summaryrefslogtreecommitdiff
path: root/drivers
AgeCommit message (Collapse)Author
2022-10-02watchdog: imx93: add watchdog timer on imx93Alice Guo
The WDOG clocks are sourced from lpo_clk, and lpo_clk is the fixed 32KHz. TOVAL contains the 16-bit value used to set the timeout period of the watchdog. When the timeout period exceeds 2 seconds, the value written to the TOVAL register is larger than 16-bit can represent. Enabling watchdog prescaler can solve this problem. Two points need to be aware of: 1. watchdog prescaler enables a fixed 256 pre-scaling of watchdog counter reference clock 2. reconfiguration takes about 55ms on imx93 Reviewed-by: Jacky Bai <ping.bai@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-8-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: imx7ulp_wdt: init wdog when it was activeJason Liu
Paired with suspend, we can only init wdog again when it was active and ping it once to avoid the watchdog timeout after it resumed. Signed-off-by: Jason Liu <jason.hui.liu@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Ye Li <ye.li@nxp.com> Reviewed-by: Jacky Bai <ping.bai@nxp.com> Tested-by: Jacky Bai <ping.bai@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-7-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: imx7ulp_wdt: Handle wdog reconfigure failureYe Li
Current driver may meet reconfigure failure caused by below reasons: 1. The wdog on iMX7ULP has different behavior after RCS valid. It needs to wait more than 2.5 wdog clock for clock sync before next reconfiguration, while imx8ulp wdog does not need such delay. 2. After unlock, there is 128 bus clock window opened for reconfiguration, but on iMX8ULP, the HW can't guarantee the latency. So it is possible the window is closed before the writing arrives to wdog. 3. If the PRES is enabled, the RCS valid time becomes x256 to the time of PRES disabled. It is about 1715ms on iMX8ULP. So We have to increase the RCS timeout and can't wait it in IRQ disabled. The patch updates the driver to handle failures 1. Using different wait for unlock and RCS. Unlock valid time is very short and only related to bus clock. It must be in IRQ disabled to avoid being interrupted in 128 clock window. But for RCS time, it is longer and ok for IRQ enabled. 2. Add retry for any reconfigure failure with default 5 times. 3. Add "fsl,imx8ulp-wdt" compatile string for iMX8ULP and afterwards platform which don't need more 2.5 wdog clock after RCS valid. For imx7ulp, add post delay of 2.5 clock after RCS valid. Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Jacky Bai <ping.bai@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-6-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: imx7ulp_wdt: Fix RCS timeout issueYe Li
According to measure on i.MX7ULP and i.MX8ULP, the RCS done needs about 3400us and 6700us respectively. So current 20us timeout is not enough. When reconfiguring is on-going, unlock and configure CS will lead to unknown result. Increase the wait timeout value to 10ms and check the return value of RCS wait to fix the issue Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Jacky Bai <ping.bai@nxp.com> Acked-by: Jason Liu <jason.hui.liu@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-5-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: imx7ulp_wdt: Check CMD32EN in wdog initYe Li
When bootloader has enabled the CMD32EN bit, switch to use 32bits unlock command to unlock the CS register. Using 32bits command will help on avoiding 16 bus cycle window violation for two 16 bits commands. Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Jacky Bai <ping.bai@nxp.com> Acked-by: Jason Liu <jason.hui.liu@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-4-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: imx7ulp: Add explict memory barrier for unlock sequenceJacky Bai
When reconfiguring the WDOG Timer of i.MX7ULP, there is a certain probability causes it to reset. The reason is that the CMD32EN of the WDOG Timer of i.MX7ULP is disabled in bootloader. The unlock sequence are two 16-bit writes to the CNT register within 16 bus clocks. Adding mb() is to guarantee that two 16-bit writes are finished within 16 bus clocks. Memory barriers cannot be added between these two 16-bit writes so that writel_relaxed is used. Suggested-by: Ye Li <ye.li@nxp.com> Signed-off-by: Jacky Bai <ping.bai@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Ye Li <ye.li@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-3-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: imx7ulp: Move suspend/resume to noirq phaseAnson Huang
The i.MX7ULP's watchdog is enabled by default when out of reset, so the resume callback which is to disable watchdog should be called earlier to avoid unexpected timeout, move suspend/resume callback to noirq phase. Signed-off-by: Anson Huang <Anson.Huang@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Jacky Bai <ping.bai@nxp.com> Tested-by: Peter Chen <peter.chen@nxp.com> Tested-by: Li Jun <jun.li@nxp.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220825083256.14565-2-alice.guo@oss.nxp.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: rti-wdt:using the pm_runtime_resume_and_get to simplify the codeye xingchen
