[virtio-dev] Re: [PATCH 1/4] iommu: Add virtio-iommu driver
On 23/03/18 08:27, Tian, Kevin wrote: >>> The host kernel needs to have *some* MSI region in place before the >>> guest can start configuring interrupts, otherwise it won't know what >>> address to give to the underlying hardware. However, as soon as the host >>> kernel has picked a region, host userspace needs to know that it can no >>> longer use addresses in that region for DMA-able guest memory. It's a >>> lot easier when the address is fixed in hardware and the host userspace >>> will never be stupid enough to try and VFIO_IOMMU_DMA_MAP it, but in >>> the >>> more general case where MSI writes undergo IOMMU address translation >>> so >>> it's an arbitrary IOVA, this has the potential to conflict with stuff >>> like guest memory hotplug. >>> >>> What we currently have is just the simplest option, with the host kernel >>> just picking something up-front and pretending to host userspace that >>> it's a fixed hardware address. There's certainly scope for it to be a >>> bit more dynamic in the sense of adding an interface to let userspace >>> move it around (before attaching any devices, at least), but I don't >>> think it's feasible for the host kernel to second-guess userspace enough >>> to make it entirely transparent like it is in the DMA API domain case. >>> >>> Of course, that's all assuming the host itself is using a virtio-iommu >>> (e.g. in a nested virt or emulation scenario). When it's purely within a >>> guest then an MSI reservation shouldn't matter so much, since the guest >>> won't be anywhere near the real hardware configuration anyway. >>> >>> Robin. >> >> Curious since anyway we are defining a new iommu architecture >> is it possible to avoid those ARM-specific burden completely? >> > > OK, after some study around those tricks below is my learning: > > - MSI_IOVA window is used only on request (iommu_dma_get > _msi_page), not meant to take effect on all architectures once > initialized. e.g. ARM GIC does it but not x86. So it is reasonable > for virtio-iommu driver to implement such capability; > > - I thought whether hardware MSI doorbell can be always reported > on virtio-iommu since it's newly defined. Looks there is a problem > if underlying IOMMU is sw-managed MSI style - valid mapping is > expected in all level of translations, meaning guest has to manage > stage-1 mapping in nested configuration since stage-1 is owned > by guest. > > Then virtio-iommu is naturally expected to report the same MSI > model as supported by underlying hardware. Below are some > further thoughts along this route (use 'IOMMU' to represent the > physical one and 'virtio-iommu' for virtual one): > > > > In the scope of current virtio-iommu spec v.6, there is no nested > consideration yet. Guest driver is expected to use MAP/UNMAP > interface on assigned endpoints. In this case the MAP requests > (IOVA->GPA) is caught and maintained within Qemu which then > further talks to VFIO to map IOVA->HPA in IOMMU. > > Qemu can learn the MSI model of IOMMU from sysfs. > > For hardware MSI doorbell (x86 and some ARM): > * Host kernel reports to Qemu as IOMMU_RESV_MSI > * Qemu report to guest as VIRTIO_IOMMU_RESV_MEM_T_MSI > * Guest takes the range as IOMMU_RESV_MSI. reserved > * Qemu MAP database has no mapping for the doorbell > * Physical IOMMU page table has no mapping for the doorbell > * MSI from passthrough device bypass IOMMU > * MSI from emulated device bypass virtio-iommu > > For software MSI doorbell (most ARM): > * Host kernel reports to Qemu as IOMMU_RESV_SW_MSI > * Qemu report to guest as VIRTIO_IOMMU_RESV_MEM_T_RESERVED > * Guest takes the range as IOMMU_RESV_RESERVED > * vGIC requests to map 'GPA of the virtual doorbell' > * a map request (IOVA->GPA) sent on endpoint > * Qemu maintains the mapping in MAP database > * but no VFIO_MAP request since it's purely virtual > * GIC requests to map 'HPA of the physical doorbell' > * e.g. triggered by VFIO enable msi > * IOMMU now includes a valid mapping (IOVA->HPA) > * MSI from emulated device go through Qemu MAP > database (IOVA->'GPA