Syscalls just point to stubs that are then generated during compile; I don't know the details, but I ran into the same problem a while back. Duane On Sun, Nov 30, 2008 at 12:19 PM, <romain.haiku@xxxxxxxxxxx> wrote: > Hello everybody, > > I am trying to follow the path of syscalls in haiku and there is a link > that I am missing. I tried, and failed, to see how is handled > "is_computer_on" from userland until the inevitable "return 1" of its > implementation. > > In the kernel I found "_user_is_computer_on" in syscall.cpp. > In the userland I found "is_computer_on" in libroot that calls > "_kern_is_computer_on". However I could not find where/how is defined the > latter one. I suppose that it is generated by some tool since I can find > it in the generated "libroot.so". > > So basically I am missing 2 parts of the chain: > - Where is the "_kern_xxx" code for all userland calls to syscalls ? > - Where is the link between the interrupt that is generated and the call > to the corresponding "_user_xxx" ? > > Btw why does the code in the kernel is prefixed "_user", and the code in > user is prefixed "_kernel" ? > > Romain. > > > > >