| Age | Commit message (Collapse) | Author |
|---|
|
This patch adds support for receive checksum offload for TCP/IPv6
and UDP/IPv6. Since receive checksum offload can't be configured
separately for IPv4 and IPv6, IFCAP_RXCSUM and IFCAP_RXCSUM_IPV6
can't be changed independently.
Reviewed by: Timo Völker
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54756
|
|
This patch adds support for transmit checksum offload for TCP/IPv6
and UDP/IPv6.
Reviewed by: Timo Völker
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54754
|
|
- s/vaues/values/
MFC after: 5 days
|
|
No functional change intended.
MFC after: 3 days
|
|
Building the LINT-NOIP kernel on amd64 with clang 21 results in a
-Werror warning similar to:
sys/dev/mxge/if_mxge.c:1846:44: error: variable 'sum' is uninitialized when passed as a const pointer argument here [-Werror,-Wuninitialized-const-pointer]
1846 | cksum_offset, sizeof(sum), (caddr_t)&sum);
| ^~~
Indeed, if both `INET` and `INET6` are undefined, `sum` is never
initialized. Initialize it to zero to silence the warning.
Reviewed by: jhibbits
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54730
|
|
Each C file is compiled separately so these functions can't be inlined
except in the file where they are defined. Since these functions aren't
used outside smartpqi_request.c, just do the simple thing and make them
private to that file.
Reported by: gcc
Reviewed by: jrhall
Fixes: c558eca47970 ("smartpqi: update to version 4660.0.2002")
Differential Revision: https://reviews.freebsd.org/D54732
|
|
The bge interface is special with respect to transmit checksumi
offloading. In the default settings, an bge interface announces TXCSUM
capabilities, but only supports TCP/IPv4 and not UDP/IPv4 due to
limitations of some of the NICs. This results in problems when the bge
interface becomes a member of a bridge. Since currently only the
TXCSUM capabilities are synced when a member is added to a bridge and
not the protocol specific capabilities, this can result in a situation
where UDP packets are sent out using a bge interface without having a
correct checksum.
To mitigate this problem, initially don't announce TXCSUM capabilities,
when UDP transmit checksum is disabled. It is still possible to enable
TXCSUM capabilities via ifconfig.
PR: 291420
Reviewed by: Timo Voelker
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54486
|
|
This patch provides two improvements for TCP/IPv4 and UDP/IPv4
transmit checksum offloading:
(1) Use *CIC_SEG instead of *CIC_FULL, since FreeBSD always provides
a pseudo header checksum.
(2) Don't make transmit IPv4 header checksum offloading a prerequisite
for TCP/IPv4 or UDP/IPv4 transmit checksum offloading.
This is the root cause of PR 291696, since right now the epair
interface does not support transmit IPv4 header checksum offloading,
but TCP/IPv4 and UDP/IPv4 transmit checksum offloading.
PR: 291696
Reviewed by: Timo Voelker
Tested by: Marek Benc
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54395
|
|
This simplifies detach/cleanup for drivers that add multiple ACPI ioctls.
Reviewed by: imp
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D54420
|
|
Fixes: 73c921ef1d44 ("nvme: Add ability to override ioq to put the request on")
|
|
Sometimes the client device needs to manage the IOQ the request goes
to. Expand the interface we have for the request to allow it to be set
for this special use case.
Sponsored by: Netflix
Reviewed by: jhb
Differential Revision: https://reviews.freebsd.org/D54714
|
|
Add support for the MacBookPro11,5 (Mid 2015, 15-inch with AMD Radeon R9 M370X GPU)
to the Apple SMC driver.
Debug testing revealed this model lacks several SMC keys present on MacBookPro11,4
(IBLC, ICMC, IC2C), that model-specific sensor definitions.
Differential Revision: https://reviews.freebsd.org/D54665
Reviewed by: adrian
|
|
The mtw(4) driver works correctly on initial boot, but fails to initialize
the MT7601U WiFi adapter after a warm reboot.
Users must either physically unplug and replug the USB adapter, or perform a
full power cycle to restore functionality, if usb power is always powered
(only a replug works)
The root cause is that warm reboot does not power-cycle USB devices,
leaving the MT7601U in a stale state from the previous session.
The MCU retains its ready flag and the device ignores initialization
commands, resulting in timeout waiting for MCU to initialize errors.
At the OS Level, pinging 1.1.1.1 will work, but the speed will be very
slow. In addition in debug mode, we see thousand of error logs.
This patch addresses the issue by:
* Performing USB re-enumeration on attach to reset the device state
* Detecting when the MCU is already marked ready (stale from previous
session) and forcing a reset of the MCU before loading firmware
* Increasing the firmware load timeout from 3s to 10s to accommodate
slower initialization after reset
* Increasing MCU ready poll attempts from 100 to 300 with longer delays
to handle devices that take longer to become ready after reset
Note: The increase was random, lower value might work.
