Kernel test robot reported incorrect type in argument 1 of readl(), but
more importantly it brought attention that MMIO accessor shouldn't be
used in this case, since req->hdr.status is part of a command-response
buffer in system memory.
Since its value may be altered by PSP outside of the scope of current
thread (somehow similar to IRQ handler case), we need to use
READ_ONCE() to ensure compiler won't optimize this call.
Fix also 'status' variable type to reflect that corresponding field in
command-response buffer is platform-independent u32.
Fixes: 78d5e9e299 ("i2c: designware: Add AMD PSP I2C bus support")
Signed-off-by: Jan Dabros <jsd@semihalf.com>
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Do not call get_trip_hyst() from thermal_genl_cmd_tz_get_trip() if
the thermal zone does not define one.
Fixes: 1ce50e7d40 ("thermal: core: genetlink support for events/cmd/sampling")
Signed-off-by: Nicolas Cavallari <nicolas.cavallari@green-communications.fr>
Cc: 5.10+ <stable@vger.kernel.org> # 5.10+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
QUP Serial engines supports data transfers thru FIFO mode, SE DMA mode
and lastly GPI DMA mode. Former two are already supported and this adds
supports for the last mode.
In GPI DMA mode, the firmware is issued commands by driver to perform
DMA and setup the serial port.
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Read reset phandle as optional instead of exclusive so that all the DT's
passing the reset phandle can be used to assert/deassert the reset line.
With this change we don't have to differentiate the RZ/G2L SoC.
With the above changes we no longer need the "renesas,riic-r9a07g044"
compatible string, so drop it from riic_i2c_dt_ids[]. No changes are
required to the r9a07g044.dtsi as we already have "renesas,riic-rz" as a
fallback compatible string.
While at it, check the return code of reset_control_deassert() as it might
fail and also add a devres action to assert the reset line.
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
Exit to userspace if setup_vmgexit_scratch() fails due to OOM or because
copying data from guest (userspace) memory failed/faulted. The OOM
scenario is clearcut, it's userspace's decision as to whether it should
terminate the guest, free memory, etc...
As for -EFAULT, arguably, any guest issue is a violation of the guest's
contract with userspace, and thus userspace needs to decide how to
proceed. E.g. userspace defines what is RAM vs. MMIO and communicates
that directly to the guest, KVM is not involved in deciding what is/isn't
RAM nor in communicating that information to the guest. If the scratch
GPA doesn't resolve to a memslot, then the guest is not honoring the
memory configuration as defined by userspace.
And if userspace unmaps an hva for whatever reason, then exiting to
userspace with -EFAULT is absolutely the right thing to do. KVM's ABI
currently sucks and doesn't provide enough information to act on the
-EFAULT, but that will hopefully be remedied in the future as there are
multiple use cases, e.g. uffd and virtiofs truncation, that shouldn't
require any work in KVM beyond returning -EFAULT with a small amount of
metadata.
KVM could define its ABI such that failure to access the scratch area is
reflected into the guest, i.e. establish a contract with userspace, but
that's undesirable as it limits KVM's options in the future, e.g. in the
potential uffd case any failure on a uaccess needs to kick out to
userspace. KVM does have several cases where it reflects these errors
into the guest, e.g. kvm_pv_clock_pairing() and Hyper-V emulation, but
KVM would preferably "fix" those instead of propagating the falsehood
that any memory failure is the guest's fault.
Lastly, returning a boolean as an "error" for that a helper that isn't
named accordingly never works out well.
Fixes: ad5b353240 ("KVM: SVM: Do not terminate SEV-ES guests on GHCB validation failure")
Cc: Alper Gun <alpergun@google.com>
Cc: Peter Gonda <pgonda@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220225205209.3881130-1-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Instead of manually disabling interrupts before invoking use
generic_handle_irq_safe() which can be invoked with enabled and disabled
interrupts.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Wolfram Sang <wsa@kernel.org>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
The i2c-i801 driver invokes i2c_handle_smbus_host_notify() from his
interrupt service routine. On PREEMPT_RT i2c-i801's handler is forced
threaded with enabled interrupts which leads to a warning by
handle_irq_event_percpu() assuming that irq_default_primary_handler()
enabled interrupts.
i2c-i801's interrupt handler can't be made non-threaded because the
interrupt line is shared with other devices.
