diff options
Diffstat (limited to 'arch/sh/kernel/cpu/sh4')
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/Makefile | 9 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/clock-sh4-202.c | 79 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/clock-sh4.c | 32 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/fpu.c | 180 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/perf_event.c | 268 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/probe.c | 174 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/setup-sh4-202.c | 124 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/setup-sh7750.c | 238 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/setup-sh7760.c | 201 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/softfloat.c | 104 | ||||
| -rw-r--r-- | arch/sh/kernel/cpu/sh4/sq.c | 54 |
11 files changed, 995 insertions, 468 deletions
diff --git a/arch/sh/kernel/cpu/sh4/Makefile b/arch/sh/kernel/cpu/sh4/Makefile index d608557c7a3..3a1dbc70983 100644 --- a/arch/sh/kernel/cpu/sh4/Makefile +++ b/arch/sh/kernel/cpu/sh4/Makefile @@ -5,9 +5,15 @@ obj-y := probe.o common.o common-y += $(addprefix ../sh3/, entry.o ex.o) +obj-$(CONFIG_HIBERNATION) += $(addprefix ../sh3/, swsusp.o) obj-$(CONFIG_SH_FPU) += fpu.o softfloat.o obj-$(CONFIG_SH_STORE_QUEUES) += sq.o +# Perf events +perf-$(CONFIG_CPU_SUBTYPE_SH7750) := perf_event.o +perf-$(CONFIG_CPU_SUBTYPE_SH7750S) := perf_event.o +perf-$(CONFIG_CPU_SUBTYPE_SH7091) := perf_event.o + # CPU subtype setup obj-$(CONFIG_CPU_SUBTYPE_SH7750) += setup-sh7750.o obj-$(CONFIG_CPU_SUBTYPE_SH7750R) += setup-sh7750.o @@ -26,4 +32,5 @@ endif # Additional clocks by subtype clock-$(CONFIG_CPU_SUBTYPE_SH4_202) += clock-sh4-202.o -obj-y += $(clock-y) +obj-y += $(clock-y) +obj-$(CONFIG_PERF_EVENTS) += $(perf-y) diff --git a/arch/sh/kernel/cpu/sh4/clock-sh4-202.c b/arch/sh/kernel/cpu/sh4/clock-sh4-202.c index a33429463e9..4b5bab5f875 100644 --- a/arch/sh/kernel/cpu/sh4/clock-sh4-202.c +++ b/arch/sh/kernel/cpu/sh4/clock-sh4-202.c @@ -12,19 +12,20 @@ #include <linux/init.h> #include <linux/kernel.h> #include <linux/err.h> +#include <linux/io.h> +#include <linux/clkdev.h> #include <asm/clock.h> #include <asm/freq.h> -#include <asm/io.h> #define CPG2_FRQCR3 0xfe0a0018 static int frqcr3_divisors[] = { 1, 2, 3, 4, 6, 8, 16 }; static int frqcr3_values[] = { 0, 1, 2, 3, 4, 5, 6 }; -static void emi_clk_recalc(struct clk *clk) +static unsigned long emi_clk_recalc(struct clk *clk) { - int idx = ctrl_inl(CPG2_FRQCR3) & 0x0007; - clk->rate = clk->parent->rate / frqcr3_divisors[idx]; + int idx = __raw_readl(CPG2_FRQCR3) & 0x0007; + return clk->parent->rate / frqcr3_divisors[idx]; } static inline int frqcr3_lookup(struct clk *clk, unsigned long rate) @@ -40,29 +41,27 @@ static inline int frqcr3_lookup(struct clk *clk, unsigned long rate) return 5; } -static struct clk_ops sh4202_emi_clk_ops = { +static struct sh_clk_ops sh4202_emi_clk_ops = { .recalc = emi_clk_recalc, }; static struct clk sh4202_emi_clk = { - .name = "emi_clk", - .flags = CLK_ALWAYS_ENABLED, + .flags = CLK_ENABLE_ON_INIT, .ops = &sh4202_emi_clk_ops, }; -static void femi_clk_recalc(struct clk *clk) +static unsigned long femi_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(CPG2_FRQCR3) >> 3) & 0x0007; - clk->rate = clk->parent->rate / frqcr3_divisors[idx]; + int idx = (__raw_readl(CPG2_FRQCR3) >> 3) & 0x0007; + return clk->parent->rate / frqcr3_divisors[idx]; } -static struct clk_ops sh4202_femi_clk_ops = { +static struct sh_clk_ops sh4202_femi_clk_ops = { .recalc = femi_clk_recalc, }; static struct clk sh4202_femi_clk = { - .name = "femi_clk", - .flags = CLK_ALWAYS_ENABLED, + .flags = CLK_ENABLE_ON_INIT, .ops = &sh4202_femi_clk_ops, }; @@ -82,18 +81,17 @@ static void shoc_clk_init(struct clk *clk) for (i = 0; i < ARRAY_SIZE(frqcr3_divisors); i++) { int divisor = frqcr3_divisors[i]; - if (clk->ops->set_rate(clk, clk->parent->rate / - divisor, 0) == 0) + if (clk->ops->set_rate(clk, clk->parent->rate / divisor) == 0) break; } WARN_ON(i == ARRAY_SIZE(frqcr3_divisors)); /* Undefined clock */ } -static void shoc_clk_recalc(struct clk *clk) +static unsigned long shoc_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(CPG2_FRQCR3) >> 6) & 0x0007; - clk->rate = clk->parent->rate / frqcr3_divisors[idx]; + int idx = (__raw_readl(CPG2_FRQCR3) >> 6) & 0x0007; + return clk->parent->rate / frqcr3_divisors[idx]; } static int shoc_clk_verify_rate(struct clk *clk, unsigned long rate) @@ -111,7 +109,7 @@ static int shoc_clk_verify_rate(struct clk *clk, unsigned long rate) return 0; } -static int shoc_clk_set_rate(struct clk *clk, unsigned long rate, int algo_id) +static int shoc_clk_set_rate(struct clk *clk, unsigned long rate) { unsigned long frqcr3; unsigned int tmp; @@ -122,25 +120,24 @@ static int shoc_clk_set_rate(struct clk *clk, unsigned long rate, int algo_id) tmp = frqcr3_lookup(clk, rate); - frqcr3 = ctrl_inl(CPG2_FRQCR3); + frqcr3 = __raw_readl(CPG2_FRQCR3); frqcr3 &= ~(0x0007 << 6); frqcr3 |= tmp << 6; - ctrl_outl(frqcr3, CPG2_FRQCR3); + __raw_writel(frqcr3, CPG2_FRQCR3); clk->rate = clk->parent->rate / frqcr3_divisors[tmp]; return 0; } -static struct clk_ops sh4202_shoc_clk_ops = { +static struct sh_clk_ops sh4202_shoc_clk_ops = { .init = shoc_clk_init, .recalc = shoc_clk_recalc, .set_rate = shoc_clk_set_rate, }; static struct clk sh4202_shoc_clk = { - .name = "shoc_clk", - .flags = CLK_ALWAYS_ENABLED, + .flags = CLK_ENABLE_ON_INIT, .ops = &sh4202_shoc_clk_ops, }; @@ -150,31 +147,31 @@ static struct clk *sh4202_onchip_clocks[] = { &sh4202_shoc_clk, }; -static int __init sh4202_clk_init(void) +static struct clk_lookup lookups[] = { + /* main clocks */ + CLKDEV_CON_ID("emi_clk", &sh4202_emi_clk), + CLKDEV_CON_ID("femi_clk", &sh4202_femi_clk), + CLKDEV_CON_ID("shoc_clk", &sh4202_shoc_clk), +}; + +int __init arch_clk_init(void) { - struct clk *clk = clk_get(NULL, "master_clk"); - int i; + struct clk *clk; + int i, ret = 0; + cpg_clk_init(); + + clk = clk_get(NULL, "master_clk"); for (i = 0; i < ARRAY_SIZE(sh4202_onchip_clocks); i++) { struct clk *clkp = sh4202_onchip_clocks[i]; clkp->parent = clk; - clk_register(clkp); - clk_enable(clkp); + ret |= clk_register(clkp); } - /* - * Now that we have the rest of the clocks registered, we need to - * force the parent clock to propagate so that these clocks will - * automatically figure out their rate. We cheat by handing the - * parent clock its current rate and forcing child propagation. - */ - clk_set_rate(clk, clk_get_rate(clk)); - clk_put(clk); - return 0; -} - -arch_initcall(sh4202_clk_init); + clkdev_add_table(lookups, ARRAY_SIZE(lookups)); + return ret; +} diff --git a/arch/sh/kernel/cpu/sh4/clock-sh4.c b/arch/sh/kernel/cpu/sh4/clock-sh4.c index dca9f87a12d..99e5ec8b483 100644 --- a/arch/sh/kernel/cpu/sh4/clock-sh4.c +++ b/arch/sh/kernel/cpu/sh4/clock-sh4.c @@ -28,51 +28,51 @@ static int pfc_divisors[] = { 2, 3, 4, 6, 8, 2, 2, 2 }; static void master_clk_init(struct clk *clk) { - clk->rate *= pfc_divisors[ctrl_inw(FRQCR) & 0x0007]; + clk->rate *= pfc_divisors[__raw_readw(FRQCR) & 0x0007]; } -static struct clk_ops sh4_master_clk_ops = { +static struct sh_clk_ops sh4_master_clk_ops = { .init = master_clk_init, }; -static void module_clk_recalc(struct clk *clk) +static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0007); - clk->rate = clk->parent->rate / pfc_divisors[idx]; + int idx = (__raw_readw(FRQCR) & 0x0007); + return clk->parent->rate / pfc_divisors[idx]; } -static struct clk_ops sh4_module_clk_ops = { +static struct sh_clk_ops sh4_module_clk_ops = { .recalc = module_clk_recalc, }; -static void bus_clk_recalc(struct clk *clk) +static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) >> 3) & 0x0007; - clk->rate = clk->parent->rate / bfc_divisors[idx]; + int idx = (__raw_readw(FRQCR) >> 3) & 0x0007; + return clk->parent->rate / bfc_divisors[idx]; } -static struct clk_ops sh4_bus_clk_ops = { +static struct sh_clk_ops