Using pm_runtime_resume_and_get() to instade of pm_runtime_get_sync and pm_runtime_put_noidle. Reported-by: Zeal Robot <zealci@zte.com.cn> Signed-off-by: ye xingchen <ye.xingchen@zte.com.cn> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220802074643.1648660-1-ye.xingchen@zte.com.cn Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: s3c2410_wdt: support exynosautov9 watchdogChanho Park
Like exynos850, exynosautov9 SoC also has two cpu watchdogs. Unfortunately, some configurations are slightly different so we need to add samsung,exynosautov9-wdt and separate drv data for those watchdogs. Signed-off-by: Chanho Park <chanho61.park@samsung.com> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220520121750.71473-3-chanho61.park@samsung.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: npcm: Enable clock if providedJonathan Neuschäfer
On the Nuvoton WPCM450 SoC, with its upcoming clock driver, peripheral clocks are individually gated and ungated. Therefore, the watchdog driver must be able to ungate the watchdog clock. Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220610072141.347795-3-j.neuschaefer@gmx.net Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: meson: keep running if already activePhilippe Boos
If the watchdog is already running (e.g.: started by bootloader) then the kernel driver should keep the watchdog active but the amlogic driver turns it off. Let the driver fix the clock rate if already active because we do not know the previous timebase value. To avoid unintentional resetting we temporarily set it to its maximum value. Then keep the enable bit if is was previously active. Signed-off-by: Philippe Boos <pboos@baylibre.com> Reviewed-by: Jerome Brunet <jbrunet@baylibre.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220801092150.4449-1-pboos@baylibre.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: armada_37xx_wdt: Fix .set_timeout callbackPali Rohár
ioctl(WDIOC_SETTIMEOUT) calls .set_timeout and .ping callbacks and it is expected that it changes current watchdog timeout. armada_37xx_wdt's .ping callback just reping counter 0 and does not touch counter 1 used for timeout. So it is needed to set counter 1 to the new value in .set_timeout callback to ensure ioctl(WDIOC_SETTIMEOUT) functionality. Fix it. Fixes: 54e3d9b518c8 ("watchdog: Add support for Armada 37xx CPU watchdog") Signed-off-by: Pali Rohár <pali@kernel.org> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Reviewed-by: Marek Behún <kabel@kernel.org> Link: https://lore.kernel.org/r/20220726085612.10672-1-pali@kernel.org Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: sa1100: make variable sa1100dog_driver staticsunliming
This symbol is not used outside of sa1100_wdt.c, so marks it static. Fixes the following warning: >> drivers/watchdog/sa1100_wdt.c:241:24: sparse: sparse: symbol 'sa1100dog_driver' was not declared. Should it be static? Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: sunliming <sunliming@kylinos.cn> Acked-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220802020819.1226454-1-sunliming@kylinos.cn Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: w83977f_wdt: Fix comment typoJason Wang
The double `we' is duplicated in the comment, remove one. Signed-off-by: Jason Wang <wangborong@cdjrlc.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220802201109.6843-1-wangborong@cdjrlc.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: eurotechwdt: Remove redundant word in commentsshaomin Deng
There is a rebundant word "we" in comments, so remove it. Signed-off-by: shaomin Deng <dengshaomin@cdjrlc.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220808153956.8374-1-dengshaomin@cdjrlc.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: Check dev_set_name() return valueBo Liu
It's possible that dev_set_name() returns -ENOMEM, catch and handle this. Signed-off-by: Bo Liu <liubo03@inspur.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220920020312.2383-1-liubo03@inspur.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: rzg2l_wdt: Add rzv2m supportPhil Edworthy
The WDT on RZ/V2M devices is basically the same as RZ/G2L, but without the parity error registers. This means the driver has to reset the hardware plus set the minimum timeout in order to do a restart and has a single interrupt. Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com> Reviewed-by: Biju Das <biju.das.jz@bp.renesas.com> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220823093233.8577-3-phil.edworthy@renesas.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: ftwdt010_wdt: fix test for platform_get_irq() failureDan Carpenter
This code assumes that platform_get_irq() function returns zero on failure. In fact, platform_get_irq() never returns zero. It returns negative error codes or positive non-zero values on success. Fixes: eca10ae6000d ("watchdog: add driver for Cortina Gemini watchdog") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/YvTgRk/ABp62/hNA@kili Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: Exar/MaxLinear XR28V38x driverDavid Müller
