On 25/03/17 18:23, Leo Yan wrote: > Coresight includes debug module and usually the module connects with CPU > debug logic. ARMv8 architecture reference manual (ARM DDI 0487A.k) has > description for related info in "Part H: External Debug". > > Chapter H7 "The Sample-based Profiling Extension" introduces several > sampling registers, e.g. we can check program counter value with > combined CPU exception level, secure state, etc. So this is helpful for > analysis CPU lockup scenarios, e.g. if one CPU has run into infinite > loop with IRQ disabled. In this case the CPU cannot switch context and > handle any interrupt (including IPIs), as the result it cannot handle > SMP call for stack dump. > > This patch is to enable coresight debug module, so firstly this driver > is to bind apb clock for debug module and this is to ensure the debug > module can be accessed from program or external debugger. And the driver > uses sample-based registers for debug purpose, e.g. when system triggers > panic, the driver will dump program counter and combined context > registers (EDCIDSR, EDVIDSR); by parsing context registers so can > quickly get to know CPU secure state, exception level, etc. > > Some of the debug module registers are located in CPU power domain, so > this requires the CPU power domain stays on when access related debug > registers, but the power management for CPU power domain is quite > dependent on SoC integration for power management. For the platforms > which with sane power controller implementations, this driver follows > the method to set EDPRCR to try to pull the CPU out of low power state > and then set 'no power down request' bit so the CPU has no chance to > lose power. > > If the SoC has not followed up this design well for power management > controller, the driver introduces module parameter "idle_constraint". > Setting this parameter for latency requirement in microseconds, finally > we can constrain all or partial idle states to ensure the CPU power > domain is enabled, this is a backup method to access coresight CPU > debug component safely. > > Signed-off-by: Leo Yan <leo.yan@xxxxxxxxxx> > --- > drivers/hwtracing/coresight/Kconfig | 11 + > drivers/hwtracing/coresight/Makefile | 1 + > drivers/hwtracing/coresight/coresight-cpu-debug.c | 704 ++++++++++++++++++++++ > 3 files changed, 716 insertions(+) > create mode 100644 drivers/hwtracing/coresight/coresight-cpu-debug.c > > diff --git a/drivers/hwtracing/coresight/Kconfig b/drivers/hwtracing/coresight/Kconfig > index 130cb21..18d7931 100644 > --- a/drivers/hwtracing/coresight/Kconfig > +++ b/drivers/hwtracing/coresight/Kconfig > @@ -89,4 +89,15 @@ config CORESIGHT_STM > logging useful software events or data coming from various entities > in the system, possibly running different OSs > > +config CORESIGHT_CPU_DEBUG > + tristate "CoreSight CPU Debug driver" > + depends on ARM || ARM64 > + depends on DEBUG_FS > + help > + This driver provides support for coresight debugging module. This > + is primarily used to dump sample-based profiling registers when > + system triggers panic, the driver will parse context registers so > + can quickly get to know program counter (PC), secure state, > + exception level, etc. > + > endif > diff --git a/drivers/hwtracing/coresight/Makefile b/drivers/hwtracing/coresight/Makefile > index af480d9..433d590 100644 > --- a/drivers/hwtracing/coresight/Makefile > +++ b/drivers/hwtracing/coresight/Makefile > @@ -16,3 +16,4 @@ obj-$(CONFIG_CORESIGHT_SOURCE_ETM4X) += coresight-etm4x.o \ > coresight-etm4x-sysfs.o > obj-$(CONFIG_CORESIGHT_QCOM_REPLICATOR) += coresight-replicator-qcom.o > obj-$(CONFIG_CORESIGHT_STM) += coresight-stm.o > +obj-$(CONFIG_CORESIGHT_CPU_DEBUG) += coresight-cpu-debug.o > diff --git a/drivers/hwtracing/coresight/coresight-cpu-debug.c b/drivers/hwtracing/coresight/coresight-cpu-debug.c > new file mode 100644 > index 0000000..fbec1d1 > --- /dev/null > +++ b/drivers/hwtracing/coresight/coresight-cpu-debug.c > @@ -0,0 +1,704 @@ > +/* > + * Copyright (c) 2017 Linaro Limited. All rights reserved. > + * > + * Author: Leo Yan <leo.yan@xxxxxxxxxx> > + * > + * This program is free software; you can redistribute it and/or modify it > + * under the terms of the GNU General Public License version 