blob: d3a6f74a94bdf29d15b9ef1db95927fae29d668c [file] [log] [blame]
Greg Kroah-Hartmanb2441312017-11-01 15:07:57 +01001# SPDX-License-Identifier: GPL-2.0
Linus Torvalds1da177e2005-04-16 15:20:36 -07002
peterz@infradead.orged004952020-07-27 14:48:52 +02003config TRACE_IRQFLAGS_NMI_SUPPORT
4 def_bool y
5
Lu Baoluaeb9dd12017-03-21 16:01:30 +08006config EARLY_PRINTK_USB
7 bool
8
Ben Collins6bcb13b2008-06-18 14:04:35 -04009config X86_VERBOSE_BOOTUP
10 bool "Enable verbose x86 bootup info messages"
11 default y
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090012 help
Ben Collins6bcb13b2008-06-18 14:04:35 -040013 Enables the informational output from the decompression stage
14 (e.g. bzImage) of the boot. If you disable this you will still
15 see errors. Disable this if you want silent bootup.
16
Linus Torvalds1da177e2005-04-16 15:20:36 -070017config EARLY_PRINTK
David Rientjes6a108a12011-01-20 14:44:16 -080018 bool "Early printk" if EXPERT
Linus Torvalds1da177e2005-04-16 15:20:36 -070019 default y
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090020 help
Linus Torvalds1da177e2005-04-16 15:20:36 -070021 Write kernel log output directly into the VGA buffer or to a serial
22 port.
23
24 This is useful for kernel debugging when your machine crashes very
25 early before the console code is initialized. For normal operation
26 it is not recommended because it looks ugly and doesn't cooperate
Lu Baoluaeb9dd12017-03-21 16:01:30 +080027 with klogd/syslogd or the X server. You should normally say N here,
Linus Torvalds1da177e2005-04-16 15:20:36 -070028 unless you want to debug such a crash.
29
Yinghai Lu5c059172008-07-24 17:29:40 -070030config EARLY_PRINTK_DBGP
31 bool "Early printk via EHCI debug port"
Ingo Molnar97499862008-07-26 17:28:11 +020032 depends on EARLY_PRINTK && PCI
Lu Baoluaeb9dd12017-03-21 16:01:30 +080033 select EARLY_PRINTK_USB
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090034 help
Yinghai Lu5c059172008-07-24 17:29:40 -070035 Write kernel log output directly into the EHCI debug port.
36
37 This is useful for kernel debugging when your machine crashes very
38 early before the console code is initialized. For normal operation
39 it is not recommended because it looks ugly and doesn't cooperate
Lu Baoluaeb9dd12017-03-21 16:01:30 +080040 with klogd/syslogd or the X server. You should normally say N here,
Yinghai Lu5c059172008-07-24 17:29:40 -070041 unless you want to debug such a crash. You need usb debug device.
42
Lu Baoluaeb9dd12017-03-21 16:01:30 +080043config EARLY_PRINTK_USB_XDBC
44 bool "Early printk via the xHCI debug port"
45 depends on EARLY_PRINTK && PCI
46 select EARLY_PRINTK_USB
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090047 help
Lu Baoluaeb9dd12017-03-21 16:01:30 +080048 Write kernel log output directly into the xHCI debug port.
49
50 One use for this feature is kernel debugging, for example when your
51 machine crashes very early before the regular console code is
52 initialized. Other uses include simpler, lockless logging instead of
53 a full-blown printk console driver + klogd.
54
55 For normal production environments this is normally not recommended,
56 because it doesn't feed events into klogd/syslogd and doesn't try to
57 print anything on the screen.
58
59 You should normally say N here, unless you want to debug early
60 crashes or need a very simple printk logging facility.
