Re: [U-Boot] Intel Galileo on u-boot

Hi Fei,
On Wed, Apr 29, 2015 at 4:42 PM, WANG FEI wangfei.jimei@gmail.com wrote:
Thanks, Meng Bin, for your big help.
I'm not familiar with u-boot, so I even don't realize the BUILD_ROM flag need to be changed, I just downloaded the u-boot-2015.04 source code and build it with following instruction.
$ make galileo_defconfig $ make all
maybe I should do it again with BUILD_ROM flag.
Yes, you need export BUILD_ROM=y before building u-boot.rom.
Back to the compiler issue, I just use the latest coreboot source code, I don't know the details of the GCC compiler flag, but I think I should have a close look into it, I do realize Quark SoC does not support MMX/SSE etc, so maybe the default coreboot compiler should be modified for this special SoC. Did you do any GCC compiler change when developing it on u-boot? What is your Host O/S(Fedora/Ubuntu 32/64 bit?)? What is the GCC version you are using?
I don't think GCC compiler version matters. I've tested a GCC 4.7.2 compiler from a 64-bit host (cross-compiling) and a GCC 4.6.1 from a 32-bit host (native build). Both work fine.
-Fei
On Tue, Apr 28, 2015 at 10:40 AM, Bin Meng bmeng.cn@gmail.com wrote:
Hi Fei,
On Tue, Apr 28, 2015 at 7:36 AM, WANG FEI wangfei.jimei@gmail.com wrote:
Meng Bin,
I remember few weeks I've built a u-boot binary for Galileo, but I remember the generated u-boot.bin is not 256/512/1024/2048K, it possible is about 384K, is it correct? I want to use the u-boot you built as a reference.
No, it should be 1MB. Did you turn on BUILD_ROM as documented in README.x86?
I've successfully built a coreboot image with CAR enabled, console function only enabled partly due to some issue, it looks the "va_arg" in printf() has something wrong, I guess it's a compiler issue - maybe Quark SoC does not support some special instruments. Did you get the similar issue when porting u-boot?
The Quark integrates a CPU which has the 486 ISA. What compiler flag did you pass when you build coreboot?
Regards, Bin
participants (1)
-
Bin Meng