2 as published by > + * the Free Software Foundation. > + * > + * This program is distributed in the hope that it will be useful, but WITHOUT > + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or > + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for > + * more details. > + * > + * You should have received a copy of the GNU General Public License along with > + * this program. If not, see <http://www.gnu.org/licenses/>. > + * > + */ > +#include <linux/amba/bus.h> > +#include <linux/coresight.h> > +#include <linux/cpu.h> > +#include <linux/debugfs.h> > +#include <linux/delay.h> > +#include <linux/device.h> > +#include <linux/err.h> > +#include <linux/init.h> > +#include <linux/io.h> > +#include <linux/kernel.h> > +#include <linux/module.h> > +#include <linux/moduleparam.h> > +#include <linux/pm_qos.h> > +#include <linux/slab.h> > +#include <linux/smp.h> > +#include <linux/types.h> > +#include <linux/uaccess.h> > + > +#include "coresight-priv.h" > + > +#define EDPCSR 0x0A0 > +#define EDCIDSR 0x0A4 > +#define EDVIDSR 0x0A8 > +#define EDPCSR_HI 0x0AC > +#define EDOSLAR 0x300 > +#define EDPRCR 0x310 > +#define EDPRSR 0x314 > +#define EDDEVID1 0xFC4 > +#define EDDEVID 0xFC8 > + > +#define EDPCSR_PROHIBITED 0xFFFFFFFF > + > +/* bits definition for EDPCSR */ > +#ifndef CONFIG_64BIT > +#define EDPCSR_THUMB BIT(0) > +#define EDPCSR_ARM_INST_MASK GENMASK(31, 2) > +#define EDPCSR_THUMB_INST_MASK GENMASK(31, 1) > +#endif > + > +/* bits definition for EDPRCR */ > +#define EDPRCR_COREPURQ BIT(3) > +#define EDPRCR_CORENPDRQ BIT(0) > + > +/* bits definition for EDPRSR */ > +#define EDPRSR_DLK BIT(6) > +#define EDPRSR_PU BIT(0) > + > +/* bits definition for EDVIDSR */ > +#define EDVIDSR_NS BIT(31) > +#define EDVIDSR_E2 BIT(30) > +#define EDVIDSR_E3 BIT(29) > +#define EDVIDSR_HV BIT(28) > +#define EDVIDSR_VMID GENMASK(7, 0) > + > +/* > + * bits definition for EDDEVID1:PSCROffset > + * > + * NOTE: armv8 and armv7 have different definition for the register, > + * so consolidate the bits definition as below: > + * > + * 0b0000 - Sample offset applies based on the instruction state, we > + * rely on EDDEVID to check if EDPCSR is implemented or not > + * 0b0001 - No offset applies. > + * 0b0010 - No offset applies, but do not use in AArch32 mode > + * > + */ > +#define EDDEVID1_PCSR_OFFSET_MASK GENMASK(3, 0) > +#define EDDEVID1_PCSR_OFFSET_INS_SET (0x0) > +#define EDDEVID1_PCSR_NO_OFFSET_DIS_AARCH32 (0x2) > + > +/* bits definition for EDDEVID */ > +#define EDDEVID_PCSAMPLE_MODE GENMASK(3, 0) > +#define EDDEVID_IMPL_NONE (0x0) > +#define EDDEVID_IMPL_EDPCSR (0x1) > +#define EDDEVID_IMPL_EDPCSR_EDCIDSR (0x2) > +#define EDDEVID_IMPL_FULL (0x3) > + > +#define DEBUG_WAIT_TIMEOUT 32 > + > +struct debug_drvdata { > + void __iomem *base; > + struct device *dev; > + int cpu; > + > + bool edpcsr_present; > + bool edcidsr_present; > + bool edvidsr_present; > + bool pc_has_offset; > + > + u32 eddevid; > + u32 eddevid1; > + > + u32 edpcsr; > + u32 edpcsr_hi; > + u32 edprcr; > + u32 edprsr; > + u32 edvidsr; > + u32 edcidsr; > +}; > + > +static DEFINE_MUTEX(debug_lock); > +static DEFINE_PER_CPU(struct debug_drvdata *, debug_drvdata); > +static int debug_count; > +static struct dentry *debug_debugfs_dir; > + > +static struct pm_qos_request debug_qos_req; > +static int idle_constraint = PM_QOS_DEFAULT_VALUE; > +module_param(idle_constraint, int, 0600); > +MODULE_PARM_DESC(idle_constraint, "Latency requirement in microseconds for CPU " > + "idle states (default is -1, which means have no limiation " > + "to CPU idle states; 0 means disabling all idle states; user " > + "can choose other platform dependent values so can disable " > + "specific idle states for the platform)"); > + NACK for this. Why you want the policy inside the driver. You can always do that from the user-space. I have mentioned it several times now. What can't you do these ? 1. echo "what_ever_latency_you_need_in_uS" > /dev/cpu_dma_latency 2. echo 1 > /sys/devices/system/cpu/cpu$cpu/cpuidle/state$state/disable (for all cpus and their states) (1) is definitely simpler way to disable deeper idle if latency = 0uS You can always warn user about that when it's enabled via debugfs/sysfs -- Regards, Sudeep -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
