cxgbe(4): Add flag to reliably stop the driver from accessing hw stats.
There are two kinds of routines in the driver that read statistics from
the hardware: the cxgbe_* variants read the per-port MPS/MAC registers
and the vi_* variants read the per-VI registers. They can be called
from the 1Hz callout or if_get_counter. All stats collection now takes
place under the callout lock and there is a new flag to indicate that
these routines should not access any hardware register.
Warner Losh [Wed, 21 Apr 2021 18:10:42 +0000 (12:10 -0600)]
pvscsi: Advertise maxio of 256k.
While the PV SCSI SG list can handle 512k of SG entries, it can only do
so for I/O that's aligned to 4k or better. newfs_msdos does unaligned
I/O, so triggers too long for host errors in cam when a 512k I/O is
attempted. Prefer power of 2 256k to the absolute maximum 508k, though
that can be revisited should the latter show to give significant
performance improvement.
MFC After: 3 days
Tested by: darius on discord (508k version of patch)
Sponsored by: Netflix
John Baldwin [Wed, 21 Apr 2021 20:57:04 +0000 (13:57 -0700)]
riscv: Clear SUM in SSTATUS for supervisor mode exceptions.
Previously, a page fault taken during copyin/out and related functions
would run the entire fault handler while permitting direct access to
user addresses. This could also leak across context switches (e.g. if
the page fault handler was preempted by an interrupt or slept for disk
I/O).
To fix, clear SUM in assembly after saving the original version of
SSTATUS in the supervisor mode trapframe.
Path MTU discovery hooks for offloaded TCP connections.
Notify the TOE driver when when an ICMP type 3 code 4 (Fragmentation
needed and DF set) message is received for an offloaded connection.
This gives the driver an opportunity to lower the path MTU for the
connection and resume transmission, much like what the kernel does for
the connections that it handles.
Mark Johnston [Wed, 21 Apr 2021 19:38:01 +0000 (15:38 -0400)]
Add required checks for unmapped mbufs in ipdivert and ipfw
Also add an M_ASSERTMAPPED() macro to verify that all mbufs in the chain
are mapped. Use it in ipfw_nat, which operates on a chain returned by
m_megapullup().
PR: 255164
Reviewed by: ae, gallatin
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D29838
Stop free() even if kvm_getprocs as we can come back but set nprocs = 0.
Check nprocs in showpigs() to ensure not try displaying with kvm_getprocs failed.
Current code can have pt with non-null after kvm_getprocs() failure.
If VOP_ADD_WRITECOUNT() or adv locking failed, so VOP_CLOSE() needs to
be called, we cannot use fp fo_close() when there is no fp. This occurs
when e.g. kernel code directly calls vn_open() instead of the open(2)
syscall.
In this case, VOP_CLOSE() can be called directly, after possible lock
upgrade.
Reported by: nvass@gmx.com
PR: 255119
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D29830
rc: make the 'linux' script explicitly load filesystem modules
This removes a minor annoyance with Linux jails, where you often
want linux_mounts_enable="NO", yet you want those filesystems
available for mounting in jails; normally mount(8) would result
in kernel automatically loading the KLD, but this doesn't work
inside jails or chroots.
Two of these tests now pass. Looking at Jenkins to find the first commit
where this behaviour changed indicates that 2fe5a79425c79f7b828acd91da66d97230925fc8 is the most likely cause.
Ka Ho Ng [Wed, 21 Apr 2021 08:19:15 +0000 (16:19 +0800)]
sndstat: nvlist schema and API definition changes
- SNDSTAT_LABEL_* are renamed to SNDST_DSPS_*, and SNDSTAT_LABEL_DSPS
becomes SNDST_DSPS.
- Centralize channel number/rate/formats into a single nvlist
The above nvlist is named "info_play" and "info_rec"
- Expose only encoding format in pfmts/rfmts. Userland has no direct
access to AFMT_ENCODING/CHANNEL/EXTCHANNEL macros, thus it serves no
meaning to expose too much information through this pair of labels.
However pminrate/rminrate, pmaxrate/rmaxrate, pfmts/rfmts are
deprecated and will be removed in future.
