GK SDK 源码库: XMIPCLinuxV100R005C00SPC030 (kernel/tools/open_source excluded)

This commit is contained in:
lai
2026-09-06 03:52:57 +08:00
commit b1928b41c0
21813 changed files with 4413081 additions and 0 deletions
@@ -0,0 +1,17 @@
# SPDX-License-Identifier: GPL-2.0+
#
# Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
extra-y = start.o
head-y := start.o
obj-y += cache.o
obj-y += cpu.o
obj-y += interrupts.o
obj-y += relocate.o
obj-y += reset.o
obj-y += ints_low.o
obj-y += init_helpers.o
obj-$(CONFIG_CMD_BOOTM) += bootm.o
lib-$(CONFIG_USE_PRIVATE_LIBGCC) += _millicodethunk.o libgcc2.o
@@ -0,0 +1,225 @@
/* SPDX-License-Identifier: GPL-2.0+ */
/*
* Copyright (C) 1995, 1997, 2007-2013 Free Software Foundation, Inc.
*/
/* ANSI concatenation macros. */
#define CONCAT1(a, b) CONCAT2(a, b)
#define CONCAT2(a, b) a ## b
/* Use the right prefix for global labels. */
#define SYM(x) CONCAT1 (__USER_LABEL_PREFIX__, x)
#ifndef WORKING_ASSEMBLER
#define abs_l abs
#define asl_l asl
#define mov_l mov
#endif
#define FUNC(X) .type SYM(X),@function
#define HIDDEN_FUNC(X) FUNC(X)` .hidden X
#define ENDFUNC0(X) .Lfe_##X: .size X,.Lfe_##X-X
#define ENDFUNC(X) ENDFUNC0(X)
.section .text
.align 4
.global SYM(__st_r13_to_r15)
.global SYM(__st_r13_to_r16)
.global SYM(__st_r13_to_r17)
.global SYM(__st_r13_to_r18)
.global SYM(__st_r13_to_r19)
.global SYM(__st_r13_to_r20)
.global SYM(__st_r13_to_r21)
.global SYM(__st_r13_to_r22)
.global SYM(__st_r13_to_r23)
.global SYM(__st_r13_to_r24)
.global SYM(__st_r13_to_r25)
HIDDEN_FUNC(__st_r13_to_r15)
HIDDEN_FUNC(__st_r13_to_r16)
HIDDEN_FUNC(__st_r13_to_r17)
HIDDEN_FUNC(__st_r13_to_r18)
HIDDEN_FUNC(__st_r13_to_r19)
HIDDEN_FUNC(__st_r13_to_r20)
HIDDEN_FUNC(__st_r13_to_r21)
HIDDEN_FUNC(__st_r13_to_r22)
HIDDEN_FUNC(__st_r13_to_r23)
HIDDEN_FUNC(__st_r13_to_r24)
HIDDEN_FUNC(__st_r13_to_r25)
.align 4
SYM(__st_r13_to_r25):
st r25, [sp,48]
SYM(__st_r13_to_r24):
st r24, [sp,44]
SYM(__st_r13_to_r23):
st r23, [sp,40]
SYM(__st_r13_to_r22):
st r22, [sp,36]
SYM(__st_r13_to_r21):
st r21, [sp,32]
SYM(__st_r13_to_r20):
st r20, [sp,28]
SYM(__st_r13_to_r19):
st r19, [sp,24]
SYM(__st_r13_to_r18):
st r18, [sp,20]
SYM(__st_r13_to_r17):
st r17, [sp,16]
SYM(__st_r13_to_r16):
st r16, [sp,12]
SYM(__st_r13_to_r15):
#ifdef __ARC700__
st r15, [sp,8] ; minimum function size to avoid stall: 6 bytes.
#else
st_s r15, [sp,8]
#endif
st_s r14, [sp,4]
j_s.d [%blink]
st_s r13, [sp,0]
ENDFUNC(__st_r13_to_r15)
ENDFUNC(__st_r13_to_r16)
ENDFUNC(__st_r13_to_r17)
ENDFUNC(__st_r13_to_r18)
ENDFUNC(__st_r13_to_r19)
ENDFUNC(__st_r13_to_r20)
ENDFUNC(__st_r13_to_r21)
ENDFUNC(__st_r13_to_r22)
ENDFUNC(__st_r13_to_r23)
ENDFUNC(__st_r13_to_r24)
ENDFUNC(__st_r13_to_r25)
.section .text
.align 4
; ==================================
; the loads
.global SYM(__ld_r13_to_r15)
.global SYM(__ld_r13_to_r16)
.global SYM(__ld_r13_to_r17)
.global SYM(__ld_r13_to_r18)
.global SYM(__ld_r13_to_r19)
.global SYM(__ld_r13_to_r20)
.global SYM(__ld_r13_to_r21)
.global SYM(__ld_r13_to_r22)
.global SYM(__ld_r13_to_r23)
.global SYM(__ld_r13_to_r24)
.global SYM(__ld_r13_to_r25)
HIDDEN_FUNC(__ld_r13_to_r15)
HIDDEN_FUNC(__ld_r13_to_r16)
HIDDEN_FUNC(__ld_r13_to_r17)
HIDDEN_FUNC(__ld_r13_to_r18)
HIDDEN_FUNC(__ld_r13_to_r19)
HIDDEN_FUNC(__ld_r13_to_r20)
HIDDEN_FUNC(__ld_r13_to_r21)
HIDDEN_FUNC(__ld_r13_to_r22)
HIDDEN_FUNC(__ld_r13_to_r23)
HIDDEN_FUNC(__ld_r13_to_r24)
HIDDEN_FUNC(__ld_r13_to_r25)
SYM(__ld_r13_to_r25):
ld r25, [sp,48]
SYM(__ld_r13_to_r24):
ld r24, [sp,44]
SYM(__ld_r13_to_r23):
ld r23, [sp,40]
SYM(__ld_r13_to_r22):
ld r22, [sp,36]
SYM(__ld_r13_to_r21):
ld r21, [sp,32]
SYM(__ld_r13_to_r20):
ld r20, [sp,28]
SYM(__ld_r13_to_r19):
ld r19, [sp,24]
SYM(__ld_r13_to_r18):
ld r18, [sp,20]
SYM(__ld_r13_to_r17):
ld r17, [sp,16]
SYM(__ld_r13_to_r16):
ld r16, [sp,12]
SYM(__ld_r13_to_r15):
#ifdef __ARC700__
ld r15, [sp,8] ; minimum function size to avoid stall: 6 bytes.
#else
ld_s r15, [sp,8]
#endif
ld_s r14, [sp,4]
j_s.d [%blink]
ld_s r13, [sp,0]
ENDFUNC(__ld_r13_to_r15)
ENDFUNC(__ld_r13_to_r16)
ENDFUNC(__ld_r13_to_r17)
ENDFUNC(__ld_r13_to_r18)
ENDFUNC(__ld_r13_to_r19)
ENDFUNC(__ld_r13_to_r20)
ENDFUNC(__ld_r13_to_r21)
ENDFUNC(__ld_r13_to_r22)
ENDFUNC(__ld_r13_to_r23)
ENDFUNC(__ld_r13_to_r24)
ENDFUNC(__ld_r13_to_r25)
.global SYM(__ld_r13_to_r14_ret)
.global SYM(__ld_r13_to_r15_ret)
.global SYM(__ld_r13_to_r16_ret)
.global SYM(__ld_r13_to_r17_ret)
.global SYM(__ld_r13_to_r18_ret)
.global SYM(__ld_r13_to_r19_ret)
.global SYM(__ld_r13_to_r20_ret)
.global SYM(__ld_r13_to_r21_ret)
.global SYM(__ld_r13_to_r22_ret)
.global SYM(__ld_r13_to_r23_ret)
.global SYM(__ld_r13_to_r24_ret)
.global SYM(__ld_r13_to_r25_ret)
HIDDEN_FUNC(__ld_r13_to_r14_ret)
HIDDEN_FUNC(__ld_r13_to_r15_ret)
HIDDEN_FUNC(__ld_r13_to_r16_ret)
HIDDEN_FUNC(__ld_r13_to_r17_ret)
HIDDEN_FUNC(__ld_r13_to_r18_ret)
HIDDEN_FUNC(__ld_r13_to_r19_ret)
HIDDEN_FUNC(__ld_r13_to_r20_ret)
HIDDEN_FUNC(__ld_r13_to_r21_ret)
HIDDEN_FUNC(__ld_r13_to_r22_ret)
HIDDEN_FUNC(__ld_r13_to_r23_ret)
HIDDEN_FUNC(__ld_r13_to_r24_ret)
HIDDEN_FUNC(__ld_r13_to_r25_ret)
.section .text
.align 4
SYM(__ld_r13_to_r25_ret):
ld r25, [sp,48]
SYM(__ld_r13_to_r24_ret):
ld r24, [sp,44]
SYM(__ld_r13_to_r23_ret):
ld r23, [sp,40]
SYM(__ld_r13_to_r22_ret):
ld r22, [sp,36]
SYM(__ld_r13_to_r21_ret):
ld r21, [sp,32]
SYM(__ld_r13_to_r20_ret):
ld r20, [sp,28]
SYM(__ld_r13_to_r19_ret):
ld r19, [sp,24]
SYM(__ld_r13_to_r18_ret):
ld r18, [sp,20]
SYM(__ld_r13_to_r17_ret):
ld r17, [sp,16]
SYM(__ld_r13_to_r16_ret):
ld r16, [sp,12]
SYM(__ld_r13_to_r15_ret):
ld r15, [sp,8]
SYM(__ld_r13_to_r14_ret):
ld blink,[sp,r12]
ld_s r14, [sp,4]
ld.ab r13, [sp,r12]
j_s.d [%blink]
add_s sp,sp,4
ENDFUNC(__ld_r13_to_r14_ret)
ENDFUNC(__ld_r13_to_r15_ret)
ENDFUNC(__ld_r13_to_r16_ret)
ENDFUNC(__ld_r13_to_r17_ret)
ENDFUNC(__ld_r13_to_r18_ret)
ENDFUNC(__ld_r13_to_r19_ret)
ENDFUNC(__ld_r13_to_r20_ret)
ENDFUNC(__ld_r13_to_r21_ret)
ENDFUNC(__ld_r13_to_r22_ret)
ENDFUNC(__ld_r13_to_r23_ret)
ENDFUNC(__ld_r13_to_r24_ret)
ENDFUNC(__ld_r13_to_r25_ret)
@@ -0,0 +1,119 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
*/
#include <irq_func.h>
#include <asm/cache.h>
#include <common.h>
DECLARE_GLOBAL_DATA_PTR;
static ulong get_sp(void)
{
ulong ret;
asm("mov %0, sp" : "=r"(ret) : );
return ret;
}
void arch_lmb_reserve(struct lmb *lmb)
{
ulong sp;
/*
* Booting a (Linux) kernel image
*
* Allocate space for command line and board info - the
* address should be as high as possible within the reach of
* the kernel (see CONFIG_SYS_BOOTMAPSZ settings), but in unused
* memory, which means far enough below the current stack
* pointer.
