| Age | Commit message (Collapse) | Author |
|---|
|
Included change:
- properly compute the batman-adv header overhead. Such
result is later used to initialize the hard_header_len
member of the soft-interface netdev object
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Now it catches the NETDEV_CHANGEMTU notification, which is signaled after
the actual change happened on the device, and returns NOTIFY_BAD, so that
the change on the device is reverted.
This might be quite costly and messy, so use the new NETDEV_PRECHANGEMTU to
catch the MTU change before the actual change happens and signal that it's
forbidden to do it.
CC: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Currently, if a device changes its mtu, first the change happens (invloving
all the side effects), and after that the NETDEV_CHANGEMTU is sent so that
other devices can catch up with the new mtu. However, if they return
NOTIFY_BAD, then the change is reverted and error returned.
This is a really long and costy operation (sometimes). To fix this, add
NETDEV_PRECHANGEMTU notification which is called prior to any change
actually happening, and if any callee returns NOTIFY_BAD - the change is
aborted. This way we're skipping all the playing with apply/revert the mtu.
CC: "David S. Miller" <davem@davemloft.net>
CC: Jiri Pirko <jiri@resnulli.us>
CC: Eric Dumazet <edumazet@google.com>
CC: Nicolas Dichtel <nicolas.dichtel@6wind.com>
CC: Cong Wang <amwang@redhat.com>
Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
* pm-cpufreq: (40 commits)
thermal: exynos: boost: Automatic enable/disable of BOOST feature (at Exynos4412)
cpufreq: exynos4x12: Change L0 driver data to CPUFREQ_BOOST_FREQ
Documentation: cpufreq / boost: Update BOOST documentation
cpufreq: exynos: Extend Exynos cpufreq driver to support boost
cpufreq / boost: Kconfig: Support for software-managed BOOST
acpi-cpufreq: Adjust the code to use the common boost attribute
cpufreq: Add boost frequency support in core
intel_pstate: Add trace point to report internal state.
cpufreq: introduce cpufreq_generic_get() routine
ARM: SA1100: Create dummy clk_get_rate() to avoid build failures
cpufreq: stats: create sysfs entries when cpufreq_stats is a module
cpufreq: stats: free table and remove sysfs entry in a single routine
cpufreq: stats: remove hotplug notifiers
cpufreq: stats: handle cpufreq_unregister_driver() and suspend/resume properly
cpufreq: speedstep: remove unused speedstep_get_state
powernow-k6: reorder frequencies
powernow-k6: correctly initialize default parameters
powernow-k6: disable cache when changing frequency
Documentation: add ABI entry for intel_pstate
cpufreq: exynos: Convert exynos-cpufreq to platform driver
...
|
|
Exynos4412)
This patch provides auto disable/enable operation for boost. It uses already
present thermal infrastructure to provide BOOST hysteresis.
The TMU data is modified to work properly with or without BOOST.
Hence, the two first trip points with corresponding clip frequencies are
adjusted.
The first one is reduced from 85 to 70 degrees and the clip frequency is
increased to 1.4 GHz from 800 MHz. This trip point is in fact responsible
for providing BOOST hysteresis. When temperature exceeds 70 deg, the maximal
non BOOST frequency for Exynos4412 is imposed.
Since the first trigger level has been "stolen" for BOOST, the second one
needs to be a compromise for the previously used two for non BOOST
configuration. The 95 deg with modified clip freq (to 400 MHz) should provide
a good balance between cooling down the overheated device and throughput on
an acceptable level.
Two last trigger levels are not modified since, they cause platform shutdown
on emergency overheat to happen.
The third trip point passage results in SW managed shut down of the system.
If the last trip point is crossed, the PMU HW generates the power off
signal.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Eduardo Valentin <eduardo.valentin@ti.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
[rjw: Changelog]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Add a special driver data flag (CPUFREQ_BOOST_FREQ) to indicate
a frequency that can be only enabled for BOOST mode.
