Merge tag 'csky-for-linus-5.3-rc1' of git://github.com/c-sky/csky-linux

Pull arch/csky pupdates from Guo Ren:
 "This round of csky subsystem gives two features (ASID algorithm
  update, Perf pmu record support) and some fixups.

  ASID updates:
   - Revert mmu ASID mechanism
   - Add new asid lib code from arm
   - Use generic asid algorithm to implement switch_mm
   - Improve tlb operation with help of asid

  Perf pmu record support:
   - Init pmu as a device
   - Add count-width property for csky pmu
   - Add pmu interrupt support
   - Fix perf record in kernel/user space
   - dt-bindings: Add csky PMU bindings

  Fixes:
   - Fixup no panic in kernel for some traps
   - Fixup some error count in 810 & 860.
   - Fixup abiv1 memset error"

* tag 'csky-for-linus-5.3-rc1' of git://github.com/c-sky/csky-linux:
  csky: Fixup abiv1 memset error
  csky: Improve tlb operation with help of asid
  csky: Use generic asid algorithm to implement switch_mm
  csky: Add new asid lib code from arm
  csky: Revert mmu ASID mechanism
  dt-bindings: csky: Add csky PMU bindings
  dt-bindings: interrupt-controller: Update csky mpintc
  csky: Fixup some error count in 810 & 860.
  csky: Fix perf record in kernel/user space
  csky: Add pmu interrupt support
  csky: Add count-width property for csky pmu
  csky: Init pmu as a device
  csky: Fixup no panic in kernel for some traps
  csky: Select intc & timer drivers

Author: torvalds

Documentation/devicetree/bindings/csky/pmu.txt

@@ -0,0 +1,38 @@
+===============================
+C-SKY Performance Monitor Units
+===============================
+
+C-SKY Performance Monitor is designed for ck807/ck810/ck860 SMP soc and
+it could count cpu's events for helping analysis performance issues.
+
+============================
+PMU node bindings definition
+============================
+
+	Description: Describes PMU
+
+	PROPERTIES
+
+	- compatible
+		Usage: required
+		Value type: <string>
+		Definition: must be "csky,csky-pmu"
+	- interrupts
+		Usage: required
+		Value type: <u32 IRQ_TYPE_XXX>
+		Definition: must be pmu irq num defined by soc
+	- count-width
+		Usage: optional
+		Value type: <u32>
+		Definition: the width of pmu counter
+
+Examples:
+---------
+#include <dt-bindings/interrupt-controller/irq.h>
+
+	pmu: performace-monitor {
+		compatible = "csky,csky-pmu";
+		interrupts = <23 IRQ_TYPE_EDGE_RISING>;
+		interrupt-parent = <&intc>;
+		count-width = <48>;
+        };

arch/csky/Kconfig

@@ -10,6 +10,9 @@ config CSKY
 	select COMMON_CLK
 	select CLKSRC_MMIO
 	select CLKSRC_OF
+	select CSKY_MPINTC if CPU_CK860
+	select CSKY_MP_TIMER if CPU_CK860
+	select CSKY_APB_INTC
 	select DMA_DIRECT_REMAP
 	select IRQ_DOMAIN
 	select HANDLE_DOMAIN_IRQ
@@ -30,6 +33,7 @@ config CSKY
 	select GENERIC_IRQ_MULTI_HANDLER
 	select GENERIC_SCHED_CLOCK
 	select GENERIC_SMP_IDLE_THREAD
+	select GX6605S_TIMER if CPU_CK610
 	select HAVE_ARCH_TRACEHOOK
 	select HAVE_ARCH_AUDITSYSCALL
 	select HAVE_DYNAMIC_FTRACE

arch/csky/abiv1/Makefile

@@ -5,5 +5,4 @@ obj-y					+= bswapsi.o
 obj-y					+= cacheflush.o
 obj-y					+= mmap.o
 obj-y					+= memcpy.o
-obj-y					+= memset.o
 obj-y					+= strksyms.o

arch/csky/abiv1/inc/abi/ckmmu.h

@@ -78,6 +78,12 @@ static inline void tlb_invalid_all(void)
 	cpwcr("cpcr8", 0x04000000);
 }
 
+
+static inline void local_tlb_invalid_all(void)
+{
+	tlb_invalid_all();
+}
+
 static inline void tlb_invalid_indexed(void)
 {
 	cpwcr("cpcr8", 0x02000000);

arch/csky/abiv1/inc/abi/string.h

@@ -7,7 +7,4 @@
 #define __HAVE_ARCH_MEMCPY
 extern void *memcpy(void *, const void *, __kernel_size_t);
 
