]> git.sur5r.net Git - openocd/commitdiff
Added support for ARMv7-M in arm io.
authorHenrik Nilsson <henrik.nilsson@bytequest.se>
Wed, 27 Feb 2013 18:16:47 +0000 (19:16 +0100)
committerSpencer Oliver <spen@spen-soft.co.uk>
Tue, 2 Apr 2013 15:11:02 +0000 (15:11 +0000)
Added support for ARMv7-M targets in arm_nandwrite and
arm_nandread.

Change-Id: Iab1d78d401f735e191c6a8519f3619035a300fae
Signed-off-by: Henrik Nilsson <henrik.nilsson@bytequest.se>
Reviewed-on: http://openocd.zylin.com/1188
Tested-by: jenkins
Reviewed-by: Freddie Chopin <freddie.chopin@gmail.com>
contrib/loaders/flash/armv7m_io.s [new file with mode: 0644]
src/flash/nand/arm_io.c

diff --git a/contrib/loaders/flash/armv7m_io.s b/contrib/loaders/flash/armv7m_io.s
new file mode 100644 (file)
index 0000000..3bd90ca
--- /dev/null
@@ -0,0 +1,60 @@
+/***************************************************************************
+ *   Copyright (C) 2013 by Henrik Nilsson                                  *
+ *   henrik.nilsson@bytequest.se                                           *
+ *                                                                         *
+ *   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.             *
+ ***************************************************************************/
+
+       .text
+       .syntax unified
+       .arch armv7-m
+       .thumb
+       .thumb_func
+
+       .align 4
+
+/* Inputs:
+ *  r0 buffer address
+ *  r1 NAND data address (byte wide)
+ *  r2 buffer length
+ */
+read:
+       ldrb    r3, [r1]
+       strb    r3, [r0], #1
+       subs    r2, r2, #1
+       bne             read
+
+done_read:
+       bkpt #0
+
+       .align 4
+
+/* Inputs:
+ *  r0 NAND data address (byte wide)
+ *  r1 buffer address
+ *  r2 buffer length
+ */
+write:
+       ldrb    r3, [r1], #1
+       strb    r3, [r0]
+       subs    r2, r2, #1
+       bne             write
+
+done_write:
+       bkpt #0
+
+       .end
+
index aab1d050f4bc8a75d6e93b5bca13f34ef47d3c58..c22954c9403360b24a43c7f65cf54b3838ec41ca 100644 (file)
@@ -28,6 +28,7 @@
 #include "arm_io.h"
 #include <helper/binarybuffer.h>
 #include <target/arm.h>
+#include <target/armv7m.h>
 #include <target/algorithm.h>
 
 /**
@@ -78,14 +79,13 @@ static int arm_code_to_working_area(struct target *target,
 
 /**
  * ARM-specific bulk write from buffer to address of 8-bit wide NAND.
- * For now this only supports ARMv4 and ARMv5 cores.
+ * For now this supports ARMv4,ARMv5 and ARMv7-M cores.
  *
  * Enhancements to target_run_algorithm() could enable:
  *   - ARMv6 and ARMv7 cores in ARM mode
  *
  * Different code fragments could handle:
- *   - Thumb2 cores like Cortex-M (needs different byteswapping)
- *   - 16-bit wide data (needs different setup too)
+ *   - 16-bit wide data (needs different setup)
  *
  * @param nand Pointer to the arm_nand_data struct that defines the I/O
  * @param data Pointer to the data to be copied to flash
@@ -95,7 +95,9 @@ static int arm_code_to_working_area(struct target *target,
 int arm_nandwrite(struct arm_nand_data *nand, uint8_t *data, int size)
 {
        struct target *target = nand->target;
-       struct arm_algorithm algo;
+       struct arm_algorithm armv4_5_algo;
+       struct armv7m_algorithm armv7m_algo;
+       void *arm_algo;
        struct arm *arm = target->arch_info;
        struct reg_param reg_params[3];
        uint32_t target_buf;
@@ -107,7 +109,7 @@ int arm_nandwrite(struct arm_nand_data *nand, uint8_t *data, int size)
         *  r1  buffer address
         *  r2  buffer length
         */
-       static const uint32_t code[] = {
+       static const uint32_t code_armv4_5[] = {
                0xe4d13001,     /* s: ldrb  r3, [r1], #1 */
                0xe5c03000,     /*    strb  r3, [r0]     */
                0xe2522001,     /*    subs  r2, r2, #1   */
@@ -117,8 +119,41 @@ int arm_nandwrite(struct arm_nand_data *nand, uint8_t *data, int size)
                0xe1200070,     /* e: bkpt  #0           */
        };
 
