Peng Li says:
====================
net: hns3: add support ethtool_ops.{set|get}_coalesce for VF
This patch-set adds ethtool_ops.{get|set}_coalesce to VF and
fix one related bug.
HNS3 PF and VF driver use the common enet layer, as the
ethtool_ops.{get|set}_coalesce to PF have upstreamed, just
need add the ops to hns3vf_ethtool_ops.
[Patch 1/2] fix a related bug for the VF ethtool_ops.{set|
get}_coalesce.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Just like PF, if the int_gl_idx of VF does not be set, the default
interrupt coalesce index of VF is 0. But it should be GL1 for TX
queues and GL0 for RX queues.
This patch adds the int_gl_idx setup for VF.
Fixes: 200ecda42598 ("net: hns3: Add HNS3 VF HCL(Hardware Compatibility Layer) Support")
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds ethtool_ops.get/set_coalesce support to VF.
Since PF and VF share the same get/set_coalesce interface,
we only need to set hns3_get/set_coalesce to the ethtool_ops
when supporting get/set_coalesce for VF.
Signed-off-by: Fuyun Liang <liangfuyun1@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Marc Kleine-Budde says:
====================
pull-request: can-next 2018-01-26
this is a pull request for net-next/master consisting of 3 patches.
The first two patches target the CAN documentation. The first is by me
and fixes pointer to location of fsl,mpc5200-mscan node in the mpc5200
documentation. The second patch is by Robert Schwebel and it converts
the plain ASCII documentation to restructured text.
The third patch is by Fabrizio Castro add the r8a774[35] support to the
rcar_can dt-bindings documentation.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Change _regulator_list_voltage() argument from regulator to
regulator_dev in order to provide better separation of core layers.
Allow calling _regulator_list_voltage() from functions, with
regulator_dev argument. This refactoring is needed in order to
implement setting voltage of coupled regulators.
Signed-off-by: Maciej Purski <m.purski@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
As of_find_regulator_by_node() is an of function it should be moved from
core.c to of_regulator.c. It provides better separation of device tree
functions from the core and allows other of_functions in of_regulator.c
to resolve device_node to regulator_dev. This will be useful for
implementation of parsing coupled regulators properties.
Declare of_find_regulator_by_node() function in internal.h as well as
regulator_class and dev_to_rdev(), as they are needed by
of_find_regulator_by_node().
Signed-off-by: Maciej Purski <m.purski@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Based on commit ec718254cb
("ixgbe: Improve performance and reduce size of ixgbe_tx_map")
This change is meant to both improve the performance and reduce the size of
ixgbevf_tx_map().
Expand the work done in the main loop by pushing first into tx_buffer.
This allows us to pull in the dma_mapping_error check, the tx_buffer value
assignment, and the initial DMA value assignment to the Tx descriptor.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Based on commit d2bead576e
("igb: Clear Rx buffer_info in configure instead of clean")
This change makes it so that instead of going through the entire ring on Rx
cleanup we only go through the region that was designated to be cleaned up
and stop when we reach the region where new allocations should start.
In addition we can avoid having to perform a memset on the Rx buffer_info
structures until we are about to start using the ring again.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We already had placehloders for failed page and buffer allocations.
Added alloc_rx_page and made sure the stats are properly updated and
exposed in ethtool.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Based on commit bd4171a5d4
("igb: update code to better handle incrementing page count")
Update the driver code so that we do bulk updates of the page reference
count instead of just incrementing it by one reference at a time. The
advantage to doing this is that we cut down on atomic operations and
this in turn should give us a slight improvement in cycles per packet.
In addition if we eventually move this over to using build_skb the gains
will be more noticeable.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Based on commit 5be5955425
("igb: update driver to make use of DMA_ATTR_SKIP_CPU_SYNC")
and
commit 7bd1759282 ("igb: Add support for DMA_ATTR_WEAK_ORDERING")
Convert the calls to dma_map/unmap_page() to the attributes version
and add DMA_ATTR_SKIP_CPU_SYNC/WEAK_ORDERING which should help
improve performance on some platforms.
Move sync_for_cpu call before we perform a prefetch to avoid
invalidating the first 128 bytes of the packet on architectures where
that call may invalidate the cache.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Based on:
commit 7ec0116c91 ("igb: Use length to determine if descriptor is done")
This change makes it so that we use the length of the packet instead of the
DD status bit to determine if a new descriptor is ready to be processed.
