Each TCP connection that uses the system default cc_newreno(4) congestion
control algorithm module leaks a "struct newreno" (8 bytes of memory) at
connection initialisation time. The NULL-pointer dereference is only germane
when using the ABE feature, which is disabled by default.
While at it:
- Defer the allocation of memory until it is actually needed given that ABE is
optional and disabled by default.
- Document the ENOMEM errno in getsockopt(2)/setsockopt(2).
- Document ENOMEM and ENOBUFS in tcp(4) as being synonymous given that they are
used interchangeably throughout the code.
- Fix a few other nits also accidentally omitted from the original patch.
Reported by: Harsh Jain on freebsd-net@
Tested by: tjh@
Differential Revision: https://reviews.freebsd.org/D15358
provisioned for NIC_ETHOFLD and the kernel has option RATELIMIT.
It is possible to use the chip's offload queues for normal NIC Tx and
not just TOE Tx. The difference is that these queues support out of
order processing of work requests and have a per-"flowid" mechanism for
tracking credits between the driver and hardware. This allows Tx for
any number of flows bound to different rate limits to be submitted to a
single Tx queue and the work requests for slow flows won't cause HOL
blocking for the rest.
Sponsored by: Chelsio Communications
by default.
This is the first of a series of commits that will add support for
RATELIMIT kernel option to the base if_cxgbe driver, for use with
ordinary NIC traffic "flows". RATELIMIT is already supported by t4_tom
for the fully-offloaded TCP connections that it handles.
Sponsored by: Chelsio Communications
Add epoch section to struct thread. We can use this to
ennable epoch counter to advance even if a section is
perpetually occupied by a thread.
Approved by: sbruno
(https://svnweb.freebsd.org/base?view=revision&revision=304232)
converting clrbuf() (which clears the entire buffer) to vfs_bio_clrbuf()
(which clears only the new pages that have been added to the buffer).
Failure to properly remove pages from the buffer cache can make
pages that appear not to need clearing to actually have bad random
data in them. See for example base r304232
(https://svnweb.freebsd.org/base?view=revision&revision=304232)
which noted the need to set B_INVAL and B_NOCACHE as well as clear
the B_CACHE flag before calling brelse() to release the buffer.
Rather than trying to find all the incomplete brelse() calls, it
is simpler, though more slightly expensive, to simply clear the
entire buffer when it is newly allocated.
PR: 213507
Submitted by: Damjan Jovanovic
Reviewed by: kib
This implements per-thread counters for PMC sampling. The thread
descriptors are stored in a list attached to the process descriptor.
These thread descriptors can store any per-thread information necessary
for current or future features. For the moment, they just store the counters
for sampling.
The thread descriptors are created when the process descriptor is created.
Additionally, thread descriptors are created or freed when threads
are started or stopped. Because the thread exit function is called in a
critical section, we can't directly free the thread descriptors. Hence,
they are freed to a cache, which is also used as a source of allocations
when needed for new threads.
Approved by: sbruno
Obtained from: jtl
Sponsored by: Juniper Networks, Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15335
When poll() is called via netmap, txsync is initially called,
and if there are no available buffers to reclaim, it waits for the driver
to notify of new buffers. Since the TX IRQ is generally not used in iflib
drivers, this ends up causing a timeout.
Work around this by having the reclaim DELAY(1) if it's initially unable
to reclaim anything, then schedule the tx task, which will spin by
continuously rescheduling itself until some buffers are reclaimed. In
general, the delay is enough to allow some buffers to be reclaimed, so
spinning is minimized.
Reported by: Johannes Lundberg <johalun0@gmail.com>
Reviewed by: sbruno
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15455
if an error is reported while pre-processing the configuration file that
the driver attempted to use.
Also, allow the user to explicitly use the built-in configuration with
hw.cxgbe.config_file="built-in"
MFC after: 2 days
Sponsored by: Chelsio Communications
There is no need to try to resume it after each smaller operations
(putchar, cursor_position, copy, fill).
The resume function already checks if the timer is armed before doing
anything, but it uses an atomic cmpset which is expensive. And resuming
the timer at the end of input processing is enough.
While here, we also skip timer resume if the input is for another
windows than the currently displayed one. I.e. if `ttyv0` is currently
displayed, any changes to `ttyv1` shouldn't resume the timer (which
would refresh `ttyv0`).
By doing the same benchmark as r333669, I get:
* vt(4), before r333669: 1500 ms
* vt(4), with this patch: 760 ms
* syscons(4): 700 ms
... to process input, instead of inside each smaller operations such as
appending a character or moving the cursor forward.
In other words, before we were doing (oversimplified):
teken_input()
<for each input character>
vtterm_putchar()
VTBUF_LOCK()
VTBUF_UNLOCK()
vtterm_cursor_position()
VTBUF_LOCK()
VTBUF_UNLOCK()
Now, we are doing:
vtterm_pre_input()
VTBUF_LOCK()
teken_input()
<for each input character>
vtterm_putchar()
vtterm_cursor_position()
vtterm_post_input()
VTBUF_UNLOCK()
The situation was even worse when the vtterm_copy() and vtterm_fill()
callbacks were involved.
The new callbacks are:
* struct terminal_class->tc_pre_input()
* struct terminal_class->tc_post_input()
They are called in teken_input(), surrounding the while() loop.
The goal is to improve input processing speed of vt(4). As a benchmark,
here is the time taken to write a text file of 360 000 lines (26 MiB) on
`ttyv0`:
* vt(4), unmodified: 1500 ms
* vt(4), with this patch: 1200 ms
* syscons(4): 700 ms
This is on a Haswell laptop with a GENERIC-NODEBUG kernel.
At the same time, the locking is changed in the vt_flush() function
which is responsible to draw the text on screen. So instead of
(indirectly) using VTBUF_LOCK() just to read and reset the dirty area
of the internal buffer, the lock is held for about the entire function,
including the drawing part.
The change is mostly visible while content is scrolling fast: before,
lines could appear garbled while scrolling because the internal buffer
was accessed without locks (once the scrolling was finished, the output
was correct). Now, the scrolling appears correct.
In the end, the locking model is closer to what syscons(4) does.
Differential Revision: https://reviews.freebsd.org/D15302
This change updates arm, arm64 and mips achitectures. Additionally, it
removes redundant checks for kdb_active where it already results in
kdb_reenter() and adds kdb_reenter() calls where they were missing.
Some architectures check the return value of kdb_trap(), but some don't.
I haven't changed any of that.
Some trap handling routines have a return code. I am not sure if I
provided correct ones for returns after kdb_reenter(). kdb_reenter
should never return unless kdb_jmpbufp is NULL for some reason.
Only compile tested for all affected architectures. There can be bugs
resulting from my poor understanding of architecture specific details.
Reported by: jhb
Reviewed by: jhb, eadler
MFC after: 4 weeks
Differential Revision: https://reviews.freebsd.org/D15431
Some of the DEBUG_BUFRING checks are racy, and can lead to
spurious assertions when run under high load. Unhook these
from INVARIANTS until the author can fix or remove them.
Reviewed by: mmacy
Sponsored by: Netflix
When a disk disappears and the periph is invalidated, any I/Os that
are pending with the controller can cause a crash when they
complete. Move to holding the softc reference count taken in dastart()
until the I/O is complete rather than only until xpt_action()
returns. (This approach was suggested by Ken Merry.) This extends
the method used in da to ada, nda, and mda.
Sponsored by: Netflix
Submitted by: Chuck Silvers
When a disk disappears and the periph is invalidated, any I/Os that
are pending with the controller can cause a crash when they
complete. Move to holding the softc reference count taken in dastart()
until the I/O is complete rather than only until xpt_action()
returns. (This approach was suggested by Ken Merry.)
Sponsored by: Netflix
Submitted by: Chuck Silvers
Differential Revision: https://reviews.freebsd.org/D15435
later devices. These caches work akin to the ones found in HDDs/SSDs
that ada(4)/da(4) also enable if existent, but likewise increase the
likelihood of data loss in case of a sudden power outage etc. On the
other hand, write performance is up to twice as high for e. g. 1 GiB
files depending on the actual chip and transfer mode employed.
For maximum data integrity, the usage of eMMC caches can be disabled
via the hw.mmcsd.cache tunable.
- Get rid of the NOP mmcsd_open().
The NFSv4 protocol requires that the server only allow reclaim of state
and not issue any new open/lock state for a grace period after booting.
The NFSv4.0 protocol required this grace period to be greater than the
lease duration (over 2minutes). For NFSv4.1, the client tells the server
that it has done reclaiming state by doing a ReclaimComplete operation.
If all NFSv4 clients are NFSv4.1, the grace period can end once all the
clients have done ReclaimComplete, shortening the time period considerably.
This patch does this. If there are any NFSv4.0 mounts, the grace period
will still be over 2minutes.
This change is only an optimization and does not affect correct operation.
Tested by: andreas.nagy@frequentis.com
MFC after: 2 months
Currently, when using dd(1) to take a VM memory image, the capture never ends,
reading zeroes when it's beyond VM system memory max address.
Return EFAULT when trying to read beyond VM system memory max address.
Reviewed by: imp, grehan, anish
Approved by: grehan
Differential Revision: https://reviews.freebsd.org/D15156
The idea is to calibrate the LAPIC timer just once and only on boot,
given that [at present] the timer constants are global and shared
between all processors.
My primary motivation is to fix a panic that can happen when dynamically
switching to lapic timer. The panic is caused by a recursion on
et_hw_mtx when printing the calibration results to console. See the
review for the details of the panic.
Also, the code should become slightly simpler and easier to read. The
previous code was racy too. Multiple processors could start calibrating
the global constants concurrently, although that seems to have been
benign.
