On 11/19/19 12:19, Michael Ellerman wrote:

Qais Yousef [off-list ref] writes:

quoted

The core device API performs extra housekeeping bits that are missing
from directly calling cpu_up/down.

See commit a6717c01ddc2 ("powerpc/rtas: use device model APIs and
serialization during LPM") for an example description of what might go
wrong.

This also prepares to make cpu_up/down a private interface for anything
but the cpu subsystem.

Signed-off-by: Qais Yousef <redacted>
CC: Benjamin Herrenschmidt <benh@kernel.crashing.org>
CC: Paul Mackerras <redacted>
CC: Michael Ellerman <mpe@ellerman.id.au>
CC: Enrico Weigelt <info@metux.net>
CC: Ram Pai <redacted>
CC: Nicholas Piggin <npiggin@gmail.com>
CC: Thiago Jung Bauermann <redacted>
CC: Christophe Leroy <redacted>
CC: Thomas Gleixner <redacted>
CC: linuxppc-dev@lists.ozlabs.org
CC: linux-kernel@vger.kernel.org
---
 arch/powerpc/kernel/machine_kexec_64.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

My initial though is "what about kdump", but that function is not called
during kdump, so there should be no issue with the extra locking leading
to deadlocks in a crash.

Acked-by: Michael Ellerman <mpe@ellerman.id.au>

Thanks.

I assume you haven't actually tested it?

Only compile tested it I'm afraid. Would appreciate if you can give it a spin.
Otherwise I'd be happy to try it out on qemu.

Cheers

--
Qais Yousef