Test Plan:
Tested on MacBook Pro (late-2015) and a MacMini with MediaTek MT7601U
USB adapter across multiple warm reboot cycles.
With the mac-mini and a another desktop, the issue happens only if
connected via an always powered usb hub port in the monitor.
The laptop don't power cycle it power.
Differential Revision: https://reviews.freebsd.org/D54659
Reviewed by: adrian
|
|
VT_SETMODE ioctl currently checks the provided signal numbers with its
own ISSIGVALID macro that uses NSIG (32) as a maximum, although the code
that will actually send the signal in sys/kern/kern_sig.c uses
_SIG_VALID which allows up to _SIG_MAXSIG (128).
This change aligns the vt code with the kernel internals and enables the
use of higher signal numbers so that applications are not limited to
SIGUSR1 and SIGUSR2 for vt release and acquire signals.
Signed-off-by: Quentin Thébault <quentin.thebault@defenso.fr>
Reviewed by: emaste, imp, kevans
Differential Revision: https://reviews.freebsd.org/D53615
|
|
Implement STYPE_SUSPEND_TO_IDLE sleep type added in c43473dc9b83
("sys/power: Generic sleep types").
This is a prerequisite for the firmware to enter the S0ix states. When
suspending to idle, the system stays in an ACPI S0 state, but the CPUs
are idled and devices are suspended/resumed before and after this as
they would be when entering any other sleep type (except for AWAKE and
POWEROFF).
Factor out do_standby, do_sleep, and add a new do_idle function for
idling the CPU (a future patch will make this an idle loop and not just
a simple cpu_idle() call). In do_idle, SCIs (interrupt 9) are enabled to
allow wake events to break the CPU out of idle.
Record all the steps made instead of just the last one in slp_state,
which allows for more flexible unwinding (will be useful to not have to
goto breakout if the SPMC entry call fails when that is committed).
A lot of this borrows from Ben Widawsky's patch: D17675. The main
functional difference with that patch is that suspend-to-idle is a
wholly separate sleep type in this one as opposed to being an
alternative implementation for s2mem (S3).
Reviewed by: emaste, olce
Approved by: olce
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D48734
|
|
State that the object charge is zero if object->cred == NULL, or equal
to the ptoa(object->size) otherwise.
Besides being much simpler, the transition to use object->size corrects
the architectural issue with the use of object->charge. The split
operations effectively carve the holes in the charged regions, but
single counter cannot properly express it. As result, coalescing
anonymous mappings cannot calculate correctly if the extended mapping
already backed by the existing object is already accounted or not [1].
To properly solve the issue, either we need to start tracking exact
charged regions in the anonymous objects, which has the significant
overhead and complications. Or give up on the slight over-accounting
and charge the whole object unconditionally, as it is done in the patch.
Reported by: mmel, pho [1]
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D54572
|
|
|
|
We use min() in most places.
No functional change intended.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
|
|
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D54130
|
|
Replace CTLFLAG_NEEDGIANT with CTLFLAG_MPSAFE for all ADC sensor
sysctls. All of the sysctl handlers are serialized by a driver mutex.
Sysctls converted:
- dev.ads111x.X.channelN.gain_index (PGA setting)
- dev.ads111x.X.channelN.rate_index (sample rate)
- dev.ads111x.X.channelN.voltage (sampled voltage)
- dev.ads111x.X.config (configuration register)
- dev.ads111x.X.lo_thresh (comparator low threshold)
- dev.ads111x.X.hi_thresh (comparator high threshold)
Reviewed by: markj
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D54620
|
|
Replace CTLFLAG_NEEDGIANT with CTLFLAG_MPSAFE for all light sensor
sysctls. All of the sysctl handlers are serialized by a driver mutex.
Sysctls converted:
- dev.isl.X.als (ambient light sensor)
- dev.isl.X.ir (infrared sensor)
- dev.isl.X.prox (proximity sensor)
- dev.isl.X.resolution (sensor resolution)
- dev.isl.X.range (sensor range)
Reviewed by: markj
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D54621
|
|
Replace CTLFLAG_NEEDGIANT with CTLFLAG_MPSAFE for all interrupt
coalescing sysctls. The driver uses jme_mtx mutex for proper
synchronization and does not require Giant lock.
Sysctls converted:
- dev.jme.X.tx_coal_to (TX coalescing timeout)
- dev.jme.X.tx_coal_pkt (TX coalescing packet count)
- dev.jme.X.rx_coal_to (RX coalescing timeout)
- dev.jme.X.rx_coal_pkt (RX coalescing packet count)
- dev.jme.X.process_limit (max RX events to process)
Reviewed by: markj
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D54618
|
|
IPsec rx uses a new CPL and this prepares eth_rx for IPsec support in
future commits.