Use generic_handle_irq_safe() which can invoked with disabled and enabled
interrupts.
Reported-by: Michael Below <below@judiz.de>
Link: https://bugs.debian.org/1002537
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
dev_err_probe() can reduce code size, makes the code easier to read
and has the added benefit of recording the defer reason for later
read out. Use it where appropriate.
This also fixes an issue, where an error message in __genpd_dev_pm_attach
was not terminated by a line break.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The function alloc_workqueue() in nintendo_hid_probe() can fail, but
there is no check of its return value. To fix this bug, its return value
should be checked with new error handling code.
Fixes: c4eae84fef ("HID: nintendo: add rumble support")
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Reviewed-by: Silvan Jegen <s.jegen@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
A PM domain managed by genpd may support multiple idlestates (power-off
states). During genpd_power_off() a genpd governor may be asked to select
one of the idlestates based upon the dev PM QoS constraints, for example.
However, there is a problem with the behaviour around this in genpd. More
precisely, a parent-domain is allowed to be powered off, no matter of what
idlestate that has been selected for the child-domain.
For the stm32mp1 platform from STMicro, this behaviour doesn't play well.
Instead, the parent-domain must not be powered off, unless the deepest
idlestate has been selected for the child-domain. As the current behaviour
in genpd is quite questionable anyway, let's simply change it into what is
needed by the stm32mp1 platform.
If it surprisingly turns out that other platforms may need a different
behaviour from genpd, then we will have to revisit this to find a way to
make it configurable.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
`CONFIG_SATA_LPM_MOBILE_POLICY` reflects a configuration to apply only to
mobile chipsets. As some desktop boards may want to use this policy by
default as well, rename the configuration item to `SATA_LPM_POLICY`.
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
`AHCI_HFLAG_IS_MOBILE` designates that a chipset should be using the
default link power management policy from a kernel configuration item.
As desktop chipsets may also be interested in this default policy
configuration, rename the flag to `AHCI_HFLAG_USE_LPM_POLICY` to more
accurately reflect that a chipset doesn't have to be mobile to adopt it.
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
This board definition was originally created for mobile devices to
designate default link power managmeent policy to influence runtime
power consumption.
As this is interesting for more than just mobile designs, rename the
board to `board_ahci_low_power` to make it clear it is about default
policy.
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
The HID core stack used to be very relaxed considering the BTN_TOOL_*
usage. With the recent commits, we should now enforce to have only one
tool at a time, meaning that we can now express that requirement in the
docs.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Multitouch devices in hybrid mode are reporting multiple times the
same collection. We should accommodate for this in our handling
of priorities by defining the slots they belong to.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
HID_QUIRK_INVERT is kind of complex to deal with and was bogus.
Furthermore, it didn't make sense to use a global per struct hid_device
quirk for something dynamic as the current state.
Store the current tool information in the report itself, and re-order
the processing of the fields to enforce having all the tablet "state"
fields before getting to In Range and other input fields.
This way, we now have all the information whether a tool is present
or not while processing In Range.
This new behavior enforces that only one tool gets forwarded to userspace
at the same time, and that if either eraser or invert is set, we enforce
BTN_TOOL_RUBBER.
Note that the release of the previous tool now happens in its own EV_SYN
report so userspace doesn't get confused by having 2 tools.
