| Message ID | 1486496525-14637-6-git-send-email-edgar.iglesias@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, 7 Feb 2017, Edgar E. Iglesias wrote: > From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> > > Implement an EEMI mediator and forward platform specific > firmware calls from guests to firmware. > > The EEMI mediator is responsible for implementing access > controls modifying or blocking calls that try to operate > on setup for devices that are not under the calling guest's > control. > > EEMI: > https://www.xilinx.com/support/documentation/user_guides/ug1200-eemi-api.pdf > > Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> > --- > xen/arch/arm/platforms/Makefile | 1 + > xen/arch/arm/platforms/xilinx-zynqmp-eemi.c | 794 +++++++++++++++++++++ > xen/arch/arm/platforms/xilinx-zynqmp.c | 18 + > xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h | 337 +++++++++ > xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h | 284 ++++++++ > 5 files changed, 1434 insertions(+) > create mode 100644 xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > create mode 100644 xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h > create mode 100644 xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h > > diff --git a/xen/arch/arm/platforms/Makefile b/xen/arch/arm/platforms/Makefile > index 49fa683..b58b71f 100644 > --- a/xen/arch/arm/platforms/Makefile > +++ b/xen/arch/arm/platforms/Makefile > @@ -8,3 +8,4 @@ obj-$(CONFIG_ARM_64) += seattle.o > obj-$(CONFIG_ARM_32) += sunxi.o > obj-$(CONFIG_ARM_64) += xgene-storm.o > obj-$(CONFIG_ARM_64) += xilinx-zynqmp.o > +obj-$(CONFIG_ARM_64) += xilinx-zynqmp-eemi.o > diff --git a/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c b/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > new file mode 100644 > index 0000000..c2c184d > --- /dev/null > +++ b/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > @@ -0,0 +1,794 @@ > +/* > + * xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > + * > + * Xilinx ZynqMP EEMI API mediator. > + * > + * Copyright (c) 2017 Xilinx Inc. > + * Written by Edgar E. Iglesias <edgar.iglesias@xilinx.com> > + * > + * This program is free software; you can redistribute it and/or modify > + * it under the terms of the GNU General Public License as published by > + * the Free Software Foundation; either version 2 of the License, or > + * (at your option) any later version. > + * > + * This program is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the > + * GNU General Public License for more details. > + */ > + > +/* > + * Mediator for the EEMI Power Mangement API. > + * > + * Refs: > + * https://www.xilinx.com/support/documentation/user_guides/ug1200-eemi-api.pdf > + * > + * Background: > + * The ZynqMP has a subsystem named the PMU with a CPU and special devices > + * dedicated to running Power Management Firmware. Other masters in the > + * system need to send requests to the PMU in order to for example: > + * * Manage power state > + * * Configure clocks > + * * Program bitstreams for the programmable logic > + * * etc > + * > + * Allthough the details of the setup are configurable, in the common case > + * the PMU lives in the Secure world. NS World cannot directly communicate > + * with it and must use proxy services from ARM Trusted Firmware to reach > + * the PMU. > + * > + * Power Management on the ZynqMP is implemented in a layered manner. > + * The PMU knows about various masters and will enforce access controls > + * based on a pre-configured partitioning. This configuration dictates > + * which devices are owned by the various masters and the PMU FW makes sure > + * that a given master cannot turn off a device that it does not own or that > + * is in use by other masters. > + * > + * The PMU is not aware of multiple execution states in masters. > + * For example, it treats the ARMv8 cores as single units and does not > + * distinguish between Secure vs NS OS:s nor does it know about Hypervisors > + * and multiple guests. It is up to software on the ARMv8 cores to present > + * a unified view of its power requirements. > + * > + * To implement this unified view, ARM Trusted Firmware at EL3 provides > + * access to the PM API via SMC calls. ARM Trusted Firmware is responsible > + * for mediating between the Secure and the NS world, rejecting SMC calls > + * that request changes that are not allowed. > + * > + * Xen running above ATF owns the NS world and is responsible for presenting > + * unified PM requests taking all guests and the hypervisor into account. > + * > + * Implementation: > + * The PM API contains different classes of calls. > + * Certain calls are harmless to expose to any guest. > + * These include calls to get the PM API Version, or to read out the version > + * of the chip we're running on. > + * > + * Other calls are more problematic. Here're some: > + * * Power requests for nodes (i.e turn on or off a given device) > + * * Reset line control (i.e reset a given device) > + * * MMIO access (i.e directly access clock and reset registers) > + * > + * Power requests for nodes: > + * In order to correctly mediate these calls, we need to know if > + * guests issuing these calls have ownership of the given device. > + * The approach taken here is to map PM API Nodes identifying > + * a device into base addresses for registers that belong to that > + * same device. > + * > + * If the guest has access to a devices registers, we give the guest > + * access to PM API calls that affect that device. This is implemented > + * by pm_node_access and domain_has_node_access(). > + * > + * Reset lines: > + * Reset lines are identified by a list of identifiers that do not relate > + * to device nodes. We can use the same approach here as for nodes, except > + * that we need a separate table mapping reset lines to base addresses. > + * This is implemented by pm_reset_access. > + * > + * MMIO access: > + * These calls allow guests to access certain memory ranges. These ranges > + * are typically protected for secure-world access only and also from > + * certain masters only, so guests cannot access them directly. > + * Registers within the memory regions affect certain nodes. In this case, > + * our imput is an address and we map that address into a node. If the ^ input > + * guest has ownership of that node, the access is allowed. > + * Some registers contain bitfields and a single register may contain > + * bits that affect multiple nodes. > + * > + * Some of the calls act on more global state. For example acting on > + * NODE_PS, which affects many a lot of nodes. This higher level > + * orchestrating is left for the hardware-domain only. > + */ > + > +#include <xen/iocap.h> > +#include <asm/platform.h> > +#include <asm/platforms/xilinx-zynqmp-eemi.h> > +#include <asm/platforms/xilinx-zynqmp-mm.h> > + > +struct pm_access > +{ > + uint64_t addr; > + bool hwdom_access; /* HW domain gets access regardless. */ If the hardware domain gets access regardless, I don't think we need this struct field, right? If fact, is_hardware_domain() is basically used as a blank check in Xen. I would remove hwdom_access, use is_hardware_domain() instead, and simply comment out any entries we don't want dom0 to have access to. For now that would be fine, but in the future we'll need a way to prevent dom0 access to something, if it has been assigned to another guest. > +}; > + > +/* > + * This table maps a node into a memory address. > + * If a guest has access to the address, it has enough control > + * over the node to grant it access to EEMI calls for that node. > + */ > +static const struct pm_access pm_node_access[] = { > + /* MM_RPU grants access to alll RPU Nodes. */ ^ all > + [NODE_RPU] = { MM_RPU }, > + [NODE_RPU_0] = { MM_RPU }, > + [NODE_RPU_1] = { MM_RPU }, > + [NODE_IPI_RPU_0] = { MM_RPU }, > + > + /* GPU nodes. */ > + [NODE_GPU] = { MM_GPU }, > + [NODE_GPU_PP_0] = { MM_GPU }, > + [NODE_GPU_PP_1] = { MM_GPU }, > + > + [NODE_USB_0] = { MM_USB3_0 }, > + [NODE_USB_1] = { MM_USB3_1 }, > + [NODE_TTC_0] = { MM_TTC0 }, > + [NODE_TTC_1] = { MM_TTC1 }, > + [NODE_TTC_2] = { MM_TTC2 }, > + [NODE_TTC_3] = { MM_TTC3 }, > + [NODE_SATA] = { MM_SATA_AHCI_HBA }, > + [NODE_ETH_0] = { MM_GEM0 }, > + [NODE_ETH_1] = { MM_GEM1 }, > + [NODE_ETH_2] = { MM_GEM2 }, > + [NODE_ETH_3] = { MM_GEM3 }, > + [NODE_UART_0] = { MM_UART0 }, > + [NODE_UART_1] = { MM_UART1 }, > + [NODE_SPI_0] = { MM_SPI0 }, > + [NODE_SPI_1] = { MM_SPI1 }, > + [NODE_I2C_0] = { MM_I2C0 }, > + [NODE_I2C_1] = { MM_I2C1 }, > + [NODE_SD_0] = { MM_SD0 }, > + [NODE_SD_1] = { MM_SD1 }, > + [NODE_DP] = { MM_DP }, > + > + /* Guest with GDMA Channel 0 gets PM access. Other guests don't. */ > + [NODE_GDMA] = { MM_GDMA_CH0 }, > + /* Guest with ADMA Channel 0 gets PM access. Other guests don't. */ > + [NODE_ADMA] = { MM_ADMA_CH0 }, > + > + [NODE_NAND] = { MM_NAND }, > + [NODE_QSPI] = { MM_QSPI }, > + [NODE_GPIO] = { MM_GPIO }, > + [NODE_CAN_0] = { MM_CAN0 }, > + [NODE_CAN_1] = { MM_CAN1 }, > + > + /* Dom0 only. */ > + [NODE_AFI] = { .hwdom_access = true }, > + [NODE_APLL] = { .hwdom_access = true }, > + [NODE_VPLL] = { .hwdom_access = true }, > + [NODE_DPLL] = { .hwdom_access = true }, > + [NODE_RPLL] = { .hwdom_access = true }, > + [NODE_IOPLL] = { .hwdom_access = true }, > + [NODE_DDR] = { .hwdom_access = true }, > + [NODE_IPI_APU] = { .hwdom_access = true }, > + [NODE_PCAP] = { .hwdom_access = true }, > + > + [NODE_PCIE] = { MM_PCIE_ATTRIB }, > + [NODE_RTC] = { MM_RTC }, > +}; > + > +/* > + * This table maps reset line IDs into a memory address. > + * If a guest has access to the address, it has enough control > + * over the affected node to grant it access to EEMI calls for > + * resetting that node. > + */ > +#define PM_RESET_IDX(n) (n - PM_RESET_PCIE_CFG) > +static const struct pm_access pm_reset_access[] = { > + [PM_RESET_IDX(PM_RESET_PCIE_CFG)] = { MM_AXIPCIE_MAIN }, > + [PM_RESET_IDX(PM_RESET_PCIE_BRIDGE)] = { MM_PCIE_ATTRIB }, > + [PM_RESET_IDX(PM_RESET_PCIE_CTRL)] = { MM_PCIE_ATTRIB }, > + > + [PM_RESET_IDX(PM_RESET_DP)] = { MM_DP }, > + [PM_RESET_IDX(PM_RESET_SWDT_CRF)] = { MM_SWDT }, > + [PM_RESET_IDX(PM_RESET_AFI_FM5)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_AFI_FM4)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_AFI_FM3)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_AFI_FM2)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_AFI_FM1)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_AFI_FM0)] = { .hwdom_access = true }, > + > + /* Channel 0 grants PM access. */ > + [PM_RESET_IDX(PM_RESET_GDMA)] = { MM_GDMA_CH0 }, > + [PM_RESET_IDX(PM_RESET_GPU_PP1)] = { MM_GPU }, > + [PM_RESET_IDX(PM_RESET_GPU_PP0)] = { MM_GPU }, > + [PM_RESET_IDX(PM_RESET_GT)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_SATA)] = { MM_SATA_AHCI_HBA }, > + > + /* We don't allow anyone to turn on/off the ACPUs. */ > + [PM_RESET_IDX(PM_RESET_ACPU3_PWRON)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU2_PWRON)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU1_PWRON)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU0_PWRON)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_APU_L2)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU3)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU2)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU1)] = { 0 }, > + [PM_RESET_IDX(PM_RESET_ACPU0)] = { 0 }, > + > + [PM_RESET_IDX(PM_RESET_DDR)] = { 0 }, > + > + [PM_RESET_IDX(PM_RESET_APM_FPD)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_SOFT)] = { .hwdom_access = true }, > + > + [PM_RESET_IDX(PM_RESET_GEM0)] = { MM_GEM0 }, > + [PM_RESET_IDX(PM_RESET_GEM1)] = { MM_GEM1 }, > + [PM_RESET_IDX(PM_RESET_GEM2)] = { MM_GEM2 }, > + [PM_RESET_IDX(PM_RESET_GEM3)] = { MM_GEM3 }, > + > + [PM_RESET_IDX(PM_RESET_QSPI)] = { MM_QSPI }, > + [PM_RESET_IDX(PM_RESET_UART0)] = { MM_UART0 }, > + [PM_RESET_IDX(PM_RESET_UART1)] = { MM_UART1 }, > + [PM_RESET_IDX(PM_RESET_SPI0)] = { MM_SPI0 }, > + [PM_RESET_IDX(PM_RESET_SPI1)] = { MM_SPI1 }, > + [PM_RESET_IDX(PM_RESET_SDIO0)] = { MM_SD0 }, > + [PM_RESET_IDX(PM_RESET_SDIO1)] = { MM_SD1 }, > + [PM_RESET_IDX(PM_RESET_CAN0)] = { MM_CAN0 }, > + [PM_RESET_IDX(PM_RESET_CAN1)] = { MM_CAN1 }, > + [PM_RESET_IDX(PM_RESET_I2C0)] = { MM_I2C0 }, > + [PM_RESET_IDX(PM_RESET_I2C1)] = { MM_I2C1 }, > + [PM_RESET_IDX(PM_RESET_TTC0)] = { MM_TTC0 }, > + [PM_RESET_IDX(PM_RESET_TTC1)] = { MM_TTC1 }, > + [PM_RESET_IDX(PM_RESET_TTC2)] = { MM_TTC2 }, > + [PM_RESET_IDX(PM_RESET_TTC3)] = { MM_TTC3 }, > + [PM_RESET_IDX(PM_RESET_SWDT_CRL)] = { MM_SWDT }, > + [PM_RESET_IDX(PM_RESET_NAND)] = { MM_NAND }, > + /* ADMA Channel 0 grants access to pull the reset signal. */ > + [PM_RESET_IDX(PM_RESET_ADMA)] = { MM_ADMA_CH0 }, > + [PM_RESET_IDX(PM_RESET_GPIO)] = { MM_GPIO }, > + /* FIXME: What is this? */ > + [PM_RESET_IDX(PM_RESET_IOU_CC)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_TIMESTAMP)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_RPU_R50)] = { MM_RPU }, > + [PM_RESET_IDX(PM_RESET_RPU_R51)] = { MM_RPU }, > + [PM_RESET_IDX(PM_RESET_RPU_AMBA)] = { MM_RPU }, > + [PM_RESET_IDX(PM_RESET_OCM)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_RPU_PGE)] = { MM_RPU }, > + > + [PM_RESET_IDX(PM_RESET_USB0_CORERESET)] = { MM_USB3_0 }, > + [PM_RESET_IDX(PM_RESET_USB0_HIBERRESET)] = { MM_USB3_0 }, > + [PM_RESET_IDX(PM_RESET_USB0_APB)] = { MM_USB3_0 }, > + > + [PM_RESET_IDX(PM_RESET_USB1_CORERESET)] = { MM_USB3_1 }, > + [PM_RESET_IDX(PM_RESET_USB1_HIBERRESET)] = { MM_USB3_1 }, > + [PM_RESET_IDX(PM_RESET_USB1_APB)] = { MM_USB3_1 }, > + > + [PM_RESET_IDX(PM_RESET_IPI)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_APM_LPD)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_RTC)] = { MM_RTC }, > + [PM_RESET_IDX(PM_RESET_SYSMON)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_AFI_FM6)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_LPD_SWDT)] = { MM_SWDT }, > + [PM_RESET_IDX(PM_RESET_FPD)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_RPU_DBG1)] = { MM_RPU }, > + [PM_RESET_IDX(PM_RESET_RPU_DBG0)] = { MM_RPU }, > + [PM_RESET_IDX(PM_RESET_DBG_LPD)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_DBG_FPD)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_APLL)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_DPLL)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_VPLL)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_IOPLL)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_RPLL)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_0)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_1)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_2)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_3)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_4)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_5)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_6)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_7)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_8)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_9)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_10)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_11)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_12)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_13)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_14)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_15)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_16)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_17)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_18)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_19)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_20)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_21)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_22)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_23)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_24)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_25)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_26)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_27)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_28)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_29)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_30)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_GPO3_PL_31)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_RPU_LS)] = { MM_RPU }, > + [PM_RESET_IDX(PM_RESET_PS_ONLY)] = { .hwdom_access = true }, > + [PM_RESET_IDX(PM_RESET_PL)] = { .hwdom_access = true }, > +}; > + > +/* > + * This table maps reset line IDs into a memory address. > + * If a guest has access to the address, it has enough control > + * over the affected node to grant it access to EEMI calls for > + * resetting that node. > + */ > +static const struct { > + uint64_t start; > + uint64_t end; /* Inclusive. If not set, single reg entry. */ > + uint32_t mask; /* Zero means no mask, i.e all bits. */ Please clarify the comment. What is this mask for exactly? > + enum pm_node_id node; > + bool hwdom_access; same here > + bool readonly; > +} pm_mmio_access[] = { > + { > + .start = MM_CRF_APB + R_CRF_APLL_CTRL, > + .end = MM_CRF_APB + R_CRF_ACPU_CTRL, > + .hwdom_access = true > + }, > + { > + .start = MM_CRL_APB + R_CRL_IOPLL_CTRL, > + .end = MM_CRL_APB + R_CRL_RPLL_TO_FPD_CTRL, > + .hwdom_access = true > + }, > + { .start = MM_CRF_APB + R_CRF_DP_VIDEO_REF_CTRL, .node = NODE_DP }, > + { .start = MM_CRF_APB + R_CRF_DP_STC_REF_CTRL, .node = NODE_DP }, > + { .start = MM_CRF_APB + R_CRF_GPU_REF_CTRL, .node = NODE_GPU }, > + { .start = MM_CRF_APB + R_CRF_SATA_REF_CTRL, .node = NODE_SATA }, > + { .start = MM_CRF_APB + R_CRF_PCIE_REF_CTRL, .node = NODE_PCIE }, > + { .start = MM_CRF_APB + R_CRF_GDMA_REF_CTRL, .node = NODE_GDMA }, > + { .start = MM_CRF_APB + R_CRF_DPDMA_REF_CTRL, .node = NODE_DP }, > + { .start = MM_CRL_APB + R_CRL_USB3_DUAL_REF_CTRL, .node = NODE_USB_0 }, > + { .start = MM_CRL_APB + R_CRL_USB0_BUS_REF_CTRL, .node = NODE_USB_0 }, > + { .start = MM_CRL_APB + R_CRL_USB1_BUS_REF_CTRL, .node = NODE_USB_1 }, > + { .start = MM_CRL_APB + R_CRL_USB1_BUS_REF_CTRL, .node = NODE_USB_1 }, > + { .start = MM_CRL_APB + R_CRL_GEM0_REF_CTRL, .node = NODE_ETH_0 }, > + { .start = MM_CRL_APB + R_CRL_GEM1_REF_CTRL, .node = NODE_ETH_1 }, > + { .start = MM_CRL_APB + R_CRL_GEM2_REF_CTRL, .node = NODE_ETH_2 }, > + { .start = MM_CRL_APB + R_CRL_GEM3_REF_CTRL, .node = NODE_ETH_3 }, > + { .start = MM_CRL_APB + R_CRL_QSPI_REF_CTRL, .node = NODE_QSPI }, > + { .start = MM_CRL_APB + R_CRL_SDIO0_REF_CTRL, .node = NODE_SD_0 }, > + { .start = MM_CRL_APB + R_CRL_SDIO1_REF_CTRL, .node = NODE_SD_1 }, > + { .start = MM_CRL_APB + R_CRL_UART0_REF_CTRL, .node = NODE_UART_0 }, > + { .start = MM_CRL_APB + R_CRL_UART1_REF_CTRL, .node = NODE_UART_1 }, > + { .start = MM_CRL_APB + R_CRL_SPI0_REF_CTRL, .node = NODE_SPI_0 }, > + { .start = MM_CRL_APB + R_CRL_SPI1_REF_CTRL, .node = NODE_SPI_1 }, > + { .start = MM_CRL_APB + R_CRL_CAN0_REF_CTRL, .node = NODE_CAN_0 }, > + { .start = MM_CRL_APB + R_CRL_CAN1_REF_CTRL, .node = NODE_CAN_1 }, > + { .start = MM_CRL_APB + R_CRL_CPU_R5_CTRL, .node = NODE_RPU }, > + { .start = MM_CRL_APB + R_CRL_IOU_SWITCH_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_CSU_PLL_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PCAP_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_LPD_SWITCH_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_LPD_LSBUS_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_DBG_LPD_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_NAND_REF_CTRL, .node = NODE_NAND }, > + { .start = MM_CRL_APB + R_CRL_ADMA_REF_CTRL, .node = NODE_ADMA }, > + { .start = MM_CRL_APB + R_CRL_PL0_REF_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL1_REF_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL2_REF_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL3_REF_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL0_THR_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL1_THR_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL2_THR_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL3_THR_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL0_THR_CNT, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL1_THR_CNT, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL2_THR_CNT, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_PL3_THR_CNT, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_GEM_TSU_REF_CTRL, .node = NODE_ETH_0 }, > + { .start = MM_CRL_APB + R_CRL_DLL_REF_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_AMS_REF_CTRL, .hwdom_access = true }, > + { .start = MM_CRL_APB + R_CRL_I2C0_REF_CTRL, .node = NODE_I2C_0 }, > + { .start = MM_CRL_APB + R_CRL_I2C1_REF_CTRL, .node = NODE_I2C_1 }, > + { .start = MM_CRL_APB + R_CRL_TIMESTAMP_REF_CTRL, .hwdom_access = true }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_WDT_CLK_SEL, .hwdom_access = true }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_CAN_MIO_CTRL, > + .mask = 0x1ff, .node = NODE_CAN_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_CAN_MIO_CTRL, > + .mask = 0x1ff << 15, .node = NODE_CAN_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, > + .mask = 0xf, .node = NODE_ETH_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, > + .mask = 0xf << 5, .node = NODE_ETH_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, > + .mask = 0xf << 10, .node = NODE_ETH_2 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, > + .mask = 0xf << 15, .node = NODE_ETH_3 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, > + .mask = 0x7 << 20, .hwdom_access = true > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_SDIO_CLK_CTRL, > + .mask = 0x7, .node = NODE_SD_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_SDIO_CLK_CTRL, > + .mask = 0x7 << 17, .node = NODE_SD_1 > + }, > + > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_CTRL_REG_SD, > + .mask = 0x1, .node = NODE_SD_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_CTRL_REG_SD, > + .mask = 0x1 << 15, .node = NODE_SD_1 > + }, > + /* A series of SD regs with the same layout. */ > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_SD_ITAPDLY, > + .end = MM_IOU_SLCR + R_IOU_SLCR_SD_CDN_CTRL, > + .mask = 0x3ff << 0, .node = NODE_SD_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_SD_ITAPDLY, > + .end = MM_IOU_SLCR + R_IOU_SLCR_SD_CDN_CTRL, > + .mask = 0x3ff << 16, .node = NODE_SD_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, > + .mask = 0x3 << 0, .node = NODE_ETH_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, > + .mask = 0x3 << 2, .node = NODE_ETH_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, > + .mask = 0x3 << 4, .node = NODE_ETH_2 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, > + .mask = 0x3 << 6, .node = NODE_ETH_3 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, > + .mask = 0x3 << 0, .node = NODE_TTC_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, > + .mask = 0x3 << 2, .node = NODE_TTC_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, > + .mask = 0x3 << 4, .node = NODE_TTC_2 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, > + .mask = 0x3 << 6, .node = NODE_TTC_3 > + }, > + > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TAPDLY_BYPASS, > + .mask = 0x3, .node = NODE_NAND > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TAPDLY_BYPASS, > + .mask = 0x1 << 2, .node = NODE_QSPI > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 0, .node = NODE_ETH_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 4, .node = NODE_ETH_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 8, .node = NODE_ETH_2 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 12, .node = NODE_ETH_3 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 16, .node = NODE_SD_0 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 20, .node = NODE_SD_1 > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 24, .node = NODE_NAND > + }, > + { > + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, > + .mask = 0xf << 28, .node = NODE_QSPI > + }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_VIDEO_PSS_CLK_SEL, .hwdom_access = true }, > + /* No access to R_IOU_SLCR_IOU_INTERCONNECT_ROUTE at all. */ > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM0, .node = NODE_ETH_0 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM1, .node = NODE_ETH_1 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM2, .node = NODE_ETH_2 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM3, .node = NODE_ETH_3 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_SD0, .node = NODE_SD_0 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_SD1, .node = NODE_SD_1 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_CAN0, .node = NODE_CAN_0 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_CAN1, .node = NODE_CAN_1 }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_LQSPI, .node = NODE_QSPI }, > + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_NAND, .node = NODE_NAND }, > + { > + .start = MM_PMU_GLOBAL + R_PMU_GLOBAL_PWR_STATE, > + .readonly = true, > + }, > + { > + .start = MM_PMU_GLOBAL + R_PMU_GLOBAL_GLOBAL_GEN_STORAGE0, > + .end = MM_PMU_GLOBAL + R_PMU_GLOBAL_PERS_GLOB_GEN_STORAGE7, > + .readonly = true, > + }, > + { > + /* Universal read-only access to CRF. Linux CCF needs this. */ > + .start = MM_CRF_APB, .end = MM_CRF_APB + 0x104, > + .readonly = true, > + }, > + { > + /* Universal read-only access to CRL. Linux CCF needs this. */ > + .start = MM_CRL_APB, .end = MM_CRL_APB + 0x284, > + .readonly = true, > + } > +}; > + > +static bool pm_check_access(const struct pm_access *acl, struct domain *d, int idx) > +{ > + unsigned long mfn; > + > + if ( acl[idx].hwdom_access && is_hardware_domain(d) ) > + return true; > + > + mfn = paddr_to_pfn(acl[idx].addr); > + if ( !mfn ) > + return false; > + > + return iomem_access_permitted(d, mfn, mfn); I would write this function as: mfn = paddr_to_pfn(acl[idx].addr); if ( mfn != 0 ) return iomem_access_permitted(d, mfn, mfn); return is_hardware_domain(d); This way if addr is valid, then the domain need to have iomem access for it. Otherwise, it needs to be the the hardware domain. > +} > + > +/* Check if a domain has access to a node. */ > +static bool domain_has_node_access(struct domain *d, enum pm_node_id node) > +{ > + if ( node <= 0 || node > ARRAY_SIZE(pm_node_access) ) > + return false; > + > + /* NODE_UNKNOWN is treated as a wildcard. */ > + if ( node == NODE_UNKNOWN ) > + return true; > + > + return pm_check_access(pm_node_access, d, node); > +} > + > +/* Check if a domain has access to a reset line. */ > +static bool domain_has_reset_access(struct domain *d, enum pm_reset rst) > +{ > + int rst_idx = PM_RESET_IDX(rst); > + > + if ( rst_idx < 0 || rst_idx > ARRAY_SIZE(pm_reset_access) ) > + return false; > + > + return pm_check_access(pm_reset_access, d, rst_idx); > +} > + > +/* > + * Check if a given domain has access to perform an indirect > + * MMIO access. > + * > + * If the provided mask is invalid, it will be fixed up. > + */ > +static bool domain_mediate_mmio_access(struct domain *d, > + bool write, uint64_t addr, > + uint32_t *mask) > +{ > + unsigned int i; > + bool ret = false; > + uint32_t prot_mask = 0; > + > + ASSERT(mask); > + > + /* > + * The hardware domain gets read access to everything. > + * Lower layers will do further filtering. > + */ > + if ( !write && is_hardware_domain(d) ) > + return true; > + /* Scan the ACL. */ > + for ( i = 0; i < ARRAY_SIZE(pm_mmio_access); i++ ) { > + bool r; > + > + /* Check if the address is OK. */ > + if ( pm_mmio_access[i].end ) { > + /* Memory range. */ > + if ( addr < pm_mmio_access[i].start ) > + continue; > + if ( addr > pm_mmio_access[i].end ) > + continue; > + } else { > + /* Single register. */ > + if ( addr != pm_mmio_access[i].start ) > + continue; > + } Would it make sense to order the entries, so that if addr > pm_mmio_access[i].start, we can return immediately? > + if ( write && pm_mmio_access[i].readonly ) > + continue; > + if ( pm_mmio_access[i].hwdom_access && !is_hardware_domain(d) ) > + continue; > + /* Unlimited access is represented by a zero mask. */ > + ASSERT( pm_mmio_access[i].mask != 0xFFFFFFFF ); This is confusing: bits are set in the mask to allow write access, except in the case of 0, which actually means 0xFFFFFFFF. Did I get this right? If so, please choose one of the following: - mask bits are set to allow write access, no exception: instead of 0, we would have 0xFFFFFFFF to give full access - mask bits are set to deny write access: 0 means full access, 0xFFFFFFFF means no access. Then you just need to apply the mask inverted: ~pm_mmio_access[i].mask > + r = domain_has_node_access(d, pm_mmio_access[i].node); > + if ( r ) { > + /* We've got access to this reg (or parts of it). */ > + ret = true; > + /* Masking only applies to writes, so reads can exit here. */ > + if ( !write ) > + break; Instead of break, I would just return ret; > + > + /* Permit write access to selected bits. */ > + prot_mask |= pm_mmio_access[i].mask ? pm_mmio_access[i].mask : ~0; > + printk("match: mask=%x prot_mask=%x\n", > + pm_mmio_access[i].mask, prot_mask); > + if ( prot_mask == 0xFFFFFFFF ) > + break; /* Full access, we're done. */ > + } else { > + if ( !pm_mmio_access[i].mask ) { > + /* > + * The entire reg is tied to a device that we don't have > + * access to. No point in continuing. > + */ > + return false; > + } > + } > + } > + > + /* Masking only applies to writes. */ > + if ( write ) > + *mask &= prot_mask; > + return