Hi guys. Its been a while since I've posted, yet I still browse the forums daily.
I'm working on an STM320F103E project which interfaces with an FPGA via memory mapping.
The code is functional, but 'for some reason' the processor is accessing memory at 0x68008000 and at 0x6800A000 as observed in ChipScope within the Xilinx FPGA.
As you know by now, I really hate not knowing everything that my code is doing, so this really upsets me.
Although the interface to the FPGA is in BANK3 at 0x68018000 & 0x68018002, the code is accessing memory that I don't know about or remember yet.
All of this code works. (No problems here) I just wanted to share the basic access performed: I don't read/write externally anywhere else (that I know) typedef volatile unsigned int vu16; typedef volatile unsigned long vu32; typedef unsigned long u32; #define BANK3_PSRAM ( (vu32)0x68000000 ) // 0x6800.0000-0x68FF.FFFF #define FPGA_BASE_ADDRESS ( BANK3_PSRAM + 0x00000000 ) // FPGA mapped base #define FPGA_ROTARY_DIRECT_DATA_LSW *( (vu32 *) (u32) ( FPGA_BASE_ADDRESS + 0x00018000 ) ) #define FPGA_ROTARY_DIRECT_DATA_MSW *( (vu32 *) (u32) ( FPGA_BASE_ADDRESS + 0x00018002 ) ) vu32 The_Encoder_Value; // 'global' for testing void FPGA_Test_RW( void ) // redacted routine for this post { vu16 rotary_encoder_lsw; // 16-bit external access to FPGA (least significant word) vu16 rotary_encoder_msw; // 16-bit external access to FPGA (most significant word) rotary_encoder_lsw = FPGA_ROTARY_DIRECT_DATA_LSW; rotary_encoder_msw = FPGA_ROTARY_DIRECT_DATA_MSW; The_Encoder_Value = ( ((u32) rotary_encoder_msw)<<16 ) | rotary_encoder_lsw; }
I'd like to set a break point on writing or reading to/from 0x68008000 & 0x6800A000, but the Expression Help doesn't explain the syntax well enough.
Yes, I did a 'little' research first, but decided to ask the forum how to do this, or better yet, provide a link that explains the Expression syntax better.
--Cpt. Vince Foster 2nd Cannon Place Fort Marcy Park, VA
I have a stupid question.
typedef volatile unsigned int vu16;
Is the size of unsigned int 16-bits on the mentioned platform?
I prefer to use uint8_t, uint16_t, uint32_t, ... when I need variables of fixed size. If the platform don't have a stdint.h file, then I declare my own "layer 0" data types.
Linq,
Yes. I did.
Thanks to you, I had to run it down in the libraries. As a result, I found that Keil's "rt_Typedef.h" defines it this way:
FILE: C:\IDEs\Keil\ARM\RL\RXT\SRC\ARM\rt_Typedef.h /* Types */ typedef char S8; typedef unsigned char U8; typedef short S16; typedef unsigned short U16; typedef int S32; // defined as 32-bits typedef unsigned int U32; // defined as 32-bits typedef long long S64; typedef unsigned long long U64; typedef unsigned char BIT; typedef unsigned int BOOL; typedef void (*FUNCP)(void); typedef void (*FUNCPP)(void *);
So, I've been doing it wrong since, well, forever. This is embarrassing. Thanks for ruining my day. I shan't forget this John Linq. You have made my list. ;-)
In my defense, I never declare an 'int', I declare a s16 or u16 for a 16-bit value. Same goes for the other types such as 'char' and 'long' etc: I work with bit-widths.
But I can see that this misinterpretation could get me in trouble when interfacing with STM/Keil library routines: a thing I avoid. Thus far, no errors have resulted. I shall be very mindful of this.
It is not a 'stupid question' at all. I think you meant to say "I've got a question, stupid..."
Per, I prefer to use this form using lower-case letters:
typedef unsigned char u8; typedef signed char s8; typedef unsigned int u16; typedef signed int s16; typedef unsigned long u32; typedef signed long s32;
But this is kind of besides the OP point.
How do I set a breakpoint on a memory access to 0x68008000 and 0x6800A0000 ? The expression looks like what? http://i.imgur.com/7vd9U.jpg
(I would also like to see a document that explains the expression parameters better than the Help File on breakpoints)
Thanks for your help on this.
This should do the trick - in the debugger command window, enter:
bs readwrite *((int *) 0x68008000) bs readwrite *((int *) 0x6800A000) bl // list the current breakpoints
With the cast around the address constant, the access size is specified (4 in this case). You might change this to 'short *' or 'char *' to specify the access size to be 2 or 1. You might also open the breakpoint dialog and look how the above breakpoints are translated and shown - just as a reference...
Thanks Gp F !
It worked. You nailed it.
I had to enter it via command-line, and not the Breakpoints Dialog box. (It shows up in the dialog box after the command-line command is entered).
Yup, it turns out that some "temporary" diagnostic code was still burried in one of my isr routines when I first started interfacing with the FPGA at 0x68008000 / 0x6800A000.
Problem solved. Thanks again.