summaryrefslogtreecommitdiff
path: root/drivers/devfreq
AgeCommit message (Collapse)Author
2025-07-14PM / devfreq: Add HiSilicon uncore frequency scaling driverJie Zhan
Add the HiSilicon uncore frequency scaling driver for Kunpeng SoCs based on the devfreq framework. The uncore domain contains shared computing resources, including system interconnects and L3 cache. The uncore frequency significantly impacts the system-wide performance as well as power consumption. This driver adds support for runtime management of uncore frequency from kernel and userspace. The main function includes setting and getting frequencies, changing frequency scaling policies, and querying the list of CPUs whose performance is significantly related to this uncore frequency domain, etc. The driver communicates with a platform controller through an ACPI PCC mailbox to take the actual actions of frequency scaling. Co-developed-by: Lifeng Zheng <zhenglifeng1@huawei.com> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com> Reviewed-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Jie Zhan <zhanjie9@hisilicon.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com> Link: https://patchwork.kernel.org/project/linux-pm/patch/20250623143401.4095045-3-zhanjie9@hisilicon.com/
2025-07-14PM / devfreq: Allow devfreq driver to add custom sysfs ABIsJie Zhan
Extend the devfreq_dev_profile to allow drivers optionally create device-specific sysfs ABIs together with other common devfreq ABIs under the devfreq device path. Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com> Reviewed-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Jie Zhan <zhanjie9@hisilicon.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com> Link: https://patchwork.kernel.org/project/linux-pm/patch/20250623143401.4095045-2-zhanjie9@hisilicon.com/
2025-07-14PM / devfreq: sun8i-a33-mbus: Simplify by using more devm functionsUwe Kleine-König
Use devm allocators for enabling the bus clock and clk_rate_exclusive_get(). This simplifies error handling and the remove callback. Reviewed-by: Chen-Yu Tsai <wens@csie.org> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com> Tested-by: Corentin LABBE <clabbe.montjoie@gmail.com> Link: https://patchwork.kernel.org/project/linux-pm/patch/20250513203908.205060-2-u.kleine-koenig@baylibre.com/ Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2025-07-14PM / devfreq: Fix a index typo in trans_statChanwoo Choi
Fixes: 4920ee6dcfaf ("PM / devfreq: Convert to use sysfs_emit_at() API") Signed-off-by: pls <pleasurefish@126.com> Link: https://patchwork.kernel.org/project/linux-pm/patch/20250515143100.17849-1-chanwoo@kernel.org/ Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2025-07-14PM / devfreq: Check governor before using governor->nameLifeng Zheng
Commit 96ffcdf239de ("PM / devfreq: Remove redundant governor_name from struct devfreq") removes governor_name and uses governor->name to replace it. But devfreq->governor may be NULL and directly using devfreq->governor->name may cause null pointer exception. Move the check of governor to before using governor->name. Fixes: 96ffcdf239de ("PM / devfreq: Remove redundant governor_name from struct devfreq") Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Link: https://lore.kernel.org/lkml/20250421030020.3108405-5-zhenglifeng1@huawei.com/ Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2025-07-14PM / devfreq: Remove redundant devfreq_get_freq_range() calling in ↵Lifeng Zheng
devfreq_add_device() The calling of devfreq_get_freq_range() in devfreq_add_device() is redundant because min_freq and max_freq are never used. Remove it. Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Link: https://lore.kernel.org/lkml/20250421030020.3108405-4-zhenglifeng1@huawei.com/ Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2025-07-14PM / devfreq: Limit max_freq with scaling_min_freqLifeng Zheng
