Replace byte-access to memory with faster word-access
authorLuca Ellero <lroluk@gmail.com>
Wed, 13 Apr 2011 18:55:19 +0000 (18:55 +0000)
committerØyvind Harboe <oyvind.harboe@zylin.com>
Wed, 13 Apr 2011 19:33:11 +0000 (21:33 +0200)
Freescale iMX53 doesn't seem to like unaligned accesses to his memory
mapped registers.
Anyway this patch makes dump_image/load_image 4X faster for every
access through APB.

Signed-off-by: Luca Ellero <lroluk@gmail.com>
src/target/cortex_a.c [changed mode: 0644->0755]

old mode 100644 (file)
new mode 100755 (executable)
index 86706cc..934f75a
@@ -1450,42 +1450,76 @@ static int cortex_a8_deassert_reset(struct target *target)
        return ERROR_OK;
 }
 
+
 static int cortex_a8_write_apb_ab_memory(struct target *target,
                 uint32_t address, uint32_t size,
                 uint32_t count, const uint8_t *buffer)
 {
+
+       /* write memory through APB-AP */
+
        int retval = ERROR_INVALID_ARGUMENTS;
        struct armv7a_common *armv7a = target_to_armv7a(target);
        struct arm *armv4_5 = &armv7a->armv4_5_common;
-       int nbytes = count * size;
-       uint32_t data;
+       int total_bytes = count * size;
+       int start_byte, nbytes_to_write, i;
        struct reg *reg;
+       union _data {
+               uint8_t uc_a[4];
+               uint32_t ui;
+       } data;
 
        if (target->state != TARGET_HALTED)
        {
                LOG_WARNING("target not halted");
                return ERROR_TARGET_NOT_HALTED;
        }
+
        reg = arm_reg_current(armv4_5, 0);
        reg->dirty = 1;
        reg = arm_reg_current(armv4_5, 1);
        reg->dirty = 1;
-       retval = cortex_a8_dap_write_coreregister_u32(target, address, 0);
+
+       retval = cortex_a8_dap_write_coreregister_u32(target, address & 0xFFFFFFFC, 0);
        if (retval != ERROR_OK)
                return retval;
 
-       while (nbytes > 0) {
-               data = *buffer++;
-               retval = cortex_a8_dap_write_coreregister_u32(target, data, 1);
+       start_byte = address & 0x3;
+
+       while (total_bytes > 0) {
+
+               nbytes_to_write = 4 - start_byte;
+               if (total_bytes < nbytes_to_write)
+                       nbytes_to_write = total_bytes; 
+                       
+               if ( nbytes_to_write != 4 ) {
+               
+                       /* execute instruction LDR r1, [r0] */
+                       retval = cortex_a8_exec_opcode(target,  ARMV4_5_LDR(1, 0), NULL);
+                       if (retval != ERROR_OK)
+                               return retval;
+
+                       retval = cortex_a8_dap_read_coreregister_u32(target, &data.ui, 1);
+                       if (retval != ERROR_OK)
+                               return retval;
+               }
+       
+               for (i = 0; i < nbytes_to_write; ++i)
+                       data.uc_a[i + start_byte] = *buffer++;
+
+               retval = cortex_a8_dap_write_coreregister_u32(target, data.ui, 1);
                if (retval != ERROR_OK)
                        return retval;
 
-               /* execute instruction STRB r1, [r0], 1 (0xe4c01001) */
-               retval = cortex_a8_exec_opcode(target, ARMV4_5_STRB_IP(1, 0) , NULL);
+               /* execute instruction STRW r1, [r0], 1 (0xe4801004) */
+               retval = cortex_a8_exec_opcode(target, ARMV4_5_STRW_IP(1, 0) , NULL);
                if (retval != ERROR_OK)
-                       return retval;
-               --nbytes;
+                               return retval;
+
+               total_bytes -= nbytes_to_write;
+               start_byte = 0;
        }
+
        return retval;
 }
 
