diff options
| -rw-r--r-- | arch/x86/Kconfig | 16 | ||||
| -rw-r--r-- | arch/x86/Kconfig.debug | 13 | ||||
| -rw-r--r-- | arch/x86/include/asm/imr.h | 60 | ||||
| -rw-r--r-- | arch/x86/platform/Makefile | 1 | ||||
| -rw-r--r-- | arch/x86/platform/intel-quark/Makefile | 2 | ||||
| -rw-r--r-- | arch/x86/platform/intel-quark/imr.c | 661 | ||||
| -rw-r--r-- | arch/x86/platform/intel-quark/imr_selftest.c | 129 | ||||
| -rw-r--r-- | drivers/platform/x86/Kconfig | 25 |
8 files changed, 907 insertions, 0 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index eb1cf898ed3c..c2fb8a87dccb 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig | |||
| @@ -488,6 +488,22 @@ config X86_INTEL_MID | |||
| 488 | Intel MID platforms are based on an Intel processor and chipset which | 488 | Intel MID platforms are based on an Intel processor and chipset which |
| 489 | consume less power than most of the x86 derivatives. | 489 | consume less power than most of the x86 derivatives. |
| 490 | 490 | ||
| 491 | config X86_INTEL_QUARK | ||
| 492 | bool "Intel Quark platform support" | ||
| 493 | depends on X86_32 | ||
| 494 | depends on X86_EXTENDED_PLATFORM | ||
| 495 | depends on X86_PLATFORM_DEVICES | ||
| 496 | depends on X86_TSC | ||
| 497 | depends on PCI | ||
| 498 | depends on PCI_GOANY | ||
| 499 | depends on X86_IO_APIC | ||
| 500 | select IOSF_MBI | ||
| 501 | select INTEL_IMR | ||
| 502 | ---help--- | ||
| 503 | Select to include support for Quark X1000 SoC. | ||
| 504 | Say Y here if you have a Quark based system such as the Arduino | ||
| 505 | compatible Intel Galileo. | ||
| 506 | |||
| 491 | config X86_INTEL_LPSS | 507 | config X86_INTEL_LPSS |
| 492 | bool "Intel Low Power Subsystem Support" | 508 | bool "Intel Low Power Subsystem Support" |
| 493 | depends on ACPI | 509 | depends on ACPI |
diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 61bd2ad94281..20028da8ae18 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug | |||
| @@ -313,6 +313,19 @@ config DEBUG_NMI_SELFTEST | |||
| 313 | 313 | ||
| 314 | If unsure, say N. | 314 | If unsure, say N. |
| 315 | 315 | ||
| 316 | config DEBUG_IMR_SELFTEST | ||
| 317 | bool "Isolated Memory Region self test" | ||
| 318 | default n | ||
| 319 | depends on INTEL_IMR | ||
| 320 | ---help--- | ||
| 321 | This option enables automated sanity testing of the IMR code. | ||
| 322 | Some simple tests are run to verify IMR bounds checking, alignment | ||
| 323 | and overlapping. This option is really only useful if you are | ||
| 324 | debugging an IMR memory map or are modifying the IMR code and want to | ||
| 325 | test your changes. | ||
| 326 | |||
| 327 | If unsure say N here. | ||
| 328 | |||
| 316 | config X86_DEBUG_STATIC_CPU_HAS | 329 | config X86_DEBUG_STATIC_CPU_HAS |
| 317 | bool "Debug alternatives" | 330 | bool "Debug alternatives" |
| 318 | depends on DEBUG_KERNEL | 331 | depends on DEBUG_KERNEL |
diff --git a/arch/x86/include/asm/imr.h b/arch/x86/include/asm/imr.h new file mode 100644 index 000000000000..cd2ce4068441 --- /dev/null +++ b/arch/x86/include/asm/imr.h | |||
| @@ -0,0 +1,60 @@ | |||
| 1 | /* | ||
| 2 | * imr.h: Isolated Memory Region API | ||
| 3 | * | ||
| 4 | * Copyright(c) 2013 Intel Corporation. | ||
| 5 | * Copyright(c) 2015 Bryan O'Donoghue <pure.logic@nexus-software.ie> | ||
| 6 | * | ||
| 7 | * This program is free software; you can redistribute it and/or | ||
| 8 | * modify it under the terms of the GNU General Public License | ||
| 9 | * as published by the Free Software Foundation; version 2 | ||
| 10 | * of the License. | ||
| 11 | */ | ||
| 12 | #ifndef _IMR_H | ||
| 13 | #define _IMR_H | ||
| 14 | |||
| 15 | #include <linux/types.h> | ||
| 16 | |||
| 17 | /* | ||
| 18 | * IMR agent access mask bits | ||
| 19 | * See section 12.7.4.7 from quark-x1000-datasheet.pdf for register | ||
| 20 | * definitions. | ||
| 21 | */ | ||
| 22 | #define IMR_ESRAM_FLUSH BIT(31) | ||
| 23 | #define IMR_CPU_SNOOP BIT(30) /* Applicable only to write */ | ||
| 24 | #define IMR_RMU BIT(29) | ||
| 25 | #define IMR_VC1_SAI_ID3 BIT(15) | ||
| 26 | #define IMR_VC1_SAI_ID2 BIT(14) | ||
| 27 | #define IMR_VC1_SAI_ID1 BIT(13) | ||
| 28 | #define IMR_VC1_SAI_ID0 BIT(12) | ||
| 29 | #define IMR_VC0_SAI_ID3 BIT(11) | ||
| 30 | #define IMR_VC0_SAI_ID2 BIT(10) | ||
| 31 | #define IMR_VC0_SAI_ID1 BIT(9) | ||
| 32 | #define IMR_VC0_SAI_ID0 BIT(8) | ||
| 33 | #define IMR_CPU_0 BIT(1) /* SMM mode */ | ||
| 34 | #define IMR_CPU BIT(0) /* Non SMM mode */ | ||
| 35 | #define IMR_ACCESS_NONE 0 | ||
| 36 | |||
| 37 | /* | ||
| 38 | * Read/Write access-all bits here include some reserved bits | ||
| 39 | * These are the values firmware uses and are accepted by hardware. | ||
| 40 | * The kernel defines read/write access-all in the same way as firmware | ||
| 41 | * in order to have a consistent and crisp definition across firmware, | ||