Reviewed by: kib, mav, jhb
MFC after: 3 weeks
Differential Revision: https://reviews.freebsd.org/D15422
If ifma_protospec is NULL when inm_lookup() is called, there
is a dereference in a NULL struct pointer. This ensures that struct is
not NULL before comparing the address.
Reported by: dumbbell
Reviewed by: sbruno
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15440
already close to the limit so increasing the kernel size may cause it to
fail to boot when it runs past the end of allocated memory.
Reported by: manu
Sponsored by: DARPA, AFRL
This is required by programs like sockstat that read variably sized
sysctls such as kern.file. The normal path has no such restriction and
the restriction was added without comment along with initial support for
freebsd32 in 2002 (r100384).
Reviewed by: kib
Sponsored by: DARPA, AFRL
Differential Revision: https://reviews.freebsd.org/D15438
I210 restore functionality if pxeboot rom is enabled on this device.
r333345 attempted to determine if this code was needed or it was some kind
of work around for a problem. Turns out, its definitely a work around for
hardware locking and synchronization that manifests itself if the option
Rom is enabled and is selected as a boot device (there was a PXE attempt).
Reviewed by: mmacy
Differential Revision: https://reviews.freebsd.org/D15439
Creating a pool with a temporary name fails when we also specify custom
dataset properties: this is because we mistakenly call
zfs_set_prop_nvlist() on the "real" pool name which, as expected,
cannot be found because the SPA is present in the namespace with the
temporary name.
Fix this by specifying the correct pool name when setting the dataset
properties.
Author: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Prakash Surya <prakash.surya@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Obtained from: ZFS on Linux, zfsonlinux/zfs@4ceb8dd6fd
MFC after: 1 week
- Driver support for hardware NAT.
- Driver support for swapmac action.
- Validate a request to create a hashfilter against the filter mask.
- Add a hashfilter config file for T5.
Sponsored by: Chelsio Communications
in the commit log of r321385 has been confirmed via the public VLI54
erratum. Thus, stop advertising DDR52 for these controllers.
Note that this change should hardly make a difference in practice as
eMMC chips from the same era as these SoCs most likely support HS200
at least, probably even up to HS400ES.
Use the new epoch based reclamation API. Now the hot paths will not
block at all, and the sx lock is used for the softc data. This fixes LORs
reported where the rwlock was obtained when the sxlock was held.
Submitted by: mmacy
Reported by: Harry Schmalzbauer <freebsd@omnilan.de>
Reviewed by: sbruno
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15355
Kernel debuggers depend on symbol names to find stack frames with a
trapframe rather than a normal stack frame. The labels used for the
shared interrupt entry point for the PTI and non-PTI cases did not
match the existing patterns confusing debuggers. Add the '.L' prefix
to mark these symbols as local so they are not visible in the symbol
table.
Reviewed by: kib
MFC after: 1 week
Sponsored by: Chelsio Communications
The length of SCTP packets is always a multiple of 4. Therefore,
ensure that the MTUs used are also a multiple of 4.
Thanks to Irene Ruengeler for providing an earlier version of this
patch.
MFC after: 1 week
r333273 and partially reverted with r333594.
Older CPUs implement addition of offsets into the page table by a
bitwise OR rather than actual addition, which only works if the table is
aligned at a multiple of its own size (they also require it to be aligned
at a multiple of 256KB). Newer ones do not have that requirement, but it
hardly matters to enforce it anyway.
The original code was failing on newer systems with huge amounts of RAM
(> 512 GB), in which the page table was 4 GB in size. Because the
bootstrap memory allocator took its alignment parameter as an int, this
turned into a 0, removing any alignment constraint at all and making
the MMU fail. The first round of this patch (r333273) fixed this case by
aligning it at 256 KB, which broke older CPUs. Fix this instead by widening
the alignment parameter.
Once a pmc owner is added to the pmc_ss_owners list it is
visible for all to see. We don't want this to happen until
setup is complete.
Reported by: mjg
Approved by: sbruno
- fix load/unload race by allocating the per-domain list structure at boot
- fix long extant vm map LOR by replacing pmc_sx sx_slock with global_epoch
to protect the liveness of elements of the pmc_ss_owners list
Reported by: pho
Approved by: sbruno
The INVARIANTS checks in epoch_wait() were intended to
prevent the block handler from returning with locks held.
What it in fact did was preventing anything except Giant
from being held across it. Check that the number of locks
held has not changed instead.
Approved by: sbruno@
In the reply to an ExchangeID operation, the NFSv4.1 server returns a
"scope" value (eir_server_scope). If this value is the same, it indicates
that two servers share state, which is never the case for FreeBSD servers.
As such, the value needs to be unique and it was without this patch.
However, I just found out that it is not supposed to change when the
server reboots and without this patch, it did change.
This patch fixes eir_server_scope so that it does not change when the
server is rebooted.
The only affect not having this patch has is that Linux clients don't
reclaim opens and locks after a server reboot, which meant they lost
any byte range locks held before the server rebooted.
