Re: [PATCH] powerpc: Pass the right cpu_spec to ->setup_cpu() on 64-bit

Re: [PATCH] powerpc: Pass the right cpu_spec to ->setup_cpu() on 64-bit

[Andreas Schwab]

You may also want to remove the misleading sentence about PTRRELOC in
the preceding comment, since both s and t were already passed through it
anyway.

Andreas.

-- 
Andreas Schwab, schwab@redhat.com
GPG Key fingerprint = D4E8 DBE3 3813 BB5D FA84  5EC7 45C6 250E 6F00 984E
"And now for something completely different."

[PATCH] powerpc: Pass the right cpu_spec to ->setup_cpu() on 64-bit

[Benjamin Herrenschmidt]

When calling setup_cpu() on 64-bit, we pass a pointer to the
cputable entry we have found. This used to be fine when cur_cpu_spec
was a pointer to that entry, but nowadays, we copy the entry into
a separate variable, and we do so before we call the setup_cpu()
callback. That means that any attempt by that callback at patching
the CPU table entry (to adjust CPU features for example) will patch
the wrong table.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
 arch/powerpc/kernel/cputable.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c
index 843abf7..fb83d36 100644
--- a/arch/powerpc/kernel/cputable.c
+++ b/arch/powerpc/kernel/cputable.c
@@ -2105,8 +2105,8 @@ static void __init setup_cpu_spec(unsigned long offset, struct cpu_spec *s)
 	 * pointer on ppc64 and booke as we are running at 0 in real mode
 	 * on ppc64 and reloc_offset is always 0 on booke.
 	 */
-	if (s->cpu_setup) {
-		s->cpu_setup(offset, s);
+	if (t->cpu_setup) {
+		t->cpu_setup(offset, t);
 	}
 #endif /* CONFIG_PPC64 || CONFIG_BOOKE */
 }