*/
sp = get_sp();
debug("## Current stack ends at 0x%08lx ", sp);
/* adjust sp by 4K to be safe */
sp -= 4096;
lmb_reserve(lmb, sp, (CONFIG_SYS_SDRAM_BASE + gd->ram_size - sp));
}
static int cleanup_before_linux(void)
{
disable_interrupts();
sync_n_cleanup_cache_all();
return 0;
}
__weak int board_prep_linux(bootm_headers_t *images) { return 0; }
/* Subcommand: PREP */
static int boot_prep_linux(bootm_headers_t *images)
{
int ret;
ret = image_setup_linux(images);
if (ret)
return ret;
return board_prep_linux(images);
}
/* Generic implementation for single core CPU */
__weak void board_jump_and_run(ulong entry, int zero, int arch, uint params)
{
void (*kernel_entry)(int zero, int arch, uint params);
kernel_entry = (void (*)(int, int, uint))entry;
kernel_entry(zero, arch, params);
}
/* Subcommand: GO */
static void boot_jump_linux(bootm_headers_t *images, int flag)
{
ulong kernel_entry;
unsigned int r0, r2;
int fake = (flag & BOOTM_STATE_OS_FAKE_GO);
kernel_entry = images->ep;
debug("## Transferring control to Linux (at address %08lx)...\n",
kernel_entry);
bootstage_mark(BOOTSTAGE_ID_RUN_OS);
printf("\nStarting kernel ...%s\n\n", fake ?
"(fake run for tracing)" : "");
bootstage_mark_name(BOOTSTAGE_ID_BOOTM_HANDOFF, "start_kernel");
if (IMAGE_ENABLE_OF_LIBFDT && images->ft_len) {
r0 = 2;
r2 = (unsigned int)images->ft_addr;
} else {
r0 = 1;
r2 = (unsigned int)env_get("bootargs");
}
cleanup_before_linux();
if (!fake)
board_jump_and_run(kernel_entry, r0, 0, r2);
}
int do_bootm_linux(int flag, int argc, char *argv[], bootm_headers_t *images)
{
/* No need for those on ARC */
if ((flag & BOOTM_STATE_OS_BD_T) || (flag & BOOTM_STATE_OS_CMDLINE))
return -1;
if (flag & BOOTM_STATE_OS_PREP)
return boot_prep_linux(images);
if (flag & (BOOTM_STATE_OS_GO | BOOTM_STATE_OS_FAKE_GO)) {
boot_jump_linux(images, flag);
return 0;
}
return -1;
}
@@ -0,0 +1,720 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
*/
#include <config.h>
#include <common.h>
#include <cpu_func.h>
#include <linux/compiler.h>
#include <linux/kernel.h>
#include <linux/log2.h>
#include <asm/arcregs.h>
#include <asm/arc-bcr.h>
#include <asm/cache.h>
/*
* [ NOTE 1 ]:
* Data cache (L1 D$ or SL$) entire invalidate operation or data cache disable
* operation may result in unexpected behavior and data loss even if we flush
* data cache right before invalidation. That may happens if we store any context
* on stack (like we store BLINK register on stack before function call).
* BLINK register is the register where return address is automatically saved
* when we do function call with instructions like 'bl'.
*
* There is the real example:
* We may hang in the next code as we store any BLINK register on stack in
* invalidate_dcache_all() function.
*
* void flush_dcache_all() {
* __dc_entire_op(OP_FLUSH);
* // Other code //
* }
*
* void invalidate_dcache_all() {
* __dc_entire_op(OP_INV);
* // Other code //
* }
*
* void foo(void) {
* flush_dcache_all();
* invalidate_dcache_all();
* }
*
* Now let's see what really happens during that code execution:
*
* foo()
* |->> call flush_dcache_all
* [return address is saved to BLINK register]
* [push BLINK] (save to stack) ![point 1]
* |->> call __dc_entire_op(OP_FLUSH)
* [return address is saved to BLINK register]
* [flush L1 D$]
* return [jump to BLINK]
* <<------
* [other flush_dcache_all code]
* [pop BLINK] (get from stack)
* return [jump to BLINK]
* <<------
* |->> call invalidate_dcache_all
* [return address is saved to BLINK register]
* [push BLINK] (save to stack) ![point 2]
* |->> call __dc_entire_op(OP_FLUSH)
* [return address is saved to BLINK register]
* [invalidate L1 D$] ![point 3]
* // Oops!!!
* // We lose return address from invalidate_dcache_all function:
* // we save it to stack and invalidate L1 D$ after that!
* return [jump to BLINK]
* <<------
* [other invalidate_dcache_all code]
* [pop BLINK] (get from stack)
* // we don't have this data in L1 dcache as we invalidated it in [point 3]
* // so we get it from next memory level (for example DDR memory)
* // but in the memory we have value which we save in [point 1], which
* // is return address from flush_dcache_all function (instead of
* // address from current invalidate_dcache_all function which we
* // saved in [point 2] !)
* return [jump to BLINK]
* <<------
* // As BLINK points to invalidate_dcache_all, we call it again and
* // loop forever.
*
* Fortunately we may fix that by using flush & invalidation of D$ with a single
* one instruction (instead of flush and invalidation instructions pair) and
* enabling force function inline with '__attribute__((always_inline))' gcc
* attribute to avoid any function call (and BLINK store) between cache flush
* and disable.
*
*
* [ NOTE 2 ]:
* As of today we only support the following cache configurations on ARC.
* Other configurations may exist in HW (for example, since version 3.0 HS
* supports SL$ (L2 system level cache) disable) but we don't support it in SW.