Limit max_freq in devfreq_get_freq_range() with scaling_min_freq to avoid showing an unreachable freq when reading it. Use macro clamp to simplify code. Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Link: https://lore.kernel.org/lkml/20250421030020.3108405-3-zhenglifeng1@huawei.com/ Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2025-07-14PM / devfreq: governor: Replace sscanf() with kstrtoul() in set_freq_store()Lifeng Zheng
Replace sscanf() with kstrtoul() in set_freq_store() and check the result to avoid invalid input. Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Link: https://lore.kernel.org/lkml/20250421030020.3108405-2-zhenglifeng1@huawei.com/ Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2025-02-18PM / devfreq: rockchip-dfi: Switch to use hrtimer_setup()Nam Cao
hrtimer_setup() takes the callback function pointer as argument and initializes the timer completely. Replace hrtimer_init() and the open coded initialization of hrtimer::function with the new setup mechanism. Patch was created by using Coccinelle. Signed-off-by: Nam Cao <namcao@linutronix.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Link: https://lore.kernel.org/all/dfb35fb45eacdba7b8767052aa8c29ca157ac1b0.1738746821.git.namcao@linutronix.de
2025-01-13Merge tag 'devfreq-next-for-6.14' of ↵Rafael J. Wysocki
ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/chanwoo/linux Merge devfreq updates for 6.14 from Chanwoo Choi: "- Call of_node_put() only once in devfreq_event_get_edev_by_phandle() on devfreq-event.c - Remove unused function parameter of exynos_bus_parse_of() on exynos-bus.c" * tag 'devfreq-next-for-6.14' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/chanwoo/linux: PM / devfreq: exynos: remove unused function parameter PM / devfreq: event: Call of_node_put() only once in devfreq_event_get_edev_by_phandle()
2025-01-11PM / devfreq: exynos: remove unused function parameterJeongjun Park
exynos_bus_parse_of() still declares a parameter struct device_node that is not used yet. This parameter is unnecessary and should be removed. Signed-off-by: Jeongjun Park <aha310510@gmail.com>
2024-12-08PM / devfreq: event: Call of_node_put() only once in ↵Markus Elfring
devfreq_event_get_edev_by_phandle() An of_node_put(node) call was immediately used after a null pointer check for the local variable “edev” at the end of this function implementation. Thus call such a function only once instead directly before the check. This issue was transformed by using the Coccinelle software. Link: https://lore.kernel.org/lkml/0f103384-376c-41f0-a35c-8ad98327d6cb@web.de/ Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-12-01Get rid of 'remove_new' relic from platform driver structLinus Torvalds
The continual trickle of small conversion patches is grating on me, and is really not helping. Just get rid of the 'remove_new' member function, which is just an alias for the plain 'remove', and had a comment to that effect: /* * .remove_new() is a relic from a prototype conversion of .remove(). * New drivers are supposed to implement .remove(). Once all drivers are * converted to not use .remove_new any more, it will be dropped. */ This was just a tree-wide 'sed' script that replaced '.remove_new' with '.remove', with some care taken to turn a subsequent tab into two tabs to make things line up. I did do some minimal manual whitespace adjustment for places that used spaces to line things up. Then I just removed the old (sic) .remove_new member function, and this is the end result. No more unnecessary conversion noise. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2024-09-05PM / devfreq: imx-bus: Use of_property_present()Rob Herring (Arm)
Use of_property_present() to test for property presence rather than of_get_property(). This is part of a larger effort to remove callers of of_get_property() and similar functions. of_get_property() leaks the DT property data pointer which is a problem for dynamically allocated nodes which may be freed. Link: https://lore.kernel.org/lkml/20240731191312.1710417-11-robh@kernel.org/ Acked-by: Peng Fan <peng.fan@nxp.com> Signed-off-by: Rob Herring (Arm) <robh@kernel.org> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-09-05PM / devfreq: exynos: Use Use devm_clk_get_enabled() helpersAnand Moon