of virtual doorbell') and then hit vGIC > * MSI from passthrough device go through IOMMU > (IOVA->'HPA of physical doorbell') and then hit GIC > > In this case, host doorbell is treated as reserved resource in > guest side. Guest has its own sw-management for virtual > doorbell which is only used for emulated device. two paths > are completely separated. > > If above captures the right flow, current v0.6 spec is complete > regarding to required function definition. Yes I think this summarizes well the current state or SW/HW MSI > Then comes nested case, with two level page tables (stage-1 > and stage-2) in IOMMU. stage-1 is for IOVA->GPA and stage-2 > is for GPA->HPA. VFIO map/unmap happens on stage-2, > while stage-1 is directly managed by guest (and bound to > IOMMU which enables nested translation from IOVA->GPA > ->HPA). > > For hardware MSI, there is nothing special compared to >
[virtio-dev] RE: [PATCH 1/4] iommu: Add virtio-iommu driver
> From: Tian, Kevin > Sent: Thursday, March 22, 2018 6:06 PM > > > From: Robin Murphy [mailto:robin.mur...@arm.com] > > Sent: Wednesday, March 21, 2018 10:24 PM > > > > On 21/03/18 13:14, Jean-Philippe Brucker wrote: > > > On 21/03/18 06:43, Tian, Kevin wrote: > > > [...] > > >>> + > > >>> +#include > > >>> + > > >>> +#define MSI_IOVA_BASE 0x800 > > >>> +#define MSI_IOVA_LENGTH0x10 > > >> > > >> this is ARM specific, and according to virtio-iommu spec isn't it > > >> better probed on the endpoint instead of hard-coding here? > > > > > > These values are arbitrary, not really ARM-specific even if ARM is the > > > only user yet: we're just reserving a random IOVA region for mapping > > MSIs. > > > It is hard-coded because of the way iommu-dma.c works, but I don't > > quite > > > remember why that allocation isn't dynamic. > > > > The host kernel needs to have *some* MSI region in place before the > > guest can start configuring interrupts, otherwise it won't know what > > address to give to the underlying hardware. However, as soon as the host > > kernel has picked a region, host userspace needs to know that it can no > > longer use addresses in that region for DMA-able guest memory. It's a > > lot easier when the address is fixed in hardware and the host userspace > > will never be stupid enough to try and VFIO_IOMMU_DMA_MAP it, but in > > the > > more general case where MSI writes undergo IOMMU address translation > > so > > it's an arbitrary IOVA, this has the potential to conflict with stuff > > like guest memory hotplug. > > > > What we currently have is just the simplest option, with the host kernel > > just picking something up-front and pretending to host userspace that > > it's a fixed hardware address. There's certainly scope for it to be a > > bit more dynamic in the sense of adding an interface to let userspace > > move it around (before attaching any devices, at least), but I don't > > think it's feasible for the host kernel to second-guess userspace enough > > to make it entirely transparent like it is in the DMA API domain case. > > > > Of course, that's all assuming the host itself is using a virtio-iommu > > (e.g. in a nested virt or emulation scenario). When it's purely within a > > guest then an MSI reservation shouldn't matter so much, since the guest > > won't be anywhere near the real hardware configuration anyway. > > > > Robin. > > Curious since anyway we are defining a new iommu architecture > is it possible to avoid those ARM-specific burden completely? > OK, after some study around those tricks below is my learning: - MSI_IOVA window is used only on request (iommu_dma_get _msi_page), not meant to take effect on all architectures once initialized. e.g. ARM GIC does it but not x86. So it is reasonable for virtio-iommu driver to implement such capability; - I thought whether hardware MSI doorbell can be always reported on virtio-iommu since it's newly defined. Looks there is a problem if underlying IOMMU is sw-managed MSI style - valid mapping is expected in all level of translations, meaning guest has to manage stage-1 mapping in nested configuration since stage-1 is owned