linux_dsm_epyc7002/include/soc/arc/mcip.h

123 lines
2.8 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: GPL-2.0-only */
/*
* ARConnect IP Support (Multi core enabler: Cross core IPI, RTC ...)
*
* Copyright (C) 2014-15 Synopsys, Inc. (www.synopsys.com)
*/
#ifndef __SOC_ARC_MCIP_H
#define __SOC_ARC_MCIP_H
#include <soc/arc/aux.h>
#define ARC_REG_MCIP_BCR 0x0d0
#define ARC_REG_MCIP_IDU_BCR 0x0D5
ARC: mcip: halt GFRC counter when ARC cores halt In SMP systems, GFRC is used for clocksource. However by default the counter keeps running even when core is halted (say when debugging via a JTAG debugger). This confuses Linux timekeeping and triggers flase RCU stall splat such as below: | [ARCLinux]# while true; do ./shm_open_23-1.run-test ; done | Running with 1000 processes for 1000 objects | hrtimer: interrupt took 485060 ns | | create_cnt: 1000 | Running with 1000 processes for 1000 objects | [ARCLinux]# INFO: rcu_preempt self-detected stall on CPU | 2-...: (1 GPs behind) idle=a01/1/0 softirq=135770/135773 fqs=0 | INFO: rcu_preempt detected stalls on CPUs/tasks: | 0-...: (1 GPs behind) idle=71e/0/0 softirq=135264/135264 fqs=0 | 2-...: (1 GPs behind) idle=a01/1/0 softirq=135770/135773 fqs=0 | 3-...: (1 GPs behind) idle=4e0/0/0 softirq=134304/134304 fqs=0 | (detected by 1, t=13648 jiffies, g=31493, c=31492, q=1) Starting from ARC HS v3.0 it's possible to tie GFRC to state of up-to 4 ARC cores with help of GFRC's CORE register where we set a mask for cores which state we need to rely on. We update cpu mask every time new cpu came online instead of using hardcoded one or using mask generated from "possible_cpus" as we want it set correctly even if we run kernel on HW which has fewer cores than expected (or we launch kernel via debugger and kick fever cores than HW has) Note that GFRC halts when all cores have halted and thus relies on programming of Inter-Core-dEbug register to halt all cores when one halts. Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com> Signed-off-by: Vineet Gupta <vgupta@synopsys.com> [vgupta: rewrote changelog]
2018-02-23 23:41:52 +07:00
#define ARC_REG_GFRC_BUILD 0x0D6
#define ARC_REG_MCIP_CMD 0x600
#define ARC_REG_MCIP_WDATA 0x601
#define ARC_REG_MCIP_READBACK 0x602
struct mcip_cmd {
#ifdef CONFIG_CPU_BIG_ENDIAN
unsigned int pad:8, param:16, cmd:8;
#else
unsigned int cmd:8, param:16, pad:8;
#endif
#define CMD_INTRPT_GENERATE_IRQ 0x01
#define CMD_INTRPT_GENERATE_ACK 0x02
#define CMD_INTRPT_READ_STATUS 0x03
#define CMD_INTRPT_CHECK_SOURCE 0x04
/* Semaphore Commands */
#define CMD_SEMA_CLAIM_AND_READ 0x11
#define CMD_SEMA_RELEASE 0x12
#define CMD_DEBUG_SET_MASK 0x34
#define CMD_DEBUG_READ_MASK 0x35
#define CMD_DEBUG_SET_SELECT 0x36
#define CMD_DEBUG_READ_SELECT 0x37
#define CMD_GFRC_READ_LO 0x42
#define CMD_GFRC_READ_HI 0x43