These changes are tested in the following hid-tools regression tests:
https://gitlab.freedesktop.org/libevdev/hid-tools/-/merge_requests/127
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
When a device exposes both Invert and InRange, Invert must be processed
before InRange. If we keep the order of the device and we process them
out of order, InRange will first set BTN_TOOL_PEN, and then Invert will
set BTN_TOOL_RUBBER. Userspace knows how to deal with that situation,
but fixing it in the kernel is now easier.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Now that we have a list of fields/usages by priority order,
walk through that list to process the inputs instead of using the
order provided by the manufacturer.
Note that this changes the way we update the values in the struct
hid_field:
Previously, once a field was processed, we updated the new values.
Now we need to wait for the entire report to be processed to update
the values.
I don't think it will be an issue: because we were relying on the device
ordering, there were no guarantees to have a field stored before an other.
Which is why we introduced .report() in drivers to have those values
updated.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This will be used in a later commit:
we build a list of input fields (and usage_index) that is ordered based
on a usage priority.
Changing the usage priority allows to re-order the processed list, meaning
that we can enforce some usages to be process before others.
For instance, before processing InRange in the HID tablets, we need to
know if we are using the eraser (side or button). Enforcing a higher
(lower number) priority for Invert allows to force the input stack to
process that field before.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
It actually makes sense to clamp the value to its boundaries before
doing further processing.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Instead of using multiple `if (a == b)`, use the switch statement
which has been done exactly for that.
There should be no functional change (I don't expect moving down
HID_QUIRK_{X|Y}_INVERT having any impact.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Some devices (Elan, Synaptics...) are sometimes not setting a physical
in their finger collections. hid-input will consider them to be pen
devices, leading to some wrong behavior in user space.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This is a preparatory patch for being able to process the usages
out of order. We split the retrieval of the data in a separate function
and also split out the processing of the usages depending if the field
is an array or a variable.
No functional changes from this patch.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
I had to go twice through the history to get a grasp at this code.
De-duplicate the various tests in one common helper to make it
more explicit.
Note that the `HID_UP_KEYBOARD + 1` condition is tested through
https://gitlab.freedesktop.org/libevdev/hid-tools/-/merge_requests/121
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This is a preparation patch for rethinking the generic processing
of HID reports.
We can actually pre-allocate all of our memory instead of dynamically
allocating/freeing it whenever we parse a report.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: Ping Cheng <ping.cheng@wacom.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
* rfkill
- add missing rfill_soft_blocked() when disabled
* cfg80211
- handle a nla_memdup() failure correctly
- fix CONFIG_CFG80211_EXTRA_REGDB_KEYDIR typo in
Makefile
* mac80211
- fix EAPOL handling in 802.3 RX path
- reject setting up aggregation sessions before
connection is authorized to avoid timeouts or
similar
- handle some SAE authentication steps correctly
- fix AC selection in mesh forwarding
* iwlwifi
- remove TWT support as it causes firmware crashes
when the AP isn't behaving correctly
- check debugfs pointer before dereferncing it
-----BEGIN PGP SIGNATURE-----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=Q7qF
-----END PGP SIGNATURE-----
Merge tag 'wireless-for-net-2022-03-01' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless
johannes Berg says:
====================
Some last-minute fixes:
* rfkill
- add missing rfill_soft_blocked() when disabled
* cfg80211
- handle a nla_memdup() failure correctly
- fix CONFIG_CFG80211_EXTRA_REGDB_KEYDIR typo in
Makefile
* mac80211
- fix EAPOL handling in 802.3 RX path
- reject setting up aggregation sessions before
connection is authorized to avoid timeouts or
similar
- handle some SAE authentication steps correctly
- fix AC selection in mesh forwarding
* iwlwifi
- remove TWT support as it causes firmware crashes
when the AP isn't behaving correctly
- check debugfs pointer before dereferncing it
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Booting msm8939 on tip-of-tree I encountered the following error.
[ 1.212340] qnoc-msm8939 580000.interconnect_mm: can't request region for resource [mem 0x00580000-0x0059407f]
[ 1.212391] qnoc-msm8939 580000.interconnect_mm: Cannot ioremap interconnect bus resource
[ 1.221524] qnoc-msm8939: probe of 580000.interconnect_mm failed with error -16
Initially I thought this was a bug with the interconnect driver but,
examining it a bit more I realized the DTS I was working with based on
downstream, declares snoc and snoc_mm as existing at the same address
range.