| 42 | * bootloader and kernel. | ||
| 43 | */ | ||
| 44 | #define IMR_READ_ACCESS_ALL 0xBFFFFFFF | ||
| 45 | #define IMR_WRITE_ACCESS_ALL 0xFFFFFFFF | ||
| 46 | |||
| 47 | /* Number of IMRs provided by Quark X1000 SoC */ | ||
| 48 | #define QUARK_X1000_IMR_MAX 0x08 | ||
| 49 | #define QUARK_X1000_IMR_REGBASE 0x40 | ||
| 50 | |||
| 51 | /* IMR alignment bits - only bits 31:10 are checked for IMR validity */ | ||
| 52 | #define IMR_ALIGN 0x400 | ||
| 53 | #define IMR_MASK (IMR_ALIGN - 1) | ||
| 54 | |||
| 55 | int imr_add_range(phys_addr_t base, size_t size, | ||
| 56 | unsigned int rmask, unsigned int wmask, bool lock); | ||
| 57 | |||
| 58 | int imr_remove_range(phys_addr_t base, size_t size); | ||
| 59 | |||
| 60 | #endif /* _IMR_H */ | ||
diff --git a/arch/x86/platform/Makefile b/arch/x86/platform/Makefile index 85afde1fa3e5..a62e0be3a2f1 100644 --- a/arch/x86/platform/Makefile +++ b/arch/x86/platform/Makefile | |||
| @@ -5,6 +5,7 @@ obj-y += geode/ | |||
| 5 | obj-y += goldfish/ | 5 | obj-y += goldfish/ |
| 6 | obj-y += iris/ | 6 | obj-y += iris/ |
| 7 | obj-y += intel-mid/ | 7 | obj-y += intel-mid/ |
| 8 | obj-y += intel-quark/ | ||
| 8 | obj-y += olpc/ | 9 | obj-y += olpc/ |
| 9 | obj-y += scx200/ | 10 | obj-y += scx200/ |
| 10 | obj-y += sfi/ | 11 | obj-y += sfi/ |
diff --git a/arch/x86/platform/intel-quark/Makefile b/arch/x86/platform/intel-quark/Makefile new file mode 100644 index 000000000000..9cc57ed36022 --- /dev/null +++ b/arch/x86/platform/intel-quark/Makefile | |||
| @@ -0,0 +1,2 @@ | |||
| 1 | obj-$(CONFIG_INTEL_IMR) += imr.o | ||
| 2 | obj-$(CONFIG_DEBUG_IMR_SELFTEST) += imr_selftest.o | ||
diff --git a/arch/x86/platform/intel-quark/imr.c b/arch/x86/platform/intel-quark/imr.c new file mode 100644 index 000000000000..0ee619f9fcb7 --- /dev/null +++ b/arch/x86/platform/intel-quark/imr.c | |||
| @@ -0,0 +1,661 @@ | |||
| 1 | /** | ||
| 2 | * imr.c | ||
| 3 | * | ||
| 4 | * Copyright(c) 2013 Intel Corporation. | ||
| 5 | * Copyright(c) 2015 Bryan O'Donoghue <pure.logic@nexus-software.ie> | ||
| 6 | * | ||
| 7 | * IMR registers define an isolated region of memory that can | ||
| 8 | * be masked to prohibit certain system agents from accessing memory. | ||
| 9 | * When a device behind a masked port performs an access - snooped or | ||
| 10 | * not, an IMR may optionally prevent that transaction from changing | ||
| 11 | * the state of memory or from getting correct data in response to the | ||
| 12 | * operation. | ||
| 13 | * | ||
| 14 | * Write data will be dropped and reads will return 0xFFFFFFFF, the | ||
| 15 | * system will reset and system BIOS will print out an error message to | ||
| 16 | * inform the user that an IMR has been violated. | ||
| 17 | * | ||
| 18 | * This code is based on the Linux MTRR code and reference code from | ||
| 19 | * Intel's Quark BSP EFI, Linux and grub code. | ||
| 20 | * | ||
| 21 | * See quark-x1000-datasheet.pdf for register definitions. | ||
| 22 | * http://www.intel.com/content/dam/www/public/us/en/documents/datasheets/quark-x1000-datasheet.pdf | ||
| 23 | */ | ||
| 24 | |||
| 25 | #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt | ||
| 26 | |||
| 27 | #include <asm-generic/sections.h> | ||
| 28 | #include <asm/cpu_device_id.h> | ||
| 29 | #include <asm/imr.h> | ||
| 30 | #include <asm/iosf_mbi.h> | ||
| 31 | #include <linux/debugfs.h> | ||
| 32 | #include <linux/init.h> | ||
| 33 | #include <linux/mm.h> | ||
| 34 | #include <linux/module.h> | ||
| 35 | #include <linux/types.h> | ||
| 36 | |||
| 37 | struct imr_device { | ||
| 38 | struct dentry *file; | ||
| 39 | bool init; | ||
| 40 | struct mutex lock; | ||
| 41 | int max_imr; | ||
| 42 | int reg_base; | ||
| 43 | }; | ||
| 44 | |||
| 45 | static struct imr_device imr_dev; | ||
| 46 | |||
| 47 | /* | ||
| 48 | * IMR read/write mask control registers. | ||
| 49 | * See quark-x1000-datasheet.pdf sections 12.7.4.5 and 12.7.4.6 for | ||
| 50 | * bit definitions. | ||
| 51 | * | ||
| 52 | * addr_hi | ||
| 53 | * 31 Lock bit | ||
| 54 | * 30:24 Reserved | ||
| 55 | * 23:2 1 KiB aligned lo address | ||
| 56 | * 1:0 Reserved | ||
| 57 | * | ||
| 58 | * addr_hi | ||
| 59 | * 31:24 Reserved | ||
| 60 | * 23:2 1 KiB aligned hi address | ||
| 61 | * 1:0 Reserved | ||
| 62 | */ | ||
| 63 | #define IMR_LOCK BIT(31) | ||
| 64 | |||
| 65 | struct imr_regs { | ||
| 66 | u32 addr_lo; | ||
| 67 | u32 addr_hi; | ||
| 68 | u32 rmask; | ||
| 69 | u32 wmask; | ||
| 70 | }; | ||
| 71 | |||
| 72 | #define IMR_NUM_REGS (sizeof(struct imr_regs)/sizeof(u32)) | ||
| 73 | #define IMR_SHIFT 8 | ||
| 74 | #define imr_to_phys(x) ((x) << IMR_SHIFT) | ||
| 75 | #define phys_to_imr(x) ((x) >> IMR_SHIFT) | ||
| 76 | |||
| 77 | /** | ||
| 78 | * imr_is_enabled - true if an IMR is enabled false otherwise. | ||
| 79 | * | ||
