There are 3 counters modified all the time in this structure - one for
keeping the structure alive, one for preventing unmount and one for
tracking active writers. Exact values of these counters are very rarely
needed, which makes them a prime candidate for conversion to a per-cpu
scheme, resulting in much better performance.
Sample benchmark performing fstatfs (modifying 2 out of 3 counters) on
a 104-way 2 socket Skylake system:
before: 852393 ops/s
after: 76682077 ops/s
Reviewed by: kib, jeff
Tested by: pho
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21637
New primitive is introduced to denote sections can operate locklessly
on aspects of struct mount, but which can also be disabled if necessary.
This provides an opportunity to start scaling common case modifications
while providing stable state of the struct when facing unmount, write
suspendion or other events.
mnt_ref is the first counter to start being managed in this manner with
the intent to make it per-cpu.
Reviewed by: kib, jeff
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21425
Where open(2) is expected to fail, the tests should assert or expect that
its return value is -1. These tests all accepted too much but happened to
pass anyway.
Reported by: Coverity
Coverity CID: 1404512, 1404378, 1404504, 1404483
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Assert that the refcount value is not VPRC_BLOCKED in vm_page_drop().
VPRC_BLOCKED is a refcount flag used to indicate that a thread is
tearing down mappings of a page. When set, it causes attempts to wire a
page via a pmap lookup to fail. It should never represent the last
reference to a page, so assert this.
Fix a race in vm_page_dequeue_deferred_free() after r352110.
This function loaded the page's queue index before setting PGA_DEQUEUE.
In this window the page daemon may have deactivated the page, updating
its queue index. Make the operation atomic using vm_page_pqstate_cmpset();
the page daemon will not modify the page once it observes that PGA_DEQUEUE
is set.
fusefs: fix some minor issues with fuse_vnode_setparent
* When unparenting a vnode, actually clear the flag. AFAIK this is basically
a no-op because we only unparent a vnode when reclaiming it or when
unlinking.
* There's no need to call fuse_vnode_setparent during reclaim, because we're
about to free the vnode data anyway.
Reviewed by: emaste
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21630
Since TX interrupt is generated when THRE is set, wait for TEMT set means
wait for full character transmission time. At low speeds that may take
awhile, burning CPU time while holding sc_hwmtx lock, also congested.
ian [Sun, 15 Sep 2019 19:38:15 +0000 (19:38 +0000)]
Apply a runtime patch to the FDT data for imx6 to fix iomuxc problems.
The latest imported FDT data defines a node for an iomuxc-gpr device,
which we don't support (or need, right now) in addition to the usual
iomuxc device. Unfortunately, the dts improperly assigns overlapping
ranges of mmio space to both devices. The -gpr device is also a syscon
and simple_mfd device.
At runtime the simple_mfd driver attaches for the iomuxc-gpr node, then
when the real iomuxc driver comes along later, it fails to attach because
it tries to allocate its register space, and it's already partially in
use by the bogus instance of simple_mfd.
This change works around the problem by simply disabling the node for
the iomuxc-gpr device, since we don't need it for anything.
When the IP layer calls back into the SCTP layer to perform the SCTP
checksum computation, do not assume that the IP header chain and the
SCTP common header are in contiguous memory although the SCTP lays
out the mbuf chains that way. If there are IP-level options inserted
by the IP layer, the constraint is not fulfilled anymore.
This issues was found by running syzkaller. Thanks to markj@ who is
running an instance which also provides kernel dumps. This allowed me
to find this issue.
A future change to posixshm to add file sealing (in DIFF_21391[0] and child)
will move locking out of shm_dotruncate as kern_shm_open() will require the
lock to be held across the dotruncate until the seal is actually applied.
For this, the cookie is passed into shm_dotruncate_locked which asserts
RCA_WLOCKED.
[0] Name changed to protect the innocent, hopefully, from getting autoclosed
due to this reference...
ian [Sun, 15 Sep 2019 01:02:01 +0000 (01:02 +0000)]
Make the ti_sysc device quiet. It's an internal utility pseudo-device
that makes the upstream FDT data work right, so we don't need to see a
couple dozen instances of it spam the dmesg at boot time unless it's a
verbose boot.
ian [Sat, 14 Sep 2019 19:33:36 +0000 (19:33 +0000)]
Create a mechanism for encoding a system errno into the IIC_Exxxxx space.