It only affects NFSv4.1 mounts and the FreeBSD NFSv4.1 client was not
affected by this bug.
MFC after: 1 week
- GC the _nopreempt routines
- to really benefit we'd need a separate routine
- they're not currently in use
- they complicate the API for no benefit at this time
- check that we're actually in a epoch section at exit
- handle epoch_call() early in boot
- Fix copyright declaration language
Approved by: sbruno@
vm_page_queue(), added in r333256, generalizes vm_pageout_page_queued(),
so use it instead. No functional change intended.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D15402
For a fairly rare case of a client doing an ExchangeID after a hard reboot,
the old confirmed clientid still exists, but some clients use a new
co_verifier. For this case, the server was not freeing up the sessions on
the old confirmed clientid.
This patch fixes this case. It also adds two LIST_INIT() macros, which are
actually no-ops, since the structure is malloc()d with M_ZERO so the pointer
is already set to NULL.
It should have minimal impact, since the only way I could exercise this
code path was by doing a hard power cycle (pulling the plus) on a machine
running Linux with a NFSv4.1 mount on the server.
Originally spotted during testing of the ESXi 6.5 client.
Tested by: andreas.nagy@frequentis.com
MFC after: 2 months
When an NFSv4.1 session is busy due to a callback being in progress,
nfsrv_freesession() should return NFSERR_BACKCHANBUSY instead of NFS_OK.
The only effect this has is that the DestroySession operation will report
the failure for this case and this probably has little or no effect on a
client. Spotted by inspection and no failures related to this have been
reported.
MFC after: 2 months
- Create getblkx(9) variant of getblk(9) which can return error.
- Add GB_NOSPARSE flag for getblk()/getblkx() which requests that BMAP
was performed before the buffer is created, and EJUSTRETURN returned
in case the requested block does not exist.
- Make ffs_read() use GB_NOSPARSE to avoid instantiating buffer (and
allocating the pages for it), copying from zero_region instead.
The end result is less page allocations and buffer recycling when a
hole is read, which is important for some benchmarks.
Requested and reviewed by: jeff
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D14917
It appears that domain information is set correctly independent
of whether or not NUMA is defined. However, there is no memory
backing secondary domains leading to allocation failure.
Reported by: pho@, np@
Approved by: sbruno@
unwind_frame() may be instrumented by FBT, leading to recursion into
dtrace_probe(). Manually inline unwind_frame() as we do with stack
unwinding code for other architectures.
Submitted by: Domagoj Stolfa
Reviewed by: manu
MFC after: 1 week
Sponsored by: DARPA / AFRL
Differential Revision: https://reviews.freebsd.org/D15359
Don't enable regulator on attach but dealt with them on power_up/power_off
Only set the voltage for the signaling regulator since I don't have boards
that can change the supply voltage.
Enable 1.8v signaling voltage.
Only do a reset of the controller at attach and init it at power_up.
We use to enable some interrupts in reset, only enable the interrupts
we are interested in when doing a request.
While here remove the regulators handling in power_on as it is very wrong
and will be dealt with in another commit.
Tested on: A31, A64
is executed on the right stack already. No copy from the entry stack
to the kstack must be performed for vm86 bios call code to function.
To access the pcb flags on kernel entry, unconditionally switch to
kernel address space if vm86 mode is detected.
This fixes very early vm86 bios calls, typically done when boot is
performed by boot2 without loader, and kernel falls back to BIOS calls
to get SMAP.
Reported by: bde
Sponsored by: The FreeBSD Foundation
PCB_VM86CALL pcb flag not set should be treated same as return to
userspace.
Most important, the address space must be switched. This fixes
usermode vm86 operations after the 4/4 split.
Sponsored by: The FreeBSD Foundation
exception code is copied to the trampoline.
The correct value is then copied to trampoline automatically, so
tramp_idleptd_reloced can be eliminated.
This will allow to use the same exception entry code to handle traps
from vm86 bios calls on early boot stage, as after the trampoline is
configured.
Sponsored by: The FreeBSD Foundation
Record common_tssd, the descriptor to be written in GDT to point to
the common TSS, before LTR is executed. The LTR instruction sets the
loaded descriptor type to 386 TSS busy, which traps on reloads.
Sponsored by: The FreeBSD Foundation
On non-trivial SMP systems the contention on the pmc_owner mutex leads
to a substantial number of samples captured being from the pmc process
itself. This change a) makes buffers larger to avoid contention on the
global list b) makes the working sample buffer per cpu.
