as a mitigation against branch speculation attacks, so we can disable
retpoline PLTs again.
ok deraadt@
jsg [Mon, 12 Feb 2024 02:57:14 +0000 (02:57 +0000)]
Intel and AMD use different cpuid bits for MSR_PRED_CMD IBPB
ok guenther@
guenther [Mon, 12 Feb 2024 01:18:17 +0000 (01:18 +0000)]
Retpolines are an anti-pattern for IBT, so we need to shift protecting
userspace from cross-process BTI to the kernel. Have each CPU track
the last pmap run on in userspace and the last vmm VCPU in guest-mode
and use the IBPB msr to flush predictors right before running in
userspace on a different pmap or entering guest-mode on a different
VCPU. Codepatch-nop the userspace bits and conditionalize the vmm
bits to keep working if IBPB isn't supported.
ok deraadt@ kettenis@
kn [Sun, 11 Feb 2024 21:56:10 +0000 (21:56 +0000)]
Enable disk encryption in unattended installations
Interactively keeps using bioctl(8)'s own prompt, in unattended mode
ask_passphrase() ensures non-empty responses or fails.
Unlike user passwords, autoinstall(8) only supports plaintext passphrases:
Encrypt the root disk with a (p)assphrase or (k)eydisk = passphrase
New passphrase = secret
Make sure to trust the install network or use a pre-configured key disk:
Encrypt the root disk with a (p)assphrase or (k)eydisk = keydisk
Which disk contains the key disk = sd2
Which sd2 partition is the key disk = a
initial diff from Chris Narkiewicz
OK afresh1
Feedback sthen
mvs [Sun, 11 Feb 2024 21:36:49 +0000 (21:36 +0000)]
Release `sb_mtx' mutex(9) before sbunlock().
ok bluhm
bluhm [Sun, 11 Feb 2024 21:29:12 +0000 (21:29 +0000)]
Remove needless includes of netinet6/ip6_var.h header in userland.
OK millert@
mvs [Sun, 11 Feb 2024 18:14:26 +0000 (18:14 +0000)]
Use `sb_mtx' instead of `inp_mtx' in receive path for inet sockets.
In soreceve(), we only touch `so_rcv' socket buffer, which has it's own
`sb_mtx' mutex(9) for protection. So, we can avoid solock() in this
path - it's enough to hold `sb_mtx' in soreceive() and around
corresponding sbappend*(). But not right now :)
This time we use shared netlock for some inet sockets in the soreceive()
path. To protect `so_rcv' buffer we use `inp_mtx' mutex(9) and the
pru_lock() to acquire this mutex(9) in socket layer. But the `inp_mtx'
mutex belongs to the PCB. We initialize socket before PCB, tcp(4)
sockets could exist without PCB, so use `sb_mtx' mutex(9) to protect
sockbuf stuff.
This diff mechanically replaces `inp_mtx' by `sb_mtx' in the receive
path. Only for sockets which already use `inp_mtx'. All other sockets
left as is. They will be converted later.
Since the `sb_mtx' is optional, the new SB_MTXLOCK flag introduced. If
this flag is set on `sb_flags', the `sb_mtx' mutex(9) should be taken.
New sb_mtx_lock() and sb_mtx_unlock() was introduced to hide this check.
They are temporary and will be replaced by mtx_enter() when all this
area will be converted to `sb_mtx' mutex(9).
Also, the new sbmtxassertlocked() function introduced to throw
corresponding assertion for SB_MTXLOCK marked buffers. This time only
sbappendaddr() calls it. This function is also temporary and will be
replaced by MTX_ASSERT_LOCKED() later.
ok bluhm
kettenis [Sun, 11 Feb 2024 16:01:09 +0000 (16:01 +0000)]
Fix potential out-of-bounds array access.
Found by smatch, reported by and ok jsg@
op [Sun, 11 Feb 2024 09:24:26 +0000 (09:24 +0000)]
unify smtpd and makemap table parser
These are supposed to parse the same file format but have subtle
difference in the handling of comments, continuation lines and escaping.
