Hi, Lothar Best Regards! Anson Huang > -----Original Message----- > From: Lothar Waßmann [mailto:LW@xxxxxxxxxxxxxxxxxxx] > Sent: 2018年11月28日 17:01 > To: Anson Huang <anson.huang@xxxxxxx> > Cc: a.zummo@xxxxxxxxxxxx; alexandre.belloni@xxxxxxxxxxx; > robh+dt@xxxxxxxxxx; mark.rutland@xxxxxxx; catalin.marinas@xxxxxxx; > will.deacon@xxxxxxx; Aisheng DONG <aisheng.dong@xxxxxxx>; Andy Gross > <andy.gross@xxxxxxxxxx>; arnd@xxxxxxxx; horms+renesas@xxxxxxxxxxxx; > heiko@xxxxxxxxx; bjorn.andersson@xxxxxxxxxx; enric.balletbo@xxxxxxxxxxxxx; > amit.kucheria@xxxxxxxxxx; ezequiel@xxxxxxxxxxxxx; olof@xxxxxxxxx; > shawnguo@xxxxxxxxxx; linux-rtc@xxxxxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx; > linux-kernel@xxxxxxxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; > dl-linux-imx <linux-imx@xxxxxxx> > Subject: Re: [PATCH V2 2/4] rtc: add i.MX system controller RTC support > > Hi, > > On Wed, 28 Nov 2018 05:21:11 +0000 Anson Huang wrote: > > i.MX8QXP is an ARMv8 SoC which has a Cortex-M4 system controller > > inside, the system controller is in charge of controlling power, clock > > and secure rtc etc.. > > > > This patch adds i.MX system controller RTC driver support, Linux > > kernel has to communicate with system controller via MU (message unit) > > IPC to set/get RTC time and other alarm functions, since the RTC set > > time needs to be done in secure EL3 mode (required by system > > controller firmware) and alarm functions needs to be done with general > > MU IRQ handle, these depend on other components which are NOT ready, > > so this patch ONLY enables the RTC time read. > > > > Signed-off-by: Anson Huang <Anson.Huang@xxxxxxx> > > --- > > changes since V1: > > - sort module in Makefile alphabetically; > > - remove boiler plate license text; > > - use devm_rtc_allocate_device and rtc_register_device to register the > RTC. > > drivers/rtc/Kconfig | 6 +++ > > drivers/rtc/Makefile | 1 + > > drivers/rtc/rtc-imx-sc.c | 107 > > +++++++++++++++++++++++++++++++++++++++++++++++ > > 3 files changed, 114 insertions(+) > > create mode 100644 drivers/rtc/rtc-imx-sc.c > > > > diff --git a/drivers/rtc/Kconfig b/drivers/rtc/Kconfig index > > a819ef0..3b9642e 100644 > > --- a/drivers/rtc/Kconfig > > +++ b/drivers/rtc/Kconfig > > @@ -1677,6 +1677,12 @@ config RTC_DRV_SNVS > > This driver can also be built as a module, if so, the module > > will be called "rtc-snvs". > > > > +config RTC_DRV_IMX_SC > > + tristate "NXP i.MX System Controller RTC support" > > + help > > + If you say yes here you get support for the NXP i.MX System > > + Controller RTC module. > > + > > config RTC_DRV_SIRFSOC > > tristate "SiRFSOC RTC" > > depends on ARCH_SIRF > > diff --git a/drivers/rtc/Makefile b/drivers/rtc/Makefile index > > 290c173..f97c05e 100644 > > --- a/drivers/rtc/Makefile > > +++ b/drivers/rtc/Makefile > > @@ -75,6 +75,7 @@ obj-$(CONFIG_RTC_DRV_GOLDFISH) += > rtc-goldfish.o > > obj-$(CONFIG_RTC_DRV_HID_SENSOR_TIME) += rtc-hid-sensor-time.o > > obj-$(CONFIG_RTC_DRV_HYM8563) += rtc-hym8563.o > > obj-$(CONFIG_RTC_DRV_IMXDI) += rtc-imxdi.o > > +obj-$(CONFIG_RTC_DRV_IMX_SC) += rtc-imx-sc.o > > obj-$(CONFIG_RTC_DRV_ISL12022) += rtc-isl12022.o > > obj-$(CONFIG_RTC_DRV_ISL12026) += rtc-isl12026.o > > obj-$(CONFIG_RTC_DRV_ISL1208) += rtc-isl1208.o > > diff --git a/drivers/rtc/rtc-imx-sc.c b/drivers/rtc/rtc-imx-sc.c new > > file mode 100644 index 0000000..7212e38 > > --- /dev/null > > +++ b/drivers/rtc/rtc-imx-sc.c > > @@ -0,0 +1,107 @@ > > +// SPDX-License-Identifier: GPL-2.0+ > > +/* > > + * Copyright 2018 NXP. > > + */ > > + > > +#include <linux/init.h> > > +#include <linux/io.h> > > +#include <linux/kernel.h> > > +#include <linux/firmware/imx/sci.h> > > +#include <linux/module.h> > > +#include <linux/of.h> > > +#include <linux/platform_device.h> > > +#include <linux/rtc.h> > > + > > +#define IMX_SC_TIMER_FUNC_GET_RTC_SEC1970 9 > > +#define IMX_SC_TIMER_FUNC_SET_RTC_TIME 6 > > + > > +struct imx_sc_ipc *rtc_ipc_handle; > > +struct rtc_device *rtc; > > + > Shouldn't these be static? > A generic name like 'rtc' as a global variable name is definitely a very bad idea. Agreed, will improve them in V3 patch set. > > > +struct imx_sc_msg_req_timer_get_rtc_time { > > + struct imx_sc_rpc_msg hdr; > > +} __packed; > > + > > +struct imx_sc_msg_resp_timer_get_rtc_time { > > + struct imx_sc_rpc_msg hdr; > > + u32 time; > > +} __packed; > > + > > +static int imx_sc_rtc_read_time(struct device *dev, struct rtc_time > > +*tm) { > > + struct imx_sc_msg_resp_timer_get_rtc_time *resp; > > + struct imx_sc_msg_req_timer_get_rtc_time msg; > > + struct imx_sc_rpc_msg *hdr = &msg.hdr; > > + int ret; > > + > > + hdr->ver = IMX_SC_RPC_VERSION; > > + hdr->svc = IMX_SC_RPC_SVC_TIMER; > > + hdr->func = IMX_SC_TIMER_FUNC_GET_RTC_SEC1970; > > + hdr->size = 1; > > + > > + ret = imx_scu_call_rpc(rtc_ipc_handle, &msg, true); > > + if (ret) { > > + pr_err("read rtc time failed, ret %d\n", ret); > > + return ret; > > + } > > + > > + resp = (struct imx_sc_msg_resp_timer_get_rtc_time *)&msg; > > + rtc_time_to_tm(resp->time, tm); > > + > > + return 0; > > +} > > + > > +static const struct rtc_class_ops imx_sc_rtc_ops = { > > + .read_time = imx_sc_rtc_read_time, > > +}; > > + > > +static int imx_sc_rtc_probe(struct platform_device *pdev) { > > + int ret; > > + > > + ret = imx_scu_get_handle(&rtc_ipc_handle); > > + if (ret) { > > + if (ret == -EPROBE_DEFER) > > + return ret; > > + > > + dev_err(&pdev->dev, "failed to get ipc handle: %d!\n", ret); > > + return ret; > > + } > > + > > + rtc = devm_rtc_allocate_device(&pdev->dev); > > + if (IS_ERR(rtc)) { > > + ret = PTR_ERR(rtc); > > + return ret; > > + } > > + > > + rtc->ops = &imx_sc_rtc_ops; > > + rtc->range_min = 0; > > + rtc->range_max = U32_MAX; > > + > > + ret = rtc_register_device(rtc); > > + if (ret) { > > + dev_err(&pdev->dev, "failed to register rtc: %d\n", ret); > > + return ret; > > + } > > + > > + return 0; > > +} > > + > > +static const struct of_device_id imx_sc_dt_ids[] = { > > + { .compatible = "nxp,imx8qxp-sc-rtc", }, > > + {}, > s/{},/{}/ > The trailing comma is usually added to make patches that add more entries to > the array simpler and help avoid patch conflicts. > Since the empty entry MUST be the LAST entry in this array the comma should > NOT be added here. Without the comma, adding new entries after the EOL > marker will generate a compile error. Thanks for review, will improve it in V3 patch set. Anson. > > > Lothar Waßmann
