/*************************************************************************** * Copyright (C) 2006, 2007 by Dominic Rath * * Dominic.Rath@gmx.de * * * * Copyright (C) 2007,2008 Øyvind Harboe * * oyvind.harboe@zylin.com * * * * Copyright (C) 2009 Michael Schwingen * * michael@schwingen.org * * * * This program is free software; you can redistribute it and/or modify * * it under the terms of the GNU General Public License as published by * * the Free Software Foundation; either version 2 of the License, or * * (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * * GNU General Public License for more details. * * * * You should have received a copy of the GNU General Public License * * along with this program; if not, write to the * * Free Software Foundation, Inc., * * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. * ***************************************************************************/ #ifdef HAVE_CONFIG_H #include "config.h" #endif #include "breakpoints.h" #include "xscale.h" #include "target_type.h" #include "arm_jtag.h" #include "arm_simulator.h" #include "arm_disassembler.h" #include "time_support.h" #include "register.h" #include "image.h" /* * Important XScale documents available as of October 2009 include: * * Intel XScale® Core Developer’s Manual, January 2004 * Order Number: 273473-002 * This has a chapter detailing debug facilities, and punts some * details to chip-specific microarchitecture documents. * * Hot-Debug for Intel XScale® Core Debug White Paper, May 2005 * Document Number: 273539-005 * Less detailed than the developer's manual, but summarizes those * missing details (for most XScales) and gives LOTS of notes about * debugger/handler interaction issues. Presents a simpler reset * and load-handler sequence than the arch doc. (Note, OpenOCD * doesn't currently support "Hot-Debug" as defined there.) * * Chip-specific microarchitecture documents may also be useful. */ /* forward declarations */ static int xscale_resume(struct target *, int current, uint32_t address, int handle_breakpoints, int debug_execution); static int xscale_debug_entry(struct target *); static int xscale_restore_context(struct target *); static int xscale_get_reg(struct reg *reg); static int xscale_set_reg(struct reg *reg, uint8_t *buf); static int xscale_set_breakpoint(struct target *, struct breakpoint *); static int xscale_set_watchpoint(struct target *, struct watchpoint *); static int xscale_unset_breakpoint(struct target *, struct breakpoint *); static int xscale_read_trace(struct target *); /* This XScale "debug handler" is loaded into the processor's * mini-ICache, which is 2K of code writable only via JTAG. * * FIXME the OpenOCD "bin2char" utility currently doesn't handle * binary files cleanly. It's string oriented, and terminates them * with a NUL character. Better would be to generate the constants * and let other code decide names, scoping, and other housekeeping. */ static /* unsigned const char xscale_debug_handler[] = ... */ #include "xscale_debug.h" static char *const xscale_reg_list[] = { "XSCALE_MAINID", /* 0 */ "XSCALE_CACHETYPE", "XSCALE_CTRL", "XSCALE_AUXCTRL", "XSCALE_TTB", "XSCALE_DAC", "XSCALE_FSR", "XSCALE_FAR", "XSCALE_PID", "XSCALE_CPACCESS", "XSCALE_IBCR0", /* 10 */ "XSCALE_IBCR1", "XSCALE_DBR0", "XSCALE_DBR1", "XSCALE_DBCON", "XSCALE_TBREG", "XSCALE_CHKPT0", "XSCALE_CHKPT1", "XSCALE_DCSR", "XSCALE_TX", "XSCALE_RX", /* 20 */ "XSCALE_TXRXCTRL", }; static const struct xscale_reg xscale_reg_arch_info[] = { {XSCALE_MAINID, NULL}, {XSCALE_CACHETYPE, NULL}, {XSCALE_CTRL, NULL}, {XSCALE_AUXCTRL, NULL}, {XSCALE_TTB, NULL}, {XSCALE_DAC, NULL}, {XSCALE_FSR, NULL}, {XSCALE_FAR, NULL}, {XSCALE_PID, NULL}, {XSCALE_CPACCESS, NULL}, {XSCALE_IBCR0, NULL}, {XSCALE_IBCR1, NULL}, {XSCALE_DBR0, NULL}, {XSCALE_DBR1, NULL}, {XSCALE_DBCON, NULL}, {XSCALE_TBREG, NULL}, {XSCALE_CHKPT0, NULL}, {XSCALE_CHKPT1, NULL}, {XSCALE_DCSR, NULL}, /* DCSR accessed via JTAG or SW */ {-1, NULL}, /* TX accessed via JTAG */ {-1, NULL}, /* RX accessed via JTAG */ {-1, NULL}, /* TXRXCTRL implicit access via JTAG */ }; /* convenience wrapper to access XScale specific registers */ static int xscale_set_reg_u32(struct reg *reg, uint32_t value) { uint8_t buf[4]; buf_set_u32(buf, 0, 32, value); return xscale_set_reg(reg, buf); } static const char xscale_not[] = "target is not an XScale"; static int xscale_verify_pointer(struct command_context *cmd_ctx, struct xscale_common *xscale) { if (xscale->common_magic != XSCALE_COMMON_MAGIC) { command_print(cmd_ctx, xscale_not); return ERROR_TARGET_INVALID; } return ERROR_OK; } static int xscale_jtag_set_instr(struct jtag_tap *tap, uint32_t new_instr) { if (tap == NULL) return ERROR_FAIL; if (buf_get_u32(tap->cur_instr, 0, tap->ir_length) != new_instr) { struct scan_field field; uint8_t scratch[4]; memset(&field, 0, sizeof field); field.tap = tap; field.num_bits = tap->ir_length; field.out_value = scratch; buf_set_u32(field.out_value, 0, field.num_bits, new_instr); jtag_add_ir_scan(1, &field, jtag_get_end_state()); } return ERROR_OK; } static int xscale_read_dcsr(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); int retval; struct scan_field fields[3]; uint8_t field0 = 0x0; uint8_t field0_check_value = 0x2; uint8_t field0_check_mask = 0x7; uint8_t field2 = 0x0; uint8_t field2_check_value = 0x0; uint8_t field2_check_mask = 0x1; jtag_set_end_state(TAP_DRPAUSE); xscale_jtag_set_instr(target->tap, XSCALE_SELDCSR << xscale->xscale_variant); buf_set_u32(&field0, 1, 1, xscale->hold_rst); buf_set_u32(&field0, 2, 1, xscale->external_debug_break); memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 3; fields[0].out_value = &field0; uint8_t tmp; fields[0].in_value = &tmp; fields[1].tap = target->tap; fields[1].num_bits = 32; fields[1].in_value = xscale->reg_cache->reg_list[XSCALE_DCSR].value; fields[2].tap = target->tap; fields[2].num_bits = 1; fields[2].out_value = &field2; uint8_t tmp2; fields[2].in_value = &tmp2; jtag_add_dr_scan(3, fields, jtag_get_end_state()); jtag_check_value_mask(fields + 0, &field0_check_value, &field0_check_mask); jtag_check_value_mask(fields + 2, &field2_check_value, &field2_check_mask); if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while reading DCSR"); return retval; } xscale->reg_cache->reg_list[XSCALE_DCSR].dirty = 0; xscale->reg_cache->reg_list[XSCALE_DCSR].valid = 1; /* write the register with the value we just read * on this second pass, only the first bit of field0 is guaranteed to be 0) */ field0_check_mask = 0x1; fields[1].out_value = xscale->reg_cache->reg_list[XSCALE_DCSR].value; fields[1].in_value = NULL; jtag_set_end_state(TAP_IDLE); jtag_add_dr_scan(3, fields, jtag_get_end_state()); /* DANGER!!! this must be here. It will make sure that the arguments * to jtag_set_check_value() does not go out of scope! */ return jtag_execute_queue(); } static void xscale_getbuf(jtag_callback_data_t arg) { uint8_t *in = (uint8_t *)arg; *((uint32_t *)in) = buf_get_u32(in, 0, 32); } static int xscale_receive(struct target *target, uint32_t *buffer, int num_words) { if (num_words == 0) return ERROR_INVALID_ARGUMENTS; struct xscale_common *xscale = target_to_xscale(target); int retval = ERROR_OK; tap_state_t path[3]; struct scan_field fields[3]; uint8_t *field0 = malloc(num_words * 1); uint8_t field0_check_value = 0x2; uint8_t field0_check_mask = 0x6; uint32_t *field1 = malloc(num_words * 4); uint8_t field2_check_value = 0x0; uint8_t field2_check_mask = 0x1; int words_done = 0; int words_scheduled = 0; int i; path[0] = TAP_DRSELECT; path[1] = TAP_DRCAPTURE; path[2] = TAP_DRSHIFT; memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 3; fields[0].check_value = &field0_check_value; fields[0].check_mask = &field0_check_mask; fields[1].tap = target->tap; fields[1].num_bits = 32; fields[2].tap = target->tap; fields[2].num_bits = 1; fields[2].check_value = &field2_check_value; fields[2].check_mask = &field2_check_mask; jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_DBGTX << xscale->xscale_variant); jtag_add_runtest(1, jtag_get_end_state()); /* ensures that we're in the TAP_IDLE state as the above could be a no-op */ /* repeat until all words have been collected */ int attempts = 0; while (words_done < num_words) { /* schedule reads */ words_scheduled = 0; for (i = words_done; i < num_words; i++) { fields[0].in_value = &field0[i]; jtag_add_pathmove(3, path); fields[1].in_value = (uint8_t *)(field1 + i); jtag_add_dr_scan_check(3, fields, jtag_set_end_state(TAP_IDLE)); jtag_add_callback(xscale_getbuf, (jtag_callback_data_t)(field1 + i)); words_scheduled++; } if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while receiving data from debug handler"); break; } /* examine results */ for (i = words_done; i < num_words; i++) { if (!(field0[0] & 1)) { /* move backwards if necessary */ int j; for (j = i; j < num_words - 1; j++) { field0[j] = field0[j + 1]; field1[j] = field1[j + 1]; } words_scheduled--; } } if (words_scheduled == 0) { if (attempts++==1000) { LOG_ERROR("Failed to receiving data from debug handler after 1000 attempts"); retval = ERROR_TARGET_TIMEOUT; break; } } words_done += words_scheduled; } for (i = 0; i < num_words; i++) *(buffer++) = buf_get_u32((uint8_t*)&field1[i], 0, 32); free(field1); return retval; } static int xscale_read_tx(struct target *target, int consume) { struct xscale_common *xscale = target_to_xscale(target); tap_state_t path[3]; tap_state_t noconsume_path[6]; int retval; struct timeval timeout, now; struct scan_field fields[3]; uint8_t field0_in = 0x0; uint8_t field0_check_value = 0x2; uint8_t field0_check_mask = 0x6; uint8_t field2_check_value = 0x0; uint8_t field2_check_mask = 0x1; jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_DBGTX << xscale->xscale_variant); path[0] = TAP_DRSELECT; path[1] = TAP_DRCAPTURE; path[2] = TAP_DRSHIFT; noconsume_path[0] = TAP_DRSELECT; noconsume_path[1] = TAP_DRCAPTURE; noconsume_path[2] = TAP_DREXIT1; noconsume_path[3] = TAP_DRPAUSE; noconsume_path[4] = TAP_DREXIT2; noconsume_path[5] = TAP_DRSHIFT; memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 3; fields[0].in_value = &field0_in; fields[1].tap = target->tap; fields[1].num_bits = 32; fields[1].in_value = xscale->reg_cache->reg_list[XSCALE_TX].value; fields[2].tap = target->tap; fields[2].num_bits = 1; uint8_t tmp; fields[2].in_value = &tmp; gettimeofday(&timeout, NULL); timeval_add_time(&timeout, 1, 0); for (;;) { /* if we want to consume the register content (i.e. clear TX_READY), * we have to go straight from Capture-DR to Shift-DR * otherwise, we go from Capture-DR to Exit1-DR to Pause-DR */ if (consume) jtag_add_pathmove(3, path); else { jtag_add_pathmove(ARRAY_SIZE(noconsume_path), noconsume_path); } jtag_add_dr_scan(3, fields, jtag_set_end_state(TAP_IDLE)); jtag_check_value_mask(fields + 0, &field0_check_value, &field0_check_mask); jtag_check_value_mask(fields + 2, &field2_check_value, &field2_check_mask); if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while reading TX"); return ERROR_TARGET_TIMEOUT; } gettimeofday(&now, NULL); if ((now.tv_sec > timeout.tv_sec) || ((now.tv_sec == timeout.tv_sec)&& (now.tv_usec > timeout.tv_usec))) { LOG_ERROR("time out reading TX register"); return ERROR_TARGET_TIMEOUT; } if (!