* Configuration 1:
* ______________________
* | |
* | ARC CPU |
* |______________________|
* ___|___ ___|___
* | | | |
* | L1 I$ | | L1 D$ |
* |_______| |_______|
* on/off on/off
* ___|______________|____
* | |
* | main memory |
* |______________________|
*
* Configuration 2:
* ______________________
* | |
* | ARC CPU |
* |______________________|
* ___|___ ___|___
* | | | |
* | L1 I$ | | L1 D$ |
* |_______| |_______|
* on/off on/off
* ___|______________|____
* | |
* | L2 (SL$) |
* |______________________|
* always must be on
* ___|______________|____
* | |
* | main memory |
* |______________________|
*
* Configuration 3:
* ______________________
* | |
* | ARC CPU |
* |______________________|
* ___|___ ___|___
* | | | |
* | L1 I$ | | L1 D$ |
* |_______| |_______|
* on/off must be on
* ___|______________|____ _______
* | | | |
* | L2 (SL$) |-----| IOC |
* |______________________| |_______|
* always must be on on/off
* ___|______________|____
* | |
* | main memory |
* |______________________|
*/
DECLARE_GLOBAL_DATA_PTR;
/* Bit values in IC_CTRL */
#define IC_CTRL_CACHE_DISABLE BIT(0)
/* Bit values in DC_CTRL */
#define DC_CTRL_CACHE_DISABLE BIT(0)
#define DC_CTRL_INV_MODE_FLUSH BIT(6)
#define DC_CTRL_FLUSH_STATUS BIT(8)
#define OP_INV BIT(0)
#define OP_FLUSH BIT(1)
#define OP_FLUSH_N_INV (OP_FLUSH | OP_INV)
/* Bit val in SLC_CONTROL */
#define SLC_CTRL_DIS 0x001
#define SLC_CTRL_IM 0x040
#define SLC_CTRL_BUSY 0x100
#define SLC_CTRL_RGN_OP_INV 0x200
#define CACHE_LINE_MASK (~(gd->arch.l1_line_sz - 1))
/*
* We don't want to use '__always_inline' macro here as it can be redefined
* to simple 'inline' in some cases which breaks stuff. See [ NOTE 1 ] for more
* details about the reasons we need to use always_inline functions.
*/
#define inlined_cachefunc inline __attribute__((always_inline))
static inlined_cachefunc void __ic_entire_invalidate(void);
static inlined_cachefunc void __dc_entire_op(const int cacheop);
static inline bool pae_exists(void)
{
/* TODO: should we compare mmu version from BCR and from CONFIG? */
#if (CONFIG_ARC_MMU_VER >= 4)
union bcr_mmu_4 mmu4;
mmu4.word = read_aux_reg(ARC_AUX_MMU_BCR);
if (mmu4.fields.pae)
return true;
#endif /* (CONFIG_ARC_MMU_VER >= 4) */
return false;
}
static inlined_cachefunc bool icache_exists(void)
{
union bcr_di_cache ibcr;
ibcr.word = read_aux_reg(ARC_BCR_IC_BUILD);
return !!ibcr.fields.ver;
}
static inlined_cachefunc bool icache_enabled(void)
{
if (!icache_exists())
return false;
return !(read_aux_reg(ARC_AUX_IC_CTRL) & IC_CTRL_CACHE_DISABLE);
}
static inlined_cachefunc bool dcache_exists(void)
{
union bcr_di_cache dbcr;
dbcr.word = read_aux_reg(ARC_BCR_DC_BUILD);
return !!dbcr.fields.ver;
}
static inlined_cachefunc bool dcache_enabled(void)
{
if (!dcache_exists())
return false;
return !(read_aux_reg(ARC_AUX_DC_CTRL) & DC_CTRL_CACHE_DISABLE);
}
static inlined_cachefunc bool slc_exists(void)
{
if (is_isa_arcv2()) {
union bcr_generic sbcr;
sbcr.word = read_aux_reg(ARC_BCR_SLC);
return !!sbcr.fields.ver;
}
return false;
}
static inlined_cachefunc bool slc_data_bypass(void)
{
/*
* If L1 data cache is disabled SL$ is bypassed and all load/store
* requests are sent directly to main memory.
*/
return !dcache_enabled();
}
static inline bool ioc_exists(void)
{
if (is_isa_arcv2()) {
union bcr_clust_cfg cbcr;
cbcr.word = read_aux_reg(ARC_BCR_CLUSTER);
return cbcr.fields.c;
}
return false;
}
static inline bool ioc_enabled(void)
{
/*
* We check only CONFIG option instead of IOC HW state check as IOC
* must be disabled by default.
*/
if (is_ioc_enabled())
return ioc_exists();
return false;
}
static inlined_cachefunc void __slc_entire_op(const int op)
{
unsigned int ctrl;
if (!slc_exists())
return;
ctrl = read_aux_reg(ARC_AUX_SLC_CTRL);
if (!(op & OP_FLUSH)) /* i.e. OP_INV */
ctrl &= ~SLC_CTRL_IM; /* clear IM: Disable flush before Inv */
else
ctrl |= SLC_CTRL_IM;
write_aux_reg(ARC_AUX_SLC_CTRL, ctrl);
if (op & OP_INV) /* Inv or flush-n-inv use same cmd reg */
write_aux_reg(ARC_AUX_SLC_INVALIDATE, 0x1);
else
write_aux_reg(ARC_AUX_SLC_FLUSH, 0x1);
/* Make sure "busy" bit reports correct stataus, see STAR 9001165532 */
read_aux_reg(ARC_AUX_SLC_CTRL);
/* Important to wait for flush to complete */
while (read_aux_reg(ARC_AUX_SLC_CTRL) & SLC_CTRL_BUSY);
}
static void slc_upper_region_init(void)
{
/*
* ARC_AUX_SLC_RGN_START1 and ARC_AUX_SLC_RGN_END1 register exist
* only if PAE exists in current HW. So we had to check pae_exist
* before using them.
*/
if (!pae_exists())
return;
/*
* ARC_AUX_SLC_RGN_END1 and ARC_AUX_SLC_RGN_START1 are always == 0
* as we don't use PAE40.
*/
write_aux_reg(ARC_AUX_SLC_RGN_END1, 0);
write_aux_reg(ARC_AUX_SLC_RGN_START1, 0);
}
static void __slc_rgn_op(unsigned long paddr, unsigned long sz, const int op)
{
#ifdef CONFIG_ISA_ARCV2
unsigned int ctrl;
unsigned long end;
if (!slc_exists())
return;
/*
* The Region Flush operation is specified by CTRL.RGN_OP[11..9]
* - b'000 (default) is Flush,
* - b'001 is Invalidate if CTRL.IM == 0
* - b'001 is Flush-n-Invalidate if CTRL.IM == 1
*/
ctrl = read_aux_reg(ARC_AUX_SLC_CTRL);
/* Don't rely on default value of IM bit */
if (!(op & OP_FLUSH)) /* i.e. OP_INV */
ctrl &= ~SLC_CTRL_IM; /* clear IM: Disable flush before Inv */
else
ctrl |= SLC_CTRL_IM;
if (op & OP_INV)
ctrl |= SLC_CTRL_RGN_OP_INV; /* Inv or flush-n-inv */
else
ctrl &= ~SLC_CTRL_RGN_OP_INV;
write_aux_reg(ARC_AUX_SLC_CTRL, ctrl);
/*
* Lower bits are ignored, no need to clip
* END needs to be setup before START (latter triggers the operation)
* END can't be same as START, so add (l2_line_sz - 1) to sz
*/
end = paddr + sz + gd->arch.slc_line_sz - 1;
/*
* Upper addresses (ARC_AUX_SLC_RGN_END1 and ARC_AUX_SLC_RGN_START1)
* are always == 0 as we don't use PAE40, so we only setup lower ones
* (ARC_AUX_SLC_RGN_END and ARC_AUX_SLC_RGN_START)
*/
write_aux_reg(ARC_AUX_SLC_RGN_END, end);
write_aux_reg(ARC_AUX_SLC_RGN_START, paddr);
/* Make sure "busy" bit reports correct stataus, see STAR 9001165532 */
read_aux_reg(ARC_AUX_SLC_CTRL);
while (read_aux_reg(ARC_AUX_SLC_CTRL) & SLC_CTRL_BUSY);
#endif /* CONFIG_ISA_ARCV2 */
}
static void arc_ioc_setup(void)
{
/* IOC Aperture start is equal to DDR start */
unsigned int ap_base = CONFIG_SYS_SDRAM_BASE;
/* IOC Aperture size is equal to DDR size */
long ap_size = CONFIG_SYS_SDRAM_SIZE;
/* Unsupported configuration. See [ NOTE 2 ] for more details. */
if (!slc_exists())
panic("Try to enable IOC but SLC is not present");
/* Unsupported configuration. See [ NOTE 2 ] for more details. */
if (!dcache_enabled())
panic("Try to enable IOC but L1 D$ is disabled");
if (!is_power_of_2(ap_size) || ap_size < 4096)
panic("IOC Aperture size must be power of 2 and bigger 4Kib");
/* IOC Aperture start must be aligned to the size of the aperture */
if (ap_base % ap_size != 0)
panic("IOC Aperture start must be aligned to the size of the aperture");
flush_n_invalidate_dcache_all();
/*
* IOC Aperture size decoded as 2 ^ (SIZE + 2) KB,
* so setting 0x11 implies 512M, 0x12 implies 1G...
