On Apr 21, 2008, at 3:19 PM, Gabriel Paubert wrote:
On Mon, Apr 21, 2008 at 02:57:47PM -0500, Kumar Gala wrote:
603 CPUs have the same issue that some 750 CPUs have in that they can crash in funny ways if a store from an FPU register instruction is executed on a register that has never been initialized since power on. This patch fixes it by making sure all FP registers have been properly initialized at kernel
boot.

Signed-off-by: Kumar Gala <[EMAIL PROTECTED]>
---

Paul, I've put this into my powerpc-next tree since its a pretty trivial
patch.

- k

arch/powerpc/kernel/cpu_setup_6xx.S |    9 ++++++++-
1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/arch/powerpc/kernel/cpu_setup_6xx.S b/arch/powerpc/ kernel/cpu_setup_6xx.S
index f1ee0b3..4a1c3cd 100644
--- a/arch/powerpc/kernel/cpu_setup_6xx.S
+++ b/arch/powerpc/kernel/cpu_setup_6xx.S
@@ -17,7 +17,14 @@
#include <asm/cache.h>

_GLOBAL(__setup_cpu_603)
-       b       setup_common_caches
+       mflr    r4
+BEGIN_FTR_SECTION
+       bl      __init_fpu_registers
+END_FTR_SECTION_IFCLR(CPU_FTR_FPU_UNAVAILABLE)
+       bl      __init_fpu_registers

I don't understand that code.

What does the second bl __init_fpu_regiters do?

that was a copy/paste error.  (see v2 of the patch).

As I understand it, the first one is conditional on the presence
of an FPU (reasonable). But the secone one makes no sense.

Finally wouldn't it be simpler to initialize the FPU on all CPU
that have it?

that would require a change to the entry point of all these functions. Its not worth the savings.

+       bl      setup_common_caches
+       mtlr    r4
+       blr

You can use tail call elimination and save one instruction there :-)

I think we can find better places to save code size. :)

- k
_______________________________________________
Linuxppc-dev mailing list
[email protected]
https://ozlabs.org/mailman/listinfo/linuxppc-dev

Reply via email to