1 # SPDX-License-Identifier: GPL-2.0
4 select ARCH_32BIT_OFF_T
5 select ARCH_HAS_BINFMT_FLAT if !MMU
6 select ARCH_HAS_DMA_PREP_COHERENT if MMU
7 select ARCH_HAS_SYNC_DMA_FOR_CPU if MMU
8 select ARCH_HAS_SYNC_DMA_FOR_DEVICE if MMU
9 select ARCH_HAS_DMA_SET_UNCACHED if MMU
10 select ARCH_USE_QUEUED_RWLOCKS
11 select ARCH_USE_QUEUED_SPINLOCKS
12 select ARCH_WANT_FRAME_POINTERS
13 select ARCH_WANT_IPC_PARSE_VERSION
14 select BUILDTIME_TABLE_SORT
15 select CLONE_BACKWARDS
17 select DMA_REMAP if MMU
18 select GENERIC_ATOMIC64
19 select GENERIC_CLOCKEVENTS
20 select GENERIC_IRQ_SHOW
21 select GENERIC_PCI_IOMAP
22 select GENERIC_SCHED_CLOCK
23 select GENERIC_STRNCPY_FROM_USER if KASAN
24 select HAVE_ARCH_JUMP_LABEL if !XIP_KERNEL
25 select HAVE_ARCH_KASAN if MMU && !XIP_KERNEL
26 select HAVE_ARCH_TRACEHOOK
27 select HAVE_COPY_THREAD_TLS
28 select HAVE_DEBUG_KMEMLEAK
29 select HAVE_DMA_CONTIGUOUS
30 select HAVE_EXIT_THREAD
31 select HAVE_FUNCTION_TRACER
32 select HAVE_FUTEX_CMPXCHG if !MMU
33 select HAVE_HW_BREAKPOINT if PERF_EVENTS
34 select HAVE_IRQ_TIME_ACCOUNTING
37 select HAVE_PERF_EVENTS
38 select HAVE_STACKPROTECTOR
39 select HAVE_SYSCALL_TRACEPOINTS
41 select MODULES_USE_ELF_RELA
42 select PERF_USE_VMALLOC
45 Xtensa processors are 32-bit RISC machines designed by Tensilica
46 primarily for embedded systems. These processors are both
47 configurable and extensible. The Linux port to the Xtensa
48 architecture supports all processor configurations and extensions,
49 with reasonable minimum requirements. The Xtensa Linux project has
50 a home page at <http://www.linux-xtensa.org/>.
52 config GENERIC_HWEIGHT
55 config ARCH_HAS_ILOG2_U32
58 config ARCH_HAS_ILOG2_U64
68 config LOCKDEP_SUPPORT
71 config STACKTRACE_SUPPORT
74 config TRACE_IRQFLAGS_SUPPORT
80 config HAVE_XTENSA_GPIO32
83 config KASAN_SHADOW_OFFSET
87 menu "Processor type and features"
90 prompt "Xtensa Processor Configuration"
91 default XTENSA_VARIANT_FSF
93 config XTENSA_VARIANT_FSF
94 bool "fsf - default (not generic) configuration"
97 config XTENSA_VARIANT_DC232B
98 bool "dc232b - Diamond 232L Standard Core Rev.B (LE)"
100 select HAVE_XTENSA_GPIO32
102 This variant refers to Tensilica's Diamond 232L Standard core Rev.B (LE).
104 config XTENSA_VARIANT_DC233C
105 bool "dc233c - Diamond 233L Standard Core Rev.C (LE)"
107 select HAVE_XTENSA_GPIO32
109 This variant refers to Tensilica's Diamond 233L Standard core Rev.C (LE).
111 config XTENSA_VARIANT_CUSTOM
112 bool "Custom Xtensa processor configuration"
113 select HAVE_XTENSA_GPIO32
115 Select this variant to use a custom Xtensa processor configuration.
116 You will be prompted for a processor variant CORENAME.
119 config XTENSA_VARIANT_CUSTOM_NAME
120 string "Xtensa Processor Custom Core Variant Name"
121 depends on XTENSA_VARIANT_CUSTOM
123 Provide the name of a custom Xtensa processor variant.
124 This CORENAME selects arch/xtensa/variant/CORENAME.
125 Don't forget you have to select MMU if you have one.
127 config XTENSA_VARIANT_NAME
129 default "dc232b" if XTENSA_VARIANT_DC232B
130 default "dc233c" if XTENSA_VARIANT_DC233C
131 default "fsf" if XTENSA_VARIANT_FSF
132 default XTENSA_VARIANT_CUSTOM_NAME if XTENSA_VARIANT_CUSTOM
134 config XTENSA_VARIANT_MMU
135 bool "Core variant has a Full MMU (TLB, Pages, Protection, etc)"
136 depends on XTENSA_VARIANT_CUSTOM
140 Build a Conventional Kernel with full MMU support,
141 ie: it supports a TLB with auto-loading, page protection.