Simple driver for the watchdog present in some Exar/MaxLinear UART chips. Please see https://www.maxlinear.com/product/interface/uarts/lpc-uarts/xr28v384 for more info. Signed-off-by: David Müller <d.mueller@elsoft.ch> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220914094605.93377-1-d.mueller@elsoft.ch Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: ftwdt010_wdt: implement _restart() functionSergei Antonov
Implement ftwdt010_wdt_restart(). It enables watchdog with timeout = 0 and disabled IRQ. Since it needs code similar to ftwdt010_wdt_start(), add a new function ftwdt010_enable() and move common code there. Suggested-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Sergei Antonov <saproj@gmail.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220829090436.452742-1-saproj@gmail.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: w83627hf_wdt: add bootstatus supportHenning Schild
The status bit in the status and control register can tell us whether the last reboot was caused by the watchdog. Make sure to take that into the bootstatus before clearing it. Signed-off-by: Henning Schild <henning.schild@siemens.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220824152448.7736-1-henning.schild@siemens.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: aspeed_wdt: Reorder output signal register configurationChin-Ting Kuo
If the output driving type is push-pull mode, the output polarity should be selected in advance. Otherwise, an unexpected value will be output at the moment of changing to push-pull mode. Thus, output polarity, WDT18[31], must be configured before changing driving type, WDT18[30]. Signed-off-by: Chin-Ting Kuo <chin-ting_kuo@aspeedtech.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Tested-by: Bonnie Lo <Bonnie_Lo@wiwynn.com> Reviewed-by: Joel Stanley <joel@jms.id.au> Link: https://lore.kernel.org/r/20220819094905.1962513-1-chin-ting_kuo@aspeedtech.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog/hpwdt: Enable HP_WATCHDOG for ARM64 systems.Jerry Hoemann
Enable HP_WATCHDOG for ARM64 systems. HPWDT_NMI_DECODING requires X86 as NMI handlers are X86 specific. Signed-off-by: Jerry Hoemann <jerry.hoemann@hpe.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220820202821.1263837-3-jerry.hoemann@hpe.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog/hpwdt: Include nmi.h only if CONFIG_HPWDT_NMI_DECODINGJerry Hoemann
Fixes: d48b0e173715 ("x86, nmi, drivers: Fix nmi splitup build bug") Arm64 does not support NMI and has no <asm/nmi.h>. Include <asm/nmi.h> only if CONFIG_HPWDT_NMI_DECODING is defined to avoid build failure on non-existent header file on Arm64. Signed-off-by: Jerry Hoemann <jerry.hoemann@hpe.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20220820202821.1263837-2-jerry.hoemann@hpe.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: wdat_wdt: Set the min and max timeout values properlyJean Delvare
The wdat_wdt driver is misusing the min_hw_heartbeat_ms field. This field should only be used when the hardware watchdog device should not be pinged more frequently than a specific period. The ACPI WDAT "Minimum Count" field, on the other hand, specifies the minimum timeout value that can be set. This corresponds to the min_timeout field in Linux's watchdog infrastructure. Setting min_hw_heartbeat_ms instead can cause pings to the hardware to be delayed when there is no reason for that, eventually leading to unexpected firing of the watchdog timer (and thus unexpected reboot). Since commit 6d72c7ac9fbe ("watchdog: wdat_wdt: Using the existing function to check parameter timeout"), min_timeout is being set too, but to the arbitrary value of 1 second, which doesn't make sense and allows setting timeout values lower that the ACPI WDAT "Minimum Count" field. I'm also changing max_hw_heartbeat_ms to max_timeout for symmetry, although the use of this one isn't fundamentally wrong, but there is also no reason to enable the software-driven ping mechanism for the wdat_wdt driver. Signed-off-by: Jean Delvare <jdelvare@suse.de> Fixes: 058dfc767008 ("ACPI / watchdog: Add support for WDAT hardware watchdog") Fixes: 6d72c7ac9fbe ("watchdog: wdat_wdt: Using the existing function to check parameter timeout") Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Cc: Wim Van Sebroeck <wim@linux-watchdog.org> Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Cc: Liu Xinpeng <liuxp11@chinatelecom.cn> Link: https://lore.kernel.org/r/20220823154713.023ee771@endymion.delvare Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-02watchdog: bd9576_wdt: switch to using devm_fwnode_gpiod_get()Dmitry Torokhov
I would like to stop exporting OF-specific devm_gpiod_get_from_of_node() so that gpiolib can be cleaned a bit, so let's switch to the generic fwnode property API. While at it, switch the rest of the calls to read properties in bd9576_wdt_probe() to the generic device property API as well. Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Link: https://lore.kernel.org/r/20220903-gpiod_get_from_of_node-remove-v1-10-b29adfb27a6c@gmail.com Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
2022-10-01Merge branch 'for-6.1/nvdimm' into libnvdimm-for-nextDan Williams
Add v6.1 content on top of some straggling updates that missed v6.0.
