forked from Imagelibrary/rtems
5eb7da97cd
<e.norum@sk.sympatico.ca> and concerns from Thomas Doerfler <td@imd.m.ISAR.de> when he submitted the patch: Since enabling XON/XOFF has such a major performance hit on `smart' output devices I think it should be *off* by default. I think some thought should be given to adding hooks for hardware that can support XON/XOFF without software intervention, or for hardware like the 68360 SCC's that can use large buffers, but still handle special characters immediately. The patch you sent is a very good start, though. I just think that the software flow control should be off -- to match the way the serial I/O support has worked up until now.
--
-- $Id$
--
Overview of newlib support (newlib is from CYGNUS)
Each task can have its own libc state including:
open stdio files
strtok
multi precision arithmetic state
etc.
This is implemented by a reentrancy data structure for each task.
When a task is "started" (in RTEMS sense) the reentrancy structure
is allocated. Its address is stored in notepad[NOTEPAD_LAST].
When task is switched to, the value of global variable _impure_ptr
is changed to the value of the new tasks reentrancy structure.
When a task is deleted
atexit() processing (for that task) happens
task's stdio buffers are flushed
When exit(3) is called
calling task's atexit processing done
global libc state atexit processing done
(this will include any atexit routines installed by drivers)
executive is shutdown
causes a context switch back to bsp land
NOTE:
libc extension are installed by bsp_libc_init()
iff we are using clock interrupts.
This hack is necessary to allow the tmtests to avoid
timing the extensions.