Update some comments/function headers

This commit is contained in:
Gregory Nutt 2014-09-11 17:15:26 -06:00
parent ba7d5acf94
commit 67838239c4
2 changed files with 16 additions and 7 deletions

View file

@ -70,6 +70,15 @@
/****************************************************************************
* Name: up_fullcontextrestore
*
* Description:
* Restore the specified task context. Full prototype is:
*
* void up_fullcontextrestore(uint32_t *restoreregs) noreturn_function;
*
* Return:
* None
*
****************************************************************************/
.globl up_fullcontextrestore
@ -83,7 +92,7 @@ up_fullcontextrestore:
#ifdef CONFIG_ARCH_FPU
/* First, restore the floating point registers. Lets do this before we
* restore the arm registers so that we have plentry of registers to
* restore the ARM registers so that we have plenty of registers to
* work with.
*/

View file

@ -128,12 +128,12 @@ Issues
and the apps/ directory. That should not be the case; the nuttx/ logic
should be completely independent of apps/ logic (but not vice versa).
2. The current implementation will not work in the CONFIG_KERNEL_BUILD.
This is because of the little proxy logic (function pcode_proxy() in the
file pcode.c). (a) That logic would attempt to link with P-code logic
that resides in user space. That will not work. And (2) that proxy
would be started in user mode but in the kernel address space which will
certainly crash immediately.
2. The current implementation will not work in the CONFIG_BUILD_PROTECTED or
CONFIG_BUILD_KERNEL configurations. That is because of the little proxy
logic (function pcode_proxy() in the file pcode.c). (a) That logic would
attempt to link with P-code logic that resides in user space. That will
not work. And (2) that proxy would be started in user mode but in the
kernel address space which will certainly crash immediately.
The general idea to fix both of these problems is as follows: