AI Trajectory


Trajectory: (Expand/Collapse All)
Seq Timestamp Type Name Duration
0/0 2026/10/02 04:43 flow patch-triage
0m Args:
null

Results:
{
  "EnableConfigs": null,
  "FocusSymbols": [],
  "KMSANReasoning": "",
  "NeedsKMSAN": false,
  "Reasoning": "The patch series introduces support for the Arm Realm Management Extension (RME) and Realm Management Monitor (RMM) / Realm Management Interface (RMI), which is exclusive to the arm64 architecture. The driver code under drivers/firmware/arm_rmm/ depends on ARCH_SUPPORTS_RMM, which is only selected by arm64 and cannot be compiled on amd64. The only generic kernel change is the addition of a weak stub arch_hibernation_available() returning true in kernel/power/hibernate.c, which produces no functional change on amd64. Therefore, all executable logic modified by the patch is exclusive to architectures other than the target architecture (amd64).",
  "WorthFuzzing": false
}

1/1 2026/10/02 04:43 action read-patch-diff
0m Args:
null

Results:
{
  "PatchDiff": "commit c0c6b7f46a5bcf42fd906c3277c7c85b5bdd6abe\nAuthor: syz-cluster \u003ctriage@syzkaller.com\u003e\nDate:   Fri Oct 2 04:42:57 2026 +0000\n\n    syz-cluster: applied patch under review\n\ndiff --git a/MAINTAINERS b/MAINTAINERS\nindex 6215fcb077705..1646012706823 100644\n--- a/MAINTAINERS\n+++ b/MAINTAINERS\n@@ -3951,8 +3951,11 @@ S:\tMaintained\n T:\tgit git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git\n F:\tDocumentation/arch/arm64/\n F:\tarch/arm64/\n+F:\tdrivers/firmware/arm_rmm/\n F:\tdrivers/virt/coco/arm-cca-guest/\n F:\tdrivers/virt/coco/pkvm-guest/\n+F:\tinclude/linux/arm-rmi-cmds.h\n+F:\tinclude/linux/arm-smccc-rmi.h\n F:\ttools/testing/selftests/arm64/\n X:\tarch/arm64/boot/dts/\n X:\tarch/arm64/configs/defconfig\ndiff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig\nindex b5a51b0ef9440..ff9565d3ffa59 100644\n--- a/arch/arm64/Kconfig\n+++ b/arch/arm64/Kconfig\n@@ -38,6 +38,7 @@ config ARM64\n \tselect ARCH_HAS_MEMBARRIER_SYNC_CORE\n \tselect ARCH_HAS_MEM_ENCRYPT\n \tselect ARCH_SUPPORTS_MSEAL_SYSTEM_MAPPINGS\n+\tselect ARCH_SUPPORTS_RMM\n \tselect ARCH_HAS_NMI_SAFE_THIS_CPU_OPS\n \tselect ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE\n \tselect ARCH_HAS_NONLEAF_PMD_YOUNG if ARM64_HAFT\ndiff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c\nindex 32102c3912fa7..e8b29983b0021 100644\n--- a/arch/arm64/kernel/cpufeature.c\n+++ b/arch/arm64/kernel/cpufeature.c\n@@ -293,6 +293,7 @@ static const struct arm64_ftr_bits ftr_id_aa64isar3[] = {\n static const struct arm64_ftr_bits ftr_id_aa64pfr0[] = {\n \tARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_CSV3_SHIFT, 4, 0),\n \tARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_CSV2_SHIFT, 4, 0),\n+\tARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_RME_SHIFT, 4, 0),\n \tARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_DIT_SHIFT, 4, 0),\n \tARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_AMU_SHIFT, 4, 0),\n \tARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_MPAM_SHIFT, 4, 0),\ndiff --git a/arch/arm64/kernel/hibernate.c b/arch/arm64/kernel/hibernate.c\nindex 7bf1174277772..08e03d24b93a8 100644\n--- a/arch/arm64/kernel/hibernate.c\n+++ b/arch/arm64/kernel/hibernate.c\n@@ -10,6 +10,8 @@\n  * Copyright (C) 2006 Rafael J. Wysocki \u003crjw@sisk.pl\u003e\n  */\n #define pr_fmt(x) \"hibernate: \" x\n+\n+#include \u003clinux/arm-rmi-cmds.h\u003e\n #include \u003clinux/cpu.h\u003e\n #include \u003clinux/kvm_host.h\u003e\n #include \u003clinux/pm.h\u003e\n@@ -105,6 +107,18 @@ void notrace restore_processor_state(void)\n {\n }\n \n+bool arch_hibernation_available(void)\n+{\n+\t/*\n+\t * If we have activated the RMM, there could be pages that are\n+\t * delegated to the RMM. Trying to save them to the image will be fatal.\n+\t * Also, we donate pages to the RMM at activation and restoring data\n+\t * to those pages are going to be fatal.\n+\t * Hence, disable the hibernation when the RMM is active\n+\t */\n+\treturn !cpus_are_stuck_in_kernel() \u0026\u0026 !is_rmm_active();\n+}\n+\n int arch_hibernation_header_save(void *addr, unsigned int max_size)\n {\n \tstruct arch_hibernate_hdr *hdr = addr;\ndiff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c\nindex 8f9bc2327dc85..48f343704cb54 100644\n--- a/arch/arm64/kernel/machine_kexec.c\n+++ b/arch/arm64/kernel/machine_kexec.c\n@@ -6,6 +6,7 @@\n  * Copyright (C) Huawei Futurewei Technologies.\n  */\n \n+#include \u003clinux/arm-rmi-cmds.h\u003e\n #include \u003clinux/interrupt.h\u003e\n #include \u003clinux/irq.h\u003e\n #include \u003clinux/kernel.h\u003e\n@@ -59,6 +60,16 @@ int machine_kexec_prepare(struct kimage *kimage)\n \t\treturn -EBUSY;\n \t}\n \n+\t/*\n+\t * We will be able to allow kdump to proceed, once we have the support\n+\t * for handling GPF from vmcore accesses to delegated pages. Until then\n+\t * block kexec completely.\n+\t */\n+\tif (is_rmm_active()) {\n+\t\tpr_err(\"Can't kexec: RMM is active.\\n\");\n+\t\treturn -EBUSY;\n+\t}\n+\n \treturn 0;\n }\n \ndiff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c\nindex 44aae52c473d7..f0ef348f725b6 100644\n--- a/arch/arm64/kvm/sys_regs.c\n+++ b/arch/arm64/kvm/sys_regs.c\n@@ -2163,6 +2163,8 @@ static u64 sanitise_id_aa64pfr0_el1(const struct kvm_vcpu *vcpu, u64 val)\n \t */\n \tval \u0026= ~ID_AA64PFR0_EL1_MPAM_MASK;\n \n+\tval \u0026= ~ID_AA64PFR0_EL1_RME_MASK;\n+\n \treturn val;\n }\n \ndiff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig\nindex b7cc11e4fbfa6..62660bf520a8d 100644\n--- a/drivers/firmware/Kconfig\n+++ b/drivers/firmware/Kconfig\n@@ -310,5 +310,6 @@ source \"drivers/firmware/samsung/Kconfig\"\n source \"drivers/firmware/smccc/Kconfig\"\n source \"drivers/firmware/tegra/Kconfig\"\n source \"drivers/firmware/xilinx/Kconfig\"\n+source \"drivers/firmware/arm_rmm/Kconfig\"\n \n endmenu\ndiff --git a/drivers/firmware/Makefile b/drivers/firmware/Makefile\nindex be46f1e1dc77f..196a650ccf025 100644\n--- a/drivers/firmware/Makefile\n+++ b/drivers/firmware/Makefile\n@@ -39,3 +39,4 @@ obj-y\t\t\t\t+= samsung/\n obj-y\t\t\t\t+= smccc/\n obj-y\t\t\t\t+= tegra/\n obj-y\t\t\t\t+= xilinx/\n+obj-y\t\t\t\t+= arm_rmm/\ndiff --git a/drivers/firmware/arm_rmm/Kconfig b/drivers/firmware/arm_rmm/Kconfig\nnew file mode 100644\nindex 0000000000000..a37cba6647360\n--- /dev/null\n+++ b/drivers/firmware/arm_rmm/Kconfig\n@@ -0,0 +1,25 @@\n+\n+config ARCH_SUPPORTS_RMM\n+\tbool\n+\n+config ARM_RMM_RMI\n+\tbool \"Realm Management Interface (RMI) Support\"\n+\tdepends on ARCH_SUPPORTS_RMM\n+\thelp\n+\t  Support the Realm Management Monitor (RMM) on Arm systems that\n+\t  implement the Realm Management Extension (RME), as defined by the\n+\t  Arm Confidential Compute Architecture.\n+\n+\t  The RMM runs at EL2 in the Realm world and provides the Realm\n+\t  Management Interface (RMI) used by a Normal World host to create,\n+\t  manage and run protected virtual machines called Realms. The RMM can\n+\t  also act as a TSM, as defined by the PCIe TDISP and can manage the\n+\t  PCI IDE setup for securing the PCIe links.\n+\n+\t  This option builds the host-side RMI support used by KVM to detect a\n+\t  compatible RMM, configure it, manage delegated memory and enable\n+\t  Realm guests.\n+\n+\t  Selecting this option does not by itself make Realm guests available:\n+\t  the system must also provide RME-capable hardware and firmware with a\n+\t  compatible RMM implementation.\ndiff --git a/drivers/firmware/arm_rmm/Makefile b/drivers/firmware/arm_rmm/Makefile\nnew file mode 100644\nindex 0000000000000..25e6ad80d3a19\n--- /dev/null\n+++ b/drivers/firmware/arm_rmm/Makefile\n@@ -0,0 +1 @@\n+obj-$(CONFIG_ARM_RMM_RMI)\t+= rmi.o\ndiff --git a/drivers/firmware/arm_rmm/rmi.c b/drivers/firmware/arm_rmm/rmi.c\nnew file mode 100644\nindex 0000000000000..a01ab69a64c7c\n--- /dev/null\n+++ b/drivers/firmware/arm_rmm/rmi.c\n@@ -0,0 +1,1128 @@\n+// SPDX-License-Identifier: GPL-2.0\n+/*\n+ * Copyright (C) 2023-2026 ARM Ltd.\n+ */\n+\n+#include \u003clinux/arm-rmi-cmds.h\u003e\n+#include \u003clinux/cpufeature.h\u003e\n+#include \u003clinux/memblock.h\u003e\n+#include \u003clinux/memory.h\u003e\n+#include \u003clinux/mmzone.h\u003e\n+#include \u003clinux/slab.h\u003e\n+\n+#include \u003casm/memory.h\u003e\n+#include \u003casm/pgtable-hwdef.h\u003e\n+\n+/* RMM v2.0 defines RmiFeatureRegister0 to RmiFeatureRegister4. */\n+static unsigned long rmi_feat_reg_cache[5] __ro_after_init;\n+static bool arm64_rmm_active;\n+static bool arm64_rmi_is_available;\n+\n+/* is_rmm_active: Returns if the RMM is ACTIVE */\n+bool is_rmm_active(void)\n+{\n+\treturn READ_ONCE(arm64_rmm_active);\n+}\n+EXPORT_SYMBOL_GPL(is_rmm_active);\n+\n+/*\n+ * is_rmi_available: Returns if the RMM is ACTIVE and is configured with static\n+ * carveout for metadata nad L1GPTs.\n+ */\n+bool is_rmi_available(void)\n+{\n+\treturn READ_ONCE(arm64_rmi_is_available);\n+}\n+EXPORT_SYMBOL_GPL(is_rmi_available);\n+\n+/*\n+ * rmi_granule_range_delegate() - Delegate granules\n+ * @base: PA of the first granule of the range\n+ * @top: PA of the first granule after the range\n+ * @out_top: PA of the first granule not delegated\n+ *\n+ * Delegate a range of granule for use by the realm world. If the entire range\n+ * was delegated then @out_top == @top, otherwise the function should be called\n+ * again with @base == @out_top.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static long rmi_granule_range_delegate(unsigned long base,\n+\t\t\t\t       unsigned long top,\n+\t\t\t\t       unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_GRANULE_RANGE_DELEGATE, base, top\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_top)\n+\t\t*out_top = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+/*\n+ * rmi_delegate_range: Delegate a physically contiguous range.\n+ * We iterate over the range until we hit an error. So we may\n+ * return an error, but with a partially delegated range. The\n+ * caller must always look at the @out_phys to figure out, how\n+ * much progress was made.\n+ *\n+ * @phys:\tBase of the physical address range\n+ * @size:\tSize of the physical address range\n+ * @out_phys:\tTop of the range that was completed. This is always\n+ *\t\tvalid, irrespective of the result.\n+ *\n+ * Returns RMI_SUCCESS on successful completion. Otherwise, returns\n+ * the Linux error number or the RMI status code as described\n+ * by the RMM spec for RMI_GRANULE_DELEGATE_RANGE or RMI_BLOCKED.\n+ */\n+int rmi_delegate_range(phys_addr_t phys,\n+\t\t       unsigned long size,\n+\t\t       phys_addr_t *out_phys)\n+{\n+\tlong ret = 0;\n+\tunsigned long top = phys + size;\n+\tunsigned long out_top;\n+\n+\twhile (phys \u003c top) {\n+\t\tret = rmi_granule_range_delegate(phys, top, \u0026out_top);\n+\n+\t\tif (ret != RMI_SUCCESS)\n+\t\t\tbreak;\n+\t\t/*\n+\t\t * Buggy RMM ? Let the caller handle the failure.\n+\t\t * We can't know how far the RMM delegated in this\n+\t\t * iteration, so we return the best known good limit.\n+\t\t * RMM can deal with granules already in \"undelegated\"\n+\t\t * in a given range. So, it is fine for the caller to\n+\t\t * try the range we return.\n+\t\t */\n+\t\tif (WARN_ON(out_top \u003c= phys)) {\n+\t\t\tret = -ENXIO;\n+\t\t\tbreak;\n+\t\t}\n+\t\tphys = out_top;\n+\t}\n+\n+\tif (out_phys)\n+\t\t*out_phys = phys;\n+\n+\treturn ret;\n+}\n+EXPORT_SYMBOL_GPL(rmi_delegate_range);\n+\n+/*\n+ * rmi_granule_range_undelegate() - Undelegate a range of granules\n+ * @base: Base PA of the target range\n+ * @top: Top PA of the target range\n+ * @out_top: Returns the top PA of range whose state is undelegated\n+ *\n+ * Undelegate a range of granules to allow use by the normal world. Will fail\n+ * if the granules are in use by RMM. RMM can ignore granules that are already\n+ * undelegated and thus is safe to be called on a range with a mix of delegated\n+ * and undelegated granules.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static long rmi_granule_range_undelegate(unsigned long base,\n+\t\t\t\t\t\tunsigned long top,\n+\t\t\t\t\t\tunsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_GRANULE_RANGE_UNDELEGATE, base, top\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_top)\n+\t\t*out_top = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+int rmi_undelegate_range(phys_addr_t phys,\n+\t\t\t unsigned long size)\n+{\n+\tlong ret = 0;\n+\tunsigned long top = phys + size;\n+\tunsigned long out_top;\n+\n+\twhile (phys \u003c top) {\n+\t\tret = rmi_granule_range_undelegate(phys, top, \u0026out_top);\n+\n+\t\tif (ret != RMI_SUCCESS)\n+\t\t\tbreak;\n+\t\t/* Buggy RMM ? Let the caller leak the pages */\n+\t\tif (WARN_ON(out_top \u003c= phys))\n+\t\t\treturn -ENXIO;\n+\t\tphys = out_top;\n+\t}\n+\n+\treturn ret;\n+}\n+EXPORT_SYMBOL_GPL(rmi_undelegate_range);\n+\n+/*\n+ * rmi_free_delegated_page: Undelegate and free a page that has been previously\n+ * delegated to the Realm world. If we are unable to undelegate it, the page is\n+ * leaked.\n+ * NOTE: Do not use this helper if the page could be concurrently operated by\n+ * another thread, as it may get leaked if the undelegation fails due to RMI_BLOCKED\n+ */\n+int rmi_free_delegated_page(phys_addr_t phys)\n+{\n+\tif (WARN_ON_ONCE(rmi_undelegate_page(phys))) {\n+\t\t/* Undelegate failed: leak the page */\n+\t\treturn -EBUSY;\n+\t}\n+\n+\tfree_page((unsigned long)phys_to_virt(phys));\n+\n+\treturn 0;\n+}\n+EXPORT_SYMBOL_GPL(rmi_free_delegated_page);\n+\n+/*\n+ * Convert the RmiAddrBlockSize to actual size. This is used in RmiDonateReq\n+ * and RmiAddrRangeDesc*.\n+ */\n+static unsigned long rmi_addr_block_size_to_bytes(unsigned long block_size_fld)\n+{\n+\treturn BIT(ARM64_HW_PGTABLE_LEVEL_SHIFT(3 - block_size_fld));\n+}\n+\n+/*\n+ * free_addr_range: Free memory described by the address range entry, which may\n+ *\t\t    be partially consumed by RMM.\n+ *\n+ * @entry: RMI_ADDR_RANGE descriptor\n+ * @consumed_size: Page aligned size consumed by the RMM from the address range.\n+ *\n+ * If the state of the address is DELEGATED, undelegate it back, before freeing.\n+ * Leaks the memory if we cannot undelegate the range.\n+ */\n+static void free_addr_range(unsigned long entry, unsigned long consumed_size)\n+{\n+\tunsigned long phys = RMI_ADDR_RANGE_ADDR(entry);\n+\tunsigned long block_size_fld = RMI_ADDR_RANGE_BLOCK_SIZE(entry);\n+\tunsigned long count = RMI_ADDR_RANGE_COUNT(entry);\n+\tunsigned long state = RMI_ADDR_RANGE_STATE(entry);\n+\tunsigned long size = rmi_addr_block_size_to_bytes(block_size_fld) * count;\n+\n+\tWARN_ON(!PAGE_ALIGNED(phys) || !PAGE_ALIGNED(consumed_size));\n+\n+\t/* We shouldn't see this in reclaim path, leak it for now  */\n+\tif (WARN_ON((state != RMI_OP_MEM_DELEGATED) \u0026\u0026\n+\t\t    (state != RMI_OP_MEM_UNDELEGATED)))\n+\t\treturn;\n+\n+\t/* Adjust the address and size for partially consumed entry */\n+\tphys += consumed_size;\n+\tsize -= consumed_size;\n+\t/*\n+\t * Undelegate the pages back if required. If we can't\n+\t * change them back, leak the pages.\n+\t */\n+\tif (state == RMI_OP_MEM_DELEGATED \u0026\u0026\n+\t    WARN_ON(rmi_undelegate_range(phys, size)))\n+\t\treturn;\n+\tfree_pages_exact(phys_to_virt(phys), size);\n+}\n+\n+static void rmi_op_continue(unsigned long sro_handle, unsigned long flags,\n+\t\t\t    struct arm_smccc_1_2_regs *out_regs)\n+{\n+\t*out_regs = (struct arm_smccc_1_2_regs) {\n+\t\tSMC_RMI_OP_CONTINUE, sro_handle, flags\n+\t};\n+\n+\trmi_smccc_invoke(out_regs);\n+}\n+\n+static void rmi_op_cancel(unsigned long sro_handle,\n+\t\t\t  struct arm_smccc_1_2_regs *out_regs)\n+{\n+\t*out_regs = (struct arm_smccc_1_2_regs) {\n+\t\tSMC_RMI_OP_CANCEL, sro_handle\n+\t};\n+\n+\trmi_smccc_invoke(out_regs);\n+}\n+\n+static void rmi_op_mem_donate(unsigned long sro_handle, unsigned long list_addr,\n+\t\t\t      unsigned long list_count, unsigned long flags,\n+\t\t\t      struct arm_smccc_1_2_regs *out_regs)\n+{\n+\t*out_regs = (struct arm_smccc_1_2_regs) {\n+\t\tSMC_RMI_OP_MEM_DONATE, sro_handle, list_addr, list_count, flags\n+\t};\n+\n+\t/*\n+\t * The output donated count (a1) is always valid, irrespective\n+\t * of the return result. i.e., 0 if there was an error\n+\t */\n+\trmi_smccc_invoke(out_regs);\n+}\n+\n+static void rmi_op_mem_reclaim(unsigned long sro_handle,\n+\t\t\t       unsigned long list_addr,\n+\t\t\t       unsigned long list_count,\n+\t\t\t       struct arm_smccc_1_2_regs *out_regs)\n+{\n+\t*out_regs = (struct arm_smccc_1_2_regs) {\n+\t\tSMC_RMI_OP_MEM_RECLAIM, sro_handle, list_addr, list_count\n+\t};\n+\n+\trmi_smccc_invoke(out_regs);\n+}\n+\n+static int rmi_sro_ensure_capacity(struct rmi_sro_state *sro,\n+\t\t\t\t   unsigned long count)\n+{\n+\tif (WARN_ON_ONCE(sro-\u003eaddr_count \u003e RMI_MAX_ADDR_LIST))\n+\t\treturn -EOVERFLOW;\n+\n+\tif (count \u003e RMI_MAX_ADDR_LIST - sro-\u003eaddr_count)\n+\t\treturn -ENOSPC;\n+\n+\treturn 0;\n+}\n+\n+static int rmi_sro_donate_contig(struct rmi_sro_state *sro,\n+\t\t\t\t unsigned long sro_handle,\n+\t\t\t\t unsigned long donatereq,\n+\t\t\t\t struct arm_smccc_1_2_regs *out_regs,\n+\t\t\t\t gfp_t gfp)\n+{\n+\tunsigned long block_size_fld = RMI_DONATE_BLOCK_SIZE(donatereq);\n+\tunsigned long block_size = rmi_addr_block_size_to_bytes(block_size_fld);\n+\tunsigned long count = RMI_DONATE_COUNT(donatereq);\n+\tunsigned long state = RMI_DONATE_STATE(donatereq);\n+\tunsigned long size = block_size * count;\n+\tunsigned long addr_range;\n+\tunsigned long donated_size;\n+\tint ret;\n+\tvoid *virt;\n+\tphys_addr_t phys;\n+\n+\t/*\n+\t * The RMM specification requires contiguous allocations are always a\n+\t * power of 2\n+\t */\n+\tif (WARN_ON_ONCE(!is_power_of_2(size)))\n+\t\treturn -EINVAL;\n+\t/*\n+\t * RMM clamps the Maximum value of RmiOpMemDonateReq:count to prevent\n+\t * overflow in the RMI_ADDR_RANGE_COUNT field.\n+\t */\n+\tif (WARN_ON_ONCE(count \u003e FIELD_MAX(RMI_ADDR_RANGE_COUNT_MASK)))\n+\t\treturn -EINVAL;\n+\n+\t/*\n+\t * We can't allocate blocks more than what the buddy allocator can\n+\t * satisfy. TODO: Add support for larger allocations.\n+\t */\n+\tif (get_order(size) \u003e MAX_PAGE_ORDER)\n+\t\treturn -ENOMEM;\n+\n+\t/* Reuse the cached address range if we have one */\n+\tfor (int i = 0; i \u003c sro-\u003eaddr_count; i++) {\n+\t\tunsigned long entry = sro-\u003eaddr_list[i];\n+\n+\t\tif (RMI_ADDR_RANGE_BLOCK_SIZE(entry) == block_size_fld \u0026\u0026\n+\t\t    RMI_ADDR_RANGE_COUNT(entry) == count \u0026\u0026\n+\t\t    RMI_ADDR_RANGE_STATE(entry) == state \u0026\u0026\n+\t\t    IS_ALIGNED(RMI_ADDR_RANGE_ADDR(entry), size)) {\n+\t\t\tsro-\u003eaddr_count--;\n+\t\t\tswap(sro-\u003eaddr_list[sro-\u003eaddr_count],\n+\t\t\t     sro-\u003eaddr_list[i]);\n+\n+\t\t\tgoto mem_donate;\n+\t\t}\n+\t}\n+\n+\tret = rmi_sro_ensure_capacity(sro, 1);\n+\tif (ret)\n+\t\treturn ret;\n+\n+\tvirt = alloc_pages_exact(size, gfp);\n+\tif (!virt)\n+\t\treturn -ENOMEM;\n+\tphys = virt_to_phys(virt);\n+\n+\tif (state == RMI_OP_MEM_DELEGATED) {\n+\t\tphys_addr_t delegated_phys;\n+\n+\t\tif (rmi_delegate_range(phys, size, \u0026delegated_phys)) {\n+\t\t\tif (!rmi_undelegate_range(phys, delegated_phys - phys))\n+\t\t\t\tfree_pages_exact(virt, size);\n+\t\t\treturn -ENXIO;\n+\t\t}\n+\t}\n+\n+\taddr_range = phys \u0026 RMI_ADDR_RANGE_ADDR_MASK;\n+\tFIELD_MODIFY(RMI_ADDR_RANGE_BLOCK_SIZE_MASK, \u0026addr_range, block_size_fld);\n+\tFIELD_MODIFY(RMI_ADDR_RANGE_COUNT_MASK, \u0026addr_range, count);\n+\tFIELD_MODIFY(RMI_ADDR_RANGE_STATE_MASK, \u0026addr_range, state);\n+\n+\tsro-\u003eaddr_list[sro-\u003eaddr_count] = addr_range;\n+\n+mem_donate:\n+\trmi_op_mem_donate(sro_handle,\n+\t\t\t  virt_to_phys(\u0026sro-\u003eaddr_list[sro-\u003eaddr_count]), 1,\n+\t\t\t  0, out_regs);\n+\tdonated_size = out_regs-\u003ea1 \u003c\u003c PAGE_SHIFT;\n+\n+\tif (WARN_ON(out_regs-\u003ea1 \u003e (size \u003e\u003e PAGE_SHIFT)))\n+\t\tdonated_size = size;\n+\n+\t/* All granules consumed by the RMM */\n+\tif (donated_size == size)\n+\t\treturn 0;\n+\t/* No granules were consumed by the RMM, cache them */\n+\tif (donated_size == 0) {\n+\t\tsro-\u003eaddr_count++;\n+\t\treturn 0;\n+\t}\n+\n+\t/* The granules were partially consumed, reclaim the unused ones. */\n+\tfree_addr_range(sro-\u003eaddr_list[sro-\u003eaddr_count], donated_size);\n+\n+\treturn 0;\n+}\n+\n+static int rmi_sro_donate_noncontig(struct rmi_sro_state *sro,\n+\t\t\t\t    unsigned long sro_handle,\n+\t\t\t\t    unsigned long donatereq,\n+\t\t\t\t    struct arm_smccc_1_2_regs *out_regs,\n+\t\t\t\t    gfp_t gfp)\n+{\n+\tunsigned long block_size_fld = RMI_DONATE_BLOCK_SIZE(donatereq);\n+\tunsigned long block_size = rmi_addr_block_size_to_bytes(block_size_fld);\n+\tunsigned long count = RMI_DONATE_COUNT(donatereq);\n+\tunsigned long state = RMI_DONATE_STATE(donatereq);\n+\tunsigned long found = 0;\n+\tunsigned long donated_granules;\n+\tunsigned long granules_per_block = block_size \u003e\u003e PAGE_SHIFT;\n+\tunsigned long consumed_blocks;\n+\tint addr_list_start = sro-\u003eaddr_count;\n+\tint ret, i, src;\n+\n+\t/*\n+\t * We can't allocate blocks more than what the buddy allocator can\n+\t * satisfy. TODO: Add support for larger allocations.\n+\t */\n+\tif (get_order(block_size) \u003e MAX_PAGE_ORDER)\n+\t\treturn -ENOMEM;\n+\t/*\n+\t * Maximum value of RmiOpMemDonateReq:count is 2^(PAGE_SHIFT-2) - 1.\n+\t * But we further clamp it down by the number of entries we can do\n+\t * in one go. The RMM can request  the remaining in the next iteration.\n+\t */\n+\tif (count \u003e RMI_MAX_ADDR_LIST)\n+\t\tcount = RMI_MAX_ADDR_LIST;\n+\n+\t/* Gather the suitable entries to the end of the list */\n+\ti = 0;\n+\twhile (i \u003c  addr_list_start \u0026\u0026 found \u003c count) {\n+\t\tunsigned long entry = sro-\u003eaddr_list[i];\n+\n+\t\tif (RMI_ADDR_RANGE_BLOCK_SIZE(entry) == block_size_fld \u0026\u0026\n+\t\t    RMI_ADDR_RANGE_COUNT(entry) == 1 \u0026\u0026\n+\t\t    RMI_ADDR_RANGE_STATE(entry) == state) {\n+\t\t\taddr_list_start--;\n+\t\t\tswap(sro-\u003eaddr_list[addr_list_start],\n+\t\t\t     sro-\u003eaddr_list[i]);\n+\t\t\tfound++;\n+\t\t\t/* Continue from the swapped in entry */\n+\t\t\tcontinue;\n+\t\t}\n+\t\t/* Skip past the entry */\n+\t\ti++;\n+\t}\n+\n+\tret = rmi_sro_ensure_capacity(sro, count - found);\n+\tif (ret) {\n+\t\t/* If we have found some entries, donate them and try again */\n+\t\tif (found)\n+\t\t\tgoto mem_donate;\n+\t\t/* Otherwise free up the list and start again */\n+\t\trmi_sro_free(sro);\n+\t\t/* Reset the addr_list_start to match sro-\u003eaddr_count */\n+\t\taddr_list_start = 0;\n+\t}\n+\n+\tfor (; found \u003c count; found++) {\n+\t\tunsigned long addr_range;\n+\t\tvoid *virt = alloc_pages_exact(block_size, gfp);\n+\t\tphys_addr_t phys;\n+\n+\t\tif (!virt)\n+\t\t\treturn -ENOMEM;\n+\n+\t\tphys = virt_to_phys(virt);\n+\n+\t\tif (state == RMI_OP_MEM_DELEGATED) {\n+\t\t\tphys_addr_t delegated_phys;\n+\n+\t\t\tif (rmi_delegate_range(phys, block_size, \u0026delegated_phys)) {\n+\t\t\t\tif (!rmi_undelegate_range(phys, delegated_phys - phys))\n+\t\t\t\t\tfree_pages_exact(virt, block_size);\n+\t\t\t\treturn -ENXIO;\n+\t\t\t}\n+\t\t}\n+\n+\t\taddr_range = phys \u0026 RMI_ADDR_RANGE_ADDR_MASK;\n+\t\tFIELD_MODIFY(RMI_ADDR_RANGE_BLOCK_SIZE_MASK, \u0026addr_range, block_size_fld);\n+\t\tFIELD_MODIFY(RMI_ADDR_RANGE_COUNT_MASK, \u0026addr_range, 1);\n+\t\tFIELD_MODIFY(RMI_ADDR_RANGE_STATE_MASK, \u0026addr_range, state);\n+\n+\t\tsro-\u003eaddr_list[sro-\u003eaddr_count++] = addr_range;\n+\t}\n+\n+mem_donate:\n+\trmi_op_mem_donate(sro_handle,\n+\t\t\t  virt_to_phys(\u0026sro-\u003eaddr_list[addr_list_start]),\n+\t\t\t  found, 0, out_regs);\n+\n+\tdonated_granules = out_regs-\u003ea1;\n+\t/*\n+\t * The RMM shouldn't report more granules than we provided, but clamp\n+\t * just in case.\n+\t */\n+\tif (WARN_ON_ONCE(donated_granules \u003e found * granules_per_block))\n+\t\tdonated_granules = found * granules_per_block;\n+\n+\t/*\n+\t * The RMM reports the consumed memory in terms of granules, but we\n+\t * track in the address lists in block-sized ranges. So divide to get\n+\t * the number of (complete) consumed blocks.\n+\t */\n+\tconsumed_blocks = donated_granules / granules_per_block;\n+\tif (donated_granules % granules_per_block) {\n+\t\t/*\n+\t\t * A block has been partially consumed, the start is owned by\n+\t\t * the RMM, the tail is owned by the host\n+\t\t */\n+\t\tunsigned long entry =\n+\t\t\tsro-\u003eaddr_list[addr_list_start + consumed_blocks];\n+\t\tunsigned long donated_size =\n+\t\t\t(donated_granules % granules_per_block) \u003c\u003c PAGE_SHIFT;\n+\n+\t\tfree_addr_range(entry, donated_size);\n+\t\t/*\n+\t\t * This block is now fully 'consumed' (either held by the RMM or\n+\t\t * freed)\n+\t\t */\n+\t\tconsumed_blocks++;\n+\t}\n+\n+\t/*\n+\t * Keep just the blocks the RMM didn't use in addr_list\n+\t * RMM claimed consumed_blocks entries from addr_list_start.\n+\t * Move the entries left out at the end i.e.,\n+\t * [ addr_list_start + consumed_blocks, addr_list_start + found)\n+\t * to the rest of the valid entries and adjust the addr_count to\n+\t * reflect the available entries.\n+\t */\n+\tfor (i = 0, src = addr_list_start + consumed_blocks;\n+\t\ti \u003c found - consumed_blocks; i++)\n+\t\tsro-\u003eaddr_list[addr_list_start + i] = sro-\u003eaddr_list[src + i];\n+\n+\tsro-\u003eaddr_count -= consumed_blocks;\n+\n+\treturn 0;\n+}\n+\n+static int rmi_sro_donate(struct rmi_sro_state *sro,\n+\t\t\t  unsigned long sro_handle,\n+\t\t\t  unsigned long donatereq,\n+\t\t\t  struct arm_smccc_1_2_regs *regs,\n+\t\t\t  gfp_t gfp)\n+{\n+\tunsigned long state = RMI_DONATE_STATE(donatereq);\n+\n+\tif (WARN_ON_ONCE(!RMI_DONATE_COUNT(donatereq)))\n+\t\treturn -EINVAL;\n+\n+\t/*\n+\t * We do not support RMI_OP_MEM_CONDITIONAL yet. This is only required\n+\t * for use in RMI_GRANULE_TRACKING_SET, which we don't support yet.