Run pmcstat in the background (default event rate of 64k):
pmcstat -S UNHALTED_CORE_CYCLES -O /dev/null sleep 600 &
Before:
make -j96 buildkernel -s >&/dev/null 3336.68s user 24684.10s system 7442% cpu 6:16.50 total
After:
make -j96 buildkernel -s >&/dev/null 2697.82s user 1347.35s system 6058% cpu 1:06.77 total
For more realistic overhead measurement set the sample rate for ~2khz
on a 2.1Ghz processor:
pmcstat -n 1050000 -S UNHALTED_CORE_CYCLES -O /dev/null sleep 6000 &
Collecting 10 samples of `make -j96 buildkernel` from each:
x before
+ after
real time:
N Min Max Median Avg Stddev
x 10 76.4 127.62 84.845 88.577 15.100031
+ 10 59.71 60.79 60.135 60.179 0.29957192
Difference at 95.0% confidence
-28.398 +/- 10.0344
-32.0602% +/- 7.69825%
(Student's t, pooled s = 10.6794)
system time:
N Min Max Median Avg Stddev
x 10 2277.96 6948.53 2949.47 3341.492 1385.2677
+ 10 1038.7 1081.06 1070.555 1064.017 15.85404
Difference at 95.0% confidence
-2277.47 +/- 920.425
-68.1574% +/- 8.77623%
(Student's t, pooled s = 979.596)
x no pmc
+ pmc running
real time:
HEAD:
N Min Max Median Avg Stddev
x 10 58.38 59.15 58.86 58.847 0.22504567
+ 10 76.4 127.62 84.845 88.577 15.100031
Difference at 95.0% confidence
29.73 +/- 10.0335
50.5208% +/- 17.0525%
(Student's t, pooled s = 10.6785)
patched:
N Min Max Median Avg Stddev
x 10 58.38 59.15 58.86 58.847 0.22504567
+ 10 59.71 60.79 60.135 60.179 0.29957192
Difference at 95.0% confidence
1.332 +/- 0.248939
2.2635% +/- 0.426506%
(Student's t, pooled s = 0.264942)
system time:
HEAD:
N Min Max Median Avg Stddev
x 10 1010.15 1073.31 1025.465 1031.524 18.135705
+ 10 2277.96 6948.53 2949.47 3341.492 1385.2677
Difference at 95.0% confidence
2309.97 +/- 920.443
223.937% +/- 89.3039%
(Student's t, pooled s = 979.616)
patched:
N Min Max Median Avg Stddev
x 10 1010.15 1073.31 1025.465 1031.524 18.135705
+ 10 1038.7 1081.06 1070.555 1064.017 15.85404
Difference at 95.0% confidence
32.493 +/- 16.0042
3.15% +/- 1.5794%
(Student's t, pooled s = 17.0331)
Reviewed by: jeff@
Approved by: sbruno@
Differential Revision: https://reviews.freebsd.org/D15155
The Linux client now uses the TestStateID operation, so this patch adds
support for it to the NFSv4.1 server. The FreeBSD client never uses this
operation, so it should not be affected.
MFC after: 2 months
r333175 updated the join_group functions, but not the leave_group ones.
Reviewed by: sbruno
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15393
Part 3 of many ...
The VPC framework relies heavily on cloning pseudo interfaces
(vmnics, vpc switch, vcpswitch port, hostif, vxlan if, etc).
This pulls in that piece. Some ancillary changes get pulled
in as a side effect.
Reviewed by: shurd@
Approved by: sbruno@
Sponsored by: Joyent, Inc.
Differential Revision: https://reviews.freebsd.org/D15347
This hardware isn't totally ancient, about equal to a mxge(4) or mlx4en(4),
but the company was sold to Exar which then promptly exited the Ethernet
business so the card was commercially available for under 2 years. On deep
search, the only usage of these cards I found was by the importing of the
driver. There are code quality issues identified by Brooks and Hiren and
no visible use nor maintainership that warrant removal from FreeBSD 12.0.
Submitted by: kbowling
Reviewed by: gnn brooks
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15363
It is guaranteed that if_ipsec(4) interface is used only for tunnel
mode IPsec, i.e. decrypted and decapsultaed packet has its own IP header.
Thus we can consider it as new packet and clear the protocols flags.
This allows ICMP/ICMPv6 properly handle errors that may cause this packet.
PR: 228108
MFC after: 1 week
sys/dev/ocs/ocs_os.c:
In ocs_thread_create(), use sizeof(*thread) (instead of
sizeof(thread)) as the size argument to memset so that we clear
out the entire thread structure instead of just a few bytes of it.
Submitted by: jtl
MFC after: 3 days
LAN7800 USB 3.1 to 10/100/1000 Ethernet with PHY
LAN7801 USB 3.1 to 10/100/1000 Ethernet with RGMII interface
Also update manufacturer name for the Vendor ID. Microchip acquired
SMSC in May 2012.
Sponsored by: The FreeBSD Foundation
This aligns the section itself compared to other sections, does not change
internal alignment of fields stored inside. This may or may not come later.
The motivation is partially combating adverse effects of the adjacent cache
line prefetcher. Without the annotation part of read_mostly section was on
the line of fire.
- initialize the pcpu STAILQ in the NUMA case
- don't enqueue the callback task if there isn't sufficient work to be done
Reported by: pho@
Approved by: sbruno@
Current UMA internals are not suited for efficient operation in
multi-socket environments. In particular there is very common use of
MAXCPU arrays and other fields which are not always properly aligned and
are not local for target threads (apart from the first node of course).