Errors are communicated between the i2c controller layer and upper layers
(iicbus and slave device drivers) using a set of IIC_Exxxxxx constants which
effectively define a private number space separate from (and having values
that conflict with) the system errno number space. Sometimes it is necessary
to report a plain old system error (especially EINTR) from the controller or
bus layer and have that value make it back across the syscall interface
intact.
I initially considered replicating a few "crucial" errno values with similar
names and new numbers, e.g., IIC_EINTR, IIC_ERESTART, etc. It seemed like
that had the potential to grow over time until many of the errno names were
duplicated into the IIC_Exxxxx space.
So instead, this defines a mechanism to "encode" an errno into the IIC_Exxxx
space by setting the high bit and putting the errno into the lower-order
bits; a new errno2iic() function does this. The existing iic2errno()
recognizes the encoded values and extracts the original errno out of the
encoded value. An interesting wrinkle occurs with the pseudo-error values
such as ERESTART -- they aleady have the high bit set, and turning it off
would be the wrong thing to do. Instead, iic2errno() recognizes that lots of
high bits are on (i.e., it's a negative number near to zero) and just
returns that value as-is.
Thus, existing drivers continue to work without needing any changes, and
there is now a way to return errno values from the lower layers. The first
use of that is in iicbus_poll() which does mtx_sleep() with the PCATCH flag,
and needs to return the errno from that up the call chain.
Introduce arb(3), the Array-based Red-Black Tree macros: similar
to the traditional tree(3) RB trees, but using an array (preallocated,
linear chunk of memory) to store the tree.
This avoids allocation overhead, improves memory locality,
and makes it trivially easy to share/transfer/copy the entire tree
without the need for marshalling. The downside is that the size
is fixed at initialization time; there is no mechanism to resize
it.
This is one of the dependencies for the new stats(3) framework
(https://reviews.freebsd.org/D20477).
Make pseudofs(9) create directory entries in order, instead
of the reverse.
This fixes Linux sysctl(8) binary - it assumes the first two
directory entries are always "." and "..". There might be other
Linux apps affected by this.
NB it might be a good idea to rewrite it using queue(3).
Reviewed by: kib
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21550
- Fix the statement that big5 is a de facto standard of Traditional Chinese
text [1]
- Add a BUGS section describes the problem of big5 and suggests use utf8
PR: 189095
Submitted by: Brennan Vincent <brennan@umanwizard.com> [1]
Reviewed by: Ting-Wei Lan <lantw44@gmail.com>
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21622
This command will trigger a reload of the configuration from disk. This is
useful if you've changed currdev from recovery media to local disk as much
as I have over the past ~2 hours and are tired of the extra keystrokes.
This is really just a glorified shortcut, but reload-conf is likely easier
to remember for other people and does save some keystrokes when reloading
the configuration. It is also resilient to the underlying config method
changing interface, but this is unlikely to happen.
powerpc64/powernv: Add opal NVRAM driver for PowerNV systems
Add a very basic NVRAM driver for OPAL which can be used by the IBM
powerpc-utils nvram utility, not to be confused with the base nvram utility,
which only operates on powermac_nvram.
The IBM utility handles all partitions itself, treating the nvram device as
a plain store.
An alternative would be to manage partitions in the kernel, and augment the
base nvram utility to deal with different backing stores, but that
complicates the driver significantly. Instead, present the same interface
IBM's utlity expects, and we get the usage for free.
riscv trap_pfault: remove unneeded hold of the process around vm_fault() call.
This is re-appearance of the nop code removed from other arches in r287625.
Reviewed by: alc (as part of the larger patch), markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
DIfferential revision: https://reviews.freebsd.org/D21645
1. If we release the last usecount we take ownership of the hold count, which
means the vnode will remain allocated until we vdrop it.
2. If someone else vrefs they will find no usecount and will proceed to add
their own hold count.
3. No code has a problem with v_usecount transitioning to 0 without the
interlock
These facts combined mean we can fetchadd instead of having a cmpset loop.