((!(field0_in & 1)) && consume)) { goto done; } if (debug_level >= 3) { LOG_DEBUG("waiting 100ms"); alive_sleep(100); /* avoid flooding the logs */ } else { keep_alive(); } } done: if (!(field0_in & 1)) return ERROR_TARGET_RESOURCE_NOT_AVAILABLE; return ERROR_OK; } static int xscale_write_rx(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); int retval; struct timeval timeout, now; struct scan_field fields[3]; uint8_t field0_out = 0x0; uint8_t field0_in = 0x0; uint8_t field0_check_value = 0x2; uint8_t field0_check_mask = 0x6; uint8_t field2 = 0x0; uint8_t field2_check_value = 0x0; uint8_t field2_check_mask = 0x1; jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_DBGRX << xscale->xscale_variant); memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 3; fields[0].out_value = &field0_out; fields[0].in_value = &field0_in; fields[1].tap = target->tap; fields[1].num_bits = 32; fields[1].out_value = xscale->reg_cache->reg_list[XSCALE_RX].value; fields[2].tap = target->tap; fields[2].num_bits = 1; fields[2].out_value = &field2; uint8_t tmp; fields[2].in_value = &tmp; gettimeofday(&timeout, NULL); timeval_add_time(&timeout, 1, 0); /* poll until rx_read is low */ LOG_DEBUG("polling RX"); for (;;) { jtag_add_dr_scan(3, fields, jtag_set_end_state(TAP_IDLE)); jtag_check_value_mask(fields + 0, &field0_check_value, &field0_check_mask); jtag_check_value_mask(fields + 2, &field2_check_value, &field2_check_mask); if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while writing RX"); return retval; } gettimeofday(&now, NULL); if ((now.tv_sec > timeout.tv_sec) || ((now.tv_sec == timeout.tv_sec)&& (now.tv_usec > timeout.tv_usec))) { LOG_ERROR("time out writing RX register"); return ERROR_TARGET_TIMEOUT; } if (!(field0_in & 1)) goto done; if (debug_level >= 3) { LOG_DEBUG("waiting 100ms"); alive_sleep(100); /* avoid flooding the logs */ } else { keep_alive(); } } done: /* set rx_valid */ field2 = 0x1; jtag_add_dr_scan(3, fields, jtag_set_end_state(TAP_IDLE)); if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while writing RX"); return retval; } return ERROR_OK; } /* send count elements of size byte to the debug handler */ static int xscale_send(struct target *target, uint8_t *buffer, int count, int size) { struct xscale_common *xscale = target_to_xscale(target); uint32_t t[3]; int bits[3]; int retval; int done_count = 0; jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_DBGRX << xscale->xscale_variant); bits[0]=3; t[0]=0; bits[1]=32; t[2]=1; bits[2]=1; int endianness = target->endianness; while (done_count++ < count) { switch (size) { case 4: if (endianness == TARGET_LITTLE_ENDIAN) { t[1]=le_to_h_u32(buffer); } else { t[1]=be_to_h_u32(buffer); } break; case 2: if (endianness == TARGET_LITTLE_ENDIAN) { t[1]=le_to_h_u16(buffer); } else { t[1]=be_to_h_u16(buffer); } break; case 1: t[1]=buffer[0]; break; default: LOG_ERROR("BUG: size neither 4, 2 nor 1"); return ERROR_INVALID_ARGUMENTS; } jtag_add_dr_out(target->tap, 3, bits, t, jtag_set_end_state(TAP_IDLE)); buffer += size; } if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while sending data to debug handler"); return retval; } return ERROR_OK; } static int xscale_send_u32(struct target *target, uint32_t value) { struct xscale_common *xscale = target_to_xscale(target); buf_set_u32(xscale->reg_cache->reg_list[XSCALE_RX].value, 0, 32, value); return xscale_write_rx(target); } static int xscale_write_dcsr(struct target *target, int hold_rst, int ext_dbg_brk) { struct xscale_common *xscale = target_to_xscale(target); int retval; struct scan_field fields[3]; uint8_t field0 = 0x0; uint8_t field0_check_value = 0x2; uint8_t field0_check_mask = 0x7; uint8_t field2 = 0x0; uint8_t field2_check_value = 0x0; uint8_t field2_check_mask = 0x1; if (hold_rst != -1) xscale->hold_rst = hold_rst; if (ext_dbg_brk != -1) xscale->external_debug_break = ext_dbg_brk; jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_SELDCSR << xscale->xscale_variant); buf_set_u32(&field0, 1, 1, xscale->hold_rst); buf_set_u32(&field0, 2, 1, xscale->external_debug_break); memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 3; fields[0].out_value = &field0; uint8_t tmp; fields[0].in_value = &tmp; fields[1].tap = target->tap; fields[1].num_bits = 32; fields[1].out_value = xscale->reg_cache->reg_list[XSCALE_DCSR].value; fields[2].tap = target->tap; fields[2].num_bits = 1; fields[2].out_value = &field2; uint8_t tmp2; fields[2].in_value = &tmp2; jtag_add_dr_scan(3, fields, jtag_get_end_state()); jtag_check_value_mask(fields + 0, &field0_check_value, &field0_check_mask); jtag_check_value_mask(fields + 2, &field2_check_value, &field2_check_mask); if ((retval = jtag_execute_queue()) != ERROR_OK) { LOG_ERROR("JTAG error while writing DCSR"); return retval; } xscale->reg_cache->reg_list[XSCALE_DCSR].dirty = 0; xscale->reg_cache->reg_list[XSCALE_DCSR].valid = 1; return ERROR_OK; } /* parity of the number of bits 0 if even; 1 if odd. for 32 bit words */ static unsigned int parity (unsigned int v) { // unsigned int ov = v; v ^= v >> 16; v ^= v >> 8; v ^= v >> 4; v &= 0xf; // LOG_DEBUG("parity of 0x%x is %i", ov, (0x6996 >> v) & 1); return (0x6996 >> v) & 1; } static int xscale_load_ic(struct target *target, uint32_t va, uint32_t buffer[8]) { struct xscale_common *xscale = target_to_xscale(target); uint8_t packet[4]; uint8_t cmd; int word; struct scan_field fields[2]; LOG_DEBUG("loading miniIC at 0x%8.8" PRIx32 "", va); /* LDIC into IR */ jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_LDIC << xscale->xscale_variant); /* CMD is b011 to load a cacheline into the Mini ICache. * Loading into the main ICache is deprecated, and unused. * It's followed by three zero bits, and 27 address bits. */ buf_set_u32(&cmd, 0, 6, 0x3); /* virtual address of desired cache line */ buf_set_u32(packet, 0, 27, va >> 5); memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 6; fields[0].out_value = &cmd; fields[1].tap = target->tap; fields[1].num_bits = 27; fields[1].out_value = packet; jtag_add_dr_scan(2, fields, jtag_get_end_state()); /* rest of packet is a cacheline: 8 instructions, with parity */ fields[0].num_bits = 32; fields[0].out_value = packet; fields[1].num_bits = 1; fields[1].out_value = &cmd; for (word = 0; word < 8; word++) { buf_set_u32(packet, 0, 32, buffer[word]); uint32_t value; memcpy(&value, packet, sizeof(uint32_t)); cmd = parity(value); jtag_add_dr_scan(2, fields, jtag_get_end_state()); } return jtag_execute_queue(); } static int xscale_invalidate_ic_line(struct target *target, uint32_t va) { struct xscale_common *xscale = target_to_xscale(target); uint8_t packet[4]; uint8_t cmd; struct scan_field fields[2]; jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_LDIC << xscale->xscale_variant); /* CMD for invalidate IC line b000, bits [6:4] b000 */ buf_set_u32(&cmd, 0, 6, 0x0); /* virtual address of desired cache line */ buf_set_u32(packet, 0, 27, va >> 5); memset(&fields, 0, sizeof fields); fields[0].tap = target->tap; fields[0].num_bits = 6; fields[0].out_value = &cmd; fields[1].tap = target->tap; fields[1].num_bits = 27; fields[1].out_value = packet; jtag_add_dr_scan(2, fields, jtag_get_end_state()); return ERROR_OK; } static int xscale_update_vectors(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); int i; int retval; uint32_t low_reset_branch, high_reset_branch; for (i = 1; i < 8; i++) { /* if there's a static vector specified for this exception, override */ if (xscale->static_high_vectors_set & (1 << i)) { xscale->high_vectors[i] = xscale->static_high_vectors[i]; } else { retval = target_read_u32(target, 0xffff0000 + 4*i, &xscale->high_vectors[i]); if (retval == ERROR_TARGET_TIMEOUT) return retval; if (retval != ERROR_OK) { /* Some of these reads will fail as part of normal execution */ xscale->high_vectors[i] = ARMV4_5_B(0xfffffe, 0); } } } for (i = 1; i < 8; i++) { if (xscale->static_low_vectors_set & (1 << i)) { xscale->low_vectors[i] = xscale->static_low_vectors[i]; } else { retval = target_read_u32(target, 0x0 + 4*i, &xscale->low_vectors[i]); if (retval == ERROR_TARGET_TIMEOUT) return retval; if (retval != ERROR_OK) { /* Some of these reads will fail as part of normal execution */ xscale->low_vectors[i] = ARMV4_5_B(0xfffffe, 0); } } } /* calculate branches to debug handler */ low_reset_branch = (xscale->handler_address + 0x20 - 0x0 - 0x8) >> 2; high_reset_branch = (xscale->handler_address + 0x20 - 0xffff0000 - 0x8) >> 2; xscale->low_vectors[0] = ARMV4_5_B((low_reset_branch & 0xffffff), 0); xscale->high_vectors[0] = ARMV4_5_B((high_reset_branch & 0xffffff), 0); /* invalidate and load exception vectors in mini i-cache */ xscale_invalidate_ic_line(target, 0x0); xscale_invalidate_ic_line(target, 0xffff0000); xscale_load_ic(target, 0x0, xscale->low_vectors); xscale_load_ic(target, 0xffff0000, xscale->high_vectors); return ERROR_OK; } static int xscale_arch_state(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; static const char *state[] = { "disabled", "enabled" }; static const char *arch_dbg_reason[] = { "", "\n(processor reset)", "\n(trace buffer full)" }; if (armv4_5->common_magic != ARMV4_5_COMMON_MAGIC) { LOG_ERROR("BUG: called for a non-ARMv4/5 target"); return ERROR_INVALID_ARGUMENTS; } LOG_USER("target halted in %s state due to %s, current mode: %s\n" "cpsr: 0x%8.8" PRIx32 " pc: 0x%8.8" PRIx32 "\n" "MMU: %s, D-Cache: %s, I-Cache: %s" "%s", armv4_5_state_strings[armv4_5->core_state], Jim_Nvp_value2name_simple(nvp_target_debug_reason, target->debug_reason)->name , arm_mode_name(armv4_5->core_mode), buf_get_u32(armv4_5->cpsr->value, 0, 32), buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32), state[xscale->armv4_5_mmu.mmu_enabled], state[xscale->armv4_5_mmu.armv4_5_cache.d_u_cache_enabled], state[xscale->armv4_5_mmu.armv4_5_cache.i_cache_enabled], arch_dbg_reason[xscale->arch_debug_reason]); return ERROR_OK; } static int xscale_poll(struct target *target) { int retval = ERROR_OK; if ((target->state == TARGET_RUNNING) || (target->state == TARGET_DEBUG_RUNNING)) { enum target_state previous_state = target->state; if ((retval = xscale_read_tx(target, 0)) == ERROR_OK) { /* there's data to read from the tx register, we entered debug state */ target->state = TARGET_HALTED; /* process debug entry, fetching current mode regs */ retval = xscale_debug_entry(target); } else if (retval != ERROR_TARGET_RESOURCE_NOT_AVAILABLE) { LOG_USER("error while polling TX register, reset CPU"); /* here we "lie" so GDB won't get stuck and a reset can be perfomed */ target->state = TARGET_HALTED; } /* debug_entry could have overwritten target state (i.e. immediate resume) * don't signal event handlers in that case */ if (target->state != TARGET_HALTED) return ERROR_OK; /* if target was running, signal that we halted * otherwise we reentered from debug execution */ if (previous_state == TARGET_RUNNING) target_call_event_callbacks(target, TARGET_EVENT_HALTED); else target_call_event_callbacks(target, TARGET_EVENT_DEBUG_HALTED); } return retval; } static int xscale_debug_entry(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; uint32_t