When we were developing the DTS for 8939 we weren't using the common rpm
interconnect driver so we never saw the ioremap collision.
Taking a hard look at the qcom documentation as well as the downstream code
we see that yes downstream declares snoc and snoc_mm separately but, also
at the same overlapping address.
The qcom documentation for performance points for msm8936/msm8939 snoc,
deliniates snoc_mm as simply two new performance-points i.e. a faster GPLL0
vote associated with new multi-media devices attached to the snoc.
In other words the snoc had two new RPM vote indices added to it, to
represent the higher performance clocks, should one of the multi-media IP
blocks call for it.
We can fix the ioremap collision and still represent the two higher
performance point clock votes by coalsecing snoc and snoc_mm into snoc. The
DTS clock references will take care of the appropriate votes.
V2:
Made snoc-mm into a child of snoc - Dmitry
Removes redundant msm8939_snoc_mm_regmap_config - bod
Link: https://lore.kernel.org/r/20220129032735.2410936-1-bryan.odonoghue@linaro.org
Signed-off-by: Georgi Djakov <djakov@kernel.org>
The driver creates the top row map sysfs attribute in input_configured()
method; unfortunately we do not have a callback that is executed when HID
interface is unbound, thus we are leaking these sysfs attributes, for
example when device is disconnected.
To fix it let's switch to managed version of adding sysfs attributes which
will ensure that they are destroyed when the driver is unbound.
Fixes: 14c9c014ba ("HID: add vivaldi HID driver")
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Tested-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Add support for multiple frame input devices and their parameters to
the UC-Logic driver. This prepares for creating a separate input device
for Huion HS610 virtual touch ring reports.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Put general hardware version before everything else in uclogic_rdesc_
namespace.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Use the term "frame" instead of "buttonpad" for consistency, in UC-Logic
driver.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Allow to set the report ID in UCLOGIC_RDESC_FRAME_BYTES instead of
using a hardcoded value.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Simplify the UCLOGIC_RDESC_BUTTONPAD_BYTES macro by passing as param
the size of the report to pad to in bytes.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Switch from matching UC-Logic subreport with a mask to a full value to
support detecting Huion HS610 touch dial reports. Those would match the
button mask otherwise.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Replace a single pen_frame_flag in struct uclogic_params with
subreport_list in struct uclogic_params_pen to prepare for handling more
subreports in Huion HS610.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Remove support for pen usage masking from hid-uclogic. Disable whole
interfaces instead. Most of those interfaces are useless, and if there
is one which has an unused pen usage, but also has useful reports, its
report descriptor should be rewritten instead.
This simplifies the code and the data structures.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: José Expósito <jose.exposito89@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Dust Li says:
====================
net/smc: some datapath performance optimizations
This series tries to improve the performance of SMC in datapath.
- patch #1, add sysctl interface to support tuning the behaviour of
SMC in container environment.
- patch #2/#3, add autocorking support which is very efficient for small
messages without trade-off for latency.
- patch #4, send directly on setting TCP_NODELAY, without wake up the
TX worker, this make it consistent with clearing TCP_CORK.
- patch #5, this correct the setting of RMB window update limit, so
we don't send CDC messages to update peer's RMB window too frequently
in some cases.
- patch #6, implemented something like NAPI in SMC, decrease the number
of hardirq when busy.
- patch #7, this moves TX work doing in the BH to the user context when
sock_lock is hold by user.
With this patchset applied, we can get a good performance gain:
- qperf tcp_bw test has shown a great improvement. Other benchmarks like
'netperf TCP_STREAM' or 'sockperf throughput' has similar result.
- In my testing environment, running qperf tcp_bw and tcp_lat, SMC behaves
better then TCP in most all message size.