sh4_bus_clk_ops = { .recalc = bus_clk_recalc, }; -static void cpu_clk_recalc(struct clk *clk) +static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) >> 6) & 0x0007; - clk->rate = clk->parent->rate / ifc_divisors[idx]; + int idx = (__raw_readw(FRQCR) >> 6) & 0x0007; + return clk->parent->rate / ifc_divisors[idx]; } -static struct clk_ops sh4_cpu_clk_ops = { +static struct sh_clk_ops sh4_cpu_clk_ops = { .recalc = cpu_clk_recalc, }; -static struct clk_ops *sh4_clk_ops[] = { +static struct sh_clk_ops *sh4_clk_ops[] = { &sh4_master_clk_ops, &sh4_module_clk_ops, &sh4_bus_clk_ops, &sh4_cpu_clk_ops, }; -void __init arch_init_clk_ops(struct clk_ops **ops, int idx) +void __init arch_init_clk_ops(struct sh_clk_ops **ops, int idx) { if (idx < ARRAY_SIZE(sh4_clk_ops)) *ops = sh4_clk_ops[idx]; diff --git a/arch/sh/kernel/cpu/sh4/fpu.c b/arch/sh/kernel/cpu/sh4/fpu.c index 8020796139f..69ab4d3c8d4 100644 --- a/arch/sh/kernel/cpu/sh4/fpu.c +++ b/arch/sh/kernel/cpu/sh4/fpu.c @@ -13,10 +13,10 @@ #include <linux/sched.h> #include <linux/signal.h> #include <linux/io.h> -#include <asm/cpu/fpu.h> +#include <cpu/fpu.h> #include <asm/processor.h> -#include <asm/system.h> #include <asm/fpu.h> +#include <asm/traps.h> /* The PR (precision) bit in the FP Status Register must be clear when * an frchg instruction is executed, otherwise the instruction is undefined. @@ -36,18 +36,16 @@ extern unsigned long int float32_add(unsigned long int a, unsigned long int b); extern unsigned long long float64_sub(unsigned long long a, unsigned long long b); extern unsigned long int float32_sub(unsigned long int a, unsigned long int b); - +extern unsigned long int float64_to_float32(unsigned long long a); static unsigned int fpu_exception_flags; /* * Save FPU registers onto task structure. - * Assume called with FPU enabled (SR.FD=0). */ -void save_fpu(struct task_struct *tsk, struct pt_regs *regs) +void save_fpu(struct task_struct *tsk) { unsigned long dummy; - clear_tsk_thread_flag(tsk, TIF_USEDFPU); enable_fpu(); asm volatile ("sts.l fpul, @-%0\n\t" "sts.l fpscr, @-%0\n\t" @@ -87,15 +85,14 @@ void save_fpu(struct task_struct *tsk, struct pt_regs *regs) "fmov.s fr1, @-%0\n\t" "fmov.s fr0, @-%0\n\t" "lds %3, fpscr\n\t":"=r" (dummy) - :"0"((char *)(&tsk->thread.fpu.hard.status)), + :"0"((char *)(&tsk->thread.xstate->hardfpu.status)), "r"(FPSCR_RCHG), "r"(FPSCR_INIT) :"memory"); disable_fpu(); - release_fpu(regs); } -static void restore_fpu(struct task_struct *tsk) +void restore_fpu(struct task_struct *tsk) { unsigned long dummy; @@ -138,62 +135,11 @@ static void restore_fpu(struct task_struct *tsk) "lds.l @%0+, fpscr\n\t" "lds.l @%0+, fpul\n\t" :"=r" (dummy) - :"0"(&tsk->thread.fpu), "r"(FPSCR_RCHG) + :"0" (tsk->thread.xstate), "r" (FPSCR_RCHG) :"memory"); disable_fpu(); } -/* - * Load the FPU with signalling NANS. This bit pattern we're using - * has the property that no matter wether considered as single or as - * double precision represents signaling NANS. - */ - -static void fpu_init(void) -{ - enable_fpu(); - asm volatile ( "lds %0, fpul\n\t" - "lds %1, fpscr\n\t" - "fsts fpul, fr0\n\t" - "fsts fpul, fr1\n\t" - "fsts fpul, fr2\n\t" - "fsts fpul, fr3\n\t" - "fsts fpul, fr4\n\t" - "fsts fpul, fr5\n\t" - "fsts fpul, fr6\n\t" - "fsts fpul, fr7\n\t" - "fsts fpul, fr8\n\t" - "fsts fpul, fr9\n\t" - "fsts fpul, fr10\n\t" - "fsts fpul, fr11\n\t" - "fsts fpul, fr12\n\t" - "fsts fpul, fr13\n\t" - "fsts fpul, fr14\n\t" - "fsts fpul, fr15\n\t" - "frchg\n\t" - "fsts fpul, fr0\n\t" - "fsts fpul, fr1\n\t" - "fsts fpul, fr2\n\t" - "fsts fpul, fr3\n\t" - "fsts fpul, fr4\n\t" - "fsts fpul, fr5\n\t" - "fsts fpul, fr6\n\t" - "fsts fpul, fr7\n\t" - "fsts fpul, fr8\n\t" - "fsts fpul, fr9\n\t" - "fsts fpul, fr10\n\t" - "fsts fpul, fr11\n\t" - "fsts fpul, fr12\n\t" - "fsts fpul, fr13\n\t" - "fsts fpul, fr14\n\t" - "fsts fpul, fr15\n\t" - "frchg\n\t" - "lds %2, fpscr\n\t" - : /* no output */ - :"r" (0), "r"(FPSCR_RCHG), "r"(FPSCR_INIT)); - disable_fpu(); -} - /** * denormal_to_double - Given denormalized float number, * store double float @@ -285,10 +231,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* fcnvsd */ struct task_struct *tsk = current; - save_fpu(tsk, regs); - if ((tsk->thread.fpu.hard.fpscr & FPSCR_CAUSE_ERROR)) + if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_CAUSE_ERROR)) /* FPU error */ - denormal_to_double(&tsk->thread.fpu.hard, + denormal_to_double(&tsk->thread.xstate->hardfpu, (finsn >> 8) & 0xf); else return 0; @@ -304,9 +249,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & FPSCR_DBL_PRECISION; if ((fpscr & FPSCR_CAUSE_ERROR) @@ -316,18 +261,18 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* FPU error because of denormal (doubles) */ llx = ((long long)hx << 32) - | tsk->thread.fpu.hard.fp_regs[n + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[n + 1]; lly = ((long long)hy << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; llx = float64_mul(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n + 1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_CAUSE_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { /* FPU error because of denormal (floats) */ hx = float32_mul(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -342,9 +287,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & FPSCR_DBL_PRECISION; if ((fpscr & FPSCR_CAUSE_ERROR) @@ -354,15 +299,15 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* FPU error because of denormal (doubles) */ llx = ((long long)hx << 32) - | tsk->thread.fpu.hard.fp_regs[n + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[n + 1]; lly = ((long long)hy << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; if ((finsn & 0xf00f) == 0xf000) llx = float64_add(llx, lly); else llx = float64_sub(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n + 1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_CAUSE_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { @@ -371,7 +316,7 @@ static int ieee_fpe_handler(struct pt_regs *regs) hx = float32_add(hx, hy); else hx = float32_sub(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -386,9 +331,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & FPSCR_DBL_PRECISION; if ((fpscr & FPSCR_CAUSE_ERROR) @@ -398,20 +343,43 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* FPU error because of denormal (doubles) */ llx = ((long long)hx << 32) - | tsk->thread.fpu.hard.fp_regs[n + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[n + 1]; lly = ((long long)hy << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; llx = float64_div(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n + 1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_CAUSE_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { /* FPU error because of denormal (floats) */ hx = float32_div(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; + } else + return 0; + + regs->pc = nextpc; + return 1; + } else if ((finsn & 0xf0bd) == 0xf0bd) { + /* fcnvds - double to single precision convert */ + struct task_struct *tsk = current; + int m; + unsigned int hx; + + m = (finsn >> 8) & 0x7; + hx = tsk->thread.xstate->hardfpu.fp_regs[m]; + + if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_CAUSE_ERROR) + && ((hx & 0x7fffffff) < 0x00100000)) { + /* subnormal double to float conversion */ + long long llx; + + llx = ((long long)tsk->thread.xstate->hardfpu.fp_regs[m] << 32) + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; + + tsk->thread.xstate->hardfpu.fpul = float64_to_float32(llx); } else return 0; @@ -430,7 +398,7 @@ void float_raise(unsigned int flags) int float_rounding_mode(void) { struct task_struct *tsk = current; - int roundingMode = FPSCR_ROUNDING_MODE(tsk->thread.fpu.hard.fpscr); + int roundingMode = FPSCR_ROUNDING_MODE(tsk->thread.xstate->hardfpu.fpscr); return roundingMode; } @@ -439,19 +407,19 @@ BUILD_TRAP_HANDLER(fpu_error) struct task_struct *tsk = current; TRAP_HANDLER_DECL; - save_fpu(tsk, regs); + __unlazy_fpu(tsk, regs); fpu_exception_flags = 0; if (ieee_fpe_handler(regs)) { - tsk->thread.fpu.hard.fpscr &= + tsk->thread.xstate->hardfpu.fpscr &= ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK); - tsk->thread.fpu.hard.fpscr |= fpu_exception_flags; + tsk->thread.xstate->hardfpu.fpscr |= fpu_exception_flags; /* Set the FPSCR flag as well as cause bits - simply * replicate the cause */ - tsk->thread.fpu.hard.fpscr |= (fpu_exception_flags >> 10); + tsk->thread.xstate->hardfpu.fpscr |= (fpu_exception_flags >> 10); grab_fpu(regs); restore_fpu(tsk); - set_tsk_thread_flag(tsk, TIF_USEDFPU); - if ((((tsk->thread.fpu.hard.fpscr & FPSCR_ENABLE_MASK) >> 7) & + task_thread_info(tsk)->status |= TS_USEDFPU; + if ((((tsk->thread.xstate->hardfpu.fpscr & FPSCR_ENABLE_MASK) >> 7) & (fpu_exception_flags >> 2)) == 0) { return; } @@ -459,25 +427,3 @@ BUILD_TRAP_HANDLER(fpu_error) force_sig(SIGFPE, tsk); } - -BUILD_TRAP_HANDLER(fpu_state_restore) -{ - struct task_struct *tsk = current; - TRAP_HANDLER_DECL; - - grab_fpu(regs); - if (!user_mode(regs)) { - printk(KERN_ERR "BUG: FPU is used in kernel mode.