On 7/24/05, mark brown <markbrowni1000@xxxxxxxxx> wrote: > > if you call kmalloc, it gives you a virtual addr. That > address is constructed by adding PAGE_OFFSET to the > founded phys addr. and kmalloc returns this addr to > the caller. I looked at kmalloc func. and didnt see > this "adding stuff". Perhaps there is a different > mechanism instead of adding. I dont know. I am > wondering only this adding stuff. > I havn't looked what kmalloc do, but I can say that the memory allocation and virtual/physical address mappings are done using page tables which are built during the boot time (I think it starts from init_page function in arch/i386/mm/init.c) and the virtual address we got from memory allocation functions is actually mappings for the page tables to finally point to the page table entry (PTE) which tells processor about the physical address ........... I want to make this more clear that processor in protected / paging enabled mode not deals with the physical addresses rather it takes the virtual address and then through the page global table entry address which was loaded in CR3 register, it moves along the table entries and then finally reach required to the PTE entry from where if takes 24bits and then adds the 12-LSBs from the virtual address to this PTE entry to generate the corresponding physical address ..... If the above explaination is not clear to u then u can consult Intel's IA32 System Programming Guide's Chapter 3 or so ........... -- Fawad Lateef -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