This frequency will be used only for limited time, since running with
it for too long may cause the target device to overheat.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
[rjw: Changelog]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Since the support for software and hardware controlled boosting has
been added, update the corresponding documentation.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The cpufreq_driver's boost_supported flag is true only when boost
support is explicitly enabled. Boost related attributes are exported
only under the same condition.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Add CONFIG_CPU_FREQ_BOOST_SW Kconfig option such that software-managed
boost is enabled only after selecting "EXYNOS Frequency Overclocking -
Software". It also depends on the thermal subsystem to be compiled in,
which is necessary for disabling boost and cooling down the device when
overheating is detected.
Software-managed boost _MUST_ _NOT_ be enabled without thermal subsystem
with properly defined overheating temperature thresholds.
This option doesn't affect the x86's hardware-driven boost support
in the acpi-cpufreq driver.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
[rjw: Subject and changelog]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Modify acpi-cpufreq's hardware-based boost solution to work with the
common cpufreq boost framework.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
[rjw: Subject and changelog]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
This commit adds boost frequency support in cpufreq core (Hardware &
Software). Some SoCs (like Exynos4 - e.g. 4x12) allow setting frequency
above its normal operation limits. Such mode shall be only used for a
short time.
Overclocking (boost) support is essentially provided by platform
dependent cpufreq driver.
This commit unifies support for SW and HW (Intel) overclocking solutions
in the core cpufreq driver. Previously the "boost" sysfs attribute was
defined in the ACPI processor driver code. By default boost is disabled.
One global attribute is available at: /sys/devices/system/cpu/cpufreq/boost.
It only shows up when cpufreq driver supports overclocking.
Under the hood frequencies dedicated for boosting are marked with a
special flag (CPUFREQ_BOOST_FREQ) at driver's frequency table.
It is the user's concern to enable/disable overclocking with a proper call
to sysfs.
The cpufreq_boost_trigger_state() function is defined non static on purpose.
It is used later with thermal subsystem to provide automatic enable/disable
of the BOOST feature.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Add perf trace event "power:pstate_sample" to report driver state to
aid in diagnosing issues reported against intel_pstate.
Signed-off-by: Dirk Brandewie <dirk.j.brandewie@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
CPUFreq drivers that use clock frameworks interface,i.e. clk_get_rate(),
to get CPUs clk rate, have similar sort of code used in most of them.
This patch adds a generic ->get() which will do the same thing for them.
All those drivers are required to now is to set .get to cpufreq_generic_get()
and set their clk pointer in policy->clk during ->init().
Acked-by: Hans-Christian Egtvedt <egtvedt@samfundet.no>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
There are some parts of common kernel which would be using routines
like clk_get_rate() on some platforms. Currently, they wouldn't be
called for SA1100 boards, but they are needed for successful kernel
compilation.
Create a dummy clk_get_rate() routine for SA1100 which can be called
by the cpufreq core. More dummy routines might be added later if
necessary.
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Arnd Bergmann <arnd.bergmann@linaro.org>
Reported-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
[rjw: Changelog]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
When cpufreq_stats is compiled in as a module, cpufreq driver would
have already been registered. And so the CPUFREQ_CREATE_POLICY
notifiers wouldn't be called for it. Hence no sysfs entries for stats. :(
This patch calls cpufreq_stats_create_table() for each online CPU from
cpufreq_stats_init() and so if policy is already created for CPUx then
we will register sysfs stats for it.
When its not compiled as module, we will return early as policy wouldn't
be found for any of the CPUs.
Acked-by: Nicolas Pitre <nico@linaro.org>
Tested-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
We don't have code paths now where we need to do these two things
separately, so it is better do them in a single routine. Just as
they are allocated in a single routine.
Acked-by: Nicolas Pitre <nico@linaro.org>
Tested-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Either CPUs are hot-unplugged or suspend/resume occurs, cpufreq core
will send notifications to cpufreq-stats and stats structure and sysfs
entries would be correctly handled..
And so we don't actually need hotcpu notifiers in cpufreq-stats anymore.
We were only handling cpu hot-unplug events here and that are already
taken care of by POLICY notifiers.
Acked-by: Nicolas Pitre <nico@linaro.org>
Tested-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
There are several problems with cpufreq stats in the way it handles
cpufreq_unregister_driver() and suspend/resume..
- We must not lose data collected so far when suspend/resume happens
and so stats directories must not be removed/allocated during these
operations, which is done currently.
- cpufreq_stat has registered notifiers with both cpufreq and hotplug.