This commit keeps ioctls ABI compatibility with __FreeBSD_version 1400006 for now. In future the compat ABI with 1400006 will be removed
once audio/virtual_oss is rebuilt.
Sponsored by: The FreeBSD Foundation
Reviewed by: hselasky
Approved by: philip (mentor)
Differential Revision: https://reviews.freebsd.org/D29770
Currently, PCB caching mechanism relies on the rib generation
counter (rnh_gen) to invalidate cached nhops/LLE entries.
With certain fib algorithms, it is now possible that the
datapath lookup state applies RIB changes with some delay.
In that scenario, PCB cache will invalidate on the RIB change,
but the new lookup may result in the same nexthop being returned.
When fib algo finally gets in sync with the RIB changes, PCB cache
will not receive any notification and will end up caching the stale data.
To fix this, introduce additional counter, rnh_gen_rib, which is used
only when FIB_ALGO is enabled.
This counter is incremented by the control plane. Each time when fib algo
synchronises with the RIB, it updates rnh_gen to the current rnh_gen_rib value.
Address multiple issues with strict rtsock message validation.
D28668 "normalisation" approach was based on the assumption that
we always have at least "standard" sockaddr len.
It turned out to be false - certain older applications like quagga
or routed abuse sin[6]_len field and set it to the offset to the
first fully-zero bit in the mask. It is impossible to normalise
such sockaddrs without reallocation.
With that in mind, change the approach to use a distinct memory
buffer for the altered sockaddrs. This allows supporting the older
software while maintaining the guarantee on the "standard" sockaddrs.
PR: 255273,255089
Differential Revision: https://reviews.freebsd.org/D29826
MFC after: 3 days
John Baldwin [Tue, 20 Apr 2021 20:22:11 +0000 (13:22 -0700)]
etcupdate: Gracefully handle SIGINT when building trees.
Run the 'build_tree' function inside of a subshell and trap SIGINT to
return an error to the caller. This allows callers to gracefully
cleanup a partially created tree.
While here, redirect stdout/stderr of the subshell to the log file
instead of applying redirections individually to each command executed
while building the tree.
John Baldwin [Tue, 20 Apr 2021 20:21:42 +0000 (13:21 -0700)]
etcupdate: Always extract to a temporary tree.
etcupdate has had a somewhat nasty race condition since its creation
in that its state machine can get very confused if it is interrupted
while building the tree to compare against. This is exacerbated by
the fact that etcupdate doesn't emit any output while building the
tree which can take several seconds (especially in recent years with
the addition of the tree-wide buildconfig/installconfig passes).
To mitigate this, always install a new tree into a temporary directory
created via mktemp as was previously done only for dry-runs via -n.
The existing trees are only rotated and the new tree installed as
/var/db/etcupdate/current after the update command has completed.
iwnstats: fix build with clang and allow install under /usr/local/sbin
iwnstats was not compiling because of some issues raised by the clang
compiler due to -Werror. As a tool it is not connected to world build.
Add missing field "barker_mrc" initialization in struct
iwn_sensitivity_limits for -Wmissing-field-initializers, remove unused
pointer *is on iwn_stats_*_print functions and unused variables for
-Wunused-parameter and -Wunused-variable.
The value for field "barker_mrc" of struct iwn2030_sensitivity_limits
was obtained from linux 3.2 wireless/iwlwifi driver code (iwl-2000.c:115
.barker_corr_th_min_mrc = 390).
Also set BINDIR in Makefile to make it possible to install under
/usr/local/sbin/iwnstats as it require super user.
Reviewed by: adrian
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D29800
Gleb Smirnoff [Mon, 22 Mar 2021 22:32:57 +0000 (15:32 -0700)]
tcp_hostcache: use SMR for lookups, mutex(9) for updates.
In certain cases, e.g. a SYN-flood from a limited set of hosts,
the TCP hostcache becomes the main contention point. To solve
that, this change introduces lockless lookups on the hostcache.
The cache remains a hash, however buckets are now CK_SLIST. For
updates a bucket mutex is obtained, for read an SMR section is
entered.
tcp_input: always request read-locking of PCB for any pure SYN segment.