*/
write_aux_reg(ARC_AUX_IO_COH_AP0_SIZE,
order_base_2(ap_size / 1024) - 2);
write_aux_reg(ARC_AUX_IO_COH_AP0_BASE, ap_base >> 12);
write_aux_reg(ARC_AUX_IO_COH_PARTIAL, 1);
write_aux_reg(ARC_AUX_IO_COH_ENABLE, 1);
}
static void read_decode_cache_bcr_arcv2(void)
{
#ifdef CONFIG_ISA_ARCV2
union bcr_slc_cfg slc_cfg;
if (slc_exists()) {
slc_cfg.word = read_aux_reg(ARC_AUX_SLC_CONFIG);
gd->arch.slc_line_sz = (slc_cfg.fields.lsz == 0) ? 128 : 64;
/*
* We don't support configuration where L1 I$ or L1 D$ is
* absent but SL$ exists. See [ NOTE 2 ] for more details.
*/
if (!icache_exists() || !dcache_exists())
panic("Unsupported cache configuration: SLC exists but one of L1 caches is absent");
}
#endif /* CONFIG_ISA_ARCV2 */
}
void read_decode_cache_bcr(void)
{
int dc_line_sz = 0, ic_line_sz = 0;
union bcr_di_cache ibcr, dbcr;
/*
* We don't care much about I$ line length really as there're
* no per-line ops on I$ instead we only do full invalidation of it
* on occasion of relocation and right before jumping to the OS.
* Still we check insane config with zero-encoded line length in
* presense of version field in I$ BCR. Just in case.
*/
ibcr.word = read_aux_reg(ARC_BCR_IC_BUILD);
if (ibcr.fields.ver) {
ic_line_sz = 8 << ibcr.fields.line_len;
if (!ic_line_sz)
panic("Instruction exists but line length is 0\n");
}
dbcr.word = read_aux_reg(ARC_BCR_DC_BUILD);
if (dbcr.fields.ver) {
gd->arch.l1_line_sz = dc_line_sz = 16 << dbcr.fields.line_len;
if (!dc_line_sz)
panic("Data cache exists but line length is 0\n");
}
}
void cache_init(void)
{
read_decode_cache_bcr();
if (is_isa_arcv2())
read_decode_cache_bcr_arcv2();
if (is_isa_arcv2() && ioc_enabled())
arc_ioc_setup();
if (is_isa_arcv2() && slc_exists())
slc_upper_region_init();
}
int icache_status(void)
{
return icache_enabled();
}
void icache_enable(void)
{
if (icache_exists())
write_aux_reg(ARC_AUX_IC_CTRL, read_aux_reg(ARC_AUX_IC_CTRL) &
~IC_CTRL_CACHE_DISABLE);
}
void icache_disable(void)
{
if (!icache_exists())
return;
__ic_entire_invalidate();
write_aux_reg(ARC_AUX_IC_CTRL, read_aux_reg(ARC_AUX_IC_CTRL) |
IC_CTRL_CACHE_DISABLE);
}
/* IC supports only invalidation */
static inlined_cachefunc void __ic_entire_invalidate(void)
{
if (!icache_enabled())
return;
/* Any write to IC_IVIC register triggers invalidation of entire I$ */
write_aux_reg(ARC_AUX_IC_IVIC, 1);
/*
* As per ARC HS databook (see chapter 5.3.3.2)
* it is required to add 3 NOPs after each write to IC_IVIC.
*/
__builtin_arc_nop();
__builtin_arc_nop();
__builtin_arc_nop();
read_aux_reg(ARC_AUX_IC_CTRL); /* blocks */
}
void invalidate_icache_all(void)
{
__ic_entire_invalidate();
/*
* If SL$ is bypassed for data it is used only for instructions,
* so we need to invalidate it too.
* TODO: HS 3.0 supports SLC disable so we need to check slc
* enable/disable status here.
*/
if (is_isa_arcv2() && slc_data_bypass())
__slc_entire_op(OP_INV);
}
int dcache_status(void)
{
return dcache_enabled();
}
void dcache_enable(void)
{
if (!dcache_exists())
return;
write_aux_reg(ARC_AUX_DC_CTRL, read_aux_reg(ARC_AUX_DC_CTRL) &
~(DC_CTRL_INV_MODE_FLUSH | DC_CTRL_CACHE_DISABLE));
}
void dcache_disable(void)
{
if (!dcache_exists())
return;
__dc_entire_op(OP_FLUSH_N_INV);
/*
* As SLC will be bypassed for data after L1 D$ disable we need to
* flush it first before L1 D$ disable. Also we invalidate SLC to
* avoid any inconsistent data problems after enabling L1 D$ again with
* dcache_enable function.
*/
if (is_isa_arcv2())
__slc_entire_op(OP_FLUSH_N_INV);
write_aux_reg(ARC_AUX_DC_CTRL, read_aux_reg(ARC_AUX_DC_CTRL) |
DC_CTRL_CACHE_DISABLE);
}
/* Common Helper for Line Operations on D-cache */
static inline void __dcache_line_loop(unsigned long paddr, unsigned long sz,
const int cacheop)
{
unsigned int aux_cmd;
int num_lines;
/* d$ cmd: INV (discard or wback-n-discard) OR FLUSH (wback) */
aux_cmd = cacheop & OP_INV ? ARC_AUX_DC_IVDL : ARC_AUX_DC_FLDL;
sz += paddr & ~CACHE_LINE_MASK;
paddr &= CACHE_LINE_MASK;
num_lines = DIV_ROUND_UP(sz, gd->arch.l1_line_sz);
while (num_lines-- > 0) {
#if (CONFIG_ARC_MMU_VER == 3)
write_aux_reg(ARC_AUX_DC_PTAG, paddr);
#endif
write_aux_reg(aux_cmd, paddr);
paddr += gd->arch.l1_line_sz;
}
}
static inlined_cachefunc void __before_dc_op(const int op)
{
unsigned int ctrl;
ctrl = read_aux_reg(ARC_AUX_DC_CTRL);
/* IM bit implies flush-n-inv, instead of vanilla inv */
if (op == OP_INV)
ctrl &= ~DC_CTRL_INV_MODE_FLUSH;
else
ctrl |= DC_CTRL_INV_MODE_FLUSH;
write_aux_reg(ARC_AUX_DC_CTRL, ctrl);
}
static inlined_cachefunc void __after_dc_op(const int op)
{
if (op & OP_FLUSH) /* flush / flush-n-inv both wait */
while (read_aux_reg(ARC_AUX_DC_CTRL) & DC_CTRL_FLUSH_STATUS);
}
static inlined_cachefunc void __dc_entire_op(const int cacheop)
{
int aux;
if (!dcache_enabled())
return;
__before_dc_op(cacheop);
if (cacheop & OP_INV) /* Inv or flush-n-inv use same cmd reg */
aux = ARC_AUX_DC_IVDC;
else
aux = ARC_AUX_DC_FLSH;
write_aux_reg(aux, 0x1);
__after_dc_op(cacheop);
}
static inline void __dc_line_op(unsigned long paddr, unsigned long sz,
const int cacheop)
{
if (!dcache_enabled())
return;
__before_dc_op(cacheop);
__dcache_line_loop(paddr, sz, cacheop);
__after_dc_op(cacheop);
}
void invalidate_dcache_range(unsigned long start, unsigned long end)
{
if (start >= end)
return;
/*
* ARCv1 -> call __dc_line_op
* ARCv2 && L1 D$ disabled -> nothing
* ARCv2 && L1 D$ enabled && IOC enabled -> nothing
* ARCv2 && L1 D$ enabled && no IOC -> call __dc_line_op; call __slc_rgn_op
*/
if (!is_isa_arcv2() || !ioc_enabled())
__dc_line_op(start, end - start, OP_INV);
if (is_isa_arcv2() && !ioc_enabled() && !slc_data_bypass())
__slc_rgn_op(start, end - start, OP_INV);
}
void flush_dcache_range(unsigned long start, unsigned long end)
{
if (start >= end)
return;
/*
* ARCv1 -> call __dc_line_op
* ARCv2 && L1 D$ disabled -> nothing
* ARCv2 && L1 D$ enabled && IOC enabled -> nothing
* ARCv2 && L1 D$ enabled && no IOC -> call __dc_line_op; call __slc_rgn_op
*/
if (!is_isa_arcv2() || !ioc_enabled())
__dc_line_op(start, end - start, OP_FLUSH);
if (is_isa_arcv2() && !ioc_enabled() && !slc_data_bypass())
__slc_rgn_op(start, end - start, OP_FLUSH);
}
void flush_cache(unsigned long start, unsigned long size)
{
flush_dcache_range(start, start + size);
}
/*
* As invalidate_dcache_all() is not used in generic U-Boot code and as we
* don't need it in arch/arc code alone (invalidate without flush) we implement
* flush_n_invalidate_dcache_all (flush and invalidate in 1 operation) because
* it's much safer. See [ NOTE 1 ] for more details.
