From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Thu, 9 May 2019 13:37:01 +0200 From: Cornelia Huck Subject: Re: [PATCH 07/10] s390/airq: use DMA memory for adapter interrupts Message-ID: <20190509133701.3e93299b.cohuck@redhat.com> In-Reply-To: <20190426183245.37939-8-pasic@linux.ibm.com> References: <20190426183245.37939-1-pasic@linux.ibm.com> <20190426183245.37939-8-pasic@linux.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: kvm-owner@vger.kernel.org List-Archive: List-Post: To: Halil Pasic Cc: kvm@vger.kernel.org, linux-s390@vger.kernel.org, Martin Schwidefsky , Sebastian Ott , virtualization@lists.linux-foundation.org, "Michael S. Tsirkin" , Christoph Hellwig , Thomas Huth , Christian Borntraeger , Viktor Mihajlovski , Vasily Gorbik , Janosch Frank , Claudio Imbrenda , Farhan Ali , Eric Farman List-ID: On Fri, 26 Apr 2019 20:32:42 +0200 Halil Pasic wrote: > Protected virtualization guests have to use shared pages for airq > notifier bit vectors, because hypervisor needs to write these bits. > > Let us make sure we allocate DMA memory for the notifier bit vectors. > > Signed-off-by: Halil Pasic > --- > arch/s390/include/asm/airq.h | 2 ++ > drivers/s390/cio/airq.c | 18 ++++++++++++++---- > 2 files changed, 16 insertions(+), 4 deletions(-) As an aside, there are some other devices that use adapter interrupts as well (pci, ap, qdio). How does that interact with their needs? Do they continue to work on non-protected virt guests (kvm or z/VM), and can they be accommodated if support for them on protected guests is added in the future? (For some of the indicator bit handling, I suspect millicode takes care of it anyway, but at least for pci, there's some hypervisor action to be aware of.) Also, as another aside, has this been tested with a regular guest under tcg as well? If not, can you provide a branch for quick verification somewhere? From mboxrd@z Thu Jan 1 00:00:00 1970 From: Cornelia Huck Subject: Re: [PATCH 07/10] s390/airq: use DMA memory for adapter interrupts Date: Thu, 9 May 2019 13:37:01 +0200 Message-ID: <20190509133701.3e93299b.cohuck@redhat.com> References: <20190426183245.37939-1-pasic@linux.ibm.com> <20190426183245.37939-8-pasic@linux.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20190426183245.37939-8-pasic@linux.ibm.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: virtualization-bounces@lists.linux-foundation.org Errors-To: virtualization-bounces@lists.linux-foundation.org To: Halil Pasic Cc: Vasily Gorbik , linux-s390@vger.kernel.org, Thomas Huth , Claudio Imbrenda , kvm@vger.kernel.org, Sebastian Ott , "Michael S. Tsirkin" , Farhan Ali , Eric Farman , virtualization@lists.linux-foundation.org, Christoph Hellwig , Martin Schwidefsky , Viktor Mihajlovski , Janosch Frank List-Id: virtualization@lists.linuxfoundation.org On Fri, 26 Apr 2019 20:32:42 +0200 Halil Pasic wrote: > Protected virtualization guests have to use shared pages for airq > notifier bit vectors, because hypervisor needs to write these bits. > > Let us make sure we allocate DMA memory for the notifier bit vectors. > > Signed-off-by: Halil Pasic > --- > arch/s390/include/asm/airq.h | 2 ++ > drivers/s390/cio/airq.c | 18 ++++++++++++++---- > 2 files changed, 16 insertions(+), 4 deletions(-) As an aside, there are some other devices that use adapter interrupts as well (pci, ap, qdio). How does that interact with their needs? Do they continue to work on non-protected virt guests (kvm or z/VM), and can they be accommodated if support for them on protected guests is added in the future? (For some of the indicator bit handling, I suspect millicode takes care of it anyway, but at least for pci, there's some hypervisor action to be aware of.) Also, as another aside, has this been tested with a regular guest under tcg as well? If not, can you provide a branch for quick verification somewhere?