summaryrefslogtreecommitdiff
path: root/drivers
AgeCommit message (Collapse)Author
2021-02-08scsi: mpt3sas: Additional diagnostic buffer query interfaceSuganath Prabu S
When a host trace buffer is released, applications never know for what reason the buffer is released. Add a new IOCTL MPT3ADDNLDIAGQUERY to provide the trigger information due to which the diag buffer is released. Link: https://lore.kernel.org/r/20210204033724.1345-2-suganath-prabu.subramani@broadcom.com Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-02-08scsi: mpt3sas: Add support for shared host tagset for CPU hotplugSreekanth Reddy
MPT Fusion adapters can steer completions to individual queues and we now have support for shared host-wide tags in the I/O stack. The addition of the host-wide tags allows us to enable multiqueue support for MPT Fusion adapters. Once host-wise tags are enabled, the CPU hotplug feature is also supported. Allow use of host-wide tags to be disabled through the "host_tagset_enable" module parameter. Once we do not have any major performance regressions using host-wide tags, we will drop the hand-crafted interrupt affinity settings. Performance is meeting expectations. About 3.1M IOPS using 24 Drive SSD on Aero controllers. Link: https://lore.kernel.org/r/20210202095832.23072-1-sreekanth.reddy@broadcom.com Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-02-08scsi: mpt3sas: Fix ReplyPostFree pool allocationSreekanth Reddy
Currently the driver allocates memory for ReplyPostFree queues in chunks of 16. In resource constrained environments--such as VM with 1 GB RAM and 2 CPUs--memory allocation for ReplyPostFree pools may fail because the driver tries to allocate a memory for 16 ReplyPostFree queues even though the actual number needed is 2. Change the driver to allocate memory for only the actual number of queues needed if the ReplyPostFree queue count is less than 16. Link: https://lore.kernel.org/r/20210201141522.25363-1-sreekanth.reddy@broadcom.com Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-02-08scsi: pmcraid: Fix 'ioarcb' alignment warningArnd Bergmann
Building with 'make W=1' enables -Wpacked-not-aligned, and this warns about pmcraid because of incompatible alignment constraints for pmcraid_passthrough_ioctl_buffer: drivers/scsi/pmcraid.h:1044:1: warning: alignment 1 of 'struct pmcraid_passthrough_ioctl_buffer' is less than 32 [-Wpacked-not-aligned] 1044 | } __attribute__ ((packed)); | ^ drivers/scsi/pmcraid.h:1041:24: warning: 'ioarcb' offset 16 in 'struct pmcraid_passthrough_ioctl_buffer' isn't aligned to 32 [-Wpacked-not-aligned] 1041 | struct pmcraid_ioarcb ioarcb; The inner structure is documented as having 32 byte alignment here, but is starts at a 16 byte offset in the outer structure, so it's never actually aligned, as the outer structure is also marked 'packed'. Lee Jones point this out as one of the last files that need to be changed before the warning can be enabled by default. Change the annotations in a way that avoids the warning but leaves the layout unchanged, by removing the packing on the inner structure and adding it to the outer one. The one-byte request_buffer[] array should have been a flexible array member here, which is how I change it to avoid extra padding from the alignment attribute. Link: https://lore.kernel.org/r/20210204163020.3286210-1-arnd@kernel.org Cc: Lee Jones <lee.jones@linaro.org> Reviewed-by: Lee Jones <lee.jones@linaro.org> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-02-08scsi: scsi_debug: Fix a memory leakMaurizio Lombardi
The sdebug_q_arr pointer must be freed when the module is unloaded. $ cat /sys/kernel/debug/kmemleak unreferenced object 0xffff888e1cfb0000 (size 4096): comm "modprobe", pid 165555, jiffies 4325987516 (age 685.194s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000458f4f5d>] 0xffffffffc06702d9 [<000000003edc4b1f>] do_one_initcall+0xe9/0x57d [<00000000da7d518c>] do_init_module+0x1d1/0x6f0 [<000000009a6a9248>] load_module+0x36bd/0x4f50 [<00000000ddb0c3ce>] __do_sys_init_module+0x1db/0x260 [<000000009532db57>] do_syscall_64+0xa5/0x420 [<000000002916b13d>] entry_SYSCALL_64_after_hwframe+0x6a/0xdf Fixes: 87c715dcde63 ("scsi: scsi_debug: Add per_host_store option") Link: https://lore.kernel.org/r/20210208111734.34034-1-mlombard@redhat.com Acked-by: Douglas Gilbert <dgilbert@interlog.com> Signed-off-by: Maurizio Lombardi <mlombard@redhat.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-02-08clk: xilinx: move xlnx_vcu clock driver from socMichael Tretter
The xlnx_vcu driver is actually a clock controller driver which provides clocks that can be used by a driver for the encoder/decoder units. There is no reason to keep this driver in soc. Move the driver to clk. NOTE: The register mapping actually contains registers for AXI performance monitoring, but these are not used by the driver. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-16-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: fix alignment to open parenthesisMichael Tretter