The devm_clk_get_enabled() helpers: - call devm_clk_get() - call clk_prepare_enable() and register what is needed in order to call clk_disable_unprepare() when needed, as a managed resource. This simplifies the code and avoids the calls to clk_disable_unprepare(). While at it, use dev_err_probe consistently, and use its return value to return the error code. Link: https://lore.kernel.org/lkml/20240510094034.12493-1-linux.amoon@gmail.com/ Signed-off-by: Anand Moon <linux.amoon@gmail.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-09-05PM/devfreq: governor: add missing MODULE_DESCRIPTION() macrosJeff Johnson
make allmodconfig && make W=1 C=1 reports: WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/devfreq/governor_simpleondemand.o WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/devfreq/governor_performance.o WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/devfreq/governor_powersave.o WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/devfreq/governor_userspace.o Add all missing invocations of the MODULE_DESCRIPTION() macro. Link: https://lore.kernel.org/lkml/20240605-md-drivers-devfreq-v1-1-d01ae91b907e@quicinc.com/ Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-05-09PM / devfreq: exynos: Use DEFINE_SIMPLE_DEV_PM_OPS for PM functionsAnand Moon
This macro has the advantage over SET_SYSTEM_SLEEP_PM_OPS that we don't have to care about when the functions are actually used. Also make use of pm_sleep_ptr() to discard all PM_SLEEP related stuff if CONFIG_PM_SLEEP isn't enabled. Link: https://lore.kernel.org/lkml/20240417044459.1908-2-linux.amoon@gmail.com/ Signed-off-by: Anand Moon <linux.amoon@gmail.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-05-08PM / devfreq: rk3399_dmc: Convert to platform remove callback returning voidUwe Kleine-König
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-05-08PM / devfreq: sun8i-a33-mbus: Convert to platform remove callback returning voidUwe Kleine-König
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-05-08PM / devfreq: mtk-cci: Convert to platform remove callback returning voidUwe Kleine-König
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-05-08PM / devfreq: exynos-ppmu: Convert to platform remove callback returning voidUwe Kleine-König
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2024-05-08PM / devfreq: exynos-nocp: Convert to platform remove callback returning voidUwe Kleine-König
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-12-19PM / devfreq: Synchronize devfreq_monitor_[start/stop]Mukesh Ojha
There is a chance if a frequent switch of the governor done in a loop result in timer list corruption where timer cancel being done from two place one from cancel_delayed_work_sync() and followed by expire_timers() can be seen from the traces[1]. while true do echo "simple_ondemand" > /sys/class/devfreq/1d84000.ufshc/governor echo "performance" > /sys/class/devfreq/1d84000.ufshc/governor done It looks to be issue with devfreq driver where device_monitor_[start/stop] need to synchronized so that delayed work should get corrupted while it is either being queued or running or being cancelled. Let's use polling flag and devfreq lock to synchronize the queueing the timer instance twice and work data being corrupted. [1] ... .. <idle>-0 [003] 9436.209662: timer_cancel timer=0xffffff80444f0428 <idle>-0 [003] 9436.209664: timer_expire_entry timer=0xffffff80444f0428 now=0x10022da1c function=__typeid__ZTSFvP10timer_listE_global_addr baseclk=0x10022da1c <idle>-0 [003] 9436.209718: timer_expire_exit timer=0xffffff80444f0428 kworker/u16:6-14217 [003] 9436.209863: timer_start timer=0xffffff80444f0428 function=__typeid__ZTSFvP10timer_listE_global_addr expires=0x10022da2b now=0x10022da1c flags=182452227 vendor.xxxyyy.ha-1593 [004] 9436.209888: timer_cancel timer=0xffffff80444f0428 vendor.xxxyyy.ha-1593 [004] 9436.216390: timer_init timer=0xffffff80444f0428 vendor.xxxyyy.ha-1593 [004] 9436.216392: timer_start timer=0xffffff80444f0428 function=__typeid__ZTSFvP10timer_listE_global_addr