It adds sysfs stats directory with a cpufreq notifier: CPUFREQ_NOTIFY
and removes this directory with a notifier from hotplug core.
In case cpufreq_unregister_driver() is called (on rmmod cpufreq driver),
stats directories per cpu aren't removed as CPUs are still online. The
only call cpufreq_stats gets is cpufreq_stats_update_policy_cpu() for
all CPUs except the last of each policy. And pointer to stat information
is stored in the entry for last CPU in the per-cpu cpufreq_stats_table.
But policy structure would be freed inside cpufreq core and so that will
result in memory leak inside cpufreq stats (as we are never freeing
memory for stats).
Now if we again insert the module cpufreq_register_driver() will be
called and we will again allocate stats data and put it on for first
CPU of every policy. In case we only have a single CPU per policy, we
will return with a error from cpufreq_stats_create_table() due to this
code:
if (per_cpu(cpufreq_stats_table, cpu))
return -EBUSY;
And so probably cpufreq stats directory would not show up anymore (as
it was added inside last policies->kobj which doesn't exist anymore).
I haven't tested it, though. Also the values in stats files wouldn't
be refreshed as we are using the earlier stats structure.
- CPUFREQ_NOTIFY is called from cpufreq_set_policy() which is called for
scenarios where we don't really want cpufreq_stat_notifier_policy() to get
called. For example whenever we are changing anything related to a policy:
min/max/current freq, etc. cpufreq_set_policy() is called and so cpufreq
stats is notified. Where we don't do any useful stuff other than simply
returning with -EBUSY from cpufreq_stats_create_table(). And so this
isn't the right notifier that cpufreq stats..
Due to all above reasons this patch does following changes:
- Add new notifiers CPUFREQ_CREATE_POLICY and CPUFREQ_REMOVE_POLICY,
which are only called when policy is created/destroyed. They aren't
called for suspend/resume paths..
- Use these notifiers in cpufreq_stat_notifier_policy() to create/destory
stats sysfs entries. And so cpufreq_unregister_driver() or suspend/resume
shouldn't be a problem for cpufreq_stats.
- Return early from cpufreq_stat_cpu_callback() for suspend/resume sequence,
so that we don't free stats structure.
Acked-by: Nicolas Pitre <nico@linaro.org>
Tested-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The only caller of speedstep_get_state() was removed in commit d4019f0a92ab
("cpufreq: move freq change notifications to cpufreq core"). So building
speedstep-smi.o now triggers a GCC warning:
drivers/cpufreq/speedstep-smi.c:148:12: warning: 'speedstep_get_state' defined but not used [-Wunused-function]
Remove this unused function.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
* acpi-tools:
ACPICA: acpidump: Update MAINTAINERS file to include tools folder for ACPI/ACPICA.
ACPICA: acpidump: Enable tools Makefile to include acpi tools.
ACPICA: acpidump: Cleanup tools/power/acpi makefiles.
* pm-tools:
PM / tools: new tool for suspend/resume performance optimization
cpupower: Fix sscanf robustness in cpufreq-set
|
|
* acpi-modules:
platform: introduce OF style 'modalias' support for platform bus
ACPI: fix module autoloading for ACPI enumerated devices
ACPI: add module autoloading support for ACPI enumerated devices
ACPI: fix create_modalias() return value handling
|
|
Fix a problem that, the platform bus supports the OF style modalias
in .uevent() call, but not in its device 'modalias' sysfs attribute.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
* acpi-init:
ACPI / init: Run acpi_early_init() before timekeeping_init()
* acpi-hotplug:
ACPI / memhotplug: add parameter to disable memory hotplug
|
|
* pm-clk:
PM / clock_ops: report clock errors from clk_enable()
PM / clock_ops: check return of clk_enable() in pm_clk_resume()
PM / clock_ops: fix up clk prepare/unprepare count
|
|
* pm-cpuidle:
intel_idle: remove superfluous dev->state_count initialization
intel_idle: do C1E promotion disable quirk for hotplugged CPUs
ACPI / cpuidle: remove dev->state_count setting
ACPI / cpuidle: fix max idle state handling with hotplug CPU support
POWERPC: pseries: cpuidle: use the common cpuidle_[un]register() routines
POWERPC: pseries: cpuidle: remove superfluous dev->state_count initialization
ARM: EXYNOS: cpuidle: fix AFTR mode check
|
|
This tool is designed to assist kernel and OS developers in optimizing
their linux stack's suspend/resume time. Using a kernel image built with a
few extra options enabled, the tool will execute a suspend and will
capture dmesg and ftrace data until resume is complete. This data is
transformed into a device timeline and a callgraph to give a quick and
detailed view of which devices and callbacks are taking the most time in
suspend/resume. The output is a single html file which can be viewed in
firefox or chrome.