\n+\t */\n+\tif (WARN_ON_ONCE((state != RMI_OP_MEM_DELEGATED) \u0026\u0026\n+\t\t\t (state != RMI_OP_MEM_UNDELEGATED)))\n+\t\treturn -EINVAL;\n+\n+\tif (RMI_DONATE_CONTIG(donatereq) == RMI_OP_MEM_CONTIG)\n+\t\treturn rmi_sro_donate_contig(sro, sro_handle, donatereq, regs, gfp);\n+\telse\n+\t\treturn rmi_sro_donate_noncontig(sro, sro_handle, donatereq, regs, gfp);\n+}\n+\n+static int rmi_sro_reclaim(struct rmi_sro_state *sro,\n+\t\t\t   unsigned long sro_handle,\n+\t\t\t   struct arm_smccc_1_2_regs *out_regs)\n+{\n+\tunsigned long capacity;\n+\n+\t/*\n+\t * We don't do a partial free of the entries. So for now free the\n+\t * entire address list as we prepare to reclaim more from the RMM.\n+\t */\n+\tif (rmi_sro_ensure_capacity(sro, 1))\n+\t\trmi_sro_free(sro);\n+\n+\tcapacity = RMI_MAX_ADDR_LIST - sro-\u003eaddr_count;\n+\n+\trmi_op_mem_reclaim(sro_handle,\n+\t\t\t   virt_to_phys(\u0026sro-\u003eaddr_list[sro-\u003eaddr_count]),\n+\t\t\t   capacity, out_regs);\n+\n+\t/*\n+\t * RMI_OP_MEM_RECLAIM always return RMI_INCOMPLETE, except when the\n+\t * input parameters were invalid.\n+\t */\n+\tif (WARN_ON_ONCE(RMI_RESULT_STATUS(out_regs-\u003ea0) != RMI_INCOMPLETE))\n+\t\treturn -EINVAL;\n+\tif (WARN_ON_ONCE(out_regs-\u003ea1 \u003e capacity))\n+\t\tout_regs-\u003ea1 = capacity;\n+\n+\tsro-\u003eaddr_count += out_regs-\u003ea1;\n+\n+\treturn 0;\n+}\n+\n+void rmi_sro_free(struct rmi_sro_state *sro)\n+{\n+\t/* Handle the worse */\n+\tif (WARN_ON(sro-\u003eaddr_count \u003c 0))\n+\t\treturn;\n+\n+\tif (WARN_ON(sro-\u003eaddr_count \u003e RMI_MAX_ADDR_LIST))\n+\t\tsro-\u003eaddr_count = RMI_MAX_ADDR_LIST;\n+\n+\tfor (int i = 0; i \u003c sro-\u003eaddr_count; i++)\n+\t\tfree_addr_range(sro-\u003eaddr_list[i], 0);\n+\n+\tsro-\u003eaddr_count = 0;\n+}\n+EXPORT_SYMBOL_GPL(rmi_sro_free);\n+\n+long rmi_sro_memxfer_execute(struct rmi_sro_state *sro, gfp_t gfp)\n+{\n+\tstruct arm_smccc_1_2_regs *regs = \u0026sro-\u003eregs;\n+\tbool cancelled = false;\n+\tunsigned long sro_handle;\n+\n+\trmi_smccc_invoke(regs);\n+\n+\tsro_handle = regs-\u003ea1;\n+\twhile (RMI_RESULT_STATUS(regs-\u003ea0) == RMI_INCOMPLETE) {\n+\t\tbool can_cancel = RMI_RESULT_CAN_CANCEL(regs-\u003ea0) == RMI_OP_CAN_CANCEL;\n+\t\tint ret = 0;\n+\n+\t\tswitch (RMI_RESULT_MEMREQ(regs-\u003ea0)) {\n+\t\tcase RMI_OP_MEM_REQ_NONE:\n+\t\t\trmi_op_continue(sro_handle, RMI_CONTINUE_KEEP_GOING,\n+\t\t\t\t\tregs);\n+\t\t\tbreak;\n+\t\tcase RMI_OP_MEM_REQ_DONATE:\n+\t\t\tret = rmi_sro_donate(sro, sro_handle, regs-\u003ea2, regs,\n+\t\t\t\t\t     gfp);\n+\t\t\tbreak;\n+\t\tcase RMI_OP_MEM_REQ_RECLAIM:\n+\t\t\tret = rmi_sro_reclaim(sro, sro_handle, regs);\n+\t\t\tbreak;\n+\t\tdefault:\n+\t\t\tWARN_ON_ONCE(1);\n+\t\t\tret = -ENXIO;\n+\t\t}\n+\n+\t\tif (ret) {\n+\t\t\t/*\n+\t\t\t * All memory donating SROs must be cancellable. So a\n+\t\t\t * failure in memory allocation shouldn't be an issue.\n+\t\t\t * However, if we encounter a random failure (e.g.,\n+\t\t\t * buggy RMM), don't loop forever, just give up.\n+\t\t\t */\n+\t\t\tif (WARN_ON_ONCE(!can_cancel))\n+\t\t\t\treturn ret;\n+\t\t\t/*\n+\t\t\t * If we have already cancelled, and came back here due\n+\t\t\t * to an error in MEMREQ, then there is no point\n+\t\t\t * in going in loops.\n+\t\t\t */\n+\t\t\tif (WARN_ON_ONCE(cancelled))\n+\t\t\t\tbreak;\n+\t\t\trmi_op_cancel(sro_handle, regs);\n+\t\t\tcancelled = true;\n+\n+\t\t\tif (WARN_ON_ONCE(RMI_RESULT_STATUS(regs-\u003ea0) != RMI_INCOMPLETE))\n+\t\t\t\treturn ret;\n+\t\t}\n+\t}\n+\n+\tif (cancelled)\n+\t\treturn -ECANCELED;\n+\n+\treturn regs-\u003ea0;\n+}\n+EXPORT_SYMBOL_GPL(rmi_sro_memxfer_execute);\n+\n+/*\n+ * rmi_sro_execute: Execute an RMI command that is stateful but not memory\n+ * transferring. Takes regs, filled with the FIDs and the arguments in place.\n+ *\n+ * Returns :\n+ *  -ECANCELED - If the operation had to be aborted and SRO was cancellable.\n+ *  Otherwise, returns the result of the RMI command.\n+ */\n+long rmi_sro_execute(struct arm_smccc_1_2_regs *regs)\n+{\n+\tbool cancelled = false;\n+\tunsigned long sro_handle;\n+\n+\trmi_smccc_invoke(regs);\n+\n+\tsro_handle = regs-\u003ea1;\n+\twhile (RMI_RESULT_STATUS(regs-\u003ea0) == RMI_INCOMPLETE) {\n+\t\tbool can_cancel = RMI_RESULT_CAN_CANCEL(regs-\u003ea0) == RMI_OP_CAN_CANCEL;\n+\n+\t\tswitch (RMI_RESULT_MEMREQ(regs-\u003ea0)) {\n+\t\tcase RMI_OP_MEM_REQ_NONE:\n+\t\t\trmi_op_continue(sro_handle, RMI_CONTINUE_KEEP_GOING,\n+\t\t\t\t\tregs);\n+\t\t\tbreak;\n+\t\tdefault:\n+\t\t\tWARN_ON_ONCE(1);\n+\t\t\tif (!can_cancel)\n+\t\t\t\treturn regs-\u003ea0;\n+\t\t\t/*\n+\t\t\t * We can't get here normally, but handle this anyway\n+\t\t\t * for a buggy RMM implementation.\n+\t\t\t */\n+\t\t\tif (cancelled)\n+\t\t\t\treturn -ECANCELED;\n+\t\t\trmi_op_cancel(sro_handle, regs);\n+\t\t\tcancelled = true;\n+\t\t}\n+\t}\n+\n+\tif (cancelled)\n+\t\treturn -ECANCELED;\n+\n+\treturn regs-\u003ea0;\n+}\n+EXPORT_SYMBOL_GPL(rmi_sro_execute);\n+\n+static int rmi_check_version(void)\n+{\n+\tunsigned short version_major, version_minor;\n+\tunsigned long host_version = RMI_ABI_VERSION(RMI_ABI_MAJOR_VERSION,\n+\t\t\t\t\t\t     RMI_ABI_MINOR_VERSION);\n+\tunsigned long aa64pfr0 = read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1);\n+\tstruct arm_smccc_1_2_regs res = {\n+\t\tSMC_RMI_VERSION, host_version,\n+\t};\n+\n+\t/* If RME isn't supported, then RMI can't be */\n+\tif (cpuid_feature_extract_unsigned_field(aa64pfr0, ID_AA64PFR0_EL1_RME_SHIFT) == 0)\n+\t\treturn -ENXIO;\n+\n+\trmi_smccc_invoke(\u0026res);\n+\tif (res.a0 == SMCCC_RET_NOT_SUPPORTED)\n+\t\treturn -ENXIO;\n+\n+\tversion_major = RMI_ABI_VERSION_GET_MAJOR(res.a1);\n+\tversion_minor = RMI_ABI_VERSION_GET_MINOR(res.a1);\n+\n+\tif (res.a0 != RMI_SUCCESS) {\n+\t\tunsigned short high_version_major, high_version_minor;\n+\n+\t\thigh_version_major = RMI_ABI_VERSION_GET_MAJOR(res.a2);\n+\t\thigh_version_minor = RMI_ABI_VERSION_GET_MINOR(res.a2);\n+\n+\t\tpr_err(\"Unsupported RMI ABI (v%d.%d - v%d.%d) we want v%d.%d\\n\",\n+\t\t       version_major, version_minor,\n+\t\t       high_version_major, high_version_minor,\n+\t\t       RMI_ABI_MAJOR_VERSION,\n+\t\t       RMI_ABI_MINOR_VERSION);\n+\t\treturn -ENXIO;\n+\t}\n+\n+\tpr_info(\"RMI ABI version %d.%d\\n\", version_major, version_minor);\n+\n+\treturn 0;\n+}\n+\n+static int rmi_read_features(void)\n+{\n+\t/*\n+\t * Since we've negotiated a compatible version these feature registers\n+\t * should always be available\n+\t */\n+\tfor (int i = 0; i \u003c ARRAY_SIZE(rmi_feat_reg_cache); i++) {\n+\t\tstruct arm_smccc_1_2_regs args = {\n+\t\t\tSMC_RMI_FEATURES, i,\n+\t\t};\n+\n+\t\trmi_smccc_invoke(\u0026args);\n+\t\tif (WARN_ON(args.a0 != RMI_SUCCESS))\n+\t\t\treturn -EINVAL;\n+\n+\t\trmi_feat_reg_cache[i] = args.a1;\n+\t}\n+\n+\treturn 0;\n+}\n+\n+unsigned long rmi_feat_reg(unsigned int index)\n+{\n+\tif (WARN_ON(index \u003e= ARRAY_SIZE(rmi_feat_reg_cache)))\n+\t\treturn 0;\n+\n+\treturn rmi_feat_reg_cache[index];\n+}\n+EXPORT_SYMBOL_GPL(rmi_feat_reg);\n+\n+static int rmi_rmm_config_set(unsigned long cfg_ptr)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RMM_CONFIG_SET, cfg_ptr,\n+\t};\n+\n+\trmi_smccc_invoke(\u0026regs);\n+\n+\treturn regs.a0;\n+}\n+\n+static int rmi_configure(void)\n+{\n+\tunsigned long granule_feature;\n+\tunsigned long granule_size;\n+\tint ret;\n+\n+\tswitch (PAGE_SIZE) {\n+\tcase SZ_4K:\n+\t\tgranule_size = RMI_GRANULE_SIZE_4KB;\n+\t\tgranule_feature = RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_4KB;\n+\t\tbreak;\n+\tcase SZ_16K:\n+\t\tgranule_size = RMI_GRANULE_SIZE_16KB;\n+\t\tgranule_feature = RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_16KB;\n+\t\tbreak;\n+\tcase SZ_64K:\n+\t\tgranule_size = RMI_GRANULE_SIZE_64KB;\n+\t\tgranule_feature = RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_64KB;\n+\t\tbreak;\n+\tdefault:\n+\t\tBUILD_BUG();\n+\t}\n+\n+\tif (!(rmi_feat_reg(1) \u0026 granule_feature)) {\n+\t\tpr_err(\"RMM does not support %luKB granules\\n\",\n+\t\t       PAGE_SIZE \u003e\u003e 10);\n+\t\treturn -ENXIO;\n+\t}\n+\n+\tstruct rmm_config *config __free(free_page) =\n+\t\t(struct rmm_config *)get_zeroed_page(GFP_KERNEL);\n+\n+\tif (!config) {\n+\t\tpr_err(\"Unable to allocate memory for RMM config\\n\");\n+\t\treturn -ENOMEM;\n+\t}\n+\n+\tconfig-\u003ermi_granule_size = granule_size;\n+\n+\t/*\n+\t * For now we set the tracking_region_size to 0 which is the only option\n+\t * for 4KB PAGE_SIZE (1GB for 4KB PAGE_SIZE, 32MB/512MB for 16KB/64KB).\n+\t * TODO: Support other tracking sizes via Kconfig option for other\n+\t * PAGE_SIZES\n+\t */\n+\tconfig-\u003etracking_region_size = 0;\n+\n+\tret = rmi_rmm_config_set(virt_to_phys(config));\n+\tif (ret) {\n+\t\tpr_err(\"RMM config set failed (%d)\\n\", ret);\n+\t\treturn -EINVAL;\n+\t}\n+\n+\treturn 0;\n+}\n+\n+/**\n+ * rmi_granule_tracking_get() - Get configuration of a Granule tracking region\n+ * @start: Base PA of the tracking region\n+ * @end: End of the PA region\n+ * @out_category: Memory category\n+ * @out_state: Tracking region state\n+ * @out_top: Top of the memory region\n+ *\n+ * Return: RMI return code\n+ */\n+static int rmi_granule_tracking_get(unsigned long start,\n+\t\t\t\t    unsigned long end,\n+\t\t\t\t    unsigned long *out_category,\n+\t\t\t\t    unsigned long *out_state,\n+\t\t\t\t    unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_GRANULE_TRACKING_GET, start, end,\n+\t};\n+\n+\trmi_smccc_invoke(\u0026regs);\n+\n+\tif (regs.a0 != RMI_SUCCESS)\n+\t\treturn regs.a0;\n+\n+\tif (out_category)\n+\t\t*out_category = regs.a1;\n+\tif (out_state)\n+\t\t*out_state = regs.a2;\n+\tif (out_top)\n+\t\t*out_top = regs.a3;\n+\n+\treturn RMI_SUCCESS;\n+}\n+\n+/*\n+ * Make sure the area is tracked by RMM at FINE granularity.\n+ * We do not support changing the tracking yet.\n+ */\n+static int rmi_verify_memory_tracking(phys_addr_t start, phys_addr_t end)\n+{\n+\twhile (start \u003c end) {\n+\t\tunsigned long ret, category, state, next;\n+\n+\t\tret = rmi_granule_tracking_get(start, end, \u0026category, \u0026state, \u0026next);\n+\t\tif (ret != RMI_SUCCESS) {\n+\t\t\tpr_err(\"RMI_GRANULE_TRACKING_GET failed for %llx-%llx: (%lu)\\n\",\n+\t\t\t\tstart, end, ret);\n+\t\t\treturn -ENXIO;\n+\t\t}\n+\n+\t\tif (WARN_ON(next \u003c= start))\n+\t\t\treturn -ENXIO;\n+\n+\t\tif (state != RMI_TRACKING_FINE ||\n+\t\t    category != RMI_MEM_CATEGORY_CONVENTIONAL) {\n+\t\t\t/* TODO: Set granule tracking in this case */\n+\t\t\tpr_err(\"Granule tracking for region isn't fine/conventional: %llx-%lx\\n\",\n+\t\t\t       start, next);\n+\t\t\treturn -ENODEV;\n+\t\t}\n+\t\tstart = next;\n+\t}\n+\n+\treturn 0;\n+}\n+\n+/*\n+ * rmi_gpt_info - Query the GPT info for the given PAR.\n+ * @start: Base of the physical address region\n+ * @end: Top of the physical address region\n+ * @out_top: Top of the physical address region for which the GPT @out_gpt_par_state\n+ *\t\tis valid\n+ * @out_gpt_par_state: State of the GPT covered by [start, out_top)\n+ */\n+static long rmi_gpt_info(unsigned long start, unsigned long end,\n+\t\t\t unsigned long *out_top,\n+\t\t\t unsigned long *out_gpt_par_state)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_GPT_INFO, start, end,\n+\t};\n+\n+\trmi_smccc_invoke(\u0026regs);\n+\tif (regs.a0 != RMI_SUCCESS)\n+\t\treturn regs.a0;\n+\n+\tif (out_top)\n+\t\t*out_top = regs.a1;\n+\tif (out_gpt_par_state)\n+\t\t*out_gpt_par_state = regs.a2;\n+\n+\treturn RMI_SUCCESS;\n+}\n+\n+/*\n+ * We do not support creating L1 GPTs yet. So, make sure that all the regions\n+ * are managed by the firmware.\n+ */\n+static int rmi_verify_gpt_firmware_managed(phys_addr_t start, phys_addr_t end)\n+{\n+\tunsigned long l0gpt_sz;\n+\tunsigned long next, par_state;\n+\n+\tl0gpt_sz = 1UL \u003c\u003c (30 + FIELD_GET(RMI_FEATURE_REGISTER_1_L0GPTSZ,\n+\t\t\t\t\t  rmi_feat_reg(1)));\n+\tstart = ALIGN_DOWN(start, l0gpt_sz);\n+\tend = ALIGN(end, l0gpt_sz);\n+\n+\twhile (start \u003c end) {\n+\t\tlong ret = rmi_gpt_info(start, end, \u0026next, \u0026par_state);\n+\n+\t\tif (ret != RMI_SUCCESS) {\n+\t\t\tpr_err(\"RMI_GPT_INFO failed for %llx-%llx: (%ld)\\n\",\n+\t\t\t\tstart, end, ret);\n+\t\t\treturn -ENXIO;\n+\t\t}\n+\n+\t\tif (WARN_ON(next \u003c= start))\n+\t\t\treturn -ENXIO;\n+\n+\t\tif (par_state != RMI_GPT_PAR_PLAT) {\n+\t\t\tpr_err(\"GPT for the region is not managed by firmware %llx-%lx\\n\",\n+\t\t\t\tstart, next);\n+\t\t\treturn -ENODEV;\n+\t\t}\n+\t\tstart = next;\n+\t}\n+\n+\treturn 0;\n+}\n+\n+static int rmi_prepare_memory(phys_addr_t start, phys_addr_t end)\n+{\n+\tint ret;\n+\n+\tif (start \u003e= end)\n+\t\treturn -EINVAL;\n+\n+\tret = rmi_verify_memory_tracking(start, end);\n+\tif (ret)\n+\t\treturn ret;\n+\n+\treturn rmi_verify_gpt_firmware_managed(start, end);\n+}\n+\n+static int rmi_init_metadata(void)\n+{\n+\tphys_addr_t start, end;\n+\tstruct memblock_region *r;\n+\n+\tfor_each_mem_region(r) {\n+\t\tint ret;\n+\n+\t\t/* Firmware-reserved NOMAP regions are not usable system RAM */\n+\t\tif (memblock_is_nomap(r))\n+\t\t\tcontinue;\n+\n+\t\tstart = PAGE_ALIGN(r-\u003ebase);\n+\t\t/*\n+\t\t * We always deal with PAGE_SIZE and if in the odd case this\n+\t\t * region boundary is not PAGE aligned, we stick to the page\n+\t\t * that we can use from the region.\n+\t\t */\n+\t\tend = PAGE_ALIGN_DOWN(r-\u003ebase + r-\u003esize);\n+\t\t/* Too small ? */\n+\t\tif (start \u003e= end)\n+\t\t\tcontinue;\n+\n+\t\tret = rmi_prepare_memory(start, end);\n+\t\tif (ret)\n+\t\t\treturn ret;\n+\t}\n+\n+\treturn 0;\n+}\n+\n+static int rmi_memory_notifier(struct notifier_block *nb,\n+\t\t\t       unsigned long action, void *data)\n+{\n+\tstruct memory_notify *arg = data;\n+\tphys_addr_t start, end;\n+\tint ret;\n+\n+\tif (action != MEM_GOING_ONLINE)\n+\t\treturn NOTIFY_DONE;\n+\n+\tstart = PFN_PHYS(arg-\u003estart_pfn);\n+\tend = PFN_PHYS(arg-\u003estart_pfn + arg-\u003enr_pages);\n+\n+\t/*\n+\t * If the hotplugged memory region is in ZONE_MOVABLE, neither the\n+\t * guest-memfd nor the kernel allocations can come from there. So we are\n+\t * fine to skip this.\n+\t * For boot time memory, we don't allow this discount, as the\n+\t * decision to move boot memory pages to ZONE_MOVABLE is purely\n+\t * a kernel choice and the firmware must have decided what it covers.\n+\t */\n+\tif (is_zone_movable_page(pfn_to_page(arg-\u003estart_pfn)))\n+\t\treturn NOTIFY_DONE;\n+\n+\tret = rmi_prepare_memory(start, end);\n+\n+\treturn notifier_from_errno(ret);\n+}\n+\n+static struct notifier_block rmi_memory_nb = {\n+\t.notifier_call = rmi_memory_notifier,\n+};\n+\n+static int rmi_init_memory(void)\n+{\n+\tint ret;\n+\n+\tret = rmi_init_metadata();\n+\tif (ret)\n+\t\treturn ret;\n+\n+\treturn register_memory_notifier(\u0026rmi_memory_nb);\n+}\n+\n+static int __init arm64_init_rmi(void)\n+{\n+\tint ret;\n+\n+\t/* If we can't agree on the RMI ABI version, don't proceed further */\n+\tret = rmi_check_version();\n+\tif (ret)\n+\t\treturn ret;\n+\n+\tret = rmi_read_features();\n+\tif (ret)\n+\t\treturn ret;\n+\n+\tret = rmi_configure();\n+\tif (ret)\n+\t\treturn ret;\n+\n+\t/* Activate the RMM */\n+\tstruct rmi_sro_state *sro __free(kfree) = kmalloc_obj(*sro);\n+\tif (!sro)\n+\t\treturn -ENOMEM;\n+\n+\tret = rmi_sro_memxfer_cmd(sro, GFP_KERNEL, SMC_RMI_RMM_ACTIVATE);\n+\tif (ret) {\n+\t\tpr_err(\"RMM activate failed (%d)\\n\", ret);\n+\t\tret = ret \u003c 0 ? ret : -ENXIO;\n+\t\treturn ret;\n+\t}\n+\n+\tWRITE_ONCE(arm64_rmm_active, true);\n+\n+\tret = rmi_init_memory();\n+\tif (ret) {\n+\t\t/* Deactivate the RMM */\n+\t\tif (!WARN_ON(rmi_sro_memxfer_cmd(sro, GFP_KERNEL, SMC_RMI_RMM_DEACTIVATE)))\n+\t\t\tWRITE_ONCE(arm64_rmm_active, false);\n+\t\treturn ret;\n+\t}\n+\n+\t/* Mark the RMM is usable */\n+\tWRITE_ONCE(arm64_rmi_is_available, true);\n+\tpr_info(\"RMI configured\\n\");\n+\n+\treturn 0;\n+}\n+\n+/*\n+ * Note arm64_init_rmi() must be called before kvm_init_rmi() otherwise KVM\n+ * will not support realm guests. subsys_initcall() is called before\n+ * module_init() (used for KVM) so this is OK.\n+ */\n+subsys_initcall(arm64_init_rmi);\ndiff --git a/include/linux/arm-rmi-cmds.h b/include/linux/arm-rmi-cmds.h\nnew file mode 100644\nindex 0000000000000..e3503aba646e2\n--- /dev/null\n+++ b/include/linux/arm-rmi-cmds.h\n@@ -0,0 +1,578 @@\n+/* SPDX-License-Identifier: GPL-2.0 */\n+/*\n+ * Copyright (C) 2026 ARM Ltd.\n+ */\n+\n+#ifndef __LINUX_ARM_RMI_CMDS_H_\n+#define __LINUX_ARM_RMI_CMDS_H_\n+\n+#include \u003clinux/arm-smccc-rmi.h\u003e\n+#include \u003clinux/bug.h\u003e\n+#include \u003clinux/gfp.h\u003e\n+#include \u003clinux/processor.h\u003e\n+#include \u003clinux/string.h\u003e\n+#include \u003clinux/types.h\u003e\n+\n+#define RMI_MAX_ADDR_LIST\t256\n+\n+struct rmi_sro_state {\n+\tstruct arm_smccc_1_2_regs regs;\n+\tint addr_count;\n+\tunsigned long addr_list[RMI_MAX_ADDR_LIST];\n+};\n+\n+#define RMM_BLOCKED_RETRY_COUNT\t\t2\n+/*\n+ * rmi_smccc_invoke: Invoke the RMI call and return the results, retrying the\n+ * command when status is RMI_BUSY. If we encounter RMI_BLOCKED, we retry\n+ * it one more time before we give up. The caller is supposed to handle the\n+ * result and reissue if required.\n+ *\n+ * We don't expect to see RMI_BLOCKED on a practical system, except when\n+ * there are parallel requests that results in long standing operation,\n+ * with one blocking the other.\n+ *\n+ * @regs: Input parameters filled in. Updated with the ouptput results\n+ * after the call.\n+ */\n+static inline void rmi_smccc_invoke(struct arm_smccc_1_2_regs *regs)\n+{\n+\tstruct arm_smccc_1_2_regs args = *regs;\n+\tlong status;\n+\tint i = 0;\n+\n+\twhile (i \u003c RMM_BLOCKED_RETRY_COUNT) {\n+\t\tarm_smccc_1_2_invoke(\u0026args, regs);\n+\n+\t\tstatus = RMI_RESULT_STATUS(regs-\u003ea0);\n+\t\tif (status != RMI_BUSY \u0026\u0026 status != RMI_BLOCKED)\n+\t\t\tbreak;\n+\t\tif (status == RMI_BLOCKED)\n+\t\t\ti++;\n+\t\tcpu_relax();\n+\t}\n+}\n+\n+/*\n+ * rmi_smccc_invoke_once: Invoke the RMI call and return the results. Do not\n+ * retry the command. Let the caller deal with RMI_BUSY or RMI_BLOCKED.\n+ */\n+static inline void rmi_smccc_invoke_once(struct arm_smccc_1_2_regs *regs)\n+{\n+\tstruct arm_smccc_1_2_regs args = *regs;\n+\n+\tarm_smccc_1_2_invoke(\u0026args, regs);\n+}\n+\n+unsigned long rmi_feat_reg(unsigned int index);\n+\n+int rmi_delegate_range(phys_addr_t phys, unsigned long size,\n+\t\t       phys_addr_t *out_phys);\n+int rmi_undelegate_range(phys_addr_t phys, unsigned long size);\n+int rmi_free_delegated_page(phys_addr_t phys);\n+\n+static inline int rmi_delegate_page(phys_addr_t phys)\n+{\n+\treturn rmi_delegate_range(phys, PAGE_SIZE, NULL);\n+}\n+\n+static inline int rmi_undelegate_page(phys_addr_t phys)\n+{\n+\treturn rmi_undelegate_range(phys, PAGE_SIZE);\n+}\n+\n+long rmi_sro_memxfer_execute(struct rmi_sro_state *sro, gfp_t gfp);\n+void rmi_sro_free(struct rmi_sro_state *sro);\n+long rmi_sro_execute(struct arm_smccc_1_2_regs *regs);\n+\n+/*\n+ * Resetting the addr_count is sufficient to ignore the addr_list contents.\n+ */\n+#define rmi_sro_memxfer_cmd(sro, gfp, ...) ({\t\t\t\t\\\n+\tstruct rmi_sro_state *__sro = (sro);\t\t\t\t\\\n+\t__sro-\u003eaddr_count = 0;\t\t\t\t\t\t\\\n+\t__sro-\u003eregs = (struct arm_smccc_1_2_regs){ __VA_ARGS__ };\t\\\n+\tlong __ret = rmi_sro_memxfer_execute(__sro, gfp);\t\t\\\n+\trmi_sro_free(__sro);\t\t\t\t\t\t\\\n+\t__ret;\t\t\t\t\t\t\t\t\\\n+})\n+\n+#ifdef CONFIG_ARM_RMM_RMI\n+\n+bool is_rmm_active(void);\n+bool is_rmi_available(void);\n+\n+#else\n+\n+static inline bool is_rmm_active(void)\n+{\n+\treturn false;\n+}\n+\n+static inline bool is_rmi_available(void)\n+{\n+\treturn false;\n+}\n+#endif\t/* CONFIG_ARM_RMM_RMI */\n+\n+\n+/**\n+ * rmi_rtt_data_map_init() - Create a mapping at protected IPA, copying contents\n+ *\t\t\t     from a given non-secure source granule.\n+ * @rd: PA of the RD\n+ * @data: PA of the target granule mapped in the guest\n+ * @ipa: IPA at which the granule @data will be mapped in the guest\n+ * @src: PA of the source granule with contents\n+ * @flags: RMI_MEASURE_CONTENT if the contents should be measured\n+ *\n+ * Create a mapping from Protected IPA space to conventional memory, copying\n+ * contents from a Non-secure Granule provided by the caller.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_data_map_init(unsigned long rd, unsigned long data,\n+\t\t\t\t\t unsigned long ipa, unsigned long src,\n+\t\t\t\t\t unsigned long flags)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_DATA_MAP_INIT, rd, data, ipa, src, flags\n+\t};\n+\n+\treturn rmi_sro_execute(\u0026regs);\n+}\n+\n+/**\n+ * rmi_rtt_data_map() - Create mappings in protected IPA range with unknown contents\n+ * @rd: PA of the RD\n+ * @base: Base of the target IPA range\n+ * @top: Top of the target IPA range\n+ * @flags: Flags\n+ * @oaddr: Output address set descriptor\n+ * @out_top: Top address of range which was processed.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_data_map(unsigned long rd,\n+\t\t\t\t    unsigned long base,\n+\t\t\t\t    unsigned long top,\n+\t\t\t\t    unsigned long flags,\n+\t\t\t\t    unsigned long oaddr,\n+\t\t\t\t    unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_DATA_MAP, rd, base, top, flags, oaddr\n+\t};\n+\tlong ret;\n+\n+\tret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_top)\n+\t\t*out_top = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+/**\n+ * rmi_rtt_data_unmap() - Remove mappings to conventional memory at a protected\n+ *\t\t\t  IPA range\n+ * @rd: PA of the RD\n+ * @base: Base of the target IPA range\n+ * @top: Top of the target IPA range\n+ * @flags: Flags\n+ * @oaddr: Output address set descriptor\n+ * @out_top: Returns top IPA of range which has been unmapped\n+ * @out_range: Output address range\n+ * @out_count: Number of entries in output address list\n+ *\n+ * Removes mappings to convention memory with a target Protected IPA range.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_data_unmap(unsigned long rd,\n+\t\t\t\t      unsigned long base,\n+\t\t\t\t      unsigned long top,\n+\t\t\t\t      unsigned long flags,\n+\t\t\t\t      unsigned long oaddr,\n+\t\t\t\t      unsigned long *out_top,\n+\t\t\t\t      unsigned long *out_range,\n+\t\t\t\t      unsigned long *out_count)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_DATA_UNMAP, rd, base, top, flags, oaddr\n+\t};\n+\tlong ret;\n+\n+\tret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret != RMI_SUCCESS)\n+\t\treturn ret;\n+\n+\tif (out_top)\n+\t\t*out_top = regs.a1;\n+\tif (out_range)\n+\t\t*out_range = regs.a2;\n+\tif (out_count)\n+\t\t*out_count = regs.a3;\n+\n+\treturn RMI_SUCCESS;\n+}\n+\n+/**\n+ * rmi_psci_complete() - Complete pending PSCI command\n+ * @calling_rec: PA of the calling REC\n+ * @status: Status of the PSCI request\n+ *\n+ * Completes a pending PSCI command.\n+ *\n+ * Return: RMI return code\n+ */\n+static inline long rmi_psci_complete(unsigned long calling_rec,\n+\t\t\t\t     unsigned long status)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_PSCI_COMPLETE, calling_rec, status,\n+\t};\n+\n+\trmi_smccc_invoke(\u0026regs);\n+\treturn regs.a0;\n+}\n+\n+/**\n+ * rmi_realm_activate() - Activate a realm\n+ * @rd: PA of the RD\n+ *\n+ * Mark a realm as Active, signalling that creation is completed, allowing\n+ * execution of the realm.\n+ *\n+ * Return: RMI return code\n+ */\n+static inline long rmi_realm_activate(unsigned long rd)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_REALM_ACTIVATE, rd,\n+\t};\n+\n+\trmi_smccc_invoke(\u0026regs);\n+\treturn regs.a0;\n+}\n+\n+/**\n+ * rmi_realm_create() - Create a realm\n+ * @rd: PA of the RD\n+ * @params: PA of realm parameters\n+ * @sro: Preallocated SRO context\n+ *\n+ * Create a new realm using the given parameters.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_realm_create(unsigned long rd, unsigned long params,\n+\t\t\t\t    struct rmi_sro_state *sro)\n+{\n+\treturn rmi_sro_memxfer_cmd(sro, GFP_KERNEL,\n+\t\t\t\t   SMC_RMI_REALM_CREATE, rd, params);\n+}\n+\n+/**\n+ * rmi_realm_terminate() - Terminate a realm\n+ * @rd: PA of the RD\n+ * @sro: Preallocated SRO context\n+ *\n+ * Terminates a realm, moving it into a ZOMBIE state\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_realm_terminate(unsigned long rd,\n+\t\t\t\t       struct rmi_sro_state *sro)\n+{\n+\treturn rmi_sro_memxfer_cmd(sro, GFP_KERNEL,\n+\t\t\t\t   SMC_RMI_REALM_TERMINATE, rd);\n+}\n+\n+/**\n+ * rmi_realm_destroy() - Destroy a realm\n+ * @rd: PA of the RD\n+ * @sro: Preallocated SRO context\n+ *\n+ * Destroys a realm, all objects belonging to the realm must be destroyed first.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_realm_destroy(unsigned long rd,\n+\t\t\t\t     struct rmi_sro_state *sro)\n+{\n+\treturn rmi_sro_memxfer_cmd(sro, GFP_KERNEL,\n+\t\t\t\t   SMC_RMI_REALM_DESTROY, rd);\n+}\n+\n+/**\n+ * rmi_rec_create() - Create a REC\n+ * @rd: PA of the RD\n+ * @rec: PA of the target REC\n+ * @params: PA of REC parameters\n+ * @sro: Allocated SRO context to be used\n+ *\n+ * Create a REC using the parameters specified in the struct rec_params pointed\n+ * to by @params.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rec_create(unsigned long rd,\n+\t\t\t\t  unsigned long rec,\n+\t\t\t\t  unsigned long params,\n+\t\t\t\t  struct rmi_sro_state *sro)\n+{\n+\treturn rmi_sro_memxfer_cmd(sro, GFP_KERNEL,\n+\t\t\t\t   SMC_RMI_REC_CREATE, rd, rec, params);\n+}\n+\n+/**\n+ * rmi_rec_destroy() - Destroy a REC\n+ * @rec: PA of the target REC\n+ * @sro: Allocated SRO context to be used\n+ *\n+ * Destroys a REC. The REC must not be running.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rec_destroy(unsigned long rec,\n+\t\t\t\t   struct rmi_sro_state *sro)\n+{\n+\treturn rmi_sro_memxfer_cmd(sro, GFP_KERNEL, SMC_RMI_REC_DESTROY, rec);\n+}\n+\n+/**\n+ * rmi_rec_enter() - Enter a REC\n+ * @rec: PA of the target REC\n+ * @run_ptr: PA of RecRun structure\n+ *\n+ * Starts (or continues) execution within a REC.