Tidy up the comment and code around M_TSTMP while here.
MFC after: 1 week
Sponsored by: Chelsio Communications
|
|
Obtained from: Chelsio Communications
MFC after: 1 week
Sponsored by: Chelsio Communications
|
|
This fixes a build error with -O0 and matches all the other pqisrc_is_*
routines in the file.
--- kernel.full ---
ld: error: undefined symbol: pqisrc_is_supported_write
>>> referenced by smartpqi_request.c:1799 (/.../smartpqi_request.c:1799)
>>> smartpqi_request.o:(pqisrc_build_scsi_cmd_raidbypass)
|
|
The 1166:0241 PCI device has 8 ports instead of 4.
Signed-off-by: Dmitry Borisov <di.sean@protonmail.com>
Reviewed by: imp, jlduran
Pull Request: https://github.com/freebsd/freebsd-src/pull/1883
|
|
All sysctls except thermal sensor readings and the handlerevents
whitelist-separated string are bitmasks or small integers. This avoids some
bitmasks being erroneously displayed as negative values.
Signed-off-by: Quentin Thébault <quentin.thebault@defenso.fr>
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1824
|
|
The apst_data tunable allows APST configuration to be adjusted
during controller initialization. It accepts an array of encoded
integers, each defining specific transition parameters.
Relnotes: YES
Signed-off-by: Alexey Sukhoguzov <sap@eseipi.net>
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1444
|
|
APST is an optional NVMe power-saving feature that allows devices
to autonomously enter higher non-operational power states after a
certain amount of idle time, reducing the controller's overall power
consumption.
Signed-off-by: Alexey Sukhoguzov <sap@eseipi.net>
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1444
|
|
Signed-off-by: Alexey Sukhoguzov <sap@eseipi.net>
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1444
|
|
This is equivalent of sdio_readw and sdio_writew in linuxkpi
Signed-off-by: Minsoo Choo <minsoochoo0122@proton.me>
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/1951
|
|
This updates the smartpqi driver to Microchip's
latest available public release.
Reviewed by: imp
Approved by: imp
Sponsored by: Microchip Technology Inc.
Differential Revision: https://reviews.freebsd.org/D52507
|
|
The driver and applications currently use hard-coded numeric ioctl command
opcodes. These opcodes are interpreted as having the IOC_IN direction (data
copied from the user application to the driver), regardless of the actual packet
size. Consequently, when the packet size is zero and the direction is set to
IOC_IN, the kernel fails these ioctls if COMPAT is disabled.
While the driver and applications should ideally set the direction correctly—
for example, using IOC_VOID when the packet size is zero—the driver will now
be updated to define ioctl opcodes using the _IOC macro to support both
IOC_VOID and IOC_IN. This change ensures backward compatibility with older
applications that exclusively use IOC_IN.
Reviewed by: gallatin
Differential Revision: https://reviews.freebsd.org/D54601
MFC after: 3 days
|
|
cam::mpr:complete(union ccb *, struct mpr_command *, u_int, u32);
Where u_int is scsas->flags u32 is the device_info.
This can't be done as an fbt because the data needed for it isn't
present out a function boundary.
Sponsored by: Netflix
|
|
cam::mps:complete(union ccb *, struct mps_command *, u_int, u32);
Where u_int is scsas->flags u32 is the device_info.
This can't be done as an fbt because the data needed for it isn't
present out a function boundary.
Sponsored by: Netflix
|
|
These are left over from the $FreeBSD$ stuff.
Sponsored by: Netflix
|
|
Replace bzero with the equivalent memset(x,0,x) since the latter is
available in the kshim environment and we have a soft goal of migrating
to standard interfaces when there's a reason...
Sponsored by: Netflix
|
|
Required when KTR is configured.
Remove the pcpu.h include while here, as it seems to be unneeded.
Reported by: Jenkins
Fixes: 5f13d6b60740 ("vmm: Move common accessors and vm_eventinfo into sys/dev/vmm")
|
|
Now that struct vm and struct vcpu are defined in headers, provide
inline accessors. We could just remove the accessors outright, but they
don't hurt and it would result in unneeded churn.
As a part of this, consolidate definitions related to struct
vm_eventinfo as well. I'm not sure if struct vm_eventinfo is really
needed anymore, now that vmmops_run implementations can directly access
vm and vcpu fields, but this can be resolved later.
No functional change intended.
MFC after: 2 months
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D53586
|
|
Now that the machine-independent fields of struct vm and struct vcpu are
available in a header, we can move lots of duplicated code into
sys/dev/vmm/vmm_vm.c. This change does exactly that.