ARC: mcip: halt GFRC counter when ARC cores halt In SMP systems, GFRC is used for clocksource. However by default the counter keeps running even when core is halted (say when debugging via a JTAG debugger). This confuses Linux timekeeping and triggers flase RCU stall splat such as below: | [ARCLinux]# while true; do ./shm_open_23-1.run-test ; done | Running with 1000 processes for 1000 objects | hrtimer: interrupt took 485060 ns | | create_cnt: 1000 | Running with 1000 processes for 1000 objects | [ARCLinux]# INFO: rcu_preempt self-detected stall on CPU | 2-...: (1 GPs behind) idle=a01/1/0 softirq=135770/135773 fqs=0 | INFO: rcu_preempt detected stalls on CPUs/tasks: | 0-...: (1 GPs behind) idle=71e/0/0 softirq=135264/135264 fqs=0 | 2-...: (1 GPs behind) idle=a01/1/0 softirq=135770/135773 fqs=0 | 3-...: (1 GPs behind) idle=4e0/0/0 softirq=134304/134304 fqs=0 | (detected by 1, t=13648 jiffies, g=31493, c=31492, q=1) Starting from ARC HS v3.0 it's possible to tie GFRC to state of up-to 4 ARC cores with help of GFRC's CORE register where we set a mask for cores which state we need to rely on. We update cpu mask every time new cpu came online instead of using hardcoded one or using mask generated from "possible_cpus" as we want it set correctly even if we run kernel on HW which has fewer cores than expected (or we launch kernel via debugger and kick fever cores than HW has) Note that GFRC halts when all cores have halted and thus relies on programming of Inter-Core-dEbug register to halt all cores when one halts. Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com> Signed-off-by: Vineet Gupta <vgupta@synopsys.com> [vgupta: rewrote changelog]
2018-02-23 23:41:52 +07:00
#define CMD_GFRC_SET_CORE 0x47
#define CMD_GFRC_READ_CORE 0x48
#define CMD_IDU_ENABLE 0x71
#define CMD_IDU_DISABLE 0x72
#define CMD_IDU_SET_MODE 0x74
#define CMD_IDU_SET_DEST 0x76
#define CMD_IDU_SET_MASK 0x7C
#define IDU_M_TRIG_LEVEL 0x0
#define IDU_M_TRIG_EDGE 0x1
#define IDU_M_DISTRI_RR 0x0
#define IDU_M_DISTRI_DEST 0x2
};
struct mcip_bcr {
#ifdef CONFIG_CPU_BIG_ENDIAN
unsigned int pad4:6, pw_dom:1, pad3:1,
idu:1, pad2:1, num_cores:6,
pad:1, gfrc:1, dbg:1, pw:1,
msg:1, sem:1, ipi:1, slv:1,
ver:8;
#else
unsigned int ver:8,
slv:1, ipi:1, sem:1, msg:1,
pw:1, dbg:1, gfrc:1, pad:1,
num_cores:6, pad2:1, idu:1,
pad3:1, pw_dom:1, pad4:6;
#endif
};
struct mcip_idu_bcr {
#ifdef CONFIG_CPU_BIG_ENDIAN
unsigned int pad:21, cirqnum:3, ver:8;
#else
unsigned int ver:8, cirqnum:3, pad:21;
#endif
};
/*
* Build register for IDU contains not an actual number of supported common
* interrupts but an exponent of 2 which must be multiplied by 4 to
* get a number of supported common interrupts.
*/
#define mcip_idu_bcr_to_nr_irqs(bcr) (4 * (1 << (bcr).cirqnum))
/*
* MCIP programming model
*
* - Simple commands write {cmd:8,param:16} to MCIP_CMD aux reg
* (param could be irq, common_irq, core_id ...)
* - More involved commands setup MCIP_WDATA with cmd specific data
* before invoking the simple command
*/
static inline void __mcip_cmd(unsigned int cmd, unsigned int param)
{
struct mcip_cmd buf;
buf.pad = 0;
buf.cmd = cmd;
buf.param = param;
WRITE_AUX(ARC_REG_MCIP_CMD, buf);
}
/*
* Setup additional data for a cmd
* Callers need to lock to ensure atomicity
*/
static inline void __mcip_cmd_data(unsigned int cmd, unsigned int param,
unsigned int data)
{
write_aux_reg(ARC_REG_MCIP_WDATA, data);
__mcip_cmd(cmd, param);
}
#endif