This is further rework of 08d9c920275. Now we carry the knowledge of
lock type all the way through tcp_input() and also into tcp_twcheck().
Ideally the rlocking for pure SYNs should propagate all the way into
the alternative TCP stacks, but not yet today.
This should close a race when socket is bind(2)-ed but not yet
listen(2)-ed and a SYN-packet arrives racing with listen(2), discovered
recently by pho@.
tcp: Deal with DSACKs, and adjust rescue hole on success.
When a rescue retransmission is successful, rather than
inserting new holes to the left of it, adjust the old
rescue entry to cover the missed sequence space.
Also, as snd_fack may be stale by that point, pull it forward
in order to never create a hole left of snd_una/th_ack.
Finally, with DSACKs, tcp_sack_doack() may be called
with new full ACKs but a DSACK block. Account for this
eventuality properly to keep sacked_bytes >= 0.
MFC after: 3 days
Reviewed By: kbowling, tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D29835
This change makes the TCP LRO code more generic and flexible with regards
to supporting multiple different TCP encapsulation protocols and in general
lays the ground for broader TCP LRO support. The main job of the TCP LRO code is
to merge TCP packets for the same flow, to reduce the number of calls to upper
layers. This reduces CPU and increases performance, due to being able to send
larger TSO offloaded data chunks at a time. Basically the TCP LRO makes it
possible to avoid per-packet interaction by the host CPU.
Because the current TCP LRO code was tightly bound and optimized for TCP/IP
over ethernet only, several larger changes were needed. Also a minor bug was
fixed in the flushing mechanism for inactive entries, where the expire time,
"le->mtime" was not always properly set.
To avoid having to re-run time consuming regression tests for every change,
it was chosen to squash the following list of changes into a single commit:
- Refactor parsing of all address information into the "lro_parser" structure.
This easily allows to reuse parsing code for inner headers.
- Speedup header data comparison. Don't compare field by field, but
instead use an unsigned long array, where the fields get packed.
- Refactor the IPv4/TCP/UDP checksum computations, so that they may be computed
recursivly, only applying deltas as the result of updating payload data.
- Make smaller inline functions doing one operation at a time instead of
big functions having repeated code.
- Refactor the TCP ACK compression code to only execute once
per TCP LRO flush. This gives a minor performance improvement and
keeps the code simple.
- Use sbintime() for all time-keeping. This change also fixes flushing
of inactive entries.
- Try to shrink the size of the LRO entry, because it is frequently zeroed.
- Removed unused TCP LRO macros.
- Cleanup unused TCP LRO statistics counters while at it.
- Try to use __predict_true() and predict_false() to optimise CPU branch
predictions.
Bump the __FreeBSD_version due to changing the "lro_ctrl" structure.
Rick Macklem [Tue, 20 Apr 2021 00:51:07 +0000 (17:51 -0700)]
nfsd: fix stripe size reply for the File Layout pNFS server
At a recent testing event I found out that I had misinterpreted
RFC5661 where it describes the stripe size in the File Layout's
nfl_util field. This patch fixes the pNFS File Layout server
so that it returns the correct value to the NFSv4.1/4.2 pNFS
enabled client.
This affects almost no one, since pNFS server configurations
are rare and the extant pNFS aware NFS clients seemed to
function correctly despite the erroneous stripe size.
It *might* be needed for correct behaviour if a recent
Linux client mounts a FreeBSD pNFS server configuration
that is using File Layout (non-mirrored configuration).
Alex Richardson [Mon, 19 Apr 2021 23:22:45 +0000 (00:22 +0100)]
Allow lib/msun/logarithm_test to pass on ld128 platforms
For some reason the ld128 log1pl() implementation is less accurate than
logl(), but does at least guarantee precision >= the ld80 implementation.
Mark log1p_accuracy_tests as XFAIL for ld128 and increase the log1p tolerance
to the ld80 equivalent in accuracy_tests to avoid losing test coverage for
the other functions.