2022-10-02i2c: pci1xxxx: prevent signed integer overflowWolfram Sang
Some constants need 'UL' markings, otherwise they are shifted into the sign bit. Fixes: 361693697249 ("i2c: microchip: pci1xxxx: Add driver for I2C host controller in multifunction endpoint of pci1xxxx switch") Signed-off-by: Wolfram Sang <wsa@kernel.org>
2022-10-02i2c: acpi: Replace zero-length array with DECLARE_FLEX_ARRAY() helperGustavo A. R. Silva
Zero-length arrays are deprecated and we are moving towards adopting C99 flexible-array members, instead. So, replace zero-length arrays declarations in anonymous union with the new DECLARE_FLEX_ARRAY() helper macro. This helper allows for flexible-array members in unions. Link: https://github.com/KSPP/linux/issues/193 Link: https://github.com/KSPP/linux/issues/218 Link: https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Wolfram Sang <wsa@kernel.org>
2022-10-02i2c: i801: Prefer async probeMani Milani
This i801 driver probe can take more than ~190ms in some devices, since the "i2c_register_spd()" call was added inside "i801_probe_optional_slaves()". Prefer async probe so that other drivers can be probed and boot can continue in parallel while this driver loads, to reduce boot time. There is no reason to block other drivers from probing while this driver is loading. Signed-off-by: Mani Milani <mani@chromium.org> Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com> Reviewed-by: Jean Delvare <jdelvare@suse.de> Signed-off-by: Wolfram Sang <wsa@kernel.org>
2022-10-02i2c: davinci: fix PM disable depth imbalance in davinci_i2c_probeZhang Qilong
The pm_runtime_enable will increase power disable depth. Thus a pairing decrement is needed on the error handling path to keep it balanced according to context. Fixes: 17f88151ff190 ("i2c: davinci: Add PM Runtime Support") Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com> Reviewed-by: Bartosz Golaszewski <brgl@bgdev.pl> Signed-off-by: Wolfram Sang <wsa@kernel.org>
2022-10-02i2c: designware-pci: Use standard pattern for memory allocationAndy Shevchenko
The pattern foo = kmalloc(sizeof(*foo), GFP_KERNEL); has an advantage when foo type is changed. Since we are planning a such, better to be prepared by using standard pattern for memory allocation. Signed-off-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>
2022-10-02i2c: designware-pci: Group AMD NAVI quirk parts togetherAndy Shevchenko
The code is ogranized in a way that all related parts to the certain platform quirk go together. This is not the case for AMD NAVI. Shuffle code to make it happen. While at it, drop the frequency definition and use hard coded value as it's done for other platforms and add a comment to the PCI ID list. Signed-off-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>
2022-10-01random: fix typos in get_random_bytes() commentWilliam Zijl
Remove extra whitespace and add a missing word to a sentence describing get_random_bytes(). Signed-off-by: William Zijl <postmaster@gusted.xyz> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
2022-10-01net/mlx5e: xsk: Use queue indices starting from 0 for XSK queuesMaxim Mikityanskiy
In the initial implementation of XSK in mlx5e, XSK RQs coexisted with regular RQs in the same channel. The main idea was to allow RSS work the same for regular traffic, without need to reconfigure RSS to exclude XSK queues. However, this scheme didn't prove to be beneficial, mainly because of incompatibility with other vendors. Some tools don't properly support using higher indices for XSK queues, some tools get confused with the double amount of RQs exposed in sysfs. Some use cases are purely XSK, and allocating the same amount of unused regular RQs is a waste of resources. This commit changes the queuing scheme to the standard one, where XSK RQs replace regular RQs on the channels where XSK sockets are open. Two RQs still exist in the channel to allow failsafe disable of XSK, but only one is exposed at a time. The next commit will achieve the desired memory save by flushing the buffers when the regular RQ is unused. As the result of this transition: 1. It's possible to use RSS contexts over XSK RQs. 2. It's possible to dedicate all queues to XSK. 3. When XSK RQs coexist with regular RQs, the admin should make sure no unwanted traffic goes into XSK RQs by either excluding them from RSS or settings up the XDP program to return XDP_PASS for non-XSK traffic. 4. When using a mixed fleet of mlx5e devices and other netdevs, the same configuration can be applied. If the application supports the fallback to copy mode on unsupported drivers, it will work too. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Introduce the mlx5e_flush_rq functionMaxim Mikityanskiy