| 80 | * Determines if an IMR is enabled based on address range and read/write | ||
| 81 | * mask. An IMR set with an address range set to zero and a read/write | ||
| 82 | * access mask set to all is considered to be disabled. An IMR in any | ||
| 83 | * other state - for example set to zero but without read/write access | ||
| 84 | * all is considered to be enabled. This definition of disabled is how | ||
| 85 | * firmware switches off an IMR and is maintained in kernel for | ||
| 86 | * consistency. | ||
| 87 | * | ||
| 88 | * @imr: pointer to IMR descriptor. | ||
| 89 | * @return: true if IMR enabled false if disabled. | ||
| 90 | */ | ||
| 91 | static inline int imr_is_enabled(struct imr_regs *imr) | ||
| 92 | { | ||
| 93 | return !(imr->rmask == IMR_READ_ACCESS_ALL && | ||
| 94 | imr->wmask == IMR_WRITE_ACCESS_ALL && | ||
| 95 | imr_to_phys(imr->addr_lo) == 0 && | ||
| 96 | imr_to_phys(imr->addr_hi) == 0); | ||
| 97 | } | ||
| 98 | |||
| 99 | /** | ||
| 100 | * imr_read - read an IMR at a given index. | ||
| 101 | * | ||
| 102 | * Requires caller to hold imr mutex. | ||
| 103 | * | ||
| 104 | * @idev: pointer to imr_device structure. | ||
| 105 | * @imr_id: IMR entry to read. | ||
| 106 | * @imr: IMR structure representing address and access masks. | ||
| 107 | * @return: 0 on success or error code passed from mbi_iosf on failure. | ||
| 108 | */ | ||
| 109 | static int imr_read(struct imr_device *idev, u32 imr_id, struct imr_regs *imr) | ||
| 110 | { | ||
| 111 | u32 reg = imr_id * IMR_NUM_REGS + idev->reg_base; | ||
| 112 | int ret; | ||
| 113 | |||
| 114 | ret = iosf_mbi_read(QRK_MBI_UNIT_MM, QRK_MBI_MM_READ, | ||
| 115 | reg++, &imr->addr_lo); | ||
| 116 | if (ret) | ||
| 117 | return ret; | ||
| 118 | |||
| 119 | ret = iosf_mbi_read(QRK_MBI_UNIT_MM, QRK_MBI_MM_READ, | ||
| 120 | reg++, &imr->addr_hi); | ||
| 121 | if (ret) | ||
| 122 | return ret; | ||
| 123 | |||
| 124 | ret = iosf_mbi_read(QRK_MBI_UNIT_MM, QRK_MBI_MM_READ, | ||
| 125 | reg++, &imr->rmask); | ||
| 126 | if (ret) | ||
| 127 | return ret; | ||
| 128 | |||
| 129 | return iosf_mbi_read(QRK_MBI_UNIT_MM, QRK_MBI_MM_READ, | ||
| 130 | reg++, &imr->wmask); | ||
| 131 | } | ||
| 132 | |||
| 133 | /** | ||
| 134 | * imr_write - write an IMR at a given index. | ||
| 135 | * | ||
| 136 | * Requires caller to hold imr mutex. | ||
| 137 | * Note lock bits need to be written independently of address bits. | ||
| 138 | * | ||
| 139 | * @idev: pointer to imr_device structure. | ||
| 140 | * @imr_id: IMR entry to write. | ||
| 141 | * @imr: IMR structure representing address and access masks. | ||
| 142 | * @lock: indicates if the IMR lock bit should be applied. | ||
| 143 | * @return: 0 on success or error code passed from mbi_iosf on failure. | ||
| 144 | */ | ||
| 145 | static int imr_write(struct imr_device *idev, u32 imr_id, | ||
| 146 | struct imr_regs *imr, bool lock) | ||
| 147 | { | ||
| 148 | unsigned long flags; | ||
| 149 | u32 reg = imr_id * IMR_NUM_REGS + idev->reg_base; | ||
| 150 | int ret; | ||
| 151 | |||
| 152 | local_irq_save(flags); | ||
| 153 | |||
| 154 | ret = iosf_mbi_write(QRK_MBI_UNIT_MM, QRK_MBI_MM_WRITE, reg++, | ||
| 155 | imr->addr_lo); | ||
| 156 | if (ret) | ||
| 157 | goto failed; | ||
| 158 | |||
| 159 | ret = iosf_mbi_write(QRK_MBI_UNIT_MM, QRK_MBI_MM_WRITE, | ||
| 160 | reg++, imr->addr_hi); | ||
| 161 | if (ret) | ||
| 162 | goto failed; | ||
| 163 | |||
| 164 | ret = iosf_mbi_write(QRK_MBI_UNIT_MM, QRK_MBI_MM_WRITE, | ||
| 165 | reg++, imr->rmask); | ||
| 166 | if (ret) | ||
| 167 | goto failed; | ||
| 168 | |||
| 169 | ret = iosf_mbi_write(QRK_MBI_UNIT_MM, QRK_MBI_MM_WRITE, | ||
| 170 | reg++, imr->wmask); | ||
| 171 | if (ret) | ||
| 172 | goto failed; | ||
| 173 | |||
| 174 | /* Lock bit must be set separately to addr_lo address bits. */ | ||
| 175 | if (lock) { | ||
| 176 | imr->addr_lo |= IMR_LOCK; | ||
| 177 | ret = iosf_mbi_write(QRK_MBI_UNIT_MM, QRK_MBI_MM_WRITE, | ||
| 178 | reg - IMR_NUM_REGS, imr->addr_lo); | ||
| 179 | if (ret) | ||
| 180 | goto failed; | ||
| 181 | } | ||
| 182 | |||
| 183 | local_irq_restore(flags); | ||
| 184 | return 0; | ||
| 185 | failed: | ||
| 186 | /* | ||
| 187 | * If writing to the IOSF failed then we're in an unknown state, | ||
| 188 | * likely a very bad state. An IMR in an invalid state will almost | ||
| 189 | * certainly lead to a memory access violation. | ||
| 190 | */ | ||
| 191 | local_irq_restore(flags); | ||
| 192 | WARN(ret, "IOSF-MBI write fail range 0x%08x-0x%08x unreliable\n", | ||
| 193 | imr_to_phys(imr->addr_lo), imr_to_phys(imr->addr_hi) + IMR_MASK); | ||
| 194 | |||
| 195 | return ret; | ||
| 196 | } | ||
| 197 | |||
| 198 | /** | ||
| 199 | * imr_dbgfs_state_show - print state of IMR registers. | ||
| 200 | * | ||
| 201 | * @s: pointer to seq_file for output. | ||
| 202 | * @unused: unused parameter. | ||