143 config XTENSA_VARIANT_HAVE_PERF_EVENTS
144 bool "Core variant has Performance Monitor Module"
145 depends on XTENSA_VARIANT_CUSTOM
148 Enable if core variant has Performance Monitor Module with
149 External Registers Interface.
153 config XTENSA_FAKE_NMI
154 bool "Treat PMM IRQ as NMI"
155 depends on XTENSA_VARIANT_HAVE_PERF_EVENTS
158 If PMM IRQ is the only IRQ at EXCM level it is safe to
159 treat it as NMI, which improves accuracy of profiling.
161 If there are other interrupts at or above PMM IRQ priority level
162 but not above the EXCM level, PMM IRQ still may be treated as NMI,
163 but only if these IRQs are not used. There will be a build warning
164 saying that this is not safe, and a bugcheck if one of these IRQs
169 config XTENSA_UNALIGNED_USER
170 bool "Unaligned memory access in user space"
172 The Xtensa architecture currently does not handle unaligned
173 memory accesses in hardware but through an exception handler.
174 Per default, unaligned memory accesses are disabled in user space.
176 Say Y here to enable unaligned memory access in user space.
179 bool "System Supports SMP (MX)"
180 depends on XTENSA_VARIANT_CUSTOM
183 This option is used to indicate that the system-on-a-chip (SOC)
184 supports Multiprocessing. Multiprocessor support implemented above
185 the CPU core definition and currently needs to be selected manually.
187 Multiprocessor support is implemented with external cache and
188 interrupt controllers.
190 The MX interrupt distributer adds Interprocessor Interrupts
191 and causes the IRQ numbers to be increased by 4 for devices
192 like the open cores ethernet driver and the serial interface.
194 You still have to select "Enable SMP" to enable SMP on this SOC.
197 bool "Enable Symmetric multi-processing support"
199 select GENERIC_SMP_IDLE_THREAD
201 Enabled SMP Software; allows more than one CPU/CORE
202 to be activated during startup.
206 int "Maximum number of CPUs (2-32)"
211 bool "Enable CPU hotplug support"
214 Say Y here to allow turning CPUs off and on. CPUs can be
215 controlled through /sys/devices/system/cpu.
217 Say N if you want to disable CPU hotplug.
219 config FAST_SYSCALL_XTENSA
220 bool "Enable fast atomic syscalls"
223 fast_syscall_xtensa is a syscall that can make atomic operations
224 on UP kernel when processor has no s32c1i support.
226 This syscall is deprecated. It may have issues when called with
227 invalid arguments. It is provided only for backwards compatibility.
228 Only enable it if your userspace software requires it.
232 config FAST_SYSCALL_SPILL_REGISTERS
233 bool "Enable spill registers syscall"
236 fast_syscall_spill_registers is a syscall that spills all active
237 register windows of a calling userspace task onto its stack.
239 This syscall is deprecated. It may have issues when called with
240 invalid arguments. It is provided only for backwards compatibility.
241 Only enable it if your userspace software requires it.
245 config USER_ABI_CALL0
249 prompt "Userspace ABI"
250 default USER_ABI_DEFAULT
252 Select supported userspace ABI.
254 If unsure, choose the default ABI.
256 config USER_ABI_DEFAULT
257 bool "Default ABI only"
259 Assume default userspace ABI. For XEA2 cores it is windowed ABI.
260 call0 ABI binaries may be run on such kernel, but signal delivery
261 will not work correctly for them.
263 config USER_ABI_CALL0_ONLY
264 bool "Call0 ABI only"
265 select USER_ABI_CALL0
267 Select this option to support only call0 ABI in userspace.
268 Windowed ABI binaries will crash with a segfault caused by
269 an illegal instruction exception on the first 'entry' opcode.
271 Choose this option if you're planning to run only user code
272 built with call0 ABI.