\n+ *\n+ * Return: RMI return code\n+ */\n+static inline long rmi_rec_enter(unsigned long rec, unsigned long run_ptr)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_REC_ENTER, rec, run_ptr,\n+\t};\n+\n+\trmi_smccc_invoke_once(\u0026regs);\n+\treturn regs.a0;\n+}\n+\n+/**\n+ * rmi_rtt_create() - Creates an RTT\n+ * @rd: PA of the RD\n+ * @rtt: PA of the target RTT\n+ * @ipa: Base of the IPA range described by the RTT\n+ * @level: Depth of the RTT within the tree\n+ *\n+ * Creates an RTT (Realm Translation Table) at the specified level for the\n+ * translation of the specified address within the realm.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_create(unsigned long rd, unsigned long rtt,\n+\t\t\t\t  unsigned long ipa, long level)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_CREATE, rd, rtt, ipa, level\n+\t};\n+\n+\treturn rmi_sro_execute(\u0026regs);\n+}\n+\n+/**\n+ * rmi_rtt_destroy() - Destroy an RTT\n+ * @rd: PA of the RD\n+ * @ipa: Base of the IPA range described by the RTT\n+ * @level: RTT level\n+ * @out_rtt: Pointer to write the PA of the RTT which was destroyed\n+ * @out_top: Pointer to write the top IPA of non-live RTT entries, from entry\n+ * at which the RTT walk terminated.\n+ *\n+ * Destroys an RTT. The RTT must be non-live, i.e. none of the entries in the\n+ * table are in ASSIGNED or TABLE state.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code.\n+ */\n+static inline long rmi_rtt_destroy(unsigned long rd,\n+\t\t\t\t   unsigned long ipa,\n+\t\t\t\t   long level,\n+\t\t\t\t   unsigned long *out_rtt,\n+\t\t\t\t   unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_DESTROY, rd, ipa, level\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tswitch (RMI_RESULT_STATUS(ret)) {\n+\tcase RMI_SUCCESS:\n+\t\tif (out_rtt)\n+\t\t\t*out_rtt = regs.a1;\n+\t\tfallthrough;\n+\tcase RMI_ERROR_RTT:\n+\t\tif (out_top)\n+\t\t\t*out_top = regs.a2;\n+\t\tbreak;\n+\tdefault:\n+\t\tbreak;\n+\t}\n+\n+\treturn ret;\n+}\n+\n+/**\n+ * rmi_rtt_fold() - Fold an RTT\n+ * @rd: PA of the RD\n+ * @ipa: Base of the IPA range described by the RTT\n+ * @level: Depth of the RTT within the tree\n+ * @out_rtt: Pointer to write the PA of the RTT which was destroyed\n+ *\n+ * Folds an RTT. If all entries with the RTT are 'homogeneous' the RTT can be\n+ * folded into the parent and the RTT destroyed.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_fold(unsigned long rd, unsigned long ipa,\n+\t\t\t\tlong level, unsigned long *out_rtt)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_FOLD, rd, ipa, level\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_rtt)\n+\t\t*out_rtt = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+/**\n+ * rmi_rtt_init_ripas() - Set RIPAS for new realm\n+ * @rd: PA of the RD\n+ * @base: Base of target IPA region\n+ * @top: Top of target IPA region\n+ * @out_top: Top IPA of range whose RIPAS was modified\n+ *\n+ * Sets the RIPAS of a target IPA range to RAM, for a realm in the NEW state.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_init_ripas(unsigned long rd, unsigned long base,\n+\t\t\t\t      unsigned long top, unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_INIT_RIPAS, rd, base, top\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_top)\n+\t\t*out_top = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+/**\n+ * rmi_rtt_unprot_map() - Map unprotected granules into a realm\n+ * @rd: PA of the RD\n+ * @base: Base IPA of the mapping\n+ * @top: Top of the target IPA range\n+ * @flags: Flags\n+ * @oaddr: Output address set descriptor\n+ * @out_top: Top IPA of range which has been mapped\n+ *\n+ * Create mappings to memory within a target unprotected IPA range.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_unprot_map(unsigned long rd,\n+\t\t\t\t      unsigned long base,\n+\t\t\t\t      unsigned long top,\n+\t\t\t\t      unsigned long flags,\n+\t\t\t\t      unsigned long oaddr,\n+\t\t\t\t      unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_UNPROT_MAP, rd, base, top, flags, oaddr\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_top)\n+\t\t*out_top = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+/**\n+ * rmi_rtt_set_ripas() - Set RIPAS for an running realm\n+ * @rd: PA of the RD\n+ * @rec: PA of the REC making the request\n+ * @base: Base of target IPA region\n+ * @top: Top of target IPA region\n+ * @out_top: Pointer to write top IPA of range whose RIPAS was modified\n+ *\n+ * Completes a request made by the realm to change the RIPAS of a target IPA\n+ * range.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_set_ripas(unsigned long rd, unsigned long rec,\n+\t\t\t\t     unsigned long base, unsigned long top,\n+\t\t\t\t     unsigned long *out_top)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_SET_RIPAS, rd, rec, base, top\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret == RMI_SUCCESS \u0026\u0026 out_top)\n+\t\t*out_top = regs.a1;\n+\n+\treturn ret;\n+}\n+\n+/**\n+ * rmi_rtt_unprot_unmap() - Remove mappings within an unprotected IPA range\n+ * @rd: PA of the RD\n+ * @base: Base IPA of the mapping\n+ * @top: Top of the target IPA range\n+ * @flags: Flags\n+ * @oaddr: Output address set descriptor\n+ * @out_top: Top IPA which has been unmapped\n+ * @out_range: Output address range\n+ * @out_count: Number of entries in output address list\n+ *\n+ * Removes mappings to memory within a target unprotected IPA range.\n+ *\n+ * Return: 0 on success, positive RMI result code or negative Linux error code\n+ */\n+static inline long rmi_rtt_unprot_unmap(unsigned long rd,\n+\t\t\t\t\tunsigned long base,\n+\t\t\t\t\tunsigned long top,\n+\t\t\t\t\tunsigned long flags,\n+\t\t\t\t\tunsigned long oaddr,\n+\t\t\t\t\tunsigned long *out_top,\n+\t\t\t\t\tunsigned long *out_range,\n+\t\t\t\t\tunsigned long *out_count)\n+{\n+\tstruct arm_smccc_1_2_regs regs = {\n+\t\tSMC_RMI_RTT_UNPROT_UNMAP, rd, base, top, flags, oaddr\n+\t};\n+\tlong ret = rmi_sro_execute(\u0026regs);\n+\n+\tif (ret != RMI_SUCCESS)\n+\t\treturn ret;\n+\n+\tif (out_top)\n+\t\t*out_top = regs.a1;\n+\tif (out_range)\n+\t\t*out_range = regs.a2;\n+\tif (out_count)\n+\t\t*out_count = regs.a3;\n+\n+\treturn RMI_SUCCESS;\n+}\n+\n+#endif\t/* __LINUX_ARM_RMI_CMDS_H_ */\ndiff --git a/include/linux/arm-smccc-rmi.h b/include/linux/arm-smccc-rmi.h\nnew file mode 100644\nindex 0000000000000..655726b374624\n--- /dev/null\n+++ b/include/linux/arm-smccc-rmi.h\n@@ -0,0 +1,516 @@\n+/* SPDX-License-Identifier: GPL-2.0 */\n+/*\n+ * Copyright (C) 2023-2026 ARM Ltd.\n+ *\n+ * The values and structures in this file are from the Realm Management Monitor\n+ * specification (DEN0137) version 2.0-bet3:\n+ * https://developer.arm.com/documentation/den0137/2-0bet3/\n+ */\n+\n+#ifndef __LINUX_ARM_SMCCC_RMI_H_\n+#define __LINUX_ARM_SMCCC_RMI_H_\n+\n+#include \u003clinux/arm-smccc.h\u003e\n+#include \u003clinux/bitfield.h\u003e\n+#include \u003clinux/bits.h\u003e\n+#include \u003clinux/build_bug.h\u003e\n+#include \u003clinux/sizes.h\u003e\n+#include \u003clinux/stddef.h\u003e\n+\n+#include \u003casm/page.h\u003e\n+\n+#define SMC_RMI_CALL(func)\t\t\t\t\\\n+\tARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,\t\t\\\n+\t\t\t   ARM_SMCCC_SMC_64,\t\t\\\n+\t\t\t   ARM_SMCCC_OWNER_STANDARD,\t\\\n+\t\t\t   (func))\n+\n+#define SMC_RMI_VERSION\t\t\t\tSMC_RMI_CALL(0x0150)\n+\n+#define SMC_RMI_RTT_DATA_MAP_INIT\t\tSMC_RMI_CALL(0x0153)\n+\n+#define SMC_RMI_REALM_ACTIVATE\t\t\tSMC_RMI_CALL(0x0157)\n+#define SMC_RMI_REALM_CREATE\t\t\tSMC_RMI_CALL(0x0158)\n+#define SMC_RMI_REALM_DESTROY\t\t\tSMC_RMI_CALL(0x0159)\n+#define SMC_RMI_REC_CREATE\t\t\tSMC_RMI_CALL(0x015a)\n+#define SMC_RMI_REC_DESTROY\t\t\tSMC_RMI_CALL(0x015b)\n+#define SMC_RMI_REC_ENTER\t\t\tSMC_RMI_CALL(0x015c)\n+#define SMC_RMI_RTT_CREATE\t\t\tSMC_RMI_CALL(0x015d)\n+#define SMC_RMI_RTT_DESTROY\t\t\tSMC_RMI_CALL(0x015e)\n+\n+#define SMC_RMI_RTT_READ_ENTRY\t\t\tSMC_RMI_CALL(0x0161)\n+\n+#define SMC_RMI_RTT_DEV_VALIDATE\t\tSMC_RMI_CALL(0x0163)\n+#define SMC_RMI_PSCI_COMPLETE\t\t\tSMC_RMI_CALL(0x0164)\n+#define SMC_RMI_FEATURES\t\t\tSMC_RMI_CALL(0x0165)\n+#define SMC_RMI_RTT_FOLD\t\t\tSMC_RMI_CALL(0x0166)\n+\n+#define SMC_RMI_RTT_INIT_RIPAS\t\t\tSMC_RMI_CALL(0x0168)\n+#define SMC_RMI_RTT_SET_RIPAS\t\t\tSMC_RMI_CALL(0x0169)\n+#define SMC_RMI_VSMMU_CREATE\t\t\tSMC_RMI_CALL(0x016a)\n+#define SMC_RMI_VSMMU_DESTROY\t\t\tSMC_RMI_CALL(0x016b)\n+\n+#define SMC_RMI_RMM_CONFIG_SET\t\t\tSMC_RMI_CALL(0x016e)\n+#define SMC_RMI_PSMMU_IRQ_NOTIFY\t\tSMC_RMI_CALL(0x016f)\n+#define SMC_RMI_ATTEST_PLAT_TOKEN_REFRESH\tSMC_RMI_CALL(0x0170)\n+\n+#define SMC_RMI_PDEV_ABORT\t\t\tSMC_RMI_CALL(0x0174)\n+#define SMC_RMI_PDEV_COMMUNICATE\t\tSMC_RMI_CALL(0x0175)\n+#define SMC_RMI_PDEV_CREATE\t\t\tSMC_RMI_CALL(0x0176)\n+#define SMC_RMI_PDEV_DESTROY\t\t\tSMC_RMI_CALL(0x0177)\n+#define SMC_RMI_PDEV_GET_STATE\t\t\tSMC_RMI_CALL(0x0178)\n+\n+#define SMC_RMI_PDEV_STREAM_KEY_REFRESH\t\tSMC_RMI_CALL(0x017a)\n+#define SMC_RMI_PDEV_SET_PUBKEY\t\t\tSMC_RMI_CALL(0x017b)\n+#define SMC_RMI_PDEV_STOP\t\t\tSMC_RMI_CALL(0x017c)\n+#define SMC_RMI_RTT_AUX_CREATE\t\t\tSMC_RMI_CALL(0x017d)\n+#define SMC_RMI_RTT_AUX_DESTROY\t\t\tSMC_RMI_CALL(0x017e)\n+#define SMC_RMI_RTT_AUX_FOLD\t\t\tSMC_RMI_CALL(0x017f)\n+\n+#define SMC_RMI_VDEV_ABORT\t\t\tSMC_RMI_CALL(0x0185)\n+#define SMC_RMI_VDEV_COMMUNICATE\t\tSMC_RMI_CALL(0x0186)\n+#define SMC_RMI_VDEV_CREATE\t\t\tSMC_RMI_CALL(0x0187)\n+#define SMC_RMI_VDEV_DESTROY\t\t\tSMC_RMI_CALL(0x0188)\n+#define SMC_RMI_VDEV_GET_STATE\t\t\tSMC_RMI_CALL(0x0189)\n+#define SMC_RMI_VDEV_UNLOCK\t\t\tSMC_RMI_CALL(0x018a)\n+#define SMC_RMI_RTT_SET_S2AP\t\t\tSMC_RMI_CALL(0x018b)\n+\n+#define SMC_RMI_VDEV_GET_INTERFACE_REPORT\tSMC_RMI_CALL(0x01d0)\n+#define SMC_RMI_VDEV_GET_MEASUREMENTS\t\tSMC_RMI_CALL(0x01d1)\n+#define SMC_RMI_VDEV_LOCK\t\t\tSMC_RMI_CALL(0x01d2)\n+#define SMC_RMI_VDEV_START\t\t\tSMC_RMI_CALL(0x01d3)\n+\n+#define SMC_RMI_VSMMU_EVENT_HANDLE\t\tSMC_RMI_CALL(0x01d6)\n+#define SMC_RMI_PSMMU_ACTIVATE\t\t\tSMC_RMI_CALL(0x01d7)\n+#define SMC_RMI_PSMMU_DEACTIVATE\t\tSMC_RMI_CALL(0x01d8)\n+\n+#define SMC_RMI_PSMMU_ST_L2_CREATE\t\tSMC_RMI_CALL(0x01db)\n+#define SMC_RMI_PSMMU_ST_L2_DESTROY\t\tSMC_RMI_CALL(0x01dc)\n+#define SMC_RMI_DPT_L0_CREATE\t\t\tSMC_RMI_CALL(0x01dd)\n+#define SMC_RMI_DPT_L0_DESTROY\t\t\tSMC_RMI_CALL(0x01de)\n+#define SMC_RMI_DPT_L1_CREATE\t\t\tSMC_RMI_CALL(0x01df)\n+#define SMC_RMI_DPT_L1_DESTROY\t\t\tSMC_RMI_CALL(0x01e0)\n+#define SMC_RMI_GRANULE_TRACKING_GET\t\tSMC_RMI_CALL(0x01e1)\n+\n+#define SMC_RMI_GRANULE_TRACKING_SET\t\tSMC_RMI_CALL(0x01e3)\n+\n+#define SMC_RMI_RMM_CONFIG_GET\t\t\tSMC_RMI_CALL(0x01ec)\n+\n+#define SMC_RMI_RMM_STATE_GET\t\t\tSMC_RMI_CALL(0x01ee)\n+\n+#define SMC_RMI_PSMMU_EVENT_CONSUME\t\tSMC_RMI_CALL(0x01f0)\n+#define SMC_RMI_GRANULE_RANGE_DELEGATE\t\tSMC_RMI_CALL(0x01f1)\n+#define SMC_RMI_GRANULE_RANGE_UNDELEGATE\tSMC_RMI_CALL(0x01f2)\n+#define SMC_RMI_GPT_L1_CREATE\t\t\tSMC_RMI_CALL(0x01f3)\n+#define SMC_RMI_GPT_L1_DESTROY\t\t\tSMC_RMI_CALL(0x01f4)\n+#define SMC_RMI_RTT_DATA_MAP\t\t\tSMC_RMI_CALL(0x01f5)\n+#define SMC_RMI_RTT_DATA_UNMAP\t\t\tSMC_RMI_CALL(0x01f6)\n+#define SMC_RMI_RTT_DEV_MAP\t\t\tSMC_RMI_CALL(0x01f7)\n+#define SMC_RMI_RTT_DEV_UNMAP\t\t\tSMC_RMI_CALL(0x01f8)\n+#define SMC_RMI_RTT_ARCH_DEV_MAP\t\tSMC_RMI_CALL(0x01f9)\n+#define SMC_RMI_RTT_ARCH_DEV_UNMAP\t\tSMC_RMI_CALL(0x01fa)\n+#define SMC_RMI_RTT_UNPROT_MAP\t\t\tSMC_RMI_CALL(0x01fb)\n+#define SMC_RMI_RTT_UNPROT_UNMAP\t\tSMC_RMI_CALL(0x01fc)\n+#define SMC_RMI_RTT_AUX_PROT_MAP\t\tSMC_RMI_CALL(0x01fd)\n+#define SMC_RMI_RTT_AUX_PROT_UNMAP\t\tSMC_RMI_CALL(0x01fe)\n+#define SMC_RMI_RTT_AUX_UNPROT_MAP\t\tSMC_RMI_CALL(0x01ff)\n+#define SMC_RMI_RTT_AUX_UNPROT_UNMAP\t\tSMC_RMI_CALL(0x0200)\n+#define SMC_RMI_REALM_TERMINATE\t\t\tSMC_RMI_CALL(0x0201)\n+#define SMC_RMI_RMM_ACTIVATE\t\t\tSMC_RMI_CALL(0x0202)\n+#define SMC_RMI_OP_CONTINUE\t\t\tSMC_RMI_CALL(0x0203)\n+#define SMC_RMI_PDEV_STREAM_CONNECT\t\tSMC_RMI_CALL(0x0204)\n+#define SMC_RMI_PDEV_STREAM_DISCONNECT\t\tSMC_RMI_CALL(0x0205)\n+#define SMC_RMI_PDEV_STREAM_COMPLETE\t\tSMC_RMI_CALL(0x0206)\n+#define SMC_RMI_PDEV_STREAM_KEY_PURGE\t\tSMC_RMI_CALL(0x0207)\n+#define SMC_RMI_OP_MEM_DONATE\t\t\tSMC_RMI_CALL(0x0208)\n+#define SMC_RMI_OP_MEM_RECLAIM\t\t\tSMC_RMI_CALL(0x0209)\n+#define SMC_RMI_OP_CANCEL\t\t\tSMC_RMI_CALL(0x020a)\n+#define SMC_RMI_VSMMU_FEATURES\t\t\tSMC_RMI_CALL(0x020b)\n+#define SMC_RMI_VSMMU_CMD_GET\t\t\tSMC_RMI_CALL(0x020c)\n+#define SMC_RMI_VSMMU_CMD_COMPLETE\t\tSMC_RMI_CALL(0x020d)\n+#define SMC_RMI_PSMMU_INFO\t\t\tSMC_RMI_CALL(0x020e)\n+#define SMC_RMI_RMM_DEACTIVATE\t\t\tSMC_RMI_CALL(0x020f)\n+#define SMC_RMI_PDEV_STREAM_INFO\t\tSMC_RMI_CALL(0x0210)\n+#define SMC_RMI_GPT_INFO\t\t\tSMC_RMI_CALL(0x0211)\n+\n+#define RMI_ABI_MAJOR_VERSION\t2\n+#define RMI_ABI_MINOR_VERSION\t0\n+\n+#define RMI_ABI_VERSION_MAJOR_MASK\t\tGENMASK(30, 16)\n+#define RMI_ABI_VERSION_MINOR_MASK\t\tGENMASK(15, 0)\n+\n+#define RMI_ABI_VERSION_GET_MAJOR(v)\tFIELD_GET(RMI_ABI_VERSION_MAJOR_MASK, (v))\n+#define RMI_ABI_VERSION_GET_MINOR(v)\tFIELD_GET(RMI_ABI_VERSION_MINOR_MASK, (v))\n+#define RMI_ABI_VERSION(major, minor)\t\t\t\t\\\n+\t(FIELD_PREP(RMI_ABI_VERSION_MAJOR_MASK, (major)) |\t\\\n+\t FIELD_PREP(RMI_ABI_VERSION_MINOR_MASK, (minor)))\n+\n+/* RmiResult Type definitions */\n+#define RMI_RESULT_STATUS_MASK\t\tGENMASK(7, 0)\n+/* RmiResultDataLevel field */\n+#define RMI_RESULT_DATA_LEVEL_MASK\tGENMASK(15, 8)\n+/* RmiResultDataIncomplete fields */\n+#define RMI_RESULT_MEMREQ_MASK\t\tGENMASK(9, 8)\n+#define RMI_RESULT_CAN_CANCEL_MASK\tBIT(10)\n+\n+#define RMI_RESULT_STATUS(ret)\t\tFIELD_GET(RMI_RESULT_STATUS_MASK, ret)\n+#define RMI_RESULT_DATA_LEVEL(ret)\tFIELD_GET(RMI_RESULT_DATA_LEVEL_MASK, ret)\n+#define RMI_RESULT_MEMREQ(ret)\t\tFIELD_GET(RMI_RESULT_MEMREQ_MASK, ret)\n+#define RMI_RESULT_CAN_CANCEL(ret)\tFIELD_GET(RMI_RESULT_CAN_CANCEL_MASK, ret)\n+\n+#define RMI_SUCCESS\t\t\t0\n+#define RMI_ERROR_INPUT\t\t\t1\n+#define RMI_ERROR_REALM\t\t\t2\n+#define RMI_ERROR_REC\t\t\t3\n+#define RMI_ERROR_RTT\t\t\t4\n+#define RMI_ERROR_NOT_SUPPORTED\t\t5\n+#define RMI_ERROR_DEVICE\t\t6\n+#define RMI_ERROR_RTT_AUX\t\t7\n+#define RMI_ERROR_PSMMU_ST\t\t8\n+#define RMI_ERROR_DPT\t\t\t9\n+#define RMI_BUSY\t\t\t10\n+#define RMI_ERROR_GLOBAL\t\t11\n+#define RMI_ERROR_TRACKING\t\t12\n+#define RMI_INCOMPLETE\t\t\t13\n+#define RMI_BLOCKED\t\t\t14\n+#define RMI_ERROR_GPT\t\t\t15\n+#define RMI_ERROR_GRANULE\t\t16\n+\n+#define RMI_CONTINUE_KEEP_GOING\t\t0\n+#define RMI_CONTINUE_STOP\t\t1\n+\n+#define RMI_OP_MEM_REQ_NONE\t\t0\n+#define RMI_OP_MEM_REQ_DONATE\t\t1\n+#define RMI_OP_MEM_REQ_RECLAIM\t\t2\n+\n+#define RMI_OP_CANNOT_CANCEL\t\t0\n+#define RMI_OP_CAN_CANCEL\t\t1\n+\n+#define RMI_DONATE_STATE_MASK\t\tGENMASK(18, 17)\n+#define RMI_DONATE_CONTIG_MASK\t\tBIT(16)\n+#define RMI_DONATE_COUNT_MASK\t\tGENMASK(15, 2)\n+#define RMI_DONATE_BLOCK_SIZE_MASK\tGENMASK(1, 0)\n+\n+#define RMI_DONATE_STATE(req)\t\tFIELD_GET(RMI_DONATE_STATE_MASK, req)\n+#define RMI_DONATE_CONTIG(req)\t\tFIELD_GET(RMI_DONATE_CONTIG_MASK, req)\n+#define RMI_DONATE_COUNT(req)\t\tFIELD_GET(RMI_DONATE_COUNT_MASK, req)\n+#define RMI_DONATE_BLOCK_SIZE(req)\tFIELD_GET(RMI_DONATE_BLOCK_SIZE_MASK, req)\n+\n+#define RMI_OP_MEM_DELEGATED\t\t0\n+#define RMI_OP_MEM_UNDELEGATED\t\t1\n+#define RMI_OP_MEM_CONDITIONAL\t\t2\n+\n+#define RMI_OP_MEM_NON_CONTIG\t\t0\n+#define RMI_OP_MEM_CONTIG\t\t1\n+\n+#define RMI_ADDR_TYPE_NONE\t\t0\n+#define RMI_ADDR_TYPE_SINGLE\t\t1\n+#define RMI_ADDR_TYPE_LIST\t\t2\n+\n+#define RMI_ADDR_RANGE_STATE_MASK\tGENMASK_ULL(63, 62)\n+#define RMI_ADDR_RANGE_ADDR_MASK\tGENMASK_ULL(51, PAGE_SHIFT)\n+#define RMI_ADDR_RANGE_COUNT_MASK\tGENMASK(PAGE_SHIFT - 1, 2)\n+#define RMI_ADDR_RANGE_BLOCK_SIZE_MASK\tGENMASK(1, 0)\n+\n+#define RMI_ADDR_RANGE_STATE(r)\t\tFIELD_GET(RMI_ADDR_RANGE_STATE_MASK, (r))\n+#define RMI_ADDR_RANGE_ADDR(r)\t\t((r) \u0026 RMI_ADDR_RANGE_ADDR_MASK)\n+#define RMI_ADDR_RANGE_COUNT(r)\t\tFIELD_GET(RMI_ADDR_RANGE_COUNT_MASK, (r))\n+#define RMI_ADDR_RANGE_BLOCK_SIZE(r)\tFIELD_GET(RMI_ADDR_RANGE_BLOCK_SIZE_MASK, (r))\n+\n+enum rmi_ripas {\n+\tRMI_EMPTY = 0,\n+\tRMI_RAM = 1,\n+\tRMI_DESTROYED = 2,\n+\tRMI_DEV = 3,\n+};\n+\n+#define RMI_NO_MEASURE_CONTENT\t0\n+#define RMI_MEASURE_CONTENT\t1\n+\n+#define RMI_FEATURE_REGISTER_0_S2OASZ\t\t\tGENMASK_ULL(40, 33)\n+#define RMI_FEATURE_REGISTER_0_L0GPT_BLOCK_DELEGATE\tBIT_ULL(32)\n+#define RMI_FEATURE_REGISTER_0_PMU_NUM_CTRS\t\tGENMASK(31, 27)\n+#define RMI_FEATURE_REGISTER_0_PMU\t\t\tBIT(26)\n+#define RMI_FEATURE_REGISTER_0_NUM_WPS\t\t\tGENMASK(25, 20)\n+#define RMI_FEATURE_REGISTER_0_NUM_BPS\t\t\tGENMASK(19, 14)\n+#define RMI_FEATURE_REGISTER_0_SVE_VL\t\t\tGENMASK(13, 10)\n+#define RMI_FEATURE_REGISTER_0_SVE\t\t\tBIT(9)\n+#define RMI_FEATURE_REGISTER_0_LPA2\t\t\tBIT(8)\n+#define RMI_FEATURE_REGISTER_0_S2SZ\t\t\tGENMASK(7, 0)\n+\n+#define RMI_FEATURE_REGISTER_1_PPS\t\t\tGENMASK(16, 14)\n+#define RMI_FEATURE_REGISTER_1_L0GPTSZ\t\t\tGENMASK(13, 10)\n+#define RMI_FEATURE_REGISTER_1_MAX_RECS_ORDER\t\tGENMASK(9, 6)\n+#define RMI_FEATURE_REGISTER_1_HASH_SHA_512\t\tBIT(5)\n+#define RMI_FEATURE_REGISTER_1_HASH_SHA_384\t\tBIT(4)\n+#define RMI_FEATURE_REGISTER_1_HASH_SHA_256\t\tBIT(3)\n+#define RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_64KB\t\tBIT(2)\n+#define RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_16KB\t\tBIT(1)\n+#define RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_4KB\t\tBIT(0)\n+\n+#define RMI_FEATURE_REGISTER_2_REALM_MAX_VDEVS_ORDER\tGENMASK(14, 10)\n+#define RMI_FEATURE_REGISTER_2_NON_TEE_STREAM\t\tBIT(9)\n+#define RMI_FEATURE_REGISTER_2_VDEV_KROU\t\tBIT(8)\n+#define RMI_FEATURE_REGISTER_2_PDEV_MAX_VDEVS_ORDER\tGENMASK(7, 4)\n+#define RMI_FEATURE_REGISTER_2_ATS\t\t\tBIT(3)\n+#define RMI_FEATURE_REGISTER_2_VSMMU\t\t\tBIT(2)\n+#define RMI_FEATURE_REGISTER_2_DA_COH\t\t\tBIT(1)\n+#define RMI_FEATURE_REGISTER_2_DA\t\t\tBIT(0)\n+\n+#define RMI_FEATURE_REGISTER_3_RTT_S2AP_INDIRECT\tBIT(6)\n+#define RMI_FEATURE_REGISTER_3_RTT_PLANE\t\tGENMASK(5, 4)\n+#define RMI_FEATURE_REGISTER_3_MAX_NUM_AUX_PLANES\tGENMASK(3, 0)\n+\n+#define RMI_FEATURE_REGISTER_4_MEC_COUNT\t\tGENMASK_ULL(63, 0)\n+\n+#define RMI_MEM_CATEGORY_CONVENTIONAL\t\t0\n+#define RMI_MEM_CATEGORY_DEV_NCOH\t\t1\n+#define RMI_MEM_CATEGORY_DEV_COH\t\t2\n+#define RMI_MEM_CATEGORY_NONE\t\t\t3\n+\n+#define RMI_TRACKING_RESERVED\t\t\t0\n+#define RMI_TRACKING_NONE\t\t\t1\n+#define RMI_TRACKING_FINE\t\t\t2\n+#define RMI_TRACKING_COARSE\t\t\t3\n+#define RMI_TRACKING_INTERMEDIATE\t\t4\n+\n+#define RMI_GRANULE_SIZE_4KB\t0\n+#define RMI_GRANULE_SIZE_16KB\t1\n+#define RMI_GRANULE_SIZE_64KB\t2\n+\n+#define RMI_GPT_PAR_RESERVED           0U\n+#define RMI_GPT_PAR_PLAT               1U\n+#define RMI_GPT_PAR_HOST_NOT_CREATED   2U\n+#define RMI_GPT_PAR_HOST_CREATED       3U\n+\n+/*\n+ * Note many of these fields are smaller than u64 but all fields have u64\n+ * alignment, so use u64 to ensure correct alignment.\n+ */\n+struct rmm_config {\n+\tunion { /* 0x0 */\n+\t\tstruct {\n+\t\t\tu64 tracking_region_size;\n+\t\t\tu64 rmi_granule_size;\n+\t\t};\n+\t\tu8 sizer[SZ_4K];\n+\t};\n+};\n+\n+static_assert(sizeof(struct rmm_config) == SZ_4K);\n+\n+#define RMI_REALM_PARAM_FLAG_MEC_POLICY_MASK\tGENMASK(8, 7)\n+#define RMI_REALM_PARAM_FLAG_LFA_POLICY_MASK\tGENMASK(6, 5)\n+#define RMI_REALM_PARAM_FLAG_DA\t\t\tBIT(3)\n+#define RMI_REALM_PARAM_FLAG_PMU\t\tBIT(2)\n+#define RMI_REALM_PARAM_FLAG_SVE\t\tBIT(1)\n+\n+#define RMI_HASH_SHA_256\t\t\t0\n+#define RMI_HASH_SHA_512\t\t\t1\n+#define RMI_HASH_SHA_384\t\t\t2\n+\n+struct realm_params {\n+\tunion { /* 0x0 */\n+\t\tstruct {\n+\t\t\tu64 flags0;\n+\t\t\tu64 s2sz;\n+\t\t\tu64 sve_vl;\n+\t\t\tu64 num_bps;\n+\t\t\tu64 num_wps;\n+\t\t\tu64 pmu_num_ctrs;\n+\t\t\tu64 hash_algo;\n+\t\t\tu64 num_aux_planes;\n+\t\t};\n+\t\tu8 sizer0[0x400];\n+\t};\n+\tunion { /* 0x400 */\n+\t\tstruct {\n+\t\t\tu8 rpv[64];\n+\t\t\tu64 ats_plane;\n+\t\t};\n+\t\tu8 sizer1[0x400];\n+\t};\n+\tunion { /* 0x800 */\n+\t\tstruct {\n+\t\t\tu64 sizer2;\n+\t\t\tu64 rtt_base;\n+\t\t\ts64 rtt_level_start;\n+\t\t\tu64 rtt_num_start;\n+\t\t\tu64 flags1;\n+\t\t\tu64 max_num_vdevs;\n+\t\t};\n+\t\tu8 sizer3[0x700];\n+\t};\n+\tunion { /* 0xf00 */\n+\t\tstruct {\n+\t\t\tu8 sizer4[0x80];\n+\t\t\tu64 aux_rtt_base[3];\n+\t\t};\n+\t\tu8 sizer5[0x100];\n+\t};\n+};\n+\n+static_assert(offsetof(struct realm_params, rpv) == 0x400);\n+static_assert(offsetof(struct realm_params, rtt_base) == 0x808);\n+static_assert(offsetof(struct realm_params, aux_rtt_base) == 0xf80);\n+static_assert(sizeof(struct realm_params) == SZ_4K);\n+\n+/*\n+ * The number of GPRs (starting from X0) that are configured by the host when\n+ * a REC is created.\n+ */\n+#define REC_CREATE_NR_GPRS\t\t8\n+\n+#define REC_PARAMS_FLAG_RUNNABLE\tBIT(0)\n+\n+struct rec_params {\n+\tunion { /* 0x0 */\n+\t\tu64 flags;\n+\t\tu8 sizer0[0x100];\n+\t};\n+\tunion { /* 0x100 */\n+\t\tu64 mpidr;\n+\t\tu8 sizer1[0x100];\n+\t};\n+\tunion { /* 0x200 */\n+\t\tu64 pc;\n+\t\tu8 sizer2[0x100];\n+\t};\n+\tunion { /* 0x300 */\n+\t\tu64 gprs[REC_CREATE_NR_GPRS];\n+\t\tu8 sizer3[0xd00];\n+\t};\n+};\n+\n+static_assert(offsetof(struct rec_params, mpidr) == 0x100);\n+static_assert(offsetof(struct rec_params, pc) == 0x200);\n+static_assert(offsetof(struct rec_params, gprs) == 0x300);\n+static_assert(sizeof(struct rec_params) == SZ_4K);\n+\n+#define REC_ENTER_FLAG_FORCE_P0\t\tBIT(7)\n+#define REC_ENTER_FLAG_DEV_MEM_RESPONSE\tBIT(6)\n+#define REC_ENTER_FLAG_S2AP_RESPONSE\tBIT(5)\n+#define REC_ENTER_FLAG_RIPAS_RESPONSE\tBIT(4)\n+#define REC_ENTER_FLAG_TRAP_WFE\t\tBIT(3)\n+#define REC_ENTER_FLAG_TRAP_WFI\t\tBIT(2)\n+#define REC_ENTER_FLAG_INJECT_SEA\tBIT(1)\n+#define REC_ENTER_FLAG_EMULATED_MMIO\tBIT(0)\n+\n+#define REC_RUN_GPRS\t\t\t31\n+\n+struct rec_enter {\n+\tunion { /* 0x000 */\n+\t\tu64 flags;\n+\t\tu8 sizer0[0x200];\n+\t};\n+\tunion { /* 0x200 */\n+\t\tu64 gprs[REC_RUN_GPRS];\n+\t\tu8 sizer1[0x600];\n+\t};\n+};\n+\n+static_assert(offsetof(struct rec_enter, gprs) == 0x200);\n+static_assert(sizeof(struct rec_enter) == SZ_2K);\n+\n+#define RMI_EXIT_SYNC\t\t\t0x00\n+#define RMI_EXIT_IRQ\t\t\t0x01\n+#define RMI_EXIT_FIQ\t\t\t0x02\n+#define RMI_EXIT_PSCI\t\t\t0x03\n+#define RMI_EXIT_RIPAS_CHANGE\t\t0x04\n+#define RMI_EXIT_HOST_CALL\t\t0x05\n+#define RMI_EXIT_SERROR\t\t\t0x06\n+#define RMI_EXIT_S2AP_CHANGE\t\t0x07\n+#define RMI_EXIT_VDEV_VALIDATE_MAPPING\t0x08\n+#define RMI_EXIT_VSMMU_COMMAND\t\t0x0a\n+\n+struct rec_exit {\n+\tunion { /* 0x000 */\n+\t\tu8 exit_reason;\n+\t\tu8 sizer0[0x100];\n+\t};\n+\tunion { /* 0x100 */\n+\t\tstruct {\n+\t\t\tu64 esr;\n+\t\t\tu64 far;\n+\t\t\tu64 hpfar;\n+\t\t\tu64 rtt_tree;\n+\t\t};\n+\t\tu8 sizer1[0x100];\n+\t};\n+\tunion { /* 0x200 */\n+\t\tu64 gprs[REC_RUN_GPRS];\n+\t\tu8 sizer2[0x200];\n+\t};\n+\tunion { /* 0x400 */\n+\t\tstruct {\n+\t\t\tu64 cntp_ctl;\n+\t\t\tu64 cntp_cval;\n+\t\t\tu64 cntv_ctl;\n+\t\t\tu64 cntv_cval;\n+\t\t};\n+\t\tu8 sizer4[0x100];\n+\t};\n+\tunion { /* 0x500 */\n+\t\tstruct {\n+\t\t\tu64 ripas_base;\n+\t\t\tu64 ripas_top;\n+\t\t\tu8 ripas_value;\n+\t\t\tu8 sizer5[0xf];\n+\t\t\tu64 s2ap_base;\n+\t\t\tu64 s2ap_top;\n+\t\t\tu64 vdev_id_1;\n+\t\t\tu64 vdev_id_2;\n+\t\t\tu64 dev_mem_base;\n+\t\t\tu64 dev_mem_top;\n+\t\t\tu64 dev_mem_pa;\n+\t\t};\n+\t\tu8 sizer6[0x100];\n+\t};\n+\tunion { /* 0x600 */\n+\t\tstruct {\n+\t\t\tu16 imm;\n+\t\t\tu8 sizer7[0x6];\n+\t\t\tu64 plane;\n+\t\t};\n+\t\tu8 sizer8[0x100];\n+\t};\n+\tunion { /* 0x700 */\n+\t\tstruct {\n+\t\t\tu8 pmu_ovf_status;\n+\t\t\tu8 sizer9[0xf];\n+\t\t\tu64 vsmmu;\n+\t\t};\n+\t\tu8 sizer10[0x100];\n+\t};\n+};\n+\n+static_assert(sizeof(struct rec_exit) == SZ_2K);\n+static_assert(offsetof(struct rec_exit, esr) == 0x100);\n+static_assert(offsetof(struct rec_exit, gprs) == 0x200);\n+static_assert(offsetof(struct rec_exit, cntp_ctl) == 0x400);\n+static_assert(offsetof(struct rec_exit, ripas_base) == 0x500);\n+static_assert(offsetof(struct rec_exit, imm) == 0x600);\n+static_assert(offsetof(struct rec_exit, pmu_ovf_status) == 0x700);\n+\n+struct rec_run {\n+\tstruct rec_enter enter;\n+\tstruct rec_exit exit;\n+};\n+\n+static_assert(sizeof(struct rec_run) == SZ_4K);\n+\n+/* RMI_RTT_UNPROT_MAP_FLAGS definitions */\n+#define RMI_RTT_UNPROT_MAP_FLAGS_S2AP_MASK\t\tGENMASK(22, 19)\n+#define RMI_RTT_UNPROT_MAP_FLAGS_MEMATTR_MASK\t\tGENMASK(18, 16)\n+#define RMI_RTT_UNPROT_MAP_FLAGS_LIST_COUNT_MASK\tGENMASK(15, 2)\n+#define RMI_RTT_UNPROT_MAP_FLAGS_OADDR_TYPE_MASK\tGENMASK(1, 0)\n+\n+/* RMI_RTT_PROT_MAP_FLAGS definitions */\n+#define RMI_RTT_PROT_MAP_FLAGS_LIST_COUNT_MASK\t\tGENMASK(15, 2)\n+#define RMI_RTT_PROT_MAP_FLAGS_OADDR_TYPE_MASK\t\tGENMASK(1, 0)\n+\n+/* S2AP Direct Encodings, used in RMI_RTT_UNPROT_MAP_FLAGS_S2AP */\n+#define RMI_S2AP_DIRECT_READ\t\t\tBIT(1)\n+#define RMI_S2AP_DIRECT_WRITE\t\t\tBIT(0)\n+\n+#endif /* __LINUX_ARM_SMCCC_RMI_H_ */\ndiff --git a/include/linux/suspend.h b/include/linux/suspend.h\nindex b02876f1ae38a..a3815027773c5 100644\n--- a/include/linux/suspend.h\n+++ b/include/linux/suspend.h\n@@ -401,6 +401,7 @@ int hibernate_quiet_exec(int (*func)(void *data), void *data);\n int hibernate_resume_nonboot_cpu_disable(void);\n int arch_hibernation_header_save(void *addr, unsigned int max_size);\n int arch_hibernation_header_restore(void *addr);\n+bool arch_hibernation_available(void);\n \n #else /* CONFIG_HIBERNATION */\n static inline void register_nosave_region(unsigned long b, unsigned long e) {}\ndiff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c\nindex d2479c69d71a4..9d9d53828542f 100644\n--- a/kernel/power/hibernate.c\n+++ b/kernel/power/hibernate.c\n@@ -106,11 +106,17 @@ bool hibernation_in_progress(void)\n \treturn !atomic_read(\u0026hibernate_atomic);\n }\n \n+__weak bool arch_hibernation_available(void)\n+{\n+\treturn true;\n+}\n+\n bool hibernation_available(void)\n {\n \treturn nohibernate == 0 \u0026\u0026\n \t\t!security_locked_down(LOCKDOWN_HIBERNATION) \u0026\u0026\n-\t\t!secretmem_active() \u0026\u0026 !cxl_mem_active();\n+\t\t!secretmem_active() \u0026\u0026 !cxl_mem_active() \u0026\u0026\n+\t\tarch_hibernation_available();\n }\n \n /**\n"
}