Fixes the following checkpatch check: CHECK: Alignment should match open parenthesis #610: FILE: drivers/soc/xilinx/xlnx_vcu.c:610: + xvcu->vcu_slcr_ba = devm_ioremap(&pdev->dev, res->start, + resource_size(res)); Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-15-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: fix repeated word the in commentMichael Tretter
Fixes the following checkpatch warning: WARNING: Possible repeated word: 'the' #703: FILE: drivers/soc/xilinx/xlnx_vcu.c:703: + /* Add the the Gasket isolation and put the VCU in reset. */ Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-14-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: use bitfields for register definitionMichael Tretter
This makes the register accesses more readable and is closer to what is usually used in the kernel. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Reviewed-by: Stephen Boyd <sboyd@kernel.org> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-13-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: remove calculation of PLL configurationMichael Tretter
As the consumers are now responsible for setting the clock rate via clock framework, the clock rate is now calculated using round_rate and the driver does not need to calculate the clock rate beforehand. Remove the code that calculates the PLL configuration. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-12-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: make the PLL configurableMichael Tretter
Do not configure the PLL when probing the driver, but register the clock in the clock framework and do the configuration based on the respective callbacks. This is necessary to allow the consumers, i.e., encoder and decoder drivers, of the xlnx_vcu clock provider to set the clock rate and actually enable the clocks without relying on some pre-configuration. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-11-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: make pll post divider explicitMichael Tretter
According to the downstream driver documentation due to timing constraints the output divider of the PLL has to be set to 1/2. Add a helper function for that check instead of burying the code in one large setup function. The bit is undocumented and marked as reserved in the register reference. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-10-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: implement clock provider for output clocksMichael Tretter
The VCU System-Level Control uses an internal PLL to drive the core and MCU clock for the allegro encoder and decoder based on an external PL clock. In order be able to ensure that the clocks are enabled and to get their rate from other drivers, the module must implement a clock provider and register the clocks at the common clock framework. Other drivers are then able to access the clock via devicetree bindings. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-9-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: register PLL as fixed rate clockMichael Tretter
Currently, xvcu_pll_set_rate configures the PLL to a clock rate that is pre-calculated when probing the driver. To still make the clock framework aware of the PLL and to allow to configure other clocks based on the PLL rate, register the PLL as a fixed rate clock. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-8-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: implement PLL disableMichael Tretter
The disabling of the PLL is not fully implemented, because according to the ZynqMP register reference the RESET, POR_IN and PWR_POR bits have to be set to bring the PLL into reset. Set the bits to disable the PLL. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-7-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: add helpers for configuring PLLMichael Tretter
The xvcu_set_vcu_pll_info function sets the rate of the PLL and enables it, which makes it difficult to cleanly convert the driver to the common clock framework. Split the function and add separate functions for setting the rate, enabling the clock and disabling the clock. Also move the enable of the reference clock from probe to the helper that enables the PLL. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-6-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: add helper to wait for PLL lockedMichael Tretter
Extract a helper function to wait until the PLL is locked. Also, disabling the bypass was buried in the exit path on the wait loop. Separate the different steps and add a helper function to make the code more readable. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-5-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08soc: xilinx: vcu: drop coreclk from struct xlnx_vcuMichael Tretter