274 config USER_ABI_CALL0_PROBE
275 bool "Support both windowed and call0 ABI by probing"
276 select USER_ABI_CALL0
278 Select this option to support both windowed and call0 userspace
279 ABIs. When enabled all processes are started with PS.WOE disabled
280 and a fast user exception handler for an illegal instruction is
281 used to turn on PS.WOE bit on the first 'entry' opcode executed by
284 This option should be enabled for the kernel that must support
285 both call0 and windowed ABIs in userspace at the same time.
287 Note that Xtensa ISA does not guarantee that entry opcode will
288 raise an illegal instruction exception on cores with XEA2 when
289 PS.WOE is disabled, check whether the target core supports it.
295 config XTENSA_CALIBRATE_CCOUNT
298 On some platforms (XT2000, for example), the CPU clock rate can
299 vary. The frequency can be determined, however, by measuring
300 against a well known, fixed frequency, such as an UART oscillator.
302 config SERIAL_CONSOLE
305 config PLATFORM_HAVE_XIP
308 menu "Platform options"
311 prompt "Xtensa System Type"
312 default XTENSA_PLATFORM_ISS
314 config XTENSA_PLATFORM_ISS
316 select XTENSA_CALIBRATE_CCOUNT
317 select SERIAL_CONSOLE
319 ISS is an acronym for Tensilica's Instruction Set Simulator.
321 config XTENSA_PLATFORM_XT2000
325 XT2000 is the name of Tensilica's feature-rich emulation platform.
326 This hardware is capable of running a full Linux distribution.
328 config XTENSA_PLATFORM_XTFPGA
330 select ETHOC if ETHERNET
331 select PLATFORM_WANT_DEFAULT_MEM if !MMU
332 select SERIAL_CONSOLE
333 select XTENSA_CALIBRATE_CCOUNT
334 select PLATFORM_HAVE_XIP
336 XTFPGA is the name of Tensilica board family (LX60, LX110, LX200, ML605).
337 This hardware is capable of running a full Linux distribution.
341 config PLATFORM_NR_IRQS
343 default 3 if XTENSA_PLATFORM_XT2000
346 config XTENSA_CPU_CLOCK
347 int "CPU clock rate [MHz]"
348 depends on !XTENSA_CALIBRATE_CCOUNT
351 config GENERIC_CALIBRATE_DELAY
352 bool "Auto calibration of the BogoMIPS value"
354 The BogoMIPS value can easily be derived from the CPU frequency.
357 bool "Default bootloader kernel arguments"
360 string "Initial kernel command string"
361 depends on CMDLINE_BOOL
362 default "console=ttyS0,38400 root=/dev/ram"
364 On some architectures (EBSA110 and CATS), there is currently no way
365 for the boot loader to pass arguments to the kernel. For these
366 architectures, you should supply some command-line options at build
367 time by entering them here. As a minimum, you should specify the
368 memory size and the root device (e.g., mem=64M root=/dev/nfs).
371 bool "Flattened Device Tree support"
373 select OF_EARLY_FLATTREE
375 Include support for flattened device tree machine descriptions.
377 config BUILTIN_DTB_SOURCE
378 string "DTB to build into the kernel image"
381 config PARSE_BOOTPARAM
382 bool "Parse bootparam block"
385 Parse parameters passed to the kernel from the bootloader. It may
386 be disabled if the kernel is known to run without the bootloader.
390 config BLK_DEV_SIMDISK
391 tristate "Host file-based simulated block device support"
393 depends on XTENSA_PLATFORM_ISS && BLOCK
395 Create block devices that map to files in the host file system.
396 Device binding to host file may be changed at runtime via proc
397 interface provided the device is not in use.
399 config BLK_DEV_SIMDISK_COUNT
400 int "Number of host file-based simulated block devices"
402 depends on BLK_DEV_SIMDISK
405 This is the default minimal number of created block devices.
406 Kernel/module parameter 'simdisk_count' may be used to change this
407 value at runtime. More file names (but no more than 10) may be
408 specified as parameters, simdisk_count grows accordingly.
410 config SIMDISK0_FILENAME
411 string "Host filename for the first simulated device"
412 depends on BLK_DEV_SIMDISK = y
415 Attach a first simdisk to a host file. Conventionally, this file
416 contains a root file system.
418 config SIMDISK1_FILENAME
419 string "Host filename for the second simulated device"
420 depends on BLK_DEV_SIMDISK = y && BLK_DEV_SIMDISK_COUNT != 1
423 Another simulated disk in a host file for a buildroot-independent
427 bool "Enable XTFPGA LCD driver"
428 depends on XTENSA_PLATFORM_XTFPGA
431 There's a 2x16 LCD on most of XTFPGA boards, kernel may output
432 progress messages there during bootup/shutdown. It may be useful
433 during board bringup.