Reviewed by: kib (previous version)
Tested by: pho
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21528
This version bring many fixes regarding unicode support
It also adds proper support for filename completion (we do not need our custom
patches anymore)
Improves the libreadline compatibility
Note that the same work was done by Yuichiro Naito in
https://reviews.freebsd.org/D21196 the main difference is in this case we have
reimported libedit in contrib to fix a long standing mess in the previous merges
which prevented a proper update workflow. (discussed long ago with pfg@)
The only difference with upstream libedit is we have added a compatibility shim
for the _elf_fn_sh_complete function which we previously added to support quoting
in filename completion and is not needed anymore.
This was added to continue supported old /bin/sh binaries and not break backward
compatibility (as discussed with jilles@)
Those scripts are without copyright and license assignement since their creation
After grabbing information from The various authors and contributors assign
proper license header and copyrights.
This has been reported by yuripv in his work on integrating those in Illumos!
[jail] removal by jid doesn't trigger pre/post stop scripts
This commit fixes bug: command "jail -r" didn't trigger pre/post stop
commands (and others) defined in config file if jid is specified insted of
name. Also it adds basic tests for usr.sbin/jail to avoid regression.
Reviewed by: jamie, kevans, ray
MFC after: 5 days
Differential Revision: https://reviews.freebsd.org/D21328
This only lists the changed type and not other attributes so that it
matches the behavior of -C as done in r66747 for fmtree. The NetBSD
-ff implementation was copied from fmtree.
[Bug 240473] add support of Ralink RT5390R Wireless Card
This commit adds PCI ID of Ralink RT5390R into ids table of driver ral.
Tests show stability of card during day. Network speed is reasonable (
around several megabytes per second).
PR: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=240473
Reported by: zetrotrack000@gmail.com
Reviewed by: ray
Approved by: ray
MFC after: 5 days
Differential Revision: https://reviews.freebsd.org/D21604
Use REFCOUNT_COUNT() to obtain refcount where appropriate.
Refcount waiting will set some flag bits in the refcount value.
Make sure these bits get cleared by using the REFCOUNT_COUNT()
macro to obtain the actual refcount.
Get the readline header from the installed header instead of the from the source
location.
With newer import of libedit, the path to be able to access readline/readline.h
will also include header which name will conflict with some expected by ntp in
another path and end up breaking the build.
SIOCSIFNAME: Do nothing if we're not actually changing
Instead of throwing EEXIST, just succeed if the name isn't actually
changing. We don't need to trigger departure or any of that because there's
no change from consumers' perspective.
As I like to forget: static kenv var formatting is actually such that an
empty environment would be double null bytes. We should make sure that a
non-zero buffer has at least enough for this, though most of the current
usage is with a 4k buffer.
kenv: assert that an empty static buffer passed in is "empty"
Garbage in the passed-in buffer can cause problems if any attempts to read
the kenv are inadvertently made between init_static_kenv and the first
kern_setenv -- assuming there is one.
This is cheap and easy, so do it. This also helps rule out some class of
bugs as one tries to debug; tunables fetch from the static environment up
until SI_SUB_KMEM + 1, and many of these buffers are global ~4k buffers that
rely on BSS clearing while others just grab a page of free memory and use it
(e.g. xen).
ig4(4): Fix SDA HOLD time set too low on Skylake controllers
Execution of "Soft reset" command (IG4_REG_RESETS_SKL) at controller init
stage sets SDA_HOLD register value to 0x0001 which is often too low for
normal operation.
Set SDA_HOLD back to 28 after reset to restore controller functionality.
PR: 240339
Reported by: imp, GregV, et al.
MFC after: 3 days
cem [Wed, 11 Sep 2019 21:24:14 +0000 (21:24 +0000)]
buf: Add B_INVALONERR flag to discard data
Setting the B_INVALONERR flag before a synchronous write causes the buf
cache to forcibly invalidate contents if the write fails (BIO_ERROR).
This is intended to be used to allow layers above the buffer cache to make
more informed decisions about when discarding dirty buffers without
successful write is acceptable.
As a proof of concept, use in msdosfs to handle failures to mark the on-disk
'dirty' bit during rw mount or ro->rw update.