pc; uint32_t buffer[10]; int i; int retval; uint32_t moe; /* clear external dbg break (will be written on next DCSR read) */ xscale->external_debug_break = 0; if ((retval = xscale_read_dcsr(target)) != ERROR_OK) return retval; /* get r0, pc, r1 to r7 and cpsr */ if ((retval = xscale_receive(target, buffer, 10)) != ERROR_OK) return retval; /* move r0 from buffer to register cache */ buf_set_u32(armv4_5->core_cache->reg_list[0].value, 0, 32, buffer[0]); armv4_5->core_cache->reg_list[0].dirty = 1; armv4_5->core_cache->reg_list[0].valid = 1; LOG_DEBUG("r0: 0x%8.8" PRIx32 "", buffer[0]); /* move pc from buffer to register cache */ buf_set_u32(armv4_5->core_cache->reg_list[15].value, 0, 32, buffer[1]); armv4_5->core_cache->reg_list[15].dirty = 1; armv4_5->core_cache->reg_list[15].valid = 1; LOG_DEBUG("pc: 0x%8.8" PRIx32 "", buffer[1]); /* move data from buffer to register cache */ for (i = 1; i <= 7; i++) { buf_set_u32(armv4_5->core_cache->reg_list[i].value, 0, 32, buffer[1 + i]); armv4_5->core_cache->reg_list[i].dirty = 1; armv4_5->core_cache->reg_list[i].valid = 1; LOG_DEBUG("r%i: 0x%8.8" PRIx32 "", i, buffer[i + 1]); } arm_set_cpsr(armv4_5, buffer[9]); LOG_DEBUG("cpsr: 0x%8.8" PRIx32 "", buffer[9]); if (!is_arm_mode(armv4_5->core_mode)) { target->state = TARGET_UNKNOWN; LOG_ERROR("cpsr contains invalid mode value - communication failure"); return ERROR_TARGET_FAILURE; } LOG_DEBUG("target entered debug state in %s mode", arm_mode_name(armv4_5->core_mode)); /* get banked registers, r8 to r14, and spsr if not in USR/SYS mode */ if ((armv4_5->core_mode != ARMV4_5_MODE_USR) && (armv4_5->core_mode != ARMV4_5_MODE_SYS)) { xscale_receive(target, buffer, 8); buf_set_u32(ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, 16).value, 0, 32, buffer[7]); ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, 16).dirty = 0; ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, 16).valid = 1; } else { /* r8 to r14, but no spsr */ xscale_receive(target, buffer, 7); } /* move data from buffer to register cache */ for (i = 8; i <= 14; i++) { buf_set_u32(ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, i).value, 0, 32, buffer[i - 8]); ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, i).dirty = 0; ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, i).valid = 1; } /* examine debug reason */ xscale_read_dcsr(target); moe = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 2, 3); /* stored PC (for calculating fixup) */ pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); switch (moe) { case 0x0: /* Processor reset */ target->debug_reason = DBG_REASON_DBGRQ; xscale->arch_debug_reason = XSCALE_DBG_REASON_RESET; pc -= 4; break; case 0x1: /* Instruction breakpoint hit */ target->debug_reason = DBG_REASON_BREAKPOINT; xscale->arch_debug_reason = XSCALE_DBG_REASON_GENERIC; pc -= 4; break; case 0x2: /* Data breakpoint hit */ target->debug_reason = DBG_REASON_WATCHPOINT; xscale->arch_debug_reason = XSCALE_DBG_REASON_GENERIC; pc -= 4; break; case 0x3: /* BKPT instruction executed */ target->debug_reason = DBG_REASON_BREAKPOINT; xscale->arch_debug_reason = XSCALE_DBG_REASON_GENERIC; pc -= 4; break; case 0x4: /* Ext. debug event */ target->debug_reason = DBG_REASON_DBGRQ; xscale->arch_debug_reason = XSCALE_DBG_REASON_GENERIC; pc -= 4; break; case 0x5: /* Vector trap occured */ target->debug_reason = DBG_REASON_BREAKPOINT; xscale->arch_debug_reason = XSCALE_DBG_REASON_GENERIC; pc -= 4; break; case 0x6: /* Trace buffer full break */ target->debug_reason = DBG_REASON_DBGRQ; xscale->arch_debug_reason = XSCALE_DBG_REASON_TB_FULL; pc -= 4; break; case 0x7: /* Reserved (may flag Hot-Debug support) */ default: LOG_ERROR("Method of Entry is 'Reserved'"); exit(-1); break; } /* apply PC fixup */ buf_set_u32(armv4_5->core_cache->reg_list[15].value, 0, 32, pc); /* on the first debug entry, identify cache type */ if (xscale->armv4_5_mmu.armv4_5_cache.ctype == -1) { uint32_t cache_type_reg; /* read cp15 cache type register */ xscale_get_reg(&xscale->reg_cache->reg_list[XSCALE_CACHETYPE]); cache_type_reg = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_CACHETYPE].value, 0, 32); armv4_5_identify_cache(cache_type_reg, &xscale->armv4_5_mmu.armv4_5_cache); } /* examine MMU and Cache settings */ /* read cp15 control register */ xscale_get_reg(&xscale->reg_cache->reg_list[XSCALE_CTRL]); xscale->cp15_control_reg = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_CTRL].value, 0, 32); xscale->armv4_5_mmu.mmu_enabled = (xscale->cp15_control_reg & 0x1U) ? 1 : 0; xscale->armv4_5_mmu.armv4_5_cache.d_u_cache_enabled = (xscale->cp15_control_reg & 0x4U) ? 1 : 0; xscale->armv4_5_mmu.armv4_5_cache.i_cache_enabled = (xscale->cp15_control_reg & 0x1000U) ? 1 : 0; /* tracing enabled, read collected trace data */ if (xscale->trace.buffer_enabled) { xscale_read_trace(target); xscale->trace.buffer_fill--; /* resume if we're still collecting trace data */ if ((xscale->arch_debug_reason == XSCALE_DBG_REASON_TB_FULL) && (xscale->trace.buffer_fill > 0)) { xscale_resume(target, 1, 0x0, 1, 0); } else { xscale->trace.buffer_enabled = 0; } } return ERROR_OK; } static int xscale_halt(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); LOG_DEBUG("target->state: %s", target_state_name(target)); if (target->state == TARGET_HALTED) { LOG_DEBUG("target was already halted"); return ERROR_OK; } else if (target->state == TARGET_UNKNOWN) { /* this must not happen for a xscale target */ LOG_ERROR("target was in unknown state when halt was requested"); return ERROR_TARGET_INVALID; } else if (target->state == TARGET_RESET) { LOG_DEBUG("target->state == TARGET_RESET"); } else { /* assert external dbg break */ xscale->external_debug_break = 1; xscale_read_dcsr(target); target->debug_reason = DBG_REASON_DBGRQ; } return ERROR_OK; } static int xscale_enable_single_step(struct target *target, uint32_t next_pc) { struct xscale_common *xscale = target_to_xscale(target); struct reg *ibcr0 = &xscale->reg_cache->reg_list[XSCALE_IBCR0]; int retval; if (xscale->ibcr0_used) { struct breakpoint *ibcr0_bp = breakpoint_find(target, buf_get_u32(ibcr0->value, 0, 32) & 0xfffffffe); if (ibcr0_bp) { xscale_unset_breakpoint(target, ibcr0_bp); } else { LOG_ERROR("BUG: xscale->ibcr0_used is set, but no breakpoint with that address found"); exit(-1); } } if ((retval = xscale_set_reg_u32(ibcr0, next_pc | 0x1)) != ERROR_OK) return retval; return ERROR_OK; } static int xscale_disable_single_step(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); struct reg *ibcr0 = &xscale->reg_cache->reg_list[XSCALE_IBCR0]; int retval; if ((retval = xscale_set_reg_u32(ibcr0, 0x0)) != ERROR_OK) return retval; return ERROR_OK; } static void xscale_enable_watchpoints(struct target *target) { struct watchpoint *watchpoint = target->watchpoints; while (watchpoint) { if (watchpoint->set == 0) xscale_set_watchpoint(target, watchpoint); watchpoint = watchpoint->next; } } static void xscale_enable_breakpoints(struct target *target) { struct breakpoint *breakpoint = target->breakpoints; /* set any pending breakpoints */ while (breakpoint) { if (breakpoint->set == 0) xscale_set_breakpoint(target, breakpoint); breakpoint = breakpoint->next; } } static int xscale_resume(struct target *target, int current, uint32_t address, int handle_breakpoints, int debug_execution) { struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; struct breakpoint *breakpoint = target->breakpoints; uint32_t current_pc; int retval; int i; LOG_DEBUG("-"); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } if (!debug_execution) { target_free_all_working_areas(target); } /* update vector tables */ if ((retval = xscale_update_vectors(target)) != ERROR_OK) return retval; /* current = 1: continue on current pc, otherwise continue at
*/ if (!current) buf_set_u32(armv4_5->core_cache->reg_list[15].value, 0, 32, address); current_pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); /* if we're at the reset vector, we have to simulate the branch */ if (current_pc == 0x0) { arm_simulate_step(target, NULL); current_pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); } /* the front-end may request us not to handle breakpoints */ if (handle_breakpoints) { if ((breakpoint = breakpoint_find(target, buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)))) { uint32_t next_pc; /* there's a breakpoint at the current PC, we have to step over it */ LOG_DEBUG("unset breakpoint at 0x%8.8" PRIx32 "", breakpoint->address); xscale_unset_breakpoint(target, breakpoint); /* calculate PC of next instruction */ if ((retval = arm_simulate_step(target, &next_pc)) != ERROR_OK) { uint32_t current_opcode; target_read_u32(target, current_pc, ¤t_opcode); LOG_ERROR("BUG: couldn't calculate PC of next instruction, current opcode was 0x%8.8" PRIx32 "", current_opcode); } LOG_DEBUG("enable single-step"); xscale_enable_single_step(target, next_pc); /* restore banked registers */ xscale_restore_context(target); /* send resume request (command 0x30 or 0x31) * clean the trace buffer if it is to be enabled (0x62) */ if (xscale->trace.buffer_enabled) { xscale_send_u32(target, 0x62); xscale_send_u32(target, 0x31); } else xscale_send_u32(target, 0x30); /* send CPSR */ xscale_send_u32(target, buf_get_u32(armv4_5->cpsr->value, 0, 32)); LOG_DEBUG("writing cpsr with value 0x%8.8" PRIx32, buf_get_u32(armv4_5->cpsr->value, 0, 32)); for (i = 7; i >= 0; i--) { /* send register */ xscale_send_u32(target, buf_get_u32(armv4_5->core_cache->reg_list[i].value, 0, 32)); LOG_DEBUG("writing r%i with value 0x%8.8" PRIx32 "", i, buf_get_u32(armv4_5->core_cache->reg_list[i].value, 0, 32)); } /* send PC */ xscale_send_u32(target, buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)); LOG_DEBUG("writing PC with value 0x%8.8" PRIx32 "", buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)); /* wait for and process debug entry */ xscale_debug_entry(target); LOG_DEBUG("disable single-step"); xscale_disable_single_step(target); LOG_DEBUG("set breakpoint at 0x%8.8" PRIx32 "", breakpoint->address); xscale_set_breakpoint(target, breakpoint); } } /* enable any pending breakpoints and watchpoints */ xscale_enable_breakpoints(target); xscale_enable_watchpoints(target); /* restore banked registers */ xscale_restore_context(target); /* send resume request (command 0x30 or 0x31) * clean the trace buffer if it is to be enabled (0x62) */ if (xscale->trace.buffer_enabled) { xscale_send_u32(target, 0x62); xscale_send_u32(target, 0x31); } else xscale_send_u32(target, 0x30); /* send CPSR */ xscale_send_u32(target, buf_get_u32(armv4_5->cpsr->value, 0, 32)); LOG_DEBUG("writing cpsr with value 0x%8.8" PRIx32, buf_get_u32(armv4_5->cpsr->value, 0, 32)); for (i = 7; i >= 0; i--) { /* send register */ xscale_send_u32(target, buf_get_u32(armv4_5->core_cache->reg_list[i].value, 0, 32)); LOG_DEBUG("writing r%i with value 0x%8.8" PRIx32 "", i, buf_get_u32(armv4_5->core_cache->reg_list[i].value, 0, 32)); } /* send PC */ xscale_send_u32(target, buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)); LOG_DEBUG("writing PC with value 0x%8.8" PRIx32 "", buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)); target->debug_reason = DBG_REASON_NOTHALTED; if (!debug_execution) { /* registers are now invalid */ register_cache_invalidate(armv4_5->core_cache); target->state = TARGET_RUNNING; target_call_event_callbacks(target, TARGET_EVENT_RESUMED); } else { target->state = TARGET_DEBUG_RUNNING; target_call_event_callbacks(target, TARGET_EVENT_DEBUG_RESUMED); } LOG_DEBUG("target resumed"); return ERROR_OK; } static int xscale_step_inner(struct target *target, int current, uint32_t address, int handle_breakpoints) { struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; uint32_t next_pc; int retval; int i; target->debug_reason = DBG_REASON_SINGLESTEP; /* calculate PC of next instruction */ if ((retval = arm_simulate_step(target, &next_pc)) != ERROR_OK) { uint32_t current_opcode, current_pc; current_pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); target_read_u32(target, current_pc, ¤t_opcode); LOG_ERROR("BUG: couldn't calculate PC of next instruction, current opcode was 0x%8.8" PRIx32 "", current_opcode); return retval; } LOG_DEBUG("enable single-step"); if ((retval = xscale_enable_single_step(target, next_pc)) != ERROR_OK) return retval; /* restore banked registers */ if ((retval = xscale_restore_context(target)) != ERROR_OK) return retval; /* send resume request (command 0x30 or 0x31) * clean the trace buffer if it is to be enabled (0x62) */ if (xscale->trace.buffer_enabled) { if ((retval = xscale_send_u32(target, 0x62)) != ERROR_OK) return retval; if ((retval = xscale_send_u32(target, 0x31)) != ERROR_OK) return retval; } else if ((retval = xscale_send_u32(target, 0x30)) != ERROR_OK) return retval; /* send CPSR */ retval = xscale_send_u32(target, buf_get_u32(armv4_5->cpsr->value, 0, 32)); if (retval != ERROR_OK) return retval; LOG_DEBUG("writing cpsr with value 0x%8.8" PRIx32, buf_get_u32(armv4_5->cpsr->value, 0, 32)); for (i = 7; i >= 0; i--) { /* send register */ if ((retval = xscale_send_u32(target, buf_get_u32(armv4_5->core_cache->reg_list[i].value, 0, 32))) != ERROR_OK) return retval; LOG_DEBUG("writing r%i with value 0x%8.8" PRIx32 "", i, buf_get_u32(armv4_5->core_cache->reg_list[i].value, 0, 32)); } /* send PC */ if ((retval = xscale_send_u32(target, buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32))) != ERROR_OK) return retval; LOG_DEBUG("writing PC with value 0x%8.8" PRIx32, buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)); target_call_event_callbacks(target, TARGET_EVENT_RESUMED); /* registers are now invalid */ register_cache_invalidate(armv4_5->core_cache); /* wait for and process debug entry */ if ((retval = xscale_debug_entry(target)) != ERROR_OK) return retval; LOG_DEBUG("disable single-step"); if ((retval = xscale_disable_single_step(target)) != ERROR_OK) return retval; target_call_event_callbacks(target, TARGET_EVENT_HALTED); return ERROR_OK; } static int xscale_step(struct target *target, int current, uint32_t address, int handle_breakpoints) { struct arm *armv4_5 = target_to_armv4_5(target); struct breakpoint *breakpoint = target->breakpoints; uint32_t current_pc; int retval; if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } /* current = 1: continue on current pc, otherwise continue at
*/ if (!current) buf_set_u32(armv4_5->core_cache->reg_list[15].value, 0, 32, address); current_pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); /* if we're at the reset vector, we have to simulate the step */ if (current_pc == 0x0) { if ((retval = arm_simulate_step(target, NULL)) != ERROR_OK) return retval; current_pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); target->debug_reason = DBG_REASON_SINGLESTEP; target_call_event_callbacks(target, TARGET_EVENT_HALTED); return ERROR_OK; } /* the front-end may request us not to handle breakpoints */ if (handle_breakpoints) if ((breakpoint = breakpoint_find(target, buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32)))) { if ((retval = xscale_unset_breakpoint(target, breakpoint)) != ERROR_OK) return retval; } retval = xscale_step_inner(target, current, address, handle_breakpoints); if (breakpoint) { xscale_set_breakpoint(target, breakpoint); } LOG_DEBUG("target stepped"); return ERROR_OK; } static int xscale_assert_reset(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); LOG_DEBUG("target->state: %s", target_state_name(target)); /* select DCSR instruction (set endstate to R-T-I to ensure we don't * end up in T-L-R, which would reset JTAG */ jtag_set_end_state(TAP_IDLE); xscale_jtag_set_instr(target->tap, XSCALE_SELDCSR << xscale->xscale_variant); /* set Hold reset, Halt mode and Trap Reset */ buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 30, 1, 0x1); buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 16, 1, 0x1); xscale_write_dcsr(target, 1, 0); /* select BYPASS, because having DCSR selected caused problems on the PXA27x */ xscale_jtag_set_instr(target->tap, 0x7f); jtag_execute_queue(); /* assert reset */ jtag_add_reset(0, 1); /* sleep 1ms, to be sure we fulfill any requirements */ jtag_add_sleep(1000); jtag_execute_queue(); target->state = TARGET_RESET; if (target->reset_halt) { int retval; if ((retval = target_halt(target)) != ERROR_OK) return retval; } return ERROR_OK; } static int xscale_deassert_reset(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); struct breakpoint *breakpoint = target->breakpoints; LOG_DEBUG("-"); xscale->ibcr_available = 2; xscale->ibcr0_used = 0; xscale->ibcr1_used = 0; xscale->dbr_available = 2; xscale->dbr0_used = 0; xscale->dbr1_used = 0; /* mark all hardware breakpoints as unset */ while (breakpoint) { if (breakpoint->type == BKPT_HARD) { breakpoint->set = 0; } breakpoint = breakpoint->next; } register_cache_invalidate(xscale->armv4_5_common.core_cache); /* FIXME mark hardware watchpoints got unset too. Also, * at least some of the XScale registers are invalid... */ /* * REVISIT: *assumes* we had a SRST+TRST reset so the mini-icache * contents got invalidated. Safer to force that, so writing new * contents can't ever fail.. */ { uint32_t address; unsigned buf_cnt; const uint8_t *buffer = xscale_debug_handler; int retval; /* release SRST */ jtag_add_reset(0, 0); /* wait 300ms; 150 and 100ms were not enough */ jtag_add_sleep(300*1000); jtag_add_runtest(2030, jtag_set_end_state(TAP_IDLE)); jtag_execute_queue(); /* set Hold reset, Halt mode and Trap Reset */ buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 30, 1, 0x1); buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 16, 1, 0x1); xscale_write_dcsr(target, 1, 0); /* Load the debug handler into the mini-icache. Since * it's using halt mode (not monitor mode), it runs in * "Special Debug State" for access to registers, memory, * coprocessors, trace data, etc. */ address = xscale->handler_address; for (unsigned binary_size = sizeof xscale_debug_handler - 1; binary_size > 0; binary_size -= buf_cnt, buffer += buf_cnt) { uint32_t cache_line[8]; unsigned i; buf_cnt = binary_size; if (buf_cnt > 32) buf_cnt = 32; for (i = 0; i < buf_cnt; i += 4) { /* convert LE buffer to host-endian uint32_t */ cache_line[i / 4] = le_to_h_u32(&buffer[i]); } for (; i < 32; i += 4) { cache_line[i / 4] = 0xe1a08008; } /* only load addresses other than the reset vectors */ if ((address % 0x400) != 0x0) { retval = xscale_load_ic(target, address, cache_line); if (retval != ERROR_OK) return retval; } address += buf_cnt; }; retval = xscale_load_ic(target, 0x0, xscale->low_vectors); if (retval != ERROR_OK) return retval; retval = xscale_load_ic(target, 0xffff0000, xscale->high_vectors); if (retval != ERROR_OK) return retval; jtag_add_runtest(30, jtag_set_end_state(TAP_IDLE)); jtag_add_sleep(100000); /* set Hold reset, Halt mode and Trap Reset */ buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 30, 1, 0x1); buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 16, 1, 0x1); xscale_write_dcsr(target, 1, 0); /* clear Hold reset to let the target run (should enter debug handler) */ xscale_write_dcsr(target, 0, 1); target->state = TARGET_RUNNING; if (!target->reset_halt) { jtag_add_sleep(10000); /* we should have entered debug now */ xscale_debug_entry(target); target->state = TARGET_HALTED; /* resume the target */ xscale_resume(target, 1, 0x0, 1, 0); } } return ERROR_OK; } static int xscale_read_core_reg(struct target *target, struct reg *r, int num, enum armv4_5_mode mode) { /** \todo add debug handler support for core register reads */ LOG_ERROR("not implemented"); return ERROR_OK; } static int xscale_write_core_reg(struct target *target, struct reg *r, int num, enum armv4_5_mode mode, uint32_t value) { /** \todo add debug handler support for core register writes */ LOG_ERROR("not implemented"); return ERROR_OK; } static int xscale_full_context(struct target *target) { struct arm *armv4_5 = target_to_armv4_5(target); uint32_t *buffer; int i, j; LOG_DEBUG("-"); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } buffer = malloc(4 * 8); /* iterate through processor modes (FIQ, IRQ, SVC, ABT, UND and SYS) * we can't enter User mode on an XScale (unpredictable), * but User shares registers with SYS */ for (i = 1; i < 7; i++) { int valid = 1; /* check if there are invalid registers in the current mode */ for (j = 0; j <= 16; j++) { if (ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), j).valid == 0) valid = 0; } if (!valid) { uint32_t tmp_cpsr; /* request banked registers */ xscale_send_u32(target, 0x0); tmp_cpsr = 0x0; tmp_cpsr |= armv4_5_number_to_mode(i); tmp_cpsr |= 0xc0; /* I/F bits */ /* send CPSR for desired mode */ xscale_send_u32(target, tmp_cpsr); /* get banked registers, r8 to r14, and spsr if not in USR/SYS mode */ if ((armv4_5_number_to_mode(i) != ARMV4_5_MODE_USR) && (armv4_5_number_to_mode(i) != ARMV4_5_MODE_SYS)) { xscale_receive(target, buffer, 8); buf_set_u32(ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, 16).value, 0, 32, buffer[7]); ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), 16).dirty = 0; ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), 16).valid = 1; } else { xscale_receive(target, buffer, 7); } /* move data from buffer to register cache */ for (j = 8; j <= 14; j++) { buf_set_u32(ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), j).value, 0, 32, buffer[j - 8]); ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), j).dirty = 0; ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), j).valid = 1; } } } free(buffer); return ERROR_OK; } static int xscale_restore_context(struct target *target) { struct arm *armv4_5 = target_to_armv4_5(target); int i, j; if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } /* iterate through processor modes (FIQ, IRQ, SVC, ABT, UND and SYS) * we can't enter User mode on an XScale (unpredictable), * but User shares registers with SYS */ for (i = 1; i < 7; i++) { int dirty = 0; /* check if there are invalid registers in the current mode */ for (j = 8; j <= 14; j++) { if (ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), j).dirty == 1) dirty = 1; } /* if not USR/SYS, check if the SPSR needs to be written */ if ((armv4_5_number_to_mode(i) != ARMV4_5_MODE_USR) && (armv4_5_number_to_mode(i) != ARMV4_5_MODE_SYS)) { if (ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), 16).dirty == 1) dirty = 1; } if (dirty) { uint32_t tmp_cpsr; /* send banked registers */ xscale_send_u32(target, 0x1); tmp_cpsr = 0x0; tmp_cpsr |= armv4_5_number_to_mode(i); tmp_cpsr |= 0xc0; /* I/F bits */ /* send CPSR for desired mode */ xscale_send_u32(target, tmp_cpsr); /* send banked registers, r8 to r14, and spsr if not in USR/SYS mode */ for (j = 8; j <= 14; j++) { xscale_send_u32(target, buf_get_u32(ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, j).value, 0, 32)); ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), j).dirty = 0; } if ((armv4_5_number_to_mode(i) != ARMV4_5_MODE_USR) && (armv4_5_number_to_mode(i) != ARMV4_5_MODE_SYS)) { xscale_send_u32(target, buf_get_u32(ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5->core_mode, 16).value, 0, 32)); ARMV4_5_CORE_REG_MODE(armv4_5->core_cache, armv4_5_number_to_mode(i), 16).dirty = 0; } } } return ERROR_OK; } static int xscale_read_memory(struct target *target, uint32_t address, uint32_t size, uint32_t count, uint8_t *buffer) { struct xscale_common *xscale = target_to_xscale(target); uint32_t *buf32; uint32_t i; int retval; LOG_DEBUG("address: 0x%8.8" PRIx32 ", size: 0x%8.8" PRIx32 ", count: 0x%8.8" PRIx32, address, size, count); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } /* sanitize arguments */ if (((size != 4) && (size != 2) && (size != 1)) || (count == 0) || !