+       /* Inputs:
+        *  r0  NAND data address (byte wide)
+        *  r1  buffer address
+        *  r2  buffer length
+        *
+        * see contrib/loaders/flash/armv7m_io.s for src
+        */
+       static const uint32_t code_armv7m[] = {
+               0x3b01f811,
+               0x3a017003,
+               0xaffaf47f,
+               0xbf00be00,
+       };
+
+       int target_code_size = 0;
+       const uint32_t *target_code_src = NULL;
+
+       /* set up algorithm */
+       if (is_armv7m(target_to_armv7m(target))) {  /* armv7m target */
+               armv7m_algo.common_magic = ARMV7M_COMMON_MAGIC;
+               armv7m_algo.core_mode = ARM_MODE_THREAD;
+               arm_algo = &armv7m_algo;
+               target_code_size = sizeof(code_armv7m);
+               target_code_src = code_armv7m;
+       } else {
+               armv4_5_algo.common_magic = ARM_COMMON_MAGIC;
+               armv4_5_algo.core_mode = ARM_MODE_SVC;
+               armv4_5_algo.core_state = ARM_STATE_ARM;
+               arm_algo = &armv4_5_algo;
+               target_code_size = sizeof(code_armv4_5);
+               target_code_src = code_armv4_5;
+       }
+
        if (nand->op != ARM_NAND_WRITE || !nand->copy_area) {
-               retval = arm_code_to_working_area(target, code, sizeof(code),
+               retval = arm_code_to_working_area(target, target_code_src, target_code_size,
                                nand->chunk_size, &nand->copy_area);
                if (retval != ERROR_OK)
                        return retval;
@@ -127,16 +162,12 @@ int arm_nandwrite(struct arm_nand_data *nand, uint8_t *data, int size)
        nand->op = ARM_NAND_WRITE;
 
        /* copy data to work area */
-       target_buf = nand->copy_area->address + sizeof(code);
+       target_buf = nand->copy_area->address + target_code_size;
        retval = target_write_buffer(target, target_buf, size, data);
        if (retval != ERROR_OK)
                return retval;
 
-       /* set up algorithm and parameters */
-       algo.common_magic = ARM_COMMON_MAGIC;
-       algo.core_mode = ARM_MODE_SVC;
-       algo.core_state = ARM_STATE_ARM;
-
+       /* set up parameters */
        init_reg_param(&reg_params[0], "r0", 32, PARAM_IN);
        init_reg_param(&reg_params[1], "r1", 32, PARAM_IN);
        init_reg_param(&reg_params[2], "r2", 32, PARAM_IN);
@@ -147,11 +178,11 @@ int arm_nandwrite(struct arm_nand_data *nand, uint8_t *data, int size)
 
        /* armv4 must exit using a hardware breakpoint */
        if (arm->is_armv4)
-               exit_var = nand->copy_area->address + sizeof(code) - 4;
+               exit_var = nand->copy_area->address + target_code_size - 4;
 
        /* use alg to write data from work area to NAND chip */
        retval = target_run_algorithm(target, 0, NULL, 3, reg_params,
-                       nand->copy_area->address, exit_var, 1000, &algo);
+                       nand->copy_area->address, exit_var, 1000, arm_algo);
        if (retval != ERROR_OK)
                LOG_ERROR("error executing hosted NAND write");
 
