2010/10/8 Juan Rafael Garcia Blanco <jgblanco.mail@xxxxxxxxx>: > Well, I think it applies to all architectures, but I'm not sure, so take > with care. Also, I think this is something everybody knows, sorry. > The OS is mapped in every process address space, but its pages are protected > (check out this http://www.logix.cz/michal/doc/i386/chp05-02.htm#F-05-10). A > process cannot jump to a kernel address, because the process executes in > user mode and kernel addresses are therefore protected from being executed > by a user process. Then, a system call occurs and the kernel starts > executing through an interrupt (the famous 0x80 vector number). I think you are just confused by the parameters how to deliver to the kernel space. Because system call code runs in the kernel space and caller runs in the user space, they use the different the stack. On i386, the parameters of a system call are transported via registers. The system call number goes into %eax, the first parameter in %ebx, the second in %ecx, the third in %edx, the fourth in %esi, the fifth in %edi, the sixth in %ebp. Parameters more than six, there will be on register to save the address of the structure in the user space, and in the system call routine will copy the parameters into the kernel stack. The book <<Understanding the Linux Kernel By Daniel Pierre Bovet, Marco Cesatí>> will give you more details. http://books.google.com.hk/books?id=cbbMrRNiC4cC&pg=PT618&dq=system+call+parameter&hl=en&ei=AH-uTIbmBoayvgPOotndBg&sa=X&oi=book_result&ct=result&resnum=1&ved=0CCgQ6AEwAA#v=onepage&q=system%20call%20parameter&f=false >The active > address space is the caller one, but now, since the kernel executes in > privileged, any address of any page can be reached. The kernel can throw to > the bus addresses from user space, cause it is actually using the process > address space. > I was confused mostly because I thought the kernel as a process, and I > didn't remember how a page table entry looks like. Besides, I have learnt > something: during an interrupt from an external device for example, the > kernel cannot use addresses from any user process memory map, because the > current address space may not be the correct one. > Thank you, and sorry for my English > On Oct 7, 2010, at 6:38 PM, hiren panchasara wrote: > > > On Thu, Oct 7, 2010 at 9:28 AM, Juan Rafael Garcia Blanco > <jgblanco.mail@xxxxxxxxx> wrote: >> >> Hi, it has just been solved, thank you > > Would you care to share? > > Thanks. >> >> On Oct 7, 2010, at 5:55 PM, Juan Rafael Garcia Blanco wrote: >> >> > Hi. >> > During the boot process, Linux is loaded in main memory. When a system >> > call is done, the OS executes again. If the system call is accompanied by >> > some parameter, which is an address of the user process memory map, how can >> > the kernel access that variable which is in user process memory map? I mean, >> > when the kernel fires up an address, how is it translated? I can't find an >> > answer to my question in books nor in the internet. >> > >> > Thank you very much >> >> >> -- >> To unsubscribe from this list: send an email with >> "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx >> Please read the FAQ at http://kernelnewbies.org/FAQ >> > > > -- Best Regards Lin -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