ret; > +} > + > +bool zynqmp_eemi_mediate(register_t fid, > + register_t a0, > + register_t a1, > + register_t a2, > + register_t a3, > + register_t a4, > + register_t a5, > + register_t *ret) I think you can declare them as uint64_t, given that the API actually requires 64bit registers. Also, I would probably just pass "struct cpu_user_regs *regs" as argument for simplicity. > +{ > + bool is_mmio_write = false; > + unsigned int pm_fn = fid & 0xFFFF; > + uint32_t pm_arg[4]; > + > + /* Decode pm args. */ > + pm_arg[0] = a0; > + pm_arg[1] = a0 >> 32; > + pm_arg[2] = a1; > + pm_arg[3] = a1 >> 32; > + > + ASSERT(ret); > + > + switch (pm_fn) code style > + { > + /* > + * We can't allow CPUs to suspend without Xen knowing about it. > + * We accept but ignore the request and wait for the guest to issue > + * a WFI which Xen will trap and act accordingly upon. > + */ > + case PM_SELF_SUSPEND: > + ret[0] = PM_RET_SUCCESS; > + goto done; This is something in addition of the guest calling WFI, right? In that case we don't actually need to do anything to implement it, correct? > + case PM_GET_NODE_STATUS: > + /* API for PUs. */ > + case PM_REQ_SUSPEND: > + case PM_FORCE_POWERDOWN: > + case PM_ABORT_SUSPEND: > + case PM_REQ_WAKEUP: > + case PM_SET_WAKEUP_SOURCE: > + /* API for slaves. */ > + case PM_REQ_NODE: > + case PM_RELEASE_NODE: > + case PM_SET_REQUIREMENT: > + case PM_SET_MAX_LATENCY: I cannot easily tell what is the permission check for these functions. To which pm_node_access entries do they correspond to? > + if ( !domain_has_node_access(current->domain, pm_arg[0]) ) { code style (bracket on newline) > + printk("zynqmp-pm: fn=%d No access to node %d\n", pm_fn, pm_arg[0]); > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + case PM_RESET_ASSERT: > + case PM_RESET_GET_STATUS: > + if ( !domain_has_reset_access(current->domain, pm_arg[0]) ) { code style > + printk("zynqmp-pm: fn=%d No access to reset %d\n", pm_fn, pm_arg[0]); > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + /* These calls are safe and always allowed. */ > + case ZYNQMP_SIP_SVC_CALL_COUNT: > + case ZYNQMP_SIP_SVC_UID: > + case ZYNQMP_SIP_SVC_VERSION: > + case PM_GET_API_VERSION: > + case PM_GET_CHIPID: > + goto forward_to_fw; > + > + /* Mediated MMIO access. */ > + case PM_MMIO_WRITE: > + is_mmio_write = true; > + /* Fallthrough. */ > + case PM_MMIO_READ: > + if ( !domain_mediate_mmio_access(current->domain, > + is_mmio_write, > + pm_arg[0], &pm_arg[1]) ) { code style > + printk("eemi: fn=%d No access to MMIO %s %x\n", > + pm_fn, is_mmio_write ? "write" : "read", pm_arg[0]); > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + /* Exclusive to Dom0. */ > + case PM_INIT: > + case PM_SET_CONFIGURATION: > + case PM_FPGA_LOAD: > + case PM_FPGA_GET_STATUS: > + if ( !is_hardware_domain(current->domain) ) { Code style. > + printk("eemi: fn=%d No access", pm_fn); > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + /* These calls are never allowed. */ > + case PM_SYSTEM_SHUTDOWN: > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + > + default: > + printk("zynqmp-pm: Unhandled PM Call: %d\n", (u32)fid); > + return false; > + } > + > +forward_to_fw: > + /* Re-encode pm args. */ > + a0 = ((uint64_t)pm_arg[1] << 32) | pm_arg[0]; > + a1 = ((uint64_t)pm_arg[3] << 32) | pm_arg[2]; > + call_smcc64(fid, a0, a1, a2, a3, a4, a5, ret); > +done: > + return true; > +} > + > +/* > + * Local variables: > + * mode: C > + * c-file-style: "BSD" > + * c-basic-offset: 4 > + * indent-tabs-mode: nil > + * End: > + */ > diff --git a/xen/arch/arm/platforms/xilinx-zynqmp.c b/xen/arch/arm/platforms/xilinx-zynqmp.c > index 2adee91..d826282 100644 > --- a/xen/arch/arm/platforms/xilinx-zynqmp.c > +++ b/xen/arch/arm/platforms/xilinx-zynqmp.c > @@ -18,6 +18,7 @@ > */ > > #include <asm/platform.h> > +#include <asm/platforms/xilinx-zynqmp-eemi.h> > > static const char * const zynqmp_dt_compat[] __initconst = > { > @@ -32,8 +33,25 @@ static const struct dt_device_match zynqmp_blacklist_dev[] __initconst = > { /* sentinel */ }, > }; > > +bool zynqmp_hvc(struct cpu_user_regs *regs) > +{ > + register_t ret[4] = { 0 }; > + > + if ( !zynqmp_eemi_mediate(regs->x0, regs->x1, regs->x2, regs->x3, > + regs->x4, regs->x5, regs->x6, ret) ) > + return false; > + > + /* Transfer return values into guest registers. */ > + regs->x0 = ret[0]; > + regs->x1 = ret[1]; > + regs->x2 = ret[2]; > + regs->x3 = ret[3]; > + return true; > +} > + > PLATFORM_START(xgene_storm, "Xilinx ZynqMP") > .compatible = zynqmp_dt_compat, > + .hvc = zynqmp_hvc, > .blacklist_dev = zynqmp_blacklist_dev, > PLATFORM_END > > diff --git a/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h b/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h > new file mode 100644 > index 0000000..774593c > --- /dev/null > +++ b/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h > @@ -0,0 +1,337 @@ > +/* Xen specifics: */ > +/* Main PM Mediator entry. */ > +bool zynqmp_eemi_mediate(register_t fid, > + register_t a0, > + register_t a1, > + register_t a2, > + register_t a3, > + register_t a4, > + register_t a5, > + register_t *ret); > + > +/* Service calls. */ > +#define SVC_MASK 0x3F000000 > +#define SVC_SIP 0x02000000 /* SoC Implementation Specific. */ > +#define PM_SIP_SVC 0xC2000000 > + > +/* SMC function IDs for SiP Service queries */ > +#define ZYNQMP_SIP_SVC_CALL_COUNT 0xff00 > +#define ZYNQMP_SIP_SVC_UID 0xff01 > +#define ZYNQMP_SIP_SVC_VERSION 0xff03 > + > +/* Generic PM defs: */ > +/* > + * Version number is a 32bit value, like: > + * (PM_VERSION_MAJOR << 16) | PM_VERSION_MINOR > + */ > +#define PM_VERSION_MAJOR 0 > +#define PM_VERSION_MINOR 2 > + > +#define PM_VERSION ((PM_VERSION_MAJOR << 16) | PM_VERSION_MINOR) > + > +/* Capabilities for RAM */ > +#define PM_CAP_ACCESS 0x1U > +#define PM_CAP_CONTEXT 0x2U > + > +#define MAX_LATENCY (~0U) > +#define MAX_QOS 100U > + > +/* State arguments of the self suspend */ > +#define PM_STATE_CPU_IDLE 0x0U > +#define PM_STATE_SUSPEND_TO_RAM 0xFU > + > +enum pm_api_id { > + /* Miscellaneous API functions: */ Code style: no tabs please > + PM_GET_API_VERSION = 1, /* Do not change or move */ > + PM_SET_CONFIGURATION, > + PM_GET_NODE_STATUS, > + PM_GET_OP_CHARACTERISTIC, > + PM_REGISTER_NOTIFIER, > + /* API for suspending of PUs: */ > + PM_REQ_SUSPEND, > + PM_SELF_SUSPEND, > + PM_FORCE_POWERDOWN, > + PM_ABORT_SUSPEND, > + PM_REQ_WAKEUP, > + PM_SET_WAKEUP_SOURCE, > + PM_SYSTEM_SHUTDOWN, > + /* API for managing PM slaves: */ > + PM_REQ_NODE, > + PM_RELEASE_NODE, > + PM_SET_REQUIREMENT, > + PM_SET_MAX_LATENCY, > + /* Direct control API functions: */ > + PM_RESET_ASSERT, > + PM_RESET_GET_STATUS, > + PM_MMIO_WRITE, > + PM_MMIO_READ, > + PM_INIT, > + PM_FPGA_LOAD, > + PM_FPGA_GET_STATUS, > + PM_GET_CHIPID, > + PM_API_MAX > +}; > + > +enum pm_node_id { > + NODE_UNKNOWN = 0, > + NODE_APU, > + NODE_APU_0, > + NODE_APU_1, > + NODE_APU_2, > + NODE_APU_3, > + NODE_RPU, > + NODE_RPU_0, > + NODE_RPU_1, > + NODE_PL, > + NODE_FPD, > + NODE_OCM_BANK_0, > + NODE_OCM_BANK_1, > + NODE_OCM_BANK_2, > + NODE_OCM_BANK_3, > + NODE_TCM_0_A, > + NODE_TCM_0_B, > + NODE_TCM_1_A, > + NODE_TCM_1_B, > + NODE_L2, > + NODE_GPU_PP_0, > + NODE_GPU_PP_1, > + NODE_USB_0, > + NODE_USB_1, > + NODE_TTC_0, > + NODE_TTC_1, > + NODE_TTC_2, > + NODE_TTC_3, > + NODE_SATA, > + NODE_ETH_0, > + NODE_ETH_1, > + NODE_ETH_2, > + NODE_ETH_3, > + NODE_UART_0, > + NODE_UART_1, > + NODE_SPI_0, > + NODE_SPI_1, > + NODE_I2C_0, > + NODE_I2C_1, > + NODE_SD_0, > + NODE_SD_1, > + NODE_DP, > + NODE_GDMA, > + NODE_ADMA, > + NODE_NAND, > + NODE_QSPI, > + NODE_GPIO, > + NODE_CAN_0, > + NODE_CAN_1, > + NODE_AFI, > + NODE_APLL, > + NODE_VPLL, > + NODE_DPLL, > + NODE_RPLL, > + NODE_IOPLL, > + NODE_DDR, > + NODE_IPI_APU, > + NODE_IPI_RPU_0, > + NODE_GPU, > + NODE_PCIE, > + NODE_PCAP, > + NODE_RTC, > + NODE_MAX > +}; > + > +enum pm_request_ack { > + REQ_ACK_NO = 1, > + REQ_ACK_BLOCKING, > + REQ_ACK_NON_BLOCKING, > +}; > + > +enum pm_abort_reason { > + ABORT_REASON_WKUP_EVENT = 100, > + ABORT_REASON_PU_BUSY, > + ABORT_REASON_NO_PWRDN, > + ABORT_REASON_UNKNOWN, > +}; > + > +enum pm_suspend_reason { > + SUSPEND_REASON_PU_REQ = 201, > + SUSPEND_REASON_ALERT, > + SUSPEND_REASON_SYS_SHUTDOWN, > +}; > + > +enum pm_ram_state { > + PM_RAM_STATE_OFF = 1, > + PM_RAM_STATE_RETENTION, > + PM_RAM_STATE_ON, > +}; > + > +enum pm_opchar_type { > + PM_OPCHAR_TYPE_POWER = 1, > + PM_OPCHAR_TYPE_TEMP, > + PM_OPCHAR_TYPE_LATENCY, > +}; > + > +/** > + * @PM_RET_SUCCESS: success > + * @PM_RET_ERROR_ARGS: illegal arguments provided > + * @PM_RET_ERROR_ACCESS: access rights violation > + * @PM_RET_ERROR_TIMEOUT: timeout in communication with PMU > + * @PM_RET_ERROR_NOTSUPPORTED: feature not supported > + * @PM_RET_ERROR_PROC: node is not a processor node > + * @PM_RET_ERROR_API_ID: illegal API ID > + * @PM_RET_ERROR_OTHER: other error > + */ > +enum pm_ret_status { > + PM_RET_SUCCESS, > + PM_RET_ERROR_ARGS, > + PM_RET_ERROR_ACCESS, > + PM_RET_ERROR_TIMEOUT, > + PM_RET_ERROR_NOTSUPPORTED, > + PM_RET_ERROR_PROC, > + PM_RET_ERROR_API_ID, > + PM_RET_ERROR_FAILURE, > + PM_RET_ERROR_COMMUNIC, > + PM_RET_ERROR_DOUBLEREQ, > + PM_RET_ERROR_OTHER, > +}; > + > +/** > + * @PM_INITIAL_BOOT: boot is a fresh system startup > + * @PM_RESUME: boot is a resume > + * @PM_BOOT_ERROR: error, boot cause cannot be identified > + */ > +enum pm_boot_status { > + PM_INITIAL_BOOT, > + PM_RESUME, > + PM_BOOT_ERROR, > +}; > + > +enum pm_shutdown_type { > + PMF_SHUTDOWN_TYPE_SHUTDOWN, > + PMF_SHUTDOWN_TYPE_RESET, > +}; > + > +enum pm_shutdown_subtype { > + PMF_SHUTDOWN_SUBTYPE_SUBSYSTEM, > + PMF_SHUTDOWN_SUBTYPE_PS_ONLY, > + PMF_SHUTDOWN_SUBTYPE_SYSTEM, > +}; > + > +enum pm_reset { > + PM_RESET_START = 999, > + PM_RESET_PCIE_CFG, > + PM_RESET_PCIE_BRIDGE, > + PM_RESET_PCIE_CTRL, > + PM_RESET_DP, > + PM_RESET_SWDT_CRF, > + PM_RESET_AFI_FM5, > + PM_RESET_AFI_FM4, > + PM_RESET_AFI_FM3, > + PM_RESET_AFI_FM2, > + PM_RESET_AFI_FM1, > + PM_RESET_AFI_FM0, > + PM_RESET_GDMA, > + PM_RESET_GPU_PP1, > + PM_RESET_GPU_PP0, > + PM_RESET_GPU, > + PM_RESET_GT, > + PM_RESET_SATA, > + PM_RESET_ACPU3_PWRON, > + PM_RESET_ACPU2_PWRON, > + PM_RESET_ACPU1_PWRON, > + PM_RESET_ACPU0_PWRON, > + PM_RESET_APU_L2, > + PM_RESET_ACPU3, > + PM_RESET_ACPU2, > + PM_RESET_ACPU1, > + PM_RESET_ACPU0, > + PM_RESET_DDR, > + PM_RESET_APM_FPD, > + PM_RESET_SOFT, > + PM_RESET_GEM0, > + PM_RESET_GEM1, > + PM_RESET_GEM2, > + PM_RESET_GEM3, > + PM_RESET_QSPI, > + PM_RESET_UART0, > + PM_RESET_UART1, > + PM_RESET_SPI0, > + PM_RESET_SPI1, > + PM_RESET_SDIO0, > + PM_RESET_SDIO1, > + PM_RESET_CAN0, > + PM_RESET_CAN1, > + PM_RESET_I2C0, > + PM_RESET_I2C1, > + PM_RESET_TTC0, > + PM_RESET_TTC1, > + PM_RESET_TTC2, > + PM_RESET_TTC3, > + PM_RESET_SWDT_CRL, > + PM_RESET_NAND, > + PM_RESET_ADMA, > + PM_RESET_GPIO, > + PM_RESET_IOU_CC, > + PM_RESET_TIMESTAMP, > + PM_RESET_RPU_R50, > + PM_RESET_RPU_R51, > + PM_RESET_RPU_AMBA, > + PM_RESET_OCM, > + PM_RESET_RPU_PGE, > + PM_RESET_USB0_CORERESET, > + PM_RESET_USB1_CORERESET, > + PM_RESET_USB0_HIBERRESET, > + PM_RESET_USB1_HIBERRESET, > + PM_RESET_USB0_APB, > + PM_RESET_USB1_APB, > + PM_RESET_IPI, > + PM_RESET_APM_LPD, > + PM_RESET_RTC, > + PM_RESET_SYSMON, > + PM_RESET_AFI_FM6, > + PM_RESET_LPD_SWDT, > + PM_RESET_FPD, > + PM_RESET_RPU_DBG1, > + PM_RESET_RPU_DBG0, > + PM_RESET_DBG_LPD, > + PM_RESET_DBG_FPD, > + PM_RESET_APLL, > + PM_RESET_DPLL, > + PM_RESET_VPLL, > + PM_RESET_IOPLL, > + PM_RESET_RPLL, > + PM_RESET_GPO3_PL_0, > + PM_RESET_GPO3_PL_1, > + PM_RESET_GPO3_PL_2, > + PM_RESET_GPO3_PL_3, > + PM_RESET_GPO3_PL_4, > + PM_RESET_GPO3_PL_5, > + PM_RESET_GPO3_PL_6, > + PM_RESET_GPO3_PL_7, > + PM_RESET_GPO3_PL_8, > + PM_RESET_GPO3_PL_9, > + PM_RESET_GPO3_PL_10, > + PM_RESET_GPO3_PL_11, > + PM_RESET_GPO3_PL_12, > + PM_RESET_GPO3_PL_13, > + PM_RESET_GPO3_PL_14, > + PM_RESET_GPO3_PL_15, > + PM_RESET_GPO3_PL_16, > + PM_RESET_GPO3_PL_17, > + PM_RESET_GPO3_PL_18, > + PM_RESET_GPO3_PL_19, > + PM_RESET_GPO3_PL_20, > + PM_RESET_GPO3_PL_21, > + PM_RESET_GPO3_PL_22, > + PM_RESET_GPO3_PL_23, > + PM_RESET_GPO3_PL_24, > + PM_RESET_GPO3_PL_25, > + PM_RESET_GPO3_PL_26, > + PM_RESET_GPO3_PL_27, > + PM_RESET_GPO3_PL_28, > + PM_RESET_GPO3_PL_29, > + PM_RESET_GPO3_PL_30, > + PM_RESET_GPO3_PL_31, > + PM_RESET_RPU_LS, > + PM_RESET_PS_ONLY, > + PM_RESET_PL, > + PM_RESET_END > +}; > diff --git a/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h b/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h > new file mode 100644 > index 0000000..4fb1695 > --- /dev/null > +++ b/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h > @@ -0,0 +1,284 @@ > +/* > + * Merge of various auto-generated memory map and register > + * definition files. > + */ > + > +/* Selected set of memory-map definitions: */ > +#define MM_IOU_S 0xff000000 > +#define MM_IPI 0xff300000 > +#define MM_LPD_SLCR 0xff410000 > +#define MM_LPD_SLCR_SECURE 0xff4b0000 > +#define MM_CRL_APB 0xff5e0000 > +#define MM_OCM 0xff960000 > +#define MM_LPD_XPPU 0xff980000 > +#define MM_RPU 0xff9a0000 > +#define MM_AFIFM6 0xff9b0000 > +#define MM_LPD_XPPU_SINK 0xff9c0000 > +#define MM_USB3_0 0xff9d0000 > +#define MM_USB3_1 0xff9e0000 > +#define MM_APM_INTC_OCM 0xffa00000 > +#define MM_APM_LPD_FPD 0xffa10000 > +#define MM_AMS_CTRL 0xffa50000 > +#define MM_AMS_PS_SYSMON 0xffa50800 > +#define MM_AMS_PL_SYSMON 0xffa50c00 > +#define MM_RTC 0xffa60000 > +#define MM_OCM_XMPU_CFG 0xffa70000 > +#define MM_ADMA_CH0 0xffa80000 > +#define MM_ADMA_CH1 0xffa90000 > +#define MM_ADMA_CH2 0xffaa0000 > +#define MM_ADMA_CH3 0xffab0000 > +#define MM_ADMA_CH4 0xffac0000 > +#define MM_ADMA_CH5 0xffad0000 > +#define MM_ADMA_CH6 0xffae0000 > + > +#define MM_IOU_GPV 0xfe0fffff > +#define MM_LPD_GPV 0xfe1fffff > +#define MM_USB3_0_XHCI 0xfe2fffff > +#define MM_USB3_1_XHCI 0xfe3fffff > +#define MM_CORESIGHT_SOC_ROM 0xfe80ffff > +#define MM_CORESIGHT_SOC_TSGEN 0xfe90ffff > +#define MM_CORESIGHT_SOC_FUNN_0 0xfe91ffff > +#define MM_CORESIGHT_SOC_FUNN_1 0xfe92ffff > +#define MM_CORESIGHT_SOC_FUNN_2 0xfe93ffff > +#define MM_CORESIGHT_SOC_ETF_1 0xfe94ffff > +#define MM_CORESIGHT_SOC_ETF_2 0xfe95ffff > +#define MM_CORESIGHT_SOC_REPLIC 0xfe96ffff > +#define MM_CORESIGHT_SOC_ETR 0xfe97ffff > +#define MM_CORESIGHT_SOC_TPIU 0xfe98ffff > +#define MM_CORESIGHT_SOC_CTI_0 0xfe99ffff > +#define MM_CORESIGHT_SOC_CTI_1 0xfe9affff > +#define MM_CORESIGHT_SOC_CTI_2 0xfe9bffff > +#define MM_CORESIGHT_SOC_STM 0xfe9cffff > +#define MM_CORESIGHT_SOC_FTM 0xfe9dffff > +#define MM_CORESIGHT_SOC_ATM_0 0xfe9effff > +#define MM_CORESIGHT_SOC_ATM_1 0xfe9fffff > +#define MM_CORESIGHT_R5_ROM 0xfebe0fff > +#define MM_CORESIGHT_R5_DBG_0 0xfebf0fff > +#define MM_CORESIGHT_R5_PMU_0 0xfebf1fff > +#define MM_CORESIGHT_R5_DBG_1 0xfebf2fff > +#define MM_CORESIGHT_R5_PMU_1 0xfebf3fff > +#define