• State Suggested Answer
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 10 subscribers
  • Views 2301 views
  • Users 0 members are here
Options
Related
How was your experience today?
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

How to generalize SVC calls for N arguments and get a return value?

sailorarm

Hello everyone,

So I am constructing my simple kernel on an ARM CortexM3. To system calls I have created a macro that can take the SVC immediate and arguments. 

#define syscall_1(code, args) { \
asm volatile ("MOV %%r0, %0 " \
: \
: "r" (args) \
: "%r0"); \
asm volatile ("svc %[immediate]" \
: \
:[immediate] "I" (code) \
: ); \
}
#define syscall_2(code, args1, args2) { \
asm volatile ("MOV %%r0, %0\n\t" \
"MOV %%r1, %1\n\t" \
: \
: "r" (args1), \
"r" (args2) \
: "%r0", \
"%r1"); \
asm volatile ("svc %[immediate]" \
: \
:[immediate] "I" (code) \
: ); \
}

So I use a C wrapper to manage the callbacks as widely documented around, like on the ARM Cortex M3 Guide and other manuals. But first and most importantly, how would I get a return for this system call?  I am storing it at a global, but I would like to be able to do something like ret = syscall(....).

And also, any way to create a single macro to adapt to the number of arguments? 

  • 0

    The RTX kernel from ARM has svc macros that achieve something very similar to what you want/need.  I did post on this thread several weeks ago but my post was deleted, after being flagged as abusive, which I couldn't fathom.  It may have been because I linked so certain files, so I won't do that again.  let's just say

    in the CMSIS_5 arm repo on github, if you traverse to the CMSIS / RTOS / RTX / Source directory, see rtx_core_cm.h.  It has svc macros for several different compilers. 

    I use these macros in my own code.  I added some svc calls to RTX with the goal of achieving a somewhat posix i/o model, e.g. open,close,read,write and in particular select.  being able to wait for input data from many devices at same time without having to resort to a multi-threaded model is a big win, IMHO.