*/
void flush_n_invalidate_dcache_all(void)
{
__dc_entire_op(OP_FLUSH_N_INV);
if (is_isa_arcv2() && !slc_data_bypass())
__slc_entire_op(OP_FLUSH_N_INV);
}
void flush_dcache_all(void)
{
__dc_entire_op(OP_FLUSH);
if (is_isa_arcv2() && !slc_data_bypass())
__slc_entire_op(OP_FLUSH);
}
/*
* This is function to cleanup all caches (and therefore sync I/D caches) which
* can be used for cleanup before linux launch or to sync caches during
* relocation.
*/
void sync_n_cleanup_cache_all(void)
{
__dc_entire_op(OP_FLUSH_N_INV);
/*
* If SL$ is bypassed for data it is used only for instructions,
* and we shouldn't flush it. So invalidate it instead of flush_n_inv.
*/
if (is_isa_arcv2()) {
if (slc_data_bypass())
__slc_entire_op(OP_INV);
else
__slc_entire_op(OP_FLUSH_N_INV);
}
__ic_entire_invalidate();
}
@@ -0,0 +1,229 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2014, 2018 Synopsys, Inc. All rights reserved.
*/
#include <common.h>
#include <malloc.h>
#include <vsprintf.h>
#include <asm/arcregs.h>
#include <asm/cache.h>
DECLARE_GLOBAL_DATA_PTR;
int arch_cpu_init(void)
{
timer_init();
gd->cpu_clk = CONFIG_SYS_CLK_FREQ;
gd->ram_size = CONFIG_SYS_SDRAM_SIZE;
cache_init();
return 0;
}
int arch_early_init_r(void)
{
gd->bd->bi_memstart = CONFIG_SYS_SDRAM_BASE;
gd->bd->bi_memsize = CONFIG_SYS_SDRAM_SIZE;
return 0;
}
/* This is a dummy function on arc */
int dram_init(void)
{
return 0;
}
#ifdef CONFIG_DISPLAY_CPUINFO
const char *arc_700_version(int arcver, char *name, int name_len)
{
const char *arc_ver;
switch (arcver) {
case 0x32:
arc_ver = "v4.4-4.5";
break;
case 0x33:
arc_ver = "v4.6-v4.9";
break;
case 0x34:
arc_ver = "v4.10";
break;
case 0x35:
arc_ver = "v4.11";
break;
default:
arc_ver = "unknown version";
}
snprintf(name, name_len, "ARC 700 %s", arc_ver);
return name;
}
struct em_template_t {
const bool cache;
const bool dsp;
const bool xymem;
const char name[8];
};
static const struct em_template_t em_versions[] = {
{false, false, false, "EM4"},
{true, false, false, "EM6"},
{false, true, false, "EM5D"},
{true, true, false, "EM7D"},
{false, true, true, "EM9D"},
{true, true, true, "EM11D"},
};
const char *arc_em_version(int arcver, char *name, int name_len)
{
const char *arc_name = "EM";
const char *arc_ver;
bool cache = ARC_FEATURE_EXISTS(ARC_BCR_IC_BUILD);
bool dsp = ARC_FEATURE_EXISTS(ARC_AUX_DSP_BUILD);
bool xymem = ARC_FEATURE_EXISTS(ARC_AUX_XY_BUILD);
int i;
for (i = 0; i < sizeof(em_versions) / sizeof(struct em_template_t); i++) {
if (em_versions[i].cache == cache &&
em_versions[i].dsp == dsp &&
em_versions[i].xymem == xymem) {
arc_name = em_versions[i].name;
break;
}
}
switch (arcver) {
case 0x41:
arc_ver = "v1.1a";
break;
case 0x42:
arc_ver = "v3.0";
break;
case 0x43:
arc_ver = "v4.0";
break;
case 0x44:
arc_ver = "v5.0";
break;
default:
arc_ver = "unknown version";
}
snprintf(name, name_len, "ARC %s %s", arc_name, arc_ver);
return name;
}
struct hs_template_t {
const bool cache;
const bool mmu;
const bool dual_issue;
const bool dsp;
const char name[8];
};
static const struct hs_template_t hs_versions[] = {
{false, false, false, false, "HS34"},
{true, false, false, false, "HS36"},
{true, true, false, false, "HS38"},
{false, false, true, false, "HS44"},
{true, false, true, false, "HS46"},
{true, true, true, false, "HS48"},
{false, false, true, true, "HS45D"},
{true, false, true, true, "HS47D"},
};
const char *arc_hs_version(int arcver, char *name, int name_len)
{
const char *arc_name = "HS";
const char *arc_ver;
bool cache = ARC_FEATURE_EXISTS(ARC_BCR_IC_BUILD);
bool dsp = ARC_FEATURE_EXISTS(ARC_AUX_DSP_BUILD);
bool mmu = !!read_aux_reg(ARC_AUX_MMU_BCR);
bool dual_issue = arcver == 0x54 ? true : false;
int i;
for (i = 0; i < sizeof(hs_versions) / sizeof(struct hs_template_t); i++) {
if (hs_versions[i].cache == cache &&
hs_versions[i].mmu == mmu &&
hs_versions[i].dual_issue == dual_issue &&
hs_versions[i].dsp == dsp) {
arc_name = hs_versions[i].name;
break;
}
}
switch (arcver) {
case 0x50:
arc_ver = "v1.0";
break;
case 0x51:
arc_ver = "v2.0";
break;
case 0x52:
arc_ver = "v2.1c";
break;
case 0x53:
arc_ver = "v3.0";
break;
case 0x54:
arc_ver = "v4.0";
break;
default:
arc_ver = "unknown version";
}
snprintf(name, name_len, "ARC %s %s", arc_name, arc_ver);
return name;
}
const char *decode_identity(void)
{
#define MAX_CPU_NAME_LEN 64
int arcver = read_aux_reg(ARC_AUX_IDENTITY) & 0xff;
char *name = malloc(MAX_CPU_NAME_LEN);
if (arcver >= 0x50)
return arc_hs_version(arcver, name, MAX_CPU_NAME_LEN);
else if (arcver >= 0x40)
return arc_em_version(arcver, name, MAX_CPU_NAME_LEN);
else if (arcver >= 0x30)
return arc_700_version(arcver, name, MAX_CPU_NAME_LEN);
else
return "Unknown ARC core";
}
const char *decode_subsystem(void)
{
int subsys_type = read_aux_reg(ARC_AUX_SUBSYS_BUILD) & GENMASK(3, 0);
switch (subsys_type) {
case 0: return NULL;
case 2: return "ARC Sensor & Control IP Subsystem";
case 3: return "ARC Data Fusion IP Subsystem";
case 4: return "ARC Secure Subsystem";
default: return "Unknown subsystem";
};
}
__weak int print_cpuinfo(void)
{
const char *subsys_name = decode_subsystem();
char mhz[8];
printf("CPU: %s at %s MHz\n", decode_identity(),
strmhz(mhz, gd->cpu_clk));
if (subsys_name)
printf("Subsys:%s\n", subsys_name);
return 0;
}
#endif /* CONFIG_DISPLAY_CPUINFO */
@@ -0,0 +1,14 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2015 Synopsys, Inc. All rights reserved.