Extending this to other filesystems is left as future work.
getsockopt.2: clarify that SO_TIMESTAMP is not 100% reliable
When SO_TIMESTAMP is set, the kernel will attempt to attach a timestamp as
ancillary data to each IP datagram that is received on the socket. However,
it may fail, for example due to insufficient memory. In that case the
packet will still be received but not timestamp will be attached.
Reviewed by: kib
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D21607
fusefs: Fix iosize for FUSE_WRITE in 7.8 compat mode
When communicating with a FUSE server that implements version 7.8 (or older)
of the FUSE protocol, the FUSE_WRITE request structure is 16 bytes shorter
than normal. The protocol version check wasn't applied universally, leading
to an extra 16 bytes being sent to such servers. The extra bytes were
allocated and bzero()d, so there was no information disclosure.
Reviewed by: emaste
MFC after: 3 days
MFC-With: r350665
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21557
ping: Verify whether a datagram timestamp was actually received.
ping(8) uses SO_TIMESTAMP, which attaches a timestamp to each IP datagram at
the time it's received by the kernel. Except that occasionally it doesn't.
Add a check to see whether such a timestamp was actually set before trying
to read it. This fixes segfaults that can happen when the kernel doesn't
attach a timestamp.
The bug has always existed, but prior to r351461 it manifested as an
implausible round-trip-time, not a segfault.
Reported by: pho
MFC after: 3 days
MFC-With: 351461
Avoid unneeded call to arc4random() in syncache_add()
Don't call arc4random() unconditionally to initialize sc_iss, and
then when syncookies are enabled, just overwrite it with the
return value from from syncookie_generate(). Instead, only call
arc4random() to initialize sc_iss when syncookies are not
enabled.
Note that on a system under a syn flood attack, arc4random()
becomes quite expensive, and the chacha_poly crypto that it calls
is one of the more expensive things happening on the
system. Removing this unneeded arc4random() call reduces CPU from
about 40% to about 35% in my test scenario (Broadwell Xeon, 6Mpps
syn flood attack).
Avoid the use of the non-portable -D argument to ls.
This was used to store the mtime of the source file in a commment in a
generated header file. This is of little-to-no diagnostic value and
the result doesn't even end up in the source tree.
riscv: Small fix to CPU compatibility identification
fdt_is_compatible_strict() inspects the first compatible property.
We need to inspect the following properties for 'riscv'.
ofw_bus_node_is_compatible() does a recursive search.
This patch fixes "Can't find CPU" error message when bootverbose = true.
With the recent commit of ktls, we no longer have a
sb_tls_flags, its just the sb_flags. Also the ratelimit
code, now that the defintion is in sockbuf.h, does not
need the ktls.h file (or its predecessor).
- make abday, day, abmon, mon, am_pm output quoting match linux
- workaround localeconv() issue for mon_grouping and grouping (PR172215)
- for other values not available in default locale, output -1 instead of
127 (CHAR_MAX) as returned by localeconv()
With these changes, output of `locale` and `locale -k` for all keywords
specified by POSIX exactly matches the linux one.
fw_stub.awk: use @generated tag in generated files
Multiple tools use @generated to identify generated files (for example,
in a review Phabricator will by default hide diffs in enerated files).
Use the @generated tag in makesyscalls.sh as we've done for other
generated files.
Assume all the short args have optional args so allocate space for the
':'. It's slightly wasteful, but much easier (and the savings in bytes
at runtime would be tiny, but the code to do it larger).
This command simply returns 0 at the moment and explicitly takes no
arguments. This should be used by utilities wanting to see if bectl can
operate on the system they're running, or with a specific root (`bectl -r`).
It may grow more checks than "will libbe successfully init" in the future,
but for now this is enough as that checks for the dataset mounted at "/" and
that it looks capable of being a BE root (e.g. it's not a top-level dataset)
bectl commands can now specify if they want to be silent, and this will turn
off libbe_print_on_error so they can control the output as needed. This is
already used in `bectl check`, and may be turned on in the future for some
other commands where libbe errors are better suppressed as the failure mode
may be obvious.
projects / FreeBSD / FreeBSD.git / log