Add a function to flush an RQ: clean up descriptors, release pages and reset the RQ. This procedure is used by the recovery flow, and it will also be used in a following commit to free some memory when switching a channel to the XSK mode. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Support XDP metadata on XSK RQsMaxim Mikityanskiy
Add support for XDP metadata on XSK RQs for cross-program communication. The driver no longer calls xdp_set_data_meta_invalid and copies the metadata to a newly allocated SKB on XDP_PASS. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Optimize RQ page deallocationMaxim Mikityanskiy
mlx5e_free_rx_mpwqe loops over all pages of a MPWQE, calling mlx5e_page_release for ones that are not scheduled for XDP_TX or XDP_REDIRECT; and mlx5e_page_release checks whether it's an XSK RQ or a regular one for each page/XSK frame. This check can be moved outside the loop to reduce the number of branches. mlx5e_free_rx_wqe loops over all fragments, calling mlx5e_page_release for the ones that are last in a page; and mlx5e_page_release checks whether it's an XSK RQ or a regular one for each fragment. Using the fact that XSK doesn't support multiple fragments, it can be optimized for both XSK and regular usages: 1. Make an early check for XSK and call its deallocator directly, saving 3 branches (loop condition, frag->last_in_page and selection of deallocator). 2. Call the regular deallocator directly in the non-XSK case, saving a branch per fragment, except the first one. After the changes, mlx5e_page_release is removed, as there are no callers left. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Call mlx5e_page_release_dynamic directly where possibleMaxim Mikityanskiy
mlx5e_page_release calls the appropriate deallocator depending on whether it's an XSK RQ or a regular one. Some flows that call this function are not compatible with XSK, so they can call the non-XSK deallocator directly to save a branch. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Use non-XSK page allocator in SHAMPOMaxim Mikityanskiy
The SHAMPO flow is not compatible with XSK, it can call the page pool allocator directly to save a branch. mlx5e_page_alloc is removed, as it's no longer used in any flow. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Use xsk_buff_alloc_batch on striding RQMaxim Mikityanskiy
XSK provides a function to allocate frames in batches for more efficient processing. This commit starts using this function on striding RQ and creates an optimized flow for XSK. A side effect is an opportunity to optimize the regular RX flow by dropping branching for XSK cases. Performance improvement is up to 6.4% in the aligned mode and up to 7.5% in the unaligned mode. Aligned mode, 2048-byte frames: 12.9 Mpps -> 13.8 Mpps Aligned mode, 4096-byte frames: 11.8 Mpps -> 12.5 Mpps Unaligned mode, 2048-byte frames: 11.9 Mpps -> 12.8 Mpps Unaligned mode, 3072-byte frames: 11.4 Mpps -> 12.1 Mpps Unaligned mode, 4096-byte frames: 11.0 Mpps -> 11.2 Mpps CPU: Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Use xsk_buff_alloc_batch on legacy RQMaxim Mikityanskiy
XSK provides a function to allocate frames in batches for more efficient processing. This commit starts using this function on legacy RQ, adding a special case for XSK. The new branch introduced basically replaces the branch that was removed from the same place a few commits before. A check is made that DMA sync is not needed, because the batching allocator falls back to returning one frame when DMA sync is needed, and this is best handled by the loop in the standard case. Performance improvement is up to 8% in the aligned mode and up to 9% in the unaligned mode. Aligned mode, 2048-byte frames: 12.8 Mpps -> 13.5 Mpps Aligned mode, 4096-byte frames: 11.5 Mpps -> 12.4 Mpps Unaligned mode, 2048-byte frames: 12.2 Mpps -> 13.4 Mpps Unaligned mode, 3072-byte frames: 11.6 Mpps -> 12.5 Mpps Unaligned mode, 4096-byte frames: 11.2 Mpps -> 12.2 Mpps CPU: Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Split out WQE allocation for legacy XSK RQMaxim Mikityanskiy