| 203 | * @return: 0 on success or error code passed from mbi_iosf on failure. | ||
| 204 | */ | ||
| 205 | static int imr_dbgfs_state_show(struct seq_file *s, void *unused) | ||
| 206 | { | ||
| 207 | phys_addr_t base; | ||
| 208 | phys_addr_t end; | ||
| 209 | int i; | ||
| 210 | struct imr_device *idev = s->private; | ||
| 211 | struct imr_regs imr; | ||
| 212 | size_t size; | ||
| 213 | int ret = -ENODEV; | ||
| 214 | |||
| 215 | mutex_lock(&idev->lock); | ||
| 216 | |||
| 217 | for (i = 0; i < idev->max_imr; i++) { | ||
| 218 | |||
| 219 | ret = imr_read(idev, i, &imr); | ||
| 220 | if (ret) | ||
| 221 | break; | ||
| 222 | |||
| 223 | /* | ||
| 224 | * Remember to add IMR_ALIGN bytes to size to indicate the | ||
| 225 | * inherent IMR_ALIGN size bytes contained in the masked away | ||
| 226 | * lower ten bits. | ||
| 227 | */ | ||
| 228 | if (imr_is_enabled(&imr)) { | ||
| 229 | base = imr_to_phys(imr.addr_lo); | ||
| 230 | end = imr_to_phys(imr.addr_hi) + IMR_MASK; | ||
| 231 | } else { | ||
| 232 | base = 0; | ||
| 233 | end = 0; | ||
| 234 | } | ||
| 235 | size = end - base; | ||
| 236 | seq_printf(s, "imr%02i: base=%pa, end=%pa, size=0x%08zx " | ||
| 237 | "rmask=0x%08x, wmask=0x%08x, %s, %s\n", i, | ||
| 238 | &base, &end, size, imr.rmask, imr.wmask, | ||
| 239 | imr_is_enabled(&imr) ? "enabled " : "disabled", | ||
| 240 | imr.addr_lo & IMR_LOCK ? "locked" : "unlocked"); | ||
| 241 | } | ||
| 242 | |||
| 243 | mutex_unlock(&idev->lock); | ||
| 244 | return ret; | ||
| 245 | } | ||
| 246 | |||
| 247 | /** | ||
| 248 | * imr_state_open - debugfs open callback. | ||
| 249 | * | ||
| 250 | * @inode: pointer to struct inode. | ||
| 251 | * @file: pointer to struct file. | ||
| 252 | * @return: result of single open. | ||
| 253 | */ | ||
| 254 | static int imr_state_open(struct inode *inode, struct file *file) | ||
| 255 | { | ||
| 256 | return single_open(file, imr_dbgfs_state_show, inode->i_private); | ||
| 257 | } | ||
| 258 | |||
| 259 | static const struct file_operations imr_state_ops = { | ||
| 260 | .open = imr_state_open, | ||
| 261 | .read = seq_read, | ||
| 262 | .llseek = seq_lseek, | ||
| 263 | .release = single_release, | ||
| 264 | }; | ||
| 265 | |||
| 266 | /** | ||
| 267 | * imr_debugfs_register - register debugfs hooks. | ||
| 268 | * | ||
| 269 | * @idev: pointer to imr_device structure. | ||
| 270 | * @return: 0 on success - errno on failure. | ||
| 271 | */ | ||
| 272 | static int imr_debugfs_register(struct imr_device *idev) | ||
| 273 | { | ||
| 274 | idev->file = debugfs_create_file("imr_state", S_IFREG | S_IRUGO, NULL, | ||
| 275 | idev, &imr_state_ops); | ||
| 276 | return PTR_ERR_OR_ZERO(idev->file); | ||
| 277 | } | ||
| 278 | |||
| 279 | /** | ||
| 280 | * imr_debugfs_unregister - unregister debugfs hooks. | ||
| 281 | * | ||
| 282 | * @idev: pointer to imr_device structure. | ||
| 283 | * @return: | ||
| 284 | */ | ||
| 285 | static void imr_debugfs_unregister(struct imr_device *idev) | ||
| 286 | { | ||
| 287 | debugfs_remove(idev->file); | ||
| 288 | } | ||
| 289 | |||
| 290 | /** | ||
| 291 | * imr_check_params - check passed address range IMR alignment and non-zero size | ||
| 292 | * | ||
| 293 | * @base: base address of intended IMR. | ||
| 294 | * @size: size of intended IMR. | ||
| 295 | * @return: zero on valid range -EINVAL on unaligned base/size. | ||
| 296 | */ | ||
| 297 | static int imr_check_params(phys_addr_t base, size_t size) | ||
| 298 | { | ||
| 299 | if ((base & IMR_MASK) || (size & IMR_MASK)) { | ||
| 300 | pr_err("base %pa size 0x%08zx must align to 1KiB\n", | ||
| 301 | &base, size); | ||
| 302 | return -EINVAL; | ||
| 303 | } | ||
| 304 | if (size == 0) | ||
| 305 | return -EINVAL; | ||
| 306 | |||
| 307 | return 0; | ||
| 308 | } | ||
| 309 | |||
| 310 | /** | ||
| 311 | * imr_raw_size - account for the IMR_ALIGN bytes that addr_hi appends. | ||
| 312 | * | ||
| 313 | * IMR addr_hi has a built in offset of plus IMR_ALIGN (0x400) bytes from the | ||
| 314 | * value in the register. We need to subtract IMR_ALIGN bytes from input sizes | ||
| 315 | * as a result. | ||
| 316 | * | ||
| 317 | * @size: input size bytes. | ||
| 318 | * @return: reduced size. | ||
| 319 | */ | ||
| 320 | static inline size_t imr_raw_size(size_t size) | ||
| 321 | { | ||
| 322 | return size - IMR_ALIGN; | ||
| 323 | } | ||
| 324 | |||
| 325 | /** | ||
| 326 | * imr_address_overlap - detects an address overlap. | ||
| 327 | * | ||
| 328 | * @addr: address to check against an existing IMR. | ||
| 329 | * @imr: imr being checked. | ||
| 330 | * @return: true for overlap false for no overlap. | ||
| 331 | */ | ||
| 332 | static inline int imr_address_overlap(phys_addr_t addr, struct imr_regs *imr) | ||
| 333 | { | ||
| 334 | return addr >= imr_to_phys(imr->addr_lo) && addr <= imr_to_phys(imr->addr_hi); | ||
