Hello Colin,
I don't see the point of making an advanced disassembler out of
OpenOCD. I think the main purpose of adding the routine was to do some
spot checks during debugging. In fact, it's probably not really
necessary once gdb interface was implemented, as gdb can do
disassembly itself.
If you need to disassemble a lot of code "offline", it's much better
to use specialized tools such as objdump or IDA Pro or whatnot.
On Sun, Oct 25, 2009 at 01:36, Colin Howarth <co...@howarth.de> wrote:
> Hello all,
> I'm not entirely sure this is appropriate here, it being an OpenOCD
> developer's list and not an ARM developer's list as such :-(
> However, I was wondering how many of you actually code in ARM assembler? Do
> you avoid it wherever you can, using a "high-level" language and gcc?
> For those that do do assembly code, do you have a Good (TM) disassembler?
>
> The armv4_5 disassemble command is fine, in that it disassembles, but as
> I mentioned a while back, it doesn't produce what I would have written as an
> assembly code routine :-)
>
> So, I've written a quick 2-pass assembly code post-processor in Perl, and
> was wondering if anyone else had any interest in it.
> So far, it does this:
> a) identifies subroutine calls (ie. BL)
> b) identifies returns (e.g. BX R14; LDMFD R13!, {r0, r1, r15}; )
> c) identifies labels (e.g BNE 0x00001234)
> d) identifies PC-relative addressed data
> e) locates the definition of the base register (e.g. LDR r0, [r15, #0x18];
> ... STR r0, [r2] )
> It prints out the marked up assembly code on the second pass.
>
> Thus:
>
> 0x0000072c 0xe59f3088 LDR r3, [r15, #0x88]
> 0x00000730 0xe1540003 CMP r4, r3
> 0x00000734 0x1a000009 BNE 0x00000760
> 0x00000738 0xe3a00501 MOV r0, #0x400000
> 0x0000073c 0xe3a01001 MOV r1, #0x1
> 0x00000740 0xeb000174 BL 0x00000d18
> 0x00000744 0xe3a00501 MOV r0, #0x400000
> 0x00000748 0xe3a01000 MOV r1, #0x0
> 0x0000074c 0xeb000171 BL 0x00000d18
> 0x00000750 0xe3a02000 MOV r2, #0x0
> 0x00000754 0xe59f303c LDR r3, [r15, #0x3c]
> 0x00000758 0xe58320ac STR r2, [r3, #0xac]
> 0x0000075c 0xe89da810 LDMIA r13, {r4, r11, r13, r15}
> 0x00000760 0xe59f3058 LDR r3, [r15, #0x58]
> 0x00000764 0xe1540003 CMP r4, r3
> 0x00000768 0x189da810 LDMNEIA r13, {r4, r11, r13, r15}
> 0x0000076c 0xe3a00502 MOV r0, #0x800000
> 0x00000770 0xe3a01001 MOV r1, #0x1
> 0x00000774 0xeb000167 BL 0x00000d18
> 0x00000778 0xe3a00502 MOV r0, #0x800000
> 0x0000077c 0xe3a01000 MOV r1, #0x0
> 0x00000780 0xeb000164 BL 0x00000d18
> 0x00000784 0xe3a02000 MOV r2, #0x0
> 0x00000788 0xe59f3008 LDR r3, [r15, #0x8]
> 0x0000078c 0xe58320ac STR r2, [r3, #0xac]
> 0x00000790 0xe89da810 LDMIA r13, {r4, r11, r13, r15}
> 0x00000794 0x58006000 STMPLDA r0, {r13, r14
> 0x00000798 0x5c002000 UNDEFINED INSTRUCTION
> 0x0000079c 0x58007000 STMPLDA r0, {r12, r13, r14}
>
> Becomes:
>
> 0x0000072c 0xe59f3088 LDR r3, [r15, #0x88] ; data at
> 0x000007bc: 0x5800e000
> 0x00000730 0xe1540003 CMP r4, r3
> 0x00000734 0x1a000009 BNE 0x00000760 ; branch to
> label_26 at: 0x00000760
> 0x00000738 0xe3a00501 MOV r0, #0x400000
> 0x0000073c 0xe3a01001 MOV r1, #0x1
> 0x00000740 0xeb000174 BL 0x00000d18 ; call
> subroutine_7 at: 0x00000d18
> 0x00000744 0xe3a00501 MOV r0, #0x400000