Allocation of XSK frames on legacy RQ may be made more efficient with a specialized routine that relies on certain assumptions, such as there is only one fragment, allocation units (XSK frames) are not shared among multiple packets. It reduces the number of branches both in the XSK code and in the regular RQ, because with this approach there is only a single check whether it's an XSK or regular RQ. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Remove the outer loop when allocating legacy RQ WQEsMaxim Mikityanskiy
Legacy RQ WQEs are allocated in a loop in small batches (8 WQEs). As partial batches are allowed, there is no point to have a loop in a loop, so the outer loop is removed, and the batch size is increased up to the total number of WQEs to allocate, still not smaller than 8. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Use partial batches in legacy RQ with XSKMaxim Mikityanskiy
The previous commit allowed allocating WQE batches in legacy RQ partially, however, XSK still checks whether there are enough frames in the fill ring. Remove this check to allow to allocate batches partially also with XSK. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Use partial batches in legacy RQMaxim Mikityanskiy
Legacy RQ allocates WQEs in batches. If the batch allocation fails, the pages of the allocated part are released. This commit changes this behavior to allow to use the pages that have been already allocated. After this change, we need to be careful about indexing rq->wqe.frags[]. The WQ size is a power of two that divides by wqe_bulk (8), and the old code used whole bulks, which allowed to use indices [8*K; 8*K+7] without overflowing. Now that the bulks may be partial, the range can start at any location (not only at 8*K), so we need to wrap them around to avoid out-of-bounds array access. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Make the wqe_index_mask calculation more exactMaxim Mikityanskiy
The old calculation of wqe_index_mask may give false positives, i.e. request bulking of pairs of WQEs when not strictly needed, for example, when the first fragment size is equal to the PAGE_SIZE, bulking is not needed, even if the number of fragments is odd. Make the calculation more exact to cut false positives. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: Introduce wqe_index_mask for legacy RQMaxim Mikityanskiy
When fragments of different WQEs share the same page, mlx5e_post_rx_wqes must wait until the old WQE stops using the page, only then the new WQE can allocate the new page. Essentially, it means that if WQE index i is still in use, the allocation must stop before `i % bulk`, where bulk is the number of WQEs that may share the same page. As bulk is always a power of two, `i % bulk = i & (bulk - 1)`, and the new wqe_index_mask field will be equal to `bulk - 1`. At the same time, wqe_bulk remains for optimization purposes and stores `max(bulk, 8)`, which allows to skip the allocation until we have at least 8 WQEs free. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Drop the check for XSK state in mlx5e_xsk_wakeupMaxim Mikityanskiy
The MLX5E_CHANNEL_STATE_XSK flag checked in mlx5e_xsk_wakeup indicates that XSK queues are open, but not necessarily activated. This check is not very useful, because: 0. Both XSK setup and netdev state transitions take the same state_lock mutex, so they can't happen at the same time. 1. If the netdev is up, xsk_is_bound can return true only when MLX5E_CHANNEL_STATE_XSK is set on the corresponding channel. mlx5e_xsk_wakeup is only called when xsk_is_bound is true. 2. If the XSK socket is bound, and the netdev is going up or down, mlx5e_xsk_wakeup can take one of two branches, depending on the return value of napi_if_scheduled_mark_missed: 2.1. True means one of two things: either NAPI was enabled at this point, which means MLX5E_CHANNEL_STATE_XSK was also set; or NAPI was disabled, and nothing really happened. 2.2. False means that NAPI was enabled by this point, which also implies MLX5E_CHANNEL_STATE_XSK was set. Additionally, mlx5e_xsk_wakeup contains a following check for MLX5E_SQ_STATE_ENABLED on async_icosq, and this flag implies MLX5E_CHANNEL_STATE_XSK too on XSK channels. As checking this flag doesn't cut any flows, remove the check. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2022-10-01net/mlx5e: xsk: Use mlx5e_trigger_napi_icosq for XSK wakeupMaxim Mikityanskiy
mlx5e_xsk_wakeup triggers an IRQ by posting a NOP to async_icosq, taking a spinlock to protect from concurrent access. There is already a function that does the same: mlx5e_trigger_napi_icosq. Use this function in mlx5e_xsk_wakeup. Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
generated by cgit (git 2.34.1) at 2025-09-13 10:44:38 +0000