Converge both to the simpler smtpd parser which doesn't handle
continuation lines nor escaping, and support comments only at the start
of the line.
improvements and ok millert@
tb [Sun, 11 Feb 2024 08:40:37 +0000 (08:40 +0000)]
sync with src
tb [Sun, 11 Feb 2024 08:40:06 +0000 (08:40 +0000)]
libz: more windows ifdef turd shining from upstream
jmc [Sun, 11 Feb 2024 06:40:46 +0000 (06:40 +0000)]
recue -> reduce (in comment); from andrius v
gkoehler [Sun, 11 Feb 2024 03:57:10 +0000 (03:57 +0000)]
In pkg-config, accept both "CFlags:" and "Cflags:"
matthieu@ noticed that "pkg-config libpkgconf --cflags" failed because
libpkgconf.pc has "CFlags:" with upper-case 'F', but we accepted only
"Cflags:" with lower-case 'f'. freedesktop.org's pkg-config accepts
both "C[Ff]lags", but is otherwise sensitive to case.
Accept "CFlags" by mapping it to "Cflags" when reading the file. Fail
if the file has more than one "C[Ff]lags".
ok millert@ matthieu@
jsg [Sun, 11 Feb 2024 01:31:28 +0000 (01:31 +0000)]
firefall -> firewall, from Joel Carnat
bluhm [Sun, 11 Feb 2024 01:27:45 +0000 (01:27 +0000)]
Remove include netinet6/ip6_var.h from netinet/in_pcb.h.
OK mvs@
bluhm [Sun, 11 Feb 2024 01:26:17 +0000 (01:26 +0000)]
libexpat minor bump to 14.1
bluhm [Sun, 11 Feb 2024 01:01:32 +0000 (01:01 +0000)]
Update libexpat regress to version 2.6.0.
bluhm [Sun, 11 Feb 2024 00:56:28 +0000 (00:56 +0000)]
Update libexpat to version 2.6.0.
This fixes CVE-2023-52425. OpenBSD is not affected by CVE-2023-52426.
Relevant for OpenBSD are security fixes #789 #814, bug fixes #753
#812 #813, other changes #771 #788 #764 #765, and examples, docs,
compiler warnings, clang-tidy, tests. Only a minor library bump
is necessary, this has been discussed with tb@ guenther@ kettenis@.
OK deraadt@
deraadt [Sat, 10 Feb 2024 16:47:46 +0000 (16:47 +0000)]
grow arm64 iso media again
deraadt [Sat, 10 Feb 2024 15:29:04 +0000 (15:29 +0000)]
If anything goes wrong with reading the 'sysctl hw.ucomnames', act
like it is the empty string, rather than considering it an error.
ok krw
deraadt [Sat, 10 Feb 2024 15:28:16 +0000 (15:28 +0000)]
On kernels without ucom(4) support, 'sysctl hw.ucomnames' should return
the empty string, rather than error.
ok krw
dv [Sat, 10 Feb 2024 12:31:16 +0000 (12:31 +0000)]
Prevent use of uninitialized byte in vmd's mmio decoder.
The mmio code isn't wired in, but if the ModRM decode fails, byte
may be used with an uninitialized value. Properly return an error,
but initialize the byte variable as well.
Found by smatch, reported by and ok jsg@.
naddy [Sat, 10 Feb 2024 11:28:52 +0000 (11:28 +0000)]
clean sshd random relinking kit; ok miod@
jsg [Sat, 10 Feb 2024 07:10:13 +0000 (07:10 +0000)]
fix off-by-one when printing fr_arg
found by "buffer overflow 'fp64->fr_arg' 6 <= 6" smatch error
ok miod@ claudio@
tb [Sat, 10 Feb 2024 06:10:41 +0000 (06:10 +0000)]
libz: sync with base
tb [Sat, 10 Feb 2024 06:10:04 +0000 (06:10 +0000)]
libz: sync with upstream
- fix type of local variable in deflate_stored()
- more Windows compat shuffling
- wrap overlong line in gzread
dv [Sat, 10 Feb 2024 02:19:12 +0000 (02:19 +0000)]
Fix locked address interfaces in vmd(8).
Before comparing the amount of bytes read to the size of a packet
struct, make sure the fd being read was actually the packet injection
pipe(2). Locked address interfaces force using the same copy-based
approach used for the internal dhcp service for "local" interfaces
but were accidentally being treated as reads from the pipe(2) and
not the tap.