2/1 2026/10/02 04:43 agent patch-evaluator
0m Model:
core

Args:
null

Results:
{
  "EnableConfigs": null,
  "FocusSymbols": [],
  "Reasoning": "The patch series introduces support for the Arm Realm Management Extension (RME) and Realm Management Monitor (RMM) / Realm Management Interface (RMI), which is exclusive to the arm64 architecture. The driver code under drivers/firmware/arm_rmm/ depends on ARCH_SUPPORTS_RMM, which is only selected by arm64 and cannot be compiled on amd64. The only generic kernel change is the addition of a weak stub arch_hibernation_available() returning true in kernel/power/hibernate.c, which produces no functional change on amd64. Therefore, all executable logic modified by the patch is exclusive to architectures other than the target architecture (amd64).",
  "WorthFuzzing": false
}

Instruction:
You are an expert Linux kernel maintainer and security engineer.
Your job is to review a provided patch series and evaluate whether it warrants fuzzing with syzkaller.

IMPORTANT: The changes have ALREADY been applied and committed as the HEAD commit in
your workspace. Do NOT rely on internal assumptions. You must actively use your code access
tools to inspect the actual source code, callers, and surrounding context.

================================================================================
1. CORE TRIAGE PHILOSOPHY
================================================================================
The goal of patch fuzzing is to discover crashes, regressions, exposed latent bugs,
and newly triggered assertions introduced by the patch series.

- REACHABILITY IS THE PRIMARY GATE:
  Fuzzing can only discover bugs in code that can actually execute in standard virtualized
  environments (GCE or QEMU, utilizing software-emulated devices like USB gadgets, netdev, tun/tap).
  If the modified code is structurally unreachable (see Section 2), it MUST NOT be fuzzed,
  regardless of whether it adds assertions or complex logic.

- DO NOT BLINDLY TRUST "NO FUNCTIONAL CHANGE" (NFCI) OR "REFACTORING" CLAIMS:
  Patch authors routinely label changes as "cleanups", "refactorings", or state
  "No functional change intended". Do NOT take these claims at face value.
  Code refactorings that rearrange logic, introduce helper functions, or alter state management
  in core subsystems frequently introduce subtle semantic shifts or uncover latent kernel bugs.
  If reachable executable code is modified or refactored, it MUST be fuzzed.

- NEW OR MODIFIED ASSERTIONS IN REACHABLE CODE MUST BE FUZZED:
  When a patch introduces or modifies runtime checks or assertions (e.g., WARN_ON*, VM_WARN_ON*,
  BUG_ON*, lockdep_assert*) in reachable code paths, it enforces new or stricter invariants.
  Even if the author believes the invariant always holds, fuzzing is essential to verify whether
  an unusual sequence of operations can violate it.

================================================================================
2. WHEN TO RETURN WorthFuzzing=false (NEGATIVE CRITERIA)
================================================================================
Return WorthFuzzing=false ONLY IF all modified code falls strictly into one or more of these categories:

- Non-kernel and non-executable changes:
  * Modifications to Documentation/, comments, or spelling fixes.
  * User-space directories, self-tests, samples, or scripts (e.g., tools/, samples/, scripts/, usr/)
    that do not affect the compiled kernel image (vmlinux) or kernel modules.
  * Purely decorative logging (e.g., message strings in pr_err, printk, dev_info) or tracepoints
    that do not alter control flow or data structures.
  * Build system or Kconfig changes that do not alter compiled C logic.
- Structurally unreachable hardware:
  * Vendor-specific PCIe switches, SmartNICs, or GPU drivers (e.g., mlxsw, pds_core, qed,
    ionic, amdgpu) requiring physical ASIC/PCIe cards not emulated in standard QEMU.
- Unreachable execution paths:
  * Driver teardown callbacks (.remove, .shutdown, pci_unregister_driver) executed only during
    physical PCI hot-unplug or manual sysfs driver unbinding.
  * Code paths exclusive to architectures other than the target architecture.

================================================================================
3. WHEN TO RETURN WorthFuzzing=true (POSITIVE CRITERIA)
================================================================================
Return WorthFuzzing=true whenever the patch touches reachable executable code, including:
- Core Subsystems:
  * Any logic modifications in memory management (mm/), synchronization/locking (kernel/locking/),
    BPF, scheduler, core networking, VFS, or syscall handling.
- Refactorings and Code Cleanups:
  * Any restructuring of reachable data structures, helper abstractions, or algorithm flows.
- Runtime Assertions and Defensive Checks:
  * Any introduction or alteration of assertions (WARN_ON*, VM_WARN_ON*, BUG_ON*, etc.) in reachable paths.
- Reachable Drivers and Protocols:
  * Drivers accessible via virtual buses (virtio, USB gadget, loopback, netlink, binder, sockets, etc.).

================================================================================
4. EXTRACTING FocusSymbols (PREVENTING DILUTION)
================================================================================
When WorthFuzzing=true, you must extract specific kernel functions into FocusSymbols to guide the fuzzer:

- AVOID UBIQUITOUS LIFECYCLE HOT-PATHS:
  Do NOT list generic, ubiquitous functions called by almost every program in the corpus
  (including, but not limited to: general memory allocators and deallocators, page fault
  and trap handlers, or core synchronization primitives; this is not an exhaustive list).
  Listing ubiquitous functions causes the fuzzer to classify thousands of unrelated tests as "focused",
  which severely dilutes fuzzing effort away from the actual changes.

- TARGET SPECIFIC FEATURE LOGIC AND ENTRYPOINTS:
  List functions that specifically implement the logic being added or altered, or direct API entrypoints
  for the subsystem feature under review.

- HANDLING STATIC INLINE FUNCTIONS IN HEADERS (.h):
  Compiler-inlined static functions (such as static inlines in mm/*.h or include/linux/*.h) lack
  distinct symbol addresses in vmlinux and cannot be targeted directly by symbol coverage filters.
  If the changes are primarily in static inline helpers, identify non-static, feature-specific caller
  functions in .c files that exercise them (avoiding ubiquitous lifecycle wrappers).

================================================================================
5. IDENTIFYING EnableConfigs
================================================================================
Identify any specific CONFIG_ options required to properly compile and reach the modified code:
- Inspect Kconfig files and #ifdef guards; do not make assumptions.
- Check "depends on" lines in Kconfig to include any non-standard parent subsystem configs needed.
- Strip any 'CONFIG_' prefix (e.g., return "NET_IPV4" instead of "CONFIG_NET_IPV4").
Prefer calling several tools at the same time to save round-trips.


Use set-results tool to provide results of the analysis.
It must be called exactly once before the final reply.
Ignore results of this tool.

Prompt:
Target architecture: amd64

For your convenience, here is the diff of the changes:
commit c0c6b7f46a5bcf42fd906c3277c7c85b5bdd6abe
Author: syz-cluster <triage@syzkaller.com>
Date:   Fri Oct 2 04:42:57 2026 +0000

    syz-cluster: applied patch under review

diff --git a/MAINTAINERS b/MAINTAINERS
index 6215fcb077705..1646012706823 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -3951,8 +3951,11 @@ S:	Maintained
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git
 F:	Documentation/arch/arm64/
 F:	arch/arm64/
+F:	drivers/firmware/arm_rmm/
 F:	drivers/virt/coco/arm-cca-guest/
 F:	drivers/virt/coco/pkvm-guest/
+F:	include/linux/arm-rmi-cmds.h
+F:	include/linux/arm-smccc-rmi.h
 F:	tools/testing/selftests/arm64/
 X:	arch/arm64/boot/dts/
 X:	arch/arm64/configs/defconfig
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index b5a51b0ef9440..ff9565d3ffa59 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -38,6 +38,7 @@ config ARM64
 	select ARCH_HAS_MEMBARRIER_SYNC_CORE
 	select ARCH_HAS_MEM_ENCRYPT
 	select ARCH_SUPPORTS_MSEAL_SYSTEM_MAPPINGS
+	select ARCH_SUPPORTS_RMM
 	select ARCH_HAS_NMI_SAFE_THIS_CPU_OPS
 	select ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
 	select ARCH_HAS_NONLEAF_PMD_YOUNG if ARM64_HAFT
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 32102c3912fa7..e8b29983b0021 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -293,6 +293,7 @@ static const struct arm64_ftr_bits ftr_id_aa64isar3[] = {
 static const struct arm64_ftr_bits ftr_id_aa64pfr0[] = {
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_CSV3_SHIFT, 4, 0),
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_CSV2_SHIFT, 4, 0),
+	ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_RME_SHIFT, 4, 0),
 	ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_DIT_SHIFT, 4, 0),
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_AMU_SHIFT, 4, 0),
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_MPAM_SHIFT, 4, 0),
diff --git a/arch/arm64/kernel/hibernate.c b/arch/arm64/kernel/hibernate.c
index 7bf1174277772..08e03d24b93a8 100644
--- a/arch/arm64/kernel/hibernate.c
+++ b/arch/arm64/kernel/hibernate.c
@@ -10,6 +10,8 @@
  * Copyright (C) 2006 Rafael J. Wysocki <rjw@sisk.pl>
  */
 #define pr_fmt(x) "hibernate: " x
+
+#include <linux/arm-rmi-cmds.h>
 #include <linux/cpu.h>
 #include <linux/kvm_host.h>
 #include <linux/pm.h>
@@ -105,6 +107,18 @@ void notrace restore_processor_state(void)
 {
 }
 
+bool arch_hibernation_available(void)
+{
+	/*
+	 * If we have activated the RMM, there could be pages that are
+	 * delegated to the RMM. Trying to save them to the image will be fatal.
+	 * Also, we donate pages to the RMM at activation and restoring data
+	 * to those pages are going to be fatal.
+	 * Hence, disable the hibernation when the RMM is active
+	 */
+	return !cpus_are_stuck_in_kernel() && !is_rmm_active();
+}
+
 int arch_hibernation_header_save(void *addr, unsigned int max_size)
 {
 	struct arch_hibernate_hdr *hdr = addr;
diff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c
index 8f9bc2327dc85..48f343704cb54 100644
--- a/arch/arm64/kernel/machine_kexec.c
+++ b/arch/arm64/kernel/machine_kexec.c
@@ -6,6 +6,7 @@
  * Copyright (C) Huawei Futurewei Technologies.
  */
 
+#include <linux/arm-rmi-cmds.h>
 #include <linux/interrupt.h>
 #include <linux/irq.h>
 #include <linux/kernel.h>
@@ -59,6 +60,16 @@ int machine_kexec_prepare(struct kimage *kimage)
 		return -EBUSY;
 	}
 
+	/*
+	 * We will be able to allow kdump to proceed, once we have the support
+	 * for handling GPF from vmcore accesses to delegated pages. Until then
+	 * block kexec completely.
+	 */
+	if (is_rmm_active()) {
+		pr_err("Can't kexec: RMM is active.\n");
+		return -EBUSY;
+	}
+
 	return 0;
 }
 
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 44aae52c473d7..f0ef348f725b6 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -2163,6 +2163,8 @@ static u64 sanitise_id_aa64pfr0_el1(const struct kvm_vcpu *vcpu, u64 val)
 	 */
 	val &= ~ID_AA64PFR0_EL1_MPAM_MASK;
 
+	val &= ~ID_AA64PFR0_EL1_RME_MASK;
+
 	return val;
 }
 
diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig
index b7cc11e4fbfa6..62660bf520a8d 100644
--- a/drivers/firmware/Kconfig
+++ b/drivers/firmware/Kconfig
@@ -310,5 +310,6 @@ source "drivers/firmware/samsung/Kconfig"
 source "drivers/firmware/smccc/Kconfig"
 source "drivers/firmware/tegra/Kconfig"
 source "drivers/firmware/xilinx/Kconfig"
+source "drivers/firmware/arm_rmm/Kconfig"
 