References: https://01.org/suspendresume
Signed-off-by: Todd Brandt <todd.e.brandt@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fix from Catalin Marinas:
"Revert "arm64: Fix memory shareability attribute for ioremap_wc/cache"
We noticed that it breaks ioremap (and earlyprintk) with 64K page
configuration"
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
Revert "arm64: Fix memory shareability attribute for ioremap_wc/cache"
|
|
Commit 74e72f894d56 ("lib/percpu_counter.c: fix __percpu_counter_add()")
looked very plausible, but its arithmetic was badly wrong: obvious once
you see the fix, but maddening to get there from the weird tmpfs ENOSPCs
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Ming Lei <tom.leiming@gmail.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Shaohua Li <shli@fusionio.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Fan Du <fan.du@windriver.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Ever since this driver was merged the following code was included:
if (skb->len < MISR)
skb->len = MISR;
MISR is defined to 0x3C which is also equivalent to ETH_ZLEN, but use
ETH_ZLEN directly which is exactly what we want to be checking for.
Reported-by: Marc Volovic <marcv@ezchip.com>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
On newer and faster machines (Vortex X86DX) using the r6040 driver, it
was noticed that the driver was returning an error during probing traced
down to being the MDIO bus probing and the inability to complete a MDIO
read operation in time. It turns out that the MDIO operations on these
faster machines usually complete after ~2140 iterations which is bigger
than 2048 (MAC_DEF_TIMEOUT) and results in spurious timeouts depending
on the system load.
Update r6040_phy_read() and r6040_phy_write() to include a 1
micro second delay in each busy-looping iteration of the loop which is a
much safer operation than incrementing MAC_DEF_TIMEOUT.
Reported-by: Nils Koehler <nils.koehler@ibt-interfaces.de>
Reported-by: Daniel Goertzen <daniel.goertzen@gmail.com>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This patch adds support for IPv6 checksum offload and GSO when those
features are available in the backend.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When initializing a gro_list for a packet, first check the rxhash of
the incoming skb against that of the skb's in the list. This should be
a very strong inidicator of whether the flow is going to be matched,
and potentially allows a lot of other checks to be short circuited.
Use skb_hash_raw so that we don't force the hash to be calculated.
Tested by running netperf 200 TCP_STREAMs between two machines with
GRO, HW rxhash, and 1G. Saw no performance degration, slight reduction
of time in dev_gro_receive.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Function to just return skb->rxhash without checking to see if it needs
to be recomputed.
Signed-off-by: Tom Herbert <therbert@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Remove unused function vxge_hw_vpath_vid_get
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Use 'make namespacecheck' to code that could be declared static.
After that remove code that is not being used.
Compile tested only.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
recently, dm9000 codes have many checkpatch errors and warnings:
WARNING: please, no space before tabs
3: FILE: dm9000.c:3:
+ * ^ICopyright (C) 1997 Sten Wang$
WARNING: please, no space before tabs
5: FILE: dm9000.c:5:
+ * ^IThis program is free software; you can redistribute it and/or$
WARNING: please, no space before tabs
6: FILE: dm9000.c:6:
+ * ^Imodify it under the terms of the GNU General Public License$
WARNING: please, no space before tabs
7: FILE: dm9000.c:7:
+ * ^Ias published by the Free Software Foundation; either version 2$
WARNING: please, no space before tabs
8: FILE: dm9000.c:8:
+ * ^Iof the License, or (at your option) any later version.$
WARNING: please, no space before tabs
10: FILE: dm9000.c:10:
+ * ^IThis program is distributed in the hope that it will be useful,$
WARNING: please, no space before tabs
11: FILE: dm9000.c:11:
+ * ^Ibut WITHOUT ANY WARRANTY; without even the implied warranty of$
WARNING: please, no space before tabs
12: FILE: dm9000.c:12:
+ * ^IMERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the$
WARNING: please, no space before tabs
13: FILE: dm9000.c:13:
+ * ^IGNU General Public License for more details.$
WARNING: do not add new typedefs
97: FILE: dm9000.c:97:
+typedef struct board_info {
ERROR: spaces prohibited around that ':' (ctx:WxV)
113: FILE: dm9000.c:113:
+ unsigned int in_suspend :1;
^
ERROR: spaces prohibited around that ':' (ctx:WxV)
114: FILE: dm9000.c:114:
+ unsigned int wake_supported :1;
^
This patch fixes important errors in it.