expires=0x10022da2c now=0x10022da1d flags=186646532 vendor.xxxyyy.ha-1593 [005] 9436.220992: timer_cancel timer=0xffffff80444f0428 xxxyyyTraceManag-7795 [004] 9436.261641: timer_cancel timer=0xffffff80444f0428 [2] 9436.261653][ C4] Unable to handle kernel paging request at virtual address dead00000000012a [ 9436.261664][ C4] Mem abort info: [ 9436.261666][ C4] ESR = 0x96000044 [ 9436.261669][ C4] EC = 0x25: DABT (current EL), IL = 32 bits [ 9436.261671][ C4] SET = 0, FnV = 0 [ 9436.261673][ C4] EA = 0, S1PTW = 0 [ 9436.261675][ C4] Data abort info: [ 9436.261677][ C4] ISV = 0, ISS = 0x00000044 [ 9436.261680][ C4] CM = 0, WnR = 1 [ 9436.261682][ C4] [dead00000000012a] address between user and kernel address ranges [ 9436.261685][ C4] Internal error: Oops: 96000044 [#1] PREEMPT SMP [ 9436.261701][ C4] Skip md ftrace buffer dump for: 0x3a982d0 ... [ 9436.262138][ C4] CPU: 4 PID: 7795 Comm: TraceManag Tainted: G S W O 5.10.149-android12-9-o-g17f915d29d0c #1 [ 9436.262141][ C4] Hardware name: Qualcomm Technologies, Inc. (DT) [ 9436.262144][ C4] pstate: 22400085 (nzCv daIf +PAN -UAO +TCO BTYPE=--) [ 9436.262161][ C4] pc : expire_timers+0x9c/0x438 [ 9436.262164][ C4] lr : expire_timers+0x2a4/0x438 [ 9436.262168][ C4] sp : ffffffc010023dd0 [ 9436.262171][ C4] x29: ffffffc010023df0 x28: ffffffd0636fdc18 [ 9436.262178][ C4] x27: ffffffd063569dd0 x26: ffffffd063536008 [ 9436.262182][ C4] x25: 0000000000000001 x24: ffffff88f7c69280 [ 9436.262185][ C4] x23: 00000000000000e0 x22: dead000000000122 [ 9436.262188][ C4] x21: 000000010022da29 x20: ffffff8af72b4e80 [ 9436.262191][ C4] x19: ffffffc010023e50 x18: ffffffc010025038 [ 9436.262195][ C4] x17: 0000000000000240 x16: 0000000000000201 [ 9436.262199][ C4] x15: ffffffffffffffff x14: ffffff889f3c3100 [ 9436.262203][ C4] x13: ffffff889f3c3100 x12: 00000000049f56b8 [ 9436.262207][ C4] x11: 00000000049f56b8 x10: 00000000ffffffff [ 9436.262212][ C4] x9 : ffffffc010023e50 x8 : dead000000000122 [ 9436.262216][ C4] x7 : ffffffffffffffff x6 : ffffffc0100239d8 [ 9436.262220][ C4] x5 : 0000000000000000 x4 : 0000000000000101 [ 9436.262223][ C4] x3 : 0000000000000080 x2 : ffffff889edc155c [ 9436.262227][ C4] x1 : ffffff8001005200 x0 : ffffff80444f0428 [ 9436.262232][ C4] Call trace: [ 9436.262236][ C4] expire_timers+0x9c/0x438 [ 9436.262240][ C4] __run_timers+0x1f0/0x330 [ 9436.262245][ C4] run_timer_softirq+0x28/0x58 [ 9436.262255][ C4] efi_header_end+0x168/0x5ec [ 9436.262265][ C4] __irq_exit_rcu+0x108/0x124 [ 9436.262274][ C4] __handle_domain_irq+0x118/0x1e4 [ 9436.262282][ C4] gic_handle_irq.30369+0x6c/0x2bc [ 9436.262286][ C4] el0_irq_naked+0x60/0x6c Link: https://lore.kernel.org/all/1700860318-4025-1-git-send-email-quic_mojha@quicinc.com/ Reported-by: Joyyoung Huang <huangzaiyang@oppo.com> Acked-by: MyungJoo Ham <myungjoo.ham@samsung.com> Signed-off-by: Mukesh Ojha <quic_mojha@quicinc.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-11-13PM / devfreq: Convert to use sysfs_emit_at() APIChristian Marangi
Follow the advice of the Documentation/filesystems/sysfs.rst and show() should only use sysfs_emit() or sysfs_emit_at() when formatting the value to be returned to user space. Link: https://lore.kernel.org/all/20231024183016.14648-3-ansuelsmth@gmail.com/ Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-11-13PM / devfreq: Fix buffer overflow in trans_stat_showChristian Marangi
Fix buffer overflow in trans_stat_show(). Convert simple snprintf to the more secure scnprintf with size of PAGE_SIZE. Add condition checking if we are exceeding PAGE_SIZE and exit early from loop. Also add at the end a warning that we exceeded PAGE_SIZE and that stats is disabled. Return -EFBIG in the case where we don't have enough space to write the full transition table. Also document in the ABI that this function can return -EFBIG error. Link: https://lore.kernel.org/all/20231024183016.14648-2-ansuelsmth@gmail.com/ Cc: stable@vger.kernel.org Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218041 Fixes: e552bbaf5b98 ("PM / devfreq: Add sysfs node for representing frequency transition information.") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: add support for RK3588Sascha Hauer
