Strange data behaviour
Hi,
I'm trying to get my LPC2468 based board to work. I've some problems with the external memory databus.
I would like to know the settings of internal registers to see whether I've initialized them correctly, so I tried making a function that prints the content over the serial link. I'm still in the lowlevel_init function, so I do not have the U-boot puts functions available.
I wrote the following function to convert a long to a HEX string:
-------------------
void print_long(unsigned long data) {
char i;
char
hex_data[]={'0','1','2','3','4','5','6','7','8','9','A','B','C','D','E','F'};
for(i=0; i<8; i++) {
while((U0LSR & (1<<5)) == 0); /* Wait for empty U0THR */
// U0THR = hex_data[(0xDEADBEEF>>i*4)&0xF];
U0THR = hex_data[(0xDEADBEEF>>4)&0xF]; }
}
---------------
When I run it likes this I get 8 E's. Which is what I expect. When I run it with the commented-out line, I get back 8 0x0's. So it seems that the output is only correct when it is constant.
Does anyone have a clue on why that is?
Kind regards,
Remco Poelstra
Dear Remco Poelstra,
Quoted 9 lines · In message <[email protected]> you wrote:
I'm trying to get my LPC2468 based board to work. I've some problems with the external memory databus.
I would like to know the settings of internal registers to see whether I've initialized them correctly, so I tried making a function that prints the content over the serial link. I'm still in thelowlevel_initfunction, so I do not have the U-boot puts functions available. I wrote the following function to convert a long to a HEX string:
-------------------
Quoted 8 lines
void print_long(unsigned long data) { char i; charhex_data[]={'0','1','2','3','4','5','6','7','8','9','A','B','C','D','E','F'};
for(i=0; i<8; i++) { while((U0LSR & (1<<5)) == 0); /* Wait for empty U0THR */// U0THR = hex_data[(0xDEADBEEF>>i*4)&0xF];
For clearity of the code I would write that as "...>>(i*4)...", but that's not relevant here.
Quoted 8 lines
U0THR = hex_data[(0xDEADBEEF>>4)&0xF];}
}
---------------
When I run it likes this I get 8 E's. Which is what I expect. When I run it with the commented-out line, I get back 8 0x0's. So it seems that the output is only correct when it is constant.
Does anyone have a clue on why that is?
What exactly is U0THR ?
The only definition I can find in the U-Boot sources is here:
include/asm-arm/arch-lpc2292/lpc2292_registers.h:#define U0THR 0xE000C000
but that is obviously not what you are using.
You probably forget the effects of compiler optimization and/or caching here.
I bet you are using a plain pointer access without a "volatile", which is essential here. Maybe even some form of "sync" is needed. In any case, you should use appropriate accessor functions to access device registers.
Best regards,
Wolfgang Denk
Signature · 4 lines
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de Never call a man a fool. Borrow from him.
Wolfgang Denk schreef:
Quoted 10 lines
Quoted 8 lines
U0THR = hex_data[(0xDEADBEEF>>4)&0xF];}
}
---------------
When I run it likes this I get 8 E's. Which is what I expect. When I run it with the commented-out line, I get back 8 0x0's. So it seems that the output is only correct when it is constant.
Does anyone have a clue on why that is?What exactly is U0THR ?
Hi,
Thanks for your reply.
For my processor, U0THR is defnied as:
#define U0THR (*(volatile unsigned char *)(UART0_BASE_ADDR +0x00))
UART0_BASE_ADDR is equal to 0xE000C000
The register itself represents the top of the UART0 transmit FIFO.
Quoted 7 lines
The only definition I can find in the U-Boot sources is here:
include/asm-arm/arch-lpc2292/lpc2292_registers.h:#define U0THR
0xE000C000but that is obviously not what you are using.
Quoted 8 lines
You probably forget the effects of compiler optimization and/or caching here.
I bet you are using a plain pointer access without a "volatile", which is essential here. Maybe even some form of "sync" is needed. In any case, you should use appropriate accessor functions to access device registers.
I do not know any accessor functions.
What do you mean with a "sync"?
Kind regards,
Remco Poelstra
Quoted 7 lines · On Wed, 18 Feb 2009 09:19:08 +0100 Remco Poelstra…
Wolfgang Denk schreef:
Quoted 4 lines
I bet you are using a plain pointer access without a "volatile", which is essential here. Maybe even some form of "sync" is needed. In any case, you should use appropriate accessor functions to access device registers.
I do not know any accessor functions.
things like out_be32 (at least that's what it's called on ppc arch).
Quoted 1 line
What do you mean with a "sync"?
it's a processor instruction that ensures that memory and/or i/o accesses occur exactly when you specify they occur. For this reason, a sync instruction is included as a part of the accessor functions.
Kim