 endmenu
diff --git a/drivers/firmware/Makefile b/drivers/firmware/Makefile
index be46f1e1dc77f..196a650ccf025 100644
--- a/drivers/firmware/Makefile
+++ b/drivers/firmware/Makefile
@@ -39,3 +39,4 @@ obj-y				+= samsung/
 obj-y				+= smccc/
 obj-y				+= tegra/
 obj-y				+= xilinx/
+obj-y				+= arm_rmm/
diff --git a/drivers/firmware/arm_rmm/Kconfig b/drivers/firmware/arm_rmm/Kconfig
new file mode 100644
index 0000000000000..a37cba6647360
--- /dev/null
+++ b/drivers/firmware/arm_rmm/Kconfig
@@ -0,0 +1,25 @@
+
+config ARCH_SUPPORTS_RMM
+	bool
+
+config ARM_RMM_RMI
+	bool "Realm Management Interface (RMI) Support"
+	depends on ARCH_SUPPORTS_RMM
+	help
+	  Support the Realm Management Monitor (RMM) on Arm systems that
+	  implement the Realm Management Extension (RME), as defined by the
+	  Arm Confidential Compute Architecture.
+
+	  The RMM runs at EL2 in the Realm world and provides the Realm
+	  Management Interface (RMI) used by a Normal World host to create,
+	  manage and run protected virtual machines called Realms. The RMM can
+	  also act as a TSM, as defined by the PCIe TDISP and can manage the
+	  PCI IDE setup for securing the PCIe links.
+
+	  This option builds the host-side RMI support used by KVM to detect a
+	  compatible RMM, configure it, manage delegated memory and enable
+	  Realm guests.
+
+	  Selecting this option does not by itself make Realm guests available:
+	  the system must also provide RME-capable hardware and firmware with a
+	  compatible RMM implementation.
diff --git a/drivers/firmware/arm_rmm/Makefile b/drivers/firmware/arm_rmm/Makefile
new file mode 100644
index 0000000000000..25e6ad80d3a19
--- /dev/null
+++ b/drivers/firmware/arm_rmm/Makefile
@@ -0,0 +1 @@
+obj-$(CONFIG_ARM_RMM_RMI)	+= rmi.o
diff --git a/drivers/firmware/arm_rmm/rmi.c b/drivers/firmware/arm_rmm/rmi.c
new file mode 100644
index 0000000000000..a01ab69a64c7c
--- /dev/null
+++ b/drivers/firmware/arm_rmm/rmi.c
@@ -0,0 +1,1128 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2023-2026 ARM Ltd.
+ */
+
+#include <linux/arm-rmi-cmds.h>
+#include <linux/cpufeature.h>
+#include <linux/memblock.h>
+#include <linux/memory.h>
+#include <linux/mmzone.h>
+#include <linux/slab.h>
+
+#include <asm/memory.h>
+#include <asm/pgtable-hwdef.h>
+
+/* RMM v2.0 defines RmiFeatureRegister0 to RmiFeatureRegister4. */
+static unsigned long rmi_feat_reg_cache[5] __ro_after_init;
+static bool arm64_rmm_active;
+static bool arm64_rmi_is_available;
+
+/* is_rmm_active: Returns if the RMM is ACTIVE */
+bool is_rmm_active(void)
+{
+	return READ_ONCE(arm64_rmm_active);
+}
+EXPORT_SYMBOL_GPL(is_rmm_active);
+
+/*
+ * is_rmi_available: Returns if the RMM is ACTIVE and is configured with static
+ * carveout for metadata nad L1GPTs.
+ */
+bool is_rmi_available(void)
+{
+	return READ_ONCE(arm64_rmi_is_available);
+}
+EXPORT_SYMBOL_GPL(is_rmi_available);
+
+/*
+ * rmi_granule_range_delegate() - Delegate granules
+ * @base: PA of the first granule of the range
+ * @top: PA of the first granule after the range
+ * @out_top: PA of the first granule not delegated
+ *
+ * Delegate a range of granule for use by the realm world. If the entire range
+ * was delegated then @out_top == @top, otherwise the function should be called
+ * again with @base == @out_top.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static long rmi_granule_range_delegate(unsigned long base,
+				       unsigned long top,
+				       unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_GRANULE_RANGE_DELEGATE, base, top
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_top)
+		*out_top = regs.a1;
+
+	return ret;
+}
+
+/*
+ * rmi_delegate_range: Delegate a physically contiguous range.
+ * We iterate over the range until we hit an error. So we may
+ * return an error, but with a partially delegated range. The
+ * caller must always look at the @out_phys to figure out, how
+ * much progress was made.
+ *
+ * @phys:	Base of the physical address range
+ * @size:	Size of the physical address range
+ * @out_phys:	Top of the range that was completed. This is always
+ *		valid, irrespective of the result.
+ *
+ * Returns RMI_SUCCESS on successful completion. Otherwise, returns
+ * the Linux error number or the RMI status code as described
+ * by the RMM spec for RMI_GRANULE_DELEGATE_RANGE or RMI_BLOCKED.
+ */
+int rmi_delegate_range(phys_addr_t phys,
+		       unsigned long size,
+		       phys_addr_t *out_phys)
+{
+	long ret = 0;
+	unsigned long top = phys + size;
+	unsigned long out_top;
+
+	while (phys < top) {
+		ret = rmi_granule_range_delegate(phys, top, &out_top);
+
+		if (ret != RMI_SUCCESS)
+			break;
+		/*
+		 * Buggy RMM ? Let the caller handle the failure.
+		 * We can't know how far the RMM delegated in this
+		 * iteration, so we return the best known good limit.
+		 * RMM can deal with granules already in "undelegated"
+		 * in a given range. So, it is fine for the caller to
+		 * try the range we return.
+		 */
+		if (WARN_ON(out_top <= phys)) {
+			ret = -ENXIO;
+			break;
+		}
+		phys = out_top;
+	}
+
+	if (out_phys)
+		*out_phys = phys;
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(rmi_delegate_range);
+
+/*
+ * rmi_granule_range_undelegate() - Undelegate a range of granules
+ * @base: Base PA of the target range
+ * @top: Top PA of the target range
+ * @out_top: Returns the top PA of range whose state is undelegated
+ *
+ * Undelegate a range of granules to allow use by the normal world. Will fail
+ * if the granules are in use by RMM. RMM can ignore granules that are already
+ * undelegated and thus is safe to be called on a range with a mix of delegated
+ * and undelegated granules.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static long rmi_granule_range_undelegate(unsigned long base,
+						unsigned long top,
+						unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_GRANULE_RANGE_UNDELEGATE, base, top
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_top)
+		*out_top = regs.a1;
+
+	return ret;
+}
+
+int rmi_undelegate_range(phys_addr_t phys,
+			 unsigned long size)
+{
+	long ret = 0;
+	unsigned long top = phys + size;
+	unsigned long out_top;
+
+	while (phys < top) {
+		ret = rmi_granule_range_undelegate(phys, top, &out_top);
+
+		if (ret != RMI_SUCCESS)
+			break;
+		/* Buggy RMM ? Let the caller leak the pages */
+		if (WARN_ON(out_top <= phys))
+			return -ENXIO;
+		phys = out_top;
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(rmi_undelegate_range);
+
+/*
+ * rmi_free_delegated_page: Undelegate and free a page that has been previously
+ * delegated to the Realm world. If we are unable to undelegate it, the page is
+ * leaked.
+ * NOTE: Do not use this helper if the page could be concurrently operated by
+ * another thread, as it may get leaked if the undelegation fails due to RMI_BLOCKED
+ */
+int rmi_free_delegated_page(phys_addr_t phys)
+{
+	if (WARN_ON_ONCE(rmi_undelegate_page(phys))) {
+		/* Undelegate failed: leak the page */
+		return -EBUSY;
+	}
+
+	free_page((unsigned long)phys_to_virt(phys));
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(rmi_free_delegated_page);
+
+/*
+ * Convert the RmiAddrBlockSize to actual size. This is used in RmiDonateReq
+ * and RmiAddrRangeDesc*.
+ */
+static unsigned long rmi_addr_block_size_to_bytes(unsigned long block_size_fld)
+{
+	return BIT(ARM64_HW_PGTABLE_LEVEL_SHIFT(3 - block_size_fld));
+}
+
+/*
+ * free_addr_range: Free memory described by the address range entry, which may
+ *		    be partially consumed by RMM.
+ *
+ * @entry: RMI_ADDR_RANGE descriptor
+ * @consumed_size: Page aligned size consumed by the RMM from the address range.
+ *
+ * If the state of the address is DELEGATED, undelegate it back, before freeing.
+ * Leaks the memory if we cannot undelegate the range.
+ */
+static void free_addr_range(unsigned long entry, unsigned long consumed_size)
+{
+	unsigned long phys = RMI_ADDR_RANGE_ADDR(entry);
+	unsigned long block_size_fld = RMI_ADDR_RANGE_BLOCK_SIZE(entry);
+	unsigned long count = RMI_ADDR_RANGE_COUNT(entry);
+	unsigned long state = RMI_ADDR_RANGE_STATE(entry);
+	unsigned long size = rmi_addr_block_size_to_bytes(block_size_fld) * count;
+
+	WARN_ON(!PAGE_ALIGNED(phys) || !PAGE_ALIGNED(consumed_size));
+
+	/* We shouldn't see this in reclaim path, leak it for now  */
+	if (WARN_ON((state != RMI_OP_MEM_DELEGATED) &&
+		    (state != RMI_OP_MEM_UNDELEGATED)))
+		return;
+
+	/* Adjust the address and size for partially consumed entry */
+	phys += consumed_size;
+	size -= consumed_size;
+	/*
+	 * Undelegate the pages back if required. If we can't
+	 * change them back, leak the pages.
+	 */
+	if (state == RMI_OP_MEM_DELEGATED &&
+	    WARN_ON(rmi_undelegate_range(phys, size)))
+		return;
+	free_pages_exact(phys_to_virt(phys), size);
+}
+
+static void rmi_op_continue(unsigned long sro_handle, unsigned long flags,
+			    struct arm_smccc_1_2_regs *out_regs)
+{
+	*out_regs = (struct arm_smccc_1_2_regs) {
+		SMC_RMI_OP_CONTINUE, sro_handle, flags
+	};
+
+	rmi_smccc_invoke(out_regs);
+}
+
+static void rmi_op_cancel(unsigned long sro_handle,
+			  struct arm_smccc_1_2_regs *out_regs)
+{
+	*out_regs = (struct arm_smccc_1_2_regs) {
+		SMC_RMI_OP_CANCEL, sro_handle
+	};
+
+	rmi_smccc_invoke(out_regs);
+}
+
+static void rmi_op_mem_donate(unsigned long sro_handle, unsigned long list_addr,
+			      unsigned long list_count, unsigned long flags,
+			      struct arm_smccc_1_2_regs *out_regs)
+{
+	*out_regs = (struct arm_smccc_1_2_regs) {
+		SMC_RMI_OP_MEM_DONATE, sro_handle, list_addr, list_count, flags
+	};
+
+	/*
+	 * The output donated count (a1) is always valid, irrespective
+	 * of the return result. i.e., 0 if there was an error
+	 */
+	rmi_smccc_invoke(out_regs);
+}
+
+static void rmi_op_mem_reclaim(unsigned long sro_handle,
+			       unsigned long list_addr,
+			       unsigned long list_count,
+			       struct arm_smccc_1_2_regs *out_regs)
+{
+	*out_regs = (struct arm_smccc_1_2_regs) {
+		SMC_RMI_OP_MEM_RECLAIM, sro_handle, list_addr, list_count
+	};
+
+	rmi_smccc_invoke(out_regs);
+}
+
+static int rmi_sro_ensure_capacity(struct rmi_sro_state *sro,
+				   unsigned long count)
+{
+	if (WARN_ON_ONCE(sro->addr_count > RMI_MAX_ADDR_LIST))
+		return -EOVERFLOW;
+
+	if (count > RMI_MAX_ADDR_LIST - sro->addr_count)
+		return -ENOSPC;
+
+	return 0;
+}
+
+static int rmi_sro_donate_contig(struct rmi_sro_state *sro,
+				 unsigned long sro_handle,
+				 unsigned long donatereq,
+				 struct arm_smccc_1_2_regs *out_regs,
+				 gfp_t gfp)
+{
+	unsigned long block_size_fld = RMI_DONATE_BLOCK_SIZE(donatereq);
+	unsigned long block_size = rmi_addr_block_size_to_bytes(block_size_fld);
+	unsigned long count = RMI_DONATE_COUNT(donatereq);
+	unsigned long state = RMI_DONATE_STATE(donatereq);
+	unsigned long size = block_size * count;
+	unsigned long addr_range;
+	unsigned long donated_size;
+	int ret;
+	void *virt;
+	phys_addr_t phys;
+
+	/*
+	 * The RMM specification requires contiguous allocations are always a
+	 * power of 2
+	 */
+	if (WARN_ON_ONCE(!is_power_of_2(size)))
+		return -EINVAL;
+	/*
+	 * RMM clamps the Maximum value of RmiOpMemDonateReq:count to prevent
+	 * overflow in the RMI_ADDR_RANGE_COUNT field.
+	 */
+	if (WARN_ON_ONCE(count > FIELD_MAX(RMI_ADDR_RANGE_COUNT_MASK)))
+		return -EINVAL;
+
+	/*
+	 * We can't allocate blocks more than what the buddy allocator can
+	 * satisfy. TODO: Add support for larger allocations.
+	 */
+	if (get_order(size) > MAX_PAGE_ORDER)
+		return -ENOMEM;
+
+	/* Reuse the cached address range if we have one */
+	for (int i = 0; i < sro->addr_count; i++) {
+		unsigned long entry = sro->addr_list[i];
+
+		if (RMI_ADDR_RANGE_BLOCK_SIZE(entry) == block_size_fld &&
+		    RMI_ADDR_RANGE_COUNT(entry) == count &&
+		    RMI_ADDR_RANGE_STATE(entry) == state &&
+		    IS_ALIGNED(RMI_ADDR_RANGE_ADDR(entry), size)) {
+			sro->addr_count--;
+			swap(sro->addr_list[sro->addr_count],
+			     sro->addr_list[i]);
+
+			goto mem_donate;
+		}
+	}
+
+	ret = rmi_sro_ensure_capacity(sro, 1);
+	if (ret)
+		return ret;
+
+	virt = alloc_pages_exact(size, gfp);
+	if (!virt)
+		return -ENOMEM;
+	phys = virt_to_phys(virt);
+
+	if (state == RMI_OP_MEM_DELEGATED) {
+		phys_addr_t delegated_phys;
+
+		if (rmi_delegate_range(phys, size, &delegated_phys)) {
+			if (!rmi_undelegate_range(phys, delegated_phys - phys))
+				free_pages_exact(virt, size);
+			return -ENXIO;
+		}
+	}
+
+	addr_range = phys & RMI_ADDR_RANGE_ADDR_MASK;
+	FIELD_MODIFY(RMI_ADDR_RANGE_BLOCK_SIZE_MASK, &addr_range, block_size_fld);
+	FIELD_MODIFY(RMI_ADDR_RANGE_COUNT_MASK, &addr_range, count);
+	FIELD_MODIFY(RMI_ADDR_RANGE_STATE_MASK, &addr_range, state);
+
+	sro->addr_list[sro->addr_count] = addr_range;
+
+mem_donate:
+	rmi_op_mem_donate(sro_handle,
+			  virt_to_phys(&sro->addr_list[sro->addr_count]), 1,
+			  0, out_regs);
+	donated_size = out_regs->a1 << PAGE_SHIFT;
+
+	if (WARN_ON(out_regs->a1 > (size >> PAGE_SHIFT)))
+		donated_size = size;
+
+	/* All granules consumed by the RMM */
+	if (donated_size == size)
+		return 0;
+	/* No granules were consumed by the RMM, cache them */
+	if (donated_size == 0) {
+		sro->addr_count++;
+		return 0;
+	}
+
+	/* The granules were partially consumed, reclaim the unused ones. */
+	free_addr_range(sro->addr_list[sro->addr_count], donated_size);
+
+	return 0;
+}
+
+static int rmi_sro_donate_noncontig(struct rmi_sro_state *sro,
+				    unsigned long sro_handle,
+				    unsigned long donatereq,
+				    struct arm_smccc_1_2_regs *out_regs,
+				    gfp_t gfp)
+{
+	unsigned long block_size_fld = RMI_DONATE_BLOCK_SIZE(donatereq);
+	unsigned long block_size = rmi_addr_block_size_to_bytes(block_size_fld);
+	unsigned long count = RMI_DONATE_COUNT(donatereq);
+	unsigned long state = RMI_DONATE_STATE(donatereq);
+	unsigned long found = 0;
+	unsigned long donated_granules;
+	unsigned long granules_per_block = block_size >> PAGE_SHIFT;
+	unsigned long consumed_blocks;
+	int addr_list_start = sro->addr_count;
+	int ret, i, src;
+
+	/*
+	 * We can't allocate blocks more than what the buddy allocator can
+	 * satisfy. TODO: Add support for larger allocations.
+	 */
+	if (get_order(block_size) > MAX_PAGE_ORDER)
+		return -ENOMEM;
+	/*
+	 * Maximum value of RmiOpMemDonateReq:count is 2^(PAGE_SHIFT-2) - 1.
+	 * But we further clamp it down by the number of entries we can do
+	 * in one go. The RMM can request  the remaining in the next iteration.
+	 */
+	if (count > RMI_MAX_ADDR_LIST)
+		count = RMI_MAX_ADDR_LIST;
+
+	/* Gather the suitable entries to the end of the list */
+	i = 0;
+	while (i <  addr_list_start && found < count) {
+		unsigned long entry = sro->addr_list[i];
+
+		if (RMI_ADDR_RANGE_BLOCK_SIZE(entry) == block_size_fld &&
+		    RMI_ADDR_RANGE_COUNT(entry) == 1 &&
+		    RMI_ADDR_RANGE_STATE(entry) == state) {
+			addr_list_start--;
+			swap(sro->addr_list[addr_list_start],
+			     sro->addr_list[i]);
+			found++;
+			/* Continue from the swapped in entry */
+			continue;
+		}
+		/* Skip past the entry */
+		i++;
+	}
+
+	ret = rmi_sro_ensure_capacity(sro, count - found);
+	if (ret) {
+		/* If we have found some entries, donate them and try again */
+		if (found)
+			goto mem_donate;
+		/* Otherwise free up the list and start again */
+		rmi_sro_free(sro);
+		/* Reset the addr_list_start to match sro->addr_count */
+		addr_list_start = 0;
+	}
+
+	for (; found < count; found++) {
+		unsigned long addr_range;
+		void *virt = alloc_pages_exact(block_size, gfp);
+		phys_addr_t phys;
+
+		if (!virt)
+			return -ENOMEM;
+
+		phys = virt_to_phys(virt);
+
+		if (state == RMI_OP_MEM_DELEGATED) {
+			phys_addr_t delegated_phys;
+
+			if (rmi_delegate_range(phys, block_size, &delegated_phys)) {
+				if (!rmi_undelegate_range(phys, delegated_phys - phys))
+					free_pages_exact(virt, block_size);
+				return -ENXIO;
+			}
+		}
+
+		addr_range = phys & RMI_ADDR_RANGE_ADDR_MASK;
+		FIELD_MODIFY(RMI_ADDR_RANGE_BLOCK_SIZE_MASK, &addr_range, block_size_fld);
+		FIELD_MODIFY(RMI_ADDR_RANGE_COUNT_MASK, &addr_range, 1);
+		FIELD_MODIFY(RMI_ADDR_RANGE_STATE_MASK, &addr_range, state);
+
+		sro->addr_list[sro->addr_count++] = addr_range;
+	}
+
+mem_donate:
+	rmi_op_mem_donate(sro_handle,
+			  virt_to_phys(&sro->addr_list[addr_list_start]),
+			  found, 0, out_regs);
+
+	donated_granules = out_regs->a1;
+	/*
+	 * The RMM shouldn't report more granules than we provided, but clamp
+	 * just in case.
+	 */
+	if (WARN_ON_ONCE(donated_granules > found * granules_per_block))
+		donated_granules = found * granules_per_block;
+
+	/*
+	 * The RMM reports the consumed memory in terms of granules, but we
+	 * track in the address lists in block-sized ranges. So divide to get
+	 * the number of (complete) consumed blocks.
+	 */
+	consumed_blocks = donated_granules / granules_per_block;
+	if (donated_granules % granules_per_block) {
+		/*
+		 * A block has been partially consumed, the start is owned by
+		 * the RMM, the tail is owned by the host
+		 */
+		unsigned long entry =
+			sro->addr_list[addr_list_start + consumed_blocks];
+		unsigned long donated_size =
+			(donated_granules % granules_per_block) << PAGE_SHIFT;
+
+		free_addr_range(entry, donated_size);
+		/*
+		 * This block is now fully 'consumed' (either held by the RMM or
+		 * freed)
+		 */
+		consumed_blocks++;
+	}
+
+	/*
+	 * Keep just the blocks the RMM didn't use in addr_list
+	 * RMM claimed consumed_blocks entries from addr_list_start.
+	 * Move the entries left out at the end i.e.,
+	 * [ addr_list_start + consumed_blocks, addr_list_start + found)
+	 * to the rest of the valid entries and adjust the addr_count to
+	 * reflect the available entries.
+	 */
+	for (i = 0, src = addr_list_start + consumed_blocks;
+		i < found - consumed_blocks; i++)
+		sro->addr_list[addr_list_start + i] = sro->addr_list[src + i];
+
+	sro->addr_count -= consumed_blocks;
+
+	return 0;
+}
+
+static int rmi_sro_donate(struct rmi_sro_state *sro,
+			  unsigned long sro_handle,
+			  unsigned long donatereq,
+			  struct arm_smccc_1_2_regs *regs,
+			  gfp_t gfp)
+{
+	unsigned long state = RMI_DONATE_STATE(donatereq);
+
+	if (WARN_ON_ONCE(!RMI_DONATE_COUNT(donatereq)))
+		return -EINVAL;
+
+	/*
+	 * We do not support RMI_OP_MEM_CONDITIONAL yet. This is only required
+	 * for use in RMI_GRANULE_TRACKING_SET, which we don't support yet.
+	 */
+	if (WARN_ON_ONCE((state != RMI_OP_MEM_DELEGATED) &&
+			 (state != RMI_OP_MEM_UNDELEGATED)))
+		return -EINVAL;
+
+	if (RMI_DONATE_CONTIG(donatereq) == RMI_OP_MEM_CONTIG)
+		return rmi_sro_donate_contig(sro, sro_handle, donatereq, regs, gfp);
+	else
+		return rmi_sro_donate_noncontig(sro, sro_handle, donatereq, regs, gfp);
+}
+
+static int rmi_sro_reclaim(struct rmi_sro_state *sro,
+			   unsigned long sro_handle,
+			   struct arm_smccc_1_2_regs *out_regs)
+{
+	unsigned long capacity;
+
+	/*
+	 * We don't do a partial free of the entries. So for now free the
+	 * entire address list as we prepare to reclaim more from the RMM.
+	 */
+	if (rmi_sro_ensure_capacity(sro, 1))
+		rmi_sro_free(sro);
+
+	capacity = RMI_MAX_ADDR_LIST - sro->addr_count;
+
+	rmi_op_mem_reclaim(sro_handle,
+			   virt_to_phys(&sro->addr_list[sro->addr_count]),
+			   capacity, out_regs);
+
+	/*
+	 * RMI_OP_MEM_RECLAIM always return RMI_INCOMPLETE, except when the
+	 * input parameters were invalid.
+	 */
+	if (WARN_ON_ONCE(RMI_RESULT_STATUS(out_regs->a0) != RMI_INCOMPLETE))
+		return -EINVAL;
+	if (WARN_ON_ONCE(out_regs->a1 > capacity))
+		out_regs->a1 = capacity;
+
+	sro->addr_count += out_regs->a1;
+
+	return 0;
+}
+
+void rmi_sro_free(struct rmi_sro_state *sro)
+{
+	/* Handle the worse */
+	if (WARN_ON(sro->addr_count < 0))
+		return;
+
+	if (WARN_ON(sro->addr_count > RMI_MAX_ADDR_LIST))
+		sro->addr_count = RMI_MAX_ADDR_LIST;
+
+	for (int i = 0; i < sro->addr_count; i++)
+		free_addr_range(sro->addr_list[i], 0);
+
+	sro->addr_count = 0;
+}
+EXPORT_SYMBOL_GPL(rmi_sro_free);
+
+long rmi_sro_memxfer_execute(struct rmi_sro_state *sro, gfp_t gfp)
+{
+	struct arm_smccc_1_2_regs *regs = &sro->regs;
+	bool cancelled = false;
+	unsigned long sro_handle;
+
+	rmi_smccc_invoke(regs);
+
+	sro_handle = regs->a1;
+	while (RMI_RESULT_STATUS(regs->a0) == RMI_INCOMPLETE) {
+		bool can_cancel = RMI_RESULT_CAN_CANCEL(regs->a0) == RMI_OP_CAN_CANCEL;
+		int ret = 0;
+
+		switch (RMI_RESULT_MEMREQ(regs->a0)) {
+		case RMI_OP_MEM_REQ_NONE:
+			rmi_op_continue(sro_handle, RMI_CONTINUE_KEEP_GOING,
+					regs);
+			break;
+		case RMI_OP_MEM_REQ_DONATE:
+			ret = rmi_sro_donate(sro, sro_handle, regs->a2, regs,
+					     gfp);
+			break;
+		case RMI_OP_MEM_REQ_RECLAIM:
+			ret = rmi_sro_reclaim(sro, sro_handle, regs);
+			break;
+		default:
+			WARN_ON_ONCE(1);
+			ret = -ENXIO;
+		}
+
+		if (ret) {
+			/*
+			 * All memory donating SROs must be cancellable. So a
+			 * failure in memory allocation shouldn't be an issue.
+			 * However, if we encounter a random failure (e.g.,
+			 * buggy RMM), don't loop forever, just give up.
+			 */
+			if (WARN_ON_ONCE(!can_cancel))
+				return ret;
+			/*
+			 * If we have already cancelled, and came back here due
+			 * to an error in MEMREQ, then there is no point
+			 * in going in loops.
+			 */
+			if (WARN_ON_ONCE(cancelled))
+				break;
+			rmi_op_cancel(sro_handle, regs);
+			cancelled = true;
+
+			if (WARN_ON_ONCE(RMI_RESULT_STATUS(regs->a0) != RMI_INCOMPLETE))
+				return ret;
+		}
+	}
+
+	if (cancelled)
+		return -ECANCELED;
+
+	return regs->a0;
+}
+EXPORT_SYMBOL_GPL(rmi_sro_memxfer_execute);
+
+/*
+ * rmi_sro_execute: Execute an RMI command that is stateful but not memory
+ * transferring. Takes regs, filled with the FIDs and the arguments in place.
+ *
+ * Returns :
+ *  -ECANCELED - If the operation had to be aborted and SRO was cancellable.
+ *  Otherwise, returns the result of the RMI command.