61
Borislav Petkov11cc8512014-01-18 12:48:15 +010062config EFI_PGT_DUMP
63 bool "Dump the EFI pagetable"
Stephen Smalleye1a58322015-10-05 12:55:20 -040064 depends on EFI
Steven Price2ae27132020-02-03 17:36:24 -080065 select PTDUMP_CORE
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090066 help
Borislav Petkov11cc8512014-01-18 12:48:15 +010067 Enable this if you want to dump the EFI page table before
68 enabling virtual mode. This can be used to debug miscellaneous
69 issues with the mapping of the EFI runtime regions into that
70 table.
71
Alex Shi3df32122012-06-28 09:02:20 +080072config DEBUG_TLBFLUSH
73 bool "Set upper limit of TLB entries to flush one-by-one"
Paul Bollecd69aa62013-03-14 12:54:37 +010074 depends on DEBUG_KERNEL
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090075 help
Alex Shi3df32122012-06-28 09:02:20 +080076
77 X86-only for now.
78
79 This option allows the user to tune the amount of TLB entries the
80 kernel flushes one-by-one instead of doing a full TLB flush. In
81 certain situations, the former is cheaper. This is controlled by the
82 tlb_flushall_shift knob under /sys/kernel/debug/x86. If you set it
83 to -1, the code flushes the whole TLB unconditionally. Otherwise,
84 for positive values of it, the kernel will use single TLB entry
85 invalidating instructions according to the following formula:
86
87 flush_entries <= active_tlb_entries / 2^tlb_flushall_shift
88
89 If in doubt, say "N".
90
Randy Dunlapd013a272007-10-24 15:50:43 -070091config IOMMU_DEBUG
92 bool "Enable IOMMU debugging"
Joerg Roedel966396d2007-10-24 12:49:48 +020093 depends on GART_IOMMU && DEBUG_KERNEL
Randy Dunlapd013a272007-10-24 15:50:43 -070094 depends on X86_64
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +090095 help
Randy Dunlapd013a272007-10-24 15:50:43 -070096 Force the IOMMU to on even when you have less than 4GB of
97 memory and add debugging code. On overflow always panic. And
98 allow to enable IOMMU leak tracing. Can be disabled at boot
99 time with iommu=noforce. This will also enable scatter gather
100 list merging. Currently not recommended for production
101 code. When you use it make sure you have a big enough
102 IOMMU/AGP aperture. Most of the options enabled by this can
103 be set more finegrained using the iommu= command line
Mauro Carvalho Chehabcb1aaeb2019-06-07 15:54:32 -0300104 options. See Documentation/x86/x86_64/boot-options.rst for more
Randy Dunlapd013a272007-10-24 15:50:43 -0700105 details.
106
107config IOMMU_LEAK
108 bool "IOMMU leak tracing"
FUJITA Tomonori19c1a6f2009-04-14 09:43:19 +0900109 depends on IOMMU_DEBUG && DMA_API_DEBUG
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900110 help
Randy Dunlapd013a272007-10-24 15:50:43 -0700111 Add a simple leak tracer to the IOMMU code. This is useful when you
112 are debugging a buggy device driver that leaks IOMMU mappings.
113
Pekka Paalanen6bc5c362009-01-03 21:23:51 +0200114config HAVE_MMIOTRACE_SUPPORT
115 def_bool y
Pekka Paalanen8b7d89d2008-05-12 21:20:56 +0200116
Masami Hiramatsuca0e9ba2009-08-13 16:34:21 -0400117config X86_DECODER_SELFTEST
OGAWA Hirofumicbe5c342009-12-06 20:14:29 +0900118 bool "x86 instruction decoder selftest"
Masami Hiramatsu77208042019-09-01 12:03:08 +0900119 depends on DEBUG_KERNEL && INSTRUCTION_DECODER
Ingo Molnarf8f20232014-02-05 06:51:37 +0100120 depends on !COMPILE_TEST
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900121 help
Masami Hiramatsuca0e9ba2009-08-13 16:34:21 -0400122 Perform x86 instruction decoder selftests at build time.
