Re: Reach-out for help getting qemu-e1000-demo config right+working

2020-01-13 Thread Jan Kiszka
On 13.01.20 10:29, Thorsten Schulz wrote: Hej group, I am stuck getting the e1000-demo working in qemu based on jailhouse-images/wip/update. Due to a prior glitch, I am still on PREEMPT-RT (./build-image.sh --rt --latest) if that makes a difference. Upfront maybe as a "bug-note" /if/

[PATCH v3 3/5] arm64: ti-pvu: Add support for ti-pvu iommu unit

2020-01-13 Thread nikhil.nd via Jailhouse
From: Nikhil Devshatwar Add support for Texas Instrument's Peripheral Virtualization Unit * Define a new IOMMU type and extra fields in the platform_data * Add new cofig option CONFIG_IOMMU_TI_PVU * Integrate with the arm iommu support such that multiple types of IOMMU can be supported.

[PATCH v3 2/5] configs: Move amd specific fields in separate struct

2020-01-13 Thread nikhil.nd via Jailhouse
From: Nikhil Devshatwar Create a union for all vendor specific fields and move the amd specific fields in separate struct. Also update the amd unit references of these fields. This is to handle multiple iommu devices and their custom fields separately. Signed-off-by: Nikhil Devshatwar ---

[PATCH v3 4/5] configs: arm64: k3-j721e-evm: Add PVU IOMMU devices in platform_data

2020-01-13 Thread nikhil.nd via Jailhouse
From: Nikhil Devshatwar J721e device has 3 instance of PVU which can be used as IOMMU. Each PVU has a config region and a TLB region where the memory mapping information is stored. Describe these as part of the root cell's platform_data. Signed-off-by: Nikhil Devshatwar --- Notes: Changes

[PATCH v3 1/5] core: Update cell_state while destroying the cell

2020-01-13 Thread nikhil.nd via Jailhouse
From: Nikhil Devshatwar Update the cell_state to SHUT_DOWN as part of the cell_destroy This will make sure that the memory_unmap calls and unit's cell_exit calls can see the correct status of the cell. Signed-off-by: Nikhil Devshatwar --- hypervisor/control.c | 2 ++ 1 file changed, 2

[PATCH v3 0/5] Add support for Texas Instrument's Peripheral Virtualization Unit

2020-01-13 Thread nikhil.nd via Jailhouse
From: Nikhil Devshatwar This series adds support for TI PVU as an iommu unit. PVU is a 2nd stage only IOMMU which provides realtime address translation. J721e has 3 instances of PVU and all the DMA traffic can be routed via PVU when running inside a virtual machine. Nikhil Devshatwar (5):

[PATCH v3 5/5] configs: arm64: k3-j721e-evm: Add stream ids for devices behind IOMMU

2020-01-13 Thread nikhil.nd via Jailhouse
From: Nikhil Devshatwar Add stream_ids for peripherals which are behind IOMMU instances. PVU and SMMU-V3 sets up memory mapping for all of these contexts for correct 2nd stage translation. Signed-off-by: Nikhil Devshatwar --- configs/arm64/k3-j721e-evm-linux-demo.c | 7 +++

Re: [PATCH v3 0/5] Add support for Texas Instrument's Peripheral Virtualization Unit

2020-01-13 Thread 'Nikhil Devshatwar' via Jailhouse
On 13/01/20 6:08 pm, Jan Kiszka wrote: On 13.01.20 11:46, nikhil...@ti.com wrote: From: Nikhil Devshatwar This series adds support for TI PVU as an iommu unit. PVU is a 2nd stage only IOMMU which provides realtime address translation. J721e has 3 instances of PVU and all the DMA traffic

[PATCH v4] arm64: ti-pvu: Add support for ti-pvu iommu unit

2020-01-13 Thread Jan Kiszka
From: Nikhil Devshatwar Add support for Texas Instrument's Peripheral Virtualization Unit * Define a new IOMMU type and extra fields in the platform_data * Add new cofig option CONFIG_IOMMU_TI_PVU * Integrate with the arm iommu support such that multiple types of IOMMU can be supported.