+ */
+long rmi_sro_execute(struct arm_smccc_1_2_regs *regs)
+{
+	bool cancelled = false;
+	unsigned long sro_handle;
+
+	rmi_smccc_invoke(regs);
+
+	sro_handle = regs->a1;
+	while (RMI_RESULT_STATUS(regs->a0) == RMI_INCOMPLETE) {
+		bool can_cancel = RMI_RESULT_CAN_CANCEL(regs->a0) == RMI_OP_CAN_CANCEL;
+
+		switch (RMI_RESULT_MEMREQ(regs->a0)) {
+		case RMI_OP_MEM_REQ_NONE:
+			rmi_op_continue(sro_handle, RMI_CONTINUE_KEEP_GOING,
+					regs);
+			break;
+		default:
+			WARN_ON_ONCE(1);
+			if (!can_cancel)
+				return regs->a0;
+			/*
+			 * We can't get here normally, but handle this anyway
+			 * for a buggy RMM implementation.
+			 */
+			if (cancelled)
+				return -ECANCELED;
+			rmi_op_cancel(sro_handle, regs);
+			cancelled = true;
+		}
+	}
+
+	if (cancelled)
+		return -ECANCELED;
+
+	return regs->a0;
+}
+EXPORT_SYMBOL_GPL(rmi_sro_execute);
+
+static int rmi_check_version(void)
+{
+	unsigned short version_major, version_minor;
+	unsigned long host_version = RMI_ABI_VERSION(RMI_ABI_MAJOR_VERSION,
+						     RMI_ABI_MINOR_VERSION);
+	unsigned long aa64pfr0 = read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1);
+	struct arm_smccc_1_2_regs res = {
+		SMC_RMI_VERSION, host_version,
+	};
+
+	/* If RME isn't supported, then RMI can't be */
+	if (cpuid_feature_extract_unsigned_field(aa64pfr0, ID_AA64PFR0_EL1_RME_SHIFT) == 0)
+		return -ENXIO;
+
+	rmi_smccc_invoke(&res);
+	if (res.a0 == SMCCC_RET_NOT_SUPPORTED)
+		return -ENXIO;
+
+	version_major = RMI_ABI_VERSION_GET_MAJOR(res.a1);
+	version_minor = RMI_ABI_VERSION_GET_MINOR(res.a1);
+
+	if (res.a0 != RMI_SUCCESS) {
+		unsigned short high_version_major, high_version_minor;
+
+		high_version_major = RMI_ABI_VERSION_GET_MAJOR(res.a2);
+		high_version_minor = RMI_ABI_VERSION_GET_MINOR(res.a2);
+
+		pr_err("Unsupported RMI ABI (v%d.%d - v%d.%d) we want v%d.%d\n",
+		       version_major, version_minor,
+		       high_version_major, high_version_minor,
+		       RMI_ABI_MAJOR_VERSION,
+		       RMI_ABI_MINOR_VERSION);
+		return -ENXIO;
+	}
+
+	pr_info("RMI ABI version %d.%d\n", version_major, version_minor);
+
+	return 0;
+}
+
+static int rmi_read_features(void)
+{
+	/*
+	 * Since we've negotiated a compatible version these feature registers
+	 * should always be available
+	 */
+	for (int i = 0; i < ARRAY_SIZE(rmi_feat_reg_cache); i++) {
+		struct arm_smccc_1_2_regs args = {
+			SMC_RMI_FEATURES, i,
+		};
+
+		rmi_smccc_invoke(&args);
+		if (WARN_ON(args.a0 != RMI_SUCCESS))
+			return -EINVAL;
+
+		rmi_feat_reg_cache[i] = args.a1;
+	}
+
+	return 0;
+}
+
+unsigned long rmi_feat_reg(unsigned int index)
+{
+	if (WARN_ON(index >= ARRAY_SIZE(rmi_feat_reg_cache)))
+		return 0;
+
+	return rmi_feat_reg_cache[index];
+}
+EXPORT_SYMBOL_GPL(rmi_feat_reg);
+
+static int rmi_rmm_config_set(unsigned long cfg_ptr)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RMM_CONFIG_SET, cfg_ptr,
+	};
+
+	rmi_smccc_invoke(&regs);
+
+	return regs.a0;
+}
+
+static int rmi_configure(void)
+{
+	unsigned long granule_feature;
+	unsigned long granule_size;
+	int ret;
+
+	switch (PAGE_SIZE) {
+	case SZ_4K:
+		granule_size = RMI_GRANULE_SIZE_4KB;
+		granule_feature = RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_4KB;
+		break;
+	case SZ_16K:
+		granule_size = RMI_GRANULE_SIZE_16KB;
+		granule_feature = RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_16KB;
+		break;
+	case SZ_64K:
+		granule_size = RMI_GRANULE_SIZE_64KB;
+		granule_feature = RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_64KB;
+		break;
+	default:
+		BUILD_BUG();
+	}
+
+	if (!(rmi_feat_reg(1) & granule_feature)) {
+		pr_err("RMM does not support %luKB granules\n",
+		       PAGE_SIZE >> 10);
+		return -ENXIO;
+	}
+
+	struct rmm_config *config __free(free_page) =
+		(struct rmm_config *)get_zeroed_page(GFP_KERNEL);
+
+	if (!config) {
+		pr_err("Unable to allocate memory for RMM config\n");
+		return -ENOMEM;
+	}
+
+	config->rmi_granule_size = granule_size;
+
+	/*
+	 * For now we set the tracking_region_size to 0 which is the only option
+	 * for 4KB PAGE_SIZE (1GB for 4KB PAGE_SIZE, 32MB/512MB for 16KB/64KB).
+	 * TODO: Support other tracking sizes via Kconfig option for other
+	 * PAGE_SIZES
+	 */
+	config->tracking_region_size = 0;
+
+	ret = rmi_rmm_config_set(virt_to_phys(config));
+	if (ret) {
+		pr_err("RMM config set failed (%d)\n", ret);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * rmi_granule_tracking_get() - Get configuration of a Granule tracking region
+ * @start: Base PA of the tracking region
+ * @end: End of the PA region
+ * @out_category: Memory category
+ * @out_state: Tracking region state
+ * @out_top: Top of the memory region
+ *
+ * Return: RMI return code
+ */
+static int rmi_granule_tracking_get(unsigned long start,
+				    unsigned long end,
+				    unsigned long *out_category,
+				    unsigned long *out_state,
+				    unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_GRANULE_TRACKING_GET, start, end,
+	};
+
+	rmi_smccc_invoke(&regs);
+
+	if (regs.a0 != RMI_SUCCESS)
+		return regs.a0;
+
+	if (out_category)
+		*out_category = regs.a1;
+	if (out_state)
+		*out_state = regs.a2;
+	if (out_top)
+		*out_top = regs.a3;
+
+	return RMI_SUCCESS;
+}
+
+/*
+ * Make sure the area is tracked by RMM at FINE granularity.
+ * We do not support changing the tracking yet.
+ */
+static int rmi_verify_memory_tracking(phys_addr_t start, phys_addr_t end)
+{
+	while (start < end) {
+		unsigned long ret, category, state, next;
+
+		ret = rmi_granule_tracking_get(start, end, &category, &state, &next);
+		if (ret != RMI_SUCCESS) {
+			pr_err("RMI_GRANULE_TRACKING_GET failed for %llx-%llx: (%lu)\n",
+				start, end, ret);
+			return -ENXIO;
+		}
+
+		if (WARN_ON(next <= start))
+			return -ENXIO;
+
+		if (state != RMI_TRACKING_FINE ||
+		    category != RMI_MEM_CATEGORY_CONVENTIONAL) {
+			/* TODO: Set granule tracking in this case */
+			pr_err("Granule tracking for region isn't fine/conventional: %llx-%lx\n",
+			       start, next);
+			return -ENODEV;
+		}
+		start = next;
+	}
+
+	return 0;
+}
+
+/*
+ * rmi_gpt_info - Query the GPT info for the given PAR.
+ * @start: Base of the physical address region
+ * @end: Top of the physical address region
+ * @out_top: Top of the physical address region for which the GPT @out_gpt_par_state
+ *		is valid
+ * @out_gpt_par_state: State of the GPT covered by [start, out_top)
+ */
+static long rmi_gpt_info(unsigned long start, unsigned long end,
+			 unsigned long *out_top,
+			 unsigned long *out_gpt_par_state)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_GPT_INFO, start, end,
+	};
+
+	rmi_smccc_invoke(&regs);
+	if (regs.a0 != RMI_SUCCESS)
+		return regs.a0;
+
+	if (out_top)
+		*out_top = regs.a1;
+	if (out_gpt_par_state)
+		*out_gpt_par_state = regs.a2;
+
+	return RMI_SUCCESS;
+}
+
+/*
+ * We do not support creating L1 GPTs yet. So, make sure that all the regions
+ * are managed by the firmware.
+ */
+static int rmi_verify_gpt_firmware_managed(phys_addr_t start, phys_addr_t end)
+{
+	unsigned long l0gpt_sz;
+	unsigned long next, par_state;
+
+	l0gpt_sz = 1UL << (30 + FIELD_GET(RMI_FEATURE_REGISTER_1_L0GPTSZ,
+					  rmi_feat_reg(1)));
+	start = ALIGN_DOWN(start, l0gpt_sz);
+	end = ALIGN(end, l0gpt_sz);
+
+	while (start < end) {
+		long ret = rmi_gpt_info(start, end, &next, &par_state);
+
+		if (ret != RMI_SUCCESS) {
+			pr_err("RMI_GPT_INFO failed for %llx-%llx: (%ld)\n",
+				start, end, ret);
+			return -ENXIO;
+		}
+
+		if (WARN_ON(next <= start))
+			return -ENXIO;
+
+		if (par_state != RMI_GPT_PAR_PLAT) {
+			pr_err("GPT for the region is not managed by firmware %llx-%lx\n",
+				start, next);
+			return -ENODEV;
+		}
+		start = next;
+	}
+
+	return 0;
+}
+
+static int rmi_prepare_memory(phys_addr_t start, phys_addr_t end)
+{
+	int ret;
+
+	if (start >= end)
+		return -EINVAL;
+
+	ret = rmi_verify_memory_tracking(start, end);
+	if (ret)
+		return ret;
+
+	return rmi_verify_gpt_firmware_managed(start, end);
+}
+
+static int rmi_init_metadata(void)
+{
+	phys_addr_t start, end;
+	struct memblock_region *r;
+
+	for_each_mem_region(r) {
+		int ret;
+
+		/* Firmware-reserved NOMAP regions are not usable system RAM */
+		if (memblock_is_nomap(r))
+			continue;
+
+		start = PAGE_ALIGN(r->base);
+		/*
+		 * We always deal with PAGE_SIZE and if in the odd case this
+		 * region boundary is not PAGE aligned, we stick to the page
+		 * that we can use from the region.
+		 */
+		end = PAGE_ALIGN_DOWN(r->base + r->size);
+		/* Too small ? */
+		if (start >= end)
+			continue;
+
+		ret = rmi_prepare_memory(start, end);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int rmi_memory_notifier(struct notifier_block *nb,
+			       unsigned long action, void *data)
+{
+	struct memory_notify *arg = data;
+	phys_addr_t start, end;
+	int ret;
+
+	if (action != MEM_GOING_ONLINE)
+		return NOTIFY_DONE;
+
+	start = PFN_PHYS(arg->start_pfn);
+	end = PFN_PHYS(arg->start_pfn + arg->nr_pages);
+
+	/*
+	 * If the hotplugged memory region is in ZONE_MOVABLE, neither the
+	 * guest-memfd nor the kernel allocations can come from there. So we are
+	 * fine to skip this.
+	 * For boot time memory, we don't allow this discount, as the
+	 * decision to move boot memory pages to ZONE_MOVABLE is purely
+	 * a kernel choice and the firmware must have decided what it covers.
+	 */
+	if (is_zone_movable_page(pfn_to_page(arg->start_pfn)))
+		return NOTIFY_DONE;
+
+	ret = rmi_prepare_memory(start, end);
+
+	return notifier_from_errno(ret);
+}
+
+static struct notifier_block rmi_memory_nb = {
+	.notifier_call = rmi_memory_notifier,
+};
+
+static int rmi_init_memory(void)
+{
+	int ret;
+
+	ret = rmi_init_metadata();
+	if (ret)
+		return ret;
+
+	return register_memory_notifier(&rmi_memory_nb);
+}
+
+static int __init arm64_init_rmi(void)
+{
+	int ret;
+
+	/* If we can't agree on the RMI ABI version, don't proceed further */
+	ret = rmi_check_version();
+	if (ret)
+		return ret;
+
+	ret = rmi_read_features();
+	if (ret)
+		return ret;
+
+	ret = rmi_configure();
+	if (ret)
+		return ret;
+
+	/* Activate the RMM */
+	struct rmi_sro_state *sro __free(kfree) = kmalloc_obj(*sro);
+	if (!sro)
+		return -ENOMEM;
+
+	ret = rmi_sro_memxfer_cmd(sro, GFP_KERNEL, SMC_RMI_RMM_ACTIVATE);
+	if (ret) {
+		pr_err("RMM activate failed (%d)\n", ret);
+		ret = ret < 0 ? ret : -ENXIO;
+		return ret;
+	}
+
+	WRITE_ONCE(arm64_rmm_active, true);
+
+	ret = rmi_init_memory();
+	if (ret) {
+		/* Deactivate the RMM */
+		if (!WARN_ON(rmi_sro_memxfer_cmd(sro, GFP_KERNEL, SMC_RMI_RMM_DEACTIVATE)))
+			WRITE_ONCE(arm64_rmm_active, false);
+		return ret;
+	}
+
+	/* Mark the RMM is usable */
+	WRITE_ONCE(arm64_rmi_is_available, true);
+	pr_info("RMI configured\n");
+
+	return 0;
+}
+
+/*
+ * Note arm64_init_rmi() must be called before kvm_init_rmi() otherwise KVM
+ * will not support realm guests. subsys_initcall() is called before
+ * module_init() (used for KVM) so this is OK.
+ */
+subsys_initcall(arm64_init_rmi);
diff --git a/include/linux/arm-rmi-cmds.h b/include/linux/arm-rmi-cmds.h
new file mode 100644
index 0000000000000..e3503aba646e2
--- /dev/null
+++ b/include/linux/arm-rmi-cmds.h
@@ -0,0 +1,578 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2026 ARM Ltd.
+ */
+
+#ifndef __LINUX_ARM_RMI_CMDS_H_
+#define __LINUX_ARM_RMI_CMDS_H_
+
+#include <linux/arm-smccc-rmi.h>
+#include <linux/bug.h>
+#include <linux/gfp.h>
+#include <linux/processor.h>
+#include <linux/string.h>
+#include <linux/types.h>
+
+#define RMI_MAX_ADDR_LIST	256
+
+struct rmi_sro_state {
+	struct arm_smccc_1_2_regs regs;
+	int addr_count;
+	unsigned long addr_list[RMI_MAX_ADDR_LIST];
+};
+
+#define RMM_BLOCKED_RETRY_COUNT		2
+/*
+ * rmi_smccc_invoke: Invoke the RMI call and return the results, retrying the
+ * command when status is RMI_BUSY. If we encounter RMI_BLOCKED, we retry
+ * it one more time before we give up. The caller is supposed to handle the
+ * result and reissue if required.
+ *
+ * We don't expect to see RMI_BLOCKED on a practical system, except when
+ * there are parallel requests that results in long standing operation,
+ * with one blocking the other.
+ *
+ * @regs: Input parameters filled in. Updated with the ouptput results
+ * after the call.
+ */
+static inline void rmi_smccc_invoke(struct arm_smccc_1_2_regs *regs)
+{
+	struct arm_smccc_1_2_regs args = *regs;
+	long status;
+	int i = 0;
+
+	while (i < RMM_BLOCKED_RETRY_COUNT) {
+		arm_smccc_1_2_invoke(&args, regs);
+
+		status = RMI_RESULT_STATUS(regs->a0);
+		if (status != RMI_BUSY && status != RMI_BLOCKED)
+			break;
+		if (status == RMI_BLOCKED)
+			i++;
+		cpu_relax();
+	}
+}
+
+/*
+ * rmi_smccc_invoke_once: Invoke the RMI call and return the results. Do not
+ * retry the command. Let the caller deal with RMI_BUSY or RMI_BLOCKED.
+ */
+static inline void rmi_smccc_invoke_once(struct arm_smccc_1_2_regs *regs)
+{
+	struct arm_smccc_1_2_regs args = *regs;
+
+	arm_smccc_1_2_invoke(&args, regs);
+}
+
+unsigned long rmi_feat_reg(unsigned int index);
+
+int rmi_delegate_range(phys_addr_t phys, unsigned long size,
+		       phys_addr_t *out_phys);
+int rmi_undelegate_range(phys_addr_t phys, unsigned long size);
+int rmi_free_delegated_page(phys_addr_t phys);
+
+static inline int rmi_delegate_page(phys_addr_t phys)
+{
+	return rmi_delegate_range(phys, PAGE_SIZE, NULL);
+}
+
+static inline int rmi_undelegate_page(phys_addr_t phys)
+{
+	return rmi_undelegate_range(phys, PAGE_SIZE);
+}
+
+long rmi_sro_memxfer_execute(struct rmi_sro_state *sro, gfp_t gfp);
+void rmi_sro_free(struct rmi_sro_state *sro);
+long rmi_sro_execute(struct arm_smccc_1_2_regs *regs);
+
+/*
+ * Resetting the addr_count is sufficient to ignore the addr_list contents.
+ */
+#define rmi_sro_memxfer_cmd(sro, gfp, ...) ({				\
+	struct rmi_sro_state *__sro = (sro);				\
+	__sro->addr_count = 0;						\
+	__sro->regs = (struct arm_smccc_1_2_regs){ __VA_ARGS__ };	\
+	long __ret = rmi_sro_memxfer_execute(__sro, gfp);		\
+	rmi_sro_free(__sro);						\
+	__ret;								\
+})
+
+#ifdef CONFIG_ARM_RMM_RMI
+
+bool is_rmm_active(void);
+bool is_rmi_available(void);
+
+#else
+
+static inline bool is_rmm_active(void)
+{
+	return false;
+}
+
+static inline bool is_rmi_available(void)
+{
+	return false;
+}
+#endif	/* CONFIG_ARM_RMM_RMI */
+
+
+/**
+ * rmi_rtt_data_map_init() - Create a mapping at protected IPA, copying contents
+ *			     from a given non-secure source granule.
+ * @rd: PA of the RD
+ * @data: PA of the target granule mapped in the guest
+ * @ipa: IPA at which the granule @data will be mapped in the guest
+ * @src: PA of the source granule with contents
+ * @flags: RMI_MEASURE_CONTENT if the contents should be measured
+ *
+ * Create a mapping from Protected IPA space to conventional memory, copying
+ * contents from a Non-secure Granule provided by the caller.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_data_map_init(unsigned long rd, unsigned long data,
+					 unsigned long ipa, unsigned long src,
+					 unsigned long flags)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_DATA_MAP_INIT, rd, data, ipa, src, flags
+	};
+
+	return rmi_sro_execute(&regs);
+}
+
+/**
+ * rmi_rtt_data_map() - Create mappings in protected IPA range with unknown contents
+ * @rd: PA of the RD
+ * @base: Base of the target IPA range
+ * @top: Top of the target IPA range
+ * @flags: Flags
+ * @oaddr: Output address set descriptor
+ * @out_top: Top address of range which was processed.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_data_map(unsigned long rd,
+				    unsigned long base,
+				    unsigned long top,
+				    unsigned long flags,
+				    unsigned long oaddr,
+				    unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_DATA_MAP, rd, base, top, flags, oaddr
+	};
+	long ret;
+
+	ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_top)
+		*out_top = regs.a1;
+
+	return ret;
+}
+
+/**
+ * rmi_rtt_data_unmap() - Remove mappings to conventional memory at a protected
+ *			  IPA range
+ * @rd: PA of the RD
+ * @base: Base of the target IPA range
+ * @top: Top of the target IPA range
+ * @flags: Flags
+ * @oaddr: Output address set descriptor
+ * @out_top: Returns top IPA of range which has been unmapped
+ * @out_range: Output address range
+ * @out_count: Number of entries in output address list
+ *
+ * Removes mappings to convention memory with a target Protected IPA range.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_data_unmap(unsigned long rd,
+				      unsigned long base,
+				      unsigned long top,
+				      unsigned long flags,
+				      unsigned long oaddr,
+				      unsigned long *out_top,
+				      unsigned long *out_range,
+				      unsigned long *out_count)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_DATA_UNMAP, rd, base, top, flags, oaddr
+	};
+	long ret;
+
+	ret = rmi_sro_execute(&regs);
+
+	if (ret != RMI_SUCCESS)
+		return ret;
+
+	if (out_top)
+		*out_top = regs.a1;
+	if (out_range)
+		*out_range = regs.a2;
+	if (out_count)
+		*out_count = regs.a3;
+
+	return RMI_SUCCESS;
+}
+
+/**
+ * rmi_psci_complete() - Complete pending PSCI command
+ * @calling_rec: PA of the calling REC
+ * @status: Status of the PSCI request
+ *
+ * Completes a pending PSCI command.
+ *
+ * Return: RMI return code
+ */
+static inline long rmi_psci_complete(unsigned long calling_rec,
+				     unsigned long status)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_PSCI_COMPLETE, calling_rec, status,
+	};
+
+	rmi_smccc_invoke(&regs);
+	return regs.a0;
+}
+
+/**
+ * rmi_realm_activate() - Activate a realm
+ * @rd: PA of the RD
+ *
+ * Mark a realm as Active, signalling that creation is completed, allowing
+ * execution of the realm.
+ *
+ * Return: RMI return code
+ */
+static inline long rmi_realm_activate(unsigned long rd)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_REALM_ACTIVATE, rd,
+	};
+
+	rmi_smccc_invoke(&regs);
+	return regs.a0;
+}
+
+/**
+ * rmi_realm_create() - Create a realm
+ * @rd: PA of the RD
+ * @params: PA of realm parameters
+ * @sro: Preallocated SRO context
+ *
+ * Create a new realm using the given parameters.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_realm_create(unsigned long rd, unsigned long params,
+				    struct rmi_sro_state *sro)
+{
+	return rmi_sro_memxfer_cmd(sro, GFP_KERNEL,
+				   SMC_RMI_REALM_CREATE, rd, params);
+}
+
+/**
+ * rmi_realm_terminate() - Terminate a realm
+ * @rd: PA of the RD
+ * @sro: Preallocated SRO context
+ *
+ * Terminates a realm, moving it into a ZOMBIE state
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_realm_terminate(unsigned long rd,
+				       struct rmi_sro_state *sro)
+{
+	return rmi_sro_memxfer_cmd(sro, GFP_KERNEL,
+				   SMC_RMI_REALM_TERMINATE, rd);
+}
+
+/**
+ * rmi_realm_destroy() - Destroy a realm
+ * @rd: PA of the RD
+ * @sro: Preallocated SRO context
+ *
+ * Destroys a realm, all objects belonging to the realm must be destroyed first.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_realm_destroy(unsigned long rd,
+				     struct rmi_sro_state *sro)
+{
+	return rmi_sro_memxfer_cmd(sro, GFP_KERNEL,
+				   SMC_RMI_REALM_DESTROY, rd);
+}
+
+/**
+ * rmi_rec_create() - Create a REC
+ * @rd: PA of the RD
+ * @rec: PA of the target REC
+ * @params: PA of REC parameters
+ * @sro: Allocated SRO context to be used
+ *
+ * Create a REC using the parameters specified in the struct rec_params pointed
+ * to by @params.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rec_create(unsigned long rd,
+				  unsigned long rec,
+				  unsigned long params,
+				  struct rmi_sro_state *sro)
+{
+	return rmi_sro_memxfer_cmd(sro, GFP_KERNEL,
+				   SMC_RMI_REC_CREATE, rd, rec, params);
+}
+
+/**
+ * rmi_rec_destroy() - Destroy a REC
+ * @rec: PA of the target REC
+ * @sro: Allocated SRO context to be used
+ *
+ * Destroys a REC. The REC must not be running.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rec_destroy(unsigned long rec,
+				   struct rmi_sro_state *sro)
+{
+	return rmi_sro_memxfer_cmd(sro, GFP_KERNEL, SMC_RMI_REC_DESTROY, rec);
+}
+
+/**
+ * rmi_rec_enter() - Enter a REC
+ * @rec: PA of the target REC
+ * @run_ptr: PA of RecRun structure
+ *
+ * Starts (or continues) execution within a REC.
+ *
+ * Return: RMI return code
+ */
+static inline long rmi_rec_enter(unsigned long rec, unsigned long run_ptr)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_REC_ENTER, rec, run_ptr,
+	};
+
+	rmi_smccc_invoke_once(&regs);
+	return regs.a0;
+}
+
+/**
+ * rmi_rtt_create() - Creates an RTT
+ * @rd: PA of the RD
+ * @rtt: PA of the target RTT
+ * @ipa: Base of the IPA range described by the RTT
+ * @level: Depth of the RTT within the tree
+ *
+ * Creates an RTT (Realm Translation Table) at the specified level for the
+ * translation of the specified address within the realm.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_create(unsigned long rd, unsigned long rtt,
+				  unsigned long ipa, long level)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_CREATE, rd, rtt, ipa, level
+	};
+
+	return rmi_sro_execute(&regs);
+}
+
+/**
+ * rmi_rtt_destroy() - Destroy an RTT
+ * @rd: PA of the RD
+ * @ipa: Base of the IPA range described by the RTT
+ * @level: RTT level
+ * @out_rtt: Pointer to write the PA of the RTT which was destroyed
+ * @out_top: Pointer to write the top IPA of non-live RTT entries, from entry
+ * at which the RTT walk terminated.
+ *
+ * Destroys an RTT. The RTT must be non-live, i.e. none of the entries in the
+ * table are in ASSIGNED or TABLE state.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code.
+ */
+static inline long rmi_rtt_destroy(unsigned long rd,
+				   unsigned long ipa,
+				   long level,
+				   unsigned long *out_rtt,
+				   unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_DESTROY, rd, ipa, level