This broke networking for any locked address interfaces in vmd(8).
Reported by and ok kn@
dv [Sat, 10 Feb 2024 02:10:41 +0000 (02:10 +0000)]
Set vmd virtio device fds to -1 on close after fork.
After the recent vmd(8) commit to clean up file descriptor lifecycles,
virtio disks with multiple file descriptors (QCOW2 images with at
least one base) would fail to initialize when booted with a network
device.
Use the new fd closing routine in the vm process for virtio devices
to close the device fds and set to -1, removing buggy copying and
closing of fds.
Additionally, close the vm/device sync and async channels when
closing a device's fds.
Issue reported by and ok kn@
tb [Fri, 9 Feb 2024 19:15:53 +0000 (19:15 +0000)]
pull in another upstream tweak for windows
cheloha [Fri, 9 Feb 2024 17:42:18 +0000 (17:42 +0000)]
dt(4): move interval/profile entry points to dedicated clockintr callback
To improve the utility of dt(4)'s interval and profile probes we need
to move the probe entry points from the fixed-frequency hardclock() to
a dedicated clock interrupt callback so that the probes can fire at
arbitrary frequencies.
- Remove entry points for interval/profile probes from hardclock().
- Merge dt_prov_profile_enter(), dt_prov_interval_enter(), and
dt_prov_profile_fire() into one function, dt_clock(). This is
the now-unified callback for interval/profile probes. dt_clock()
will consume multiple events during a single execution if it is
delayed, but on platforms with high quality interrupt clocks this
should be rare.
- Each struct dt_pcb gets its own clockintr handle, dp_clockintr.
- In struct dt_pcb, replace dp_maxtick/dp_nticks with dp_nsecs,
the PCB's sampling period. Aynchronous probes must initialize
dp_nsecs to a non-zero value during dtpv_alloc().
- In struct dt_pcb, replace dp_cpuid with dp_cpu so that
dt_ioctl_record_start() knows where to bind the PCB's
dp_clockintr.
- dt_ioctl_record_start() binds, staggers, and starts all
interval/profile PCBs on the given dt_softc. Each dp_clockintr
is given a reference to its enclosing PCB so that dt_clock()
doesn't need to search for it. The staggering sort-of simulates
the current behavior under hardclock().
- dt_ioctl_record_stop() unbinds all interval/profile PCBs. The
CL_BARRIER ensures that dp_clockintr's PCB reference is not in
use by dt_clock() so that the PCB may be safely freed upon
return from dt_ioctl_record_stop(). Blocking while holding
dt_lock is not ideal, but in practice blocking in this spot is
rare and dt_clock() completes quickly on all but the oldest
hardware. An extremely unlucky thread could block for every
interval/profile PCB on the softc, but this is implausible.
DT_FA_PROFILE values are up-to-date for amd64, i386, and macppc.
Somebody with the right hardware needs to check-and-maybe-fix the
values on octeon, powerpc64, and sparc64.
Joint effort with mpi@.
Thread: https://marc.info/?l=openbsd-tech&m=
170629371821879&w=2
ok mpi@
cheloha [Fri, 9 Feb 2024 16:52:58 +0000 (16:52 +0000)]
clockintr: add clockintr_unbind()
The clockintr_unbind() function cancels any pending execution of the
given clock interrupt object's callback and severs the binding between
the object and its host CPU. Upon return from clockintr_unbind(), the
clock interrupt object may be rebound with a call to clockintr_bind().
The optional CL_BARRIER flag tells clockintr_unbind() to block if the
clockintr's callback function is executing at the moment of the call.
This is useful when the clockintr's arg is a shared reference and the
caller needs to be certain the reference is inactive.
Now that clockintrs can be bound and unbound repeatedly, there is more
room for error. To help catch programmer errors, clockintr_unbind()
sets cl_queue to NULL. Calls to other API functions after a clockintr
is unbound will then fault on a NULL dereference. clockintr_bind()
also KASSERTs that cl_queue is NULL to ensure the clockintr is not
already bound. These checks are not perfect, but they do catch some
common errors.
With input from mpi@.