Signed-off-by: Barry Song <Baohua.Song@csr.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In PF_PACKET's packet mmap(), we can avoid using one atomic_inc()
and one atomic_dec() call in skb destructor and use a percpu
reference count instead in order to determine if packets are
still pending to be sent out. Micro-benchmark with [1] that has
been slightly modified (that is, protcol = 0 in socket(2) and
bind(2)), example on a rather crappy testing machine; I expect
it to scale and have even better results on bigger machines:
./packet_mm_tx -s7000 -m7200 -z700000 em1, avg over 2500 runs:
With patch: 4,022,015 cyc
Without patch: 4,812,994 cyc
time ./packet_mm_tx -s64 -c10000000 em1 > /dev/null, stable:
With patch:
real 1m32.241s
user 0m0.287s
sys 1m29.316s
Without patch:
real 1m38.386s
user 0m0.265s
sys 1m35.572s
In function tpacket_snd(), it is okay to use packet_read_pending()
since in fast-path we short-circuit the condition already with
ph != NULL, since we have next frames to process. In case we have
MSG_DONTWAIT, we also do not execute this path as need_wait is
false here anyway, and in case of _no_ MSG_DONTWAIT flag, it is
okay to call a packet_read_pending(), because when we ever reach
that path, we're done processing outgoing frames anyway and only
look if there are skbs still outstanding to be orphaned. We can
stay lockless in this percpu counter since it's acceptable when we
reach this path for the sum to be imprecise first, but we'll level
out at 0 after all pending frames have reached the skb destructor
eventually through tx reclaim. When people pin a tx process to
particular CPUs, we expect overflows to happen in the reference
counter as on one CPU we expect heavy increase; and distributed
through ksoftirqd on all CPUs a decrease, for example. As
David Laight points out, since the C language doesn't define the
result of signed int overflow (i.e. rather than wrap, it is
allowed to saturate as a possible outcome), we have to use
unsigned int as reference count. The sum over all CPUs when tx
is complete will result in 0 again.
The BUG_ON() in tpacket_destruct_skb() we can remove as well. It
can _only_ be set from inside tpacket_snd() path and we made sure
to increase tx_ring.pending in any case before we called po->xmit(skb).
So testing for tx_ring.pending == 0 is not too useful. Instead, it
would rather have been useful to test if lower layers didn't orphan
the skb so that we're missing ring slots being put back to
TP_STATUS_AVAILABLE. But such a bug will be caught in user space
already as we end up realizing that we do not have any
TP_STATUS_AVAILABLE slots left anymore. Therefore, we're all set.
Btw, in case of RX_RING path, we do not make use of the pending
member, therefore we also don't need to use up any percpu memory
here. Also note that __alloc_percpu() already returns a zero-filled
percpu area, so initialization is done already.
[1] http://wiki.ipxwarzone.com/index.php5?title=Linux_packet_mmap
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In tpacket_snd(), when we've discovered a first frame that is
not in status TP_STATUS_SEND_REQUEST, and return a NULL buffer,
we exit the send routine in case of MSG_DONTWAIT, since we've
finished traversing the mmaped send ring buffer and don't care
about pending frames.
While doing so, we still unconditionally call an expensive
schedule() in the packet_current_frame() "error" path, which
is unnecessary in this case since it's enough to just quit
the function.