437 config XTFPGA_LCD_BASE_ADDR
438 hex "XTFPGA LCD base address"
439 depends on XTFPGA_LCD
442 Base address of the LCD controller inside KIO region.
443 Different boards from XTFPGA family have LCD controller at different
444 addresses. Please consult prototyping user guide for your board for
445 the correct address. Wrong address here may lead to hardware lockup.
447 config XTFPGA_LCD_8BIT_ACCESS
448 bool "Use 8-bit access to XTFPGA LCD"
449 depends on XTFPGA_LCD
452 LCD may be connected with 4- or 8-bit interface, 8-bit access may
453 only be used with 8-bit interface. Please consult prototyping user
454 guide for your board for the correct interface width.
456 comment "Kernel memory layout"
458 config INITIALIZE_XTENSA_MMU_INSIDE_VMLINUX
459 bool "Initialize Xtensa MMU inside the Linux kernel code"
460 depends on !XTENSA_VARIANT_FSF && !XTENSA_VARIANT_DC232B
461 default y if XTENSA_VARIANT_DC233C || XTENSA_VARIANT_CUSTOM
463 Earlier version initialized the MMU in the exception vector
464 before jumping to _startup in head.S and had an advantage that
465 it was possible to place a software breakpoint at 'reset' and
466 then enter your normal kernel breakpoints once the MMU was mapped
467 to the kernel mappings (0XC0000000).
469 This unfortunately won't work for U-Boot and likely also wont
470 work for using KEXEC to have a hot kernel ready for doing a
473 So now the MMU is initialized in head.S but it's necessary to
474 use hardware breakpoints (gdb 'hbreak' cmd) to break at _startup.
475 xt-gdb can't place a Software Breakpoint in the 0XD region prior
476 to mapping the MMU and after mapping even if the area of low memory
477 was mapped gdb wouldn't remove the breakpoint on hitting it as the
478 PC wouldn't match. Since Hardware Breakpoints are recommended for
479 Linux configurations it seems reasonable to just assume they exist
480 and leave this older mechanism for unfortunate souls that choose
481 not to follow Tensilica's recommendation.
483 Selecting this will cause U-Boot to set the KERNEL Load and Entry
484 address at 0x00003000 instead of the mapped std of 0xD0003000.
489 bool "Kernel Execute-In-Place from ROM"
490 depends on PLATFORM_HAVE_XIP
492 Execute-In-Place allows the kernel to run from non-volatile storage
493 directly addressable by the CPU, such as NOR flash. This saves RAM
494 space since the text section of the kernel is not loaded from flash
495 to RAM. Read-write sections, such as the data section and stack,
496 are still copied to RAM. The XIP kernel is not compressed since
497 it has to run directly from flash, so it will take more space to
498 store it. The flash address used to link the kernel object files,
499 and for storing it, is configuration dependent. Therefore, if you
500 say Y here, you must know the proper physical address where to
501 store the kernel image depending on your own flash memory usage.
503 Also note that the make target becomes "make xipImage" rather than
504 "make Image" or "make uImage". The final kernel binary to put in
505 ROM memory will be arch/xtensa/boot/xipImage.
509 config MEMMAP_CACHEATTR
510 hex "Cache attributes for the memory address space"
514 These cache attributes are set up for noMMU systems. Each hex digit
515 specifies cache attributes for the corresponding 512MB memory
516 region: bits 0..3 -- for addresses 0x00000000..0x1fffffff,
517 bits 4..7 -- for addresses 0x20000000..0x3fffffff, and so on.
519 Cache attribute values are specific for the MMU type.
520 For region protection MMUs:
532 3: special (c and e are illegal, f is reserved).
536 2: WB, no-write-allocate cache,
541 hex "Physical address of the KSEG mapping"
542 depends on INITIALIZE_XTENSA_MMU_INSIDE_VMLINUX && MMU
545 This is the physical address where KSEG is mapped. Please refer to
546 the chosen KSEG layout help for the required address alignment.
547 Unpacked kernel image (including vectors) must be located completely
549 Physical memory below this address is not available to linux.
551 If unsure, leave the default value here.
553 config KERNEL_VIRTUAL_ADDRESS
554 hex "Kernel virtual address"
555 depends on MMU && XIP_KERNEL
558 This is the virtual address where the XIP kernel is mapped.
559 XIP kernel may be mapped into KSEG or KIO region, virtual address
560 provided here must match kernel load address provided in
563 config KERNEL_LOAD_ADDRESS
564 hex "Kernel load address"
565 default 0x60003000 if !MMU
566 default 0x00003000 if MMU && INITIALIZE_XTENSA_MMU_INSIDE_VMLINUX
567 default 0xd0003000 if MMU && !INITIALIZE_XTENSA_MMU_INSIDE_VMLINUX
569 This is the address where the kernel is loaded.
570 It is virtual address for MMUv2 configurations and physical address
571 for all other configurations.