\n"); - return; - } - - if (used_math()) { - /* Using the FPU again. */ - restore_fpu(tsk); - } else { - /* First time FPU user. */ - fpu_init(); - set_used_math(); - } - set_tsk_thread_flag(tsk, TIF_USEDFPU); -} diff --git a/arch/sh/kernel/cpu/sh4/perf_event.c b/arch/sh/kernel/cpu/sh4/perf_event.c new file mode 100644 index 00000000000..fa4f724b295 --- /dev/null +++ b/arch/sh/kernel/cpu/sh4/perf_event.c @@ -0,0 +1,268 @@ +/* + * Performance events support for SH7750-style performance counters + * + * Copyright (C) 2009 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/io.h> +#include <linux/irq.h> +#include <linux/perf_event.h> +#include <asm/processor.h> + +#define PM_CR_BASE 0xff000084 /* 16-bit */ +#define PM_CTR_BASE 0xff100004 /* 32-bit */ + +#define PMCR(n) (PM_CR_BASE + ((n) * 0x04)) +#define PMCTRH(n) (PM_CTR_BASE + 0x00 + ((n) * 0x08)) +#define PMCTRL(n) (PM_CTR_BASE + 0x04 + ((n) * 0x08)) + +#define PMCR_PMM_MASK 0x0000003f + +#define PMCR_CLKF 0x00000100 +#define PMCR_PMCLR 0x00002000 +#define PMCR_PMST 0x00004000 +#define PMCR_PMEN 0x00008000 + +static struct sh_pmu sh7750_pmu; + +/* + * There are a number of events supported by each counter (33 in total). + * Since we have 2 counters, each counter will take the event code as it + * corresponds to the PMCR PMM setting. Each counter can be configured + * independently. + * + * Event Code Description + * ---------- ----------- + * + * 0x01 Operand read access + * 0x02 Operand write access + * 0x03 UTLB miss + * 0x04 Operand cache read miss + * 0x05 Operand cache write miss + * 0x06 Instruction fetch (w/ cache) + * 0x07 Instruction TLB miss + * 0x08 Instruction cache miss + * 0x09 All operand accesses + * 0x0a All instruction accesses + * 0x0b OC RAM operand access + * 0x0d On-chip I/O space access + * 0x0e Operand access (r/w) + * 0x0f Operand cache miss (r/w) + * 0x10 Branch instruction + * 0x11 Branch taken + * 0x12 BSR/BSRF/JSR + * 0x13 Instruction execution + * 0x14 Instruction execution in parallel + * 0x15 FPU Instruction execution + * 0x16 Interrupt + * 0x17 NMI + * 0x18 trapa instruction execution + * 0x19 UBCA match + * 0x1a UBCB match + * 0x21 Instruction cache fill + * 0x22 Operand cache fill + * 0x23 Elapsed time + * 0x24 Pipeline freeze by I-cache miss + * 0x25 Pipeline freeze by D-cache miss + * 0x27 Pipeline freeze by branch instruction + * 0x28 Pipeline freeze by CPU register + * 0x29 Pipeline freeze by FPU + */ + +static const int sh7750_general_events[] = { + [PERF_COUNT_HW_CPU_CYCLES] = 0x0023, + [PERF_COUNT_HW_INSTRUCTIONS] = 0x000a, + [PERF_COUNT_HW_CACHE_REFERENCES] = 0x0006, /* I-cache */ + [PERF_COUNT_HW_CACHE_MISSES] = 0x0008, /* I-cache */ + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x0010, + [PERF_COUNT_HW_BRANCH_MISSES] = -1, + [PERF_COUNT_HW_BUS_CYCLES] = -1, +}; + +#define C(x) PERF_COUNT_HW_CACHE_##x + +static const int sh7750_cache_events + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0001, + [ C(RESULT_MISS) ] = 0x0004, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0002, + [ C(RESULT_MISS) ] = 0x0005, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + + [ C(L1I) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0006, + [ C(RESULT_MISS) ] = 0x0008, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + + [ C(LL) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0x0003, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0x0007, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + + [ C(BPU) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + + [ C(NODE) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + +static int sh7750_event_map(int event) +{ + return sh7750_general_events[event]; +} + +static u64 sh7750_pmu_read(int idx) +{ + return (u64)((u64)(__raw_readl(PMCTRH(idx)) & 0xffff) << 32) | + __raw_readl(PMCTRL(idx)); +} + +static void sh7750_pmu_disable(struct hw_perf_event *hwc, int idx) +{ + unsigned int tmp; + + tmp = __raw_readw(PMCR(idx)); + tmp &= ~(PMCR_PMM_MASK | PMCR_PMEN); + __raw_writew(tmp, PMCR(idx)); +} + +static void sh7750_pmu_enable(struct hw_perf_event *hwc, int idx) +{ + __raw_writew(__raw_readw(PMCR(idx)) | PMCR_PMCLR, PMCR(idx)); + __raw_writew(hwc->config | PMCR_PMEN | PMCR_PMST, PMCR(idx)); +} + +static void sh7750_pmu_disable_all(void) +{ + int i; + + for (i = 0; i < sh7750_pmu.num_events; i++) + __raw_writew(__raw_readw(PMCR(i)) & ~PMCR_PMEN, PMCR(i)); +} + +static void sh7750_pmu_enable_all(void) +{ + int i; + + for (i = 0; i < sh7750_pmu.num_events; i++) + __raw_writew(__raw_readw(PMCR(i)) | PMCR_PMEN, PMCR(i)); +} + +static struct sh_pmu sh7750_pmu = { + .name = "sh7750", + .num_events = 2, + .event_map = sh7750_event_map, + .max_events = ARRAY_SIZE(sh7750_general_events), + .raw_event_mask = PMCR_PMM_MASK, + .cache_events = &sh7750_cache_events, + .read = sh7750_pmu_read, + .disable = sh7750_pmu_disable, + .enable = sh7750_pmu_enable, + .disable_all = sh7750_pmu_disable_all, + .enable_all = sh7750_pmu_enable_all, +}; + +static int __init sh7750_pmu_init(void) +{ + /* + * Make sure this CPU actually has perf counters. + */ + if (!(boot_cpu_data.flags & CPU_HAS_PERF_COUNTER)) { + pr_notice("HW perf events unsupported, software events only.\n"); + return -ENODEV; + } + + return register_sh_pmu(&sh7750_pmu); +} +early_initcall(sh7750_pmu_init); diff --git a/arch/sh/kernel/cpu/sh4/probe.c b/arch/sh/kernel/cpu/sh4/probe.c index 2e42572b1b1..a521bcf5069 100644 --- a/arch/sh/kernel/cpu/sh4/probe.c +++ b/arch/sh/kernel/cpu/sh4/probe.c @@ -15,7 +15,7 @@ #include <asm/processor.h> #include <asm/cache.h> -int __init detect_cpu_and_cache_system(void) +void cpu_probe(void) { unsigned long pvr, prr, cvr; unsigned long size; @@ -28,9 +28,9 @@ int __init detect_cpu_and_cache_system(void) [9] = (1 << 16) }; - pvr = (ctrl_inl(CCN_PVR) >> 8) & 0xffffff; - prr = (ctrl_inl(CCN_PRR) >> 4) & 0xff; - cvr = (ctrl_inl(CCN_CVR)); + pvr = (__raw_readl(CCN_PVR) >> 8) & 0xffffff; + prr = (__raw_readl(CCN_PRR) >> 4) & 0xff; + cvr = (__raw_readl(CCN_CVR)); /* * Setup some sane SH-4 defaults for the icache @@ -57,15 +57,26 @@ int __init detect_cpu_and_cache_system(void) * Setup some generic flags we can probe on SH-4A parts */ if (((pvr >> 16) & 0xff) == 0x10) { - if ((cvr & 0x10000000) == 0) + boot_cpu_data.family = CPU_FAMILY_SH4A; + + if ((cvr & 0x10000000) == 0) { boot_cpu_data.flags |= CPU_HAS_DSP; + boot_cpu_data.family = CPU_FAMILY_SH4AL_DSP; + } - boot_cpu_data.flags |= CPU_HAS_LLSC; + boot_cpu_data.flags |= CPU_HAS_LLSC | CPU_HAS_PERF_COUNTER; boot_cpu_data.cut_major = pvr & 0x7f; + + boot_cpu_data.icache.ways = 4; + boot_cpu_data.dcache.ways = 4; + } else { + /* And some SH-4 defaults.. */ + boot_cpu_data.flags |= CPU_HAS_PTEA | CPU_HAS_FPU; + boot_cpu_data.family = CPU_FAMILY_SH4; } - /* FPU detection works for everyone */ - if ((cvr & 0x20000000) == 1) + /* FPU detection works for almost everyone */ + if ((cvr & 0x20000000)) boot_cpu_data.flags |= CPU_HAS_FPU; /* Mask off the upper chip ID */ @@ -78,25 +89,20 @@ int __init detect_cpu_and_cache_system(void) switch (pvr) { case 0x205: boot_cpu_data.type = CPU_SH7750; - boot_cpu_data.flags |= CPU_HAS_P2_FLUSH_BUG | CPU_HAS_FPU | - CPU_HAS_PERF_COUNTER; + boot_cpu_data.flags |= CPU_HAS_P2_FLUSH_BUG | + CPU_HAS_PERF_COUNTER; break; case 0x206: boot_cpu_data.type = CPU_SH7750S; - boot_cpu_data.flags |= CPU_HAS_P2_FLUSH_BUG | CPU_HAS_FPU | - CPU_HAS_PERF_COUNTER; + boot_cpu_data.flags |= CPU_HAS_P2_FLUSH_BUG | + CPU_HAS_PERF_COUNTER; break; case 0x1100: boot_cpu_data.type = CPU_SH7751; - boot_cpu_data.flags |= CPU_HAS_FPU; break; case 0x2001: case 0x2004: boot_cpu_data.type = CPU_SH7770; - boot_cpu_data.icache.ways = 4; - boot_cpu_data.dcache.ways = 4; - - boot_cpu_data.flags |= CPU_HAS_FPU | CPU_HAS_LLSC; break; case 0x2006: case 0x200A: @@ -107,38 +113,27 @@ int __init detect_cpu_and_cache_system(void) else boot_cpu_data.type = CPU_SH7780; - boot_cpu_data.icache.ways = 4; - boot_cpu_data.dcache.ways = 4; - - boot_cpu_data.flags |= CPU_HAS_FPU | CPU_HAS_PERF_COUNTER | - CPU_HAS_LLSC; break; case 0x3000: case 0x3003: case 0x3009: boot_cpu_data.type = CPU_SH7343; - boot_cpu_data.icache.ways = 4; - boot_cpu_data.dcache.ways = 4; - boot_cpu_data.flags |= CPU_HAS_LLSC; break; case 0x3004: case 0x3007: boot_cpu_data.type = CPU_SH7785; - boot_cpu_data.icache.ways = 4; - boot_cpu_data.dcache.ways = 4; - boot_cpu_data.flags |= CPU_HAS_FPU | CPU_HAS_PERF_COUNTER | - CPU_HAS_LLSC; + break; + case 0x4004: + case 0x4005: + boot_cpu_data.type = CPU_SH7786; + boot_cpu_data.flags |= CPU_HAS_PTEAEX | CPU_HAS_L2_CACHE; break; case 0x3008: - boot_cpu_data.icache.ways = 4; - boot_cpu_data.dcache.ways = 4; - boot_cpu_data.flags |= CPU_HAS_LLSC; - switch (prr) { case 0x50: case 0x51: boot_cpu_data.type = CPU_SH7723; - boot_cpu_data.flags |= CPU_HAS_FPU | CPU_HAS_L2_CACHE; + boot_cpu_data.flags |= CPU_HAS_L2_CACHE; break; case 0x70: boot_cpu_data.type = CPU_SH7366; @@ -149,16 +144,33 @@ int __init detect_cpu_and_cache_system(void) break; } break; + case 0x300b: + switch (prr) { + case 0x20: + boot_cpu_data.type = CPU_SH7724; + boot_cpu_data.flags |= CPU_HAS_L2_CACHE; + break; + case 0x10: + case 0x11: + boot_cpu_data.type = CPU_SH7757; + break; + case 0xd0: + case 0x40: /* yon-ten-go */ + boot_cpu_data.type = CPU_SH7372; + break; + case 0xE0: /* 0x4E0 */ + boot_cpu_data.type = CPU_SH7734; /* SH7733/SH7734 */ + break; + + } + break; case 0x4000: /* 1st cut */ case 0x4001: /* 2nd cut */ boot_cpu_data.type = CPU_SHX3; - boot_cpu_data.icache.ways = 4; - boot_cpu_data.dcache.ways = 4; - boot_cpu_data.flags |= CPU_HAS_FPU | CPU_HAS_PERF_COUNTER | - CPU_HAS_LLSC; break; case 0x700: boot_cpu_data.type = CPU_SH4_501; + boot_cpu_data.flags &= ~CPU_HAS_FPU; boot_cpu_data.icache.ways = 2; boot_cpu_data.dcache.ways = 2; break; @@ -166,7 +178,6 @@ int __init detect_cpu_and_cache_system(void) boot_cpu_data.type = CPU_SH4_202; boot_cpu_data.icache.ways = 2; boot_cpu_data.dcache.ways = 2; - boot_cpu_data.flags |= CPU_HAS_FPU; break; case 0x500 ... 0x501: switch (prr) { @@ -184,23 +195,9 @@ int __init detect_cpu_and_cache_system(void) boot_cpu_data.icache.ways = 2; boot_cpu_data.dcache.ways = 2; - boot_cpu_data.flags |= CPU_HAS_FPU; - - break; - default: - boot_cpu_data.type = CPU_SH_NONE; break; } -#ifdef CONFIG_SH_DIRECT_MAPPED - boot_cpu_data.icache.ways = 1; - boot_cpu_data.dcache.ways = 1; -#endif - -#ifdef CONFIG_CPU_HAS_PTEA - boot_cpu_data.flags |= CPU_HAS_PTEA; -#endif - /* * On anything that's not a direct-mapped cache, look to the CVR * for I/D-cache specifics. @@ -220,44 +217,47 @@ int __init detect_cpu_and_cache_system(void) } /* - * Setup the L2 cache desc - * * SH-4A's have an optional PIPT L2. */ if (boot_cpu_data.flags & CPU_HAS_L2_CACHE) { - /* Bug if we can't decode the L2 info */ - BUG_ON(!(cvr & 0xf)); - - /* Silicon and specifications have clearly never met.. */ - cvr ^= 0xf; - /* - * Size calculation is much more sensible - * than it is for the L1. - * - * Sizes are 128KB, 258KB, 512KB, and 1MB. + * Verify that it really has something hooked up, this + * is the safety net for CPUs that have optional L2 + * support yet do not implement it. */ - size = (cvr & 0xf) << 17; - - BUG_ON(!size); - - boot_cpu_data.scache.way_incr = (1 << 16); - boot_cpu_data.scache.entry_shift = 5; - boot_cpu_data.scache.ways = 4; - boot_cpu_data.scache.linesz = L1_CACHE_BYTES; - - boot_cpu_data.scache.entry_mask = - (boot_cpu_data.scache.way_incr - - boot_cpu_data.scache.linesz); - - boot_cpu_data.scache.sets = size / - (boot_cpu_data.scache.linesz * - boot_cpu_data.scache.ways); - - boot_cpu_data.scache.way_size = - (boot_cpu_data.scache.sets * - boot_cpu_data.scache.linesz); + if ((cvr & 0xf) == 0) + boot_cpu_data.flags &= ~CPU_HAS_L2_CACHE; + else { + /* + * Silicon and specifications have clearly never + * met.. + */ + cvr ^= 0xf; + + /* + * Size calculation is much more sensible + * than it is for the L1. + * + * Sizes are 128KB, 256KB, 512KB, and 1MB. + */ + size = (cvr & 0xf) << 17; + + boot_cpu_data.scache.way_incr = (1 << 16); + boot_cpu_data.scache.entry_shift = 5; + boot_cpu_data.scache.ways = 4; + boot_cpu_data.scache.linesz = L1_CACHE_BYTES; + + boot_cpu_data.scache.entry_mask = + (boot_cpu_data.scache.way_incr - + boot_cpu_data.scache.linesz); + + boot_cpu_data.scache.sets = size / + (boot_cpu_data.scache.linesz * + boot_cpu_data.scache.ways); + + boot_cpu_data.scache.way_size = + (boot_cpu_data.scache.sets * + boot_cpu_data.scache.linesz); + } } - - return 0; } diff --git a/arch/sh/kernel/cpu/sh4/setup-sh4-202.c b/arch/sh/kernel/cpu/sh4/setup-sh4-202.c index 7371abf64f8..e7a7b3cdf68 100644 --- a/arch/sh/kernel/cpu/sh4/setup-sh4-202.c +++ b/arch/sh/kernel/cpu/sh4/setup-sh4-202.c @@ -2,6 +2,7 @@ * SH4-202 Setup * * Copyright (C) 2006 Paul Mundt + * Copyright (C) 2009 Magnus Damm * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -11,28 +12,58 @@ #include <linux/init.h> #include <linux/serial.h> #include <linux/serial_sci.h> +#include <linux/sh_timer.h> +#include <linux/sh_intc.h> +#include <linux/io.h> -static struct plat_sci_port sci_platform_data[] = { - { - .mapbase = 0xffe80000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCIF, - .irqs = { 40, 41, 43, 42 }, - }, { - .flags = 0, - } +static struct plat_sci_port scif0_platform_data = { + .flags = UPF_BOOT_AUTOCONF, + .scscr = SCSCR_RE | SCSCR_TE | SCSCR_REIE, + .type = PORT_SCIF, +}; + +static struct resource scif0_resources[] = { + DEFINE_RES_MEM(0xffe80000, 0x100), + DEFINE_RES_IRQ(evt2irq(0x700)), + DEFINE_RES_IRQ(evt2irq(0x720)), + DEFINE_RES_IRQ(evt2irq(0x760)), + DEFINE_RES_IRQ(evt2irq(0x740)), }; -static struct platform_device sci_device = { +static struct platform_device scif0_device = { .name = "sh-sci", - .id = -1, + .id = 0, + .resource = scif0_resources, + .num_resources = ARRAY_SIZE(scif0_resources), .dev = { - .platform_data = sci_platform_data, + .platform_data = &scif0_platform_data, + }, +}; + +static struct sh_timer_config tmu0_platform_data = { + .channels_mask = 7, +}; + +static struct resource tmu0_resources[] = { + DEFINE_RES_MEM(0xffd80000, 0x30), + DEFINE_RES_IRQ(evt2irq(0x400)), + DEFINE_RES_IRQ(evt2irq(0x420)), + DEFINE_RES_IRQ(evt2irq(0x440)), +}; + +static struct platform_device tmu0_device = { + .name = "sh-tmu", + .id = 0, + .dev = { + .platform_data = &tmu0_platform_data, }, + .resource = tmu0_resources, + .num_resources = ARRAY_SIZE(tmu0_resources), }; static struct platform_device *sh4202_devices[] __initdata = { - &sci_device, + &scif0_device, + &tmu0_device, }; static int __init sh4202_devices_setup(void) @@ -40,9 +71,72 @@ static int __init sh4202_devices_setup(void) return platform_add_devices(sh4202_devices, ARRAY_SIZE(sh4202_devices)); } -__initcall(sh4202_devices_setup); +arch_initcall(sh4202_devices_setup); + +static struct platform_device *sh4202_early_devices[] __initdata = { + &scif0_device, + &tmu0_device, +}; + +void __init plat_early_device_setup(void) +{ + early_platform_add_devices(sh4202_early_devices, + ARRAY_SIZE(sh4202_early_devices)); +} + +enum { + UNUSED = 0, + + /* interrupt sources */ + IRL0, IRL1, IRL2, IRL3, /* only IRLM mode supported */ + HUDI, TMU0, TMU1, TMU2, RTC, SCIF, WDT, +}; + +static struct intc_vect vectors[] __initdata = { + INTC_VECT(HUDI, 0x600), + INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420), + INTC_VECT(TMU2, 0x440), INTC_VECT(TMU2, 0x460), + INTC_VECT(RTC, 0x480), INTC_VECT(RTC, 0x4a0), + INTC_VECT(RTC, 0x4c0), + INTC_VECT(SCIF, 0x700), INTC_VECT(SCIF, 0x720), + INTC_VECT(SCIF, 0x740), INTC_VECT(SCIF, 0x760), + INTC_VECT(WDT, 0x560), +}; + +static struct intc_prio_reg prio_registers[] __initdata = { + { 0xffd00004, 0, 16, 4, /* IPRA */ { TMU0, TMU1, TMU2, RTC } }, + { 0xffd00008, 0, 16, 4, /* IPRB */ { WDT, 0, 0, 0 } }, + { 0xffd0000c, 0, 16, 4, /* IPRC */ { 0, 0, SCIF, HUDI } }, + { 0xffd00010, 0, 16, 4, /* IPRD */ { IRL0, IRL1, IRL2, IRL3 } }, +}; + +static DECLARE_INTC_DESC(intc_desc, "sh4-202", vectors, NULL, + NULL, prio_registers, NULL); + +static struct intc_vect vectors_irlm[] __initdata = { + INTC_VECT(IRL0, 0x240), INTC_VECT(IRL1, 0x2a0), + INTC_VECT(IRL2, 0x300), INTC_VECT(IRL3, 0x360), +}; + +static DECLARE_INTC_DESC(intc_desc_irlm, "sh4-202_irlm", vectors_irlm, NULL, + NULL, prio_registers, NULL); void __init plat_irq_setup(void) { - /* do nothing - all IRL interrupts are handled by the board code */ + register_intc_controller(&intc_desc); +} + +#define INTC_ICR 0xffd00000UL +#define INTC_ICR_IRLM (1<<7) + +void __init plat_irq_setup_pins(int mode) +{ + switch (mode) { + case IRQ_MODE_IRQ: /* individual interrupt mode for IRL3-0 */ + __raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); + register_intc_controller(&intc_desc_irlm); + break; + default: + BUG(); + } } diff --git a/arch/sh/kernel/cpu/sh4/setup-sh7750.c b/arch/sh/kernel/cpu/sh4/setup-sh7750.c index ec884039b91..5f08c59b9f3 100644 --- a/arch/sh/kernel/cpu/sh4/setup-sh7750.c +++ b/arch/sh/kernel/cpu/sh4/setup-sh7750.c @@ -1,5 +1,5 @@ /* - * SH7750/SH7751 Setup + * SH7091/SH7750/SH7750S/SH7750R/SH7751/SH7751R Setup * * Copyright (C) 2006 Paul Mundt * Copyright (C) 2006 Jamie Lenehan @@ -12,7 +12,10 @@ #include <linux/init.h> #include <linux/serial.h> #include <linux/io.h> +#include <linux/sh_timer.h> +#include <linux/sh_intc.h> #include <linux/serial_sci.h> +#include <generated/machtypes.h> static struct resource rtc_resources[] = { [0] = { @@ -21,18 +24,8 @@ static struct resource rtc_resources[] = { .flags = IORESOURCE_IO, }, [1] = { - /* Period IRQ */ - .start = 21, - .flags = IORESOURCE_IRQ, - }, - [2] = { - /* Carry IRQ */ - .start = 22, - .flags = IORESOURCE_IRQ, - }, - [3] = { - /* Alarm IRQ */ - .start = 20, + /* Shared Period/Carry/Alarm IRQ */ + .start = evt2irq(0x480), .flags = IORESOURCE_IRQ, }, }; @@ -44,86 +37,176 @@ static struct platform_device rtc_device = { .resource = rtc_resources, }; -static struct plat_sci_port sci_platform_data[] = { - { -#ifndef CONFIG_SH_RTS7751R2D - .mapbase = 0xffe00000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCI, - .irqs = { 23, 24, 25, 0 }, - }, { -#endif - .mapbase = 0xffe80000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCIF, - .irqs = { 40, 41, 43, 42 }, - }, { - .flags = 0, - } +static struct plat_sci_port sci_platform_data = { + .port_reg = 0xffe0001C, + .flags = UPF_BOOT_AUTOCONF, + .scscr = SCSCR_TE | SCSCR_RE, + .type = PORT_SCI, + .regshift = 2, +}; + +static struct resource sci_resources[] = { + DEFINE_RES_MEM(0xffe00000, 0x100), + DEFINE_RES_IRQ(evt2irq(0x4e0)), }; static struct platform_device sci_device = { .name = "sh-sci", - .id = -1, + .id = 0, + .resource = sci_resources, + .num_resources = ARRAY_SIZE(sci_resources), .dev = { - .platform_data = sci_platform_data, + .platform_data = &sci_platform_data, + }, +}; + +static struct plat_sci_port scif_platform_data = { + .flags = UPF_BOOT_AUTOCONF, + .scscr = SCSCR_TE | SCSCR_RE | SCSCR_REIE, + .type = PORT_SCIF, +}; + +static struct resource scif_resources[] = { + DEFINE_RES_MEM(0xffe80000, 0x100), + DEFINE_RES_IRQ(evt2irq(0x700)), +}; + +static struct platform_device scif_device = { + .name = "sh-sci", + .id = 1, + .resource = scif_resources, + .num_resources = ARRAY_SIZE(scif_resources), + .dev = { + .platform_data = &scif_platform_data, + }, +}; + +static struct sh_timer_config tmu0_platform_data = { + .channels_mask = 7, +}; + +static struct resource tmu0_resources[] = { + DEFINE_RES_MEM(0xffd80000, 0x30), + DEFINE_RES_IRQ(evt2irq(0x400)), + DEFINE_RES_IRQ(evt2irq(0x420)), + DEFINE_RES_IRQ(evt2irq(0x440)), +}; + +static struct platform_device tmu0_device = { + .name = "sh-tmu", + .id = 0, + .dev = { + .platform_data = &tmu0_platform_data, + }, + .resource = tmu0_resources, + .num_resources = ARRAY_SIZE(tmu0_resources), +}; + +/* SH7750R, SH7751 and SH7751R all have two extra timer channels */ +#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \ + defined(CONFIG_CPU_SUBTYPE_SH7751) || \ + defined(CONFIG_CPU_SUBTYPE_SH7751R) + +static struct sh_timer_config tmu1_platform_data = { + .channels_mask = 3, +}; + +static struct resource tmu1_resources[] = { + DEFINE_RES_MEM(0xfe100000, 0x20), + DEFINE_RES_IRQ(evt2irq(0xb00)), + DEFINE_RES_IRQ(evt2irq(0xb80)), +}; + +static struct platform_device tmu1_device = { + .name = "sh-tmu", + .id = 1, + .dev = { + .platform_data = &tmu1_platform_data, }, + .resource = tmu1_resources, + .num_resources = ARRAY_SIZE(tmu1_resources), }; +#endif + static struct platform_device *sh7750_devices[] __initdata = { &rtc_device, - &sci_device, + &tmu0_device, +#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \ + defined(CONFIG_CPU_SUBTYPE_SH7751) || \ + defined(CONFIG_CPU_SUBTYPE_SH7751R) + &tmu1_device, +#endif }; static int __init sh7750_devices_setup(void) { + if (mach_is_rts7751r2d()) { + platform_device_register(&scif_device); + } else { + platform_device_register(&sci_device); + platform_device_register(&scif_device); + } + return platform_add_devices(sh7750_devices, ARRAY_SIZE(sh7750_devices)); } -__initcall(sh7750_devices_setup); +arch_initcall(sh7750_devices_setup); + +static struct platform_device *sh7750_early_devices[] __initdata = { + &tmu0_device, +#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \ + defined(CONFIG_CPU_SUBTYPE_SH7751) || \ + defined(CONFIG_CPU_SUBTYPE_SH7751R) + &tmu1_device, +#endif +}; + +void __init plat_early_device_setup(void) +{ + struct platform_device *dev[1]; + + if (mach_is_rts7751r2d()) { + scif_platform_data.scscr |= SCSCR_CKE1; + dev[0] = &scif_device; + early_platform_add_devices(dev, 1); + } else { + dev[0] = &sci_device; + early_platform_add_devices(dev, 1); + dev[0] = &scif_device; + early_platform_add_devices(dev, 1); + } + + early_platform_add_devices(sh7750_early_devices, + ARRAY_SIZE(sh7750_early_devices)); +} enum { UNUSED = 0, /* interrupt sources */ IRL0, IRL1, IRL2, IRL3, /* only IRLM mode supported */ - HUDI, GPIOI, - DMAC_DMTE0, DMAC_DMTE1, DMAC_DMTE2, DMAC_DMTE3, - DMAC_DMTE4, DMAC_DMTE5, DMAC_DMTE6, DMAC_DMTE7, - DMAC_DMAE, + HUDI, GPIOI, DMAC, PCIC0_PCISERR, PCIC1_PCIERR, PCIC1_PCIPWDWN, PCIC1_PCIPWON, PCIC1_PCIDMA0, PCIC1_PCIDMA1, PCIC1_PCIDMA2, PCIC1_PCIDMA3, - TMU3, TMU4, TMU0, TMU1, TMU2_TUNI, TMU2_TICPI, - RTC_ATI, RTC_PRI, RTC_CUI, - SCI1_ERI, SCI1_RXI, SCI1_TXI, SCI1_TEI, - SCIF_ERI, SCIF_RXI, SCIF_BRI, SCIF_TXI, - WDT, - REF_RCMI, REF_ROVI, + TMU3, TMU4, TMU0, TMU1, TMU2, RTC, SCI1, SCIF, WDT, REF, /* interrupt groups */ - DMAC, PCIC1, TMU2, RTC, SCI1, SCIF, REF, + PCIC1, }; static struct intc_vect vectors[] __initdata = { INTC_VECT(HUDI, 0x600), INTC_VECT(GPIOI, 0x620), INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420), - INTC_VECT(TMU2_TUNI, 0x440), INTC_VECT(TMU2_TICPI, 0x460), - INTC_VECT(RTC_ATI, 0x480), INTC_VECT(RTC_PRI, 0x4a0), - INTC_VECT(RTC_CUI, 