Turns out the existing UMA_ALIGN macro can be used to mostly work around
the problem until the code get fixed. The current setting of 64 bytes
runs into trouble when adjacent cache line prefetcher gets to work.
An example 128-way benchmark doing a lot of malloc/frees has the following
instruction samples:
before:
kernel`lf_advlockasync+0x43b 32940
kernel`malloc+0xe5 42380
kernel`bzero+0x19 47798
kernel`spinlock_exit+0x26 60423
kernel`0xffffffff80 78238
0x0 136947
kernel`uma_zfree_arg+0x46 159594
kernel`uma_zalloc_arg+0x672 180556
kernel`uma_zfree_arg+0x2a 459923
kernel`uma_zalloc_arg+0x5ec 489910
after:
kernel`bzero+0xd 46115
kernel`lf_advlockasync+0x25f 46134
kernel`lf_advlockasync+0x38a 49078
kernel`fget_unlocked+0xd1 49942
kernel`lf_advlockasync+0x43b 55392
kernel`copyin+0x4a 56963
kernel`bzero+0x19 81983
kernel`spinlock_exit+0x26 91889
kernel`0xffffffff80 136357
0x0 239424
See the review for more details.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D15346
if_bridge has a lot of limitations that make it scale poorly to higher data
rates. In my projects/VPC branch I leverage the bridge interface between
layers for my high speed soft switch as well as for purposes of stacking
in general.
Reviewed by: sbruno@
Approved by: sbruno@
Differential Revision: https://reviews.freebsd.org/D15344
- Lend priority to preempted threads in epoch_wait to handle the case
in which we've had priority lent to us. Previously we borrowed the
priority of the lowest priority preempted thread. (pointed out by mjg@)
- Don't attempt allocate memory per-domain on powerpc, we don't currently
handle empty sockets (as is the case on jhibbits Talos' board).
- Handle deferred callbacks as pcpu lists and poll the lists periodically.
Currently the interval is 1/hz.
- Drop the thread lock when adaptive spinning. Holding the lock starves
other threads and can even lead to lockups.
- Keep a generation count pcpu so that we don't keep spining if a thread
has left and re-entered an epoch section.
- Actually removed the callback from the callback list so that we don't
double free. Sigh ...
Approved by: sbruno@
Make if_printf() use vlog() instead of vprintf(). This means it can no
longer return the number of characters printed, as it used to, but every
single call to if_printf() in the entire kernel ignores the return value
anyway; just return 0 so we don't have to change the prototype.
Consistently use if_printf() throughout sys/net/if.c, instead of a
mixture of if_printf() and log().
In ifa_maintain_loopback_route(), don't needlessly log an error if we
either failed to add a route because it already existed or failed to
remove one because it did not. We still return an error code, though.
MFC after: 1 week
While <sys/sysctl.h> includes <sys/queue.h> unconditionally, it is only
actually used in code which is conditional on _KERNEL. Make the #include
itself conditional as well, and fix userland code that uses <sys/queue.h>
for other purposes but relied on <sys/sysctl.h> to bring it in.
MFC after: 1 week
As of r333461 we require ifunc support to link a working amd64 kernel.
The default in-tree bootstrap linker is lld and it has the required
support, as does any modern out-of-tree binutils linker. The in-tree
GNU ld is from binutils 2.17.50 and it does not have ifunc support,
so produce an error rather than a broken kernel.
Reviewed by: kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D15378
Read locking is over used in the kernel to guarantee liveness. This API makes
it easy to provide livenes guarantees without atomics.
Includes epoch_test kernel module to stress test the API.
Documentation will follow initial use case.
Test case and improvements to preemption handling in response to discussion
with mjg@
Reviewed by: imp@, shurd@
Approved by: sbruno@
- Remove unused or dead store variable
- Remove unused function ctl_copyin_alloc
- Add missing curly brackets, this seems a regression in r287720
Reviewed by: jhibbits
Differential Revision: https://reviews.freebsd.org/D15383
Before this change, the VGA palette was configured to match the shell
palette (e.g. color #1 was red). There was one glitch early in boot when
the vt(4)'s VGA palette was loaded: the loader's logo would switch from
red to blue. Likewise for the "Booting..." message switching from blue
to red. That's because the loader's logo was drawed with the default VGA
palette where a few colors are swapped compared to the shell palette
(e.g. blue <-> red).
This change configures the default VGA palette during initialization and
converts input's colors from shell to VGA palette index.
There should be no visible changes, except the loader's logo which will
keep its original color.
Reviewed by: eadler
This makes it more logical:
1. It checks the requested color format
2. It fills the palette accordingly
Also vt_palette_init() is only called when needed (i.e. when the format
is `COLOR_FORMAT_RGB`).
Use LIST_FOREACH_SAFE in in6m_disconnect() since we're
deleting and freeing item from the membership list
while traversing the list.