The coreclk field is newer read after being written to xlnx_vcu. Remove the coreclk field from the xlnx_vcu and use a function local variable instead. Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-4-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08clk: divider: fix initialization with parent_hwMichael Tretter
If a driver registers a divider clock with a parent_hw instead of the parent_name, the parent_hw is ignored and the clock does not have a parent. Fix this by initializing the parents the same way they are initialized for clock gates. Fixes: ff258817137a ("clk: divider: Add support for specifying parents via DT/pointers") Signed-off-by: Michael Tretter <m.tretter@pengutronix.de> Reviewed-by: Stephen Boyd <sboyd@kernel.org> Acked-by: Michal Simek <michal.simek@xilinx.com> Link: https://lore.kernel.org/r/20210121071659.1226489-3-m.tretter@pengutronix.de Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08clk: axi-clkgen: use devm_platform_ioremap_resource() short-handAlexandru Ardelean
No major functional change. Noticed while checking the driver code that this could be used. Saves two lines. Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com> Link: https://lore.kernel.org/r/20210201151245.21845-5-alexandru.ardelean@analog.com Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08clk: clk-axiclkgen: add ZynqMP PFD and VCO limitsAlexandru Ardelean
For ZynqMP (Ultrascale) the PFD and VCO limits are different. In order to support these, this change adds a compatible string (i.e. 'adi,zynqmp-axi-clkgen-2.00.a') which will take into account for these limits and apply them. Signed-off-by: Dragos Bogdan <dragos.bogdan@analog.com> Signed-off-by: Mathias Tausen <mta@gomspace.com> Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com> Link: https://lore.kernel.org/r/20210201151245.21845-3-alexandru.ardelean@analog.com Acked-by: Moritz Fischer <mdf@kernel.org> Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08clk: axi-clkgen: replace ARCH dependencies with driver depsAlexandru Ardelean
The intent is to be able to run this driver to access the IP core in setups where FPGA board is also connected via a PCIe bus. In such cases the number of combinations explodes, where the host system can be an x86 with Xilinx Zynq/ZynqMP/Microblaze board connected via PCIe. Or even a ZynqMP board with a ZynqMP/Zynq/Microblaze connected via PCIe. To accommodate for these cases, this change removes the limitation for this driver to be compilable only on Zynq/Microblaze architectures. And adds dependencies on the mechanisms required by the driver to work (OF and HAS_IOMEM). Signed-off-by: Dragos Bogdan <dragos.bogdan@analog.com> Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com> Link: https://lore.kernel.org/r/20210201151245.21845-2-alexandru.ardelean@analog.com Acked-by: Michal Simek <michal.simek@xilinx.com> Reviewed-by: Moritz Fischer <mdf@kernel.org> Signed-off-by: Stephen Boyd <sboyd@kernel.org>
2021-02-08scsi: sd: Warn if unsupported ZBC device is probedDamien Le Moal
In sd_probe(), print a warning if CONFIG_BLK_DEV_ZONED is disabled and a TYPE_ZBC device is found. While at it, use IS_ENABLED() to test if CONFIG_BLK_DEV_ZONED is enabled instead using of a #ifdef. Link: https://lore.kernel.org/r/20210128055658.530133-1-damien.lemoal@wdc.com Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2021-02-08i40e: Log error for oversized MTU on deviceEryk Rybak
When attempting to link XDP prog with MTU larger than supported, user is not informed why XDP linking fails. Adding proper error message: "MTU too large to enable XDP". Signed-off-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com> Signed-off-by: Eryk Rybak <eryk.roch.rybak@intel.com> Tested-by: Kiran Bhandare <kiranx.bhandare@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08i40e: consolidate handling of XDP program actionsCristian Dumitrescu
Consolidate the actions performed on the packet based on the XDP program result into a separate function that is easier to read and maintain. Simplify the i40e_construct_skb_zc function, so that the input xdp buffer is always freed, regardless of whether the output skb is successfully created or not. Simplify the behavior of the i40e_clean_rx_irq_zc function, so that the current packet descriptor is dropped when function i40_construct_skb_zc returns an error as opposed to re-processing the same description on the next invocation. Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com> Tested-by: Kiran Bhandare <kiranx.bhandare@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08i40e: remove the redundant buffer info updatesCristian Dumitrescu