*/
#include <asm/cache.h>
#include <common.h>
int init_cache_f_r(void)
{
sync_n_cleanup_cache_all();
return 0;
}
@@ -0,0 +1,169 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
*/
#include <common.h>
#include <irq_func.h>
#include <asm/arcregs.h>
#include <asm/ptrace.h>
/* Bit values in STATUS32 */
#define E1_MASK (1 << 1) /* Level 1 interrupts enable */
#define E2_MASK (1 << 2) /* Level 2 interrupts enable */
int interrupt_init(void)
{
return 0;
}
/*
* returns true if interrupts had been enabled before we disabled them
*/
int disable_interrupts(void)
{
int status = read_aux_reg(ARC_AUX_STATUS32);
int state = (status & (E1_MASK | E2_MASK)) ? 1 : 0;
status &= ~(E1_MASK | E2_MASK);
/* STATUS32 register is updated indirectly with "FLAG" instruction */
__asm__("flag %0" : : "r" (status));
return state;
}
void enable_interrupts(void)
{
unsigned int status = read_aux_reg(ARC_AUX_STATUS32);
status |= E1_MASK | E2_MASK;
/* STATUS32 register is updated indirectly with "FLAG" instruction */
__asm__("flag %0" : : "r" (status));
}
static void print_reg_file(long *reg_rev, int start_num)
{
unsigned int i;
/* Print 3 registers per line */
for (i = start_num; i < start_num + 25; i++) {
printf("r%02u: 0x%08lx\t", i, (unsigned long)*reg_rev);
if (((i + 1) % 3) == 0)
printf("\n");
/* Because pt_regs has registers reversed */
reg_rev--;
}
/* Add new-line if none was inserted in the end of loop above */
if (((i + 1) % 3) != 0)
printf("\n");
}
void show_regs(struct pt_regs *regs)
{
printf("ECR:\t0x%08lx\n", regs->ecr);
printf("RET:\t0x%08lx\nBLINK:\t0x%08lx\nSTAT32:\t0x%08lx\n",
regs->ret, regs->blink, regs->status32);
printf("GP: 0x%08lx\t r25: 0x%08lx\t\n", regs->r26, regs->r25);
printf("BTA: 0x%08lx\t SP: 0x%08lx\t FP: 0x%08lx\n", regs->bta,
regs->sp, regs->fp);
printf("LPS: 0x%08lx\tLPE: 0x%08lx\tLPC: 0x%08lx\n", regs->lp_start,
regs->lp_end, regs->lp_count);
print_reg_file(&(regs->r0), 0);
}
void bad_mode(struct pt_regs *regs)
{
if (regs)
show_regs(regs);
panic("Resetting CPU ...\n");
}
void do_memory_error(unsigned long address, struct pt_regs *regs)
{
printf("Memory error exception @ 0x%lx\n", address);
bad_mode(regs);
}
void do_instruction_error(unsigned long address, struct pt_regs *regs)
{
printf("Instruction error exception @ 0x%lx\n", address);
bad_mode(regs);
}
void do_machine_check_fault(unsigned long address, struct pt_regs *regs)
{
printf("Machine check exception @ 0x%lx\n", address);
bad_mode(regs);
}
void do_interrupt_handler(void)
{
printf("Interrupt fired\n");
bad_mode(0);
}
void do_itlb_miss(struct pt_regs *regs)
{
printf("I TLB miss exception\n");
bad_mode(regs);
}
void do_dtlb_miss(struct pt_regs *regs)
{
printf("D TLB miss exception\n");
bad_mode(regs);
}
void do_tlb_prot_violation(unsigned long address, struct pt_regs *regs)
{
printf("TLB protection violation or misaligned access @ 0x%lx\n",
address);
bad_mode(regs);
}
void do_privilege_violation(struct pt_regs *regs)
{
printf("Privilege violation exception\n");
bad_mode(regs);
}
void do_trap(struct pt_regs *regs)
{
printf("Trap exception\n");
bad_mode(regs);
}
void do_extension(struct pt_regs *regs)
{
printf("Extension instruction exception\n");
bad_mode(regs);
}
#ifdef CONFIG_ISA_ARCV2
void do_swi(struct pt_regs *regs)
{
printf("Software Interrupt exception\n");
bad_mode(regs);
}
void do_divzero(unsigned long address, struct pt_regs *regs)
{
printf("Division by zero exception @ 0x%lx\n", address);
bad_mode(regs);
}
void do_dcerror(struct pt_regs *regs)
{
printf("Data cache consistency error exception\n");
bad_mode(regs);
}
void do_maligned(unsigned long address, struct pt_regs *regs)
{
printf("Misaligned data access exception @ 0x%lx\n", address);
bad_mode(regs);
}
#endif
@@ -0,0 +1,178 @@
/* SPDX-License-Identifier: GPL-2.0+ */
/*
* Copyright (C) 2013-2015 Synopsys, Inc. All rights reserved.
*/
#include <linux/linkage.h>
/*
* Note on the LD/ST addressing modes with address register write-back
*
* LD.a same as LD.aw
*
* LD.a reg1, [reg2, x] => Pre Incr
* Eff Addr for load = [reg2 + x]
*
* LD.ab reg1, [reg2, x] => Post Incr
* Eff Addr for load = [reg2]
*/
.macro PUSH reg
st.a \reg, [%sp, -4]
.endm
.macro PUSHAX aux
lr %r9, [\aux]
PUSH %r9
.endm
.macro SAVE_R1_TO_R24
PUSH %r1
PUSH %r2
PUSH %r3
PUSH %r4
PUSH %r5
PUSH %r6
PUSH %r7
PUSH %r8
PUSH %r9
PUSH %r10
PUSH %r11
PUSH %r12
PUSH %r13
PUSH %r14
PUSH %r15
PUSH %r16
PUSH %r17
PUSH %r18
PUSH %r19
PUSH %r20
PUSH %r21
PUSH %r22
PUSH %r23
PUSH %r24
.endm
.macro SAVE_ALL_SYS
/* saving %r0 to reg->r0 in advance since we read %ecr into it */
st %r0, [%sp, -8]
lr %r0, [%ecr] /* all stack addressing is manual so far */
st %r0, [%sp]
st %sp, [%sp, -4]
/* now move %sp to reg->r0 position so we can do "push" automatically */
sub %sp, %sp, 8
SAVE_R1_TO_R24
PUSH %r25
PUSH %gp
PUSH %fp
PUSH %blink
PUSHAX %eret
PUSHAX %erstatus
PUSH %lp_count
PUSHAX %lp_end
PUSHAX %lp_start
PUSHAX %erbta
.endm
.macro SAVE_EXCEPTION_SOURCE
#ifdef CONFIG_MMU
/* If MMU exists exception faulting address is loaded in EFA reg */
lr %r0, [%efa]
#else
/* Otherwise in ERET (exception return) reg */
lr %r0, [%eret]
#endif
.endm
ENTRY(memory_error)
SAVE_ALL_SYS
SAVE_EXCEPTION_SOURCE
mov %r1, %sp
j do_memory_error
ENDPROC(memory_error)
ENTRY(instruction_error)
SAVE_ALL_SYS
SAVE_EXCEPTION_SOURCE
mov %r1, %sp
j do_instruction_error
ENDPROC(instruction_error)
ENTRY(interrupt_handler)
/* Todo - save and restore CPU context when interrupts will be in use */
bl do_interrupt_handler
rtie
ENDPROC(interrupt_handler)
ENTRY(EV_MachineCheck)
SAVE_ALL_SYS
SAVE_EXCEPTION_SOURCE
mov %r1, %sp
j do_machine_check_fault
ENDPROC(EV_MachineCheck)
ENTRY(EV_TLBMissI)
SAVE_ALL_SYS
mov %r0, %sp
j do_itlb_miss
ENDPROC(EV_TLBMissI)
ENTRY(EV_TLBMissD)
SAVE_ALL_SYS
mov %r0, %sp
j do_dtlb_miss
ENDPROC(EV_TLBMissD)
ENTRY(EV_TLBProtV)
SAVE_ALL_SYS
SAVE_EXCEPTION_SOURCE
mov %r1, %sp
j do_tlb_prot_violation
ENDPROC(EV_TLBProtV)
ENTRY(EV_PrivilegeV)
SAVE_ALL_SYS
mov %r0, %sp
j do_privilege_violation
ENDPROC(EV_PrivilegeV)
ENTRY(EV_Trap)
SAVE_ALL_SYS
mov %r0, %sp
j do_trap
ENDPROC(EV_Trap)
ENTRY(EV_Extension)
SAVE_ALL_SYS
mov %r0, %sp
j do_extension
ENDPROC(EV_Extension)
#ifdef CONFIG_ISA_ARCV2
ENTRY(EV_SWI)
SAVE_ALL_SYS
mov %r0, %sp
j do_swi
ENDPROC(EV_SWI)
ENTRY(EV_DivZero)
SAVE_ALL_SYS
SAVE_EXCEPTION_SOURCE
mov %r1, %sp
j do_divzero
ENDPROC(EV_DivZero)
ENTRY(EV_DCError)
SAVE_ALL_SYS
mov %r0, %sp
j do_dcerror
ENDPROC(EV_DCError)
ENTRY(EV_Maligned)
SAVE_ALL_SYS
SAVE_EXCEPTION_SOURCE
mov %r1, %sp
j do_maligned
ENDPROC(EV_Maligned)
#endif
@@ -0,0 +1,235 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 1989-2013 Free Software Foundation, Inc.