Here are some test results with the following testing command:
client: smc_run taskset -c 1 qperf smc-server -oo msg_size:1:64K:*2 \
-t 30 -vu tcp_{bw|lat}
server: smc_run taskset -c 1 qperf
==== Bandwidth ====
MsgSize Origin SMC TCP SMC with patches
1 0.578 MB/s 2.392 MB/s(313.57%) 2.561 MB/s(342.83%)
2 1.159 MB/s 4.780 MB/s(312.53%) 5.162 MB/s(345.46%)
4 2.283 MB/s 10.266 MB/s(349.77%) 10.122 MB/s(343.46%)
8 4.668 MB/s 19.040 MB/s(307.86%) 20.521 MB/s(339.59%)
16 9.147 MB/s 38.904 MB/s(325.31%) 40.823 MB/s(346.29%)
32 18.369 MB/s 79.587 MB/s(333.25%) 80.535 MB/s(338.42%)
64 36.562 MB/s 148.668 MB/s(306.61%) 158.170 MB/s(332.60%)
128 72.961 MB/s 274.913 MB/s(276.80%) 316.217 MB/s(333.41%)
256 144.705 MB/s 512.059 MB/s(253.86%) 626.019 MB/s(332.62%)
512 288.873 MB/s 884.977 MB/s(206.35%) 1221.596 MB/s(322.88%)
1024 574.180 MB/s 1337.736 MB/s(132.98%) 2203.156 MB/s(283.70%)
2048 1095.192 MB/s 1865.952 MB/s( 70.38%) 3036.448 MB/s(177.25%)
4096 2066.157 MB/s 2380.337 MB/s( 15.21%) 3834.271 MB/s( 85.58%)
8192 3717.198 MB/s 2733.073 MB/s(-26.47%) 4904.910 MB/s( 31.95%)
16384 4742.221 MB/s 2958.693 MB/s(-37.61%) 5220.272 MB/s( 10.08%)
32768 5349.550 MB/s 3061.285 MB/s(-42.77%) 5321.865 MB/s( -0.52%)
65536 5162.919 MB/s 3731.408 MB/s(-27.73%) 5245.021 MB/s( 1.59%)
==== Latency ====
MsgSize Origin SMC TCP SMC with patches
1 10.540 us 11.938 us( 13.26%) 10.356 us( -1.75%)
2 10.996 us 11.992 us( 9.06%) 10.073 us( -8.39%)
4 10.229 us 11.687 us( 14.25%) 9.996 us( -2.28%)
8 10.203 us 11.653 us( 14.21%) 10.063 us( -1.37%)
16 10.530 us 11.313 us( 7.44%) 10.013 us( -4.91%)
32 10.241 us 11.586 us( 13.13%) 10.081 us( -1.56%)
64 10.693 us 11.652 us( 8.97%) 9.986 us( -6.61%)
128 10.597 us 11.579 us( 9.27%) 10.262 us( -3.16%)
256 10.409 us 11.957 us( 14.87%) 10.148 us( -2.51%)
512 11.088 us 12.505 us( 12.78%) 10.206 us( -7.95%)
1024 11.240 us 12.255 us( 9.03%) 10.631 us( -5.42%)
2048 11.485 us 16.970 us( 47.76%) 10.981 us( -4.39%)
4096 12.077 us 13.948 us( 15.49%) 11.847 us( -1.90%)
8192 13.683 us 16.693 us( 22.00%) 13.336 us( -2.54%)
16384 16.470 us 23.615 us( 43.38%) 16.519 us( 0.30%)
32768 22.540 us 40.966 us( 81.75%) 22.452 us( -0.39%)
65536 34.192 us 73.003 us(113.51%) 33.916 us( -0.81%)
------------
Test environment notes:
1. Testing is run on 2 VMs within the same physical host
2. The NIC is ConnectX-4Lx, using SRIOV, and passing through 2 VFs to the
2 VMs respectively.
3. To decrease jitter, VM's vCPU are binded to each physical CPU, and those
physical CPUs are all isolated using boot parameter `isolcpus=xxx`
4. The queue number are set to 1, and interrupt from the queue is binded to
CPU0 in the guest
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Send data all the way down to the RDMA device is a time
consuming operation(get a new slot, maybe do RDMA Write
and send a CDC, etc). Moving those operations from BH
to user context is good for performance.