Reviewed by: mmacy
Sponsored by: Netflix
From now on, linking amd64 kernel requires either lld or newer ld.bfd.
Reviewed by: jhb (as part of the large patch)
Discussed with: emaste
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D13838
Header length is optional hint for the ENA device. Because It is not
guaranteed that every packet header will be in the first mbuf
segment, it is better to skip passing any information. If the header
length will be indicating invalid value (different than 0), then the
packet will be dropped.
This kind situation can appear, when the UDP packet will be fragmented
by the stack in the ip_fragment() function.
Submitted by: Michal Krawczyk <mk@semihalf.com>
Reported by: Krishna Yenduri <kyenduri@brkt.com>
Obtained from: Semihalf
Sponsored by: Amazon, Inc.
It is useful to enter kdb with an escape sequence.
While here move the USB_DEBUG with the others debug options and define
nooptions USB_DEBUG for GENERIC-NODEBUG
On AWS, a network interface can get reinitialized every 30 minutes due
to the MTU being (re)set when a new DHCP lease is obtained. This can
cause packet drop, along with annoying syslog messages.
Skip setting the MTU in the ena driver if the new MTU is the same as the
old MTU. Note this fix is already in the netfront driver.
Testing: Verified ena up/down messages do not appear every 30 min in
/var/log/messages with the fix in place.
Submitted by: Krishna Yenduri <kyenduri@brkt.com>
Reviewed by: Michal Krawczyk <mk@semihalf.com>
* Change ena-com BIT macro to work on unsigned value.
To make the shifting operations safer, they should be working on
unsigned values.
* Fix a mutex not owned ASSERT panic in ENA control path.
A thread calling cv_broadcast()/cv_signal() must hold the mutex used for
cv_wait(). Fix the ENA control path code that has this problem.
Submitted by: Krishna Yenduri <kyenduri@brkt.com>
Reviewed by: Michal Krawczyk <mk@semihalf.com>
Tested by: Michal Krawczyk <mk@semihalf.com>
This had been the default behavior but was changed accidentally as part
of the recent iw_cxgbe+OFED overhaul. Fix another bug in that change
while here: the global knob affects all the adapters in the system and
should be left alone by per-adapter code.
MFC after: 3 days
Sponsored by: Chelsio Communications
Summary:
There were 2 issues that were preventing correct symbol resolution
on PowerPC/pseries:
1- memory corruption at chrp_attach() - this caused the inital
part of the symbol table to become zeroed, which would cause
the kernel linker to fail to parse it.
(this was probably zeroing out other memory parts as well)
2- DDB symbol resolution wasn't working because symtab contained
not relocated addresses but it was given relocated offsets.
Although relocating the symbol table fixed this, it broke the
linker, that already handled this case.
Thus, the fix for this consists in adding a new DDB macro:
DB_STOFFS(offs) that converts a (potentially) relocated offset
into one that can be compared with symbol table values.
PR: 227093
Submitted by: Leandro Lupori <leandro.lupori_gmail.com>
Differential Revision: https://reviews.freebsd.org/D15372
ioctl frontend ports.
This revision introduces two changes to CTL:
- Changes the way options are passed to CTL_LUN_REQ and CTL_PORT_REQ ioctls.
Removes ctl_be_arg structure and associated logic and replaces it with
nv(3)-based logic for passing in and out arguments.
- Allows creating multiple ioctl frontend ports using either ctladm(8) or
ctld(8).
New frontend ports are represented by /dev/cam/ctl<pp>.<vp> nodes, eg /dev/cam/ctl5.3.
Those device nodes respond only to CTL_IO ioctl.
New command-line options for ctladm:
# creates new ioctl frontend port with using free pp and vp=0
ctladm port -c
# creates new ioctl frontend port with pp=10 and vp=0
ctladm port -c -O pp=10
# creates new ioctl frontend port with pp=11 and vp=12
ctladm port -c -O pp=11 -O vp=12
# removes port with number 4 (it's a "targ_port" number, not pp number)
ctladm port -r -p 4
New syntax for ctl.conf:
target ... {
port ioctl/<pp>
...
}
target ... {
port ioctl/<pp>/<vp>
...
Note: Most of this work was made by jceel@, thank you.
Submitted by: jceel
Reworked by: myself
Reviewed by: mav (earlier versions and recently during the rework)
Obtained from: FreeNAS and TrueOS
Relnotes: Yes
Sponsored by: iXsystems Inc.
Differential Revision: https://reviews.freebsd.org/D9299
Rather than include a copy for memmove to call bcopy to call memcpy
(which handles overlapping copies), make memmove a strong reference to
memcpy to save the two calls.
Differential Revision: https://reviews.freebsd.org/D15374
riscv and powerpc have nearly identical bcopy.c that's
supposed to be mostly MI. Move it to the MI libkern.
Differential Revision: https://reviews.freebsd.org/D15374
request, which can be used to configure hardware NAT and swapmac.