*/
#include "libgcc2.h"
DWtype
__ashldi3(DWtype u, shift_count_type b)
{
if (b == 0)
return u;
const DWunion uu = {.ll = u};
const shift_count_type bm = W_TYPE_SIZE - b;
DWunion w;
if (bm <= 0) {
w.s.low = 0;
w.s.high = (UWtype)uu.s.low << -bm;
} else {
const UWtype carries = (UWtype) uu.s.low >> bm;
w.s.low = (UWtype)uu.s.low << b;
w.s.high = ((UWtype)uu.s.high << b) | carries;
}
return w.ll;
}
DWtype
__ashrdi3(DWtype u, shift_count_type b)
{
if (b == 0)
return u;
const DWunion uu = {.ll = u};
const shift_count_type bm = W_TYPE_SIZE - b;
DWunion w;
if (bm <= 0) {
/* w.s.high = 1..1 or 0..0 */
w.s.high = uu.s.high >> (W_TYPE_SIZE - 1);
w.s.low = uu.s.high >> -bm;
} else {
const UWtype carries = (UWtype) uu.s.high << bm;
w.s.high = uu.s.high >> b;
w.s.low = ((UWtype)uu.s.low >> b) | carries;
}
return w.ll;
}
DWtype
__lshrdi3(DWtype u, shift_count_type b)
{
if (b == 0)
return u;
const DWunion uu = {.ll = u};
const shift_count_type bm = W_TYPE_SIZE - b;
DWunion w;
if (bm <= 0) {
w.s.high = 0;
w.s.low = (UWtype)uu.s.high >> -bm;
} else {
const UWtype carries = (UWtype)uu.s.high << bm;
w.s.high = (UWtype)uu.s.high >> b;
w.s.low = ((UWtype)uu.s.low >> b) | carries;
}
return w.ll;
}
unsigned long
udivmodsi4(unsigned long num, unsigned long den, int modwanted)
{
unsigned long bit = 1;
unsigned long res = 0;
while (den < num && bit && !(den & (1L<<31))) {
den <<= 1;
bit <<= 1;
}
while (bit) {
if (num >= den) {
num -= den;
res |= bit;
}
bit >>= 1;
den >>= 1;
}
if (modwanted)
return num;
return res;
}
long
__divsi3(long a, long b)
{
int neg = 0;
long res;
if (a < 0) {
a = -a;
neg = !neg;
}
if (b < 0) {
b = -b;
neg = !neg;
}
res = udivmodsi4(a, b, 0);
if (neg)
res = -res;
return res;
}
long
__modsi3(long a, long b)
{
int neg = 0;
long res;
if (a < 0) {
a = -a;
neg = 1;
}
if (b < 0)
b = -b;
res = udivmodsi4(a, b, 1);
if (neg)
res = -res;
return res;
}
long
__udivsi3(long a, long b)
{
return udivmodsi4(a, b, 0);
}
long
__umodsi3(long a, long b)
{
return udivmodsi4(a, b, 1);
}
UDWtype
__udivmoddi4(UDWtype n, UDWtype d, UDWtype *rp)
{
UDWtype q = 0, r = n, y = d;
UWtype lz1, lz2, i, k;
/*
* Implements align divisor shift dividend method. This algorithm
* aligns the divisor under the dividend and then perform number of
* test-subtract iterations which shift the dividend left. Number of
* iterations is k + 1 where k is the number of bit positions the
* divisor must be shifted left to align it under the dividend.
* quotient bits can be saved in the rightmost positions of the
* dividend as it shifts left on each test-subtract iteration.
*/
if (y <= r) {
lz1 = __builtin_clzll(d);
lz2 = __builtin_clzll(n);
k = lz1 - lz2;
y = (y << k);
/*
* Dividend can exceed 2 ^ (width - 1) - 1 but still be less
* than the aligned divisor. Normal iteration can drops the
* high order bit of the dividend. Therefore, first
* test-subtract iteration is a special case, saving its
* quotient bit in a separate location and not shifting
* the dividend.
*/
if (r >= y) {
r = r - y;
q = (1ULL << k);
}
if (k > 0) {
y = y >> 1;
/*
* k additional iterations where k regular test
* subtract shift dividend iterations are done.
*/
i = k;
do {
if (r >= y)
r = ((r - y) << 1) + 1;
else
r = (r << 1);
i = i - 1;
} while (i != 0);
/*
* First quotient bit is combined with the quotient
* bits resulting from the k regular iterations.
*/
q = q + r;
r = r >> k;
q = q - (r << k);
}
}
if (rp)
*rp = r;
return q;
}
UDWtype
__udivdi3(UDWtype n, UDWtype d)
{
return __udivmoddi4(n, d, (UDWtype *)0);
}
@@ -0,0 +1,131 @@
/* SPDX-License-Identifier: GPL-2.0+ */
/*
* Copyright (C) 1989-2013 Free Software Foundation, Inc.
*/
#ifndef __ASM_LIBGCC_H
#define __ASM_LIBGCC_H
#define UNITS_PER_WORD 4 /* for ARC */
#define BITS_PER_UNIT 8 /* for ARC */
#define W_TYPE_SIZE (4 * BITS_PER_UNIT)
#define MIN_UNITS_PER_WORD UNITS_PER_WORD
/* Work out the largest "word" size that we can deal with on this target. */
#if MIN_UNITS_PER_WORD > 4
# define LIBGCC2_MAX_UNITS_PER_WORD 8
#elif (MIN_UNITS_PER_WORD > 2 \
|| (MIN_UNITS_PER_WORD > 1 && __SIZEOF_LONG_LONG__ > 4))
# define LIBGCC2_MAX_UNITS_PER_WORD 4
#else
# define LIBGCC2_MAX_UNITS_PER_WORD MIN_UNITS_PER_WORD
#endif
/* Work out what word size we are using for this compilation.
The value can be set on the command line. */
#ifndef LIBGCC2_UNITS_PER_WORD
#define LIBGCC2_UNITS_PER_WORD LIBGCC2_MAX_UNITS_PER_WORD
#endif
typedef int QItype __attribute__ ((mode (QI)));
typedef unsigned int UQItype __attribute__ ((mode (QI)));
typedef int HItype __attribute__ ((mode (HI)));
typedef unsigned int UHItype __attribute__ ((mode (HI)));
#if MIN_UNITS_PER_WORD > 1
/* These typedefs are usually forbidden on dsp's with UNITS_PER_WORD 1. */
typedef int SItype __attribute__ ((mode (SI)));
typedef unsigned int USItype __attribute__ ((mode (SI)));
#if __SIZEOF_LONG_LONG__ > 4
/* These typedefs are usually forbidden on archs with UNITS_PER_WORD 2. */
typedef int DItype __attribute__ ((mode (DI)));
typedef unsigned int UDItype __attribute__ ((mode (DI)));
#if MIN_UNITS_PER_WORD > 4
/* These typedefs are usually forbidden on archs with UNITS_PER_WORD 4. */
typedef int TItype __attribute__ ((mode (TI)));
typedef unsigned int UTItype __attribute__ ((mode (TI)));
#endif
#endif
#endif
#if LIBGCC2_UNITS_PER_WORD == 8
#define W_TYPE_SIZE (8 * BITS_PER_UNIT)
#define Wtype DItype
#define UWtype UDItype
#define HWtype DItype
#define UHWtype UDItype
#define DWtype TItype
#define UDWtype UTItype
#ifdef LIBGCC2_GNU_PREFIX
#define __NW(a,b) __gnu_ ## a ## di ## b
#define __NDW(a,b) __gnu_ ## a ## ti ## b
#else
#define __NW(a,b) __ ## a ## di ## b
#define __NDW(a,b) __ ## a ## ti ## b
#endif
#elif LIBGCC2_UNITS_PER_WORD == 4
#define W_TYPE_SIZE (4 * BITS_PER_UNIT)
#define Wtype SItype
#define UWtype USItype
#define HWtype SItype
#define UHWtype USItype
#define DWtype DItype
#define UDWtype UDItype
#ifdef LIBGCC2_GNU_PREFIX
#define __NW(a,b) __gnu_ ## a ## si ## b
#define __NDW(a,b) __gnu_ ## a ## di ## b
#else
#define __NW(a,b) __ ## a ## si ## b
#define __NDW(a,b) __ ## a ## di ## b
#endif
#elif LIBGCC2_UNITS_PER_WORD == 2
#define W_TYPE_SIZE (2 * BITS_PER_UNIT)
#define Wtype HItype
#define UWtype UHItype
#define HWtype HItype
#define UHWtype UHItype
#define DWtype SItype
#define UDWtype USItype
#ifdef LIBGCC2_GNU_PREFIX
#define __NW(a,b) __gnu_ ## a ## hi ## b
#define __NDW(a,b) __gnu_ ## a ## si ## b
#else
#define __NW(a,b) __ ## a ## hi ## b
#define __NDW(a,b) __ ## a ## si ## b
#endif
#else
#define W_TYPE_SIZE BITS_PER_UNIT
#define Wtype QItype
#define UWtype UQItype
#define HWtype QItype
#define UHWtype UQItype
#define DWtype HItype
#define UDWtype UHItype
#ifdef LIBGCC2_GNU_PREFIX
#define __NW(a,b) __gnu_ ## a ## qi ## b
#define __NDW(a,b) __gnu_ ## a ## hi ## b
#else
#define __NW(a,b) __ ## a ## qi ## b
#define __NDW(a,b) __ ## a ## hi ## b
#endif
#endif
typedef int shift_count_type __attribute__((mode (__libgcc_shift_count__)));
#if __BYTE_ORDER__ != __ORDER_LITTLE_ENDIAN__
