similar to: [RFC PATCH 03/12] s390/mm: force swiotlb for protected virtualization

From: Thiago Jung Bauermann <bauerman at linux.ibm.com> In order to safely use the DMA API, virtio needs to know whether DMA addresses are in fact physical addresses and for that purpose, dma_addr_is_phys_addr() is introduced. cc: Benjamin Herrenschmidt <benh at kernel.crashing.org> cc: David Gibson <david at gibson.dropbear.id.au> cc: Michael Ellerman <mpe at

On 14/10/2019 05:51, David Gibson wrote: > On Fri, Oct 11, 2019 at 06:25:18PM -0700, Ram Pai wrote: >> From: Thiago Jung Bauermann <bauerman at linux.ibm.com> >> >> In order to safely use the DMA API, virtio needs to know whether DMA >> addresses are in fact physical addresses and for that purpose, >> dma_addr_is_phys_addr() is introduced. >> >>

On Fri, Oct 11, 2019 at 06:25:18PM -0700, Ram Pai wrote: > From: Thiago Jung Bauermann <bauerman at linux.ibm.com> > > In order to safely use the DMA API, virtio needs to know whether DMA > addresses are in fact physical addresses and for that purpose, > dma_addr_is_phys_addr() is introduced. > > cc: Benjamin Herrenschmidt <benh at kernel.crashing.org> > cc:

On Fri, Oct 11, 2019 at 06:25:18PM -0700, Ram Pai wrote: > From: Thiago Jung Bauermann <bauerman at linux.ibm.com> > > In order to safely use the DMA API, virtio needs to know whether DMA > addresses are in fact physical addresses and for that purpose, > dma_addr_is_phys_addr() is introduced. > > cc: Benjamin Herrenschmidt <benh at kernel.crashing.org> > cc:

On s390, protected virtualization guests have to use bounced I/O buffers. That requires some plumbing. Let us make sure, any device that uses DMA API with direct ops correctly is spared from the problems, that a hypervisor attempting I/O to a non-shared page would bring. Signed-off-by: Halil Pasic <pasic at linux.ibm.com> --- arch/s390/Kconfig | 4 +++

On s390, protected virtualization guests have to use bounced I/O buffers. That requires some plumbing. Let us make sure, any device that uses DMA API with direct ops correctly is spared from the problems, that a hypervisor attempting I/O to a non-shared page would bring. Signed-off-by: Halil Pasic <pasic at linux.ibm.com> Reviewed-by: Claudio Imbrenda <imbrenda at linux.ibm.com> ---

On s390, protected virtualization guests have to use bounced I/O buffers. That requires some plumbing. Let us make sure, any device that uses DMA API with direct ops correctly is spared from the problems, that a hypervisor attempting I/O to a non-shared page would bring. Signed-off-by: Halil Pasic <pasic at linux.ibm.com> Reviewed-by: Claudio Imbrenda <imbrenda at linux.ibm.com> ---

From: Halil Pasic <pasic at linux.ibm.com> On s390, protected virtualization guests have to use bounced I/O buffers. That requires some plumbing. Let us make sure, any device that uses DMA API with direct ops correctly is spared from the problems, that a hypervisor attempting I/O to a non-shared page would bring. Signed-off-by: Halil Pasic <pasic at linux.ibm.com> Reviewed-by:

From: Halil Pasic <pasic at linux.ibm.com> On s390, protected virtualization guests have to use bounced I/O buffers. That requires some plumbing. Let us make sure, any device that uses DMA API with direct ops correctly is spared from the problems, that a hypervisor attempting I/O to a non-shared page would bring. Signed-off-by: Halil Pasic <pasic at linux.ibm.com> Reviewed-by:

On Wed, 8 May 2019 15:15:40 +0200 Claudio Imbrenda <imbrenda at linux.ibm.com> wrote: > On Fri, 26 Apr 2019 20:32:39 +0200 > Halil Pasic <pasic at linux.ibm.com> wrote: > > > On s390, protected virtualization guests have to use bounced I/O > > buffers. That requires some plumbing. > > > > Let us make sure, any device that uses DMA API with direct