For performance reasons, remove the redundant buffer info updates (*bi = NULL). The buffers ready to be cleaned can easily be tracked based on the ring next-to-clean variable, which is consistently updated. Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com> Tested-by: Kiran Bhandare <kiranx.bhandare@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08i40e: remove unnecessary cleaned_count updatesCristian Dumitrescu
For performance reasons, remove the redundant updates of the cleaned_count variable, as its value can be computed based on the ring next-to-clean variable, which is consistently updated. Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com> Tested-by: Kiran Bhandare <kiranx.bhandare@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08i40e: remove unnecessary memory writes of the next to clean pointerCristian Dumitrescu
For performance reasons, avoid writing the ring next-to-clean pointer value back to memory on every update, as it is not really necessary. Instead, simply read it at initialization into a local copy, update the local copy as necessary and write the local copy back to memory after the last update. Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com> Tested-by: Kiran Bhandare <kiranx.bhandare@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08mlxsw: spectrum_router: Set offload_failed flagAmit Cohen
When FIB_EVENT_ENTRY_{REPLACE, APPEND} are triggered and route insertion fails, FIB abort is triggered. After aborting, set the appropriate hardware flag to make the kernel emit RTM_NEWROUTE notification with RTM_F_OFFLOAD_FAILED flag. Signed-off-by: Amit Cohen <amcohen@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-08netdevsim: fib: Add debugfs to debug route offload failureAmit Cohen
Add "fail_route_offload" flag to disallow offloading routes. It is needed to test "offload failed" notifications. Create the flag as part of nsim_fib_create() under fib directory and set it to false by default. When FIB_EVENT_ENTRY_{REPLACE, APPEND} are triggered and "fail_route_offload" value is true, set the appropriate hardware flag to make the kernel emit RTM_NEWROUTE notification with RTM_F_OFFLOAD_FAILED flag. Signed-off-by: Amit Cohen <amcohen@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-08netdevsim: dev: Initialize FIB module after debugfsIdo Schimmel
Initialize the dummy FIB offload module after debugfs, so that the FIB module could create its own directory there. Signed-off-by: Amit Cohen <amcohen@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-08netdevsim: fib: Do not warn if route was not found for several eventsAmit Cohen
The next patch will add the ability to fail route offload controlled by debugfs variable called "fail_route_offload". If we vetoed the addition, we might get a delete or append notification for a route we do not have. Therefore, do not warn if route was not found. Signed-off-by: Amit Cohen <amcohen@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-08IPv6: Add "offload failed" indication to routesAmit Cohen
After installing a route to the kernel, user space receives an acknowledgment, which means the route was installed in the kernel, but not necessarily in hardware. The asynchronous nature of route installation in hardware can lead to a routing daemon advertising a route before it was actually installed in hardware. This can result in packet loss or mis-routed packets until the route is installed in hardware. To avoid such cases, previous patch set added the ability to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/RTM_F_TRAP flags are changed, this behavior is controlled by sysctl. With the above mentioned behavior, it is possible to know from user-space if the route was offloaded, but if the offload fails there is no indication to user-space. Following a failure, a routing daemon will wait indefinitely for a notification that will never come. This patch adds an "offload_failed" indication to IPv6 routes, so that users will have better visibility into the offload process. 'struct fib6_info' is extended with new field that indicates if route offload failed. Note that the new field is added using unused bit and therefore there is no need to increase struct size. Signed-off-by: Amit Cohen <amcohen@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-08IPv4: Add "offload failed" indication to routesAmit Cohen