Add support for the RK3588 to the driver. The RK3588 has four DDR channels with a register stride of 0x4000 between the channel registers, also it has a DDRMON_CTRL register per channel. Link: https://lore.kernel.org/all/20231018061714.3553817-20-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Acked-by: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: account for multiple DDRMON_CTRL registersSascha Hauer
The currently supported RK3399 has a set of registers per channel, but it has only a single DDRMON_CTRL register. With upcoming RK3588 this will be different, the RK3588 has a DDRMON_CTRL register per channel. Instead of expecting a single DDRMON_CTRL register, loop over the channels and write the channel specific DDRMON_CTRL register. Break out early out of the loop when there is only a single DDRMON_CTRL register like on the RK3399. Link: https://lore.kernel.org/all/20231018061714.3553817-19-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: make register stride SoC specificSascha Hauer
The currently supported RK3399 has a stride of 20 between the channel specific registers. Upcoming RK3588 has a different stride, so put the stride into driver data to make it configurable. While at it convert decimal 20 to hex 0x14 for consistency with RK3588 which has a register stride 0x4000 and we want to write that in hex as well. Link: https://lore.kernel.org/all/20231018061714.3553817-18-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Add perf supportSascha Hauer
The DFI is a unit which is suitable for measuring DDR utilization, but so far it could only be used as an event driver for the DDR frequency scaling driver. This adds perf support to the DFI driver. Usage with the 'perf' tool can look like: perf stat -a -e rockchip_ddr/cycles/,\ rockchip_ddr/read-bytes/,\ rockchip_ddr/write-bytes/,\ rockchip_ddr/bytes/ sleep 1 Performance counter stats for 'system wide': 1582524826 rockchip_ddr/cycles/ 1802.25 MB rockchip_ddr/read-bytes/ 1793.72 MB rockchip_ddr/write-bytes/ 3595.90 MB rockchip_ddr/bytes/ 1.014369709 seconds time elapsed perf support has been tested on a RK3568 and a RK3399, the latter with dual channel DDR. Link: https://lore.kernel.org/all/20231019064819.3496740-1-s.hauer@pengutronix.de/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Acked-by: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> [cw00.choi: Fix typo from 'write_acccess' to 'write_access'] Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: give variable a better nameSascha Hauer
struct dmc_count_channel::total counts the clock cycles of the DDR controller. Rename it accordingly to give the reader a better idea what this is about. While at it, at some documentation to struct dmc_count_channel. Link: https://lore.kernel.org/all/20231018061714.3553817-16-s.hauer@pengutronix.de/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Prepare for multiple usersSascha Hauer
When adding perf support later the DFI must be enabled when either of devfreq-event or perf is active. Prepare for that by adding a usage counter for the DFI. Also move enabling and disabling of the clock away from the devfreq-event specific functions to which the perf specific part won't have access. Link: https://lore.kernel.org/all/20231018061714.3553817-15-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Pass private data struct to internal functionsSascha Hauer
The internal functions do not need the struct devfreq_event_dev *, so pass them the struct rockchip_dfi *. This is a preparation for adding perf support later which doesn't have a struct devfreq_event_dev *. Link: https://lore.kernel.org/all/20231018061714.3553817-14-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Handle LPDDR4XSascha Hauer
In the DFI driver LPDDR4X can be handled in the same way as LPDDR4. Add the missing case. Link: https://lore.kernel.org/all/20231018061714.3553817-13-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Acked-by: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Handle LPDDR2 correctlySascha Hauer