> 0x00000748 0xe3a01000 MOV r1, #0x0
> 0x0000074c 0xeb000171 BL 0x00000d18 ; call
> subroutine_7 at: 0x00000d18
> 0x00000750 0xe3a02000 MOV r2, #0x0
> 0x00000754 0xe59f303c LDR r3, [r15, #0x3c] ; data at
> 0x00000798: 0x5c002000
> 0x00000758 0xe58320ac STR r2, [r3, #0xac] ; Base defined at
> 0x00000754
> 0x0000075c 0xe89da810 LDMIA r13, {r4, r11, r13, r15} ; Return from
> subroutine
> label_26:
> 0x00000760 0xe59f3058 LDR r3, [r15, #0x58] ; data at
> 0x000007c0: 0x5800f000
> 0x00000764 0xe1540003 CMP r4, r3
> 0x00000768 0x189da810 LDMNEIA r13, {r4, r11, r13, r15} ; Return from
> subroutine
> 0x0000076c 0xe3a00502 MOV r0, #0x800000
> 0x00000770 0xe3a01001 MOV r1, #0x1
> 0x00000774 0xeb000167 BL 0x00000d18 ; call
> subroutine_7 at: 0x00000d18
> 0x00000778 0xe3a00502 MOV r0, #0x800000
> 0x0000077c 0xe3a01000 MOV r1, #0x0
> 0x00000780 0xeb000164 BL 0x00000d18 ; call
> subroutine_7 at: 0x00000d18
> 0x00000784 0xe3a02000 MOV r2, #0x0
> 0x00000788 0xe59f3008 LDR r3, [r15, #0x8] ; data at
> 0x00000798: 0x5c002000
> 0x0000078c 0xe58320ac STR r2, [r3, #0xac] ; Base defined at
> 0x00000788
> 0x00000790 0xe89da810 LDMIA r13, {r4, r11, r13, r15} ; Return from
> subroutine
> 0x00000794 0x58006000 DATA <STMPLDA r0, {r13, r14}>
> 0x00000798 0x5c002000 DATA <UNDEFINED INSTRUCTION>
> 0x0000079c 0x58007000 DATA <STMPLDA r0, {r12, r13, r14}>
>
> Which should soon become:
>
> 0x0000072c 0xe59f3088 LDR r3, data_1 <0x5800e000 =? non_buffered APB0 +
> GPIO_PORT_P8>
> 0x00000730 0xe1540003 CMP r4, r3
> 0x00000734 0x1a000009 BNE label_26
> 0x00000738 0xe3a00501 MOV r0, #0x400000
> 0x0000073c 0xe3a01001 MOV r1, #0x1
> 0x00000740 0xeb000174 BL subroutine_7
> 0x00000744 0xe3a00501 MOV r0, #0x400000
> 0x00000748 0xe3a01000 MOV r1, #0x0
> 0x0000074c 0xeb000171 BL subroutine_7
> 0x00000750 0xe3a02000 MOV r2, #0x0
> 0x00000754 0xe59f303c LDR r3, data_2 <0x5c002000 =? non_buffered APB1 +
> SCU>
> 0x00000758 0xe58320ac STR r2, [NB_APB1_SCU, SCU_GPIOEMI]
> 0x0000075c 0xe89da810 LDMIA r13, {r4, r11, r13, r15} ; Return from
> subroutine
> label_26:
> 0x00000760 0xe59f3058 LDR r3, data_3 <0x5800f000 =? non_buffered APB0 +
> GPIO_PORT_P9>
> 0x00000764 0xe1540003 CMP r4, r3
> 0x00000768 0x189da810 LDMNEIA r13, {r4, r11, r13, r15} ; Return from
> subroutine
> 0x0000076c 0xe3a00502 MOV r0, #0x800000
> 0x00000770 0xe3a01001 MOV r1, #0x1
> 0x00000774 0xeb000167 BL subroutine_7
> 0x00000778 0xe3a00502 MOV r0, #0x800000
> 0x0000077c 0xe3a01000 MOV r1, #0x0
> 0x00000780 0xeb000164 BL subroutine_7
> 0x00000784 0xe3a02000 MOV r2, #0x0
> 0x00000788 0xe59f3008 LDR r3, data_4 <0x5c002000 =? non_buffered APB1 +
> SCU>
> 0x0000078c 0xe58320ac STR r2, [NB_APB1_SCU, SCU_GPIOEMI]
> 0x00000790 0xe89da810 LDMIA r13, {r4, r11, r13, r15} ; Return from
> subroutine
> 0x00000794 0x58006000 DATA <STMPLDA r0, {r13, r14}>
> 0x00000798 0x5c002000 DATA <UNDEFINED INSTRUCTION>
> 0x0000079c 0x58007000 DATA <STMPLDA r0, {r12, r13, r14}>
>
>
>
> --colin
> _______________________________________________
> Openocd-development mailing list
> Openocd-development@lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/openocd-development
>
>
--
WBR, Igor
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