@@ -174,7 +205,9 @@ int arm_nandwrite(struct arm_nand_data *nand, uint8_t *data, int size)
 int arm_nandread(struct arm_nand_data *nand, uint8_t *data, uint32_t size)
 {
        struct target *target = nand->target;
-       struct arm_algorithm algo;
+       struct arm_algorithm armv4_5_algo;
+       struct armv7m_algorithm armv7m_algo;
+       void *arm_algo;
        struct arm *arm = target->arch_info;
        struct reg_param reg_params[3];
        uint32_t target_buf;
@@ -186,7 +219,7 @@ int arm_nandread(struct arm_nand_data *nand, uint8_t *data, uint32_t size)
         *  r1  NAND data address (byte wide)
         *  r2  buffer length
         */
-       static const uint32_t code[] = {
+       static const uint32_t code_armv4_5[] = {
                0xe5d13000,     /* s: ldrb  r3, [r1]     */
                0xe4c03001,     /*    strb  r3, [r0], #1 */
                0xe2522001,     /*    subs  r2, r2, #1   */
@@ -196,22 +229,51 @@ int arm_nandread(struct arm_nand_data *nand, uint8_t *data, uint32_t size)
                0xe1200070,     /* e: bkpt  #0           */
        };
 
+       /* Inputs:
+        *  r0  buffer address
+        *  r1  NAND data address (byte wide)
+        *  r2  buffer length
+        *
+        * see contrib/loaders/flash/armv7m_io.s for src
+        */
+       static const uint32_t code_armv7m[] = {
+               0xf800780b,
+               0x3a013b01,
+               0xaffaf47f,
+               0xbf00be00,
+       };
+
+       int target_code_size = 0;
+       const uint32_t *target_code_src = NULL;
+
+       /* set up algorithm */
+       if (is_armv7m(target_to_armv7m(target))) {  /* armv7m target */
+               armv7m_algo.common_magic = ARMV7M_COMMON_MAGIC;
+               armv7m_algo.core_mode = ARM_MODE_THREAD;
+               arm_algo = &armv7m_algo;
+               target_code_size = sizeof(code_armv7m);
+               target_code_src = code_armv7m;
+       } else {
+               armv4_5_algo.common_magic = ARM_COMMON_MAGIC;
+               armv4_5_algo.core_mode = ARM_MODE_SVC;
+               armv4_5_algo.core_state = ARM_STATE_ARM;
+               arm_algo = &armv4_5_algo;
+               target_code_size = sizeof(code_armv4_5);
+               target_code_src = code_armv4_5;
+       }
+
        /* create the copy area if not yet available */
        if (nand->op != ARM_NAND_READ || !nand->copy_area) {
-               retval = arm_code_to_working_area(target, code, sizeof(code),
+               retval = arm_code_to_working_area(target, target_code_src, target_code_size,
                                nand->chunk_size, &nand->copy_area);
                if (retval != ERROR_OK)
                        return retval;
        }
 
        nand->op = ARM_NAND_READ;
-       target_buf = nand->copy_area->address + sizeof(code);
-
-       /* set up algorithm and parameters */
-       algo.common_magic = ARM_COMMON_MAGIC;
-       algo.core_mode = ARM_MODE_SVC;
-       algo.core_state = ARM_STATE_ARM;
+       target_buf = nand->copy_area->address + target_code_size;
 
+       /* set up parameters */
        init_reg_param(&reg_params[0], "r0", 32, PARAM_IN);
        init_reg_param(&reg_params[1], "r1", 32, PARAM_IN);
        init_reg_param(&reg_params[2], "r2", 32, PARAM_IN);
@@ -222,11 +284,11 @@ int arm_nandread(struct arm_nand_data *nand, uint8_t *data, uint32_t size)
 
        /* armv4 must exit using a hardware breakpoint */
        if (arm->is_armv4)
-               exit_var = nand->copy_area->address + sizeof(code) - 4;
+               exit_var = nand->copy_area->address + target_code_size - 4;
 
        /* use alg to write data from NAND chip to work area */
        retval = target_run_algorithm(target, 0, NULL, 3, reg_params,
-                       nand->copy_area->address, exit_var, 1000, &algo);
+                       nand->copy_area->address, exit_var, 1000, arm_algo);
        if (retval != ERROR_OK)
                LOG_ERROR("error executing hosted NAND read");