[PATCH] powerpc: Add isync to copy_and_flush

[PATCH] powerpc: Add isync to copy_and_flush

[Michael Neuling]

In __after_prom_start we copy the kernel down to zero in two calls to 
copy_and_flush.  After the first call (copy from 0 to copy_to_here:)
we jump to the newly copied code soon after.

Unfortunately there's no isync between the copy of this code and the
jump to it.  Hence it's possible that stale instructions could still be
in the icache or pipeline before we branch to it.

We've seen this on real machines and it's results in no console output
after:
  calling quiesce...
  returning from prom_init

The below adds an isync to ensure that the copy and flushing has
completed before any branching to the new instructions occurs.

Signed-off-by: Michael Neuling <mikey@neuling.org>
cc: stable@kernel.org
---
benh: we should get this in 3.9 ASAP.

diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S
index 0886ae6..b61363d 100644
--- a/arch/powerpc/kernel/head_64.S
+++ b/arch/powerpc/kernel/head_64.S
@@ -509,6 +509,7 @@ _GLOBAL(copy_and_flush)
 	sync
 	addi	r5,r5,8
 	addi	r6,r6,8
+	isync
 	blr
 
 .align 8

Re: [PATCH] powerpc: Add isync to copy_and_flush

[Benjamin Herrenschmidt]

On Wed, 2013-04-24 at 20:30 +1000, Michael Neuling wrote:
> benh: we should get this in 3.9 ASAP.

Considering that the bug has been there *forever* I don't think I have a
real standing to try to shove it down Linus throat as a "regression
fix" :-)

I'll put the fix in 3.10 and let it trickle down to stable.

Cheers,
Ben.

Re: [PATCH] powerpc: Add isync to copy_and_flush

[Michael Neuling]

[-- Attachment #1: Type: text/plain, Size: 248 bytes --]

> Considering that the bug has been there *forever* I don't think I have a
> real standing to try to shove it down Linus throat as a "regression
> fix" :-)

True, sorry.

> I'll put the fix in 3.10 and let it trickle down to stable.

Thanks

Mikey

[-- Attachment #2: Type: text/html, Size: 377 bytes --]