RTOS/rtems
1
0
Fork 0
forked from Imagelibrary/rtems
Code Pull Requests Activity

Never allow the FPU to be switched on for integer-only tasks

Browse Source 
(new gcc may use FP regs implicitely).
This commit is contained in:
Till Straumann
2006-06-19 19:57:01 +00:00
parent efc0cd8757
commit bbc87852db
2 changed files with 16 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
c/src/lib/libcpu/powerpc/ChangeLog
View File

@@ -1,4 +1,9 @@
2006-05-16 Ralf Corsépius <ralf.corsepius@rtems.org>
2006-06-19 Till Straumann <strauman@slac.stanford.edu>
* new-exceptions/cpu.c: Never allow the FPU to be switched
on for integer-only tasks (new gcc may use FP regs implicitely).
2006-05-16 Ralf Corsepius <ralf.corsepius@rtems.org>
* configure.ac: Use RTEMS_AMPOLISH3.

20
c/src/lib/libcpu/powerpc/new-exceptions/cpu.c
View File

@@ -104,20 +104,20 @@ void _CPU_Context_Initialize(
* time (7 July 1997), this restructuring is not being done.
*/
/* Till Straumann: For deferred FPContext save/restore, make sure integer
* tasks have no FPU access in order to catch violations.
* Otherwise, the FP registers may be corrupted.
* Since we set the_contex->msr using our current MSR,
* we must make sure MSR_FP is off if (!is_fp)...
/* Make sure integer tasks have no FPU access in order to
* catch violations. Gcc may implicitely use the FPU and
* data corruption may happen.
* Since we set the_contex->msr using our current MSR,
* we must make sure MSR_FP is off if (!is_fp)...
* Unfortunately, this means that users of vfprintf_r have to use FP
* tasks or fix vfprintf. Furthermore, users of int-only tasks
* must prevent gcc from using the FPU (currently -msoft-float is the
* only way...)
*/
#if defined(CPU_USE_DEFERRED_FP_SWITCH) && (CPU_USE_DEFERRED_FP_SWITCH==TRUE)
if ( is_fp )
#endif
the_context->msr |= PPC_MSR_FP;
#if defined(CPU_USE_DEFERRED_FP_SWITCH) && (CPU_USE_DEFERRED_FP_SWITCH==TRUE)
else
the_context->msr &= ~PPC_MSR_FP;
#endif
the_context->msr &= ~PPC_MSR_FP;
the_context->pc = (uint32_t)entry_point;