struct DWstruct {Wtype high, low;};
#else
struct DWstruct {Wtype low, high;};
#endif
/* We need this union to unpack/pack DImode values, since we don't have
any arithmetic yet. Incoming DImode parameters are stored into the
`ll' field, and the unpacked result is read from the struct `s'. */
typedef union {
struct DWstruct s;
DWtype ll;
} DWunion;
#endif /* __ASM_LIBGCC_H */
@@ -0,0 +1,135 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
*/
#include <common.h>
#include <elf.h>
#include <asm-generic/sections.h>
extern ulong __image_copy_start;
extern ulong __ivt_start;
extern ulong __ivt_end;
extern ulong __text_end;
DECLARE_GLOBAL_DATA_PTR;
int copy_uboot_to_ram(void)
{
size_t len = (size_t)&__image_copy_end - (size_t)&__image_copy_start;
if (gd->flags & GD_FLG_SKIP_RELOC)
return 0;
memcpy((void *)gd->relocaddr, (void *)&__image_copy_start, len);
return 0;
}
int clear_bss(void)
{
ulong dst_addr = (ulong)&__bss_start + gd->reloc_off;
size_t len = (size_t)&__bss_end - (size_t)&__bss_start;
memset((void *)dst_addr, 0x00, len);
return 0;
}
/*
* Base functionality is taken from x86 version with added ARC-specifics
*/
int do_elf_reloc_fixups(void)
{
Elf32_Rela *re_src = (Elf32_Rela *)(&__rel_dyn_start);
Elf32_Rela *re_end = (Elf32_Rela *)(&__rel_dyn_end);
if (gd->flags & GD_FLG_SKIP_RELOC)
return 0;
debug("Section .rela.dyn is located at %08x-%08x\n",
(unsigned int)re_src, (unsigned int)re_end);
Elf32_Addr *offset_ptr_rom;
Elf32_Addr *offset_ptr_ram;
do {
/* Get the location from the relocation entry */
offset_ptr_rom = (Elf32_Addr *)re_src->r_offset;
/* Check that the location of the relocation is in .text */
if (offset_ptr_rom >= (Elf32_Addr *)&__image_copy_start &&
offset_ptr_rom < (Elf32_Addr *)&__image_copy_end) {
unsigned int val, do_swap = 0;
/* Switch to the in-RAM version */
offset_ptr_ram = (Elf32_Addr *)((ulong)offset_ptr_rom +
gd->reloc_off);
#ifdef __LITTLE_ENDIAN__
/* If location in ".text" section swap value */
if (((u32)offset_ptr_rom >= (u32)&__text_start &&
(u32)offset_ptr_rom <= (u32)&__text_end)
#if defined(__ARC700__) || defined(__ARC600__)
|| ((u32)offset_ptr_rom >= (u32)&__ivt_start &&
(u32)offset_ptr_rom <= (u32)&__ivt_end)
#endif
)
do_swap = 1;
#endif
debug("Patching value @ %08x (relocated to %08x)%s\n",
(unsigned int)offset_ptr_rom,
(unsigned int)offset_ptr_ram,
do_swap ? ", middle-endian encoded" : "");
/*
* Use "memcpy" because target location might be
* 16-bit aligned on ARC so we may need to read
* byte-by-byte. On attempt to read entire word by
* CPU throws an exception
*/
memcpy(&val, offset_ptr_ram, sizeof(int));
if (do_swap)
val = (val << 16) | (val >> 16);
/* Check that the target points into executable */
if (val < (unsigned int)&__image_copy_start ||
val > (unsigned int)&__image_copy_end) {
/* TODO: Use panic() instead of debug()
*
* For some reason GCC might generate
* fake relocation even for LD/SC of constant
* inderectly. See an example below:
* ----------------------->8--------------------
* static int setup_mon_len(void)
* {
* gd->mon_len = (ulong)&__bss_end - CONFIG_SYS_MONITOR_BASE;
* return 0;
* }
* ----------------------->8--------------------
*
* And that's what we get in the binary:
* ----------------------->8--------------------
* 10005cb4 <setup_mon_len>:
* 10005cb4: 193c 3f80 0003 2f80 st 0x32f80,[r25,60]
* 10005cb8: R_ARC_RELATIVE *ABS*-0x10000000
* 10005cbc: 7fe0 j_s.d [blink]
* 10005cbe: 700c mov_s r0,0
* ----------------------->8--------------------
*/
debug("Relocation target %08x points outside of image\n",
val);
}
val += gd->reloc_off;
if (do_swap)
val = (val << 16) | (val >> 16);
memcpy(offset_ptr_ram, &val, sizeof(int));
}
} while (++re_src < re_end);
return 0;
}
@@ -0,0 +1,22 @@
// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
*/
#include <command.h>
#include <common.h>
__weak void reset_cpu(ulong addr)
{
/* Stop debug session here */
__builtin_arc_brk();
}
int do_reset(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[])
{
printf("Resetting the board...\n");
reset_cpu(0);
return 0;
}
@@ -0,0 +1,124 @@
/* SPDX-License-Identifier: GPL-2.0+ */
/*
* Copyright (C) 2013-2014 Synopsys, Inc. All rights reserved.
*/
#include <asm-offsets.h>
#include <config.h>
#include <linux/linkage.h>
#include <asm/arcregs.h>
ENTRY(_start)
/* Setup interrupt vector base that matches "__text_start" */
sr __ivt_start, [ARC_AUX_INTR_VEC_BASE]
; Disable/enable I-cache according to configuration
lr r5, [ARC_BCR_IC_BUILD]
breq r5, 0, 1f ; I$ doesn't exist
lr r5, [ARC_AUX_IC_CTRL]
#if !CONFIG_IS_ENABLED(SYS_ICACHE_OFF)
bclr r5, r5, 0 ; 0 - Enable, 1 is Disable
#else
bset r5, r5, 0 ; I$ exists, but is not used
#endif
sr r5, [ARC_AUX_IC_CTRL]
mov r5, 1
sr r5, [ARC_AUX_IC_IVIC]
; As per ARC HS databook (see chapter 5.3.3.2)
; it is required to add 3 NOPs after each write to IC_IVIC.
nop
nop
nop
1:
; Disable/enable D-cache according to configuration
lr r5, [ARC_BCR_DC_BUILD]
breq r5, 0, 1f ; D$ doesn't exist
lr r5, [ARC_AUX_DC_CTRL]
bclr r5, r5, 6 ; Invalidate (discard w/o wback)
#if !CONFIG_IS_ENABLED(SYS_DCACHE_OFF)
bclr r5, r5, 0 ; Enable (+Inv)
#else
bset r5, r5, 0 ; Disable (+Inv)
#endif
sr r5, [ARC_AUX_DC_CTRL]
mov r5, 1
sr r5, [ARC_AUX_DC_IVDC]
1:
#ifdef CONFIG_ISA_ARCV2
; Disable System-Level Cache (SLC)
lr r5, [ARC_BCR_SLC]
breq r5, 0, 1f ; SLC doesn't exist
lr r5, [ARC_AUX_SLC_CTRL]
bclr r5, r5, 6 ; Invalidate (discard w/o wback)
bclr r5, r5, 0 ; Enable (+Inv)
sr r5, [ARC_AUX_SLC_CTRL]
1:
#endif
#ifdef __ARC_UNALIGNED__
/*
* Enable handling of unaligned access in the CPU as by default
* this HW feature is disabled while GCC starting from 8.1.0
* unconditionally uses it for ARC HS cores.
*/
flag 1 << STATUS_AD_BIT
#endif
/* Establish C runtime stack and frame */
mov %sp, CONFIG_SYS_INIT_SP_ADDR
mov %fp, %sp
/* Allocate reserved area from current top of stack */
mov %r0, %sp
bl board_init_f_alloc_reserve
/* Set stack below reserved area, adjust frame pointer accordingly */
mov %sp, %r0
mov %fp, %sp
/* Initialize reserved area - note: r0 already contains address */
bl board_init_f_init_reserve
#ifdef CONFIG_DEBUG_UART
/* Earliest point to set up early debug uart */
bl debug_uart_init
#endif
/* Zero the one and only argument of "board_init_f" */
mov_s %r0, 0
bl board_init_f
/* We only get here if relocation is disabled by GD_FLG_SKIP_RELOC */
/* Make sure we don't lose GD overwritten by zero new GD */
mov %r0, %r25
mov %r1, 0
bl board_init_r
ENDPROC(_start)
/*
* void board_init_f_r_trampoline(stack-pointer address)
*
* This "function" does not return, instead it continues in RAM
* after relocating the monitor code.
*
* r0 = new stack-pointer
*/
ENTRY(board_init_f_r_trampoline)
/* Set up the stack- and frame-pointers */
mov %sp, %r0
mov %fp, %sp
/* Update position of intterupt vector table */
lr %r0, [ARC_AUX_INTR_VEC_BASE]
ld %r1, [%r25, GD_RELOC_OFF]
add %r0, %r0, %r1
sr %r0, [ARC_AUX_INTR_VEC_BASE]
/* Re-enter U-Boot by calling board_init_f_r */
j board_init_f_r
ENDPROC(board_init_f_r_trampoline)