After installing a route to the kernel, user space receives an acknowledgment, which means the route was installed in the kernel, but not necessarily in hardware. The asynchronous nature of route installation in hardware can lead to a routing daemon advertising a route before it was actually installed in hardware. This can result in packet loss or mis-routed packets until the route is installed in hardware. To avoid such cases, previous patch set added the ability to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/RTM_F_TRAP flags are changed, this behavior is controlled by sysctl. With the above mentioned behavior, it is possible to know from user-space if the route was offloaded, but if the offload fails there is no indication to user-space. Following a failure, a routing daemon will wait indefinitely for a notification that will never come. This patch adds an "offload_failed" indication to IPv4 routes, so that users will have better visibility into the offload process. 'struct fib_alias', and 'struct fib_rt_info' are extended with new field that indicates if route offload failed. Note that the new field is added using unused bit and therefore there is no need to increase structs size. Signed-off-by: Amit Cohen <amcohen@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-02-08RDMA/qedr: Remove in_irq() usage from debug outputSebastian Andrzej Siewior
qedr_gsi_post_send() has a debug output which prints the return value of in_irq() and irqs_disabled(). The result of the in_irq(), even if invoked from an interrupt handler, is subject to change depending on the `threadirqs' command line switch. The result of irqs_disabled() is always be 1 because the function acquires spinlock_t with spin_lock_irqsave(). Remove in_irq() and irqs_disabled() from the debug output because it provides little value. Link: https://lore.kernel.org/r/20210208193347.383254-1-bigeasy@linutronix.de Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
2021-02-08RDMA/rxe: Cleanup init_send_wqeBob Pearson
This patch changes the type of init_send_wqe in rxe_verbs.c to void since it always returns 0. It also separates out the code that copies inline data into the send wqe as copy_inline_data_to_wqe(). Link: https://lore.kernel.org/r/20210206002437.2756-1-rpearson@hpe.com Signed-off-by: Bob Pearson <rpearson@hpe.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
2021-02-08RDMA/rxe: Fix minor coding style issuesBob Pearson
checkpatch -f found 3 warnings in RDMA/rxe 1. a missing space following switch 2. return followed by else 3. use of strlcpy() instead of strscpy(). This patch fixes each of these. In ... } elseif (...) { ... return 0; } else ... The middle block can be safely moved since it is completely independent of the other code. Link: https://lore.kernel.org/r/20210205230525.49068-1-rpearson@hpe.com Signed-off-by: Bob Pearson <rpearson@hpe.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
2021-02-08RDMA/mlx5: Cleanup the synchronize_srcu() from the ODP flowYishai Hadas
Cleanup the synchronize_srcu() from the ODP flow as it was found to be a very heavy time consumer as part of dereg_mr. For example de-registration of 10000 ODP MRs each with size of 2M hugepage took 19.6 sec comparing de-registration of same number of non ODP MRs that took 172 ms. The new locking scheme uses the wait_event() mechanism which follows the use count of the MR instead of using synchronize_srcu(). By that change, the time required for the above test took 95 ms which is even better than the non ODP flow. Once fully dropped the srcu usage, had to come with a lock to protect the XA access. As part of using the above mechanism we could also clean the num_deferred_work stuff and follow the use count instead. Link: https://lore.kernel.org/r/20210202071309.2057998-1-leon@kernel.org Signed-off-by: Yishai Hadas <yishaih@nvidia.com> Signed-off-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
2021-02-08ice: Improve MSI-X fallback logicTony Nguyen
Currently if the driver is unable to get all the MSI-X vectors it wants, it falls back to the minimum configuration which equates to a single Tx/Rx traffic queue pair. Instead of using the minimum configuration, if given more vectors than the minimum, utilize those vectors for additional traffic queues after accounting for other interrupts. Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com>
2021-02-08ice: Fix trivial error messageMitch Williams
This message indicates an error on close, not open. Signed-off-by: Mitch Williams <mitch.a.williams@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: remove unnecessary castsBruce Allan
Casting a void * rvalue in an assignment is unnecessary in C; remove the casts. Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: Refactor DCB related variables out of the ice_port_info structChinh T Cao
Refactor the DCB related variables out of the ice_port_info_struct. The goal is to make the ice_port_info struct cleaner. Signed-off-by: Chinh T Cao <chinh.t.cao@intel.com> Co-developed-by: Dave Ertman <david.m.ertman@intel.com> Signed-off-by: Dave Ertman <david.m.ertman@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: fix writeback enable logicJesse Brandeburg