If the sock_lock is hold by user, we don't try to send
data out in the BH context, but just mark we should
send. Since the user will release the sock_lock soon, we
can do the sending there.
Add smc_release_cb() which will be called in release_sock()
and try send in the callback if needed.
This patch moves the sending part out from BH if sock lock
is hold by user. In my testing environment, this saves about
20% softirq in the qperf 4K tcp_bw test in the sender side
with no noticeable throughput drop.
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we are handling softirq workload, enable hardirq may
again interrupt the current routine of softirq, and then
try to raise softirq again. This only wastes CPU cycles
and won't have any real gain.
Since IB_CQ_REPORT_MISSED_EVENTS already make sure if
ib_req_notify_cq() returns 0, it is safe to wait for the
next event, with no need to poll the CQ again in this case.
This patch disables hardirq during the processing of softirq,
and re-arm the CQ after softirq is done. Somehow like NAPI.
Co-developed-by: Guangguan Wang <guangguan.wang@linux.alibaba.com>
Signed-off-by: Guangguan Wang <guangguan.wang@linux.alibaba.com>
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rmbe_update_limit is used to limit announcing receive
window updating too frequently. RFC7609 request a minimal
increase in the window size of 10% of the receive buffer
space. But current implementation used:
min_t(int, rmbe_size / 10, SOCK_MIN_SNDBUF / 2)
and SOCK_MIN_SNDBUF / 2 == 2304 Bytes, which is almost
always less then 10% of the receive buffer space.
This causes the receiver always sending CDC message to
update its consumer cursor when it consumes more then 2K
of data. And as a result, we may encounter something like
"TCP silly window syndrome" when sending 2.5~8K message.
This patch fixes this using max(rmbe_size / 10, SOCK_MIN_SNDBUF / 2).
With this patch and SMC autocorking enabled, qperf 2K/4K/8K
tcp_bw test shows 45%/75%/40% increase in throughput respectively.
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit ea785a1a573b("net/smc: Send directly when
TCP_CORK is cleared"), we don't use delayed work
to implement cork.
This patch use the same algorithm, removes the
delayed work when setting TCP_NODELAY and send
directly in setsockopt(). This also makes the
TCP_NODELAY the same as TCP.
Cc: Tony Lu <tonylu@linux.alibaba.com>
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This add a new sysctl: net.smc.autocorking_size
We can dynamically change the behaviour of autocorking
by change the value of autocorking_size.
Setting to 0 disables autocorking in SMC
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds autocorking support for SMC which could improve
throughput for small message by x3+.
The main idea is borrowed from TCP autocorking with some RDMA
specific modification:
1. The first message should never cork to make sure we won't
bring extra latency
2. If we have posted any Tx WRs to the NIC that have not
completed, cork the new messages until:
a) Receive CQE for the last Tx WR
b) We have corked enough message on the connection
3. Try to push the corked data out when we receive CQE of
the last Tx WR to prevent the corked messages hang in
the send queue.
Both SMC autocorking and TCP autocorking check the TX completion
to decide whether we should cork or not. The difference is
when we got a SMC Tx WR completion, the data have been confirmed
by the RNIC while TCP TX completion just tells us the data
have been sent out by the local NIC.
Add an atomic variable tx_pushing in smc_connection to make
sure only one can send to let it cork more and save CDC slot.
SMC autocorking should not bring extra latency since the first
message will always been sent out immediately.
The qperf tcp_bw test shows more than x4 increase under small
message size with Mellanox connectX4-Lx, same result with other
throughput benchmarks like sockperf/netperf.