-#define __HAVE_ARCH_MEMSET
-extern void *memset(void *, int, __kernel_size_t);
-
 #endif /* __ABI_CSKY_STRING_H */

arch/csky/abiv1/memset.c

@@ -4,4 +4,3 @@
 #include <linux/module.h>
 
 EXPORT_SYMBOL(memcpy);
-EXPORT_SYMBOL(memset);

arch/csky/abiv1/strksyms.c

@@ -85,6 +85,16 @@ static inline void tlb_invalid_all(void)
 #endif
 }
 
+static inline void local_tlb_invalid_all(void)
+{
+#ifdef CONFIG_CPU_HAS_TLBI
+	asm volatile("tlbi.all\n":::"memory");
+	sync_is();
+#else
+	tlb_invalid_all();
+#endif
+}
+
 static inline void tlb_invalid_indexed(void)
 {
 	mtcr("cr<8, 15>", 0x02000000);

arch/csky/abiv2/inc/abi/ckmmu.h

@@ -0,0 +1,78 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __ASM_ASM_ASID_H
+#define __ASM_ASM_ASID_H
+
+#include <linux/atomic.h>
+#include <linux/compiler.h>
+#include <linux/cpumask.h>
+#include <linux/percpu.h>
+#include <linux/spinlock.h>
+
+struct asid_info
+{
+	atomic64_t	generation;
+	unsigned long	*map;
+	atomic64_t __percpu	*active;
+	u64 __percpu		*reserved;
+	u32			bits;
+	/* Lock protecting the structure */
+	raw_spinlock_t		lock;
+	/* Which CPU requires context flush on next call */
+	cpumask_t		flush_pending;
+	/* Number of ASID allocated by context (shift value) */
+	unsigned int		ctxt_shift;
+	/* Callback to locally flush the context. */
+	void			(*flush_cpu_ctxt_cb)(void);
+};
+
+#define NUM_ASIDS(info)			(1UL << ((info)->bits))
+#define NUM_CTXT_ASIDS(info)		(NUM_ASIDS(info) >> (info)->ctxt_shift)
+
+#define active_asid(info, cpu)	*per_cpu_ptr((info)->active, cpu)
+
+void asid_new_context(struct asid_info *info, atomic64_t *pasid,
+		      unsigned int cpu, struct mm_struct *mm);
+
+/*
+ * Check the ASID is still valid for the context. If not generate a new ASID.
+ *
+ * @pasid: Pointer to the current ASID batch
+ * @cpu: current CPU ID. Must have been acquired throught get_cpu()
+ */
+static inline void asid_check_context(struct asid_info *info,
+				      atomic64_t *pasid, unsigned int cpu,
+				      struct mm_struct *mm)
+{
+	u64 asid, old_active_asid;
+
+	asid = atomic64_read(pasid);
+
+	/*
+	 * The memory ordering here is subtle.
+	 * If our active_asid is non-zero and the ASID matches the current
+	 * generation, then we update the active_asid entry with a relaxed
+	 * cmpxchg. Racing with a concurrent rollover means that either:
+	 *
+	 * - We get a zero back from the cmpxchg and end up waiting on the
+	 *   lock. Taking the lock synchronises with the rollover and so
+	 *   we are forced to see the updated generation.
+	 *
+	 * - We get a valid ASID back from the cmpxchg, which means the
+	 *   relaxed xchg in flush_context will treat us as reserved
+	 *   because atomic RmWs are totally ordered for a given location.
+	 */
+	old_active_asid = atomic64_read(&active_asid(info, cpu));
+	if (old_active_asid &&
+	    !((asid ^ atomic64_read(&info->generation)) >> info->bits) &&
+	    atomic64_cmpxchg_relaxed(&active_asid(info, cpu),
+				     old_active_asid, asid))
+		return;
+
+	asid_new_context(info, pasid, cpu, mm);
+}
+
+int asid_allocator_init(struct asid_info *info,
+			u32 bits, unsigned int asid_per_ctxt,
+			void (*flush_cpu_ctxt_cb)(void));
+
+#endif

arch/csky/include/asm/asid.h

@@ -5,7 +5,7 @@
 #define __ASM_CSKY_MMU_H
 
 typedef struct {
-	unsigned long asid[NR_CPUS];
+	atomic64_t	asid;
 	void *vdso;
 } mm_context_t;