According to the downstream driver the DDRMON_CTRL_LPDDR23 bit must be set for both LPDDR2 and LPDDR3. Add the missing LPDDR2 case and while at it turn the if/else if/else into switch/case which makes it easier to read. Link: https://lore.kernel.org/all/20231018061714.3553817-12-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Add RK3568 supportSascha Hauer
This adds RK3568 support to the DFI driver. Only iniitialization differs from the currently supported RK3399. Link: https://lore.kernel.org/all/20231018061714.3553817-11-s.hauer@pengutronix.de/ Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Acked-by: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rockchip-dfi: Clean up DDR type register definesSascha Hauer
Use the HIWORD_UPDATE() define known from other rockchip drivers to make the defines look less odd to the readers who've seen other rockchip drivers. The HIWORD registers have their functional bits in the lower 16 bits whereas the upper 16 bits contain a mask. Only the functional bits that have the corresponding mask bit set are modified during a write. Although the register writes look different, the end result should be the same, at least there's no functional change intended with this patch. Link: https://lore.kernel.org/all/20231018061714.3553817-10-s.hauer@pengutronix.de/ Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-19PM / devfreq: rk3399_dmc,dfi: generalize DDRTYPE definesSascha Hauer
The DDRTYPE defines are named to be RK3399 specific, but they can be used for other Rockchip SoCs as well, so replace the RK3399_PMUGRF_ prefix with ROCKCHIP_. They are defined in a SoC specific header file, so when generalizing the prefix also move the new defines to a SoC agnostic header file. While at it use GENMASK to define the DDRTYPE bitfield and give it a name including the full register name. Link: https://lore.kernel.org/all/20231018061714.3553817-9-s.hauer@pengutronix.de/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Acked-by: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-18PM / devfreq: rockchip-dfi: introduce channel maskSascha Hauer
Different Rockchip SoC variants have a different number of channels. Introduce a channel mask to make the number of channels configurable from SoC initialization code. Link: https://lore.kernel.org/all/20231018061714.3553817-8-s.hauer@pengutronix.de/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-18PM / devfreq: rockchip-dfi: Use free running counterSascha Hauer
The DDR_MON counters are free running counters. These are resetted to 0 when starting them over like currently done when reading the current counter values. Resetting the counters becomes a problem with perf support we want to add later, because perf needs counters that are not modified elsewhere. This patch removes resetting the counters and keeps them running instead. That means we no longer use the absolute counter values but instead compare them with the counter values we read last time. Not stopping the counters also has the impact that they are running while we are reading them. We cannot read multiple timers atomically, so the values do not exactly fit together. The effect should be negligible though as the time between two measurements is some orders of magnitude bigger than the time we need to read multiple registers. Link: https://lore.kernel.org/all/20231018061714.3553817-7-s.hauer@pengutronix.de/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-17PM / devfreq: mediatek: unlock on error in mtk_ccifreq_target()Dan Carpenter
Call mutex_unlock(&drv->reg_lock) before returning the error code. Link: https://lore.kernel.org/all/1bada9b2-d276-4123-bfdf-03d165569543@moroto.mountain/ Fixes: d2805601988f ("PM / devfreq: mediatek: protect oop in critical session") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: exynos-ppmu: Use device_get_match_data()Rob Herring