MM_CORESIGHT_R5_CTI_0 0xfebf8fff > +#define MM_CORESIGHT_R5_CTI_1 0xfebf9fff > +#define MM_CORESIGHT_R5_ETM_0 0xfebfcfff > +#define MM_CORESIGHT_R5_ETM_1 0xfebfdfff > +#define MM_CORESIGHT_A53_ROM 0xfec0ffff > +#define MM_CORESIGHT_A53_DBG_0 0xfec1ffff > +#define MM_CORESIGHT_A53_CTI_0 0xfec2ffff > +#define MM_CORESIGHT_A53_PMU_0 0xfec3ffff > +#define MM_CORESIGHT_A53_ETM_0 0xfec4ffff > +#define MM_CORESIGHT_A53_DBG_1 0xfed1ffff > +#define MM_CORESIGHT_A53_CTI_1 0xfed2ffff > +#define MM_CORESIGHT_A53_PMU_1 0xfed3ffff > +#define MM_CORESIGHT_A53_ETM_1 0xfed4ffff > +#define MM_CORESIGHT_A53_DBG_2 0xfee1ffff > +#define MM_CORESIGHT_A53_CTI_2 0xfee2ffff > +#define MM_CORESIGHT_A53_PMU_2 0xfee3ffff > +#define MM_CORESIGHT_A53_ETM_2 0xfee4ffff > +#define MM_CORESIGHT_A53_DBG_3 0xfef1ffff > +#define MM_CORESIGHT_A53_CTI_3 0xfef2ffff > +#define MM_CORESIGHT_A53_PMU_3 0xfef3ffff > + > +#define MM_DDRSS 0xfd000000 > +#define MM_SATA_AHCI_HBA 0xfd0c0000 > +#define MM_SATA_AHCI_VENDOR 0xfd0c00a0 > +#define MM_SATA_AHCI_PORT0_CNTRL 0xfd0c0100 > +#define MM_SATA_AHCI_PORT1_CNTRL 0xfd0c0180 > +#define MM_AXIPCIE_MAIN 0xfd0e0000 > +#define MM_AXIPCIE_INGRESS0 0xfd0e0800 > +#define MM_AXIPCIE_INGRESS1 0xfd0e0820 > +#define MM_AXIPCIE_INGRESS2 0xfd0e0840 > +#define MM_AXIPCIE_INGRESS3 0xfd0e0860 > +#define MM_AXIPCIE_INGRESS4 0xfd0e0880 > +#define MM_AXIPCIE_INGRESS5 0xfd0e08a0 > +#define MM_AXIPCIE_INGRESS6 0xfd0e08c0 > +#define MM_AXIPCIE_INGRESS7 0xfd0e08e0 > +#define MM_AXIPCIE_EGRESS0 0xfd0e0c00 > +#define MM_AXIPCIE_EGRESS1 0xfd0e0c20 > +#define MM_AXIPCIE_EGRESS2 0xfd0e0c40 > +#define MM_AXIPCIE_EGRESS3 0xfd0e0c60 > +#define MM_AXIPCIE_EGRESS4 0xfd0e0c80 > +#define MM_AXIPCIE_EGRESS5 0xfd0e0ca0 > +#define MM_AXIPCIE_EGRESS6 0xfd0e0cc0 > +#define MM_AXIPCIE_EGRESS7 0xfd0e0ce0 > +#define MM_AXIPCIE_DMA0 0xfd0f0000 > +#define MM_AXIPCIE_DMA1 0xfd0f0080 > +#define MM_AXIPCIE_DMA2 0xfd0f0100 > +#define MM_AXIPCIE_DMA3 0xfd0f0180 > +#define MM_AXIPCIE_MSIX_TABLE 0xfd0f1000 > +#define MM_AXIPCIE_MSIX_PBA 0xfd0f2000 > +#define MM_CRF_APB 0xfd1a0000 > +#define MM_AFIFM0 0xfd360000 > +#define MM_AFIFM1 0xfd370000 > +#define MM_AFIFM2 0xfd380000 > +#define MM_AFIFM3 0xfd390000 > +#define MM_AFIFM4 0xfd3a0000 > +#define MM_AFIFM5 0xfd3b0000 > +#define MM_SIOU 0xfd3d0000 > +#define MM_SERDES 0xfd400000 > +#define MM_PCIE_ATTRIB 0xfd480000 > +#define MM_APM_CCI_INTC 0xfd490000 > +#define MM_DP 0xfd4a0000 > +#define MM_GPU 0xfd4b0000 > +#define MM_DPDMA 0xfd4c0000 > +#define MM_WDT 0xfd4d0000 > +#define MM_FPD_XMPU_SINK 0xfd4f0000 > +#define MM_GDMA_CH0 0xfd500000 > +#define MM_GDMA_CH1 0xfd510000 > +#define MM_GDMA_CH2 0xfd520000 > +#define MM_GDMA_CH3 0xfd530000 > +#define MM_GDMA_CH4 0xfd540000 > +#define MM_GDMA_CH5 0xfd550000 > +#define MM_GDMA_CH6 0xfd560000 > +#define MM_GDMA_CH7 0xfd570000 > +#define MM_APU 0xfd5c0000 > +#define MM_FPD_XMPU_CFG 0xfd5d0000 > +#define MM_CCI_REG 0xfd5e0000 > +#define MM_SMMU_REG 0xfd5f0000 > +#define MM_FPD_SLCR 0xfd610000 > +#define MM_FPD_SLCR_SECURE 0xfd690000 > +#define MM_CCI_GPV 0xfd6e0000 > +#define MM_FPD_GPV 0xfd700000 > + > +#define MM_QSPI_Linear_Address 0xc0000000 > +#define MM_UART0 0xff000000 > +#define MM_UART1 0xff010000 > +#define MM_I2C0 0xff020000 > +#define MM_I2C1 0xff030000 > +#define MM_SPI0 0xff040000 > +#define MM_SPI1 0xff050000 > +#define MM_CAN0 0xff060000 > +#define MM_CAN1 0xff070000 > +#define MM_GPIO 0xff0a0000 > +#define MM_GEM0 0xff0b0000 > +#define MM_GEM1 0xff0c0000 > +#define MM_GEM2 0xff0d0000 > +#define MM_GEM3 0xff0e0000 > +#define MM_QSPI 0xff0f0000 > +#define MM_NAND 0xff100000 > +#define MM_TTC0 0xff110000 > +#define MM_TTC1 0xff120000 > +#define MM_TTC2 0xff130000 > +#define MM_TTC3 0xff140000 > +#define MM_SWDT 0xff150000 > +#define MM_SD0 0xff160000 > +#define MM_SD1 0xff170000 > +#define MM_IOU_SLCR 0xff180000 > +#define MM_IOU_SECURE_SLCR 0xff240000 > +#define MM_IOU_SCNTR 0xff250000 > + > +#define MM_alg_vcu_enc_top 0xa0000000 > +#define MM_alg_vcu_dec_top 0xa0020000 > + > +#define MM_PMU_GLOBAL 0xffd80000 > +#define MM_CSU 0xffca0000 > + > +/* Selected set of register definitions: */ > +#define R_CRF_APLL_CTRL 0x20 > +#define R_CRF_ACPU_CTRL 0x60 > +#define R_CRF_DP_VIDEO_REF_CTRL 0x70 > +#define R_CRF_DP_STC_REF_CTRL 0x7c > +#define R_CRF_GPU_REF_CTRL 0x84 > +#define R_CRF_SATA_REF_CTRL 0xa0 > +#define R_CRF_PCIE_REF_CTRL 0xb4 > +#define R_CRF_GDMA_REF_CTRL 0xb8 > +#define R_CRF_DPDMA_REF_CTRL 0xbc > + > +#define R_CRL_IOPLL_CTRL 0x20 > +#define R_CRL_IOPLL_CFG 0x24 > +#define R_CRL_IOPLL_FRAC_CFG 0x28 > +#define R_CRL_RPLL_CTRL 0x30 > +#define R_CRL_RPLL_CFG 0x34 > +#define R_CRL_RPLL_FRAC_CFG 0x38 > +#define R_CRL_PLL_STATUS 0x40 > +#define R_CRL_IOPLL_TO_FPD_CTRL 0x44 > +#define R_CRL_RPLL_TO_FPD_CTRL 0x48 > +#define R_CRL_USB3_DUAL_REF_CTRL 0x4c > +#define R_CRL_GEM0_REF_CTRL 0x50 > +#define R_CRL_GEM1_REF_CTRL 0x54 > +#define R_CRL_GEM2_REF_CTRL 0x58 > +#define R_CRL_GEM3_REF_CTRL 0x5c > +#define R_CRL_USB0_BUS_REF_CTRL 0x60 > +#define R_CRL_USB1_BUS_REF_CTRL 0x64 > +#define R_CRL_QSPI_REF_CTRL 0x68 > +#define R_CRL_SDIO0_REF_CTRL 0x6c > +#define R_CRL_SDIO1_REF_CTRL 0x70 > +#define R_CRL_UART0_REF_CTRL 0x74 > +#define R_CRL_UART1_REF_CTRL 0x78 > +#define R_CRL_SPI0_REF_CTRL 0x7c > +#define R_CRL_SPI1_REF_CTRL 0x80 > +#define R_CRL_CAN0_REF_CTRL 0x84 > +#define R_CRL_CAN1_REF_CTRL 0x88 > +#define R_CRL_CPU_R5_CTRL 0x90 > +#define R_CRL_IOU_SWITCH_CTRL 0x9c > +#define R_CRL_CSU_PLL_CTRL 0xa0 > +#define R_CRL_PCAP_CTRL 0xa4 > +#define R_CRL_LPD_SWITCH_CTRL 0xa8 > +#define R_CRL_LPD_LSBUS_CTRL 0xac > +#define R_CRL_DBG_LPD_CTRL 0xb0 > +#define R_CRL_NAND_REF_CTRL 0xb4 > +#define R_CRL_ADMA_REF_CTRL 0xb8 > +#define R_CRL_PL0_REF_CTRL 0xc0 > +#define R_CRL_PL1_REF_CTRL 0xc4 > +#define R_CRL_PL2_REF_CTRL 0xc8 > +#define R_CRL_PL3_REF_CTRL 0xcc > +#define R_CRL_PL0_THR_CTRL 0xd0 > +#define R_CRL_PL0_THR_CNT 0xd4 > +#define R_CRL_PL1_THR_CTRL 0xd8 > +#define R_CRL_PL1_THR_CNT 0xdc > +#define R_CRL_PL2_THR_CTRL 0xe0 > +#define R_CRL_PL2_THR_CNT 0xe4 > +#define R_CRL_PL3_THR_CTRL 0xe8 > +#define R_CRL_PL3_THR_CNT 0xfc > +#define R_CRL_GEM_TSU_REF_CTRL 0x100 > +#define R_CRL_DLL_REF_CTRL 0x104 > +#define R_CRL_AMS_REF_CTRL 0x108 > +#define R_CRL_I2C0_REF_CTRL 0x120 > +#define R_CRL_I2C1_REF_CTRL 0x124 > +#define R_CRL_TIMESTAMP_REF_CTRL 0x128 > + > +#define R_PMU_GLOBAL_GLOBAL_GEN_STORAGE0 0x30 > +#define R_PMU_GLOBAL_PERS_GLOB_GEN_STORAGE7 0x6c > + > +#define R_PMU_GLOBAL_PWR_STATE 0x100 > + > +#define R_CSU_IDCODE 0x40 > + > +#define R_IOU_SLCR_WDT_CLK_SEL 0x300 > +#define R_IOU_SLCR_CAN_MIO_CTRL 0x304 > +#define R_IOU_SLCR_GEM_CLK_CTRL 0x308 > +#define R_IOU_SLCR_SDIO_CLK_CTRL 0x30c > +#define R_IOU_SLCR_CTRL_REG_SD 0x310 > +#define R_IOU_SLCR_SD_ITAPDLY 0x314 > +#define R_IOU_SLCR_SD_OTAPDLYSEL 0x318 > +#define R_IOU_SLCR_SD_CONFIG_REG1 0x31c > +#define R_IOU_SLCR_SD_CONFIG_REG2 0x320 > +#define R_IOU_SLCR_SD_CONFIG_REG3 0x324 > +#define R_IOU_SLCR_SD_INITPRESET 0x328 > +#define R_IOU_SLCR_SD_DSPPRESET 0x32c > +#define R_IOU_SLCR_SD_HSPDPRESET 0x330 > +#define R_IOU_SLCR_SD_SDR12PRESET 0x334 > +#define R_IOU_SLCR_SD_SDR25PRESET 0x338 > +#define R_IOU_SLCR_SD_SDR50PRSET 0x33c > +#define R_IOU_SLCR_SD_SDR104PRST 0x340 > +#define R_IOU_SLCR_SD_DDR50PRESET 0x344 > +#define R_IOU_SLCR_SD_MAXCUR1P8 0x34c > +#define R_IOU_SLCR_SD_MAXCUR3P0 0x350 > +#define R_IOU_SLCR_SD_MAXCUR3P3 0x354 > +#define R_IOU_SLCR_SD_DLL_CTRL 0x358 > +#define R_IOU_SLCR_SD_CDN_CTRL 0x35c > +#define R_IOU_SLCR_GEM_CTRL 0x360 > +#define R_IOU_SLCR_IOU_TTC_APB_CLK 0x380 > +#define R_IOU_SLCR_IOU_TAPDLY_BYPASS 0x390 > +#define R_IOU_SLCR_IOU_COHERENT_CTRL 0x400 > +#define R_IOU_SLCR_VIDEO_PSS_CLK_SEL 0x404 > +#define R_IOU_SLCR_IOU_INTERCONNECT_ROUTE 0x408 > +#define R_IOU_SLCR_IOU_RAM_GEM0 0x500 > +#define R_IOU_SLCR_IOU_RAM_GEM1 0x504 > +#define R_IOU_SLCR_IOU_RAM_GEM2 0x508 > +#define R_IOU_SLCR_IOU_RAM_GEM3 0x50c > +#define R_IOU_SLCR_IOU_RAM_SD0 0x510 > +#define R_IOU_SLCR_IOU_RAM_SD1 0x514 > +#define R_IOU_SLCR_IOU_RAM_CAN0 0x518 > +#define R_IOU_SLCR_IOU_RAM_CAN1 0x51c > +#define R_IOU_SLCR_IOU_RAM_LQSPI 0x520 > +#define R_IOU_SLCR_IOU_RAM_NAND 0x524 > -- > 2.7.4 >
Hi Edgar, On 07/02/17 19:42, Edgar E. Iglesias wrote: > From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> > > Implement an EEMI mediator and forward platform specific > firmware calls from guests to firmware. > > The EEMI mediator is responsible for implementing access > controls modifying or blocking calls that try to operate > on setup for devices that are not under the calling guest's > control. > > EEMI: > https://www.xilinx.com/support/documentation/user_guides/ug1200-eemi-api.pdf > > Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> > --- > xen/arch/arm/platforms/Makefile | 1 + > xen/arch/arm/platforms/xilinx-zynqmp-eemi.c | 794 +++++++++++++++++++++ > xen/arch/arm/platforms/xilinx-zynqmp.c | 18 + > xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h | 337 +++++++++ > xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h | 284 ++++++++ > 5 files changed, 1434 insertions(+) > create mode 100644 xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > create mode 100644 xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h > create mode 100644 xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h > > diff --git a/xen/arch/arm/platforms/Makefile b/xen/arch/arm/platforms/Makefile > index 49fa683..b58b71f 100644 > --- a/xen/arch/arm/platforms/Makefile > +++ b/xen/arch/arm/platforms/Makefile > @@ -8,3 +8,4 @@ obj-$(CONFIG_ARM_64) += seattle.o > obj-$(CONFIG_ARM_32) += sunxi.o > obj-$(CONFIG_ARM_64) += xgene-storm.o > obj-$(CONFIG_ARM_64) += xilinx-zynqmp.o > +obj-$(CONFIG_ARM_64) += xilinx-zynqmp-eemi.o > diff --git a/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c b/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > new file mode 100644 > index 0000000..c2c184d > --- /dev/null > +++ b/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c As you introduce another file for Xilinx, it might be worth considering a new directory xilinx in the platforms. > @@ -0,0 +1,794 @@ > +/* > + * xen/arch/arm/platforms/xilinx-zynqmp-eemi.c > + * > + * Xilinx ZynqMP EEMI API mediator. > + * > + * Copyright (c) 2017 Xilinx Inc. > + * Written by Edgar E. Iglesias <edgar.iglesias@xilinx.com> > + * > + * This program is free software; you can redistribute it and/or modify > + * it under the terms of the GNU General Public License as published by > + * the Free Software Foundation; either version 2 of the License, or > + * (at your option) any later version. We had a bad habit on ARM to use "either version 2 of the License, or (at your option) any later version". However Xen is GPLv2 only (see CONTRIBUTING). Could you update the license to: "This program is free software; you can redistribute it and/or modify it under the terms and conditions of the GNU General Public License, version 2, as published by the Free Software Foundation." > + * > + * This program is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the > + * GNU General Public License for more details. > + */ > + > +/* > + * Mediator for the EEMI Power Mangement API. s/Mangement/Management/ > + * > + * Refs: > + * https://www.xilinx.com/support/documentation/user_guides/ug1200-eemi-api.pdf > + * > + * Background: > + * The ZynqMP has a subsystem named the PMU with a CPU and special devices > + * dedicated to running Power Management Firmware. Other masters in the > + * system need to send requests to the PMU in order to for example: > + * * Manage power state > + * * Configure clocks > + * * Program bitstreams for the programmable logic > + * * etc > + * > + * Allthough the details of the setup are configurable, in the common case s/Allthough/although/ > + * the PMU lives in the Secure world. NS World cannot directly communicate > + * with it and must use proxy services from ARM Trusted Firmware to reach > + * the PMU. > + * > + * Power Management on the ZynqMP is implemented in a layered manner. > + * The PMU knows about various masters and will enforce access controls > + * based on a pre-configured partitioning. This configuration dictates > + * which devices are owned by the various masters and the PMU FW makes sure > + * that a given master cannot turn off a device that it does not own or that > + * is in use by other masters. > + * > + * The PMU is not aware of multiple execution states in masters. > + * For example, it treats the ARMv8 cores as single units and does not > + * distinguish between Secure vs NS OS:s nor does it know about Hypervisors > + * and multiple guests. It is up to software on the ARMv8 cores to present > + * a unified view of its power requirements. > + * > + * To implement this unified view, ARM Trusted Firmware at EL3 provides > + * access to the PM API via SMC calls. ARM Trusted Firmware is responsible > + * for mediating between the Secure and the NS world, rejecting SMC calls > + * that request changes that are not allowed. > + * > + * Xen running above ATF owns the NS world and is responsible for presenting > + * unified PM requests taking all guests and the hypervisor into account. > + * > + * Implementation: > + * The PM API contains different classes of calls. > + * Certain calls are harmless to expose to any guest. > + * These include calls to get the PM API Version, or to read out the version > + * of the chip we're running on. > + * > + * Other calls are more problematic. Here're some: I don't think you can the contraction "here're". > + * * Power requests for nodes (i.e turn on or off a given device) > + * * Reset line control (i.e reset a given device) > + * * MMIO access (i.e directly access clock and reset registers) > + * > + * Power requests for nodes: > + * In order to correctly mediate these calls, we need to know if > + * guests issuing these calls have ownership of the given device. > + * The approach taken here is to map PM API Nodes identifying > + * a device into base addresses for registers that belong to that > + * same device. > + * > + * If the guest has access to a devices registers, we give the guest > + * access to PM API calls that affect that device. This is implemented > + * by pm_node_access and domain_has_node_access(). > + * > + * Reset lines: > + * Reset lines are identified by a list of identifiers that do not relate > + * to device nodes. We can use the same approach here as for nodes, except > + * that we need a separate table mapping reset lines to base addresses. > + * This is implemented by pm_reset_access. > + * > + * MMIO access: > + * These calls allow guests to access certain memory ranges. These ranges > + * are typically protected for secure-world access only and also from > + * certain masters only, so guests cannot access them directly. > + * Registers within the memory regions affect certain nodes. In this case, > + * our imput is an address and we map that address into a node. If the > + * guest has ownership of that node, the access is allowed. > + * Some registers contain bitfields and a single register may contain > + * bits that affect multiple nodes. > + * > + * Some of the calls act on more global state. For example acting on > + * NODE_PS, which affects many a lot of nodes. This higher level > + * orchestrating is left for the hardware-domain only. > + */ > + > +#include <xen/iocap.h> > +#include <asm/platform.h> > +#include <asm/platforms/xilinx-zynqmp-eemi.h> > +#include <asm/platforms/xilinx-zynqmp-mm.h> > + > +struct pm_access > +{ > + uint64_t addr; Please use paddr_t here. > + bool hwdom_access; /* HW domain gets access regardless. */ > +}; [...] > +/* > + * This table maps reset line IDs into a memory address. > + * If a guest has access to the address, it has enough control > + * over the affected node to grant it access to EEMI calls for > + * resetting that node. > + */ > +static const struct { > + uint64_t start; > + uint64_t end; /* Inclusive. If not set, single reg entry. */ For both, s/uint64_t/paddr_t/ > + uint32_t mask; /* Zero means no mask, i.e all bits. */ > + enum pm_node_id node; > + bool hwdom_access; > + bool readonly; > +} pm_mmio_access[] = { Could not we find all those information from the device-tree? > +static bool pm_check_access(const struct pm_access *acl, struct domain *d, int idx) > +{ > + unsigned long mfn; > + > + if ( acl[idx].hwdom_access && is_hardware_domain(d) ) > + return true; > + > + mfn = paddr_to_pfn(acl[idx].addr); acl[idx].addr may not be page aligned, so is there any possibility to have a same page permitted in multiple domain? > + if ( !mfn ) > + return false; > + > + return iomem_access_permitted(d, mfn, mfn); Who will give access to the MMIO region? Is it the toolstack via iomem="..."