> Subject: [PATCH 04/10] s390/mm: force swiotlb for protected virtualization > Date: Fri, 26 Apr 2019 20:32:39 +0200 > From: Halil Pasic <pasic at linux.ibm.com> > To: kvm at vger.kernel.org, linux-s390 at vger.kernel.org, Cornelia Huck <cohuck at redhat.com>, > Martin Schwidefsky <schwidefsky at de.ibm.com>, Sebastian Ott <sebott at linux.ibm.com> > CC:

On Fri, 5 Apr 2019 01:16:13 +0200 Halil Pasic <pasic at linux.ibm.com> wrote: > On s390 protected virtualization guests also have to use bounce I/O > buffers. That requires some plumbing. > > Let us make sure any device using DMA API accordingly is spared from the > problems that hypervisor attempting I/O to a non-shared secure page would > bring. I have problems

On Tue, 9 Apr 2019 12:54:16 +0200 Halil Pasic <pasic at linux.ibm.com> wrote: > On Tue, 9 Apr 2019 12:16:47 +0200 > Cornelia Huck <cohuck at redhat.com> wrote: > > > On Fri, 5 Apr 2019 01:16:13 +0200 > > Halil Pasic <pasic at linux.ibm.com> wrote: > > > > > On s390 protected virtualization guests also have to use bounce I/O > > >

On Fri, 26 Apr 2019 20:32:39 +0200 Halil Pasic <pasic at linux.ibm.com> wrote: > On s390, protected virtualization guests have to use bounced I/O > buffers. That requires some plumbing. > > Let us make sure, any device that uses DMA API with direct ops > correctly is spared from the problems, that a hypervisor attempting > I/O to a non-shared page would bring. > >

On Fri, 26 Apr 2019 20:32:39 +0200 Halil Pasic <pasic at linux.ibm.com> wrote: > On s390, protected virtualization guests have to use bounced I/O > buffers. That requires some plumbing. > > Let us make sure, any device that uses DMA API with direct ops > correctly is spared from the problems, that a hypervisor attempting > I/O to a non-shared page would bring. > >

On Fri, Apr 26, 2019 at 08:32:39PM +0200, Halil Pasic wrote: > +EXPORT_SYMBOL_GPL(set_memory_encrypted); > +EXPORT_SYMBOL_GPL(set_memory_decrypted); > +EXPORT_SYMBOL_GPL(sev_active); Why do you export these? I know x86 exports those as well, but it shoudn't be needed there either.

On Fri, 26 Apr 2019 12:27:11 -0700 Christoph Hellwig <hch at infradead.org> wrote: > On Fri, Apr 26, 2019 at 08:32:39PM +0200, Halil Pasic wrote: > > +EXPORT_SYMBOL_GPL(set_memory_encrypted); > > > +EXPORT_SYMBOL_GPL(set_memory_decrypted); > > > +EXPORT_SYMBOL_GPL(sev_active); > > Why do you export these? I know x86 exports those as well, but > it

On 29.04.19 15:59, Halil Pasic wrote: > On Fri, 26 Apr 2019 12:27:11 -0700 > Christoph Hellwig <hch at infradead.org> wrote: > >> On Fri, Apr 26, 2019 at 08:32:39PM +0200, Halil Pasic wrote: >>> +EXPORT_SYMBOL_GPL(set_memory_encrypted); >> >>> +EXPORT_SYMBOL_GPL(set_memory_decrypted); >> >>> +EXPORT_SYMBOL_GPL(sev_active); >>

On Fri, Apr 26, 2019 at 08:32:39PM +0200, Halil Pasic wrote: > +EXPORT_SYMBOL_GPL(set_memory_encrypted); > +EXPORT_SYMBOL_GPL(set_memory_decrypted); > +EXPORT_SYMBOL_GPL(sev_active); Why do you export these? I know x86 exports those as well, but it shoudn't be needed there either.

On arm64 the dma_map_ops implementation is based on the swiotlb. swiotlb-xen, used by default in dom0 on Xen, is also based on the swiotlb. Avoid calling into the default arm64 dma_map_ops functions from xen_dma_map_page, xen_dma_unmap_page, xen_dma_sync_single_for_cpu, and xen_dma_sync_single_for_device otherwise we end up calling into the swiotlb twice. When arm64 gets a non-swiotlb based