| 335 | } | ||
| 336 | |||
| 337 | /** | ||
| 338 | * imr_add_range - add an Isolated Memory Region. | ||
| 339 | * | ||
| 340 | * @base: physical base address of region aligned to 1KiB. | ||
| 341 | * @size: physical size of region in bytes must be aligned to 1KiB. | ||
| 342 | * @read_mask: read access mask. | ||
| 343 | * @write_mask: write access mask. | ||
| 344 | * @lock: indicates whether or not to permanently lock this region. | ||
| 345 | * @return: zero on success or negative value indicating error. | ||
| 346 | */ | ||
| 347 | int imr_add_range(phys_addr_t base, size_t size, | ||
| 348 | unsigned int rmask, unsigned int wmask, bool lock) | ||
| 349 | { | ||
| 350 | phys_addr_t end; | ||
| 351 | unsigned int i; | ||
| 352 | struct imr_device *idev = &imr_dev; | ||
| 353 | struct imr_regs imr; | ||
| 354 | size_t raw_size; | ||
| 355 | int reg; | ||
| 356 | int ret; | ||
| 357 | |||
| 358 | if (WARN_ONCE(idev->init == false, "driver not initialized")) | ||
| 359 | return -ENODEV; | ||
| 360 | |||
| 361 | ret = imr_check_params(base, size); | ||
| 362 | if (ret) | ||
| 363 | return ret; | ||
| 364 | |||
| 365 | /* Tweak the size value. */ | ||
| 366 | raw_size = imr_raw_size(size); | ||
| 367 | end = base + raw_size; | ||
| 368 | |||
| 369 | /* | ||
| 370 | * Check for reserved IMR value common to firmware, kernel and grub | ||
| 371 | * indicating a disabled IMR. | ||
| 372 | */ | ||
| 373 | imr.addr_lo = phys_to_imr(base); | ||
| 374 | imr.addr_hi = phys_to_imr(end); | ||
| 375 | imr.rmask = rmask; | ||
| 376 | imr.wmask = wmask; | ||
| 377 | if (!imr_is_enabled(&imr)) | ||
| 378 | return -ENOTSUPP; | ||
| 379 | |||
| 380 | mutex_lock(&idev->lock); | ||
| 381 | |||
| 382 | /* | ||
| 383 | * Find a free IMR while checking for an existing overlapping range. | ||
| 384 | * Note there's no restriction in silicon to prevent IMR overlaps. | ||
| 385 | * For the sake of simplicity and ease in defining/debugging an IMR | ||
| 386 | * memory map we exclude IMR overlaps. | ||
| 387 | */ | ||
| 388 | reg = -1; | ||
| 389 | for (i = 0; i < idev->max_imr; i++) { | ||
| 390 | ret = imr_read(idev, i, &imr); | ||
| 391 | if (ret) | ||
| 392 | goto failed; | ||
| 393 | |||
| 394 | /* Find overlap @ base or end of requested range. */ | ||
| 395 | ret = -EINVAL; | ||
| 396 | if (imr_is_enabled(&imr)) { | ||
| 397 | if (imr_address_overlap(base, &imr)) | ||
| 398 | goto failed; | ||
| 399 | if (imr_address_overlap(end, &imr)) | ||
| 400 | goto failed; | ||
| 401 | } else { | ||
| 402 | reg = i; | ||
| 403 | } | ||
| 404 | } | ||
| 405 | |||
| 406 | /* Error out if we have no free IMR entries. */ | ||
| 407 | if (reg == -1) { | ||
| 408 | ret = -ENOMEM; | ||
| 409 | goto failed; | ||
| 410 | } | ||
| 411 | |||
| 412 | pr_debug("add %d phys %pa-%pa size %zx mask 0x%08x wmask 0x%08x\n", | ||
| 413 | reg, &base, &end, raw_size, rmask, wmask); | ||
| 414 | |||
| 415 | /* Enable IMR at specified range and access mask. */ | ||
| 416 | imr.addr_lo = phys_to_imr(base); | ||
| 417 | imr.addr_hi = phys_to_imr(end); | ||
| 418 | imr.rmask = rmask; | ||
| 419 | imr.wmask = wmask; | ||
| 420 | |||
| 421 | ret = imr_write(idev, reg, &imr, lock); | ||
| 422 | if (ret < 0) { | ||
| 423 | /* | ||
| 424 | * In the highly unlikely event iosf_mbi_write failed | ||
| 425 | * attempt to rollback the IMR setup skipping the trapping | ||
| 426 | * of further IOSF write failures. | ||
| 427 | */ | ||
| 428 | imr.addr_lo = 0; | ||
| 429 | imr.addr_hi = 0; | ||
| 430 | imr.rmask = IMR_READ_ACCESS_ALL; | ||
| 431 | imr.wmask = IMR_WRITE_ACCESS_ALL; | ||
| 432 | imr_write(idev, reg, &imr, false); | ||
| 433 | } | ||
| 434 | failed: | ||
| 435 | mutex_unlock(&idev->lock); | ||
| 436 | return ret; | ||
| 437 | } | ||
| 438 | EXPORT_SYMBOL_GPL(imr_add_range); | ||
| 439 | |||
| 440 | /** | ||
| 441 | * __imr_remove_range - delete an Isolated Memory Region. | ||
| 442 | * | ||
| 443 | * This function allows you to delete an IMR by its index specified by reg or | ||
| 444 | * by address range specified by base and size respectively. If you specify an | ||
| 445 | * index on its own the base and size parameters are ignored. | ||
| 446 | * imr_remove_range(0, base, size); delete IMR at index 0 base/size ignored. | ||
| 447 | * imr_remove_range(-1, base, size); delete IMR from base to base+size. | ||
| 448 | * | ||
| 449 | * @reg: imr index to remove. | ||
| 450 | * @base: physical base address of region aligned to 1 KiB. | ||
| 451 | * @size: physical size of region in bytes aligned to 1 KiB. | ||
| 452 | * @return: -EINVAL on invalid range or out or range id | ||
| 453 | * -ENODEV if reg is valid but no IMR exists or is locked | ||
| 454 | * 0 on success. | ||
| 455 | */ | ||
| 456 | static int __imr_remove_range(int reg, phys_addr_t base, size_t size) | ||
| 457 | { | ||