? Also, giving access to the MMIO also means the guest will be able to map it. After reading the comment at the top of the file, I am not sure this is what you want. > +} > + > +/* Check if a domain has access to a node. */ > +static bool domain_has_node_access(struct domain *d, enum pm_node_id node) > +{ > + if ( node <= 0 || node > ARRAY_SIZE(pm_node_access) ) Some compiler will complain of the first check because the enum is always positive. Also, what are you trying to check given the node is an enum and should fall into the array? > + return false; > + > + /* NODE_UNKNOWN is treated as a wildcard. */ > + if ( node == NODE_UNKNOWN ) > + return true; Which means every domain will access to it. I think this is easily error-prone because NODE_UNKOWN is 0. So you may give access to anyone by mistake. Looking at the pm_mmio_access list, my understanding is NODE_UNKNOWN will be used when only the hardware domain can access it or the region is read-only. So I would use the read-only property to know whether any domain has access. > + > + return pm_check_access(pm_node_access, d, node); > +} > + > +/* Check if a domain has access to a reset line. */ > +static bool domain_has_reset_access(struct domain *d, enum pm_reset rst) > +{ > + int rst_idx = PM_RESET_IDX(rst); > + > + if ( rst_idx < 0 || rst_idx > ARRAY_SIZE(pm_reset_access) ) Same question as above for the bound check. > + return false; > + > + return pm_check_access(pm_reset_access, d, rst_idx); > +} > + > +/* > + * Check if a given domain has access to perform an indirect > + * MMIO access. > + * > + * If the provided mask is invalid, it will be fixed up. > + */ > +static bool domain_mediate_mmio_access(struct domain *d, > + bool write, uint64_t addr, s/uint64_t/paddr_t/ > + uint32_t *mask) > +{ > + unsigned int i; > + bool ret = false; > + uint32_t prot_mask = 0; > + > + ASSERT(mask); > + > + /* > + * The hardware domain gets read access to everything. > + * Lower layers will do further filtering. > + */ > + if ( !write && is_hardware_domain(d) ) > + return true; > + > + /* Scan the ACL. */ > + for ( i = 0; i < ARRAY_SIZE(pm_mmio_access); i++ ) { > + bool r; > + > + /* Check if the address is OK. */ > + if ( pm_mmio_access[i].end ) { > + /* Memory range. */ > + if ( addr < pm_mmio_access[i].start ) > + continue; > + if ( addr > pm_mmio_access[i].end ) > + continue; > + } else { > + /* Single register. */ Are the single register only a byte? If not, in the case of a region is it safe to access in a middle of a register? > + if ( addr != pm_mmio_access[i].start ) > + continue; > + } I would drop the single register case and use make .end = .start. This would avoid potentially mistake in the code. If you are concern about the readability of the array, you could use a macro. > + > + if ( write && pm_mmio_access[i].readonly ) > + continue; > + if ( pm_mmio_access[i].hwdom_access && !is_hardware_domain(d) ) > + continue; > + > + /* Unlimited access is represented by a zero mask. */ > + ASSERT( pm_mmio_access[i].mask != 0xFFFFFFFF ); Why? Would not it make the logic easier to handle unlimited access using 0xFFFFFFFF? You could use macro for that purpose. > + > + r = domain_has_node_access(d, pm_mmio_access[i].node); > + if ( r ) { > + /* We've got access to this reg (or parts of it). */ > + ret = true; > + /* Masking only applies to writes, so reads can exit here. */ If you do that a guest could potentially read some memory it is not allowed. Don't you expect sensitive data in the MMIO? > + if ( !write ) > + break; > + > + /* Permit write access to selected bits. */ > + prot_mask |= pm_mmio_access[i].mask ? pm_mmio_access[i].mask : ~0; > + printk("match: mask=%x prot_mask=%x\n", > + pm_mmio_access[i].mask, prot_mask); > + if ( prot_mask == 0xFFFFFFFF ) > + break; /* Full access, we're done. */ > + } else { > + if ( !pm_mmio_access[i].mask ) { > + /* > + * The entire reg is tied to a device that we don't have > + * access to. No point in continuing. > + */ I am not sure to understand this, above you loop over until you may get prot_mask full. But here you exist as soon as one mask may not be valid. > + return false; > + } > + } > + } > + > + /* Masking only applies to writes. */ > + if ( write ) > + *mask &= prot_mask; Newline here please. > + return ret; > +} > + > +bool zynqmp_eemi_mediate(register_t fid, > + register_t a0, > + register_t a1, > + register_t a2, > + register_t a3, > + register_t a4, > + register_t a5, > + register_t *ret) > +{ > + bool is_mmio_write = false; > + unsigned int pm_fn = fid & 0xFFFF; > + uint32_t pm_arg[4]; > + > + /* Decode pm args. */ > + pm_arg[0] = a0; > + pm_arg[1] = a0 >> 32; > + pm_arg[2] = a1; > + pm_arg[3] = a1 >> 32; > + > + ASSERT(ret); > + > + switch (pm_fn) > + { > + /* > + * We can't allow CPUs to suspend without Xen knowing about it. > + * We accept but ignore the request and wait for the guest to issue > + * a WFI which Xen will trap and act accordingly upon. > + */ > + case PM_SELF_SUSPEND: > + ret[0] = PM_RET_SUCCESS; > + goto done; > + > + case PM_GET_NODE_STATUS: > + /* API for PUs. */ What PUs stand for? > + case PM_REQ_SUSPEND: > + case PM_FORCE_POWERDOWN: > + case PM_ABORT_SUSPEND: > + case PM_REQ_WAKEUP: > + case PM_SET_WAKEUP_SOURCE: > + /* API for slaves. */ > + case PM_REQ_NODE: > + case PM_RELEASE_NODE: > + case PM_SET_REQUIREMENT: > + case PM_SET_MAX_LATENCY: > + if ( !domain_has_node_access(current->domain, pm_arg[0]) ) { Coding style: if ( ... ) { > + printk("zynqmp-pm: fn=%d No access to node %d\n", pm_fn, pm_arg[0]); Printk is not rate-limited, this means a guest could flood Xen with error message if it does not have access to the region. Please use gprintk() here, it will print the domain ID and also rate-limit the log. Also, pm_fn and pm_arg are unsigned, so please use %u. > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + case PM_RESET_ASSERT: > + case PM_RESET_GET_STATUS: > + if ( !domain_has_reset_access(current->domain, pm_arg[0]) ) { Coding style. > + printk("zynqmp-pm: fn=%d No access to reset %d\n", pm_fn, pm_arg[0]); Same as above for the printk. > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + /* These calls are safe and always allowed. */ > + case ZYNQMP_SIP_SVC_CALL_COUNT: > + case ZYNQMP_SIP_SVC_UID: > + case ZYNQMP_SIP_SVC_VERSION: > + case PM_GET_API_VERSION: > + case PM_GET_CHIPID: > + goto forward_to_fw; > + > + /* Mediated MMIO access. */ > + case PM_MMIO_WRITE: > + is_mmio_write = true; > + /* Fallthrough. */ > + case PM_MMIO_READ: > + if ( !domain_mediate_mmio_access(current->domain, > + is_mmio_write, > + pm_arg[0], &pm_arg[1]) ) { See above for the coding style. > + printk("eemi: fn=%d No access to MMIO %s %x\n", > + pm_fn, is_mmio_write ? "write" : "read", pm_arg[0]); See above for the printk. Also, please use %#x if you print hexa to not confuse with decimal. But why do you print pm_arg in hexa here and decimal above? > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + /* Exclusive to Dom0. */ > + case PM_INIT: > + case PM_SET_CONFIGURATION: > + case PM_FPGA_LOAD: > + case PM_FPGA_GET_STATUS: > + if ( !is_hardware_domain(current->domain) ) { See above for the coding style. > + printk("eemi: fn=%d No access", pm_fn); See above for the printk. > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + } > + goto forward_to_fw; > + > + /* These calls are never allowed. */ > + case PM_SYSTEM_SHUTDOWN: > + ret[0] = PM_RET_ERROR_ACCESS; > + goto done; > + > + default: > + printk("zynqmp-pm: Unhandled PM Call: %d\n", (u32)fid); See above for the printk. Also why do you cast fid? > + return false; > + } > + > +forward_to_fw: > + /* Re-encode pm args. */ > + a0 = ((uint64_t)pm_arg[1] << 32) | pm_arg[0]; > + a1 = ((uint64_t)pm_arg[3] << 32) | pm_arg[2]; > + call_smcc64(fid, a0, a1, a2, a3, a4, a5, ret); > +done: > + return true; > +} > + > +/* > + * Local variables: > + * mode: C > + * c-file-style: "BSD" > + * c-basic-offset: 4 > + * indent-tabs-mode: nil > + * End: > + */ > diff --git a/xen/arch/arm/platforms/xilinx-zynqmp.c b/xen/arch/arm/platforms/xilinx-zynqmp.c > index 2adee91..d826282 100644 > --- a/xen/arch/arm/platforms/xilinx-zynqmp.c > +++ b/xen/arch/arm/platforms/xilinx-zynqmp.c > @@ -18,6 +18,7 @@ > */ > > #include <asm/platform.h> > +#include <asm/platforms/xilinx-zynqmp-eemi.h> > > static const char * const zynqmp_dt_compat[] __initconst = > { > @@ -32,8 +33,25 @@ static const struct dt_device_match zynqmp_blacklist_dev[] __initconst = > { /* sentinel */ }, > }; > > +bool zynqmp_hvc(struct cpu_user_regs *regs) This function should be static. > +{ > + register_t ret[4] = { 0 }; > + > + if ( !zynqmp_eemi_mediate(regs->x0, regs->x1, regs->x2, regs->x3, > + regs->x4, regs->x5, regs->x6, ret) ) > + return false; > + > + /* Transfer return values into guest registers. */ > + regs->x0 = ret[0]; > + regs->x1 = ret[1]; > + regs->x2 = ret[2]; > + regs->x3 = ret[3]; Newline here please. > + return true; > +} > + > PLATFORM_START(xgene_storm, "Xilinx ZynqMP") > .compatible = zynqmp_dt_compat, > + .hvc = zynqmp_hvc, > .blacklist_dev = zynqmp_blacklist_dev, > PLATFORM_END > > diff --git a/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h b/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h > new file mode 100644 > index 0000000..774593c > --- /dev/null > +++ b/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h [...] > +/** > + * @PM_RET_SUCCESS: success > + * @PM_RET_ERROR_ARGS: illegal arguments provided > + * @PM_RET_ERROR_ACCESS: access rights violation > + * @PM_RET_ERROR_TIMEOUT: timeout in communication with PMU > + * @PM_RET_ERROR_NOTSUPPORTED: feature not supported > + * @PM_RET_ERROR_PROC: node is not a processor node > + * @PM_RET_ERROR_API_ID: illegal API ID > + * @PM_RET_ERROR_OTHER: other error > + */ > +enum pm_ret_status { > + PM_RET_SUCCESS, I looked at the spec you provide in the commit message and I cannot find those name. Is it related to EEMI_SUCCESS? If so, I would prefer if we use the name from the spec. Furthermore, it does not seem that values are provided in the spec. > + PM_RET_ERROR_ARGS, > + PM_RET_ERROR_ACCESS, > + PM_RET_ERROR_TIMEOUT, > + PM_RET_ERROR_NOTSUPPORTED, > + PM_RET_ERROR_PROC, > + PM_RET_ERROR_API_ID, > + PM_RET_ERROR_FAILURE, > + PM_RET_ERROR_COMMUNIC, > + PM_RET_ERROR_DOUBLEREQ, > + PM_RET_ERROR_OTHER, > +}; [...] > diff --git a/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h b/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h > new file mode 100644 > index 0000000..4fb1695 > --- /dev/null > +++ b/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h I feel a bit unsure to see that many hardcoded values in Xen. Is it the only way to retrieve the mmio/power information? Regards,