Re: [PATCH v3 0/5] Add support for Texas Instrument's Peripheral Virtualization Unit

2020-01-13 Thread Jan Kiszka
On 13.01.20 13:38, Jan Kiszka wrote: On 13.01.20 11:46, nikhil...@ti.com wrote: From: Nikhil Devshatwar This series adds support for TI PVU as an iommu unit. PVU is a 2nd stage only IOMMU which provides realtime address translation. J721e has 3 instances of PVU and all the DMA traffic can

[siemens/jailhouse] ce9a66: arm/arm64: Move iommu.o out of arm-common

2020-01-13 Thread Nikhil Devshatwar
Branch: refs/heads/next Home: https://github.com/siemens/jailhouse Commit: ce9a667714d92711b7950d32b8c68dc68088a72e https://github.com/siemens/jailhouse/commit/ce9a667714d92711b7950d32b8c68dc68088a72e Author: Jan Kiszka Date: 2020-01-13 (Mon, 13 Jan 2020) Changed paths:

Re: [PATCH v3 0/5] Add support for Texas Instrument's Peripheral Virtualization Unit

2020-01-13 Thread Jan Kiszka
On 13.01.20 11:46, nikhil...@ti.com wrote: From: Nikhil Devshatwar This series adds support for TI PVU as an iommu unit. PVU is a 2nd stage only IOMMU which provides realtime address translation. J721e has 3 instances of PVU and all the DMA traffic can be routed via PVU when running inside a

[siemens/jailhouse] 05b632: core: Update cell_state while destroying the cell

2020-01-13 Thread Nikhil Devshatwar
Branch: refs/heads/next Home: https://github.com/siemens/jailhouse Commit: 05b6322f057a4091be5323e33cec02c129c6006a https://github.com/siemens/jailhouse/commit/05b6322f057a4091be5323e33cec02c129c6006a Author: Nikhil Devshatwar Date: 2020-01-13 (Mon, 13 Jan 2020) Changed

Re: [PATCH v3 0/5] Add support for Texas Instrument's Peripheral Virtualization Unit

2020-01-13 Thread Jan Kiszka
On 13.01.20 14:09, Nikhil Devshatwar wrote: On 13/01/20 6:08 pm, Jan Kiszka wrote: On 13.01.20 11:46, nikhil...@ti.com wrote: From: Nikhil Devshatwar This series adds support for TI PVU as an iommu unit. PVU is a 2nd stage only IOMMU which provides realtime address translation. J721e

Re: [PATCH] arm-common: gic-v3: ensure LR writes are visible

2020-01-13 Thread Chase Conklin
On 1/12/20, 11:58 PM, "Jan Kiszka" wrote: > On 09.01.20 16:55, Chase Conklin wrote: >> The GICv3 architecture does not guarantee that writes to the list >> registers are self-synchronizing. As a result, it is possible for a >> valid interrupt to be written into a list register but have the empty

[PATCH] arm/arm64: Move iommu.o out of arm-common

2020-01-13 Thread Jan Kiszka
From: Jan Kiszka There is no IOMMU support for 32-bit arm, and it's likely to now show up there anymore. Make the iommu binding module arch-specific so that we can add calls to the arm64 variant without affecting arm. Signed-off-by: Jan Kiszka --- This will go before the PVU series, and the

Re: [PATCH v3 0/5] Add support for Texas Instrument's Peripheral Virtualization Unit

2020-01-13 Thread Jan Kiszka
On 13.01.20 14:37, Jan Kiszka wrote: On 13.01.20 13:38, Jan Kiszka wrote: On 13.01.20 11:46, nikhil...@ti.com wrote: From: Nikhil Devshatwar This series adds support for TI PVU as an iommu unit. PVU is a 2nd stage only IOMMU which provides realtime address translation. J721e has 3

Re: [PATCH v4] arm64: ti-pvu: Add support for ti-pvu iommu unit

2020-01-13 Thread 'Nikhil Devshatwar' via Jailhouse
On 13/01/20 7:07 pm, Jan Kiszka wrote: From: Nikhil Devshatwar Add support for Texas Instrument's Peripheral Virtualization Unit * Define a new IOMMU type and extra fields in the platform_data * Add new cofig option CONFIG_IOMMU_TI_PVU * Integrate with the arm iommu support such that