No functional change intended.
MFC after: 2 months
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D53585
|
|
There is quite a lot of duplication of code between amd64, arm64 and
riscv with respect to VM and vCPU state management. This is a bit
tricky to resolve since struct vm and struct vcpu are private to vmm.c
and both structures contain a mix of machine-dependent and
machine-independent fields.
To allow deduplication without also introducing a lot of churn, follow
the approach of struct pcpu and 1) lift the definitions of those
structures into a new header, sys/dev/vmm/vmm_vm.h, and 2) define
machine-dependent macros, VMM_VM_MD_FIELDS and VMM_VCPU_MD_FIELDS which
lay out the machine-dependent fields.
One disadvantage of this approach is that the two structures are no
longer private to vmm.c, but I think this is acceptable.
No functional change intended. A follow-up change will move a good deal
of machine/vmm/vmm.c into sys/dev/vmm/vmm_vm.c.
MFC after: 2 months
Sponsored by: The FreeBSD Foundation
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D53584
|
|
Allow reading of i2c addresses other than A0. A0 does provide most
information, but doesn't provide things like module temps, and optical
signal levels.
Sponsored by: Netflix
Reviewed by: sumit.saxena_broadcom.com
Differential Revision: https://reviews.freebsd.org/D54590
MFC after: 3 days
|
|
This is needed because when we add the suspend-to-idle loop, we only
want to break the BSP out of idle to process the taskqueue while keeping
all the others idled.
Currently assuming BSP to be CPU0, which may not always be the case on
non-x86.
Reviewed by: kib
Approved by: kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D54406
|
|
The slot parameter actually is used as index into the memory resource.
Fixes: c11bbc7d
|
|
a0b4d9e5bd63984bc3ad25b7f2881741e0d73980
Add support for RTL8126 chip revision 0x64a00000 from Carl Henriksson.
While here, update and add microcode for the RTL8126 chipset variant.
46810ff569a9416642fc5f8e86501028008a2e95
For RTL8125, there's no need to check whether RGE_CMD_STOPREQ has been
completed. As with other chip variants, RGE_CMD_STOPREQ never acknowledges
completion.
5729fb17b5eeba56781da0b5c0525c198c5c2cd8
remove duplicate RGE_PHYSTAT_5000MBPS test
Reviewed by: adrian
Differential Revision: https://reviews.freebsd.org/D54517
Obtained from: OpenBSD
|
|
1GB (D3116)
This is an OEM card from Fujitsu using an LSI SAS2208 ROC controller shipped
with many Fujitsu PRIMERGY servers like RX300 S7. This chip is also recognized
by mrsas(4) under the generic name for the controller chip.
Controller description: https://www.fujitsu.com/global/products/computing/servers/primergy/blades/connection/cb-pmod-110426.html
Reviewed by: ziaee
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D54520
|
|
This allows folks to enable debug statements in asmc(4) using kernel
configs via the `options ASMC_DEBUG` directive.
While here, remove a duplicate `device vt_efifb` directive in `NOTES`
as it's already handled in the `GENERIC` config
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D54511
|
|
According to section 8.3.2.2.1 of the XL710 datasheet,
FLTSTAT indicates if RSS hashing was used (or flow director
or nothing) to arrive at the flowid it is reporting.
We need to check this before reporting the flowid and setting the
rss type. Not checking this allows invalid flowids (0s) to be
reported for some TCP traffic (it seems to mostly impact incoming
connections). This leads to the inp flowid being 0, and egress
traffic being unbalanced (going to tx queue 0, and NIC 0 of an
lacp bundle).
Differential Revision: https://reviews.freebsd.org/D54442
Reviewed by: adrian, kbowling
Sponsored by: Netflix
MFC after: 2 weeks
|
|
Add per-fan manual mode control via dev.asmc.0.fan.N.manual sysctl.
Apple SMCs support manual fan control via the FS! SMC key,
a 16-bit bitmask where each bit controls one fan (0=auto, 1=manual).
This change adds a new sysctl per fan:
dev.asmc.0.fan.N.manual (0=auto, 1=manual)
When set to manual mode (1), the fan runs at the speed set via
dev.asmc.0.fan.N.targetspeed instead of automatic thermal control. When
set to auto mode (0), the SMC controls fan speed automatically.
The FS! key was already defined in asmcvar.h but not accessible.
This exposes it for debugging, testing, and advanced fan control.
Implementation uses read-modify-write to allow independent control of
each fan without affecting others.
Reviewed by: adrian, markj
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D54437
|
|
The compiled objects do not change other than a few diagnostic messages
that include __LINE__.
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D54304
|