diff --git a/xen/arch/arm/platforms/Makefile b/xen/arch/arm/platforms/Makefile index 49fa683..b58b71f 100644 --- a/xen/arch/arm/platforms/Makefile +++ b/xen/arch/arm/platforms/Makefile @@ -8,3 +8,4 @@ obj-$(CONFIG_ARM_64) += seattle.o obj-$(CONFIG_ARM_32) += sunxi.o obj-$(CONFIG_ARM_64) += xgene-storm.o obj-$(CONFIG_ARM_64) += xilinx-zynqmp.o +obj-$(CONFIG_ARM_64) += xilinx-zynqmp-eemi.o diff --git a/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c b/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c new file mode 100644 index 0000000..c2c184d --- /dev/null +++ b/xen/arch/arm/platforms/xilinx-zynqmp-eemi.c @@ -0,0 +1,794 @@ +/* + * xen/arch/arm/platforms/xilinx-zynqmp-eemi.c + * + * Xilinx ZynqMP EEMI API mediator. + * + * Copyright (c) 2017 Xilinx Inc. + * Written by Edgar E. Iglesias <edgar.iglesias@xilinx.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +/* + * Mediator for the EEMI Power Mangement API. + * + * Refs: + * https://www.xilinx.com/support/documentation/user_guides/ug1200-eemi-api.pdf + * + * Background: + * The ZynqMP has a subsystem named the PMU with a CPU and special devices + * dedicated to running Power Management Firmware. Other masters in the + * system need to send requests to the PMU in order to for example: + * * Manage power state + * * Configure clocks + * * Program bitstreams for the programmable logic + * * etc + * + * Allthough the details of the setup are configurable, in the common case + * the PMU lives in the Secure world. NS World cannot directly communicate + * with it and must use proxy services from ARM Trusted Firmware to reach + * the PMU. + * + * Power Management on the ZynqMP is implemented in a layered manner. + * The PMU knows about various masters and will enforce access controls + * based on a pre-configured partitioning. This configuration dictates + * which devices are owned by the various masters and the PMU FW makes sure + * that a given master cannot turn off a device that it does not own or that + * is in use by other masters. + * + * The PMU is not aware of multiple execution states in masters. + * For example, it treats the ARMv8 cores as single units and does not + * distinguish between Secure vs NS OS:s nor does it know about Hypervisors + * and multiple guests. It is up to software on the ARMv8 cores to present + * a unified view of its power requirements. + * + * To implement this unified view, ARM Trusted Firmware at EL3 provides + * access to the PM API via SMC calls. ARM Trusted Firmware is responsible + * for mediating between the Secure and the NS world, rejecting SMC calls + * that request changes that are not allowed. + * + * Xen running above ATF owns the NS world and is responsible for presenting + * unified PM requests taking all guests and the hypervisor into account. + * + * Implementation: + * The PM API contains different classes of calls. + * Certain calls are harmless to expose to any guest. + * These include calls to get the PM API Version, or to read out the version + * of the chip we're running on. + * + * Other calls are more problematic. Here're some: + * * Power requests for nodes (i.e turn on or off a given device) + * * Reset line control (i.e reset a given device) + * * MMIO access (i.e directly access clock and reset registers) + * + * Power requests for nodes: + * In order to correctly mediate these calls, we need to know if + * guests issuing these calls have ownership of the given device. + * The approach taken here is to map PM API Nodes identifying + * a device into base addresses for registers that belong to that + * same device. + * + * If the guest has access to a devices registers, we give the guest + * access to PM API calls that affect that device. This is implemented + * by pm_node_access and domain_has_node_access(). + * + * Reset lines: + * Reset lines are identified by a list of identifiers that do not relate + * to device nodes. We can use the same approach here as for nodes, except + * that we need a separate table mapping reset lines to base addresses. + * This is implemented by pm_reset_access. + * + * MMIO access: + * These calls allow guests to access certain memory ranges. These ranges + * are typically protected for secure-world access only and also from + * certain masters only, so guests cannot access them directly. + * Registers within the memory regions affect certain nodes. In this case, + * our imput is an address and we map that address into a node. If the + * guest has ownership of that node, the access is allowed. + * Some registers contain bitfields and a single register may contain + * bits that affect multiple nodes. + * + * Some of the calls act on more global state. For example acting on + * NODE_PS, which affects many a lot of nodes. This higher level + * orchestrating is left for the hardware-domain only. + */ + +#include <xen/iocap.h> +#include <asm/platform.h> +#include <asm/platforms/xilinx-zynqmp-eemi.h> +#include <asm/platforms/xilinx-zynqmp-mm.h> + +struct pm_access +{ + uint64_t addr; + bool hwdom_access; /* HW domain gets access regardless. */ +}; + +/* + * This table maps a node into a memory address. + * If a guest has access to the address, it has enough control + * over the node to grant it access to EEMI calls for that node. + */ +static const struct pm_access pm_node_access[] = { + /* MM_RPU grants access to alll RPU Nodes. */ + [NODE_RPU] = { MM_RPU }, + [NODE_RPU_0] = { MM_RPU }, + [NODE_RPU_1] = { MM_RPU }, + [NODE_IPI_RPU_0] = { MM_RPU }, + + /* GPU nodes. */ + [NODE_GPU] = { MM_GPU }, + [NODE_GPU_PP_0] = { MM_GPU }, + [NODE_GPU_PP_1] = { MM_GPU }, + + [NODE_USB_0] = { MM_USB3_0 }, + [NODE_USB_1] = { MM_USB3_1 }, + [NODE_TTC_0] = { MM_TTC0 }, + [NODE_TTC_1] = { MM_TTC1 }, + [NODE_TTC_2] = { MM_TTC2 }, + [NODE_TTC_3] = { MM_TTC3 }, + [NODE_SATA] = { MM_SATA_AHCI_HBA }, + [NODE_ETH_0] = { MM_GEM0 }, + [NODE_ETH_1] = { MM_GEM1 }, + [NODE_ETH_2] = { MM_GEM2 }, + [NODE_ETH_3] = { MM_GEM3 }, + [NODE_UART_0] = { MM_UART0 }, + [NODE_UART_1] = { MM_UART1 }, + [NODE_SPI_0] = { MM_SPI0 }, + [NODE_SPI_1] = { MM_SPI1 }, + [NODE_I2C_0] = { MM_I2C0 }, + [NODE_I2C_1] = { MM_I2C1 }, + [NODE_SD_0] = { MM_SD0 }, + [NODE_SD_1] = { MM_SD1 }, + [NODE_DP] = { MM_DP }, + + /* Guest with GDMA Channel 0 gets PM access. Other guests don't. */ + [NODE_GDMA] = { MM_GDMA_CH0 }, + /* Guest with ADMA Channel 0 gets PM access. Other guests don't. */ + [NODE_ADMA] = { MM_ADMA_CH0 }, + + [NODE_NAND] = { MM_NAND }, + [NODE_QSPI] = { MM_QSPI }, + [NODE_GPIO] = { MM_GPIO }, + [NODE_CAN_0] = { MM_CAN0 }, + [NODE_CAN_1] = { MM_CAN1 }, + + /* Dom0 only. */ + [NODE_AFI] = { .hwdom_access = true }, + [NODE_APLL] = { .hwdom_access = true }, + [NODE_VPLL] = { .hwdom_access = true }, + [NODE_DPLL] = { .hwdom_access = true }, + [NODE_RPLL] = { .hwdom_access = true }, + [NODE_IOPLL] = { .hwdom_access = true }, + [NODE_DDR] = { .hwdom_access = true }, + [NODE_IPI_APU] = { .hwdom_access = true }, + [NODE_PCAP] = { .hwdom_access = true }, + + [NODE_PCIE] = { MM_PCIE_ATTRIB }, + [NODE_RTC] = { MM_RTC }, +}; + +/* + * This table maps reset line IDs into a memory address. + * If a guest has access to the address, it has enough control + * over the affected node to grant it access to EEMI calls for + * resetting that node. + */ +#define PM_RESET_IDX(n) (n - PM_RESET_PCIE_CFG) +static const struct pm_access pm_reset_access[] = { + [PM_RESET_IDX(PM_RESET_PCIE_CFG)] = { MM_AXIPCIE_MAIN }, + [PM_RESET_IDX(PM_RESET_PCIE_BRIDGE)] = { MM_PCIE_ATTRIB }, + [PM_RESET_IDX(PM_RESET_PCIE_CTRL)] = { MM_PCIE_ATTRIB }, + + [PM_RESET_IDX(PM_RESET_DP)] = { MM_DP }, + [PM_RESET_IDX(PM_RESET_SWDT_CRF)] = { MM_SWDT }, + [PM_RESET_IDX(PM_RESET_AFI_FM5)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_AFI_FM4)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_AFI_FM3)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_AFI_FM2)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_AFI_FM1)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_AFI_FM0)] = { .hwdom_access = true }, + + /* Channel 0 grants PM access. */ + [PM_RESET_IDX(PM_RESET_GDMA)] = { MM_GDMA_CH0 }, + [PM_RESET_IDX(PM_RESET_GPU_PP1)] = { MM_GPU }, + [PM_RESET_IDX(PM_RESET_GPU_PP0)] = { MM_GPU }, + [PM_RESET_IDX(PM_RESET_GT)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_SATA)] = { MM_SATA_AHCI_HBA }, + + /* We don't allow anyone to turn on/off the ACPUs. */ + [PM_RESET_IDX(PM_RESET_ACPU3_PWRON)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU2_PWRON)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU1_PWRON)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU0_PWRON)] = { 0 }, + [PM_RESET_IDX(PM_RESET_APU_L2)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU3)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU2)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU1)] = { 0 }, + [PM_RESET_IDX(PM_RESET_ACPU0)] = { 0 }, + + [PM_RESET_IDX(PM_RESET_DDR)] = { 0 }, + + [PM_RESET_IDX(PM_RESET_APM_FPD)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_SOFT)] = { .hwdom_access = true }, + + [PM_RESET_IDX(PM_RESET_GEM0)] = { MM_GEM0 }, + [PM_RESET_IDX(PM_RESET_GEM1)] = { MM_GEM1 }, + [PM_RESET_IDX(PM_RESET_GEM2)] = { MM_GEM2 }, + [PM_RESET_IDX(PM_RESET_GEM3)] = { MM_GEM3 }, + + [PM_RESET_IDX(PM_RESET_QSPI)] = { MM_QSPI }, + [PM_RESET_IDX(PM_RESET_UART0)] = { MM_UART0 }, + [PM_RESET_IDX(PM_RESET_UART1)] = { MM_UART1 }, + [PM_RESET_IDX(PM_RESET_SPI0)] = { MM_SPI0 }, + [PM_RESET_IDX(PM_RESET_SPI1)] = { MM_SPI1 }, + [PM_RESET_IDX(PM_RESET_SDIO0)] = { MM_SD0 }, + [PM_RESET_IDX(PM_RESET_SDIO1)] = { MM_SD1 }, + [PM_RESET_IDX(PM_RESET_CAN0)] = { MM_CAN0 }, + [PM_RESET_IDX(PM_RESET_CAN1)] = { MM_CAN1 }, + [PM_RESET_IDX(PM_RESET_I2C0)] = { MM_I2C0 }, + [PM_RESET_IDX(PM_RESET_I2C1)] = { MM_I2C1 }, + [PM_RESET_IDX(PM_RESET_TTC0)] = { MM_TTC0 }, + [PM_RESET_IDX(PM_RESET_TTC1)] = { MM_TTC1 }, + [PM_RESET_IDX(PM_RESET_TTC2)] = { MM_TTC2 }, + [PM_RESET_IDX(PM_RESET_TTC3)] = { MM_TTC3 }, + [PM_RESET_IDX(PM_RESET_SWDT_CRL)] = { MM_SWDT }, + [PM_RESET_IDX(PM_RESET_NAND)] = { MM_NAND }, + /* ADMA Channel 0 grants access to pull the reset signal. */ + [PM_RESET_IDX(PM_RESET_ADMA)] = { MM_ADMA_CH0 }, + [PM_RESET_IDX(PM_RESET_GPIO)] = { MM_GPIO }, + /* FIXME: What is this? */ + [PM_RESET_IDX(PM_RESET_IOU_CC)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_TIMESTAMP)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_RPU_R50)] = { MM_RPU }, + [PM_RESET_IDX(PM_RESET_RPU_R51)] = { MM_RPU }, + [PM_RESET_IDX(PM_RESET_RPU_AMBA)] = { MM_RPU }, + [PM_RESET_IDX(PM_RESET_OCM)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_RPU_PGE)] = { MM_RPU }, + + [PM_RESET_IDX(PM_RESET_USB0_CORERESET)] = { MM_USB3_0 }, + [PM_RESET_IDX(PM_RESET_USB0_HIBERRESET)] = { MM_USB3_0 }, + [PM_RESET_IDX(PM_RESET_USB0_APB)] = { MM_USB3_0 }, + + [PM_RESET_IDX(PM_RESET_USB1_CORERESET)] = { MM_USB3_1 }, + [PM_RESET_IDX(PM_RESET_USB1_HIBERRESET)] = { MM_USB3_1 }, + [PM_RESET_IDX(PM_RESET_USB1_APB)] = { MM_USB3_1 }, + + [PM_RESET_IDX(PM_RESET_IPI)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_APM_LPD)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_RTC)] = { MM_RTC }, + [PM_RESET_IDX(PM_RESET_SYSMON)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_AFI_FM6)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_LPD_SWDT)] = { MM_SWDT }, + [PM_RESET_IDX(PM_RESET_FPD)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_RPU_DBG1)] = { MM_RPU }, + [PM_RESET_IDX(PM_RESET_RPU_DBG0)] = { MM_RPU }, + [PM_RESET_IDX(PM_RESET_DBG_LPD)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_DBG_FPD)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_APLL)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_DPLL)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_VPLL)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_IOPLL)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_RPLL)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_0)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_1)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_2)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_3)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_4)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_5)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_6)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_7)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_8)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_9)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_10)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_11)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_12)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_13)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_14)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_15)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_16)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_17)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_18)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_19)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_20)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_21)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_22)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_23)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_24)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_25)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_26)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_27)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_28)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_29)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_30)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_GPO3_PL_31)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_RPU_LS)] = { MM_RPU }, + [PM_RESET_IDX(PM_RESET_PS_ONLY)] = { .hwdom_access = true }, + [PM_RESET_IDX(PM_RESET_PL)] = { .hwdom_access = true }, +}; + +/* + * This table maps reset line IDs into a memory address. + * If a guest has access to the address, it has enough control + * over the affected node to grant it access to EEMI calls for + * resetting that node. + */ +static const struct { + uint64_t start; + uint64_t end; /* Inclusive. If not set, single reg entry. */ + uint32_t mask; /* Zero means no mask, i.e all bits. */ + enum pm_node_id node; + bool hwdom_access; + bool readonly; +} pm_mmio_access[] = { + { + .start = MM_CRF_APB + R_CRF_APLL_CTRL, + .end = MM_CRF_APB + R_CRF_ACPU_CTRL, + .hwdom_access = true + }, + { + .start = MM_CRL_APB + R_CRL_IOPLL_CTRL, + .end = MM_CRL_APB + R_CRL_RPLL_TO_FPD_CTRL, + .hwdom_access = true + }, + { .start = MM_CRF_APB + R_CRF_DP_VIDEO_REF_CTRL, .node = NODE_DP }, + { .start = MM_CRF_APB + R_CRF_DP_STC_REF_CTRL, .node = NODE_DP }, + { .start = MM_CRF_APB + R_CRF_GPU_REF_CTRL, .node = NODE_GPU }, + { .start = MM_CRF_APB + R_CRF_SATA_REF_CTRL, .node = NODE_SATA }, + { .start = MM_CRF_APB + R_CRF_PCIE_REF_CTRL, .node = NODE_PCIE }, + { .start = MM_CRF_APB + R_CRF_GDMA_REF_CTRL, .node = NODE_GDMA }, + { .start = MM_CRF_APB + R_CRF_DPDMA_REF_CTRL, .node = NODE_DP }, + { .start = MM_CRL_APB + R_CRL_USB3_DUAL_REF_CTRL, .node = NODE_USB_0 }, + { .start = MM_CRL_APB + R_CRL_USB0_BUS_REF_CTRL, .node = NODE_USB_0 }, + { .start = MM_CRL_APB + R_CRL_USB1_BUS_REF_CTRL, .node = NODE_USB_1 }, + { .start = MM_CRL_APB + R_CRL_USB1_BUS_REF_CTRL, .node = NODE_USB_1 }, + { .start = MM_CRL_APB + R_CRL_GEM0_REF_CTRL, .node = NODE_ETH_0 }, + { .start = MM_CRL_APB + R_CRL_GEM1_REF_CTRL, .node = NODE_ETH_1 }, + { .start = MM_CRL_APB + R_CRL_GEM2_REF_CTRL, .node = NODE_ETH_2 }, + { .start = MM_CRL_APB + R_CRL_GEM3_REF_CTRL, .node = NODE_ETH_3 }, + { .start = MM_CRL_APB + R_CRL_QSPI_REF_CTRL, .node = NODE_QSPI }, + { .start = MM_CRL_APB + R_CRL_SDIO0_REF_CTRL, .node = NODE_SD_0 }, + { .start = MM_CRL_APB + R_CRL_SDIO1_REF_CTRL, .node = NODE_SD_1 }, + { .start = MM_CRL_APB + R_CRL_UART0_REF_CTRL, .node = NODE_UART_0 }, + { .start = MM_CRL_APB + R_CRL_UART1_REF_CTRL, .node = NODE_UART_1 }, + { .start = MM_CRL_APB + R_CRL_SPI0_REF_CTRL, .node = NODE_SPI_0 }, + { .start = MM_CRL_APB + R_CRL_SPI1_REF_CTRL, .node = NODE_SPI_1 }, + { .start = MM_CRL_APB + R_CRL_CAN0_REF_CTRL, .node = NODE_CAN_0 }, + { .start = MM_CRL_APB + R_CRL_CAN1_REF_CTRL, .node = NODE_CAN_1 }, + { .start = MM_CRL_APB + R_CRL_CPU_R5_CTRL, .node = NODE_RPU }, + { .start = MM_CRL_APB + R_CRL_IOU_SWITCH_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_CSU_PLL_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PCAP_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_LPD_SWITCH_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_LPD_LSBUS_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_DBG_LPD_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_NAND_REF_CTRL, .node = NODE_NAND }, + { .start = MM_CRL_APB + R_CRL_ADMA_REF_CTRL, .node = NODE_ADMA }, + { .start = MM_CRL_APB + R_CRL_PL0_REF_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL1_REF_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL2_REF_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL3_REF_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL0_THR_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL1_THR_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL2_THR_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL3_THR_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL0_THR_CNT, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL1_THR_CNT, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL2_THR_CNT, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_PL3_THR_CNT, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_GEM_TSU_REF_CTRL, .node = NODE_ETH_0 }, + { .start = MM_CRL_APB + R_CRL_DLL_REF_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_AMS_REF_CTRL, .hwdom_access = true }, + { .start = MM_CRL_APB + R_CRL_I2C0_REF_CTRL, .node = NODE_I2C_0 }, + { .start = MM_CRL_APB + R_CRL_I2C1_REF_CTRL, .node = NODE_I2C_1 }, + { .start = MM_CRL_APB + R_CRL_TIMESTAMP_REF_CTRL, .hwdom_access = true }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_WDT_CLK_SEL, .hwdom_access = true }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_CAN_MIO_CTRL, + .mask = 0x1ff, .node = NODE_CAN_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_CAN_MIO_CTRL, + .mask = 0x1ff << 15, .node = NODE_CAN_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, + .mask = 0xf, .node = NODE_ETH_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, + .mask = 0xf << 5, .node = NODE_ETH_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, + .mask = 0xf << 10, .node = NODE_ETH_2 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, + .mask = 0xf << 15, .node = NODE_ETH_3 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CLK_CTRL, + .mask = 0x7 << 20, .hwdom_access = true + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_SDIO_CLK_CTRL, + .mask = 0x7, .node = NODE_SD_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_SDIO_CLK_CTRL, + .mask = 0x7 << 17, .node = NODE_SD_1 + }, + + { + .start = MM_IOU_SLCR + R_IOU_SLCR_CTRL_REG_SD, + .mask = 0x1, .node = NODE_SD_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_CTRL_REG_SD, + .mask = 0x1 << 15, .node = NODE_SD_1 + }, + /* A series of SD regs with the same layout. */ + { + .start = MM_IOU_SLCR + R_IOU_SLCR_SD_ITAPDLY, + .end = MM_IOU_SLCR + R_IOU_SLCR_SD_CDN_CTRL, + .mask = 0x3ff << 0, .node = NODE_SD_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_SD_ITAPDLY, + .end = MM_IOU_SLCR + R_IOU_SLCR_SD_CDN_CTRL, + .mask = 0x3ff << 16, .node = NODE_SD_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, + .mask = 0x3 << 0, .node = NODE_ETH_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, + .mask = 0x3 << 2, .node = NODE_ETH_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, + .mask = 0x3 << 4, .node = NODE_ETH_2 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_GEM_CTRL, + .mask = 0x3 << 6, .node = NODE_ETH_3 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, + .mask = 0x3 << 0, .node = NODE_TTC_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, + .mask = 0x3 << 2, .node = NODE_TTC_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, + .mask = 0x3 << 4, .node = NODE_TTC_2 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TTC_APB_CLK, + .mask = 0x3 << 6, .node = NODE_TTC_3 + }, + + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TAPDLY_BYPASS, + .mask = 0x3, .node = NODE_NAND + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_TAPDLY_BYPASS, + .mask = 0x1 << 2, .node = NODE_QSPI + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 0, .node = NODE_ETH_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 4, .node = NODE_ETH_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 8, .node = NODE_ETH_2 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 12, .node = NODE_ETH_3 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 16, .node = NODE_SD_0 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 20, .node = NODE_SD_1 + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 24, .node = NODE_NAND + }, + { + .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_COHERENT_CTRL, + .mask = 0xf << 28, .node = NODE_QSPI + }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_VIDEO_PSS_CLK_SEL, .hwdom_access = true }, + /* No access to R_IOU_SLCR_IOU_INTERCONNECT_ROUTE at all. */ + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM0, .node = NODE_ETH_0 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM1, .node = NODE_ETH_1 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM2, .node = NODE_ETH_2 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_GEM3, .node = NODE_ETH_3 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_SD0, .node = NODE_SD_0 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_SD1, .node = NODE_SD_1 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_CAN0, .node = NODE_CAN_0 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_CAN1, .node = NODE_CAN_1 }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_LQSPI, .node = NODE_QSPI }, + { .start = MM_IOU_SLCR + R_IOU_SLCR_IOU_RAM_NAND, .node = NODE_NAND }, + { + .start = MM_PMU_GLOBAL + R_PMU_GLOBAL_PWR_STATE, + .readonly = true, + }, + { + .start = MM_PMU_GLOBAL + R_PMU_GLOBAL_GLOBAL_GEN_STORAGE0, + .end = MM_PMU_GLOBAL + R_PMU_GLOBAL_PERS_GLOB_GEN_STORAGE7, + .readonly = true, + }, + { + /* Universal read-only access to CRF. Linux CCF needs this. */ + .start = MM_CRF_APB, .end = MM_CRF_APB + 0x104, + .readonly = true, + }, + { + /* Universal read-only access to CRL. Linux CCF needs this. */ + .start = MM_CRL_APB, .end = MM_CRL_APB + 0x284, + .readonly = true, + } +}; + +static bool pm_check_access(const struct pm_access *acl, struct domain *d, int idx) +{ + unsigned long mfn; + + if ( acl[idx].hwdom_access && is_hardware_domain(d) ) + return true; + + mfn = paddr_to_pfn(acl[idx].addr); + if ( !mfn ) + return false; + + return iomem_access_permitted(d, mfn, mfn); +} + +/* Check if a domain has access to a node. */ +static bool domain_has_node_access(struct domain *d, enum pm_node_id node) +{ + if ( node <= 0 || node > ARRAY_SIZE(pm_node_access) ) + return false; + + /* NODE_UNKNOWN is treated as a wildcard. */ + if ( node == NODE_UNKNOWN ) + return true; + + return pm_check_access(pm_node_access, d, node); +} + +/* Check if a domain has access to a reset line. */ +static bool domain_has_reset_access(struct domain *d, enum pm_reset rst) +{ + int rst_idx = PM_RESET_IDX(rst); + + if ( rst_idx < 0 || rst_idx > ARRAY_SIZE(pm_reset_access) ) + return false; + + return pm_check_access(pm_reset_access, d, rst_idx); +} + +/* + * Check if a given domain has access to perform an indirect + * MMIO access. + * + * If the provided mask is invalid, it will be fixed up. + */ +static bool domain_mediate_mmio_access(struct domain *d, + bool write, uint64_t addr, + uint32_t *mask) +{ + unsigned int i; + bool ret = false; + uint32_t prot_mask = 0; + + ASSERT(mask); + + /* + * The hardware domain gets read access to everything. + * Lower layers will do further filtering. + */ + if ( !write && is_hardware_domain(d) ) + return true; + + /* Scan the ACL. */ + for ( i = 0; i < ARRAY_SIZE(pm_mmio_access); i++ ) { + bool r; + + /* Check if the address is OK. */ + if ( pm_mmio_access[i].end ) { + /* Memory range. */ + if ( addr < pm_mmio_access[i].start ) + continue; + if ( addr > pm_mmio_access[i].end ) + continue; + } else { + /* Single register. */ + if ( addr != pm_mmio_access[i].start ) + continue; + } + + if ( write && pm_mmio_access[i].readonly ) + continue; + if ( pm_mmio_access[i].hwdom_access && !is_hardware_domain(d) ) + continue; + + /* Unlimited access is represented by a zero mask. */ + ASSERT( pm_mmio_access[i].mask != 0xFFFFFFFF ); + + r = domain_has_node_access(d, pm_mmio_access[i].node); + if ( r ) { + /* We've got access to this reg (or parts of it). */ + ret = true; + /* Masking only applies to writes, so reads can exit here. */ + if ( !write ) + break; + + /* Permit write access to selected bits. */ + prot_mask |= pm_mmio_access[i].mask ? pm_mmio_access[i].mask : ~0; + printk("match: mask=%x prot_mask=%x\n", + pm_mmio_access[i].mask, prot_mask); + if ( prot_mask == 0xFFFFFFFF ) + break; /* Full access, we're done. */ + } else { + if ( !pm_mmio_access[i].mask ) { + /* + * The entire reg is tied to a device that we don't have + * access to. No point in continuing. + */ + return false; + } + } + } + + /* Masking only applies to writes. */ + if ( write ) + *mask &= prot_mask; + return ret; +} + +bool zynqmp_eemi_mediate(register_t fid, + register_t a0, + register_t a1, + register_t a2, + register_t a3, + register_t a4, + register_t a5, + register_t *ret) +{ + bool is_mmio_write = false; + unsigned int pm_fn = fid & 0xFFFF; + uint32_t pm_arg[4]; + + /* Decode pm args. */ + pm_arg[0] = a0; + pm_arg[1] = a0 >> 32; + pm_arg[2] = a1; + pm_arg[3] = a1 >> 32; + + ASSERT(ret); + + switch (pm_fn) + { + /* + * We can't allow CPUs to suspend without Xen knowing about it. + * We accept but ignore the request and wait for the guest to issue + * a WFI which Xen will trap and act accordingly upon. + */ + case PM_SELF_SUSPEND: + ret[0] = PM_RET_SUCCESS; + goto done; + + case PM_GET_NODE_STATUS: + /* API for PUs. */ + case PM_REQ_SUSPEND: + case PM_FORCE_POWERDOWN: + case PM_ABORT_SUSPEND: + case PM_REQ_WAKEUP: + case PM_SET_WAKEUP_SOURCE: + /* API for slaves. */ + case PM_REQ_NODE: + case PM_RELEASE_NODE: + case PM_SET_REQUIREMENT: + case PM_SET_MAX_LATENCY: + if ( !domain_has_node_access(current->domain, pm_arg[0]) ) { + printk("zynqmp-pm: fn=%d No access to node %d\n", pm_fn, pm_arg[0]); + ret[0] = PM_RET_ERROR_ACCESS; + goto done; + } + goto forward_to_fw; + + case PM_RESET_ASSERT: + case PM_RESET_GET_STATUS: + if ( !domain_has_reset_access(current->domain, pm_arg[0]) ) { + printk("zynqmp-pm: fn=%d No access to reset %d\n", pm_fn, pm_arg[0]); + ret[0] = PM_RET_ERROR_ACCESS; + goto done; + } + goto forward_to_fw; + + /* These calls are safe and always allowed. */ + case ZYNQMP_SIP_SVC_CALL_COUNT: + case ZYNQMP_SIP_SVC_UID: + case ZYNQMP_SIP_SVC_VERSION: + case PM_GET_API_VERSION: + case PM_GET_CHIPID: + goto forward_to_fw; + + /* Mediated MMIO access. */ + case PM_MMIO_WRITE: + is_mmio_write = true; + /* Fallthrough. */ + case PM_MMIO_READ: + if ( !domain_mediate_mmio_access(current->domain, + is_mmio_write, + pm_arg[0], &pm_arg[1]) ) { + printk("eemi: fn=%d No access to MMIO %s %x\n", + pm_fn, is_mmio_write ? "write" : "read", pm_arg[0]); + ret[0] = PM_RET_ERROR_ACCESS; + goto done; + } + goto forward_to_fw; + + /* Exclusive to Dom0. */ + case PM_INIT: + case PM_SET_CONFIGURATION: + case PM_FPGA_LOAD: + case PM_FPGA_GET_STATUS: + if ( !is_hardware_domain(current->domain) ) { + printk("eemi: fn=%d No access", pm_fn); + ret[0] = PM_RET_ERROR_ACCESS; + goto done; + } + goto forward_to_fw; + + /* These calls are never allowed. */ + case PM_SYSTEM_SHUTDOWN: + ret[0] = PM_RET_ERROR_ACCESS; + goto done; + + default: + printk("zynqmp-pm: Unhandled PM Call: %d\n", (u32)fid); + return false; + } + +forward_to_fw: + /* Re-encode pm args. */ + a0 = ((uint64_t)pm_arg[1] << 32) | pm_arg[0]; + a1 = ((uint64_t)pm_arg[3] << 32) | pm_arg[2]; + call_smcc64(fid, a0, a1, a2, a3, a4, a5, ret); +done: + return true; +} + +/* + * Local variables: + * mode: C + * c-file-style: "BSD" + * c-basic-offset: 4 + * indent-tabs-mode: nil + * End: + */ diff --git a/xen/arch/arm/platforms/xilinx-zynqmp.c b/xen/arch/arm/platforms/xilinx-zynqmp.c index 2adee91..d826282 100644 --- a/xen/arch/arm/platforms/xilinx-zynqmp.c +++ b/xen/arch/arm/platforms/xilinx-zynqmp.c @@ -18,6 +18,7 @@ */ #include <asm/platform.h> +#include <asm/platforms/xilinx-zynqmp-eemi.h> static const char * const zynqmp_dt_compat[] __initconst = { @@ -32,8 +33,25 @@ static const struct dt_device_match zynqmp_blacklist_dev[] __initconst = { /* sentinel */ }, }; +bool zynqmp_hvc(struct cpu_user_regs *regs) +{ + register_t ret[4] = { 0 }; + + if ( !zynqmp_eemi_mediate(regs->x0, regs->x1, regs->x2, regs->x3, + regs->x4, regs->x5, regs->x6, ret) ) + return false; + + /* Transfer return values into guest registers. */ + regs->x0 = ret[0]; + regs->x1 = ret[1]; + regs->x2 = ret[2]; + regs->x3 = ret[3]; + return true; +} + PLATFORM_START(xgene_storm, "Xilinx ZynqMP") .compatible = zynqmp_dt_compat, + .hvc = zynqmp_hvc, .blacklist_dev = zynqmp_blacklist_dev, PLATFORM_END diff --git a/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h b/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h new file mode 100644 index 0000000..774593c --- /dev/null +++ b/xen/include/asm-arm/platforms/xilinx-zynqmp-eemi.h @@ -0,0 +1,337 @@ +/* Xen specifics: */ +/* Main PM Mediator entry. */ +bool zynqmp_eemi_mediate(register_t fid, + register_t a0, + register_t a1, + register_t a2, + register_t a3, + register_t a4, + register_t a5, + register_t *ret); + +/* Service calls. */ +#define SVC_MASK 0x3F000000 +#define SVC_SIP 0x02000000 /* SoC Implementation Specific. */ +#define PM_SIP_SVC 0xC2000000 + +/* SMC function IDs for SiP Service queries */ +#define ZYNQMP_SIP_SVC_CALL_COUNT 0xff00 +#define ZYNQMP_SIP_SVC_UID 0xff01 +#define ZYNQMP_SIP_SVC_VERSION 0xff03 + +/* Generic PM defs: */ +/* + * Version number is a 32bit value, like: + * (PM_VERSION_MAJOR << 16) | PM_VERSION_MINOR + */ +#define PM_VERSION_MAJOR 0 +#define PM_VERSION_MINOR 2 + +#define PM_VERSION ((PM_VERSION_MAJOR << 16) | PM_VERSION_MINOR) + +/* Capabilities for RAM */ +#define PM_CAP_ACCESS 0x1U +#define PM_CAP_CONTEXT 0x2U + +#define MAX_LATENCY (~0U) +#define MAX_QOS 100U + +/* State arguments of the self suspend */ +#define PM_STATE_CPU_IDLE 0x0U +#define PM_STATE_SUSPEND_TO_RAM 0xFU + +enum pm_api_id { + /* Miscellaneous API functions: */ + PM_GET_API_VERSION = 1, /* Do not change or move */ + PM_SET_CONFIGURATION, + PM_GET_NODE_STATUS, + PM_GET_OP_CHARACTERISTIC, + PM_REGISTER_NOTIFIER, + /* API for suspending of PUs: */ + PM_REQ_SUSPEND, + PM_SELF_SUSPEND, + PM_FORCE_POWERDOWN, + PM_ABORT_SUSPEND, + PM_REQ_WAKEUP, + PM_SET_WAKEUP_SOURCE, + PM_SYSTEM_SHUTDOWN, + /* API for managing PM slaves: */ + PM_REQ_NODE, + PM_RELEASE_NODE, + PM_SET_REQUIREMENT, + PM_SET_MAX_LATENCY, + /* Direct control API functions: */ + PM_RESET_ASSERT, + PM_RESET_GET_STATUS, + PM_MMIO_WRITE, + PM_MMIO_READ, + PM_INIT, + PM_FPGA_LOAD, + PM_FPGA_GET_STATUS, + PM_GET_CHIPID, + PM_API_MAX +}; + +enum pm_node_id { + NODE_UNKNOWN = 