The qperf tcp_lat test shows SMC autocorking has not increase any
ping-pong latency.
Test command:
client: smc_run taskset -c 1 qperf smc-server -oo msg_size:1:64K:*2 \
-t 30 -vu tcp_{bw|lat}
server: smc_run taskset -c 1 qperf
=== Bandwidth ====
MsgSize(Bytes) SMC-NoCork TCP SMC-AutoCorking
1 0.578 MB/s 2.392 MB/s(313.57%) 2.647 MB/s(357.72%)
2 1.159 MB/s 4.780 MB/s(312.53%) 5.153 MB/s(344.71%)
4 2.283 MB/s 10.266 MB/s(349.77%) 10.363 MB/s(354.02%)
8 4.668 MB/s 19.040 MB/s(307.86%) 21.215 MB/s(354.45%)
16 9.147 MB/s 38.904 MB/s(325.31%) 41.740 MB/s(356.32%)
32 18.369 MB/s 79.587 MB/s(333.25%) 82.392 MB/s(348.52%)
64 36.562 MB/s 148.668 MB/s(306.61%) 161.564 MB/s(341.89%)
128 72.961 MB/s 274.913 MB/s(276.80%) 325.363 MB/s(345.94%)
256 144.705 MB/s 512.059 MB/s(253.86%) 633.743 MB/s(337.96%)
512 288.873 MB/s 884.977 MB/s(206.35%) 1250.681 MB/s(332.95%)
1024 574.180 MB/s 1337.736 MB/s(132.98%) 2246.121 MB/s(291.19%)
2048 1095.192 MB/s 1865.952 MB/s( 70.38%) 2057.767 MB/s( 87.89%)
4096 2066.157 MB/s 2380.337 MB/s( 15.21%) 2173.983 MB/s( 5.22%)
8192 3717.198 MB/s 2733.073 MB/s(-26.47%) 3491.223 MB/s( -6.08%)
16384 4742.221 MB/s 2958.693 MB/s(-37.61%) 4637.692 MB/s( -2.20%)
32768 5349.550 MB/s 3061.285 MB/s(-42.77%) 5385.796 MB/s( 0.68%)
65536 5162.919 MB/s 3731.408 MB/s(-27.73%) 5223.890 MB/s( 1.18%)
==== Latency ====
MsgSize(Bytes) SMC-NoCork TCP SMC-AutoCorking
1 10.540 us 11.938 us( 13.26%) 10.573 us( 0.31%)
2 10.996 us 11.992 us( 9.06%) 10.269 us( -6.61%)
4 10.229 us 11.687 us( 14.25%) 10.240 us( 0.11%)
8 10.203 us 11.653 us( 14.21%) 10.402 us( 1.95%)
16 10.530 us 11.313 us( 7.44%) 10.599 us( 0.66%)
32 10.241 us 11.586 us( 13.13%) 10.223 us( -0.18%)
64 10.693 us 11.652 us( 8.97%) 10.251 us( -4.13%)
128 10.597 us 11.579 us( 9.27%) 10.494 us( -0.97%)
256 10.409 us 11.957 us( 14.87%) 10.710 us( 2.89%)
512 11.088 us 12.505 us( 12.78%) 10.547 us( -4.88%)
1024 11.240 us 12.255 us( 9.03%) 10.787 us( -4.03%)
2048 11.485 us 16.970 us( 47.76%) 11.256 us( -1.99%)
4096 12.077 us 13.948 us( 15.49%) 12.230 us( 1.27%)
8192 13.683 us 16.693 us( 22.00%) 13.786 us( 0.75%)
16384 16.470 us 23.615 us( 43.38%) 16.459 us( -0.07%)
32768 22.540 us 40.966 us( 81.75%) 23.284 us( 3.30%)
65536 34.192 us 73.003 us(113.51%) 34.233 us( 0.12%)
With SMC autocorking support, we can archive better throughput
than TCP in most message sizes without any latency trade-off.
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