PR: 253984
Reviewed By: ngie, dim
Differential Revision: https://reviews.freebsd.org/D29039
Alex Richardson [Mon, 19 Apr 2021 23:19:20 +0000 (00:19 +0100)]
libc/string/memset.c: Use unsigned long for stores
While most 64-bit architectures have an assembly implementation of this
file, RISC-V does not. As we now store 8 bytes instead of 4 it should speed
up RISC-V.
Alex Richardson [Mon, 19 Apr 2021 23:15:57 +0000 (00:15 +0100)]
libc/string/bcopy.c: Use intptr_t as the copy type
While most 64-bit architectures have an assembly implementation of this
file RISC-V does not. As we now copy 8 bytes instead of 4 it should speed
up RISC-V. Using intptr_t instead of int also allows using this file for
CHERI pure-capability code since trying to copy pointers using integer
loads/stores will invalidate pointers.
Alex Richardson [Mon, 19 Apr 2021 23:05:30 +0000 (00:05 +0100)]
contrib/flex: Drop local __dead2 patch
Upstream flex has added a yynoreturn, so this diff is no longer needed.
Partially reverts r181269. Also regenerate the pre-generated files that
are used for bootstrapping.
Alex Richardson [Tue, 20 Apr 2021 00:46:36 +0000 (01:46 +0100)]
bsd.compiler.mk: detect Apple Clang for cross-builds
Apple clang uses a different versioning scheme, so if we enable or
disable certain warnings for Clang 11+, those might not be supported
in Apple Clang 11+. This adds 'apple-clang' to COMPILER_FEATURES, so that
bootstrap tools Makefiles can avoid warnings on macOS.
Warner Losh [Mon, 19 Apr 2021 22:09:42 +0000 (16:09 -0600)]
config: style for '\0' and NULL
Use NULL for pointers instead of '0' (though hey are the same thing in
these cases). Ditto for using the zero character '\0' instead of a naked
0 (ditto).
Kevin Bowling [Mon, 19 Apr 2021 02:11:27 +0000 (19:11 -0700)]
e1000: Add support for [Tiger, Alder, Meteor] Lake
Add support for current and future client platform PCI IDs. These are
all I219 variants and have no known driver changes versus previous
generation client platform I219 variants.
Kevin Bowling [Fri, 16 Apr 2021 23:15:50 +0000 (16:15 -0700)]
e1000: Correct promisc multicast filter handling
There are a number of issues in the e1000 multicast filter handling
that have been present for a long time. Take the updated approach from
ixgbe(4) which does not have the issues.
The issues are outlined in the PR, in particular this solves crossing
over and under the hardware's filter limit, not programming the
hardware filter when we are above its limit, disabling SBP (show bad
packets) when the tunable is enabled and exiting promiscuous mode, and
an off-by-one error in the em_copy_maddr function.
linuxkpi: guarantee allocations of task and mm for interrupt threads
Create and use zones for task and mm. Reserve items in zones based on the
estimation of the max number of interrupts in the system. Use M_USE_RESERVE
to allow to take reserved items when allocation occurs from the interrupt
thread context.
Of course, this would only work first time we allocate the task for
interrupt thread. If interrupt is deallocated and allocated anew,
creating a new thread, it might be that zone is depleted. It still
should be good enough for practical uses.
sysctl vm.objects: report backing object and swap use
For anonymous objects, provide a handle kvo_me naming the object,
and report the handle of the backing object. This allows userspace
to deconstruct the shadow chain. Right now the handle is the address
of the object in KVA, but this is not guaranteed.
For the same anonymous objects, report the swap space used for actually
swapped out pages, in kvo_swapped field. I do not believe that it is
useful to report full 64bit counter there, so only uint32_t value is
returned, clamped to the max.
For kinfo_vmentry, report anonymous object handle backing the entry,
so that the shadow chain for the specific mapping can be deconstructed.
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D29771
pf: Optionally attempt to preserve rule counter values across ruleset updates
Usually rule counters are reset to zero on every update of the ruleset.
With keepcounters set pf will attempt to find matching rules between old
and new rulesets and preserve the rule counters.
projects / FreeBSD / FreeBSD.git / log