0x4c0), - INTC_VECT(SCI1_ERI, 0x4e0), INTC_VECT(SCI1_RXI, 0x500), - INTC_VECT(SCI1_TXI, 0x520), INTC_VECT(SCI1_TEI, 0x540), - INTC_VECT(SCIF_ERI, 0x700), INTC_VECT(SCIF_RXI, 0x720), - INTC_VECT(SCIF_BRI, 0x740), INTC_VECT(SCIF_TXI, 0x760), + INTC_VECT(TMU2, 0x440), INTC_VECT(TMU2, 0x460), + INTC_VECT(RTC, 0x480), INTC_VECT(RTC, 0x4a0), + INTC_VECT(RTC, 0x4c0), + INTC_VECT(SCI1, 0x4e0), INTC_VECT(SCI1, 0x500), + INTC_VECT(SCI1, 0x520), INTC_VECT(SCI1, 0x540), + INTC_VECT(SCIF, 0x700), INTC_VECT(SCIF, 0x720), + INTC_VECT(SCIF, 0x740), INTC_VECT(SCIF, 0x760), INTC_VECT(WDT, 0x560), - INTC_VECT(REF_RCMI, 0x580), INTC_VECT(REF_ROVI, 0x5a0), -}; - -static struct intc_group groups[] __initdata = { - INTC_GROUP(TMU2, TMU2_TUNI, TMU2_TICPI), - INTC_GROUP(RTC, RTC_ATI, RTC_PRI, RTC_CUI), - INTC_GROUP(SCI1, SCI1_ERI, SCI1_RXI, SCI1_TXI, SCI1_TEI), - INTC_GROUP(SCIF, SCIF_ERI, SCIF_RXI, SCIF_BRI, SCIF_TXI), - INTC_GROUP(REF, REF_RCMI, REF_ROVI), + INTC_VECT(REF, 0x580), INTC_VECT(REF, 0x5a0), }; static struct intc_prio_reg prio_registers[] __initdata = { @@ -136,7 +219,7 @@ static struct intc_prio_reg prio_registers[] __initdata = { PCIC1, PCIC0_PCISERR } }, }; -static DECLARE_INTC_DESC(intc_desc, "sh7750", vectors, groups, +static DECLARE_INTC_DESC(intc_desc, "sh7750", vectors, NULL, NULL, prio_registers, NULL); /* SH7750, SH7750S, SH7751 and SH7091 all have 4-channel DMA controllers */ @@ -145,39 +228,28 @@ static DECLARE_INTC_DESC(intc_desc, "sh7750", vectors, groups, defined(CONFIG_CPU_SUBTYPE_SH7751) || \ defined(CONFIG_CPU_SUBTYPE_SH7091) static struct intc_vect vectors_dma4[] __initdata = { - INTC_VECT(DMAC_DMTE0, 0x640), INTC_VECT(DMAC_DMTE1, 0x660), - INTC_VECT(DMAC_DMTE2, 0x680), INTC_VECT(DMAC_DMTE3, 0x6a0), - INTC_VECT(DMAC_DMAE, 0x6c0), -}; - -static struct intc_group groups_dma4[] __initdata = { - INTC_GROUP(DMAC, DMAC_DMTE0, DMAC_DMTE1, DMAC_DMTE2, - DMAC_DMTE3, DMAC_DMAE), + INTC_VECT(DMAC, 0x640), INTC_VECT(DMAC, 0x660), + INTC_VECT(DMAC, 0x680), INTC_VECT(DMAC, 0x6a0), + INTC_VECT(DMAC, 0x6c0), }; static DECLARE_INTC_DESC(intc_desc_dma4, "sh7750_dma4", - vectors_dma4, groups_dma4, + vectors_dma4, NULL, NULL, prio_registers, NULL); #endif /* SH7750R and SH7751R both have 8-channel DMA controllers */ #if defined(CONFIG_CPU_SUBTYPE_SH7750R) || defined(CONFIG_CPU_SUBTYPE_SH7751R) static struct intc_vect vectors_dma8[] __initdata = { - INTC_VECT(DMAC_DMTE0, 0x640), INTC_VECT(DMAC_DMTE1, 0x660), - INTC_VECT(DMAC_DMTE2, 0x680), INTC_VECT(DMAC_DMTE3, 0x6a0), - INTC_VECT(DMAC_DMTE4, 0x780), INTC_VECT(DMAC_DMTE5, 0x7a0), - INTC_VECT(DMAC_DMTE6, 0x7c0), INTC_VECT(DMAC_DMTE7, 0x7e0), - INTC_VECT(DMAC_DMAE, 0x6c0), -}; - -static struct intc_group groups_dma8[] __initdata = { - INTC_GROUP(DMAC, DMAC_DMTE0, DMAC_DMTE1, DMAC_DMTE2, - DMAC_DMTE3, DMAC_DMTE4, DMAC_DMTE5, - DMAC_DMTE6, DMAC_DMTE7, DMAC_DMAE), + INTC_VECT(DMAC, 0x640), INTC_VECT(DMAC, 0x660), + INTC_VECT(DMAC, 0x680), INTC_VECT(DMAC, 0x6a0), + INTC_VECT(DMAC, 0x780), INTC_VECT(DMAC, 0x7a0), + INTC_VECT(DMAC, 0x7c0), INTC_VECT(DMAC, 0x7e0), + INTC_VECT(DMAC, 0x6c0), }; static DECLARE_INTC_DESC(intc_desc_dma8, "sh7750_dma8", - vectors_dma8, groups_dma8, + vectors_dma8, NULL, NULL, prio_registers, NULL); #endif @@ -285,7 +357,7 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ: /* individual interrupt mode for IRL3-0 */ - ctrl_outw(ctrl_inw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); + __raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); register_intc_controller(&intc_desc_irlm); break; default: diff --git a/arch/sh/kernel/cpu/sh4/setup-sh7760.c b/arch/sh/kernel/cpu/sh4/setup-sh7760.c index 254c5c55ab9..973b736b3b9 100644 --- a/arch/sh/kernel/cpu/sh4/setup-sh7760.c +++ b/arch/sh/kernel/cpu/sh4/setup-sh7760.c @@ -10,17 +10,17 @@ #include <linux/platform_device.h> #include <linux/init.h> #include <linux/serial.h> +#include <linux/sh_timer.h> +#include <linux/sh_intc.h> #include <linux/serial_sci.h> +#include <linux/io.h> enum { UNUSED = 0, /* interrupt sources */ IRL0, IRL1, IRL2, IRL3, - HUDI, GPIOI, - DMAC_DMTE0, DMAC_DMTE1, DMAC_DMTE2, DMAC_DMTE3, - DMAC_DMTE4, DMAC_DMTE5, DMAC_DMTE6, DMAC_DMTE7, - DMAC_DMAE, + HUDI, GPIOI, DMAC, IRQ4, IRQ5, IRQ6, IRQ7, HCAN20, HCAN21, SSI0, SSI1, @@ -35,21 +35,20 @@ enum { HSPI, MMCIF0, MMCIF1, MMCIF2, MMCIF3, MFI, ADC, CMT, - TMU0, TMU1, TMU2_TUNI, TMU2_TICPI, - WDT, - REF_RCMI, REF_ROVI, + TMU0, TMU1, TMU2, + WDT, REF, /* interrupt groups */ - DMAC, DMABRG, SCIF0, SCIF1, SCIF2, SIM, MMCIF, TMU2, REF, + DMABRG, SCIF0, SCIF1, SCIF2, SIM, MMCIF, }; static struct intc_vect vectors[] __initdata = { INTC_VECT(HUDI, 0x600), INTC_VECT(GPIOI, 0x620), - INTC_VECT(DMAC_DMTE0, 0x640), INTC_VECT(DMAC_DMTE1, 0x660), - INTC_VECT(DMAC_DMTE2, 0x680), INTC_VECT(DMAC_DMTE3, 0x6a0), - INTC_VECT(DMAC_DMTE4, 0x780), INTC_VECT(DMAC_DMTE5, 0x7a0), - INTC_VECT(DMAC_DMTE6, 0x7c0), INTC_VECT(DMAC_DMTE7, 0x7e0), - INTC_VECT(DMAC_DMAE, 0x6c0), + INTC_VECT(DMAC, 0x640), INTC_VECT(DMAC, 0x660), + INTC_VECT(DMAC, 0x680), INTC_VECT(DMAC, 0x6a0), + INTC_VECT(DMAC, 0x780), INTC_VECT(DMAC, 0x7a0), + INTC_VECT(DMAC, 0x7c0), INTC_VECT(DMAC, 0x7e0), + INTC_VECT(DMAC, 0x6c0), INTC_VECT(IRQ4, 0x800), INTC_VECT(IRQ5, 0x820), INTC_VECT(IRQ6, 0x840), INTC_VECT(IRQ6, 0x860), INTC_VECT(HCAN20, 0x900), INTC_VECT(HCAN21, 0x920), @@ -73,23 +72,18 @@ static struct intc_vect vectors[] __initdata = { INTC_VECT(MFI, 0xe80), /* 0xf80 according to data sheet */ INTC_VECT(ADC, 0xf80), INTC_VECT(CMT, 0xfa0), INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420), - INTC_VECT(TMU2_TUNI, 0x440), INTC_VECT(TMU2_TICPI, 0x460), + INTC_VECT(TMU2, 0x440), INTC_VECT(TMU2, 0x460), INTC_VECT(WDT, 0x560), - INTC_VECT(REF_RCMI, 0x580), INTC_VECT(REF_ROVI, 0x5a0), + INTC_VECT(REF, 0x580), INTC_VECT(REF, 0x5a0), }; static struct intc_group groups[] __initdata = { - INTC_GROUP(DMAC, DMAC_DMTE0, DMAC_DMTE1, DMAC_DMTE2, - DMAC_DMTE3, DMAC_DMTE4, DMAC_DMTE5, - DMAC_DMTE6, DMAC_DMTE7, DMAC_DMAE), INTC_GROUP(DMABRG, DMABRG0, DMABRG1, DMABRG2), INTC_GROUP(SCIF0, SCIF0_ERI, SCIF0_RXI, SCIF0_BRI, SCIF0_TXI), INTC_GROUP(SCIF1, SCIF1_ERI, SCIF1_RXI, SCIF1_BRI, SCIF1_TXI), INTC_GROUP(SCIF2, SCIF2_ERI, SCIF2_RXI, SCIF2_BRI, SCIF2_TXI), INTC_GROUP(SIM, SIM_ERI, SIM_RXI, SIM_TXI, SIM_TEI), INTC_GROUP(MMCIF, MMCIF0, MMCIF1, MMCIF2, MMCIF3), - INTC_GROUP(TMU2, TMU2_TUNI, TMU2_TICPI), - INTC_GROUP(REF, REF_RCMI, REF_ROVI), }; static struct intc_mask_reg mask_registers[] __initdata = { @@ -133,42 +127,133 @@ static struct intc_vect vectors_irq[] __initdata = { static DECLARE_INTC_DESC(intc_desc_irq, "sh7760-irq", vectors_irq, groups, mask_registers, prio_registers, NULL); -static struct plat_sci_port sci_platform_data[] = { - { - .mapbase = 0xfe600000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCIF, - .irqs = { 52, 53, 55, 54 }, - }, { - .mapbase = 0xfe610000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCIF, - .irqs = { 72, 73, 75, 74 }, - }, { - .mapbase = 0xfe620000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCIF, - .irqs = { 76, 77, 79, 78 }, - }, { - .mapbase = 0xfe480000, - .flags = UPF_BOOT_AUTOCONF, - .type = PORT_SCI, - .irqs = { 80, 81, 82, 0 }, - }, { - .flags = 0, - } +static struct plat_sci_port scif0_platform_data = { + .flags = UPF_BOOT_AUTOCONF, + .scscr = SCSCR_RE | SCSCR_TE | SCSCR_REIE, + .type = PORT_SCIF, + .regtype = SCIx_SH4_SCIF_FIFODATA_REGTYPE, +}; + +static struct resource scif0_resources[] = { + DEFINE_RES_MEM(0xfe600000, 0x100), + DEFINE_RES_IRQ(evt2irq(0x880)), + DEFINE_RES_IRQ(evt2irq(0x8a0)), + DEFINE_RES_IRQ(evt2irq(0x8e0)), + DEFINE_RES_IRQ(evt2irq(0x8c0)), +}; + +static struct platform_device scif0_device = { + .name = "sh-sci", + .id = 0, + .resource = scif0_resources, + .num_resources = ARRAY_SIZE(scif0_resources), + .dev = { + .platform_data = &scif0_platform_data, + }, +}; + +static struct plat_sci_port scif1_platform_data = { + .flags = UPF_BOOT_AUTOCONF, + .type = PORT_SCIF, + .scscr = SCSCR_RE | SCSCR_TE | SCSCR_REIE, + .regtype = SCIx_SH4_SCIF_FIFODATA_REGTYPE, +}; + +static struct resource scif1_resources[] = { + DEFINE_RES_MEM(0xfe610000, 0x100), + DEFINE_RES_IRQ(evt2irq(0xb00)), + DEFINE_RES_IRQ(evt2irq(0xb20)), + DEFINE_RES_IRQ(evt2irq(0xb60)), + DEFINE_RES_IRQ(evt2irq(0xb40)), +}; + +static struct platform_device scif1_device = { + .name = "sh-sci", + .id = 1, + .resource = scif1_resources, + .num_resources = ARRAY_SIZE(scif1_resources), + .dev = { + .platform_data = &scif1_platform_data, + }, +}; + +static struct plat_sci_port scif2_platform_data = { + .flags = UPF_BOOT_AUTOCONF, + .scscr = SCSCR_RE | SCSCR_TE | SCSCR_REIE, + .type = PORT_SCIF, + .regtype = SCIx_SH4_SCIF_FIFODATA_REGTYPE, +}; + +static struct resource scif2_resources[] = { + DEFINE_RES_MEM(0xfe620000, 0x100), + DEFINE_RES_IRQ(evt2irq(0xb80)), + DEFINE_RES_IRQ(evt2irq(0xba0)), + DEFINE_RES_IRQ(evt2irq(0xbe0)), + DEFINE_RES_IRQ(evt2irq(0xbc0)), +}; + +static struct platform_device scif2_device = { + .name = "sh-sci", + .id = 2, + .resource = scif2_resources, + .num_resources = ARRAY_SIZE(scif2_resources), + .dev = { + .platform_data = &scif2_platform_data, + }, +}; + +static struct plat_sci_port scif3_platform_data = { + .flags = UPF_BOOT_AUTOCONF, + .scscr = SCSCR_RE | SCSCR_TE | SCSCR_REIE, + .type = PORT_SCI, + .regshift = 2, }; -static struct platform_device sci_device = { +static struct resource scif3_resources[] = { + DEFINE_RES_MEM(0xfe480000, 0x100), + DEFINE_RES_IRQ(evt2irq(0xc00)), + DEFINE_RES_IRQ(evt2irq(0xc20)), + DEFINE_RES_IRQ(evt2irq(0xc40)), +}; + +static struct platform_device scif3_device = { .name = "sh-sci", - .id = -1, + .id = 3, + .resource = scif3_resources, + .num_resources = ARRAY_SIZE(scif3_resources), .dev = { - .platform_data = sci_platform_data, + .platform_data = &scif3_platform_data, + }, +}; + +static struct sh_timer_config tmu0_platform_data = { + .channels_mask = 7, +}; + +static struct resource tmu0_resources[] = { + DEFINE_RES_MEM(0xffd80000, 0x30), + DEFINE_RES_IRQ(evt2irq(0x400)), + DEFINE_RES_IRQ(evt2irq(0x420)), + DEFINE_RES_IRQ(evt2irq(0x440)), +}; + +static struct platform_device tmu0_device = { + .name = "sh-tmu", + .id = 0, + .dev = { + .platform_data = &tmu0_platform_data, }, + .resource = tmu0_resources, + .num_resources = ARRAY_SIZE(tmu0_resources), }; + static struct platform_device *sh7760_devices[] __initdata = { - &sci_device, + &scif0_device, + &scif1_device, + &scif2_device, + &scif3_device, + &tmu0_device, }; static int __init sh7760_devices_setup(void) @@ -176,12 +261,30 @@ static int __init sh7760_devices_setup(void) return platform_add_devices(sh7760_devices, ARRAY_SIZE(sh7760_devices)); } -__initcall(sh7760_devices_setup); +arch_initcall(sh7760_devices_setup); + +static struct platform_device *sh7760_early_devices[] __initdata = { + &scif0_device, + &scif1_device, + &scif2_device, + &scif3_device, + &tmu0_device, +}; + +void __init plat_early_device_setup(void) +{ + early_platform_add_devices(sh7760_early_devices, + ARRAY_SIZE(sh7760_early_devices)); +} + +#define INTC_ICR 0xffd00000UL +#define INTC_ICR_IRLM (1 << 7) void __init plat_irq_setup_pins(int mode) { switch (mode) { case IRQ_MODE_IRQ: + __raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); register_intc_controller(&intc_desc_irq); break; default: diff --git a/arch/sh/kernel/cpu/sh4/softfloat.c b/arch/sh/kernel/cpu/sh4/softfloat.c index 7b2d337ee41..42edf2e54e8 100644 --- a/arch/sh/kernel/cpu/sh4/softfloat.c +++ b/arch/sh/kernel/cpu/sh4/softfloat.c @@ -36,7 +36,8 @@ * and Kamel Khelifi <kamel.khelifi@st.com> */ #include <linux/kernel.h> -#include <asm/cpu/fpu.h> +#include <cpu/fpu.h> +#include <asm/div64.h> #define LIT64( a ) a##LL @@ -67,16 +68,16 @@ typedef unsigned long long float64; extern void float_raise(unsigned int flags); /* in fpu.c */ extern int float_rounding_mode(void); /* in fpu.c */ -inline bits64 extractFloat64Frac(float64 a); -inline flag extractFloat64Sign(float64 a); -inline int16 extractFloat64Exp(float64 a); -inline int16 extractFloat32Exp(float32 a); -inline flag extractFloat32Sign(float32 a); -inline bits32 extractFloat32Frac(float32 a); -inline float64 packFloat64(flag zSign, int16 zExp, bits64 zSig); -inline void shift64RightJamming(bits64 a, int16 count, bits64 * zPtr); -inline float32 packFloat32(flag zSign, int16 zExp, bits32 zSig); -inline void shift32RightJamming(bits32 a, int16 count, bits32 * zPtr); +bits64 extractFloat64Frac(float64 a); +flag extractFloat64Sign(float64 a); +int16 extractFloat64Exp(float64 a); +int16 extractFloat32Exp(float32 a); +flag extractFloat32Sign(float32 a); +bits32 extractFloat32Frac(float32 a); +float64 packFloat64(flag zSign, int16 zExp, bits64 zSig); +void shift64RightJamming(bits64 a, int16 count, bits64 * zPtr); +float32 packFloat32(flag zSign, int16 zExp, bits32 zSig); +void shift32RightJamming(bits32 a, int16 count, bits32 * zPtr); float64 float64_sub(float64 a, float64 b); float32 float32_sub(float32 a, float32 b); float32 float32_add(float32 a, float32 b); @@ -85,11 +86,12 @@ float64 float64_div(float64 a, float64 b); float32 float32_div(float32 a, float32 b); float32 float32_mul(float32 a, float32 b); float64 float64_mul(float64 a, float64 b); -inline void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, +float32 float64_to_float32(float64 a); +void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, bits64 * z1Ptr); -inline void sub128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, +void sub128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, bits64 * z1Ptr); -inline void mul64To128(bits64 a, bits64 b, bits64 * z0Ptr, bits64 * z1Ptr); +void mul64To128(bits64 a, bits64 b, bits64 * z0Ptr, bits64 * z1Ptr); static int8 countLeadingZeros32(bits32 a); static int8 countLeadingZeros64(bits64 a); @@ -109,42 +111,42 @@ static bits64 estimateDiv128To64(bits64 a0, bits64 a1, bits64 b); static void normalizeFloat32Subnormal(bits32 aSig, int16 * zExpPtr, bits32 * zSigPtr); -inline bits64 extractFloat64Frac(float64 a) +bits64 extractFloat64Frac(float64 a) { return a & LIT64(0x000FFFFFFFFFFFFF); } -inline flag extractFloat64Sign(float64 a) +flag extractFloat64Sign(float64 a) { return a >> 63; } -inline int16 extractFloat64Exp(float64 a) +int16 extractFloat64Exp(float64 a) { return (a >> 52) & 0x7FF; } -inline int16 extractFloat32Exp(float32 a) +int16 extractFloat32Exp(float32 a) { return (a >> 23) & 0xFF; } -inline flag extractFloat32Sign(float32 a) +flag extractFloat32Sign(float32 a) { return a >> 31; } -inline bits32 extractFloat32Frac(float32 a) +bits32 extractFloat32Frac(float32 a) { return a & 0x007FFFFF; } -inline float64 packFloat64(flag zSign, int16 zExp, bits64 zSig) +float64 packFloat64(flag zSign, int16 zExp, bits64 zSig) { return (((bits64) zSign) << 63) + (((bits64) zExp) << 52) + zSig; } -inline void shift64RightJamming(bits64 a, int16 count, bits64 * zPtr) +void shift64RightJamming(bits64 a, int16 count, bits64 * zPtr) { bits64 z; @@ -337,12 +339,12 @@ static float64 addFloat64Sigs(float64 a, float64 b, flag zSign) } -inline float32 packFloat32(flag zSign, int16 zExp, bits32 zSig) +float32 packFloat32(flag zSign, int16 zExp, bits32 zSig) { return (((bits32) zSign) << 31) + (((bits32) zExp) << 23) + zSig; } -inline void shift32RightJamming(bits32 a, int16 count, bits32 * zPtr) +void shift32RightJamming(bits32 a, int16 count, bits32 * zPtr) { bits32 z; if (count == 0) { @@ -633,7 +635,7 @@ normalizeFloat64Subnormal(bits64 aSig, int16 * zExpPtr, bits64 * zSigPtr) *zExpPtr = 1 - shiftCount; } -inline void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, +void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, bits64 * z1Ptr) { bits64 z1; @@ -643,7 +645,7 @@ inline void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, *z0Ptr = a0 + b0 + (z1 < a1); } -inline void +void sub128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr, bits64 * z1Ptr) { @@ -655,11 +657,14 @@ static bits64 estimateDiv128To64(bits64 a0, bits64 a1, bits64 b) { bits64 b0, b1; bits64 rem0, rem1, term0, term1; - bits64 z; + bits64 z, tmp; if (b <= a0) return LIT64(0xFFFFFFFFFFFFFFFF); b0 = b >> 32; - z = (b0 << 32 <= a0) ? LIT64(0xFFFFFFFF00000000) : (a0 / b0) << 32; + tmp = a0; + do_div(tmp, b0); + + z = (b0 << 32 <= a0) ? LIT64(0xFFFFFFFF00000000) : tmp << 32; mul64To128(b, z, &term0, &term1); sub128(a0, a1, term0, term1, &rem0, &rem1); while (((sbits64) rem0) < 0) { @@ -668,11 +673,13 @@ static bits64 estimateDiv128To64(bits64 a0, bits64 a1, bits64 b) add128(rem0, rem1, b0, b1, &rem0, &rem1); } rem0 = (rem0 << 32) | (rem1 >> 32); - z |= (b0 << 32 <= rem0) ? 