Based on kernel version 4.3. Page generated on 2015-11-02 12:43 EST.
1 Booting AArch64 Linux 2 ===================== 3 4 Author: Will Deacon <email@example.com> 5 Date : 07 September 2012 6 7 This document is based on the ARM booting document by Russell King and 8 is relevant to all public releases of the AArch64 Linux kernel. 9 10 The AArch64 exception model is made up of a number of exception levels 11 (EL0 - EL3), with EL0 and EL1 having a secure and a non-secure 12 counterpart. EL2 is the hypervisor level and exists only in non-secure 13 mode. EL3 is the highest priority level and exists only in secure mode. 14 15 For the purposes of this document, we will use the term `boot loader' 16 simply to define all software that executes on the CPU(s) before control 17 is passed to the Linux kernel. This may include secure monitor and 18 hypervisor code, or it may just be a handful of instructions for 19 preparing a minimal boot environment. 20 21 Essentially, the boot loader should provide (as a minimum) the 22 following: 23 24 1. Setup and initialise the RAM 25 2. Setup the device tree 26 3. Decompress the kernel image 27 4. Call the kernel image 28 29 30 1. Setup and initialise RAM 31 --------------------------- 32 33 Requirement: MANDATORY 34 35 The boot loader is expected to find and initialise all RAM that the 36 kernel will use for volatile data storage in the system. It performs 37 this in a machine dependent manner. (It may use internal algorithms 38 to automatically locate and size all RAM, or it may use knowledge of 39 the RAM in the machine, or any other method the boot loader designer 40 sees fit.) 41 42 43 2. Setup the device tree 44 ------------------------- 45 46 Requirement: MANDATORY 47 48 The device tree blob (dtb) must be placed on an 8-byte boundary and must 49 not exceed 2 megabytes in size. Since the dtb will be mapped cacheable 50 using blocks of up to 2 megabytes in size, it must not be placed within 51 any 2M region which must be mapped with any specific attributes. 52 53 NOTE: versions prior to v4.2 also require that the DTB be placed within 54 the 512 MB region starting at text_offset bytes below the kernel Image. 55 56 3. Decompress the kernel image 57 ------------------------------ 58 59 Requirement: OPTIONAL 60 61 The AArch64 kernel does not currently provide a decompressor and 62 therefore requires decompression (gzip etc.) to be performed by the boot 63 loader if a compressed Image target (e.g. Image.gz) is used. For 64 bootloaders that do not implement this requirement, the uncompressed 65 Image target is available instead. 66 67 68 4. Call the kernel image 69 ------------------------ 70 71 Requirement: MANDATORY 72 73 The decompressed kernel image contains a 64-byte header as follows: 74 75 u32 code0; /* Executable code */ 76 u32 code1; /* Executable code */ 77 u64 text_offset; /* Image load offset, little endian */ 78 u64 image_size; /* Effective Image size, little endian */ 79 u64 flags; /* kernel flags, little endian */ 80 u64 res2 = 0; /* reserved */ 81 u64 res3 = 0; /* reserved */ 82 u64 res4 = 0; /* reserved */ 83 u32 magic = 0x644d5241; /* Magic number, little endian, "ARM\x64" */ 84 u32 res5; /* reserved (used for PE COFF offset) */ 85 86 87 Header notes: 88 89 - As of v3.17, all fields are little endian unless stated otherwise. 90 91 - code0/code1 are responsible for branching to stext. 92 93 - when booting through EFI, code0/code1 are initially skipped. 94 res5 is an offset to the PE header and the PE header has the EFI 95 entry point (efi_stub_entry). When the stub has done its work, it 96 jumps to code0 to resume the normal boot process. 97 98 - Prior to v3.17, the endianness of text_offset was not specified. In 99 these cases image_size is zero and text_offset is 0x80000 in the 100 endianness of the kernel. Where image_size is non-zero image_size is 101 little-endian and must be respected. Where image_size is zero, 102 text_offset can be assumed to be 0x80000. 103 104 - The flags field (introduced in v3.17) is a little-endian 64-bit field 105 composed as follows: 106 Bit 0: Kernel endianness. 1 if BE, 0 if LE. 107 Bits 1-63: Reserved. 108 109 - When image_size is zero, a bootloader should attempt to keep as much 110 memory as possible free for use by the kernel immediately after the 111 end of the kernel image. The amount of space required will vary 112 depending on selected features, and is effectively unbound. 113 114 The Image must be placed text_offset bytes from a 2MB aligned base 115 address near the start of usable system RAM and called there. Memory 116 below that base address is currently unusable by Linux, and therefore it 117 is strongly recommended that this location is the start of system RAM. 118 The region between the 2 MB aligned base address and the start of the 119 image has no special significance to the kernel, and may be used for 120 other purposes. 121 At least image_size bytes from the start of the image must be free for 122 use by the kernel. 