mbox series

[00/12] s390/vfio_ap: crypto pass-through for SE guests

Message ID 20230815184333.6554-1-akrowiak@linux.ibm.com (mailing list archive)
Headers show
Series s390/vfio_ap: crypto pass-through for SE guests | expand

Message

Anthony Krowiak Aug. 15, 2023, 6:43 p.m. UTC
This patch series is for the changes required in the vfio_ap device 
driver to facilitate pass-through of crypto devices to a secure 
execution guest. In particular, it is critical that no data from the
queues passed through to the SE guest is leaked when the guest is 
destroyed. There are also some new response codes returned from the
PQAP(ZAPQ) and PQAP(TAPQ) commands that have been added to the
architecture in support of pass-through of crypto devices to SE guests;
these need to be accounted for when handling the reset of queues.


Janosch Frank (1):
  s390/uv: export uv_pin_shared for direct usage

Tony Krowiak (11):
  s390/vfio-ap: No need to check the 'E' and 'I' bits in APQSW after
    TAPQ
  s390/vfio-ap: clean up irq resources if possible
  s390/vfio-ap: wait for response code 05 to clear on queue reset
  s390/vfio-ap: allow deconfigured queue to be passed through to a guest
  s390/vfio-ap: remove upper limit on wait for queue reset to complete
  s390/vfio-ap: store entire AP queue status word with the queue object
  s390/vfio-ap: use work struct to verify queue reset
  s390/vfio-ap: handle queue state change in progress on reset
  s390/vfio-ap: check for TAPQ response codes 0x35 and 0x36
  kvm: s390: export kvm_s390_pv*_is_protected functions
  s390/vfio-ap: Make sure nib is shared

 arch/s390/include/asm/kvm_host.h      |   3 +
 arch/s390/include/asm/uv.h            |   6 +
 arch/s390/kernel/uv.c                 |   3 +-
 arch/s390/kvm/kvm-s390.h              |  12 --
 arch/s390/kvm/pv.c                    |  14 +++
 drivers/s390/crypto/vfio_ap_ops.c     | 164 +++++++++++++++++---------
 drivers/s390/crypto/vfio_ap_private.h |   6 +-
 7 files changed, 135 insertions(+), 73 deletions(-)

Comments

Janosch Frank Aug. 16, 2023, 11:39 a.m. UTC | #1
On 8/15/23 20:43, Tony Krowiak wrote:
> This patch series is for the changes required in the vfio_ap device
> driver to facilitate pass-through of crypto devices to a secure
> execution guest. In particular, it is critical that no data from the
> queues passed through to the SE guest is leaked when the guest is
> destroyed. There are also some new response codes returned from the
> PQAP(ZAPQ) and PQAP(TAPQ) commands that have been added to the
> architecture in support of pass-through of crypto devices to SE guests;
> these need to be accounted for when handling the reset of queues.
> 

@Heiko: Once this has soaked a day or two, could you please apply this 
and create a feature branch that I can pull from?
Janosch Frank Aug. 16, 2023, 12:12 p.m. UTC | #2
On 8/16/23 13:39, Janosch Frank wrote:
> On 8/15/23 20:43, Tony Krowiak wrote:
>> This patch series is for the changes required in the vfio_ap device
>> driver to facilitate pass-through of crypto devices to a secure
>> execution guest. In particular, it is critical that no data from the
>> queues passed through to the SE guest is leaked when the guest is
>> destroyed. There are also some new response codes returned from the
>> PQAP(ZAPQ) and PQAP(TAPQ) commands that have been added to the
>> architecture in support of pass-through of crypto devices to SE guests;
>> these need to be accounted for when handling the reset of queues.
>>
> 
> @Heiko: Once this has soaked a day or two, could you please apply this
> and create a feature branch that I can pull from?

Sorry for the noise, for some reason I still had Heiko's old address in 
the address book. I'll delete it in a second.

Here we go again.
Heiko Carstens Aug. 18, 2023, 1:35 p.m. UTC | #3
On Wed, Aug 16, 2023 at 02:12:50PM +0200, Janosch Frank wrote:
> On 8/16/23 13:39, Janosch Frank wrote:
> > On 8/15/23 20:43, Tony Krowiak wrote:
> > > This patch series is for the changes required in the vfio_ap device
> > > driver to facilitate pass-through of crypto devices to a secure
> > > execution guest. In particular, it is critical that no data from the
> > > queues passed through to the SE guest is leaked when the guest is
> > > destroyed. There are also some new response codes returned from the
> > > PQAP(ZAPQ) and PQAP(TAPQ) commands that have been added to the
> > > architecture in support of pass-through of crypto devices to SE guests;
> > > these need to be accounted for when handling the reset of queues.
> > > 
> > 
> > @Heiko: Once this has soaked a day or two, could you please apply this
> > and create a feature branch that I can pull from?
> 
> Sorry for the noise, for some reason I still had Heiko's old address in the
> address book. I'll delete it in a second.
> 
> Here we go again.

Series is now available via the vfio-ap branch, based on rc2 like your
your branches within the kvms390 tree.

https://git.kernel.org/pub/scm/linux/kernel/git/s390/linux.git vfio-ap