Hola, could it be that some of those syscalls are the non-portable version optimized for a specific platform?
slds. gabi -- [EMAIL PROTECTED] wrote: > >Hi, > >Sorry if this is explained somewhere already. I couldn't find an >answer, so here we go. > >Considering the `print' function of libc, we see that it calls >`vfprint', which in turn calls `fmtFdFlush', which finally calls >`write' with 3 parameters. > >write(2) says that the source is to be found in /sys/src/libc/ >9syscall, but the mkfile in there defines `_write' which in turn does >nothing but push the value 20 on the accumulator and call the >interrupt 0x64. > >a) Where is `write' actually defined? (Or does `write' simply call >`_write'?) >b) Why do some syscalls have `_' prefixes while others do not? >c) How are the parameters to the libc syscall wrapper passed onto the >actual kernel-side implementation? > >Thank you for your time. > >Cheers, >Anant