The obvious advantage is that it cuts down on reads as we don't really even
need the DD bit if going from a 0 to a non-zero value on size is enough to
inform us that the packet has been completed.
In addition we only reset the Rx descriptor length for descriptor zero when
resetting a ring instead of having to do a memset with 0 over the entire
ring. By doing this we can save some time on initialization.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Based on commit 64f2525ca4 ("igb: Only DMA sync frame length")
On some architectures synching a buffer for DMA may be expensive.
Instead of the entire 2K receive buffer only synchronize the length of
the frame, which will typically be the MTU or smaller.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
get_user_pages() must be called with mmap_sem held, currently
it is not. In fact it is called under the user db_table->mutex.
To fix this we can convert gup to use the fast alternative,
and safely avoid taking mmap_sem, if possible. Furthermore
this is safe wrt to the mutex as other callers that take the
lock (unmap and alloc_db) are not called under mmap_sem
(hence possible deadlock).
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Ursula Braun says:
====================
net/smc: fixes 2018-01-26
here are some more smc patches. The first 4 patches take care about
different aspects of smc socket closing, the 5th patch improves
coding style.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Return statements in functions returning bool should use
true/false instead of 1/0.
This issue was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Closing a listen socket may hit the warning
WARN_ON(sock_owned_by_user(sk)) of tcp_close(), if the wake up of
the smc_tcp_listen_worker has not yet finished.
This patch introduces smc_close_wait_listen_clcsock() making sure
the listening internal clcsock has been closed in smc_tcp_listen_work(),
before the listening external SMC socket finishes closing.
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Increase the socket refcount during poll wait.
Take the socket lock before checking socket state.
For a listening socket return a mask independent of state SMC_ACTIVE and
cover errors or closed state as well.
Get rid of the accept_q loop in smc_accept_poll().
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RoCE device changes cause an IB event, processed in the global event
handler for the ROCE device. Problems for a certain Queue Pair cause a QP
event, processed in the QP event handler for this QP.
Among those events are port errors and other fatal device errors. All
link groups using such a port or device must be terminated in those cases.
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The error message prints are a little inconsisent, tidy them up to be a
little more consistent with current style recommendations.
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Steffen Klassert says:
====================
pull request (net-next): ipsec-next 2018-01-26
One last patch for this development cycle:
1) Add ESN support for IPSec HW offload.
From Yossef Efraim.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The soc_compr_copy callback is currently broken. Since the
changes to move the compr_ops over to the component the return
value is not correctly propagated, always returning zero on
success rather than the number of bytes copied. This causes
user-space to stall continuously reading as it does not believe
it has received any data.
Furthermore, the changes to move the compr_ops over to the
component iterate through the list of components and will call
the copy callback for any that have compressed ops. There isn't
currently any consensus on the mechanism to combine the results
of multiple copy callbacks.
To fix this issue for now halt searching the component list when
we locate a copy callback and return the result of that single
callback. Additional work should probably be done to look at the
other ops, tidy things up, and work out if we want to support
multiple components on a single compressed, but this is the only
fix required to get things working again.
Fixes: 9e7e3738ab ("ASoC: snd_soc_component_driver has snd_compr_ops")
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
David Ahern says:
====================
net/ipv6: Add support for ONLINK flag
Add support for RTNH_F_ONLINK with ipv6 routes.
First patch moves existing gateway validation into helper. The onlink
flag requires a different set of checks and the existing validation
makes ip6_route_info_create long enough.
Second patch makes the table id and lookup flag an option to
ip6_nh_lookup_table. onlink check needs to verify the gateway without
the RT6_LOOKUP_F_IFACE flag and PBR with VRF means the table id can
vary between the table the route is inserted and the VRF the egress
device is enslaved to.
Third patch adds support for RTNH_F_ONLINK.
I have a set of test cases in a format based on the framework Ido and
Jiri are working on. Once that goes in I will adapt the script and
submit.
v2
- removed table id check. Too constraining for PBR with VRF use cases
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Similar to IPv4 allow routes to be added with the RTNH_F_ONLINK flag.