by guest. Then virtio-iommu is naturally expected to report the same MSI model as supported by underlying hardware. Below are some further thoughts along this route (use 'IOMMU' to represent the physical one and 'virtio-iommu' for virtual one): In the scope of current virtio-iommu spec v.6, there is no nested consideration yet. Guest driver is expected to use MAP/UNMAP interface on assigned endpoints. In this case the MAP requests (IOVA->GPA) is caught and maintained within Qemu which then further talks to VFIO to map IOVA->HPA in IOMMU. Qemu can learn the MSI model of IOMMU from sysfs. For hardware MSI doorbell (x86 and some ARM): * Host kernel reports to Qemu as IOMMU_RESV_MSI * Qemu report to guest as VIRTIO_IOMMU_RESV_MEM_T_MSI * Guest takes the range as IOMMU_RESV_MSI. reserved * Qemu MAP database has no mapping for the doorbell * Physical IOMMU page table has no mapping for the doorbell * MSI from passthrough device bypass IOMMU * MSI from emulated device bypass virtio-iommu For software MSI doorbell (most ARM): * Host kernel reports to Qemu as IOMMU_RESV_SW_MSI * Qemu report to guest as VIRTIO_IOMMU_RESV_MEM_T_RESERVED * Guest takes the range as IOMMU_RESV_RESERVED * vGIC requests to map 'GPA of the virtual doorbell' * a map request (IOVA->GPA) sent on endpoint * Qemu maintains the mapping in MAP database * but no VFIO_MAP request since it's purely virtual * GIC requests to map 'HPA of the physical doorbell' * e.g. triggered by VFIO enable msi * IOMMU now includes a valid mapping (IOVA->HPA) * MSI from emulated device go through Qemu MAP database (IOVA->'GPA of virtual doorbell') and then hit vGIC * MSI from passthrough device go through IOMMU (IOVA->'HPA of physical doorbell') and then hit GIC In this case, host
[virtio-dev] RE: [PATCH 1/4] iommu: Add virtio-iommu driver
> From: Robin Murphy [mailto:robin.mur...@arm.com] > Sent: Wednesday, March 21, 2018 10:24 PM > > On 21/03/18 13:14, Jean-Philippe Brucker wrote: > > On 21/03/18 06:43, Tian, Kevin wrote: > > [...] > >>> + > >>> +#include > >>> + > >>> +#define MSI_IOVA_BASE0x800 > >>> +#define MSI_IOVA_LENGTH 0x10 > >> > >> this is ARM specific, and according to virtio-iommu spec isn't it > >> better probed on the endpoint instead of hard-coding here? > > > > These values are arbitrary, not really ARM-specific even if ARM is the > > only user yet: we're just reserving a random IOVA region for mapping > MSIs. > > It is hard-coded because of the way iommu-dma.c works, but I don't > quite > > remember why that allocation isn't dynamic. > > The host kernel needs to have *some* MSI region in place before the > guest can start configuring interrupts, otherwise it won't know what > address to give to the underlying hardware. However, as soon as the host > kernel has picked a region, host userspace needs to know that it can no > longer use addresses in that region for DMA-able guest memory. It's a > lot easier when the address is fixed in hardware and the host userspace > will never be stupid enough to try and VFIO_IOMMU_DMA_MAP it, but in > the > more general case where MSI writes undergo IOMMU address translation > so > it's an arbitrary IOVA, this has the potential to conflict with stuff > like guest memory hotplug. > > What we currently have is just the simplest option, with the host kernel > just picking something up-front and pretending to host userspace that > it's a fixed hardware address. There's certainly scope for it to be a > bit more dynamic in the sense of adding an interface to let userspace > move it around (before attaching any devices, at least), but I don't > think it's feasible for the host kernel to second-guess userspace enough > to make it entirely transparent like it is in the DMA API domain case. > > Of course, that's all assuming the host itself is using a virtio-iommu > (e.g. in a nested virt or emulation scenario). When it's purely within a > guest then an MSI reservation shouldn't matter so much, since the guest > won't be anywhere near the real hardware configuration anyway. > > Robin. Curious since anyway we are defining a new iommu architecture is it possible to avoid those ARM-specific burden completely? Thanks Kevin