(buffer)) return ERROR_INVALID_ARGUMENTS; if (((size == 4) && (address & 0x3u)) || ((size == 2) && (address & 0x1u))) return ERROR_TARGET_UNALIGNED_ACCESS; /* send memory read request (command 0x1n, n: access size) */ if ((retval = xscale_send_u32(target, 0x10 | size)) != ERROR_OK) return retval; /* send base address for read request */ if ((retval = xscale_send_u32(target, address)) != ERROR_OK) return retval; /* send number of requested data words */ if ((retval = xscale_send_u32(target, count)) != ERROR_OK) return retval; /* receive data from target (count times 32-bit words in host endianness) */ buf32 = malloc(4 * count); if ((retval = xscale_receive(target, buf32, count)) != ERROR_OK) return retval; /* extract data from host-endian buffer into byte stream */ for (i = 0; i < count; i++) { switch (size) { case 4: target_buffer_set_u32(target, buffer, buf32[i]); buffer += 4; break; case 2: target_buffer_set_u16(target, buffer, buf32[i] & 0xffff); buffer += 2; break; case 1: *buffer++ = buf32[i] & 0xff; break; default: LOG_ERROR("invalid read size"); return ERROR_INVALID_ARGUMENTS; } } free(buf32); /* examine DCSR, to see if Sticky Abort (SA) got set */ if ((retval = xscale_read_dcsr(target)) != ERROR_OK) return retval; if (buf_get_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 5, 1) == 1) { /* clear SA bit */ if ((retval = xscale_send_u32(target, 0x60)) != ERROR_OK) return retval; return ERROR_TARGET_DATA_ABORT; } return ERROR_OK; } static int xscale_read_phys_memory(struct target *target, uint32_t address, uint32_t size, uint32_t count, uint8_t *buffer) { /** \todo: provide a non-stub implementtion of this routine. */ LOG_ERROR("%s: %s is not implemented. Disable MMU?", target_name(target), __func__); return ERROR_FAIL; } static int xscale_write_memory(struct target *target, uint32_t address, uint32_t size, uint32_t count, uint8_t *buffer) { struct xscale_common *xscale = target_to_xscale(target); int retval; LOG_DEBUG("address: 0x%8.8" PRIx32 ", size: 0x%8.8" PRIx32 ", count: 0x%8.8" PRIx32, address, size, count); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } /* sanitize arguments */ if (((size != 4) && (size != 2) && (size != 1)) || (count == 0) || !(buffer)) return ERROR_INVALID_ARGUMENTS; if (((size == 4) && (address & 0x3u)) || ((size == 2) && (address & 0x1u))) return ERROR_TARGET_UNALIGNED_ACCESS; /* send memory write request (command 0x2n, n: access size) */ if ((retval = xscale_send_u32(target, 0x20 | size)) != ERROR_OK) return retval; /* send base address for read request */ if ((retval = xscale_send_u32(target, address)) != ERROR_OK) return retval; /* send number of requested data words to be written*/ if ((retval = xscale_send_u32(target, count)) != ERROR_OK) return retval; /* extract data from host-endian buffer into byte stream */ #if 0 for (i = 0; i < count; i++) { switch (size) { case 4: value = target_buffer_get_u32(target, buffer); xscale_send_u32(target, value); buffer += 4; break; case 2: value = target_buffer_get_u16(target, buffer); xscale_send_u32(target, value); buffer += 2; break; case 1: value = *buffer; xscale_send_u32(target, value); buffer += 1; break; default: LOG_ERROR("should never get here"); exit(-1); } } #endif if ((retval = xscale_send(target, buffer, count, size)) != ERROR_OK) return retval; /* examine DCSR, to see if Sticky Abort (SA) got set */ if ((retval = xscale_read_dcsr(target)) != ERROR_OK) return retval; if (buf_get_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 5, 1) == 1) { /* clear SA bit */ if ((retval = xscale_send_u32(target, 0x60)) != ERROR_OK) return retval; return ERROR_TARGET_DATA_ABORT; } return ERROR_OK; } static int xscale_write_phys_memory(struct target *target, uint32_t address, uint32_t size, uint32_t count, uint8_t *buffer) { /** \todo: provide a non-stub implementtion of this routine. */ LOG_ERROR("%s: %s is not implemented. Disable MMU?", target_name(target), __func__); return ERROR_FAIL; } static int xscale_bulk_write_memory(struct target *target, uint32_t address, uint32_t count, uint8_t *buffer) { return xscale_write_memory(target, address, 4, count, buffer); } static uint32_t xscale_get_ttb(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); uint32_t ttb; xscale_get_reg(&xscale->reg_cache->reg_list[XSCALE_TTB]); ttb = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_TTB].value, 0, 32); return ttb; } static void xscale_disable_mmu_caches(struct target *target, int mmu, int d_u_cache, int i_cache) { struct xscale_common *xscale = target_to_xscale(target); uint32_t cp15_control; /* read cp15 control register */ xscale_get_reg(&xscale->reg_cache->reg_list[XSCALE_CTRL]); cp15_control = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_CTRL].value, 0, 32); if (mmu) cp15_control &= ~0x1U; if (d_u_cache) { /* clean DCache */ xscale_send_u32(target, 0x50); xscale_send_u32(target, xscale->cache_clean_address); /* invalidate DCache */ xscale_send_u32(target, 0x51); cp15_control &= ~0x4U; } if (i_cache) { /* invalidate ICache */ xscale_send_u32(target, 0x52); cp15_control &= ~0x1000U; } /* write new cp15 control register */ xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_CTRL], cp15_control); /* execute cpwait to ensure outstanding operations complete */ xscale_send_u32(target, 0x53); } static void xscale_enable_mmu_caches(struct target *target, int mmu, int d_u_cache, int i_cache) { struct xscale_common *xscale = target_to_xscale(target); uint32_t cp15_control; /* read cp15 control register */ xscale_get_reg(&xscale->reg_cache->reg_list[XSCALE_CTRL]); cp15_control = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_CTRL].value, 0, 32); if (mmu) cp15_control |= 0x1U; if (d_u_cache) cp15_control |= 0x4U; if (i_cache) cp15_control |= 0x1000U; /* write new cp15 control register */ xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_CTRL], cp15_control); /* execute cpwait to ensure outstanding operations complete */ xscale_send_u32(target, 0x53); } static int xscale_set_breakpoint(struct target *target, struct breakpoint *breakpoint) { int retval; struct xscale_common *xscale = target_to_xscale(target); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } if (breakpoint->set) { LOG_WARNING("breakpoint already set"); return ERROR_OK; } if (breakpoint->type == BKPT_HARD) { uint32_t value = breakpoint->address | 1; if (!xscale->ibcr0_used) { xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_IBCR0], value); xscale->ibcr0_used = 1; breakpoint->set = 1; /* breakpoint set on first breakpoint register */ } else if (!xscale->ibcr1_used) { xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_IBCR1], value); xscale->ibcr1_used = 1; breakpoint->set = 2; /* breakpoint set on second breakpoint register */ } else { LOG_ERROR("BUG: no hardware comparator available"); return ERROR_OK; } } else if (breakpoint->type == BKPT_SOFT) { if (breakpoint->length == 4) { /* keep the original instruction in target endianness */ if ((retval = target_read_memory(target, breakpoint->address, 4, 1, breakpoint->orig_instr)) != ERROR_OK) { return retval; } /* write the original instruction in target endianness (arm7_9->arm_bkpt is host endian) */ if ((retval = target_write_u32(target, breakpoint->address, xscale->arm_bkpt)) != ERROR_OK) { return retval; } } else { /* keep the original instruction in target endianness */ if ((retval = target_read_memory(target, breakpoint->address, 2, 1, breakpoint->orig_instr)) != ERROR_OK) { return retval; } /* write the original instruction in target endianness (arm7_9->arm_bkpt is host endian) */ if ((retval = target_write_u32(target, breakpoint->address, xscale->thumb_bkpt)) != ERROR_OK) { return retval; } } breakpoint->set = 1; } return ERROR_OK; } static int xscale_add_breakpoint(struct target *target, struct breakpoint *breakpoint) { struct xscale_common *xscale = target_to_xscale(target); if ((breakpoint->type == BKPT_HARD) && (xscale->ibcr_available < 1)) { LOG_INFO("no breakpoint unit available for hardware breakpoint"); return ERROR_TARGET_RESOURCE_NOT_AVAILABLE; } if ((breakpoint->length != 2) && (breakpoint->length != 4)) { LOG_INFO("only breakpoints of two (Thumb) or four (ARM) bytes length supported"); return ERROR_TARGET_RESOURCE_NOT_AVAILABLE; } if (breakpoint->type == BKPT_HARD) { xscale->ibcr_available--; } return ERROR_OK; } static int xscale_unset_breakpoint(struct target *target, struct breakpoint *breakpoint) { int retval; struct xscale_common *xscale = target_to_xscale(target); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } if (!breakpoint->set) { LOG_WARNING("breakpoint not set"); return ERROR_OK; } if (breakpoint->type == BKPT_HARD) { if (breakpoint->set == 1) { xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_IBCR0], 0x0); xscale->ibcr0_used = 0; } else if (breakpoint->set == 2) { xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_IBCR1], 0x0); xscale->ibcr1_used = 0; } breakpoint->set = 0; } else { /* restore original instruction (kept in target endianness) */ if (breakpoint->length == 4) { if ((retval = target_write_memory(target, breakpoint->address, 4, 1, breakpoint->orig_instr)) != ERROR_OK) { return retval; } } else { if ((retval = target_write_memory(target, breakpoint->address, 2, 1, breakpoint->orig_instr)) != ERROR_OK) { return retval; } } breakpoint->set = 0; } return ERROR_OK; } static int xscale_remove_breakpoint(struct target *target, struct breakpoint *breakpoint) { struct xscale_common *xscale = target_to_xscale(target); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } if (breakpoint->set) { xscale_unset_breakpoint(target, breakpoint); } if (breakpoint->type == BKPT_HARD) xscale->ibcr_available++; return ERROR_OK; } static int xscale_set_watchpoint(struct target *target, struct watchpoint *watchpoint) { struct xscale_common *xscale = target_to_xscale(target); uint8_t enable = 0; struct reg *dbcon = &xscale->reg_cache->reg_list[XSCALE_DBCON]; uint32_t dbcon_value = buf_get_u32(dbcon->value, 0, 32); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } xscale_get_reg(dbcon); switch (watchpoint->rw) { case WPT_READ: enable = 0x3; break; case WPT_ACCESS: enable = 0x2; break; case WPT_WRITE: enable = 0x1; break; default: LOG_ERROR("BUG: watchpoint->rw neither read, write nor access"); } if (!xscale->dbr0_used) { xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_DBR0], watchpoint->address); dbcon_value |= enable; xscale_set_reg_u32(dbcon, dbcon_value); watchpoint->set = 1; xscale->dbr0_used = 1; } else if (!xscale->dbr1_used) { xscale_set_reg_u32(&xscale->reg_cache->reg_list[XSCALE_DBR1], watchpoint->address); dbcon_value |= enable << 2; xscale_set_reg_u32(dbcon, dbcon_value); watchpoint->set = 2; xscale->dbr1_used = 1; } else { LOG_ERROR("BUG: no hardware comparator available"); return ERROR_OK; } return ERROR_OK; } static int xscale_add_watchpoint(struct target *target, struct watchpoint *watchpoint) { struct xscale_common *xscale = target_to_xscale(target); if (xscale->dbr_available < 1) { return ERROR_TARGET_RESOURCE_NOT_AVAILABLE; } if ((watchpoint->length != 1) && (watchpoint->length != 2) && (watchpoint->length != 4)) { return ERROR_TARGET_RESOURCE_NOT_AVAILABLE; } xscale->dbr_available--; return ERROR_OK; } static int xscale_unset_watchpoint(struct target *target, struct watchpoint *watchpoint) { struct xscale_common *xscale = target_to_xscale(target); struct reg *dbcon = &xscale->reg_cache->reg_list[XSCALE_DBCON]; uint32_t dbcon_value = buf_get_u32(dbcon->value, 0, 32); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } if (!watchpoint->set) { LOG_WARNING("breakpoint not set"); return ERROR_OK; } if (watchpoint->set == 1) { dbcon_value &= ~0x3; xscale_set_reg_u32(dbcon, dbcon_value); xscale->dbr0_used = 0; } else if (watchpoint->set == 2) { dbcon_value &= ~0xc; xscale_set_reg_u32(dbcon, dbcon_value); xscale->dbr1_used = 0; } watchpoint->set = 0; return ERROR_OK; } static int xscale_remove_watchpoint(struct target *target, struct watchpoint *watchpoint) { struct xscale_common *xscale = target_to_xscale(target); if (target->state != TARGET_HALTED) { LOG_WARNING("target not halted"); return ERROR_TARGET_NOT_HALTED; } if (watchpoint->set) { xscale_unset_watchpoint(target, watchpoint); } xscale->dbr_available++; return ERROR_OK; } static int xscale_get_reg(struct reg *reg) { struct xscale_reg *arch_info = reg->arch_info; struct target *target = arch_info->target; struct xscale_common *xscale = target_to_xscale(target); /* DCSR, TX and RX are accessible via JTAG */ if (strcmp(reg->name, "XSCALE_DCSR") == 0) { return xscale_read_dcsr(arch_info->target); } else if (strcmp(reg->name, "XSCALE_TX") == 0) { /* 1 = consume register content */ return xscale_read_tx(arch_info->target, 1); } else if (strcmp(reg->name, "XSCALE_RX") == 0) { /* can't read from RX register (host -> debug handler) */ return ERROR_OK; } else if (strcmp(reg->name, "XSCALE_TXRXCTRL") == 0) { /* can't (explicitly) read from TXRXCTRL register */ return ERROR_OK; } else /* Other DBG registers have to be transfered by the debug handler */ { /* send CP read request (command 0x40) */ xscale_send_u32(target, 0x40); /* send CP register number */ xscale_send_u32(target, arch_info->dbg_handler_number); /* read register value */ xscale_read_tx(target, 1); buf_cpy(xscale->reg_cache->reg_list[XSCALE_TX].value, reg->value, 32); reg->dirty = 0; reg->valid = 1; } return ERROR_OK; } static int xscale_set_reg(struct reg *reg, uint8_t* buf) { struct xscale_reg *arch_info = reg->arch_info; struct target *target = arch_info->target; struct xscale_common *xscale = target_to_xscale(target); uint32_t value = buf_get_u32(buf, 0, 32); /* DCSR, TX and RX are accessible via JTAG */ if (strcmp(reg->name, "XSCALE_DCSR") == 0) { buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 0, 32, value); return xscale_write_dcsr(arch_info->target, -1, -1); } else if (strcmp(reg->name, "XSCALE_RX") == 0) { buf_set_u32(xscale->reg_cache->reg_list[XSCALE_RX].value, 0, 32, value); return xscale_write_rx(arch_info->target); } else if (strcmp(reg->name, "XSCALE_TX") == 0) { /* can't write to TX register (debug-handler -> host) */ return ERROR_OK; } else if (strcmp(reg->name, "XSCALE_TXRXCTRL") == 0) { /* can't (explicitly) write to TXRXCTRL register */ return ERROR_OK; } else /* Other DBG registers have to be transfered by the debug handler */ { /* send CP write request (command 0x41) */ xscale_send_u32(target, 0x41); /* send CP register number */ xscale_send_u32(target, arch_info->dbg_handler_number); /* send CP register value */ xscale_send_u32(target, value); buf_set_u32(reg->value, 0, 32, value); } return ERROR_OK; } static int xscale_write_dcsr_sw(struct target *target, uint32_t value) { struct xscale_common *xscale = target_to_xscale(target); struct reg *dcsr = &xscale->reg_cache->reg_list[XSCALE_DCSR]; struct xscale_reg *dcsr_arch_info = dcsr->arch_info; /* send CP write request (command 0x41) */ xscale_send_u32(target, 0x41); /* send CP register number */ xscale_send_u32(target, dcsr_arch_info->dbg_handler_number); /* send CP register value */ xscale_send_u32(target, value); buf_set_u32(dcsr->value, 0, 32, value); return ERROR_OK; } static int xscale_read_trace(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; struct xscale_trace_data **trace_data_p; /* 258 words from debug handler * 256 trace buffer entries * 2 checkpoint addresses */ uint32_t trace_buffer[258]; int is_address[256]; int i, j; if (target->state != TARGET_HALTED) { LOG_WARNING("target must be stopped to read trace data"); return ERROR_TARGET_NOT_HALTED; } /* send read trace buffer command (command 0x61) */ xscale_send_u32(target, 0x61); /* receive trace buffer content */ xscale_receive(target, trace_buffer, 258); /* parse buffer backwards to identify address entries */ for (i = 255; i >= 0; i--) { is_address[i] = 0; if (((trace_buffer[i] & 0xf0) == 0x90) || ((trace_buffer[i] & 0xf0) == 0xd0)) { if (i >= 3) is_address[--i] = 1; if (i >= 2) is_address[--i] = 1; if (i >= 1) is_address[--i] = 1; if (i >= 0) is_address[--i] = 1; } } /* search first non-zero entry */ for (j = 0; (j < 256) && (trace_buffer[j] == 0) && (!is_address[j]); j++) ; if (j == 256) { LOG_DEBUG("no trace data collected"); return ERROR_XSCALE_NO_TRACE_DATA; } for (trace_data_p = &xscale->trace.data; *trace_data_p; trace_data_p = &(*trace_data_p)->next) ; *trace_data_p = malloc(sizeof(struct xscale_trace_data)); (*trace_data_p)->next = NULL; (*trace_data_p)->chkpt0 = trace_buffer[256]; (*trace_data_p)->chkpt1 = trace_buffer[257]; (*trace_data_p)->last_instruction = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); (*trace_data_p)->entries = malloc(sizeof(struct xscale_trace_entry) * (256 - j)); (*trace_data_p)->depth = 256 - j; for (i = j; i < 256; i++) { (*trace_data_p)->entries[i - j].data = trace_buffer[i]; if (is_address[i]) (*trace_data_p)->entries[i - j].type = XSCALE_TRACE_ADDRESS; else (*trace_data_p)->entries[i - j].type = XSCALE_TRACE_MESSAGE; } return ERROR_OK; } static int xscale_read_instruction(struct target *target, struct arm_instruction *instruction) { struct xscale_common *xscale = target_to_xscale(target); int i; int section = -1; size_t size_read; uint32_t opcode; int retval; if (!xscale->trace.image) return ERROR_TRACE_IMAGE_UNAVAILABLE; /* search for the section the current instruction belongs to */ for (i = 0; i < xscale->trace.image->num_sections; i++) { if ((xscale->trace.image->sections[i].base_address <= xscale->trace.current_pc) && (xscale->trace.image->sections[i].base_address + xscale->trace.image->sections[i].size > xscale->trace.current_pc)) { section = i; break; } } if (section == -1) { /* current instruction couldn't be found in the image */ return ERROR_TRACE_INSTRUCTION_UNAVAILABLE; } if (xscale->trace.core_state == ARMV4_5_STATE_ARM) { uint8_t buf[4]; if ((retval = image_read_section(xscale->trace.image, section, xscale->trace.current_pc - xscale->trace.image->sections[section].base_address, 4, buf, &size_read)) != ERROR_OK) { LOG_ERROR("error while reading instruction: %i", retval); return ERROR_TRACE_INSTRUCTION_UNAVAILABLE; } opcode = target_buffer_get_u32(target, buf); arm_evaluate_opcode(opcode, xscale->trace.current_pc, instruction); } else if (xscale->trace.core_state == ARMV4_5_STATE_THUMB) { uint8_t buf[2]; if ((retval = image_read_section(xscale->trace.image, section, xscale->trace.current_pc - xscale->trace.image->sections[section].base_address, 2, buf, &size_read)) != ERROR_OK) { LOG_ERROR("error while reading instruction: %i", retval); return ERROR_TRACE_INSTRUCTION_UNAVAILABLE; } opcode = target_buffer_get_u16(target, buf); thumb_evaluate_opcode(opcode, xscale->trace.current_pc, instruction); } else { LOG_ERROR("BUG: unknown core state encountered"); exit(-1); } return ERROR_OK; } static int xscale_branch_address(struct xscale_trace_data *trace_data, int i, uint32_t *target) { /* if there are less than four entries prior to the indirect branch message * we can't extract the address */ if (i < 4) { return -1; } *target = (trace_data->entries[i-1].data) | (trace_data->entries[i-2].data << 8) | (trace_data->entries[i-3].data << 16) | (trace_data->entries[i-4].data << 24); return 0; } static int xscale_analyze_trace(struct target *target, struct command_context *cmd_ctx) { struct xscale_common *xscale = target_to_xscale(target); int next_pc_ok = 0; uint32_t next_pc = 0x0; struct xscale_trace_data *trace_data = xscale->trace.data; int retval; while (trace_data) { int i, chkpt; int rollover; int branch; int exception; xscale->trace.core_state = ARMV4_5_STATE_ARM; chkpt = 0; rollover = 0; for (i = 0; i < trace_data->depth; i++) { next_pc_ok = 0; branch = 0; exception = 0; if (trace_data->entries[i].type == XSCALE_TRACE_ADDRESS) continue; switch ((trace_data->entries[i].data & 0xf0) >> 4) { case 0: /* Exceptions */ case 1: case 2: case 3: case 4: case 5: case 6: case 7: exception = (trace_data->entries[i].data & 0x70) >> 4; next_pc_ok = 1; next_pc = (trace_data->entries[i].data & 0xf0) >> 2; command_print(cmd_ctx, "--- exception %i ---", (trace_data->entries[i].data & 0xf0) >> 4); break; case 8: /* Direct Branch */ branch = 1; break; case 9: /* Indirect Branch */ branch = 1; if (xscale_branch_address(trace_data, i, &next_pc) == 0) { next_pc_ok = 1; } break; case 13: /* Checkpointed Indirect Branch */ if (xscale_branch_address(trace_data, i, &next_pc) == 0) { next_pc_ok = 1; if (((chkpt == 0) && (next_pc != trace_data->chkpt0)) || ((chkpt == 1) && (next_pc != trace_data->chkpt1))) LOG_WARNING("checkpointed indirect branch target address doesn't match checkpoint"); } /* explicit fall-through */ case 12: /* Checkpointed Direct Branch */ branch = 1; if (chkpt == 0) { next_pc_ok = 1; next_pc = trace_data->chkpt0; chkpt++; } else if (chkpt == 1) { next_pc_ok = 1; next_pc = trace_data->chkpt0; chkpt++; } else { LOG_WARNING("more than two checkpointed branches encountered"); } break; case 15: /* Roll-over */ rollover++; continue; default: /* Reserved */ command_print(cmd_ctx, "--- reserved trace message ---"); LOG_ERROR("BUG: trace message %i is reserved", (trace_data->entries[i].data & 0xf0) >> 4); return ERROR_OK; } if (xscale->trace.pc_ok) { int executed = (trace_data->entries[i].data & 0xf) + rollover * 16; struct arm_instruction instruction; if ((exception == 6) || (exception == 7)) { /* IRQ or FIQ exception, no instruction executed */ executed -= 1; } while (executed-- >= 0) { if ((retval = xscale_read_instruction(target, &instruction)) != ERROR_OK) { /* can't continue tracing with no image available */ if (retval == ERROR_TRACE_IMAGE_UNAVAILABLE) { return retval; } else if (retval == ERROR_TRACE_INSTRUCTION_UNAVAILABLE) { /* TODO: handle incomplete images */ } } /* a precise abort on a load to the PC is included in the incremental * word count, other instructions causing data aborts are not included */ if ((executed == 0) && (exception == 4) && ((instruction.type >= ARM_LDR) && (instruction.type <= ARM_LDM))) { if ((instruction.type == ARM_LDM) && ((instruction.info.load_store_multiple.register_list & 0x8000) == 0)) { executed--; } else if (((instruction.type >= ARM_LDR) && (instruction.type <= ARM_LDRSH)) && (instruction.info.load_store.Rd != 15)) { executed--; } } /* only the last instruction executed * (the one that caused the control flow change) * could be a taken branch */ if (((executed == -1) && (branch == 1)) && (((instruction.type == ARM_B) || (instruction.type == ARM_BL) || (instruction.type == ARM_BLX)) && (instruction.info.b_bl_bx_blx.target_address != 0xffffffff))) { xscale->trace.current_pc = instruction.info.b_bl_bx_blx.target_address; } else { xscale->trace.current_pc += (xscale->trace.core_state == ARMV4_5_STATE_ARM) ? 4 : 2; } command_print(cmd_ctx, "%s", instruction.text); } rollover = 0; } if (next_pc_ok) { xscale->trace.current_pc = next_pc; xscale->trace.pc_ok = 1; } } for (; xscale->trace.current_pc < trace_data->last_instruction; xscale->trace.current_pc += (xscale->trace.core_state == ARMV4_5_STATE_ARM) ? 4 : 2) { struct arm_instruction instruction; if ((retval = xscale_read_instruction(target, &instruction)) != ERROR_OK) { /* can't continue tracing with no image available */ if (retval == ERROR_TRACE_IMAGE_UNAVAILABLE) { return retval; } else if (retval == ERROR_TRACE_INSTRUCTION_UNAVAILABLE) { /* TODO: handle incomplete images */ } } command_print(cmd_ctx, "%s", instruction.text); } trace_data = trace_data->next; } return ERROR_OK; } static const struct reg_arch_type xscale_reg_type = { .get = xscale_get_reg, .set = xscale_set_reg, }; static void xscale_build_reg_cache(struct target *target) { struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; struct reg_cache **cache_p = register_get_last_cache_p(&target->reg_cache); struct xscale_reg *arch_info = malloc(sizeof(xscale_reg_arch_info)); int i; int num_regs = ARRAY_SIZE(xscale_reg_arch_info); (*cache_p) = armv4_5_build_reg_cache(target, armv4_5); (*cache_p)->next = malloc(sizeof(struct reg_cache)); cache_p = &(*cache_p)->next; /* fill in values for the xscale reg cache */ (*cache_p)->name = "XScale registers"; (*cache_p)->next = NULL; (*cache_p)->reg_list = malloc(num_regs * sizeof(struct reg)); (*cache_p)->num_regs = num_regs; for (i = 0; i < num_regs; i++) { (*cache_p)->reg_list[i].name = xscale_reg_list[i]; (*cache_p)->reg_list[i].value = calloc(4, 1); (*cache_p)->reg_list[i].dirty = 0; (*cache_p)->reg_list[i].valid = 0; (*cache_p)->reg_list[i].size = 32; (*cache_p)->reg_list[i].arch_info = &arch_info[i]; (*cache_p)->reg_list[i].type = &xscale_reg_type; arch_info[i] = xscale_reg_arch_info[i]; arch_info[i].target = target; } xscale->reg_cache = (*cache_p); } static int xscale_init_target(struct command_context *cmd_ctx, struct target *target) { xscale_build_reg_cache(target); return ERROR_OK; } static int xscale_init_arch_info(struct target *target, struct xscale_common *xscale, struct jtag_tap *tap, const char *variant) { struct arm *armv4_5; uint32_t high_reset_branch, low_reset_branch; int i; armv4_5 = &xscale->armv4_5_common; /* store architecture specfic data */ xscale->common_magic = XSCALE_COMMON_MAGIC; /* we don't really *need* a variant param ... */ if (variant) { int ir_length = 0; if (strcmp(variant, "pxa250") == 0 || strcmp(variant, "pxa255") == 0 || strcmp(variant, "pxa26x") == 0) ir_length = 5; else if (strcmp(variant, "pxa27x") == 0 || strcmp(variant, "ixp42x") == 0 || strcmp(variant, "ixp45x") == 0 || strcmp(variant, "ixp46x") == 0) ir_length = 7; else if (strcmp(variant, "pxa3xx") == 0) ir_length = 11; else LOG_WARNING("%s: unrecognized variant %s", tap->dotted_name, variant); if (ir_length && ir_length != tap->ir_length) { LOG_WARNING("%s: IR length for %s is %d; fixing", tap->dotted_name, variant, ir_length); tap->ir_length = ir_length; } } /* PXA3xx shifts the JTAG instructions */ if (tap->ir_length == 11) xscale->xscale_variant = XSCALE_PXA3XX; else xscale->xscale_variant = XSCALE_IXP4XX_PXA2XX; /* the debug handler isn't installed (and thus not running) at this time */ xscale->handler_address = 0xfe000800; /* clear the vectors we keep locally for reference */ memset(xscale->low_vectors, 0, sizeof(xscale->low_vectors)); memset(xscale->high_vectors, 0, sizeof(xscale->high_vectors)); /* no user-specified vectors have been configured yet */ xscale->static_low_vectors_set = 0x0; xscale->static_high_vectors_set = 0x0; /* calculate branches to debug handler */ low_reset_branch = (xscale->handler_address + 0x20 - 0x0 - 0x8) >> 2; high_reset_branch = (xscale->handler_address + 0x20 - 0xffff0000 - 0x8) >> 2; xscale->low_vectors[0] = ARMV4_5_B((low_reset_branch & 0xffffff), 0); xscale->high_vectors[0] = ARMV4_5_B((high_reset_branch & 0xffffff), 0); for (i = 1; i <= 7; i++) { xscale->low_vectors[i] = ARMV4_5_B(0xfffffe, 0); xscale->high_vectors[i] = ARMV4_5_B(0xfffffe, 0); } /* 64kB aligned region used for DCache cleaning */ xscale->cache_clean_address = 0xfffe0000; xscale->hold_rst = 0; xscale->external_debug_break = 0; xscale->ibcr_available = 2; xscale->ibcr0_used = 0; xscale->ibcr1_used = 0; xscale->dbr_available = 2; xscale->dbr0_used = 0; xscale->dbr1_used = 0; xscale->arm_bkpt = ARMV5_BKPT(0x0); xscale->thumb_bkpt = ARMV5_T_BKPT(0x0) & 0xffff; xscale->vector_catch = 0x1; xscale->trace.capture_status = TRACE_IDLE; xscale->trace.data = NULL; xscale->trace.image = NULL; xscale->trace.buffer_enabled = 0; xscale->trace.buffer_fill = 0; /* prepare ARMv4/5 specific information */ armv4_5->arch_info = xscale; armv4_5->read_core_reg = xscale_read_core_reg; armv4_5->write_core_reg = xscale_write_core_reg; armv4_5->full_context = xscale_full_context; armv4_5_init_arch_info(target, armv4_5); xscale->armv4_5_mmu.armv4_5_cache.ctype = -1; xscale->armv4_5_mmu.get_ttb = xscale_get_ttb; xscale->armv4_5_mmu.read_memory = xscale_read_memory; xscale->armv4_5_mmu.write_memory = xscale_write_memory; xscale->armv4_5_mmu.disable_mmu_caches = xscale_disable_mmu_caches; xscale->armv4_5_mmu.enable_mmu_caches = xscale_enable_mmu_caches; xscale->armv4_5_mmu.has_tiny_pages = 1; xscale->armv4_5_mmu.mmu_enabled = 0; return ERROR_OK; } static int xscale_target_create(struct target *target, Jim_Interp *interp) { struct xscale_common *xscale; if (sizeof xscale_debug_handler - 1 > 0x800) { LOG_ERROR("debug_handler.bin: larger than 2kb"); return ERROR_FAIL; } xscale = calloc(1, sizeof(*xscale)); if (!xscale) return ERROR_FAIL; return xscale_init_arch_info(target, xscale, target->tap, target->variant); } COMMAND_HANDLER(xscale_handle_debug_handler_command) { struct target *target = NULL; struct xscale_common *xscale; int retval; uint32_t handler_address; if (CMD_ARGC < 2) { LOG_ERROR("'xscale debug_handler
' command takes two required operands"); return ERROR_OK; } if ((target = get_target(CMD_ARGV[0])) == NULL) { LOG_ERROR("target '%s' not defined", CMD_ARGV[0]); return ERROR_FAIL; } xscale = target_to_xscale(target); retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; COMMAND_PARSE_NUMBER(u32, CMD_ARGV[1], handler_address); if (((handler_address >= 0x800) && (handler_address <= 0x1fef800)) || ((handler_address >= 0xfe000800) && (handler_address <= 0xfffff800))) { xscale->handler_address = handler_address; } else { LOG_ERROR("xscale debug_handler
must be between 0x800 and 0x1fef800 or between 0xfe000800 and 0xfffff800"); return ERROR_FAIL; } return ERROR_OK; } COMMAND_HANDLER(xscale_handle_cache_clean_address_command) { struct target *target = NULL; struct xscale_common *xscale; int retval; uint32_t cache_clean_address; if (CMD_ARGC < 2) { return ERROR_COMMAND_SYNTAX_ERROR; } target = get_target(CMD_ARGV[0]); if (target == NULL) { LOG_ERROR("target '%s' not defined", CMD_ARGV[0]); return ERROR_FAIL; } xscale = target_to_xscale(target); retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; COMMAND_PARSE_NUMBER(u32, CMD_ARGV[1], cache_clean_address); if (cache_clean_address & 0xffff) { LOG_ERROR("xscale cache_clean_address
must be 64kb aligned"); } else { xscale->cache_clean_address = cache_clean_address; } return ERROR_OK; } COMMAND_HANDLER(xscale_handle_cache_info_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; return armv4_5_handle_cache_info_command(CMD_CTX, &xscale->armv4_5_mmu.armv4_5_cache); } static int xscale_virt2phys(struct target *target, uint32_t virtual, uint32_t *physical) { struct xscale_common *xscale = target_to_xscale(target); int type; uint32_t cb; int domain; uint32_t ap; if (xscale->common_magic != XSCALE_COMMON_MAGIC) { LOG_ERROR(xscale_not); return ERROR_TARGET_INVALID; } uint32_t ret = armv4_5_mmu_translate_va(target, &xscale->armv4_5_mmu, virtual, &type, &cb, &domain, &ap); if (type == -1) { return ret; } *physical = ret; return ERROR_OK; } static int xscale_mmu(struct target *target, int *enabled) { struct xscale_common *xscale = target_to_xscale(target); if (target->state != TARGET_HALTED) { LOG_ERROR("Target not halted"); return ERROR_TARGET_INVALID; } *enabled = xscale->armv4_5_mmu.mmu_enabled; return ERROR_OK; } COMMAND_HANDLER(xscale_handle_mmu_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (target->state != TARGET_HALTED) { command_print(CMD_CTX, "target must be stopped for \"%s\" command", CMD_NAME); return ERROR_OK; } if (CMD_ARGC >= 1) { bool enable; COMMAND_PARSE_ENABLE(CMD_ARGV[0], enable); if (enable) xscale_enable_mmu_caches(target, 1, 0, 0); else xscale_disable_mmu_caches(target, 1, 0, 0); xscale->armv4_5_mmu.mmu_enabled = enable; } command_print(CMD_CTX, "mmu %s", (xscale->armv4_5_mmu.mmu_enabled) ? "enabled" : "disabled"); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_idcache_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (target->state != TARGET_HALTED) { command_print(CMD_CTX, "target must be stopped for \"%s\" command", CMD_NAME); return ERROR_OK; } bool icache; COMMAND_PARSE_BOOL(CMD_NAME, icache, "icache", "dcache"); if (CMD_ARGC >= 1) { bool enable; COMMAND_PARSE_ENABLE(CMD_ARGV[0], enable); if (enable) xscale_enable_mmu_caches(target, 1, 0, 0); else xscale_disable_mmu_caches(target, 1, 0, 0); if (icache) xscale->armv4_5_mmu.armv4_5_cache.i_cache_enabled = enable; else