building bootstrap
Adam Lackorzynski
adam at os.inf.tu-dresden.de
Tue Dec 23 17:52:16 CET 2008
Hi,
On Tue Dec 23, 2008 at 15:51:07 +0800, Tsai, Tung-Chieh wrote:
> I still have another question. I set the address of kernel in
> mem_layout-xxx.cpp, it successfully get into kernel ( get
> into __main(), stop in some place I still not checking ), but
> I fill confuse on virtual memory setting.
>
> On my platform, ram is start at 0x3000,0000, so I set :
> * RAM_BASE=0x3000,0000
Ok.
> * DEFAULT_RELOC_arm=0x0
No, do not change the DEFAULT_RELOC_arm values. Those are offsets,
relative to RAM_BASE! I.e. the real link-address is
RAM_BASE+DEFAULT_RELOC_arm.
> * MOD_ADDR use default value in modules.list, i.e.,
> 0x0200,0000
That is ok.
> * Mem_layout::platform_phys_base=0x3000,0000
> * Mem_layout::Sdram_phys_base =0x3020,0000
Sdram_phys_base is 0x30000000 for you, i.e. it's the start of the RAM in
the physical address space. What is platform_phys_base?
> So fiasco kernel start at 0x3020,0000, sigma0 and roottask
> are under 0x3020,000, which are 0x3006,8000 and
> 0x3007,8000, respectively.
No, the kernel starts at 0x1000 relative to RAM, i.e. 0x30001000.
> But how does kernel know the area from 0x3010,0000 to
> 0x3020,0000 ? I see map_1mb in map_hw will map this
> area( from Mem_layout::platform_map_base to
> Mem_layout::platform_phys_base ), but this only mapping
> 1mb, where is 0x3010,0000 to 0x3020,0000 setting ?
> Or I'm missing some setting ? And I fill strange in
> Mem_layout::Sdram_phys_base to 0x3020,0000, since
> physical ram is started at 0x3000,0000, is this setting
> really correct or I misunderstand its meaning ?
Sdram_phys_base must be 0x3000,0000 as this is the location where memory
starts. For a new platform you actually only need to set Sdram_phys_base
and CONFIG_KERNEL_LOAD_ADDR to the same value (for the kernel). For
user-land you need to adapt RAM_BASE and RAM_SIZE_MB, all other values
stay the same as they're relative to RAM_BASE.
Adam
--
Adam adam at os.inf.tu-dresden.de
Lackorzynski http://os.inf.tu-dresden.de/~adam/
More information about the l4-hackers
mailing list