Write memory error @ address 0x5C002000, word access: Memory access timeout

enricol · ‎2007-11-19

Posted on November 19, 2007 at 09:47

enricol · ‎2011-05-17

Posted on May 17, 2011 at 09:33

Hi All,

I could flash my STR912 dongle with no problem before. This morning, I suddenly got the following error when I re-flashed the unit:

''Write memory error @ address 0x5C002000, word access: Memory access timeout.''

Is this caused by hardware fault (JLink or the board)?

Any help is very much appreciated.

mark9 · ‎2011-05-17

Posted on May 17, 2011 at 09:33

I had the same problem. I am able to recreate the situation that kills the chip! It appears to be a bug in the C-SPY interface, in that if the chip gets misconfigured, it is impossible to regain control.

The problem occurred when 91x_conf.h was misconfigured. I forgot to uncomment the #define _GPIO2 line. When the GPIO2 was configured, it was using 0x00000000 as the base address instead of the proper memory location ((GPIO_TypeDef *)GPIO2_BASE). When it configures the (wrong) GPIO input, it immediately crashes. Even worse, you can't exit this state using the IAR tools!

What I did to exit the state was to use Keil uVision and ULINK to completely erase the chip. Then I was able to switch back to IAR and the j-link to step through the code until I found the problem.

I can understand that misconfiguring the CPU can break the debugger, but there really should be a way for IAR/J-link to regain control by fulling erasing the chip. I can't find this feature in IAR EWARN.

Mark

yuantuh · ‎2011-05-17

Posted on May 17, 2011 at 09:33

Hi Elai,

If you are using the IAR board, please make ETM_P66 pin 2-3 close and

ETM_P67 pin 2-3 close.

Yuantu

m_tibu · ‎2011-05-17

Posted on May 17, 2011 at 09:33

I had the same problem as Mark (lakata) described.

http://www.st.com/mcu/forums-cat-4846-21.html

Unfortunately it seems that it can not be solved, using Jlink with IAR environment just trying to reprogram the flash memory, because it gives the error above.

I was not able to erase the flash memory also using a simple wiggler circuit with openOCD. With my wiggler circuit I was not able to halt the ARM core that has this problem.

I solved the problem using another Jtag interface with openOCD for ftdi chips. (Amontec Jtag-Tiny key). I think the secret is to manage correctly the reset pin (there are more then one) on the jtag interface. Maybe this Amontec Jtag interface do it correctly. I had no time to check the difference between a RESET from Jlink and Amontec key.

Maybe there is an issue in the GPIO port configuration.

Check your firmware.

Marco