The writeback enable logic was incorrectly implemented (due to misunderstanding what the side effects of the implementation would be during polling). Fix this logic issue, while implementing a new feature allowing the user to control the writeback frequency using the knobs for controlling interrupt throttling that we already have. Basically if you leave adaptive interrupts enabled, the writeback frequency will be varied even if busy_polling or if napi-poll is in use. If the interrupt rates are set to a fixed value by ethtool -C and adaptive is off, the driver will allow the user-set interrupt rate to guide how frequently the hardware will complete descriptors to the driver. Effectively the user will get a control over the hardware efficiency, allowing the choice between immediate interrupts or delayed up to a maximum of the interrupt rate, even when interrupts are disabled during polling. Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Co-developed-by: Brett Creeley <brett.creeley@intel.com> Signed-off-by: Brett Creeley <brett.creeley@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: Use PSM clock frequency to calculate RL profilesBen Shelton
The core clock frequency is currently hardcoded at 446 MHz for the RL profile calculations. This causes issues since not all devices use that clock frequency. Read the GLGEN_CLKSTAT_SRC register to determine which PSM clock frequency is selected. This ensures that the rate limiter profile calculations will be correct. Signed-off-by: Ben Shelton <benjamin.h.shelton@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: create scheduler aggregator node config and move VSIsKiran Patil
Create set scheduler aggregator node and move for VSIs into respective scheduler node. Max children per aggregator node is 64. There are two types of aggregator node(s) created. 1. dedicated node for PF and _CTRL VSIs 2. dedicated node(s) for VFs. As part of reset and rebuild, aggregator nodes are recreated and VSIs are moved to respective aggregator node. Having related VSIs in respective tree avoid starvation between PF and VF w.r.t Tx bandwidth. Co-developed-by: Tarun Singh <tarun.k.singh@intel.com> Signed-off-by: Tarun Singh <tarun.k.singh@intel.com> Co-developed-by: Victor Raj <victor.raj@intel.com> Signed-off-by: Victor Raj <victor.raj@intel.com> Co-developed-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Signed-off-by: Kiran Patil <kiran.patil@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: Add initial support framework for LAGDave Ertman
Add the framework and initial implementation for receiving and processing netdev bonding events. This is only the software support and the implementation of the HW offload for bonding support will be coming at a later time. There are some architectural gaps that need to be closed before that happens. Because this is a software only solution that supports in kernel bonding, SR-IOV is not supported with this implementation. Signed-off-by: Dave Ertman <david.m.ertman@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: Remove xsk_buff_pool from VSI structureMichal Swiatkowski
Current implementation of netdev already contains xsk_buff_pools. We no longer have to contain these structures in ice_vsi. Refactor the code to operate on netdev-provided xsk_buff_pools. Move scheduling napi on each queue to a separate function to simplify setup function. Signed-off-by: Michal Swiatkowski <michal.swiatkowski@intel.com> Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com> Tested-by: Kiran Bhandare <kiranx.bhandare@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: implement new LLDP filter commandDave Ertman
There is an issue with some NVMs where an already existent LLDP filter is blocking the creation of a filter to allow LLDP packets to be redirected to the default VSI for the interface. This is blocking all LLDP functionality based in the kernel when the FW LLDP agent is disabled (e.g. software based DCBx). Implement the new AQ command to allow adding VSI destinations to existent filters on NVM versions that support the new command. The new lldp_fltr_ctrl AQ command supports Rx filters only, so the code flow for adding filters to disable Tx of control frames will remain intact. Signed-off-by: Dave Ertman <david.m.ertman@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08ice: log message when trusted VF goes in/out of promisc modeBrett Creeley
Currently there is no message printed on the host when a VF goes in and out of promiscuous mode. This is causing confusion because this is the expected behavior based on i40e. Fix this. Signed-off-by: Brett Creeley <brett.creeley@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2021-02-08RDMA/hns: Delete redundant judgment when preparing descriptorsXinhao Liu
There is no need to use a for loop to assign values for an array of cmd descriptors which has only two elements. Link: https://lore.kernel.org/r/1612517974-31867-13-git-send-email-liweihang@huawei.com Signed-off-by: Xinhao Liu <liuxinhao5@hisilicon.com> Signed-off-by: Weihang Li <liweihang@huawei.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
generated by cgit (git 2.34.1) at 2025-10-02 09:48:09 +0000