[virtio-dev] Re: [PATCH 1/4] iommu: Add virtio-iommu driver
On 21/03/18 06:43, Tian, Kevin wrote: [...] >> + >> +#include >> + >> +#define MSI_IOVA_BASE 0x800 >> +#define MSI_IOVA_LENGTH 0x10 > > this is ARM specific, and according to virtio-iommu spec isn't it > better probed on the endpoint instead of hard-coding here? These values are arbitrary, not really ARM-specific even if ARM is the only user yet: we're just reserving a random IOVA region for mapping MSIs. It is hard-coded because of the way iommu-dma.c works, but I don't quite remember why that allocation isn't dynamic. As said on the v0.6 spec thread, I'm not sure allocating the IOVA range in the host is preferable. With nested translation the guest has to map it anyway, and I believe dealing with IOVA allocation should be left to the guest when possible. Thanks, Jean - To unsubscribe, e-mail: virtio-dev-unsubscr...@lists.oasis-open.org For additional commands, e-mail: virtio-dev-h...@lists.oasis-open.org
[virtio-dev] Re: [PATCH 1/4] iommu: Add virtio-iommu driver
On Thu, Feb 22, 2018 at 11:04:30AM +, Jean-Philippe Brucker wrote: > On 21/02/18 20:12, kbuild test robot wrote: > [...] > > config: arm64-allmodconfig (attached as .config) > [...] > >aarch64-linux-gnu-ld: arch/arm64/kernel/head.o: relocation > > R_AARCH64_ABS32 against `_kernel_offset_le_lo32' can not be used when > > making a shared object > >arch/arm64/kernel/head.o: In function `kimage_vaddr': > >(.idmap.text+0x0): dangerous relocation: unsupported relocation > > Is this related? > > >arch/arm64/kernel/head.o: In function `__primary_switch': > >(.idmap.text+0x340): dangerous relocation: unsupported relocation > >(.idmap.text+0x348): dangerous relocation: unsupported relocation > >drivers/iommu/virtio-iommu.o: In function `viommu_probe': > >virtio-iommu.c:(.text+0xbdc): undefined reference to > > `virtio_check_driver_offered_feature' > >virtio-iommu.c:(.text+0xcfc): undefined reference to > > `virtio_check_driver_offered_feature' > >virtio-iommu.c:(.text+0xe10): undefined reference to > > `virtio_check_driver_offered_feature' > >drivers/iommu/virtio-iommu.o: In function `viommu_send_reqs_sync': > >virtio-iommu.c:(.text+0x130c): undefined reference to `virtqueue_add_sgs' > >virtio-iommu.c:(.text+0x1398): undefined reference to `virtqueue_kick' > >virtio-iommu.c:(.text+0x14d4): undefined reference to `virtqueue_get_buf' > >drivers/iommu/virtio-iommu.o: In function `virtio_iommu_drv_init': > >virtio-iommu.c:(.init.text+0x1c): undefined reference to > > `register_virtio_driver' > >drivers/iommu/virtio-iommu.o: In function `virtio_iommu_drv_exit': > >>> virtio-iommu.c:(.exit.text+0x14): undefined reference to > >>> `unregister_virtio_driver' > > Right. At the moment CONFIG_VIRTIO_IOMMU is a bool instead of tristate, > because the IOMMU subsystem isn't entirely ready to have IOMMU drivers > built as modules. In addition to exporting symbols it would also needs to > hold off probing endpoints behind the IOMMU until the IOMMU driver is > loaded. At the moment (I think) it gives up once userspace is reached (see > of_iommu_driver_present). > > The above report is due to VIRTIO=m and VIRTIO_IOMMU=y. To solve it we could: > > a) Allow VIRTIO_IOMMU to be built as module by exporting a dozen IOMMU > symbols. It would be a lie. The driver wouldn't be usable because of the > probe issue discussed above, but it would build. > > b) Actually make any IOMMU driver work as module. Whilst it would > certainly be a nice feature, it's a bigger problem and I don't think it > has its place in this series. > > c) Make VIRTIO_IOMMU depend on VIRTIO_MMIO=y instead of VIRTIO_MMIO, which > I think is the sanest for now (and does work), even though distro kernels > probably all have VIRTIO=m. > > I prefer doing c) now and experiment with b) once I got some time. > > Thanks, > Jean It kind of means it's a toy for now though. So fine as long as it's out of tree. -- MST - To unsubscribe, e-mail: virtio-dev-unsubscr...@lists.oasis-open.org For additional commands, e-mail: virtio-dev-h...@lists.oasis-open.org