123 124 Any memory described to the kernel (even that below the start of the 125 image) which is not marked as reserved from the kernel (e.g., with a 126 memreserve region in the device tree) will be considered as available to 127 the kernel. 128 129 Before jumping into the kernel, the following conditions must be met: 130 131 - Quiesce all DMA capable devices so that memory does not get 132 corrupted by bogus network packets or disk data. This will save 133 you many hours of debug. 134 135 - Primary CPU general-purpose register settings 136 x0 = physical address of device tree blob (dtb) in system RAM. 137 x1 = 0 (reserved for future use) 138 x2 = 0 (reserved for future use) 139 x3 = 0 (reserved for future use) 140 141 - CPU mode 142 All forms of interrupts must be masked in PSTATE.DAIF (Debug, SError, 143 IRQ and FIQ). 144 The CPU must be in either EL2 (RECOMMENDED in order to have access to 145 the virtualisation extensions) or non-secure EL1. 146 147 - Caches, MMUs 148 The MMU must be off. 149 Instruction cache may be on or off. 150 The address range corresponding to the loaded kernel image must be 151 cleaned to the PoC. In the presence of a system cache or other 152 coherent masters with caches enabled, this will typically require 153 cache maintenance by VA rather than set/way operations. 154 System caches which respect the architected cache maintenance by VA 155 operations must be configured and may be enabled. 156 System caches which do not respect architected cache maintenance by VA 157 operations (not recommended) must be configured and disabled. 158 159 - Architected timers 160 CNTFRQ must be programmed with the timer frequency and CNTVOFF must 161 be programmed with a consistent value on all CPUs. If entering the 162 kernel at EL1, CNTHCTL_EL2 must have EL1PCTEN (bit 0) set where 163 available. 164 165 - Coherency 166 All CPUs to be booted by the kernel must be part of the same coherency 167 domain on entry to the kernel. This may require IMPLEMENTATION DEFINED 168 initialisation to enable the receiving of maintenance operations on 169 each CPU. 170 171 - System registers 172 All writable architected system registers at the exception level where 173 the kernel image will be entered must be initialised by software at a 174 higher exception level to prevent execution in an UNKNOWN state. 175 176 For systems with a GICv3 interrupt controller: 177 - If EL3 is present: 178 ICC_SRE_EL3.Enable (bit 3) must be initialiased to 0b1. 179 ICC_SRE_EL3.SRE (bit 0) must be initialised to 0b1. 180 - If the kernel is entered at EL1: 181 ICC.SRE_EL2.Enable (bit 3) must be initialised to 0b1 182 ICC_SRE_EL2.SRE (bit 0) must be initialised to 0b1. 183 184 The requirements described above for CPU mode, caches, MMUs, architected 185 timers, coherency and system registers apply to all CPUs. All CPUs must 186 enter the kernel in the same exception level. 187 188 The boot loader is expected to enter the kernel on each CPU in the 189 following manner: 190 191 - The primary CPU must jump directly to the first instruction of the 192 kernel image. The device tree blob passed by this CPU must contain 193 an 'enable-method' property for each cpu node. The supported 194 enable-methods are described below. 195 196 It is expected that the bootloader will generate these device tree 197 properties and insert them into the blob prior to kernel entry. 198 199 - CPUs with a "spin-table" enable-method must have a 'cpu-release-addr' 200 property in their cpu node. This property identifies a 201 naturally-aligned 64-bit zero-initalised memory location. 202 203 These CPUs should spin outside of the kernel in a reserved area of 204 memory (communicated to the kernel by a /memreserve/ region in the 205 device tree) polling their cpu-release-addr location, which must be 206 contained in the reserved region. A wfe instruction may be inserted 207 to reduce the overhead of the busy-loop and a sev will be issued by 208 the primary CPU. When a read of the location pointed to by the 209 cpu-release-addr returns a non-zero value, the CPU must jump to this 210 value. The value will be written as a single 64-bit little-endian 211 value, so CPUs must convert the read value to their native endianness 212 before jumping to it. 213 214 - CPUs with a "psci" enable method should remain outside of 215 the kernel (i.e. outside of the regions of memory described to the 216 kernel in the memory node, or in a reserved area of memory described 217 to the kernel by a /memreserve/ region in the device tree). The 218 kernel will issue CPU_ON calls as described in ARM document number ARM 219 DEN 0022A ("Power State Coordination Interface System Software on ARM 220 processors") to bring CPUs into the kernel. 221 222 The device tree should contain a 'psci' node, as described in 223 Documentation/devicetree/bindings/arm/psci.txt. 224 225 - Secondary CPU general-purpose register settings 226 x0 = 0 (reserved for future use) 227 x1 = 0 (reserved for future use) 228 x2 = 0 (reserved for future use) 229 x3 = 0 (reserved for future use)