Use preferred device_get_match_data() instead of of_match_device() to get the driver match data. With this, adjust the includes to explicitly include the correct headers. Link: https://lore.kernel.org/all/20231006213854.333261-1-robh@kernel.org/ Signed-off-by: Rob Herring <robh@kernel.org> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: rockchip-dfi: dfi store raw values in counter structSascha Hauer
When adding perf support to the DFI driver the perf part will need the raw counter values, so move the fixed * 4 factor to rockchip_dfi_get_event(). Link: https://lore.kernel.org/lkml/20230704093242.583575-6-s.hauer@pengutronix.de/ Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: rockchip-dfi: Add SoC specific init functionSascha Hauer
Move the RK3399 specifics to a SoC specific init function to make the way free for supporting other SoCs later. Link: https://lore.kernel.org/lkml/20230704093242.583575-5-s.hauer@pengutronix.de/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: rockchip-dfi: use consistent name for private data structSascha Hauer
The variable name for the private data struct is 'info' in some functions and 'data' in others. Both names do not give a clue what type the variable has, so consistently use 'dfi'. Link: https://lore.kernel.org/lkml/20230704093242.583575-4-s.hauer@pengutronix.de/ Reviewed-by: Heiko Stuebner <heiko@sntech.de> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: rockchip-dfi: Embed desc into private data structSascha Hauer
No need for an extra allocation, just embed the struct devfreq_event_desc into the private data struct. Link: https://lore.kernel.org/lkml/20230704093242.583575-3-s.hauer@pengutronix.de/ Reviewed-by: Heiko Stuebner <heiko@sntech.de> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: rockchip-dfi: Make pmu regmap mandatorySascha Hauer
As a matter of fact the regmap_pmu already is mandatory because it is used unconditionally in the driver. Bail out gracefully in probe() rather than crashing later. Link: https://lore.kernel.org/lkml/20230704093242.583575-2-s.hauer@pengutronix.de/ Fixes: b9d1262bca0af ("PM / devfreq: event: support rockchip dfi controller") Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: Switch to dev_pm_opp_find_freq_{ceil/floor}_indexed() APIsManivannan Sadhasivam
Some devfreq consumers like UFS driver need to work with multiple clocks through the OPP framework. For this reason, OPP framework exposes the _indexed() APIs for finding the floor/ceil of the supplied frequency of the indexed clock. So let's use them in the devfreq driver. Currently, the clock index of 0 is used which works fine for multiple as well as single clock. Link: https://lore.kernel.org/all/20231003111232.42663-3-manivannan.sadhasivam@linaro.org/ Acked-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-10-09PM / devfreq: mediatek: protect oop in critical sessionMark Tseng
mtk_ccifreq_opp_notifier is reenter function when policy0 / policy6 change freq, so mutex_lock should protect all OPP event. Link: https://lore.kernel.org/lkml/20230818021741.6173-1-chun-jen.tseng@mediatek.com/T/ Signed-off-by: Mark Tseng <chun-jen.tseng@mediatek.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-08-10PM / devfreq: Fix leak in devfreq_dev_release()Boris Brezillon
srcu_init_notifier_head() allocates resources that need to be released with a srcu_cleanup_notifier_head() call. Reported by kmemleak. Fixes: 0fe3a66410a3 ("PM / devfreq: Add new DEVFREQ_TRANSITION_NOTIFIER notifier") Cc: <stable@vger.kernel.org> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com> Reviewed-by: Dhruva Gole <d-gole@ti.com> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2023-07-24PM / devfreq: Reword the kernel-doc comment for devfreq_monitor_start() APIManivannan Sadhasivam
Current kernel-doc comment doesn't specify the default timer used for the load monitoring. Also, it uses the term "default delayed work" which could be misunderstood as "default delayer timer". So reword the comment to clearly specify the default timer and also reword the last sentence to make it more understandable. Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
generated by cgit (git 2.34.1) at 2025-09-10 19:54:38 +0000