MIPS: Loongson: Invalidate special TLBs when needed

Loongson-2 has a 4 entry itlb which is a subset of jtlb, Loongson-3 has
a 4 entry itlb and a 4 entry dtlb which are subsets of jtlb. We should
write diag register to invalidate itlb/dtlb when flushing jtlb because
itlb/dtlb are not totally transparent to software.

For Loongson-3A R2 (and newer), we should invalidate ITLB, DTLB, VTLB
and FTLB before we enable/disable FTLB.

Signed-off-by: Huacai Chen <chenhc@lemote.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: Steven J . Hill <sjhill@realitydiluted.com>
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/12753/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

arch/mips/include/asm/mipsregs.h

@@ -780,6 +780,15 @@
 /* Disable Branch Return Cache */
 #define R10K_DIAG_D_BRC		(_ULCAST_(1) << 22)
 
+/* Flush ITLB */
+#define LOONGSON_DIAG_ITLB	(_ULCAST_(1) << 2)
+/* Flush DTLB */
+#define LOONGSON_DIAG_DTLB	(_ULCAST_(1) << 3)
+/* Flush VTLB */
+#define LOONGSON_DIAG_VTLB	(_ULCAST_(1) << 12)
+/* Flush FTLB */
+#define LOONGSON_DIAG_FTLB	(_ULCAST_(1) << 13)
+
 /*
  * Coprocessor 1 (FPU) register names
  */

arch/mips/kernel/cpu-probe.c

@@ -563,6 +563,9 @@ static int set_ftlb_enable(struct cpuinfo_mips *c, int enable)
 					   << MIPS_CONF7_FTLBP_SHIFT));
 		break;
 	case CPU_LOONGSON3:
+		/* Flush ITLB, DTLB, VTLB and FTLB */
+		write_c0_diag(LOONGSON_DIAG_ITLB | LOONGSON_DIAG_DTLB |
+			      LOONGSON_DIAG_VTLB | LOONGSON_DIAG_FTLB);
 		/* Loongson-3 cores use Config6 to enable the FTLB */
 		config = read_c0_config6();
 		if (enable)

arch/mips/mm/tlb-r4k.c

@@ -28,25 +28,28 @@
 extern void build_tlb_refill_handler(void);
 
 /*
- * LOONGSON2/3 has a 4 entry itlb which is a subset of dtlb,
- * unfortunately, itlb is not totally transparent to software.
+ * LOONGSON-2 has a 4 entry itlb which is a subset of jtlb, LOONGSON-3 has
+ * a 4 entry itlb and a 4 entry dtlb which are subsets of jtlb. Unfortunately,
+ * itlb/dtlb are not totally transparent to software.
  */
-static inline void flush_itlb(void)
+static inline void flush_micro_tlb(void)
 {
 	switch (current_cpu_type()) {
 	case CPU_LOONGSON2:
+		write_c0_diag(LOONGSON_DIAG_ITLB);
+		break;
 	case CPU_LOONGSON3:
-		write_c0_diag(4);
+		write_c0_diag(LOONGSON_DIAG_ITLB | LOONGSON_DIAG_DTLB);
 		break;
 	default:
 		break;
 	}
 }
 
-static inline void flush_itlb_vm(struct vm_area_struct *vma)
+static inline void flush_micro_tlb_vm(struct vm_area_struct *vma)
 {
 	if (vma->vm_flags & VM_EXEC)
-		flush_itlb();
+		flush_micro_tlb();
 }
 
 void local_flush_tlb_all(void)
@@ -93,7 +96,7 @@ void local_flush_tlb_all(void)
 	tlbw_use_hazard();
 	write_c0_entryhi(old_ctx);
 	htw_start();
-	flush_itlb();
+	flush_micro_tlb();
 	local_irq_restore(flags);
 }
 EXPORT_SYMBOL(local_flush_tlb_all);
@@ -159,7 +162,7 @@ void local_flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
 		} else {
 			drop_mmu_context(mm, cpu);
 		}
-		flush_itlb();
+		flush_micro_tlb();
 		local_irq_restore(flags);
 	}
 }
@@ -205,7 +208,7 @@ void local_flush_tlb_kernel_range(unsigned long start, unsigned long end)
 	} else {
 		local_flush_tlb_all();
 	}
-	flush_itlb();
+	flush_micro_tlb();
 	local_irq_restore(flags);
 }
 
@@ -240,7 +243,7 @@ void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
 	finish:
 		write_c0_entryhi(oldpid);
 		htw_start();
-		flush_itlb_vm(vma);
+		flush_micro_tlb_vm(vma);
 		local_irq_restore(flags);
 	}
 }
@@ -274,7 +277,7 @@ void local_flush_tlb_one(unsigned long page)
 	}
 	write_c0_entryhi(oldpid);
 	htw_start();
-	flush_itlb();
+	flush_micro_tlb();
 	local_irq_restore(flags);
 }
 
@@ -357,7 +360,7 @@ void __update_tlb(struct vm_area_struct * vma, unsigned long address, pte_t pte)
 	}
 	tlbw_use_hazard();
 	htw_start();
-	flush_itlb_vm(vma);
+	flush_micro_tlb_vm(vma);
 	local_irq_restore(flags);
 }