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	switch (RMI_RESULT_STATUS(ret)) {
+	case RMI_SUCCESS:
+		if (out_rtt)
+			*out_rtt = regs.a1;
+		fallthrough;
+	case RMI_ERROR_RTT:
+		if (out_top)
+			*out_top = regs.a2;
+		break;
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+/**
+ * rmi_rtt_fold() - Fold an RTT
+ * @rd: PA of the RD
+ * @ipa: Base of the IPA range described by the RTT
+ * @level: Depth of the RTT within the tree
+ * @out_rtt: Pointer to write the PA of the RTT which was destroyed
+ *
+ * Folds an RTT. If all entries with the RTT are 'homogeneous' the RTT can be
+ * folded into the parent and the RTT destroyed.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_fold(unsigned long rd, unsigned long ipa,
+				long level, unsigned long *out_rtt)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_FOLD, rd, ipa, level
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_rtt)
+		*out_rtt = regs.a1;
+
+	return ret;
+}
+
+/**
+ * rmi_rtt_init_ripas() - Set RIPAS for new realm
+ * @rd: PA of the RD
+ * @base: Base of target IPA region
+ * @top: Top of target IPA region
+ * @out_top: Top IPA of range whose RIPAS was modified
+ *
+ * Sets the RIPAS of a target IPA range to RAM, for a realm in the NEW state.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_init_ripas(unsigned long rd, unsigned long base,
+				      unsigned long top, unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_INIT_RIPAS, rd, base, top
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_top)
+		*out_top = regs.a1;
+
+	return ret;
+}
+
+/**
+ * rmi_rtt_unprot_map() - Map unprotected granules into a realm
+ * @rd: PA of the RD
+ * @base: Base IPA of the mapping
+ * @top: Top of the target IPA range
+ * @flags: Flags
+ * @oaddr: Output address set descriptor
+ * @out_top: Top IPA of range which has been mapped
+ *
+ * Create mappings to memory within a target unprotected IPA range.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_unprot_map(unsigned long rd,
+				      unsigned long base,
+				      unsigned long top,
+				      unsigned long flags,
+				      unsigned long oaddr,
+				      unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_UNPROT_MAP, rd, base, top, flags, oaddr
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_top)
+		*out_top = regs.a1;
+
+	return ret;
+}
+
+/**
+ * rmi_rtt_set_ripas() - Set RIPAS for an running realm
+ * @rd: PA of the RD
+ * @rec: PA of the REC making the request
+ * @base: Base of target IPA region
+ * @top: Top of target IPA region
+ * @out_top: Pointer to write top IPA of range whose RIPAS was modified
+ *
+ * Completes a request made by the realm to change the RIPAS of a target IPA
+ * range.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_set_ripas(unsigned long rd, unsigned long rec,
+				     unsigned long base, unsigned long top,
+				     unsigned long *out_top)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_SET_RIPAS, rd, rec, base, top
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret == RMI_SUCCESS && out_top)
+		*out_top = regs.a1;
+
+	return ret;
+}
+
+/**
+ * rmi_rtt_unprot_unmap() - Remove mappings within an unprotected IPA range
+ * @rd: PA of the RD
+ * @base: Base IPA of the mapping
+ * @top: Top of the target IPA range
+ * @flags: Flags
+ * @oaddr: Output address set descriptor
+ * @out_top: Top IPA which has been unmapped
+ * @out_range: Output address range
+ * @out_count: Number of entries in output address list
+ *
+ * Removes mappings to memory within a target unprotected IPA range.
+ *
+ * Return: 0 on success, positive RMI result code or negative Linux error code
+ */
+static inline long rmi_rtt_unprot_unmap(unsigned long rd,
+					unsigned long base,
+					unsigned long top,
+					unsigned long flags,
+					unsigned long oaddr,
+					unsigned long *out_top,
+					unsigned long *out_range,
+					unsigned long *out_count)
+{
+	struct arm_smccc_1_2_regs regs = {
+		SMC_RMI_RTT_UNPROT_UNMAP, rd, base, top, flags, oaddr
+	};
+	long ret = rmi_sro_execute(&regs);
+
+	if (ret != RMI_SUCCESS)
+		return ret;
+
+	if (out_top)
+		*out_top = regs.a1;
+	if (out_range)
+		*out_range = regs.a2;
+	if (out_count)
+		*out_count = regs.a3;
+
+	return RMI_SUCCESS;
+}
+
+#endif	/* __LINUX_ARM_RMI_CMDS_H_ */
diff --git a/include/linux/arm-smccc-rmi.h b/include/linux/arm-smccc-rmi.h
new file mode 100644
index 0000000000000..655726b374624
--- /dev/null
+++ b/include/linux/arm-smccc-rmi.h
@@ -0,0 +1,516 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2023-2026 ARM Ltd.
+ *
+ * The values and structures in this file are from the Realm Management Monitor
+ * specification (DEN0137) version 2.0-bet3:
+ * https://developer.arm.com/documentation/den0137/2-0bet3/
+ */
+
+#ifndef __LINUX_ARM_SMCCC_RMI_H_
+#define __LINUX_ARM_SMCCC_RMI_H_
+
+#include <linux/arm-smccc.h>
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/build_bug.h>
+#include <linux/sizes.h>
+#include <linux/stddef.h>
+
+#include <asm/page.h>
+
+#define SMC_RMI_CALL(func)				\
+	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,		\
+			   ARM_SMCCC_SMC_64,		\
+			   ARM_SMCCC_OWNER_STANDARD,	\
+			   (func))
+
+#define SMC_RMI_VERSION				SMC_RMI_CALL(0x0150)
+
+#define SMC_RMI_RTT_DATA_MAP_INIT		SMC_RMI_CALL(0x0153)
+
+#define SMC_RMI_REALM_ACTIVATE			SMC_RMI_CALL(0x0157)
+#define SMC_RMI_REALM_CREATE			SMC_RMI_CALL(0x0158)
+#define SMC_RMI_REALM_DESTROY			SMC_RMI_CALL(0x0159)
+#define SMC_RMI_REC_CREATE			SMC_RMI_CALL(0x015a)
+#define SMC_RMI_REC_DESTROY			SMC_RMI_CALL(0x015b)
+#define SMC_RMI_REC_ENTER			SMC_RMI_CALL(0x015c)
+#define SMC_RMI_RTT_CREATE			SMC_RMI_CALL(0x015d)
+#define SMC_RMI_RTT_DESTROY			SMC_RMI_CALL(0x015e)
+
+#define SMC_RMI_RTT_READ_ENTRY			SMC_RMI_CALL(0x0161)
+
+#define SMC_RMI_RTT_DEV_VALIDATE		SMC_RMI_CALL(0x0163)
+#define SMC_RMI_PSCI_COMPLETE			SMC_RMI_CALL(0x0164)
+#define SMC_RMI_FEATURES			SMC_RMI_CALL(0x0165)
+#define SMC_RMI_RTT_FOLD			SMC_RMI_CALL(0x0166)
+
+#define SMC_RMI_RTT_INIT_RIPAS			SMC_RMI_CALL(0x0168)
+#define SMC_RMI_RTT_SET_RIPAS			SMC_RMI_CALL(0x0169)
+#define SMC_RMI_VSMMU_CREATE			SMC_RMI_CALL(0x016a)
+#define SMC_RMI_VSMMU_DESTROY			SMC_RMI_CALL(0x016b)
+
+#define SMC_RMI_RMM_CONFIG_SET			SMC_RMI_CALL(0x016e)
+#define SMC_RMI_PSMMU_IRQ_NOTIFY		SMC_RMI_CALL(0x016f)
+#define SMC_RMI_ATTEST_PLAT_TOKEN_REFRESH	SMC_RMI_CALL(0x0170)
+
+#define SMC_RMI_PDEV_ABORT			SMC_RMI_CALL(0x0174)
+#define SMC_RMI_PDEV_COMMUNICATE		SMC_RMI_CALL(0x0175)
+#define SMC_RMI_PDEV_CREATE			SMC_RMI_CALL(0x0176)
+#define SMC_RMI_PDEV_DESTROY			SMC_RMI_CALL(0x0177)
+#define SMC_RMI_PDEV_GET_STATE			SMC_RMI_CALL(0x0178)
+
+#define SMC_RMI_PDEV_STREAM_KEY_REFRESH		SMC_RMI_CALL(0x017a)
+#define SMC_RMI_PDEV_SET_PUBKEY			SMC_RMI_CALL(0x017b)
+#define SMC_RMI_PDEV_STOP			SMC_RMI_CALL(0x017c)
+#define SMC_RMI_RTT_AUX_CREATE			SMC_RMI_CALL(0x017d)
+#define SMC_RMI_RTT_AUX_DESTROY			SMC_RMI_CALL(0x017e)
+#define SMC_RMI_RTT_AUX_FOLD			SMC_RMI_CALL(0x017f)
+
+#define SMC_RMI_VDEV_ABORT			SMC_RMI_CALL(0x0185)
+#define SMC_RMI_VDEV_COMMUNICATE		SMC_RMI_CALL(0x0186)
+#define SMC_RMI_VDEV_CREATE			SMC_RMI_CALL(0x0187)
+#define SMC_RMI_VDEV_DESTROY			SMC_RMI_CALL(0x0188)
+#define SMC_RMI_VDEV_GET_STATE			SMC_RMI_CALL(0x0189)
+#define SMC_RMI_VDEV_UNLOCK			SMC_RMI_CALL(0x018a)
+#define SMC_RMI_RTT_SET_S2AP			SMC_RMI_CALL(0x018b)
+
+#define SMC_RMI_VDEV_GET_INTERFACE_REPORT	SMC_RMI_CALL(0x01d0)
+#define SMC_RMI_VDEV_GET_MEASUREMENTS		SMC_RMI_CALL(0x01d1)
+#define SMC_RMI_VDEV_LOCK			SMC_RMI_CALL(0x01d2)
+#define SMC_RMI_VDEV_START			SMC_RMI_CALL(0x01d3)
+
+#define SMC_RMI_VSMMU_EVENT_HANDLE		SMC_RMI_CALL(0x01d6)
+#define SMC_RMI_PSMMU_ACTIVATE			SMC_RMI_CALL(0x01d7)
+#define SMC_RMI_PSMMU_DEACTIVATE		SMC_RMI_CALL(0x01d8)
+
+#define SMC_RMI_PSMMU_ST_L2_CREATE		SMC_RMI_CALL(0x01db)
+#define SMC_RMI_PSMMU_ST_L2_DESTROY		SMC_RMI_CALL(0x01dc)
+#define SMC_RMI_DPT_L0_CREATE			SMC_RMI_CALL(0x01dd)
+#define SMC_RMI_DPT_L0_DESTROY			SMC_RMI_CALL(0x01de)
+#define SMC_RMI_DPT_L1_CREATE			SMC_RMI_CALL(0x01df)
+#define SMC_RMI_DPT_L1_DESTROY			SMC_RMI_CALL(0x01e0)
+#define SMC_RMI_GRANULE_TRACKING_GET		SMC_RMI_CALL(0x01e1)
+
+#define SMC_RMI_GRANULE_TRACKING_SET		SMC_RMI_CALL(0x01e3)
+
+#define SMC_RMI_RMM_CONFIG_GET			SMC_RMI_CALL(0x01ec)
+
+#define SMC_RMI_RMM_STATE_GET			SMC_RMI_CALL(0x01ee)
+
+#define SMC_RMI_PSMMU_EVENT_CONSUME		SMC_RMI_CALL(0x01f0)
+#define SMC_RMI_GRANULE_RANGE_DELEGATE		SMC_RMI_CALL(0x01f1)
+#define SMC_RMI_GRANULE_RANGE_UNDELEGATE	SMC_RMI_CALL(0x01f2)
+#define SMC_RMI_GPT_L1_CREATE			SMC_RMI_CALL(0x01f3)
+#define SMC_RMI_GPT_L1_DESTROY			SMC_RMI_CALL(0x01f4)
+#define SMC_RMI_RTT_DATA_MAP			SMC_RMI_CALL(0x01f5)
+#define SMC_RMI_RTT_DATA_UNMAP			SMC_RMI_CALL(0x01f6)
+#define SMC_RMI_RTT_DEV_MAP			SMC_RMI_CALL(0x01f7)
+#define SMC_RMI_RTT_DEV_UNMAP			SMC_RMI_CALL(0x01f8)
+#define SMC_RMI_RTT_ARCH_DEV_MAP		SMC_RMI_CALL(0x01f9)
+#define SMC_RMI_RTT_ARCH_DEV_UNMAP		SMC_RMI_CALL(0x01fa)
+#define SMC_RMI_RTT_UNPROT_MAP			SMC_RMI_CALL(0x01fb)
+#define SMC_RMI_RTT_UNPROT_UNMAP		SMC_RMI_CALL(0x01fc)
+#define SMC_RMI_RTT_AUX_PROT_MAP		SMC_RMI_CALL(0x01fd)
+#define SMC_RMI_RTT_AUX_PROT_UNMAP		SMC_RMI_CALL(0x01fe)
+#define SMC_RMI_RTT_AUX_UNPROT_MAP		SMC_RMI_CALL(0x01ff)
+#define SMC_RMI_RTT_AUX_UNPROT_UNMAP		SMC_RMI_CALL(0x0200)
+#define SMC_RMI_REALM_TERMINATE			SMC_RMI_CALL(0x0201)
+#define SMC_RMI_RMM_ACTIVATE			SMC_RMI_CALL(0x0202)
+#define SMC_RMI_OP_CONTINUE			SMC_RMI_CALL(0x0203)
+#define SMC_RMI_PDEV_STREAM_CONNECT		SMC_RMI_CALL(0x0204)
+#define SMC_RMI_PDEV_STREAM_DISCONNECT		SMC_RMI_CALL(0x0205)
+#define SMC_RMI_PDEV_STREAM_COMPLETE		SMC_RMI_CALL(0x0206)
+#define SMC_RMI_PDEV_STREAM_KEY_PURGE		SMC_RMI_CALL(0x0207)
+#define SMC_RMI_OP_MEM_DONATE			SMC_RMI_CALL(0x0208)
+#define SMC_RMI_OP_MEM_RECLAIM			SMC_RMI_CALL(0x0209)
+#define SMC_RMI_OP_CANCEL			SMC_RMI_CALL(0x020a)
+#define SMC_RMI_VSMMU_FEATURES			SMC_RMI_CALL(0x020b)
+#define SMC_RMI_VSMMU_CMD_GET			SMC_RMI_CALL(0x020c)
+#define SMC_RMI_VSMMU_CMD_COMPLETE		SMC_RMI_CALL(0x020d)
+#define SMC_RMI_PSMMU_INFO			SMC_RMI_CALL(0x020e)
+#define SMC_RMI_RMM_DEACTIVATE			SMC_RMI_CALL(0x020f)
+#define SMC_RMI_PDEV_STREAM_INFO		SMC_RMI_CALL(0x0210)
+#define SMC_RMI_GPT_INFO			SMC_RMI_CALL(0x0211)
+
+#define RMI_ABI_MAJOR_VERSION	2
+#define RMI_ABI_MINOR_VERSION	0
+
+#define RMI_ABI_VERSION_MAJOR_MASK		GENMASK(30, 16)
+#define RMI_ABI_VERSION_MINOR_MASK		GENMASK(15, 0)
+
+#define RMI_ABI_VERSION_GET_MAJOR(v)	FIELD_GET(RMI_ABI_VERSION_MAJOR_MASK, (v))
+#define RMI_ABI_VERSION_GET_MINOR(v)	FIELD_GET(RMI_ABI_VERSION_MINOR_MASK, (v))
+#define RMI_ABI_VERSION(major, minor)				\
+	(FIELD_PREP(RMI_ABI_VERSION_MAJOR_MASK, (major)) |	\
+	 FIELD_PREP(RMI_ABI_VERSION_MINOR_MASK, (minor)))
+
+/* RmiResult Type definitions */
+#define RMI_RESULT_STATUS_MASK		GENMASK(7, 0)
+/* RmiResultDataLevel field */
+#define RMI_RESULT_DATA_LEVEL_MASK	GENMASK(15, 8)
+/* RmiResultDataIncomplete fields */
+#define RMI_RESULT_MEMREQ_MASK		GENMASK(9, 8)
+#define RMI_RESULT_CAN_CANCEL_MASK	BIT(10)
+
+#define RMI_RESULT_STATUS(ret)		FIELD_GET(RMI_RESULT_STATUS_MASK, ret)
+#define RMI_RESULT_DATA_LEVEL(ret)	FIELD_GET(RMI_RESULT_DATA_LEVEL_MASK, ret)
+#define RMI_RESULT_MEMREQ(ret)		FIELD_GET(RMI_RESULT_MEMREQ_MASK, ret)
+#define RMI_RESULT_CAN_CANCEL(ret)	FIELD_GET(RMI_RESULT_CAN_CANCEL_MASK, ret)
+
+#define RMI_SUCCESS			0
+#define RMI_ERROR_INPUT			1
+#define RMI_ERROR_REALM			2
+#define RMI_ERROR_REC			3
+#define RMI_ERROR_RTT			4
+#define RMI_ERROR_NOT_SUPPORTED		5
+#define RMI_ERROR_DEVICE		6
+#define RMI_ERROR_RTT_AUX		7
+#define RMI_ERROR_PSMMU_ST		8
+#define RMI_ERROR_DPT			9
+#define RMI_BUSY			10
+#define RMI_ERROR_GLOBAL		11
+#define RMI_ERROR_TRACKING		12
+#define RMI_INCOMPLETE			13
+#define RMI_BLOCKED			14
+#define RMI_ERROR_GPT			15
+#define RMI_ERROR_GRANULE		16
+
+#define RMI_CONTINUE_KEEP_GOING		0
+#define RMI_CONTINUE_STOP		1
+
+#define RMI_OP_MEM_REQ_NONE		0
+#define RMI_OP_MEM_REQ_DONATE		1
+#define RMI_OP_MEM_REQ_RECLAIM		2
+
+#define RMI_OP_CANNOT_CANCEL		0
+#define RMI_OP_CAN_CANCEL		1
+
+#define RMI_DONATE_STATE_MASK		GENMASK(18, 17)
+#define RMI_DONATE_CONTIG_MASK		BIT(16)
+#define RMI_DONATE_COUNT_MASK		GENMASK(15, 2)
+#define RMI_DONATE_BLOCK_SIZE_MASK	GENMASK(1, 0)
+
+#define RMI_DONATE_STATE(req)		FIELD_GET(RMI_DONATE_STATE_MASK, req)
+#define RMI_DONATE_CONTIG(req)		FIELD_GET(RMI_DONATE_CONTIG_MASK, req)
+#define RMI_DONATE_COUNT(req)		FIELD_GET(RMI_DONATE_COUNT_MASK, req)
+#define RMI_DONATE_BLOCK_SIZE(req)	FIELD_GET(RMI_DONATE_BLOCK_SIZE_MASK, req)
+
+#define RMI_OP_MEM_DELEGATED		0
+#define RMI_OP_MEM_UNDELEGATED		1
+#define RMI_OP_MEM_CONDITIONAL		2
+
+#define RMI_OP_MEM_NON_CONTIG		0
+#define RMI_OP_MEM_CONTIG		1
+
+#define RMI_ADDR_TYPE_NONE		0
+#define RMI_ADDR_TYPE_SINGLE		1
+#define RMI_ADDR_TYPE_LIST		2
+
+#define RMI_ADDR_RANGE_STATE_MASK	GENMASK_ULL(63, 62)
+#define RMI_ADDR_RANGE_ADDR_MASK	GENMASK_ULL(51, PAGE_SHIFT)
+#define RMI_ADDR_RANGE_COUNT_MASK	GENMASK(PAGE_SHIFT - 1, 2)
+#define RMI_ADDR_RANGE_BLOCK_SIZE_MASK	GENMASK(1, 0)
+
+#define RMI_ADDR_RANGE_STATE(r)		FIELD_GET(RMI_ADDR_RANGE_STATE_MASK, (r))
+#define RMI_ADDR_RANGE_ADDR(r)		((r) & RMI_ADDR_RANGE_ADDR_MASK)
+#define RMI_ADDR_RANGE_COUNT(r)		FIELD_GET(RMI_ADDR_RANGE_COUNT_MASK, (r))
+#define RMI_ADDR_RANGE_BLOCK_SIZE(r)	FIELD_GET(RMI_ADDR_RANGE_BLOCK_SIZE_MASK, (r))
+
+enum rmi_ripas {
+	RMI_EMPTY = 0,
+	RMI_RAM = 1,
+	RMI_DESTROYED = 2,
+	RMI_DEV = 3,
+};
+
+#define RMI_NO_MEASURE_CONTENT	0
+#define RMI_MEASURE_CONTENT	1
+
+#define RMI_FEATURE_REGISTER_0_S2OASZ			GENMASK_ULL(40, 33)
+#define RMI_FEATURE_REGISTER_0_L0GPT_BLOCK_DELEGATE	BIT_ULL(32)
+#define RMI_FEATURE_REGISTER_0_PMU_NUM_CTRS		GENMASK(31, 27)
+#define RMI_FEATURE_REGISTER_0_PMU			BIT(26)
+#define RMI_FEATURE_REGISTER_0_NUM_WPS			GENMASK(25, 20)
+#define RMI_FEATURE_REGISTER_0_NUM_BPS			GENMASK(19, 14)
+#define RMI_FEATURE_REGISTER_0_SVE_VL			GENMASK(13, 10)
+#define RMI_FEATURE_REGISTER_0_SVE			BIT(9)
+#define RMI_FEATURE_REGISTER_0_LPA2			BIT(8)
+#define RMI_FEATURE_REGISTER_0_S2SZ			GENMASK(7, 0)
+
+#define RMI_FEATURE_REGISTER_1_PPS			GENMASK(16, 14)
+#define RMI_FEATURE_REGISTER_1_L0GPTSZ			GENMASK(13, 10)
+#define RMI_FEATURE_REGISTER_1_MAX_RECS_ORDER		GENMASK(9, 6)
+#define RMI_FEATURE_REGISTER_1_HASH_SHA_512		BIT(5)
+#define RMI_FEATURE_REGISTER_1_HASH_SHA_384		BIT(4)
+#define RMI_FEATURE_REGISTER_1_HASH_SHA_256		BIT(3)
+#define RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_64KB		BIT(2)
+#define RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_16KB		BIT(1)
+#define RMI_FEATURE_REGISTER_1_RMI_GRAN_SZ_4KB		BIT(0)
+
+#define RMI_FEATURE_REGISTER_2_REALM_MAX_VDEVS_ORDER	GENMASK(14, 10)
+#define RMI_FEATURE_REGISTER_2_NON_TEE_STREAM		BIT(9)
+#define RMI_FEATURE_REGISTER_2_VDEV_KROU		BIT(8)
+#define RMI_FEATURE_REGISTER_2_PDEV_MAX_VDEVS_ORDER	GENMASK(7, 4)
+#define RMI_FEATURE_REGISTER_2_ATS			BIT(3)
+#define RMI_FEATURE_REGISTER_2_VSMMU			BIT(2)
+#define RMI_FEATURE_REGISTER_2_DA_COH			BIT(1)
+#define RMI_FEATURE_REGISTER_2_DA			BIT(0)
+
+#define RMI_FEATURE_REGISTER_3_RTT_S2AP_INDIRECT	BIT(6)
+#define RMI_FEATURE_REGISTER_3_RTT_PLANE		GENMASK(5, 4)
+#define RMI_FEATURE_REGISTER_3_MAX_NUM_AUX_PLANES	GENMASK(3, 0)
+
+#define RMI_FEATURE_REGISTER_4_MEC_COUNT		GENMASK_ULL(63, 0)
+
+#define RMI_MEM_CATEGORY_CONVENTIONAL		0
+#define RMI_MEM_CATEGORY_DEV_NCOH		1
+#define RMI_MEM_CATEGORY_DEV_COH		2
+#define RMI_MEM_CATEGORY_NONE			3
+
+#define RMI_TRACKING_RESERVED			0
+#define RMI_TRACKING_NONE			1
+#define RMI_TRACKING_FINE			2
+#define RMI_TRACKING_COARSE			3
+#define RMI_TRACKING_INTERMEDIATE		4
+
+#define RMI_GRANULE_SIZE_4KB	0
+#define RMI_GRANULE_SIZE_16KB	1
+#define RMI_GRANULE_SIZE_64KB	2
+
+#define RMI_GPT_PAR_RESERVED           0U
+#define RMI_GPT_PAR_PLAT               1U
+#define RMI_GPT_PAR_HOST_NOT_CREATED   2U
+#define RMI_GPT_PAR_HOST_CREATED       3U
+
+/*
+ * Note many of these fields are smaller than u64 but all fields have u64
+ * alignment, so use u64 to ensure correct alignment.
+ */
+struct rmm_config {
+	union { /* 0x0 */
+		struct {
+			u64 tracking_region_size;
+			u64 rmi_granule_size;
+		};
+		u8 sizer[SZ_4K];
+	};
+};
+
+static_assert(sizeof(struct rmm_config) == SZ_4K);
+
+#define RMI_REALM_PARAM_FLAG_MEC_POLICY_MASK	GENMASK(8, 7)
+#define RMI_REALM_PARAM_FLAG_LFA_POLICY_MASK	GENMASK(6, 5)
+#define RMI_REALM_PARAM_FLAG_DA			BIT(3)
+#define RMI_REALM_PARAM_FLAG_PMU		BIT(2)
+#define RMI_REALM_PARAM_FLAG_SVE		BIT(1)
+
+#define RMI_HASH_SHA_256			0
+#define RMI_HASH_SHA_512			1
+#define RMI_HASH_SHA_384			2
+
+struct realm_params {
+	union { /* 0x0 */
+		struct {
+			u64 flags0;
+			u64 s2sz;
+			u64 sve_vl;
+			u64 num_bps;
+			u64 num_wps;
+			u64 pmu_num_ctrs;
+			u64 hash_algo;
+			u64 num_aux_planes;
+		};
+		u8 sizer0[0x400];
+	};
+	union { /* 0x400 */
+		struct {
+			u8 rpv[64];
+			u64 ats_plane;
+		};
+		u8 sizer1[0x400];
+	};
+	union { /* 0x800 */
+		struct {
+			u64 sizer2;
+			u64 rtt_base;
+			s64 rtt_level_start;
+			u64 rtt_num_start;
+			u64 flags1;
+			u64 max_num_vdevs;
+		};
+		u8 sizer3[0x700];
+	};
+	union { /* 0xf00 */
+		struct {
+			u8 sizer4[0x80];
+			u64 aux_rtt_base[3];
+		};
+		u8 sizer5[0x100];
+	};
+};
+
+static_assert(offsetof(struct realm_params, rpv) == 0x400);
+static_assert(offsetof(struct realm_params, rtt_base) == 0x808);
+static_assert(offsetof(struct realm_params, aux_rtt_base) == 0xf80);
+static_assert(sizeof(struct realm_params) == SZ_4K);
+
+/*
+ * The number of GPRs (starting from X0) that are configured by the host when
+ * a REC is created.
+ */
+#define REC_CREATE_NR_GPRS		8
+
+#define REC_PARAMS_FLAG_RUNNABLE	BIT(0)
+
+struct rec_params {
+	union { /* 0x0 */
+		u64 flags;
+		u8 sizer0[0x100];
+	};
+	union { /* 0x100 */
+		u64 mpidr;
+		u8 sizer1[0x100];
+	};
+	union { /* 0x200 */
+		u64 pc;
+		u8 sizer2[0x100];
+	};
+	union { /* 0x300 */
+		u64 gprs[REC_CREATE_NR_GPRS];
+		u8 sizer3[0xd00];
+	};
+};
+
+static_assert(offsetof(struct rec_params, mpidr) == 0x100);
+static_assert(offsetof(struct rec_params, pc) == 0x200);
+static_assert(offsetof(struct rec_params, gprs) == 0x300);
+static_assert(sizeof(struct rec_params) == SZ_4K);
+
+#define REC_ENTER_FLAG_FORCE_P0		BIT(7)
+#define REC_ENTER_FLAG_DEV_MEM_RESPONSE	BIT(6)
+#define REC_ENTER_FLAG_S2AP_RESPONSE	BIT(5)
+#define REC_ENTER_FLAG_RIPAS_RESPONSE	BIT(4)
+#define REC_ENTER_FLAG_TRAP_WFE		BIT(3)
+#define REC_ENTER_FLAG_TRAP_WFI		BIT(2)
+#define REC_ENTER_FLAG_INJECT_SEA	BIT(1)
+#define REC_ENTER_FLAG_EMULATED_MMIO	BIT(0)
+
+#define REC_RUN_GPRS			31
+
+struct rec_enter {
+	union { /* 0x000 */
+		u64 flags;
+		u8 sizer0[0x200];
+	};
+	union { /* 0x200 */
+		u64 gprs[REC_RUN_GPRS];
+		u8 sizer1[0x600];
+	};
+};
+
+static_assert(offsetof(struct rec_enter, gprs) == 0x200);
+static_assert(sizeof(struct rec_enter) == SZ_2K);
+
+#define RMI_EXIT_SYNC			0x00
+#define RMI_EXIT_IRQ			0x01
+#define RMI_EXIT_FIQ			0x02
+#define RMI_EXIT_PSCI			0x03
+#define RMI_EXIT_RIPAS_CHANGE		0x04
+#define RMI_EXIT_HOST_CALL		0x05
+#define RMI_EXIT_SERROR			0x06
+#define RMI_EXIT_S2AP_CHANGE		0x07
+#define RMI_EXIT_VDEV_VALIDATE_MAPPING	0x08
+#define RMI_EXIT_VSMMU_COMMAND		0x0a
+
+struct rec_exit {
+	union { /* 0x000 */
+		u8 exit_reason;
+		u8 sizer0[0x100];
+	};
+	union { /* 0x100 */
+		struct {
+			u64 esr;
+			u64 far;
+			u64 hpfar;
+			u64 rtt_tree;
+		};
+		u8 sizer1[0x100];
+	};
+	union { /* 0x200 */
+		u64 gprs[REC_RUN_GPRS];
+		u8 sizer2[0x200];
+	};
+	union { /* 0x400 */
+		struct {
+			u64 cntp_ctl;
+			u64 cntp_cval;
+			u64 cntv_ctl;
+			u64 cntv_cval;
+		};
+		u8 sizer4[0x100];
+	};
+	union { /* 0x500 */
+		struct {
+			u64 ripas_base;
+			u64 ripas_top;
+			u8 ripas_value;
+			u8 sizer5[0xf];
+			u64 s2ap_base;
+			u64 s2ap_top;
+			u64 vdev_id_1;
+			u64 vdev_id_2;
+			u64 dev_mem_base;
+			u64 dev_mem_top;
+			u64 dev_mem_pa;
+		};
+		u8 sizer6[0x100];
+	};
+	union { /* 0x600 */
+		struct {
+			u16 imm;
+			u8 sizer7[0x6];
+			u64 plane;
+		};
+		u8 sizer8[0x100];
+	};
+	union { /* 0x700 */
+		struct {
+			u8 pmu_ovf_status;
+			u8 sizer9[0xf];
+			u64 vsmmu;
+		};
+		u8 sizer10[0x100];
+	};
+};
+
+static_assert(sizeof(struct rec_exit) == SZ_2K);
+static_assert(offsetof(struct rec_exit, esr) == 0x100);
+static_assert(offsetof(struct rec_exit, gprs) == 0x200);
+static_assert(offsetof(struct rec_exit, cntp_ctl) == 0x400);
+static_assert(offsetof(struct rec_exit, ripas_base) == 0x500);
+static_assert(offsetof(struct rec_exit, imm) == 0x600);
+static_assert(offsetof(struct rec_exit, pmu_ovf_status) == 0x700);
+
+struct rec_run {
+	struct rec_enter enter;
+	struct rec_exit exit;
+};
+
+static_assert(sizeof(struct rec_run) == SZ_4K);
+
+/* RMI_RTT_UNPROT_MAP_FLAGS definitions */
+#define RMI_RTT_UNPROT_MAP_FLAGS_S2AP_MASK		GENMASK(22, 19)
+#define RMI_RTT_UNPROT_MAP_FLAGS_MEMATTR_MASK		GENMASK(18, 16)
+#define RMI_RTT_UNPROT_MAP_FLAGS_LIST_COUNT_MASK	GENMASK(15, 2)
+#define RMI_RTT_UNPROT_MAP_FLAGS_OADDR_TYPE_MASK	GENMASK(1, 0)
+
+/* RMI_RTT_PROT_MAP_FLAGS definitions */
+#define RMI_RTT_PROT_MAP_FLAGS_LIST_COUNT_MASK		GENMASK(15, 2)
+#define RMI_RTT_PROT_MAP_FLAGS_OADDR_TYPE_MASK		GENMASK(1, 0)
+
+/* S2AP Direct Encodings, used in RMI_RTT_UNPROT_MAP_FLAGS_S2AP */
+#define RMI_S2AP_DIRECT_READ			BIT(1)
+#define RMI_S2AP_DIRECT_WRITE			BIT(0)
+
+#endif /* __LINUX_ARM_SMCCC_RMI_H_ */
diff --git a/include/linux/suspend.h b/include/linux/suspend.h
index b02876f1ae38a..a3815027773c5 100644
--- a/include/linux/suspend.h
+++ b/include/linux/suspend.h
@@ -401,6 +401,7 @@ int hibernate_quiet_exec(int (*func)(void *data), void *data);
 int hibernate_resume_nonboot_cpu_disable(void);
 int arch_hibernation_header_save(void *addr, unsigned int max_size);
 int arch_hibernation_header_restore(void *addr);
+bool arch_hibernation_available(void);
 