The onlink option requires a gateway and a nexthop device. Any unicast
gateway is allowed (including IPv4 mapped addresses and unresolved
ones) as long as the gateway is not a local address and if it resolves
it must match the given device.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
onlink verification needs to do a lookup in potentially different
table than the table in fib6_config and without the RT6_LOOKUP_F_IFACE
flag. Change ip6_nh_lookup_table to take table id and flags as input
arguments. Both verifications want to ignore link state, so add that
flag can stay in the lookup helper.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move existing code to validate nexthop into a helper. Follow on patch
adds support for nexthops marked with onlink, and this helper keeps
the complexity of ip6_route_info_create in check.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Neil Berrington reported a double-fault on a VM with 768GB of RAM that uses
large amounts of vmalloc space with PTI enabled.
The cause is that load_new_mm_cr3() was never fixed to take the 5-level pgd
folding code into account, so, on a 4-level kernel, the pgd synchronization
logic compiles away to exactly nothing.
Interestingly, the problem doesn't trigger with nopti. I assume this is
because the kernel is mapped with global pages if we boot with nopti. The
sequence of operations when we create a new task is that we first load its
mm while still running on the old stack (which crashes if the old stack is
unmapped in the new mm unless the TLB saves us), then we call
prepare_switch_to(), and then we switch to the new stack.
prepare_switch_to() pokes the new stack directly, which will populate the
mapping through vmalloc_fault(). I assume that we're getting lucky on
non-PTI systems -- the old stack's TLB entry stays alive long enough to
make it all the way through prepare_switch_to() and switch_to() so that we
make it to a valid stack.
Fixes: b50858ce3e ("x86/mm/vmalloc: Add 5-level paging support")
Reported-and-tested-by: Neil Berrington <neil.berrington@datacore.com>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Cc: stable@vger.kernel.org
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Link: https://lkml.kernel.org/r/346541c56caed61abbe693d7d2742b4a380c5001.1516914529.git.luto@kernel.org
After upgrading an old laptop to 4.15-rc9, I found that the eth0 and
wlan0 interfaces had disappeared. It turns out that the b43 and b44
drivers require SSB_PCIHOST_POSSIBLE which depends on
PCI_DRIVERS_LEGACY, a config option that only exists on Mips.
Fixes: 58eae1416b ("ssb: Disable PCI host for PCI_DRIVERS_GENERIC")
Cc: stable@vger.org
Signed-off-by: Sven Joachim <svenjoac@gmx.de>
Reviewed-by: James Hogan <jhogan@kernel.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Local struct chip_data has two members that are not used:
- cs. Looks like was never used
- enable_dma. Became unused by the commit f89a6d8f43 ("spi: dw-mid: move
to use core SPI DMA mappings").
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
In this patch, consumers are allowed to set suspend voltage, and this
actually just set the "uV" in constraint::regulator_state, when the
regulator_suspend_late() was called by PM core through callback when
the system is entering into suspend, the regulator device would act
suspend activity then.
And it assumes that if any consumer set suspend voltage, the regulator
device should be enabled in the suspend state. And if the suspend
voltage of a regulator device for all consumers was set zero, the
regulator device would be off in the suspend state.
This patch also provides a new function hook to regulator devices for
resuming from suspend states.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Regualtor suspend/resume functions should only be called by PM suspend
core via registering dev_pm_ops, and regulator devices should implement
the callback functions. Thus, any regulator consumer shouldn't call
the regulator suspend/resume functions directly.
In order to avoid compile errors, two empty functions with the same name
still be left for the time being.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The items "disabled" and "enabled" are a little redundant, since only one
of them would be set to record if the regulator device should keep on
or be switched to off in suspend states.
So in this patch, the "disabled" was removed, only leave the "enabled":
- enabled == 1 for regulator-on-in-suspend
- enabled == 0 for regulator-off-in-suspend
- enabled == -1 means do nothing when entering suspend mode.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Some regulator consumers would like to make the regulator device
keeping a voltage range output when the system entering into
suspend states.
Making regulator voltage be an array can allow consumers to set voltage
for normal state as well as for suspend states through the same code.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Some systems need to set regulators to specific states when they enter
low power modes, especially around CPUs. There are many of these modes
depending on the particular runtime state.
Currently the regulator consumers are not granted permission to change
suspend state of regulator devices, the constraints are configured at
startup. In order to allow changes in a vlotage range, we need to add
new properties for voltage range and a flag to give permission to
change the suspend voltage and suspend on/off in suspend mode.
Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
