Hello,
in message 057001c32ffc$ef02fb80$91f59f0a@pc0752 you wrote:
In that new patch (still applied on u-boot-0.3.0 version), we made the modifications concerning DataFlash support as you advise us to do below. It is now possible to boot Linux from DataFlash with Bootm command and to access DataFlash with cp, and md commands...
Thanks.
Tell me if there is something wrong.
I've added your patches, but I had to clean up a couple of compile problems. Did you really run the MAKEALL script as required?
Please clean up ALL compiler warnings before submitting a patch! ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Also, please do not use C++ comments, and use the Linux kernel coding style for your sources.
Finally, I think there is a bug in your code. In "lib_arm/armlinux.c" you write (cast by me to avoid compiler warnings):
132 #ifdef CONFIG_HAS_DATAFLASH 133 if (addr_dataflash(addr)){ 134 read_dataflash(data, len, (char *)CFG_LOAD_ADDR); 135 data = CFG_LOAD_ADDR; 136 } 137 #endif
It is IMHO wrong to use a static, compile-time defined constant here; shouldn't this be the variable load_addr instead?
The same comment applies to "common/cmd_bootm.c"
Also, should "drivers/dataflash.c" not be moved into the at91 directory? It seems it can handle only at91 dataflash anyway? For example, it calls a function AT91F_DataFlashWrite() which is unknown outside the cpu/at91rm9200/ tree...
Checked in locally now, will push to CVS later this night.
Please check out from CVS, and check carefully. If necessary, submit a patch against CVS then (especially for the CFG_LOAD_ADDR issue).
Best regards,
Wolfgang Denk