Also, in case MSG_DONTWAIT is not set, we should rather test
for need_resched() first and do schedule() only if necessary
since meanwhile pending frames could already have finished
processing and called skb destructor.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Most people acquire PF_PACKET sockets with a protocol argument in
the socket call, e.g. libpcap does so with htons(ETH_P_ALL) for
all its sockets. Most likely, at some point in time a subsequent
bind() call will follow, e.g. in libpcap with ...
memset(&sll, 0, sizeof(sll));
sll.sll_family = AF_PACKET;
sll.sll_ifindex = ifindex;
sll.sll_protocol = htons(ETH_P_ALL);
... as arguments. What happens in the kernel is that already
in socket() syscall, we install a proto hook via register_prot_hook()
if our protocol argument is != 0. Yet, in bind() we're almost
doing the same work by doing a unregister_prot_hook() with an
expensive synchronize_net() call in case during socket() the proto
was != 0, plus follow-up register_prot_hook() with a bound device
to it this time, in order to limit traffic we get.
In the case when the protocol and user supplied device index (== 0)
does not change from socket() to bind(), we can spare us doing
the same work twice. Similarly for re-binding to the same device
and protocol. For these scenarios, we can decrease create/bind
latency from ~7447us (sock-bind-2 case) to ~89us (sock-bind-1 case)
with this patch.
Alternatively, for the first case, if people care, they should
simply create their sockets with proto == 0 argument and define
the protocol during bind() as this saves a call to synchronize_net()
as well (sock-bind-3 case).
In all other cases, we're tied to user space behaviour we must not
change, also since a bind() is not strictly required. Thus, we need
the synchronize_net() to make sure no asynchronous packet processing
paths still refer to the previous elements of po->prot_hook.
In case of mmap()ed sockets, the workflow that includes bind() is
socket() -> setsockopt(<ring>) -> bind(). In that case, a pair of
{__unregister, register}_prot_hook is being called from setsockopt()
in order to install the new protocol receive handler. Thus, when
we call bind and can skip a re-hook, we have already previously
installed the new handler. For fanout, this is handled different
entirely, so we should be good.
Timings on an i7-3520M machine:
* sock-bind-1: 89 us
* sock-bind-2: 7447 us
* sock-bind-3: 75 us
sock-bind-1:
socket(PF_PACKET, SOCK_RAW, htons(ETH_P_IP)) = 3
bind(3, {sa_family=AF_PACKET, proto=htons(ETH_P_IP), if=all(0),
pkttype=PACKET_HOST, addr(0)={0, }, 20) = 0
sock-bind-2:
socket(PF_PACKET, SOCK_RAW, htons(ETH_P_IP)) = 3
bind(3, {sa_family=AF_PACKET, proto=htons(ETH_P_IP), if=lo(1),
pkttype=PACKET_HOST, addr(0)={0, }, 20) = 0
sock-bind-3:
socket(PF_PACKET, SOCK_RAW, 0) = 3
bind(3, {sa_family=AF_PACKET, proto=htons(ETH_P_IP), if=lo(1),
pkttype=PACKET_HOST, addr(0)={0, }, 20) = 0
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Fixes: 9d8bf54 ("i40e: associate VMDq queue with VM type")
Reported-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This is a cleanup proposed by coccinelle. It replaces memcpy with struct
assignment on intel-mid's sfi layer.
Generated by: coccinelle/misc/memcpy-assign.cocci
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Link: http://lkml.kernel.org/r/1389917588-9785-1-git-send-email-david.a.cohen@linux.intel.com
Signed-off-by: David Cohen <david.a.cohen@linux.intel.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
|
|
Recent commit 438e38fadca2f6e57eeecc08326c8a95758594d4
("gre_offload: statically build GRE offloading support") added
new module_init/module_exit calls to the gre_offload.c file.
The file is obj-y and can't be anything other than built-in.
Currently it can never be built modular, so using module_init
as an alias for __initcall can be somewhat misleading.
Fix this up now, so that we can relocate module_init from
init.h into module.h in the future. If we don't do this, we'd
have to add module.h to obviously non-modular code, and that
would be a worse thing. We also make the inclusion explicit.
Note that direct use of __initcall is discouraged, vs. one
of the priority categorized subgroups. As __initcall gets
mapped onto device_initcall, our use of device_initcall
directly in this change means that the runtime impact is
zero -- it will remain at level 6 in initcall ordering.