Thread: https://marc.info/?l=openbsd-tech&m=
170629367121800&w=2
ok mpi@
jan [Fri, 9 Feb 2024 15:22:41 +0000 (15:22 +0000)]
vmx(4): add missing NVLAN checks
ok bluhm@
cheloha [Fri, 9 Feb 2024 15:06:23 +0000 (15:06 +0000)]
clockintr: refactor clockintr_cancel() into clockintr_cancel_locked()
Move the mutex-protected portions of clockintr_cancel() into a separate
function, clockintr_cancel_locked(), so that the code can be reused by
other callers.
Thread: https://marc.info/?l=openbsd-tech&m=
170629367121800&w=2
ok mpi@
bluhm [Fri, 9 Feb 2024 14:57:36 +0000 (14:57 +0000)]
Include sys/lock.h to make qwx(4) compile with GENERIC kernel.
OK stsp@
dv [Fri, 9 Feb 2024 14:52:39 +0000 (14:52 +0000)]
Only debug log dhcp packet info if packet is found.
No functional change. vmd(8) was being chatty and incorrect.
dv [Fri, 9 Feb 2024 14:35:47 +0000 (14:35 +0000)]
Tuck vmd's i8253 reset debug logs behind DPRINTF.
It's super chatty and pollutes verbose logging.
stsp [Fri, 9 Feb 2024 14:11:00 +0000 (14:11 +0000)]
implement qwx_dp_process_rxdma_err()
stsp [Fri, 9 Feb 2024 14:09:19 +0000 (14:09 +0000)]
implement qwx_dp_rx_process_wbm_err()
stsp [Fri, 9 Feb 2024 14:07:27 +0000 (14:07 +0000)]
implement qwx_dp_process_reo_status()
stsp [Fri, 9 Feb 2024 14:05:48 +0000 (14:05 +0000)]
implement qwx_dp_process_rx_err()
bluhm [Fri, 9 Feb 2024 14:02:11 +0000 (14:02 +0000)]
Route cache function returns hit or miss.
The route_cache() function can easily return whether it was a cache
hit or miss. Then the logic to perform a route lookup gets a bit
simpler. Some more complicated if (ro->ro_rt == NULL) checks still
exist elsewhere.
Also use route cache in in_pcbselsrc() instead of filling struct
route manually.
OK claudio@
job [Fri, 9 Feb 2024 13:49:41 +0000 (13:49 +0000)]
Bump release
OK tb@
claudio [Fri, 9 Feb 2024 12:56:53 +0000 (12:56 +0000)]
Convert the bgplg html pages to HTML5.
From Clemens Gößnitzer (clemens (at) goessnitzer.info)
Looks good to bentley@
bluhm [Fri, 9 Feb 2024 12:50:10 +0000 (12:50 +0000)]
Struct layout of qwx_softc should not depend on NBPFILTER.
Due to a missing #include "bpfilter.h", the size of struct qwx_softc
varied in different object files. This made the kernel crash on
arm64. To make debugging core dumps and libkvm easier, kernel
object layout should not depend on kernel config. Remove #if
NBPFILTER > 0 from struct definition.
problem analysis kettenis@
OK deraadt@ stsp@
tb [Fri, 9 Feb 2024 12:48:32 +0000 (12:48 +0000)]
bio_dump: add a test that prints all values of a single byte
jsg [Fri, 9 Feb 2024 12:45:10 +0000 (12:45 +0000)]
avoid uninitialised var use if qwx_core_fetch_bdf() errors
ok stsp@
yasuoka [Fri, 9 Feb 2024 11:59:23 +0000 (11:59 +0000)]
Don't assume MPPE-{Send,Recv}-Keys are used only for MS-CHAP (or
16 bytes length). initial diff from markus
stsp [Fri, 9 Feb 2024 11:24:52 +0000 (11:24 +0000)]
qwx pci code must include bpfilter.h, too; spotted by kettenis
bluhm [Fri, 9 Feb 2024 11:05:22 +0000 (11:05 +0000)]
Enable igc(4) on sparc64.
OK deraadt@ jan@
stsp [Fri, 9 Feb 2024 09:59:01 +0000 (09:59 +0000)]
make ifconfig display the Tx rate selected by qwx firmware
stsp [Fri, 9 Feb 2024 09:55:17 +0000 (09:55 +0000)]
add missing node reference counting to qwx's Tx completion path