Re: [PATCH v1 0/1] s390x: protvirt: SCLP interpretation

[Janosch Frank <frankja@linux.ibm.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1601 bytes --]

On 11/28/19 11:13 AM, Pierre Morel wrote:
> A new proposition:
> I think it would be wise to fork directly from handle_instruction
> instead to accept per default all instructions with with secure
> instruction interception code.
> Just in case future firmware with older QEMU.
> 
> How ever I let three dors open.
> 
> 1) This patch accepts the all B2 instructions, mostly I/O.
> Some of the instructions will not work correctly for PV until patched.
> This should be fixed, and will be, in a separate patch.
> 
> 2) The same is true for DIAG instructions.
> 
> 3) Secure notifications are separated from secure instructions and
> normal instructions interception because this case is completely new.
> For B2 instructions we do not have to do anything this just informative.
> However, one information is of interrest, a notification that
> SIGP(STOP) is sent to stop the CPUs and terminate QEMU.

Pierre, I told you this morning that I don't want this and that you
should leave this untouched until I can explain my thoughts behind the
initial patch in a f2f.
Thomas' review of your change only confirmed my concerns about this patch.

This is the wrong patch at the wrong time, which creates noise and work
for other people. Please stop and work on something else.



> 
> 
> 
> Pierre Morel (1):
>   s390x: protvirt: SCLP interpretation
> 
>  hw/s390x/sclp.c         | 18 +++++++++++++
>  include/hw/s390x/sclp.h |  2 ++
>  target/s390x/kvm.c      | 56 ++++++++++++++++++++++++++++++++++++++++-
>  3 files changed, 75 insertions(+), 1 deletion(-)
> 



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]