| 458 | phys_addr_t end; | ||
| 459 | bool found = false; | ||
| 460 | unsigned int i; | ||
| 461 | struct imr_device *idev = &imr_dev; | ||
| 462 | struct imr_regs imr; | ||
| 463 | size_t raw_size; | ||
| 464 | int ret = 0; | ||
| 465 | |||
| 466 | if (WARN_ONCE(idev->init == false, "driver not initialized")) | ||
| 467 | return -ENODEV; | ||
| 468 | |||
| 469 | /* | ||
| 470 | * Validate address range if deleting by address, else we are | ||
| 471 | * deleting by index where base and size will be ignored. | ||
| 472 | */ | ||
| 473 | if (reg == -1) { | ||
| 474 | ret = imr_check_params(base, size); | ||
| 475 | if (ret) | ||
| 476 | return ret; | ||
| 477 | } | ||
| 478 | |||
| 479 | /* Tweak the size value. */ | ||
| 480 | raw_size = imr_raw_size(size); | ||
| 481 | end = base + raw_size; | ||
| 482 | |||
| 483 | mutex_lock(&idev->lock); | ||
| 484 | |||
| 485 | if (reg >= 0) { | ||
| 486 | /* If a specific IMR is given try to use it. */ | ||
| 487 | ret = imr_read(idev, reg, &imr); | ||
| 488 | if (ret) | ||
| 489 | goto failed; | ||
| 490 | |||
| 491 | if (!imr_is_enabled(&imr) || imr.addr_lo & IMR_LOCK) { | ||
| 492 | ret = -ENODEV; | ||
| 493 | goto failed; | ||
| 494 | } | ||
| 495 | found = true; | ||
| 496 | } else { | ||
| 497 | /* Search for match based on address range. */ | ||
| 498 | for (i = 0; i < idev->max_imr; i++) { | ||
| 499 | ret = imr_read(idev, i, &imr); | ||
| 500 | if (ret) | ||
| 501 | goto failed; | ||
| 502 | |||
| 503 | if (!imr_is_enabled(&imr) || imr.addr_lo & IMR_LOCK) | ||
| 504 | continue; | ||
| 505 | |||
| 506 | if ((imr_to_phys(imr.addr_lo) == base) && | ||
| 507 | (imr_to_phys(imr.addr_hi) == end)) { | ||
| 508 | found = true; | ||
| 509 | reg = i; | ||
| 510 | break; | ||
| 511 | } | ||
| 512 | } | ||
| 513 | } | ||
| 514 | |||
| 515 | if (!found) { | ||
| 516 | ret = -ENODEV; | ||
| 517 | goto failed; | ||
| 518 | } | ||
| 519 | |||
| 520 | pr_debug("remove %d phys %pa-%pa size %zx\n", reg, &base, &end, raw_size); | ||
| 521 | |||
| 522 | /* Tear down the IMR. */ | ||
| 523 | imr.addr_lo = 0; | ||
| 524 | imr.addr_hi = 0; | ||
| 525 | imr.rmask = IMR_READ_ACCESS_ALL; | ||
| 526 | imr.wmask = IMR_WRITE_ACCESS_ALL; | ||
| 527 | |||
| 528 | ret = imr_write(idev, reg, &imr, false); | ||
| 529 | |||
| 530 | failed: | ||
| 531 | mutex_unlock(&idev->lock); | ||
| 532 | return ret; | ||
| 533 | } | ||
| 534 | |||
| 535 | /** | ||
| 536 | * imr_remove_range - delete an Isolated Memory Region by address | ||
| 537 | * | ||
| 538 | * This function allows you to delete an IMR by an address range specified | ||
| 539 | * by base and size respectively. | ||
| 540 | * imr_remove_range(base, size); delete IMR from base to base+size. | ||
| 541 | * | ||
| 542 | * @base: physical base address of region aligned to 1 KiB. | ||
| 543 | * @size: physical size of region in bytes aligned to 1 KiB. | ||
| 544 | * @return: -EINVAL on invalid range or out or range id | ||
| 545 | * -ENODEV if reg is valid but no IMR exists or is locked | ||
| 546 | * 0 on success. | ||
| 547 | */ | ||
| 548 | int imr_remove_range(phys_addr_t base, size_t size) | ||
| 549 | { | ||
| 550 | return __imr_remove_range(-1, base, size); | ||
| 551 | } | ||
| 552 | EXPORT_SYMBOL_GPL(imr_remove_range); | ||
| 553 | |||
| 554 | /** | ||
| 555 | * imr_clear - delete an Isolated Memory Region by index | ||
| 556 | * | ||
| 557 | * This function allows you to delete an IMR by an address range specified | ||
| 558 | * by the index of the IMR. Useful for initial sanitization of the IMR | ||
| 559 | * address map. | ||
| 560 | * imr_ge(base, size); delete IMR from base to base+size. | ||
| 561 | * | ||
| 562 | * @reg: imr index to remove. | ||
| 563 | * @return: -EINVAL on invalid range or out or range id | ||
| 564 | * -ENODEV if reg is valid but no IMR exists or is locked | ||
| 565 | * 0 on success. | ||
| 566 | */ | ||
| 567 | static inline int imr_clear(int reg) | ||
| 568 | { | ||
| 569 | return __imr_remove_range(reg, 0, 0); | ||
| 570 | } | ||
| 571 | |||
| 572 | /** | ||
| 573 | * imr_fixup_memmap - Tear down IMRs used during bootup. | ||
| 574 | * | ||
| 575 | * BIOS and Grub both setup IMRs around compressed kernel, initrd memory | ||
| 576 | * that need to be removed before the kernel hands out one of the IMR | ||
| 577 | * encased addresses to a downstream DMA agent such as the SD or Ethernet. | ||
| 578 | * IMRs on Galileo are setup to immediately reset the system on violation. | ||
| 579 | * As a result if you're running a root filesystem from SD - you'll need | ||
| 580 | * the boot-time IMRs torn down or you'll find seemingly random resets when | ||
| 581 | * using your filesystem. | ||
| 582 | * | ||
| 583 | * @idev: pointer to imr_device structure. | ||
| 584 | * @return: | ||
| 585 | */ | ||
| 586 | static void __init imr_fixup_memmap(struct imr_device *idev) | ||