0xFFFFFFFF : rem0 / b0; + tmp = rem0; + do_div(tmp, b0); + z |= (b0 << 32 <= rem0) ? 0xFFFFFFFF : tmp; return z; } -inline void mul64To128(bits64 a, bits64 b, bits64 * z0Ptr, bits64 * z1Ptr) +void mul64To128(bits64 a, bits64 b, bits64 * z0Ptr, bits64 * z1Ptr) { bits32 aHigh, aLow, bHigh, bLow; bits64 z0, zMiddleA, zMiddleB, z1; @@ -768,7 +775,8 @@ float32 float32_div(float32 a, float32 b) { flag aSign, bSign, zSign; int16 aExp, bExp, zExp; - bits32 aSig, bSig, zSig; + bits32 aSig, bSig; + uint64_t zSig; aSig = extractFloat32Frac(a); aExp = extractFloat32Exp(a); @@ -803,11 +811,13 @@ float32 float32_div(float32 a, float32 b) aSig >>= 1; ++zExp; } - zSig = (((bits64) aSig) << 32) / bSig; + zSig = (((bits64) aSig) << 32); + do_div(zSig, bSig); + if ((zSig & 0x3F) == 0) { zSig |= (((bits64) bSig) * zSig != ((bits64) aSig) << 32); } - return roundAndPackFloat32(zSign, zExp, zSig); + return roundAndPackFloat32(zSign, zExp, (bits32)zSig); } @@ -890,3 +900,31 @@ float64 float64_mul(float64 a, float64 b) } return roundAndPackFloat64(zSign, zExp, zSig0); } + +/* + * ------------------------------------------------------------------------------- + * Returns the result of converting the double-precision floating-point value + * `a' to the single-precision floating-point format. The conversion is + * performed according to the IEC/IEEE Standard for Binary Floating-point + * Arithmetic. + * ------------------------------------------------------------------------------- + * */ +float32 float64_to_float32(float64 a) +{ + flag aSign; + int16 aExp; + bits64 aSig; + bits32 zSig; + + aSig = extractFloat64Frac( a ); + aExp = extractFloat64Exp( a ); + aSign = extractFloat64Sign( a ); + + shift64RightJamming( aSig, 22, &aSig ); + zSig = aSig; + if ( aExp || zSig ) { + zSig |= 0x40000000; + aExp -= 0x381; + } + return roundAndPackFloat32(aSign, aExp, zSig); +} diff --git a/arch/sh/kernel/cpu/sh4/sq.c b/arch/sh/kernel/cpu/sh4/sq.c index 9561b02ade0..0a47bd3e7be 100644 --- a/arch/sh/kernel/cpu/sh4/sq.c +++ b/arch/sh/kernel/cpu/sh4/sq.c @@ -13,16 +13,17 @@ #include <linux/init.h> #include <linux/cpu.h> #include <linux/bitmap.h> -#include <linux/sysdev.h> +#include <linux/device.h> #include <linux/kernel.h> #include <linux/module.h> #include <linux/slab.h> #include <linux/vmalloc.h> #include <linux/mm.h> #include <linux/io.h> +#include <linux/prefetch.h> #include <asm/page.h> #include <asm/cacheflush.h> -#include <asm/cpu/sq.h> +#include <cpu/sq.h> struct sq_mapping; @@ -43,9 +44,9 @@ static unsigned long *sq_bitmap; #define store_queue_barrier() \ do { \ - (void)ctrl_inl(P4SEG_STORE_QUE); \ - ctrl_outl(0, P4SEG_STORE_QUE + 0); \ - ctrl_outl(0, P4SEG_STORE_QUE + 8); \ + (void)__raw_readl(P4SEG_STORE_QUE); \ + __raw_writel(0, P4SEG_STORE_QUE + 0); \ + __raw_writel(0, P4SEG_STORE_QUE + 8); \ } while (0); /** @@ -100,7 +101,7 @@ static inline void sq_mapping_list_del(struct sq_mapping *map) spin_unlock_irq(&sq_mapping_lock); } -static int __sq_remap(struct sq_mapping *map, unsigned long flags) +static int __sq_remap(struct sq_mapping *map, pgprot_t prot) { #if defined(CONFIG_MMU) struct vm_struct *vma; @@ -113,7 +114,7 @@ static int __sq_remap(struct sq_mapping *map, unsigned long flags) if (ioremap_page_range((unsigned long)vma->addr, (unsigned long)vma->addr + map->size, - vma->phys_addr, __pgprot(flags))) { + vma->phys_addr, prot)) { vunmap(vma->addr); return -EAGAIN; } @@ -123,8 +124,8 @@ static int __sq_remap(struct sq_mapping *map, unsigned long flags) * straightforward, as we can just load up each queue's QACR with * the physical address appropriately masked. */ - ctrl_outl(((map->addr >> 26) << 2) & 0x1c, SQ_QACR0); - ctrl_outl(((map->addr >> 26) << 2) & 0x1c, SQ_QACR1); + __raw_writel(((map->addr >> 26) << 2) & 0x1c, SQ_QACR0); + __raw_writel(((map->addr >> 26) << 2) & 0x1c, SQ_QACR1); #endif return 0; @@ -135,14 +136,14 @@ static int __sq_remap(struct sq_mapping *map, unsigned long flags) * @phys: Physical address of mapping. * @size: Length of mapping. * @name: User invoking mapping. - * @flags: Protection flags. + * @prot: Protection bits. * * Remaps the physical address @phys through the next available store queue * address of @size length. @name is logged at boot time as well as through * the sysfs interface. */ unsigned long sq_remap(unsigned long phys, unsigned int size, - const char *name, unsigned long flags) + const char *name, pgprot_t prot) { struct sq_mapping *map; unsigned long end; @@ -177,7 +178,7 @@ unsigned long sq_remap(unsigned long phys, unsigned int size, map->sq_addr = P4SEG_STORE_QUE + (page << PAGE_SHIFT); - ret = __sq_remap(map, pgprot_val(PAGE_KERNEL_NOCACHE) | flags); + ret = __sq_remap(map, prot); if (unlikely(ret != 0)) goto out; @@ -199,7 +200,7 @@ EXPORT_SYMBOL(sq_remap); /** * sq_unmap - Unmap a Store Queue allocation - * @map: Pre-allocated Store Queue mapping. + * @vaddr: Pre-allocated Store Queue mapping. * * Unmaps the store queue allocation @map that was previously created by * sq_remap(). Also frees up the pte that was previously inserted into @@ -309,8 +310,7 @@ static ssize_t mapping_store(const char *buf, size_t count) return -EIO; if (likely(len)) { - int ret = sq_remap(base, len, "Userspace", - pgprot_val(PAGE_SHARED)); + int ret = sq_remap(base, len, "Userspace", PAGE_SHARED); if (ret < 0) return ret; } else @@ -327,7 +327,7 @@ static struct attribute *sq_sysfs_attrs[] = { NULL, }; -static struct sysfs_ops sq_sysfs_ops = { +static const struct sysfs_ops sq_sysfs_ops = { .show = sq_sysfs_show, .store = sq_sysfs_store, }; @@ -337,9 +337,9 @@ static struct kobj_type ktype_percpu_entry = { .default_attrs = sq_sysfs_attrs, }; -static int __devinit sq_sysdev_add(struct sys_device *sysdev) +static int sq_dev_add(struct device *dev, struct subsys_interface *sif) { - unsigned int cpu = sysdev->id; + unsigned int cpu = dev->id; struct kobject *kobj; int error; @@ -348,25 +348,27 @@ static int __devinit sq_sysdev_add(struct sys_device *sysdev) return -ENOMEM; kobj = sq_kobject[cpu]; - error = kobject_init_and_add(kobj, &ktype_percpu_entry, &sysdev->kobj, + error = kobject_init_and_add(kobj, &ktype_percpu_entry, &dev->kobj, "%s", "sq"); if (!error) kobject_uevent(kobj, KOBJ_ADD); return error; } -static int __devexit sq_sysdev_remove(struct sys_device *sysdev) +static int sq_dev_remove(struct device *dev, struct subsys_interface *sif) { - unsigned int cpu = sysdev->id; + unsigned int cpu = dev->id; struct kobject *kobj = sq_kobject[cpu]; kobject_put(kobj); return 0; } -static struct sysdev_driver sq_sysdev_driver = { - .add = sq_sysdev_add, - .remove = __devexit_p(sq_sysdev_remove), +static struct subsys_interface sq_interface = { + .name = "sq", + .subsys = &cpu_subsys, + .add_dev = sq_dev_add, + .remove_dev = sq_dev_remove, }; static int __init sq_api_init(void) @@ -386,7 +388,7 @@ static int __init sq_api_init(void) if (unlikely(!sq_bitmap)) goto out; - ret = sysdev_driver_register(&cpu_sysdev_class, &sq_sysdev_driver); + ret = subsys_interface_register(&sq_interface); if (unlikely(ret != 0)) goto out; @@ -401,7 +403,7 @@ out: static void __exit sq_api_exit(void) { - sysdev_driver_unregister(&cpu_sysdev_class, &sq_sysdev_driver); + subsys_interface_unregister(&sq_interface); kfree(sq_bitmap); kmem_cache_destroy(sq_cache); } |