[virtio-dev] Re: [PATCH 1/4] iommu: Add virtio-iommu driver
On 21/02/18 20:12, kbuild test robot wrote: [...] > config: arm64-allmodconfig (attached as .config) [...] >aarch64-linux-gnu-ld: arch/arm64/kernel/head.o: relocation R_AARCH64_ABS32 > against `_kernel_offset_le_lo32' can not be used when making a shared object >arch/arm64/kernel/head.o: In function `kimage_vaddr': >(.idmap.text+0x0): dangerous relocation: unsupported relocation Is this related? >arch/arm64/kernel/head.o: In function `__primary_switch': >(.idmap.text+0x340): dangerous relocation: unsupported relocation >(.idmap.text+0x348): dangerous relocation: unsupported relocation >drivers/iommu/virtio-iommu.o: In function `viommu_probe': >virtio-iommu.c:(.text+0xbdc): undefined reference to > `virtio_check_driver_offered_feature' >virtio-iommu.c:(.text+0xcfc): undefined reference to > `virtio_check_driver_offered_feature' >virtio-iommu.c:(.text+0xe10): undefined reference to > `virtio_check_driver_offered_feature' >drivers/iommu/virtio-iommu.o: In function `viommu_send_reqs_sync': >virtio-iommu.c:(.text+0x130c): undefined reference to `virtqueue_add_sgs' >virtio-iommu.c:(.text+0x1398): undefined reference to `virtqueue_kick' >virtio-iommu.c:(.text+0x14d4): undefined reference to `virtqueue_get_buf' >drivers/iommu/virtio-iommu.o: In function `virtio_iommu_drv_init': >virtio-iommu.c:(.init.text+0x1c): undefined reference to > `register_virtio_driver' >drivers/iommu/virtio-iommu.o: In function `virtio_iommu_drv_exit': >>> virtio-iommu.c:(.exit.text+0x14): undefined reference to >>> `unregister_virtio_driver' Right. At the moment CONFIG_VIRTIO_IOMMU is a bool instead of tristate, because the IOMMU subsystem isn't entirely ready to have IOMMU drivers built as modules. In addition to exporting symbols it would also needs to hold off probing endpoints behind the IOMMU until the IOMMU driver is loaded. At the moment (I think) it gives up once userspace is reached (see of_iommu_driver_present). The above report is due to VIRTIO=m and VIRTIO_IOMMU=y. To solve it we could: a) Allow VIRTIO_IOMMU to be built as module by exporting a dozen IOMMU symbols. It would be a lie. The driver wouldn't be usable because of the probe issue discussed above, but it would build. b) Actually make any IOMMU driver work as module. Whilst it would certainly be a nice feature, it's a bigger problem and I don't think it has its place in this series. c) Make VIRTIO_IOMMU depend on VIRTIO_MMIO=y instead of VIRTIO_MMIO, which I think is the sanest for now (and does work), even though distro kernels probably all have VIRTIO=m. I prefer doing c) now and experiment with b) once I got some time. Thanks, Jean - To unsubscribe, e-mail: virtio-dev-unsubscr...@lists.oasis-open.org For additional commands, e-mail: virtio-dev-h...@lists.oasis-open.org
[virtio-dev] Re: [PATCH 1/4] iommu: Add virtio-iommu driver
On 19/02/18 12:23, Tomasz Nowicki wrote: [...] >> +static int viommu_receive_resp(struct viommu_dev *viommu, int nr_sent, >> + struct list_head *sent) >> +{ >> + >> +unsigned int len; >> +int nr_received = 0; >> +struct viommu_request *req, *pending; >> + >> +pending = list_first_entry_or_null(sent, struct viommu_request, list); >> +if (WARN_ON(!pending)) >> +return 0; >> + >> +while ((req = virtqueue_get_buf(viommu->vq, )) != NULL) { >> +if (req != pending) { >> +dev_warn(viommu->dev, "discarding stale request\n"); >> +continue; >> +} >> + >> +pending->written = len; >> + >> +if (++nr_received == nr_sent) { >> +WARN_ON(!list_is_last(>list, sent)); >> +break; >> +} else if (WARN_ON(list_is_last(>list, sent))) { >> +break; >> +} >> + >> +pending = list_next_entry(pending, list); > > We should remove current element from the pending list. There is no > guarantee we get response for each while loop so when we get back for > more the _viommu_send_reqs_sync() caller will pass pointer to the out of > date head next time. Right, I'll fix this Thanks, Jean - To unsubscribe, e-mail: virtio-dev-unsubscr...@lists.oasis-open.org For additional commands, e-mail: virtio-dev-h...@lists.oasis-open.org