All firmwares released after Jan 2017 support this work request.
Sponsored by: Chelsio Communications
If the 'n' flag is provided the provided key number will be used to
decrypt device. This can be used combined with dryrun to verify if the key
is set correctly. This can be also used to determine which key slot we want to
change on already attached device.
Reviewed by: allanjude
Differential Revision: https://reviews.freebsd.org/D15309
- Add macros to allow preinitialization of cap_rights_t.
- Convert most commonly used code paths to use preinitialized cap_rights_t.
A 3.6% speedup in fstat was measured with this change.
Reported by: mjg
Reviewed by: oshogbo
Approved by: sbruno
MFC after: 1 month
r333345 added a panic to the default case statement on the incorrect
premise that it should "never happen" when in fact it is simply a
different adapter version.
Reported by: markj
Approved by: sbruno
Adapt assembly generated by clang for memcmp and use it for <= 64 sized
compares (which are the vast majority).
Sample result of doing stats on Broadwell (% of samples):
before: 4.0 kernel bcmp cache_lookup
after : 0.7 kernel bcmp cache_lookup
The routine is most definitely still not optimal. Anyone interested in
spending time improving it is welcome to take over.
Reviewed by: kib
Evaluate cpu_stdext_feature early to have moved link_elf_ireloc() see
correct flags, most important is SMAP.
Tested by: mjg
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D15367
Use full copyright year.
Remove 'All Rights Reserved' from new file (rights holder OK'd)
Minor #ifdef motion and #endif tagging
Remove __FBSDID macro from comments
Sponsored by: Netflix
OK'd by: rrs@
Supposedly, they PG_U bits there were set to easier making some kernel
page accessible to userspace in-place. Since it was not used for the
whole existence of the amd64 pmap.c and current design of the shared
pages prefers double-mapping over the in-place access, remove PG_U
both from the direct map and KVA slots.
Reviewed by: alc, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
This PML4 page is never used for the userspace process, so there is no
security implications. But the configuration trips SMAP check, which
should be corrected.
Reviewed by: alc, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
o Modify ipfw(8) to be able set any prefix6 not just Well-Known,
and also show configured prefix6;
o relocate some definitions and macros into proper place;
o convert nat64_debug and nat64_allow_private variables to be
VNET-compatible;
o add struct nat64_config that keeps generic configuration needed
to NAT64 code;
o add nat64_check_prefix6() function to check validness of specified
by user IPv6 prefix according to RFC6052;
o use nat64_check_private_ip4() and nat64_embed_ip4() functions
instead of nat64_get_ip4() and nat64_set_ip4() macros. This allows
to use any configured IPv6 prefixes that are allowed by RFC6052;
o introduce NAT64_WKPFX flag, that is set when IPv6 prefix is
Well-Known IPv6 prefix. It is used to reduce overhead to check this;
o modify nat64lsn_cfg and nat64stl_cfg structures to use nat64_config
structure. And respectivelly modify the rest of code;
o remove now unused ro argument from nat64_output() function;
o remove __FreeBSD_version ifdef, NAT64 was not merged to older versions;
o add commented -DIPFIREWALL_NAT64_DIRECT_OUTPUT flag to module's Makefile
as example.
Obtained from: Yandex LLC
MFC after: 1 month
Sponsored by: Yandex LLC
These filters reside in the card's memory instead of its TCAM and can be
configured via a new "hashfilter" subcommand in cxgbetool. Hash and
normal TCAM filters can be used together. The hardware does an
exact-match of packet fields for hash filters, unlike the masked match
performed for TCAM filters. Any T5/T6 card with memory can support at
least half a million hash filters. The sample config file with the
driver configures 512K of these, it is possible to double this to 1
million+ in some cases.
The chip does an exact-match of fields of incoming datagrams with hash
filters and performs the action configured for the filter if it matches.
The fields to match are specified in a "filter mask" in the firmware
config file. The filter mask always includes the 5-tuple (sip, dip,
sport, dport, ipproto). It can, optionally, also include any subset of
the filter mode (see filterMode and filterMask in the firmware config
file).
For example:
filterMode = fragmentation, mpshittype, protocol, vlan, port, fcoe
filterMask = protocol, port, vlan
Exact values of the 5-tuple, the physical port, and VLAN tag would have
to be provided while setting up a hash filter with the chip
configuration above.
Hash filters support all actions supported by TCAM filters. A packet
that hits a hash filter can be dropped, let through (with optional
steering to a specific queue or RSS region), switched out of another
port (with optional L2 rewrite of DMAC, SMAC, VLAN tag), or get NAT'ed.
(Support for some of these will show up in the driver in a follow-up
commit very shortly).
Sponsored by: Chelsio Communications
KTRPOINT() checks both if we are tracing _and_ if we are recursing within
ktrace. The second condition is only ever executed if ktrace is actually
enabled. This change moves the check out of the hot path in to the functions
themselves.
Discussed with mjg@
Reported by: mjg@
Approved by: sbruno@