| 587 | { | ||
| 588 | phys_addr_t base = virt_to_phys(&_text); | ||
| 589 | size_t size = virt_to_phys(&__end_rodata) - base; | ||
| 590 | int i; | ||
| 591 | int ret; | ||
| 592 | |||
| 593 | /* Tear down all existing unlocked IMRs. */ | ||
| 594 | for (i = 0; i < idev->max_imr; i++) | ||
| 595 | imr_clear(i); | ||
| 596 | |||
| 597 | /* | ||
| 598 | * Setup a locked IMR around the physical extent of the kernel | ||
| 599 | * from the beginning of the .text secton to the end of the | ||
| 600 | * .rodata section as one physically contiguous block. | ||
| 601 | */ | ||
| 602 | ret = imr_add_range(base, size, IMR_CPU, IMR_CPU, true); | ||
| 603 | if (ret < 0) { | ||
| 604 | pr_err("unable to setup IMR for kernel: (%p - %p)\n", | ||
| 605 | &_text, &__end_rodata); | ||
| 606 | } else { | ||
| 607 | pr_info("protecting kernel .text - .rodata: %zu KiB (%p - %p)\n", | ||
| 608 | size / 1024, &_text, &__end_rodata); | ||
| 609 | } | ||
| 610 | |||
| 611 | } | ||
| 612 | |||
| 613 | static const struct x86_cpu_id imr_ids[] __initconst = { | ||
| 614 | { X86_VENDOR_INTEL, 5, 9 }, /* Intel Quark SoC X1000. */ | ||
| 615 | {} | ||
| 616 | }; | ||
| 617 | MODULE_DEVICE_TABLE(x86cpu, imr_ids); | ||
| 618 | |||
| 619 | /** | ||
| 620 | * imr_init - entry point for IMR driver. | ||
| 621 | * | ||
| 622 | * return: -ENODEV for no IMR support 0 if good to go. | ||
| 623 | */ | ||
| 624 | static int __init imr_init(void) | ||
| 625 | { | ||
| 626 | struct imr_device *idev = &imr_dev; | ||
| 627 | int ret; | ||
| 628 | |||
| 629 | if (!x86_match_cpu(imr_ids) || !iosf_mbi_available()) | ||
| 630 | return -ENODEV; | ||
| 631 | |||
| 632 | idev->max_imr = QUARK_X1000_IMR_MAX; | ||
| 633 | idev->reg_base = QUARK_X1000_IMR_REGBASE; | ||
| 634 | idev->init = true; | ||
| 635 | |||
| 636 | mutex_init(&idev->lock); | ||
| 637 | ret = imr_debugfs_register(idev); | ||
| 638 | if (ret != 0) | ||
| 639 | pr_warn("debugfs register failed!\n"); | ||
| 640 | imr_fixup_memmap(idev); | ||
| 641 | return 0; | ||
| 642 | } | ||
| 643 | |||
| 644 | /** | ||
| 645 | * imr_exit - exit point for IMR code. | ||
| 646 | * | ||
| 647 | * Deregisters debugfs, leave IMR state as-is. | ||
| 648 | * | ||
| 649 | * return: | ||
| 650 | */ | ||
| 651 | static void __exit imr_exit(void) | ||
| 652 | { | ||
| 653 | imr_debugfs_unregister(&imr_dev); | ||
| 654 | } | ||
| 655 | |||
| 656 | module_init(imr_init); | ||
| 657 | module_exit(imr_exit); | ||
| 658 | |||
| 659 | MODULE_AUTHOR("Bryan O'Donoghue <pure.logic@nexus-software.ie>"); | ||
| 660 | MODULE_DESCRIPTION("Intel Isolated Memory Region driver"); | ||
| 661 | MODULE_LICENSE("Dual BSD/GPL"); | ||
diff --git a/arch/x86/platform/intel-quark/imr_selftest.c b/arch/x86/platform/intel-quark/imr_selftest.c new file mode 100644 index 000000000000..c9a0838890e2 --- /dev/null +++ b/arch/x86/platform/intel-quark/imr_selftest.c | |||
| @@ -0,0 +1,129 @@ | |||
| 1 | /** | ||
| 2 | * imr_selftest.c | ||
| 3 | * | ||
| 4 | * Copyright(c) 2013 Intel Corporation. | ||
| 5 | * Copyright(c) 2015 Bryan O'Donoghue <pure.logic@nexus-software.ie> | ||
| 6 | * | ||
| 7 | * IMR self test. The purpose of this module is to run a set of tests on the | ||
| 8 | * IMR API to validate it's sanity. We check for overlapping, reserved | ||
| 9 | * addresses and setup/teardown sanity. | ||
| 10 | * | ||
| 11 | */ | ||
| 12 | |||
| 13 | #include <asm-generic/sections.h> | ||
| 14 | #include <asm/imr.h> | ||
| 15 | #include <linux/init.h> | ||
| 16 | #include <linux/mm.h> | ||
| 17 | #include <linux/module.h> | ||
| 18 | #include <linux/types.h> | ||
| 19 | |||
| 20 | #define SELFTEST KBUILD_MODNAME ": " | ||
| 21 | /** | ||
| 22 | * imr_self_test_result - Print result string for self test. | ||
| 23 | * | ||
| 24 | * @res: result code - true if test passed false otherwise. | ||
| 25 | * @fmt: format string. | ||
| 26 | * ... variadic argument list. | ||
| 27 | */ | ||
| 28 | static void __init imr_self_test_result(int res, const char *fmt, ...) | ||
| 29 | { | ||
| 30 | va_list vlist; | ||
| 31 | |||
| 32 | /* Print pass/fail. */ | ||
| 33 | if (res) | ||
| 34 | pr_info(SELFTEST "pass "); | ||
| 35 | else | ||
| 36 | pr_info(SELFTEST "fail "); | ||
| 37 | |||
| 38 | /* Print variable string. */ | ||
| 39 | va_start(vlist, fmt); | ||
| 40 | vprintk(fmt, vlist); | ||
| 41 | va_end(vlist); | ||
| 42 | |||
| 43 | /* Optional warning. */ | ||
| 44 | WARN(res == 0, "test failed"); | ||
| 45 | } | ||
| 46 | #undef SELFTEST | ||
| 47 | |||
| 48 | /** | ||
| 49 | * imr_self_test | ||
| 50 | * | ||
| 51 | * Verify IMR self_test with some simple tests to verify overlap, | ||
| 52 | * zero sized allocations and 1 KiB sized areas. | ||
| 53 | * | ||
| 54 | */ | ||
| 55 | static void __init imr_self_test(void) | ||
| 56 | { | ||