xscale->armv4_5_mmu.armv4_5_cache.d_u_cache_enabled = enable; } bool enabled = icache ? xscale->armv4_5_mmu.armv4_5_cache.i_cache_enabled : xscale->armv4_5_mmu.armv4_5_cache.d_u_cache_enabled; const char *msg = enabled ? "enabled" : "disabled"; command_print(CMD_CTX, "%s %s", CMD_NAME, msg); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_vector_catch_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (CMD_ARGC < 1) { command_print(CMD_CTX, "usage: xscale vector_catch [mask]"); } else { COMMAND_PARSE_NUMBER(u8, CMD_ARGV[0], xscale->vector_catch); buf_set_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 16, 8, xscale->vector_catch); xscale_write_dcsr(target, -1, -1); } command_print(CMD_CTX, "vector catch mask: 0x%2.2x", xscale->vector_catch); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_vector_table_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int err = 0; int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (CMD_ARGC == 0) /* print current settings */ { int idx; command_print(CMD_CTX, "active user-set static vectors:"); for (idx = 1; idx < 8; idx++) if (xscale->static_low_vectors_set & (1 << idx)) command_print(CMD_CTX, "low %d: 0x%" PRIx32, idx, xscale->static_low_vectors[idx]); for (idx = 1; idx < 8; idx++) if (xscale->static_high_vectors_set & (1 << idx)) command_print(CMD_CTX, "high %d: 0x%" PRIx32, idx, xscale->static_high_vectors[idx]); return ERROR_OK; } if (CMD_ARGC != 3) err = 1; else { int idx; COMMAND_PARSE_NUMBER(int, CMD_ARGV[1], idx); uint32_t vec; COMMAND_PARSE_NUMBER(u32, CMD_ARGV[2], vec); if (idx < 1 || idx >= 8) err = 1; if (!err && strcmp(CMD_ARGV[0], "low") == 0) { xscale->static_low_vectors_set |= (1<static_low_vectors[idx] = vec; } else if (!err && (strcmp(CMD_ARGV[0], "high") == 0)) { xscale->static_high_vectors_set |= (1<static_high_vectors[idx] = vec; } else err = 1; } if (err) command_print(CMD_CTX, "usage: xscale vector_table "); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_trace_buffer_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); struct arm *armv4_5 = &xscale->armv4_5_common; uint32_t dcsr_value; int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (target->state != TARGET_HALTED) { command_print(CMD_CTX, "target must be stopped for \"%s\" command", CMD_NAME); return ERROR_OK; } if ((CMD_ARGC >= 1) && (strcmp("enable", CMD_ARGV[0]) == 0)) { struct xscale_trace_data *td, *next_td; xscale->trace.buffer_enabled = 1; /* free old trace data */ td = xscale->trace.data; while (td) { next_td = td->next; if (td->entries) free(td->entries); free(td); td = next_td; } xscale->trace.data = NULL; } else if ((CMD_ARGC >= 1) && (strcmp("disable", CMD_ARGV[0]) == 0)) { xscale->trace.buffer_enabled = 0; } if ((CMD_ARGC >= 2) && (strcmp("fill", CMD_ARGV[1]) == 0)) { uint32_t fill = 1; if (CMD_ARGC >= 3) COMMAND_PARSE_NUMBER(u32, CMD_ARGV[2], fill); xscale->trace.buffer_fill = fill; } else if ((CMD_ARGC >= 2) && (strcmp("wrap", CMD_ARGV[1]) == 0)) { xscale->trace.buffer_fill = -1; } if (xscale->trace.buffer_enabled) { /* if we enable the trace buffer in fill-once * mode we know the address of the first instruction */ xscale->trace.pc_ok = 1; xscale->trace.current_pc = buf_get_u32(armv4_5->core_cache->reg_list[15].value, 0, 32); } else { /* otherwise the address is unknown, and we have no known good PC */ xscale->trace.pc_ok = 0; } command_print(CMD_CTX, "trace buffer %s (%s)", (xscale->trace.buffer_enabled) ? "enabled" : "disabled", (xscale->trace.buffer_fill > 0) ? "fill" : "wrap"); dcsr_value = buf_get_u32(xscale->reg_cache->reg_list[XSCALE_DCSR].value, 0, 32); if (xscale->trace.buffer_fill >= 0) xscale_write_dcsr_sw(target, (dcsr_value & 0xfffffffc) | 2); else xscale_write_dcsr_sw(target, dcsr_value & 0xfffffffc); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_trace_image_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval; if (CMD_ARGC < 1) { command_print(CMD_CTX, "usage: xscale trace_image [base address] [type]"); return ERROR_OK; } retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (xscale->trace.image) { image_close(xscale->trace.image); free(xscale->trace.image); command_print(CMD_CTX, "previously loaded image found and closed"); } xscale->trace.image = malloc(sizeof(struct image)); xscale->trace.image->base_address_set = 0; xscale->trace.image->start_address_set = 0; /* a base address isn't always necessary, default to 0x0 (i.e. don't relocate) */ if (CMD_ARGC >= 2) { xscale->trace.image->base_address_set = 1; COMMAND_PARSE_NUMBER(int, CMD_ARGV[1], xscale->trace.image->base_address); } else { xscale->trace.image->base_address_set = 0; } if (image_open(xscale->trace.image, CMD_ARGV[0], (CMD_ARGC >= 3) ? CMD_ARGV[2] : NULL) != ERROR_OK) { free(xscale->trace.image); xscale->trace.image = NULL; return ERROR_OK; } return ERROR_OK; } COMMAND_HANDLER(xscale_handle_dump_trace_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); struct xscale_trace_data *trace_data; struct fileio file; int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (target->state != TARGET_HALTED) { command_print(CMD_CTX, "target must be stopped for \"%s\" command", CMD_NAME); return ERROR_OK; } if (CMD_ARGC < 1) { command_print(CMD_CTX, "usage: xscale dump_trace "); return ERROR_OK; } trace_data = xscale->trace.data; if (!trace_data) { command_print(CMD_CTX, "no trace data collected"); return ERROR_OK; } if (fileio_open(&file, CMD_ARGV[0], FILEIO_WRITE, FILEIO_BINARY) != ERROR_OK) { return ERROR_OK; } while (trace_data) { int i; fileio_write_u32(&file, trace_data->chkpt0); fileio_write_u32(&file, trace_data->chkpt1); fileio_write_u32(&file, trace_data->last_instruction); fileio_write_u32(&file, trace_data->depth); for (i = 0; i < trace_data->depth; i++) fileio_write_u32(&file, trace_data->entries[i].data | ((trace_data->entries[i].type & 0xffff) << 16)); trace_data = trace_data->next; } fileio_close(&file); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_analyze_trace_buffer_command) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; xscale_analyze_trace(target, CMD_CTX); return ERROR_OK; } COMMAND_HANDLER(xscale_handle_cp15) { struct target *target = get_current_target(CMD_CTX); struct xscale_common *xscale = target_to_xscale(target); int retval; retval = xscale_verify_pointer(CMD_CTX, xscale); if (retval != ERROR_OK) return retval; if (target->state != TARGET_HALTED) { command_print(CMD_CTX, "target must be stopped for \"%s\" command", CMD_NAME); return ERROR_OK; } uint32_t reg_no = 0; struct reg *reg = NULL; if (CMD_ARGC > 0) { COMMAND_PARSE_NUMBER(u32, CMD_ARGV[0], reg_no); /*translate from xscale cp15 register no to openocd register*/ switch (reg_no) { case 0: reg_no = XSCALE_MAINID; break; case 1: reg_no = XSCALE_CTRL; break; case 2: reg_no = XSCALE_TTB; break; case 3: reg_no = XSCALE_DAC; break; case 5: reg_no = XSCALE_FSR; break; case 6: reg_no = XSCALE_FAR; break; case 13: reg_no = XSCALE_PID; break; case 15: reg_no = XSCALE_CPACCESS; break; default: command_print(CMD_CTX, "invalid register number"); return ERROR_INVALID_ARGUMENTS; } reg = &xscale->reg_cache->reg_list[reg_no]; } if (CMD_ARGC == 1) { uint32_t value; /* read cp15 control register */ xscale_get_reg(reg); value = buf_get_u32(reg->value, 0, 32); command_print(CMD_CTX, "%s (/%i): 0x%" PRIx32 "", reg->name, (int)(reg->size), value); } else if (CMD_ARGC == 2) { uint32_t value; COMMAND_PARSE_NUMBER(u32, CMD_ARGV[1], value); /* send CP write request (command 0x41) */ xscale_send_u32(target, 0x41); /* send CP register number */ xscale_send_u32(target, reg_no); /* send CP register value */ xscale_send_u32(target, value); /* execute cpwait to ensure outstanding operations complete */ xscale_send_u32(target, 0x53); } else { command_print(CMD_CTX, "usage: cp15 [register]<, [value]>"); } return ERROR_OK; } static const struct command_registration xscale_exec_command_handlers[] = { { .name = "cache_info", .handler = &xscale_handle_cache_info_command, .mode = COMMAND_EXEC, NULL, }, { .name = "mmu", .handler = &xscale_handle_mmu_command, .mode = COMMAND_EXEC, .usage = "[enable|disable]", .help = "enable or disable the MMU", }, { .name = "icache", .handler = &xscale_handle_idcache_command, .mode = COMMAND_EXEC, .usage = "[enable|disable]", .help = "enable or disable the ICache", }, { .name = "dcache", .handler = &xscale_handle_idcache_command, .mode = COMMAND_EXEC, .usage = "[enable|disable]", .help = "enable or disable the DCache", }, { .name = "vector_catch", .handler = &xscale_handle_vector_catch_command, .mode = COMMAND_EXEC, .help = "mask of vectors that should be caught", .usage = "[]", }, { .name = "vector_table", .handler = &xscale_handle_vector_table_command, .mode = COMMAND_EXEC, .usage = " ", .help = "set static code for exception handler entry", }, { .name = "trace_buffer", .handler = &xscale_handle_trace_buffer_command, .mode = COMMAND_EXEC, .usage = " [fill [n]|wrap]", }, { .name = "dump_trace", .handler = &xscale_handle_dump_trace_command, .mode = COMMAND_EXEC, .help = "dump content of trace buffer to ", .usage = "", }, { .name = "analyze_trace", .handler = &xscale_handle_analyze_trace_buffer_command, .mode = COMMAND_EXEC, .help = "analyze content of trace buffer", }, { .name = "trace_image", .handler = &xscale_handle_trace_image_command, COMMAND_EXEC, .help = "load image from [base address]", .usage = " [address] [type]", }, { .name = "cp15", .handler = &xscale_handle_cp15, .mode = COMMAND_EXEC, .help = "access coproc 15", .usage = " [value]", }, COMMAND_REGISTRATION_DONE }; static const struct command_registration xscale_any_command_handlers[] = { { .name = "debug_handler", .handler = &xscale_handle_debug_handler_command, .mode = COMMAND_ANY, .usage = "
", }, { .name = "cache_clean_address", .handler = &xscale_handle_cache_clean_address_command, .mode = COMMAND_ANY, }, { .chain = xscale_exec_command_handlers, }, COMMAND_REGISTRATION_DONE }; static const struct command_registration xscale_command_handlers[] = { { .chain = arm_command_handlers, }, { .name = "xscale", .mode = COMMAND_ANY, .help = "xscale command group", .chain = xscale_any_command_handlers, }, COMMAND_REGISTRATION_DONE }; struct target_type xscale_target = { .name = "xscale", .poll = xscale_poll, .arch_state = xscale_arch_state, .target_request_data = NULL, .halt = xscale_halt, .resume = xscale_resume, .step = xscale_step, .assert_reset = xscale_assert_reset, .deassert_reset = xscale_deassert_reset, .soft_reset_halt = NULL, .get_gdb_reg_list = armv4_5_get_gdb_reg_list, .read_memory = xscale_read_memory, .read_phys_memory = xscale_read_phys_memory, .write_memory = xscale_write_memory, .write_phys_memory = xscale_write_phys_memory, .bulk_write_memory = xscale_bulk_write_memory, .checksum_memory = arm_checksum_memory, .blank_check_memory = arm_blank_check_memory, .run_algorithm = armv4_5_run_algorithm, .add_breakpoint = xscale_add_breakpoint, .remove_breakpoint = xscale_remove_breakpoint, .add_watchpoint = xscale_add_watchpoint, .remove_watchpoint = xscale_remove_watchpoint, .commands = xscale_command_handlers, .target_create = xscale_target_create, .init_target = xscale_init_target, .virt2phys = xscale_virt2phys, .mmu = xscale_mmu };