As for the module_exit, rather than replace it with __exitcall,
we simply remove it, since it appears only UML does anything
with those, and even for UML, there is no relevant cleanup
to be done here.
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Building resource_tracker.o triggers a GCC warning:
drivers/net/ethernet/mellanox/mlx4/resource_tracker.c: In function 'mlx4_HW2SW_SRQ_wrapper':
drivers/net/ethernet/mellanox/mlx4/resource_tracker.c:3202:17: warning: 'srq' may be used uninitialized in this function [-Wmaybe-uninitialized]
atomic_dec(&srq->mtt->ref_count);
^
This is a false positive. But a cleanup of srq_res_start_move_to() can
help GCC here. The code currently uses a switch statement where a plain
if/else would do, since only two of the switch's four cases can ever
occur. Dropping that switch makes the warning go away.
While we're at it, add some missing braces, and convert state to the
correct type.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Building resource_tracker.o triggers a GCC warning:
drivers/net/ethernet/mellanox/mlx4/resource_tracker.c: In function 'mlx4_HW2SW_CQ_wrapper':
drivers/net/ethernet/mellanox/mlx4/resource_tracker.c:3019:16: warning: 'cq' may be used uninitialized in this function [-Wmaybe-uninitialized]
atomic_dec(&cq->mtt->ref_count);
^
This is a false positive. But a cleanup of cq_res_start_move_to() can
help GCC here. The code currently uses a switch statement where an
if/else construct would do too, since only two of the switch's four
cases can ever occur. Dropping that switch makes the warning go away.
While we're at it, add some missing braces.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
"disconnected" is too easily confused with "DCACHE_DISCONNECTED". I
think "unhashed" is the more precise term here.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Reviewed-by: Sage Weil <sage@inktank.com>
|
|
Commit 7509963c703b (e1000e: Fix a compile flag mis-match for
suspend/resume) moved suspend and resume hooks to be available when
CONFIG_PM is set. However, it can be set even if CONFIG_PM_SLEEP is not set
causing following warnings to be emitted:
drivers/net/ethernet/intel/e1000e/netdev.c:6178:12: warning:
‘e1000_suspend’ defined but not used [-Wunused-function]
drivers/net/ethernet/intel/e1000e/netdev.c:6185:12: warning:
‘e1000_resume’ defined but not used [-Wunused-function]
To fix this make the hooks to be available only when CONFIG_PM_SLEEP is set
and remove CONFIG_PM wrapping from driver ops because this is already
handled by SET_SYSTEM_SLEEP_PM_OPS() and SET_RUNTIME_PM_OPS().
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Dave Ertman <davidx.m.ertman@intel.com>
Cc: Aaron Brown <aaron.f.brown@intel.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Aaron Brown says:
====================
Intel Wired LAN Driver Updates
This series contains updates to ixgbe and ixgbevf.
John adds rtnl lock / unlock semantics for ixgbe_reinit_locked()
which was being called without the rtnl lock being held.
Jacob corrects an issue where ixgbevf_qv_disable function does not
set the disabled bit correctly.
From the community, Wei uses a type of struct for pci driver-specific
data in ixgbevf_suspend()
Don changes the way we store ring arrays in a manner that allows
support of multiple queues on multiple nodes and creates new ring
initialization functions for work previously done across multiple
functions - making the code closer to ixgbe and hopefully more readable.
He also fixes incorrect fiber eeprom write logic.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In this code we wanted to set the bit in IXGBE_SFF_SOFT_RS_SELECT_MASK to
the value in rs. So we really needed a logical or rather than an and, this
patch makes that change.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This patch creates new functions for ring initialization,
ixgbevf_configure_tx_ring() and ixgbevf_configure_rx_ring(). The work done
in these function previously was spread between several other functions and
this change should hopefully lead to greater readability and make the code
more like ixgbe. This patch also moves the placement of some older functions
to avoid having to write prototypes. It also promotes a couple of debug
messages to errors.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This will change how we store rings arrays in the adapter sturct.
We use to have a pointer to an array now we will be using an array
of pointers. This will allow us to support multiple queues on
muliple nodes at some point we would be able to reallocate the rings
so that each is on a local node if needed.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