| 57 | phys_addr_t base = virt_to_phys(&_text); | ||
| 58 | size_t size = virt_to_phys(&__end_rodata) - base; | ||
| 59 | const char *fmt_over = "overlapped IMR @ (0x%08lx - 0x%08lx)\n"; | ||
| 60 | int ret; | ||
| 61 | |||
| 62 | /* Test zero zero. */ | ||
| 63 | ret = imr_add_range(0, 0, 0, 0, false); | ||
| 64 | imr_self_test_result(ret < 0, "zero sized IMR\n"); | ||
| 65 | |||
| 66 | /* Test exact overlap. */ | ||
| 67 | ret = imr_add_range(base, size, IMR_CPU, IMR_CPU, false); | ||
| 68 | imr_self_test_result(ret < 0, fmt_over, __va(base), __va(base + size)); | ||
| 69 | |||
| 70 | /* Test overlap with base inside of existing. */ | ||
| 71 | base += size - IMR_ALIGN; | ||
| 72 | ret = imr_add_range(base, size, IMR_CPU, IMR_CPU, false); | ||
| 73 | imr_self_test_result(ret < 0, fmt_over, __va(base), __va(base + size)); | ||
| 74 | |||
| 75 | /* Test overlap with end inside of existing. */ | ||
| 76 | base -= size + IMR_ALIGN * 2; | ||
| 77 | ret = imr_add_range(base, size, IMR_CPU, IMR_CPU, false); | ||
| 78 | imr_self_test_result(ret < 0, fmt_over, __va(base), __va(base + size)); | ||
| 79 | |||
| 80 | /* Test that a 1 KiB IMR @ zero with read/write all will bomb out. */ | ||
| 81 | ret = imr_add_range(0, IMR_ALIGN, IMR_READ_ACCESS_ALL, | ||
| 82 | IMR_WRITE_ACCESS_ALL, false); | ||
| 83 | imr_self_test_result(ret < 0, "1KiB IMR @ 0x00000000 - access-all\n"); | ||
| 84 | |||
| 85 | /* Test that a 1 KiB IMR @ zero with CPU only will work. */ | ||
| 86 | ret = imr_add_range(0, IMR_ALIGN, IMR_CPU, IMR_CPU, false); | ||
| 87 | imr_self_test_result(ret >= 0, "1KiB IMR @ 0x00000000 - cpu-access\n"); | ||
| 88 | if (ret >= 0) { | ||
| 89 | ret = imr_remove_range(0, IMR_ALIGN); | ||
| 90 | imr_self_test_result(ret == 0, "teardown - cpu-access\n"); | ||
| 91 | } | ||
| 92 | |||
| 93 | /* Test 2 KiB works. */ | ||
| 94 | size = IMR_ALIGN * 2; | ||
| 95 | ret = imr_add_range(0, size, IMR_READ_ACCESS_ALL, | ||
| 96 | IMR_WRITE_ACCESS_ALL, false); | ||
| 97 | imr_self_test_result(ret >= 0, "2KiB IMR @ 0x00000000\n"); | ||
| 98 | if (ret >= 0) { | ||
| 99 | ret = imr_remove_range(0, size); | ||
| 100 | imr_self_test_result(ret == 0, "teardown 2KiB\n"); | ||
| 101 | } | ||
| 102 | } | ||
| 103 | |||
| 104 | /** | ||
| 105 | * imr_self_test_init - entry point for IMR driver. | ||
| 106 | * | ||
| 107 | * return: -ENODEV for no IMR support 0 if good to go. | ||
| 108 | */ | ||
| 109 | static int __init imr_self_test_init(void) | ||
| 110 | { | ||
| 111 | imr_self_test(); | ||
| 112 | return 0; | ||
| 113 | } | ||
| 114 | |||
| 115 | /** | ||
| 116 | * imr_self_test_exit - exit point for IMR code. | ||
| 117 | * | ||
| 118 | * return: | ||
| 119 | */ | ||
| 120 | static void __exit imr_self_test_exit(void) | ||
| 121 | { | ||
| 122 | } | ||
| 123 | |||
| 124 | module_init(imr_self_test_init); | ||
| 125 | module_exit(imr_self_test_exit); | ||
| 126 | |||
| 127 | MODULE_AUTHOR("Bryan O'Donoghue <pure.logic@nexus-software.ie>"); | ||
| 128 | MODULE_DESCRIPTION("Intel Isolated Memory Region self-test driver"); | ||
| 129 | MODULE_LICENSE("Dual BSD/GPL"); | ||
diff --git a/drivers/platform/x86/Kconfig b/drivers/platform/x86/Kconfig index 638e797037da..97527614141b 100644 --- a/drivers/platform/x86/Kconfig +++ b/drivers/platform/x86/Kconfig | |||
| @@ -735,6 +735,31 @@ config INTEL_IPS | |||
| 735 | functionality. If in doubt, say Y here; it will only load on | 735 | functionality. If in doubt, say Y here; it will only load on |
| 736 | supported platforms. | 736 | supported platforms. |
| 737 | 737 | ||
| 738 | config INTEL_IMR | ||
| 739 | bool "Intel Isolated Memory Region support" | ||
| 740 | default n | ||
| 741 | depends on X86_INTEL_QUARK && IOSF_MBI | ||
| 742 | ---help--- | ||
| 743 | This option provides a means to manipulate Isolated Memory Regions. | ||
| 744 | IMRs are a set of registers that define read and write access masks | ||
| 745 | to prohibit certain system agents from accessing memory with 1 KiB | ||
| 746 | granularity. | ||
| 747 | |||
| 748 | IMRs make it possible to control read/write access to an address | ||
| 749 | by hardware agents inside the SoC. Read and write masks can be | ||
| 750 | defined for: | ||
| 751 | - eSRAM flush | ||
| 752 | - Dirty CPU snoop (write only) | ||
| 753 | - RMU access | ||
| 754 | - PCI Virtual Channel 0/Virtual Channel 1 | ||
| 755 | - SMM mode | ||
| 756 | - Non SMM mode | ||
| 757 | |||
| 758 | Quark contains a set of eight IMR registers and makes use of those | ||
| 759 | registers during its bootup process. | ||
| 760 | |||
| 761 | If you are running on a Galileo/Quark say Y here. | ||
| 762 | |||
| 738 | config IBM_RTL | 763 | config IBM_RTL |
| 739 | tristate "Device driver to enable PRTL support" | 764 | tristate "Device driver to enable PRTL support" |
| 740 | depends on X86 && PCI | 765 | depends on X86 && PCI |