573 If unsure, leave the default value here.
576 prompt "Relocatable vectors location"
577 default XTENSA_VECTORS_IN_TEXT
579 Choose whether relocatable vectors are merged into the kernel .text
580 or placed separately at runtime. This option does not affect
581 configurations without VECBASE register where vectors are always
582 placed at their hardware-defined locations.
584 config XTENSA_VECTORS_IN_TEXT
585 bool "Merge relocatable vectors into kernel text"
588 This option puts relocatable vectors into the kernel .text section
589 with proper alignment.
590 This is a safe choice for most configurations.
592 config XTENSA_VECTORS_SEPARATE
593 bool "Put relocatable vectors at fixed address"
595 This option puts relocatable vectors at specific virtual address.
596 Vectors are merged with the .init data in the kernel image and
597 are copied into their designated location during kernel startup.
598 Use it to put vectors into IRAM or out of FLASH on kernels with
599 XIP-aware MTD support.
604 hex "Kernel vectors virtual address"
606 depends on XTENSA_VECTORS_SEPARATE
608 This is the virtual address of the (relocatable) vectors base.
609 It must be within KSEG if MMU is used.
612 hex "XIP kernel data virtual address"
613 depends on XIP_KERNEL
616 This is the virtual address where XIP kernel data is copied.
617 It must be within KSEG if MMU is used.
619 config PLATFORM_WANT_DEFAULT_MEM
622 config DEFAULT_MEM_START
624 prompt "PAGE_OFFSET/PHYS_OFFSET" if !MMU && PLATFORM_WANT_DEFAULT_MEM
625 default 0x60000000 if PLATFORM_WANT_DEFAULT_MEM
628 This is the base address used for both PAGE_OFFSET and PHYS_OFFSET
629 in noMMU configurations.
631 If unsure, leave the default value here.
636 default XTENSA_KSEG_MMU_V2
638 config XTENSA_KSEG_MMU_V2
639 bool "MMUv2: 128MB cached + 128MB uncached"
641 MMUv2 compatible kernel memory map: TLB way 5 maps 128MB starting
642 at KSEG_PADDR to 0xd0000000 with cache and to 0xd8000000
644 KSEG_PADDR must be aligned to 128MB.
646 config XTENSA_KSEG_256M
647 bool "256MB cached + 256MB uncached"
648 depends on INITIALIZE_XTENSA_MMU_INSIDE_VMLINUX
650 TLB way 6 maps 256MB starting at KSEG_PADDR to 0xb0000000
651 with cache and to 0xc0000000 without cache.
652 KSEG_PADDR must be aligned to 256MB.
654 config XTENSA_KSEG_512M
655 bool "512MB cached + 512MB uncached"
656 depends on INITIALIZE_XTENSA_MMU_INSIDE_VMLINUX
658 TLB way 6 maps 512MB starting at KSEG_PADDR to 0xa0000000
659 with cache and to 0xc0000000 without cache.
660 KSEG_PADDR must be aligned to 256MB.
665 bool "High Memory Support"
668 Linux can use the full amount of RAM in the system by
669 default. However, the default MMUv2 setup only maps the
670 lowermost 128 MB of memory linearly to the areas starting
671 at 0xd0000000 (cached) and 0xd8000000 (uncached).
672 When there are more than 128 MB memory in the system not
673 all of it can be "permanently mapped" by the kernel.
674 The physical memory that's not permanently mapped is called
677 If you are compiling a kernel which will never run on a
678 machine with more than 128 MB total physical RAM, answer
683 config FORCE_MAX_ZONEORDER
684 int "Maximum zone order"
687 The kernel memory allocator divides physically contiguous memory
688 blocks into "zones", where each zone is a power of two number of
689 pages. This option selects the largest power of two that the kernel
690 keeps in the memory allocator. If you need to allocate very large
691 blocks of physically contiguous memory, then you may need to
694 This config option is actually maximum order plus one. For example,
695 a value of 11 means that the largest free memory block is 2^10 pages.
699 menu "Power management options"
701 source "kernel/power/Kconfig"