@@ -1494,13 +1528,19 @@ static int cortex_a8_read_apb_ab_memory(struct target *target,
                 uint32_t address, uint32_t size,
                 uint32_t count, uint8_t *buffer)
 {
+
+       /* read memory through APB-AP */
+
        int retval = ERROR_INVALID_ARGUMENTS;
        struct armv7a_common *armv7a = target_to_armv7a(target);
        struct arm *armv4_5 = &armv7a->armv4_5_common;
-       /* read memory through APB-AP */
-       int nbytes = count * size;
-       uint32_t data;
+       int total_bytes = count * size;
+       int start_byte, nbytes_to_read, i;
        struct reg *reg;
+       union _data {
+               uint8_t uc_a[4];
+               uint32_t ui;
+       } data;
 
        if (target->state != TARGET_HALTED)
        {
@@ -1513,26 +1553,34 @@ static int cortex_a8_read_apb_ab_memory(struct target *target,
        reg = arm_reg_current(armv4_5, 1);
        reg->dirty = 1;
 
-       retval = cortex_a8_dap_write_coreregister_u32(target, address, 0);
+       retval = cortex_a8_dap_write_coreregister_u32(target, address & 0xFFFFFFFC, 0);
        if (retval != ERROR_OK)
                return retval;
 
-       while (nbytes > 0) {
+       start_byte = address & 0x3;
 
+       while (total_bytes > 0) {
 
-               /* execute instruction LDRB r1, [r0], 1 (0xe4d01001) */
-               retval = cortex_a8_exec_opcode(target, ARMV4_5_LDRB_IP(1, 0) , NULL);
+               /* execute instruction LDRW r1, [r0], 4 (0xe4901004)  */
+               retval = cortex_a8_exec_opcode(target,  ARMV4_5_LDRW_IP(1, 0), NULL);
                if (retval != ERROR_OK)
                        return retval;
 
-               retval = cortex_a8_dap_read_coreregister_u32(target, &data, 1);
+               retval = cortex_a8_dap_read_coreregister_u32(target, &data.ui, 1);
                if (retval != ERROR_OK)
                        return retval;
 
-               *buffer++ = data;
-               --nbytes;
-
+               nbytes_to_read = 4 - start_byte;
+               if (total_bytes < nbytes_to_read)
+                       nbytes_to_read = total_bytes; 
+       
+               for (i = 0; i < nbytes_to_read; ++i)
+                       *buffer++ = data.uc_a[i + start_byte];
+                       
+               total_bytes -= nbytes_to_read;
+               start_byte = 0;
        }
+
        return retval;
 }
 

Linking to existing account procedure

If you already have an account and want to add another login method you MUST first sign in with your existing account and then change URL to read https://review.openocd.org/login/?link to get to this page again but this time it'll work for linking. Thank you.

SSH host keys fingerprints

1024 SHA256:YKx8b7u5ZWdcbp7/4AeXNaqElP49m6QrwfXaqQGJAOk gerrit-code-review@openocd.zylin.com (DSA)
384 SHA256:jHIbSQa4REvwCFG4cq5LBlBLxmxSqelQPem/EXIrxjk gerrit-code-review@openocd.org (ECDSA)
521 SHA256:UAOPYkU9Fjtcao0Ul/Rrlnj/OsQvt+pgdYSZ4jOYdgs gerrit-code-review@openocd.org (ECDSA)
256 SHA256:A13M5QlnozFOvTllybRZH6vm7iSt0XLxbA48yfc2yfY gerrit-code-review@openocd.org (ECDSA)
256 SHA256:spYMBqEYoAOtK7yZBrcwE8ZpYt6b68Cfh9yEVetvbXg gerrit-code-review@openocd.org (ED25519)
+--[ED25519 256]--+
|=..              |
|+o..   .         |
|*.o   . .        |
|+B . . .         |
|Bo. = o S        |
|Oo.+ + =         |
|oB=.* = . o      |
| =+=.+   + E     |
|. .=o   . o      |
+----[SHA256]-----+
2048 SHA256:0Onrb7/PHjpo6iVZ7xQX2riKN83FJ3KGU0TvI0TaFG4 gerrit-code-review@openocd.zylin.com (RSA)