 #else /* CONFIG_HIBERNATION */
 static inline void register_nosave_region(unsigned long b, unsigned long e) {}
diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c
index d2479c69d71a4..9d9d53828542f 100644
--- a/kernel/power/hibernate.c
+++ b/kernel/power/hibernate.c
@@ -106,11 +106,17 @@ bool hibernation_in_progress(void)
 	return !atomic_read(&hibernate_atomic);
 }
 
+__weak bool arch_hibernation_available(void)
+{
+	return true;
+}
+
 bool hibernation_available(void)
 {
 	return nohibernate == 0 &&
 		!security_locked_down(LOCKDOWN_HIBERNATION) &&
-		!secretmem_active() && !cxl_mem_active();
+		!secretmem_active() && !cxl_mem_active() &&
+		arch_hibernation_available();
 }
 
 /**

3/2 2026/10/02 04:43 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 32537
							output: 21
							thoughts: 691
						

Thoughts:
**Analyzing this Patch for AMD64 Compatibility**

Okay, let's dive into this patch and the associated review request. The target architecture is listed as amd64, which immediately raises a red flag given the predominance of ARM64-specific changes. My initial concern is whether this is a genuine amd64 patch, or if something has gone awry in the targeting.

First, I'll go through the patch's changes and note the following areas that have been touched. There are many areas in the ARM64 tree, including updates to MAINTAINERS, Kconfig files that are selecting ARM-specific features, and driver code and headers in `arm_rmm`. I also see changes to the `cpufeature.c`, `hibernate.c`, `machine_kexec.c`, and `kvm/sys_regs.c` files, all within the ARM64 architecture tree. This all screams "ARM64-only."

However, I spot a few areas that catch my attention in relation to amd64, namely kernel/power/hibernate.c and include/linux/suspend.h, and drivers/firmware/arm_rmm. The addition of `__weak bool arch_hibernation_available(void)` returning `true` in `hibernate.c` and its declaration in `suspend.h` are the first points of possible amd64 relevance. It seems like the intention is to provide a weak hook for architecture-specific hibernation availability. While the definition itself is a no-op, the mere existence of the function and the call to it in `hibernation_available()` warrants closer inspection.

Then, there's `drivers/firmware/arm_rmm`. Could some of this ARM-specific Realm Management Interface (RMI) code be conditionally enabled on amd64? The Kconfig entry for `ARM_RMM_RMI` is the key: it depends on `ARCH_SUPPORTS_RMM`. I already know `ARCH_SUPPORTS_RMM` is `select`ed in `arch/arm64/Kconfig`, so it appears to be ARM64 specific... but I should not just assume anything. I need to make sure amd64 doesn't inadvertently select it somehow. I will now `grep` the entire source tree for `ARCH_SUPPORTS_RMM` to be absolutely certain it isn't defined or selected elsewhere. This will confirm whether this patch is actually targeted correctly for amd64 or if there's a wider issue in the dependency structure.



4/2 2026/10/02 04:43 tool grepper
0m Args:
{
  "Expression": "ARCH_SUPPORTS_RMM"
}

Results:
{
  "Output": "arch/arm64/Kconfig=2=config ARM64\n--\narch/arm64/Kconfig-40-\tselect ARCH_SUPPORTS_MSEAL_SYSTEM_MAPPINGS\narch/arm64/Kconfig:41:\tselect ARCH_SUPPORTS_RMM\narch/arm64/Kconfig-42-\tselect ARCH_HAS_NMI_SAFE_THIS_CPU_OPS\n--\ndrivers/firmware/arm_rmm/Kconfig-1-\ndrivers/firmware/arm_rmm/Kconfig:2:config ARCH_SUPPORTS_RMM\ndrivers/firmware/arm_rmm/Kconfig-3-\tbool\n--\ndrivers/firmware/arm_rmm/Kconfig=5=config ARM_RMM_RMI\ndrivers/firmware/arm_rmm/Kconfig-6-\tbool \"Realm Management Interface (RMI) Support\"\ndrivers/firmware/arm_rmm/Kconfig:7:\tdepends on ARCH_SUPPORTS_RMM\ndrivers/firmware/arm_rmm/Kconfig-8-\thelp\n"
}

5/2 2026/10/02 04:43 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 34053
							output: 20
							thoughts: 454
						

Thoughts:
**Analyzing the Architecture-Specific Hibernation Code**

Okay, so I'm looking at this code and immediately see a problem: `ARCH_SUPPORTS_RMM` is *only* selected by `ARM64`. That means `CONFIG_ARM_RMM_RMI` is *never* going to be enabled on amd64 (x86_64). Makes sense, RMM is ARM-specific.

Digging deeper in `drivers/firmware/arm_rmm/rmi.c`, it's chock-full of includes like `<asm/memory.h>` and `<asm/pgtable-hwdef.h>` and uses ARM64-specific symbols such as `ARM64_HW_PGTABLE_LEVEL_SHIFT`, `SYS_ID_AA64PFR0_EL1`, `ID_AA64PFR0_EL1_RME_SHIFT`, `read_sanitised_ftr_reg`, `arm_smccc_1_2_regs`... It wouldn't even *compile* on x86, clearly! This part is irrelevant to the x86 code.

Now, let's look at the generic code changes, in `kernel/power/hibernate.c`.

Here's where it gets interesting. They've added this `arch_hibernation_available()` function. On amd64, this weak function defaults to returning `true`. Therefore, the added check `&& arch_hibernation_available()` basically evaluates to `&& true`. That's completely pointless on x86/amd64! It's a no-op. So the question is: does x86 *ever* define `arch_hibernation_available()` differently? Time to do a quick grep.



6/2 2026/10/02 04:43 tool grepper
0m Args:
{
  "Expression": "arch_hibernation_available"
}

Results:
{
  "Output": "arch/arm64/kernel/hibernate.c=106=void notrace restore_processor_state(void)\n--\narch/arm64/kernel/hibernate.c-109-\narch/arm64/kernel/hibernate.c:110:bool arch_hibernation_available(void)\narch/arm64/kernel/hibernate.c-111-{\n--\ninclude/linux/suspend.h=403=int arch_hibernation_header_restore(void *addr);\ninclude/linux/suspend.h:404:bool arch_hibernation_available(void);\ninclude/linux/suspend.h-405-\n--\nkernel/power/hibernate.c=104=bool hibernation_in_progress(void)\n--\nkernel/power/hibernate.c-108-\nkernel/power/hibernate.c:109:__weak bool arch_hibernation_available(void)\nkernel/power/hibernate.c-110-{\n--\nkernel/power/hibernate.c=114=bool hibernation_available(void)\n--\nkernel/power/hibernate.c-118-\t\t!secretmem_active() \u0026\u0026 !cxl_mem_active() \u0026\u0026\nkernel/power/hibernate.c:119:\t\tarch_hibernation_available();\nkernel/power/hibernate.c-120-}\n"
}

7/2 2026/10/02 04:43 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 35244
							output: 176
							thoughts: 1056
						

Thoughts:
**WorthFuzzing Evaluation: Negative Decision**

Okay, let's break this down systematically. Section 2 of the prompt outlines when we should return `WorthFuzzing=false`. It boils down to ensuring *all* modified code falls into one or more specified negative criteria, and in this case, the relevant criterion is "Unreachable execution paths: ... Code paths exclusive to architectures other than the target architecture."

First, I need to understand the scope of the changes. The patch series touches these files: `MAINTAINERS`, `arch/arm64/*`, `drivers/firmware/arm_rmm/*`, `include/linux/arm-rmi-cmds.h`, `include/linux/arm-smccc-rmi.h`, `kernel/power/hibernate.c`, and `include/linux/suspend.h`. A quick glance immediately shows a strong ARM64 bias. I recognize several components related to the Realm Management Extension (RME), specifically the Realm Management Monitor (RMM) and Realm Management Interface (RMI) support. `drivers/firmware/arm_rmm/*` screams ARM-specific.

The crucial point is that almost all changes appear exclusively tied to ARM64. `drivers/firmware/arm_rmm/*` depends on `ARCH_SUPPORTS_RMM`, which is *only* defined for ARM64. Similarly, the include files `include/linux/arm-rmi-cmds.h` and `include/linux/arm-smccc-rmi.h` are ARM-specific and stubbed out when `!CONFIG_ARM_RMM_RMI`.

Now, let's drill down on the generic `kernel/power/hibernate.c` and `include/linux/suspend.h` modifications. `arch_hibernation_available()` is introduced with a `__weak` fallback returning `true`. On an amd64 system, this function *always* returns `true`. This means the behavior is functionally identical to the pre-patch state. There's no new logic or state introduced on amd64. It’s essentially a no-op from a fuzzing perspective, a hook for an ARM64 feature.

To be absolutely certain, I've checked the diff and the file list. Nothing else jumps out that introduces any reachable code paths on amd64. The changes are confined to: `MAINTAINERS`, `arch/arm64/*`, `drivers/firmware/*` (with ARM64-specific subdirectories), and the weak fallback in power management.

Considering the negative criteria, the target architecture is amd64. All functional code is exclusive to ARM64. The fallback in power management, while generic, introduces no behavioral change on amd64. Therefore, the "Unreachable execution paths" criterion is met.

Therefore, `WorthFuzzing` is unequivocally `false`.

The arguments for `set-results` will be:

*   `WorthFuzzing: false`
*   `Reasoning: The patch series adds support for Arm Realm Management Extension (RME) / Realm Management Monitor (RMM) and Realm Management Interface (RMI) for the ARM64 architecture. All new drivers and features (CONFIG_ARM_RMM_RMI) depend on ARCH_SUPPORTS_RMM which is only selected by ARM64 and cannot be built or executed on amd64. The only generic change is a weak fallback function arch_hibernation_available() returning true in kernel/power/hibernate.c, which introduces no behavioral change on amd64. Thus, all functional changes are exclusive to architectures other than the target architecture (amd64).`
*   `FocusSymbols: []`
*   `EnableConfigs: []`

I'm confident in this assessment.



8/2 2026/10/02 04:43 tool set-results
0m Args:
{
  "EnableConfigs": [],
  "FocusSymbols": [],
  "Reasoning": "The patch series introduces support for the Arm Realm Management Extension (RME) and Realm Management Monitor (RMM) / Realm Management Interface (RMI), which is exclusive to the arm64 architecture. The driver code under drivers/firmware/arm_rmm/ depends on ARCH_SUPPORTS_RMM, which is only selected by arm64 and cannot be compiled on amd64. The only generic kernel change is the addition of a weak stub arch_hibernation_available() returning true in kernel/power/hibernate.c, which produces no functional change on amd64. Therefore, all executable logic modified by the patch is exclusive to architectures other than the target architecture (amd64).",
  "WorthFuzzing": false
}

Results:
{
  "EnableConfigs": null,
  "FocusSymbols": [],
  "Reasoning": "The patch series introduces support for the Arm Realm Management Extension (RME) and Realm Management Monitor (RMM) / Realm Management Interface (RMI), which is exclusive to the arm64 architecture. The driver code under drivers/firmware/arm_rmm/ depends on ARCH_SUPPORTS_RMM, which is only selected by arm64 and cannot be compiled on amd64. The only generic kernel change is the addition of a weak stub arch_hibernation_available() returning true in kernel/power/hibernate.c, which produces no functional change on amd64. Therefore, all executable logic modified by the patch is exclusive to architectures other than the target architecture (amd64).",
  "WorthFuzzing": false
}

9/1 2026/10/02 04:43 action If
0m Args:
{
  "WorthFuzzing": false
}

Results:
null

LLM Calls Summary:
Total Calls Total Tokens Avg Tokens Total Duration (Seconds) Avg Duration (Seconds)
Tool Calls Summary:
Total Calls Total Duration (Seconds) Avg Duration (Seconds)