0, + NODE_APU, + NODE_APU_0, + NODE_APU_1, + NODE_APU_2, + NODE_APU_3, + NODE_RPU, + NODE_RPU_0, + NODE_RPU_1, + NODE_PL, + NODE_FPD, + NODE_OCM_BANK_0, + NODE_OCM_BANK_1, + NODE_OCM_BANK_2, + NODE_OCM_BANK_3, + NODE_TCM_0_A, + NODE_TCM_0_B, + NODE_TCM_1_A, + NODE_TCM_1_B, + NODE_L2, + NODE_GPU_PP_0, + NODE_GPU_PP_1, + NODE_USB_0, + NODE_USB_1, + NODE_TTC_0, + NODE_TTC_1, + NODE_TTC_2, + NODE_TTC_3, + NODE_SATA, + NODE_ETH_0, + NODE_ETH_1, + NODE_ETH_2, + NODE_ETH_3, + NODE_UART_0, + NODE_UART_1, + NODE_SPI_0, + NODE_SPI_1, + NODE_I2C_0, + NODE_I2C_1, + NODE_SD_0, + NODE_SD_1, + NODE_DP, + NODE_GDMA, + NODE_ADMA, + NODE_NAND, + NODE_QSPI, + NODE_GPIO, + NODE_CAN_0, + NODE_CAN_1, + NODE_AFI, + NODE_APLL, + NODE_VPLL, + NODE_DPLL, + NODE_RPLL, + NODE_IOPLL, + NODE_DDR, + NODE_IPI_APU, + NODE_IPI_RPU_0, + NODE_GPU, + NODE_PCIE, + NODE_PCAP, + NODE_RTC, + NODE_MAX +}; + +enum pm_request_ack { + REQ_ACK_NO = 1, + REQ_ACK_BLOCKING, + REQ_ACK_NON_BLOCKING, +}; + +enum pm_abort_reason { + ABORT_REASON_WKUP_EVENT = 100, + ABORT_REASON_PU_BUSY, + ABORT_REASON_NO_PWRDN, + ABORT_REASON_UNKNOWN, +}; + +enum pm_suspend_reason { + SUSPEND_REASON_PU_REQ = 201, + SUSPEND_REASON_ALERT, + SUSPEND_REASON_SYS_SHUTDOWN, +}; + +enum pm_ram_state { + PM_RAM_STATE_OFF = 1, + PM_RAM_STATE_RETENTION, + PM_RAM_STATE_ON, +}; + +enum pm_opchar_type { + PM_OPCHAR_TYPE_POWER = 1, + PM_OPCHAR_TYPE_TEMP, + PM_OPCHAR_TYPE_LATENCY, +}; + +/** + * @PM_RET_SUCCESS: success + * @PM_RET_ERROR_ARGS: illegal arguments provided + * @PM_RET_ERROR_ACCESS: access rights violation + * @PM_RET_ERROR_TIMEOUT: timeout in communication with PMU + * @PM_RET_ERROR_NOTSUPPORTED: feature not supported + * @PM_RET_ERROR_PROC: node is not a processor node + * @PM_RET_ERROR_API_ID: illegal API ID + * @PM_RET_ERROR_OTHER: other error + */ +enum pm_ret_status { + PM_RET_SUCCESS, + PM_RET_ERROR_ARGS, + PM_RET_ERROR_ACCESS, + PM_RET_ERROR_TIMEOUT, + PM_RET_ERROR_NOTSUPPORTED, + PM_RET_ERROR_PROC, + PM_RET_ERROR_API_ID, + PM_RET_ERROR_FAILURE, + PM_RET_ERROR_COMMUNIC, + PM_RET_ERROR_DOUBLEREQ, + PM_RET_ERROR_OTHER, +}; + +/** + * @PM_INITIAL_BOOT: boot is a fresh system startup + * @PM_RESUME: boot is a resume + * @PM_BOOT_ERROR: error, boot cause cannot be identified + */ +enum pm_boot_status { + PM_INITIAL_BOOT, + PM_RESUME, + PM_BOOT_ERROR, +}; + +enum pm_shutdown_type { + PMF_SHUTDOWN_TYPE_SHUTDOWN, + PMF_SHUTDOWN_TYPE_RESET, +}; + +enum pm_shutdown_subtype { + PMF_SHUTDOWN_SUBTYPE_SUBSYSTEM, + PMF_SHUTDOWN_SUBTYPE_PS_ONLY, + PMF_SHUTDOWN_SUBTYPE_SYSTEM, +}; + +enum pm_reset { + PM_RESET_START = 999, + PM_RESET_PCIE_CFG, + PM_RESET_PCIE_BRIDGE, + PM_RESET_PCIE_CTRL, + PM_RESET_DP, + PM_RESET_SWDT_CRF, + PM_RESET_AFI_FM5, + PM_RESET_AFI_FM4, + PM_RESET_AFI_FM3, + PM_RESET_AFI_FM2, + PM_RESET_AFI_FM1, + PM_RESET_AFI_FM0, + PM_RESET_GDMA, + PM_RESET_GPU_PP1, + PM_RESET_GPU_PP0, + PM_RESET_GPU, + PM_RESET_GT, + PM_RESET_SATA, + PM_RESET_ACPU3_PWRON, + PM_RESET_ACPU2_PWRON, + PM_RESET_ACPU1_PWRON, + PM_RESET_ACPU0_PWRON, + PM_RESET_APU_L2, + PM_RESET_ACPU3, + PM_RESET_ACPU2, + PM_RESET_ACPU1, + PM_RESET_ACPU0, + PM_RESET_DDR, + PM_RESET_APM_FPD, + PM_RESET_SOFT, + PM_RESET_GEM0, + PM_RESET_GEM1, + PM_RESET_GEM2, + PM_RESET_GEM3, + PM_RESET_QSPI, + PM_RESET_UART0, + PM_RESET_UART1, + PM_RESET_SPI0, + PM_RESET_SPI1, + PM_RESET_SDIO0, + PM_RESET_SDIO1, + PM_RESET_CAN0, + PM_RESET_CAN1, + PM_RESET_I2C0, + PM_RESET_I2C1, + PM_RESET_TTC0, + PM_RESET_TTC1, + PM_RESET_TTC2, + PM_RESET_TTC3, + PM_RESET_SWDT_CRL, + PM_RESET_NAND, + PM_RESET_ADMA, + PM_RESET_GPIO, + PM_RESET_IOU_CC, + PM_RESET_TIMESTAMP, + PM_RESET_RPU_R50, + PM_RESET_RPU_R51, + PM_RESET_RPU_AMBA, + PM_RESET_OCM, + PM_RESET_RPU_PGE, + PM_RESET_USB0_CORERESET, + PM_RESET_USB1_CORERESET, + PM_RESET_USB0_HIBERRESET, + PM_RESET_USB1_HIBERRESET, + PM_RESET_USB0_APB, + PM_RESET_USB1_APB, + PM_RESET_IPI, + PM_RESET_APM_LPD, + PM_RESET_RTC, + PM_RESET_SYSMON, + PM_RESET_AFI_FM6, + PM_RESET_LPD_SWDT, + PM_RESET_FPD, + PM_RESET_RPU_DBG1, + PM_RESET_RPU_DBG0, + PM_RESET_DBG_LPD, + PM_RESET_DBG_FPD, + PM_RESET_APLL, + PM_RESET_DPLL, + PM_RESET_VPLL, + PM_RESET_IOPLL, + PM_RESET_RPLL, + PM_RESET_GPO3_PL_0, + PM_RESET_GPO3_PL_1, + PM_RESET_GPO3_PL_2, + PM_RESET_GPO3_PL_3, + PM_RESET_GPO3_PL_4, + PM_RESET_GPO3_PL_5, + PM_RESET_GPO3_PL_6, + PM_RESET_GPO3_PL_7, + PM_RESET_GPO3_PL_8, + PM_RESET_GPO3_PL_9, + PM_RESET_GPO3_PL_10, + PM_RESET_GPO3_PL_11, + PM_RESET_GPO3_PL_12, + PM_RESET_GPO3_PL_13, + PM_RESET_GPO3_PL_14, + PM_RESET_GPO3_PL_15, + PM_RESET_GPO3_PL_16, + PM_RESET_GPO3_PL_17, + PM_RESET_GPO3_PL_18, + PM_RESET_GPO3_PL_19, + PM_RESET_GPO3_PL_20, + PM_RESET_GPO3_PL_21, + PM_RESET_GPO3_PL_22, + PM_RESET_GPO3_PL_23, + PM_RESET_GPO3_PL_24, + PM_RESET_GPO3_PL_25, + PM_RESET_GPO3_PL_26, + PM_RESET_GPO3_PL_27, + PM_RESET_GPO3_PL_28, + PM_RESET_GPO3_PL_29, + PM_RESET_GPO3_PL_30, + PM_RESET_GPO3_PL_31, + PM_RESET_RPU_LS, + PM_RESET_PS_ONLY, + PM_RESET_PL, + PM_RESET_END +}; diff --git a/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h b/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h new file mode 100644 index 0000000..4fb1695 --- /dev/null +++ b/xen/include/asm-arm/platforms/xilinx-zynqmp-mm.h @@ -0,0 +1,284 @@ +/* + * Merge of various auto-generated memory map and register + * definition files. + */ + +/* Selected set of memory-map definitions: */ +#define MM_IOU_S 0xff000000 +#define MM_IPI 0xff300000 +#define MM_LPD_SLCR 0xff410000 +#define MM_LPD_SLCR_SECURE 0xff4b0000 +#define MM_CRL_APB 0xff5e0000 +#define MM_OCM 0xff960000 +#define MM_LPD_XPPU 0xff980000 +#define MM_RPU 0xff9a0000 +#define MM_AFIFM6 0xff9b0000 +#define MM_LPD_XPPU_SINK 0xff9c0000 +#define MM_USB3_0 0xff9d0000 +#define MM_USB3_1 0xff9e0000 +#define MM_APM_INTC_OCM 0xffa00000 +#define MM_APM_LPD_FPD 0xffa10000 +#define MM_AMS_CTRL 0xffa50000 +#define MM_AMS_PS_SYSMON 0xffa50800 +#define MM_AMS_PL_SYSMON 0xffa50c00 +#define MM_RTC 0xffa60000 +#define MM_OCM_XMPU_CFG 0xffa70000 +#define MM_ADMA_CH0 0xffa80000 +#define MM_ADMA_CH1 0xffa90000 +#define MM_ADMA_CH2 0xffaa0000 +#define MM_ADMA_CH3 0xffab0000 +#define MM_ADMA_CH4 0xffac0000 +#define MM_ADMA_CH5 0xffad0000 +#define MM_ADMA_CH6 0xffae0000 + +#define MM_IOU_GPV 0xfe0fffff +#define MM_LPD_GPV 0xfe1fffff +#define MM_USB3_0_XHCI 0xfe2fffff +#define MM_USB3_1_XHCI 0xfe3fffff +#define MM_CORESIGHT_SOC_ROM 0xfe80ffff +#define MM_CORESIGHT_SOC_TSGEN 0xfe90ffff +#define MM_CORESIGHT_SOC_FUNN_0 0xfe91ffff +#define MM_CORESIGHT_SOC_FUNN_1 0xfe92ffff +#define MM_CORESIGHT_SOC_FUNN_2 0xfe93ffff +#define MM_CORESIGHT_SOC_ETF_1 0xfe94ffff +#define MM_CORESIGHT_SOC_ETF_2 0xfe95ffff +#define MM_CORESIGHT_SOC_REPLIC 0xfe96ffff +#define MM_CORESIGHT_SOC_ETR 0xfe97ffff +#define MM_CORESIGHT_SOC_TPIU 0xfe98ffff +#define MM_CORESIGHT_SOC_CTI_0 0xfe99ffff +#define MM_CORESIGHT_SOC_CTI_1 0xfe9affff +#define MM_CORESIGHT_SOC_CTI_2 0xfe9bffff +#define MM_CORESIGHT_SOC_STM 0xfe9cffff +#define MM_CORESIGHT_SOC_FTM 0xfe9dffff +#define MM_CORESIGHT_SOC_ATM_0 0xfe9effff +#define MM_CORESIGHT_SOC_ATM_1 0xfe9fffff +#define MM_CORESIGHT_R5_ROM 0xfebe0fff +#define MM_CORESIGHT_R5_DBG_0 0xfebf0fff +#define MM_CORESIGHT_R5_PMU_0 0xfebf1fff +#define MM_CORESIGHT_R5_DBG_1 0xfebf2fff +#define MM_CORESIGHT_R5_PMU_1 0xfebf3fff +#define MM_CORESIGHT_R5_CTI_0 0xfebf8fff +#define MM_CORESIGHT_R5_CTI_1 0xfebf9fff +#define MM_CORESIGHT_R5_ETM_0 0xfebfcfff +#define MM_CORESIGHT_R5_ETM_1 0xfebfdfff +#define MM_CORESIGHT_A53_ROM 0xfec0ffff +#define MM_CORESIGHT_A53_DBG_0 0xfec1ffff +#define MM_CORESIGHT_A53_CTI_0 0xfec2ffff +#define MM_CORESIGHT_A53_PMU_0 0xfec3ffff +#define MM_CORESIGHT_A53_ETM_0 0xfec4ffff +#define MM_CORESIGHT_A53_DBG_1 0xfed1ffff +#define MM_CORESIGHT_A53_CTI_1 0xfed2ffff +#define MM_CORESIGHT_A53_PMU_1 0xfed3ffff +#define MM_CORESIGHT_A53_ETM_1 0xfed4ffff +#define MM_CORESIGHT_A53_DBG_2 0xfee1ffff +#define MM_CORESIGHT_A53_CTI_2 0xfee2ffff +#define MM_CORESIGHT_A53_PMU_2 0xfee3ffff +#define MM_CORESIGHT_A53_ETM_2 0xfee4ffff +#define MM_CORESIGHT_A53_DBG_3 0xfef1ffff +#define MM_CORESIGHT_A53_CTI_3 0xfef2ffff +#define MM_CORESIGHT_A53_PMU_3 0xfef3ffff + +#define MM_DDRSS 0xfd000000 +#define MM_SATA_AHCI_HBA 0xfd0c0000 +#define MM_SATA_AHCI_VENDOR 0xfd0c00a0 +#define MM_SATA_AHCI_PORT0_CNTRL 0xfd0c0100 +#define MM_SATA_AHCI_PORT1_CNTRL 0xfd0c0180 +#define MM_AXIPCIE_MAIN 0xfd0e0000 +#define MM_AXIPCIE_INGRESS0 0xfd0e0800 +#define MM_AXIPCIE_INGRESS1 0xfd0e0820 +#define MM_AXIPCIE_INGRESS2 0xfd0e0840 +#define MM_AXIPCIE_INGRESS3 0xfd0e0860 +#define MM_AXIPCIE_INGRESS4 0xfd0e0880 +#define MM_AXIPCIE_INGRESS5 0xfd0e08a0 +#define MM_AXIPCIE_INGRESS6 0xfd0e08c0 +#define MM_AXIPCIE_INGRESS7 0xfd0e08e0 +#define MM_AXIPCIE_EGRESS0 0xfd0e0c00 +#define MM_AXIPCIE_EGRESS1 0xfd0e0c20 +#define MM_AXIPCIE_EGRESS2 0xfd0e0c40 +#define MM_AXIPCIE_EGRESS3 0xfd0e0c60 +#define MM_AXIPCIE_EGRESS4 0xfd0e0c80 +#define MM_AXIPCIE_EGRESS5 0xfd0e0ca0 +#define MM_AXIPCIE_EGRESS6 0xfd0e0cc0 +#define MM_AXIPCIE_EGRESS7 0xfd0e0ce0 +#define MM_AXIPCIE_DMA0 0xfd0f0000 +#define MM_AXIPCIE_DMA1 0xfd0f0080 +#define MM_AXIPCIE_DMA2 0xfd0f0100 +#define MM_AXIPCIE_DMA3 0xfd0f0180 +#define MM_AXIPCIE_MSIX_TABLE 0xfd0f1000 +#define MM_AXIPCIE_MSIX_PBA 0xfd0f2000 +#define MM_CRF_APB 0xfd1a0000 +#define MM_AFIFM0 0xfd360000 +#define MM_AFIFM1 0xfd370000 +#define MM_AFIFM2 0xfd380000 +#define MM_AFIFM3 0xfd390000 +#define MM_AFIFM4 0xfd3a0000 +#define MM_AFIFM5 0xfd3b0000 +#define MM_SIOU 0xfd3d0000 +#define MM_SERDES 0xfd400000 +#define MM_PCIE_ATTRIB 0xfd480000 +#define MM_APM_CCI_INTC 0xfd490000 +#define MM_DP 0xfd4a0000 +#define MM_GPU 0xfd4b0000 +#define MM_DPDMA 0xfd4c0000 +#define MM_WDT 0xfd4d0000 +#define MM_FPD_XMPU_SINK 0xfd4f0000 +#define MM_GDMA_CH0 0xfd500000 +#define MM_GDMA_CH1 0xfd510000 +#define MM_GDMA_CH2 0xfd520000 +#define MM_GDMA_CH3 0xfd530000 +#define MM_GDMA_CH4 0xfd540000 +#define MM_GDMA_CH5 0xfd550000 +#define MM_GDMA_CH6 0xfd560000 +#define MM_GDMA_CH7 0xfd570000 +#define MM_APU 0xfd5c0000 +#define MM_FPD_XMPU_CFG 0xfd5d0000 +#define MM_CCI_REG 0xfd5e0000 +#define MM_SMMU_REG 0xfd5f0000 +#define MM_FPD_SLCR 0xfd610000 +#define MM_FPD_SLCR_SECURE 0xfd690000 +#define MM_CCI_GPV 0xfd6e0000 +#define MM_FPD_GPV 0xfd700000 + +#define MM_QSPI_Linear_Address 0xc0000000 +#define MM_UART0 0xff000000 +#define MM_UART1 0xff010000 +#define MM_I2C0 0xff020000 +#define MM_I2C1 0xff030000 +#define MM_SPI0 0xff040000 +#define MM_SPI1 0xff050000 +#define MM_CAN0 0xff060000 +#define MM_CAN1 0xff070000 +#define MM_GPIO 0xff0a0000 +#define MM_GEM0 0xff0b0000 +#define MM_GEM1 0xff0c0000 +#define MM_GEM2 0xff0d0000 +#define MM_GEM3 0xff0e0000 +#define MM_QSPI 0xff0f0000 +#define MM_NAND 0xff100000 +#define MM_TTC0 0xff110000 +#define MM_TTC1 0xff120000 +#define MM_TTC2 0xff130000 +#define MM_TTC3 0xff140000 +#define MM_SWDT 0xff150000 +#define MM_SD0 0xff160000 +#define MM_SD1 0xff170000 +#define MM_IOU_SLCR 0xff180000 +#define MM_IOU_SECURE_SLCR 0xff240000 +#define MM_IOU_SCNTR 0xff250000 + +#define MM_alg_vcu_enc_top 0xa0000000 +#define MM_alg_vcu_dec_top 0xa0020000 + +#define MM_PMU_GLOBAL 0xffd80000 +#define MM_CSU 0xffca0000 + +/* Selected set of register definitions: */ +#define R_CRF_APLL_CTRL 0x20 +#define R_CRF_ACPU_CTRL 0x60 +#define R_CRF_DP_VIDEO_REF_CTRL 0x70 +#define R_CRF_DP_STC_REF_CTRL 0x7c +#define R_CRF_GPU_REF_CTRL 0x84 +#define R_CRF_SATA_REF_CTRL 0xa0 +#define R_CRF_PCIE_REF_CTRL 0xb4 +#define R_CRF_GDMA_REF_CTRL 0xb8 +#define R_CRF_DPDMA_REF_CTRL 0xbc + +#define R_CRL_IOPLL_CTRL 0x20 +#define R_CRL_IOPLL_CFG 0x24 +#define R_CRL_IOPLL_FRAC_CFG 0x28 +#define R_CRL_RPLL_CTRL 0x30 +#define R_CRL_RPLL_CFG 0x34 +#define R_CRL_RPLL_FRAC_CFG 0x38 +#define R_CRL_PLL_STATUS 0x40 +#define R_CRL_IOPLL_TO_FPD_CTRL 0x44 +#define R_CRL_RPLL_TO_FPD_CTRL 0x48 +#define R_CRL_USB3_DUAL_REF_CTRL 0x4c +#define R_CRL_GEM0_REF_CTRL 0x50 +#define R_CRL_GEM1_REF_CTRL 0x54 +#define R_CRL_GEM2_REF_CTRL 0x58 +#define R_CRL_GEM3_REF_CTRL 0x5c +#define R_CRL_USB0_BUS_REF_CTRL 0x60 +#define R_CRL_USB1_BUS_REF_CTRL 0x64 +#define R_CRL_QSPI_REF_CTRL 0x68 +#define R_CRL_SDIO0_REF_CTRL 0x6c +#define R_CRL_SDIO1_REF_CTRL 0x70 +#define R_CRL_UART0_REF_CTRL 0x74 +#define R_CRL_UART1_REF_CTRL 0x78 +#define R_CRL_SPI0_REF_CTRL 0x7c +#define R_CRL_SPI1_REF_CTRL 0x80 +#define R_CRL_CAN0_REF_CTRL 0x84 +#define R_CRL_CAN1_REF_CTRL 0x88 +#define R_CRL_CPU_R5_CTRL 0x90 +#define R_CRL_IOU_SWITCH_CTRL 0x9c +#define R_CRL_CSU_PLL_CTRL 0xa0 +#define R_CRL_PCAP_CTRL 0xa4 +#define R_CRL_LPD_SWITCH_CTRL 0xa8 +#define R_CRL_LPD_LSBUS_CTRL 0xac +#define R_CRL_DBG_LPD_CTRL 0xb0 +#define R_CRL_NAND_REF_CTRL 0xb4 +#define R_CRL_ADMA_REF_CTRL 0xb8 +#define R_CRL_PL0_REF_CTRL 0xc0 +#define R_CRL_PL1_REF_CTRL 0xc4 +#define R_CRL_PL2_REF_CTRL 0xc8 +#define R_CRL_PL3_REF_CTRL 0xcc +#define R_CRL_PL0_THR_CTRL 0xd0 +#define R_CRL_PL0_THR_CNT 0xd4 +#define R_CRL_PL1_THR_CTRL 0xd8 +#define R_CRL_PL1_THR_CNT 0xdc +#define R_CRL_PL2_THR_CTRL 0xe0 +#define R_CRL_PL2_THR_CNT 0xe4 +#define R_CRL_PL3_THR_CTRL 0xe8 +#define R_CRL_PL3_THR_CNT 0xfc +#define R_CRL_GEM_TSU_REF_CTRL 0x100 +#define R_CRL_DLL_REF_CTRL 0x104 +#define R_CRL_AMS_REF_CTRL 0x108 +#define R_CRL_I2C0_REF_CTRL 0x120 +#define R_CRL_I2C1_REF_CTRL 0x124 +#define R_CRL_TIMESTAMP_REF_CTRL 0x128 + +#define R_PMU_GLOBAL_GLOBAL_GEN_STORAGE0 0x30 +#define R_PMU_GLOBAL_PERS_GLOB_GEN_STORAGE7 0x6c + +#define R_PMU_GLOBAL_PWR_STATE 0x100 + +#define R_CSU_IDCODE 0x40 + +#define R_IOU_SLCR_WDT_CLK_SEL 0x300 +#define R_IOU_SLCR_CAN_MIO_CTRL 0x304 +#define R_IOU_SLCR_GEM_CLK_CTRL 0x308 +#define R_IOU_SLCR_SDIO_CLK_CTRL 0x30c +#define R_IOU_SLCR_CTRL_REG_SD 0x310 +#define R_IOU_SLCR_SD_ITAPDLY 0x314 +#define R_IOU_SLCR_SD_OTAPDLYSEL 0x318 +#define R_IOU_SLCR_SD_CONFIG_REG1 0x31c +#define R_IOU_SLCR_SD_CONFIG_REG2 0x320 +#define R_IOU_SLCR_SD_CONFIG_REG3 0x324 +#define R_IOU_SLCR_SD_INITPRESET 0x328 +#define R_IOU_SLCR_SD_DSPPRESET 0x32c +#define R_IOU_SLCR_SD_HSPDPRESET 0x330 +#define R_IOU_SLCR_SD_SDR12PRESET 0x334 +#define R_IOU_SLCR_SD_SDR25PRESET 0x338 +#define R_IOU_SLCR_SD_SDR50PRSET 0x33c +#define R_IOU_SLCR_SD_SDR104PRST 0x340 +#define R_IOU_SLCR_SD_DDR50PRESET 0x344 +#define R_IOU_SLCR_SD_MAXCUR1P8 0x34c +#define R_IOU_SLCR_SD_MAXCUR3P0 0x350 +#define R_IOU_SLCR_SD_MAXCUR3P3 0x354 +#define R_IOU_SLCR_SD_DLL_CTRL 0x358 +#define R_IOU_SLCR_SD_CDN_CTRL 0x35c +#define R_IOU_SLCR_GEM_CTRL 0x360 +#define R_IOU_SLCR_IOU_TTC_APB_CLK 0x380 +#define R_IOU_SLCR_IOU_TAPDLY_BYPASS 0x390 +#define R_IOU_SLCR_IOU_COHERENT_CTRL 0x400 +#define R_IOU_SLCR_VIDEO_PSS_CLK_SEL 0x404 +#define R_IOU_SLCR_IOU_INTERCONNECT_ROUTE 0x408 +#define R_IOU_SLCR_IOU_RAM_GEM0 0x500 +#define R_IOU_SLCR_IOU_RAM_GEM1 0x504 +#define R_IOU_SLCR_IOU_RAM_GEM2 0x508 +#define R_IOU_SLCR_IOU_RAM_GEM3 0x50c +#define R_IOU_SLCR_IOU_RAM_SD0 0x510 +#define R_IOU_SLCR_IOU_RAM_SD1 0x514 +#define R_IOU_SLCR_IOU_RAM_CAN0 0x518 +#define R_IOU_SLCR_IOU_RAM_CAN1 0x51c +#define R_IOU_SLCR_IOU_RAM_LQSPI 0x520 +#define R_IOU_SLCR_IOU_RAM_NAND 0x524