123 This option is useful for checking the sanity of x86 instruction
124 decoder code.
125 If unsure, say "N".
126
Ingo Molnar6e7c4022008-01-30 13:30:05 +0100127choice
128 prompt "IO delay type"
Ingo Molnarfd59e9e2008-02-17 20:20:24 +0100129 default IO_DELAY_0X80
Ingo Molnar6e7c4022008-01-30 13:30:05 +0100130
131config IO_DELAY_0X80
132 bool "port 0x80 based port-IO delay [recommended]"
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900133 help
Ingo Molnar6e7c4022008-01-30 13:30:05 +0100134 This is the traditional Linux IO delay used for in/out_p.
135 It is the most tested hence safest selection here.
136
137config IO_DELAY_0XED
138 bool "port 0xed based port-IO delay"
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900139 help
Ingo Molnar6e7c4022008-01-30 13:30:05 +0100140 Use port 0xed as the IO delay. This frees up port 0x80 which is
141 often used as a hardware-debug port.
142
143config IO_DELAY_UDELAY
144 bool "udelay based port-IO delay"
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900145 help
Ingo Molnar6e7c4022008-01-30 13:30:05 +0100146 Use udelay(2) as the IO delay method. This provides the delay
147 while not having any side-effect on the IO port space.
148
149config IO_DELAY_NONE
150 bool "no port-IO delay"
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900151 help
Ingo Molnar6e7c4022008-01-30 13:30:05 +0100152 No port-IO delay. Will break on old boxes that require port-IO
153 delay for certain operations. Should work on most new machines.
154
155endchoice
156
Huang, Ying6d7d7432008-01-30 13:32:51 +0100157config DEBUG_BOOT_PARAMS
158 bool "Debug boot parameters"
159 depends on DEBUG_KERNEL
160 depends on DEBUG_FS
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900161 help
Huang, Ying6d7d7432008-01-30 13:32:51 +0100162 This option will cause struct boot_params to be exported via debugfs.
163
Andi Kleen0c42f392008-01-30 13:33:42 +0100164config CPA_DEBUG
Ingo Molnar971a52d2008-02-06 22:39:45 +0100165 bool "CPA self-test code"
Ingo Molnarf316fe62008-01-30 13:34:04 +0100166 depends on DEBUG_KERNEL
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900167 help
Ingo Molnar971a52d2008-02-06 22:39:45 +0100168 Do change_page_attr() self-tests every 30 seconds.
Andi Kleen0c42f392008-01-30 13:33:42 +0100169
Andy Lutomirskia97439a2015-07-15 10:29:41 -0700170config DEBUG_ENTRY
171 bool "Debug low-level entry code"
172 depends on DEBUG_KERNEL
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900173 help
Andy Lutomirskia97439a2015-07-15 10:29:41 -0700174 This option enables sanity checks in x86's low-level entry code.
175 Some of these sanity checks may slow down kernel entries and
176 exits or otherwise impact performance.
177
Andy Lutomirskia97439a2015-07-15 10:29:41 -0700178 If unsure, say N.
179
Don Zickus99e8b9c2011-10-13 15:14:26 -0400180config DEBUG_NMI_SELFTEST
181 bool "NMI Selftest"
Don Zickus4f941c52011-12-07 16:06:30 -0500182 depends on DEBUG_KERNEL && X86_LOCAL_APIC
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900183 help
Don Zickus99e8b9c2011-10-13 15:14:26 -0400184 Enabling this option turns on a quick NMI selftest to verify
185 that the NMI behaves correctly.
186
187 This might help diagnose strange hangs that rely on NMI to
188 function properly.
189
190 If unsure, say N.
191
Bryan O'Donoghue28a375d2015-01-30 16:29:38 +0000192config DEBUG_IMR_SELFTEST
193 bool "Isolated Memory Region self test"
Bryan O'Donoghue28a375d2015-01-30 16:29:38 +0000194 depends on INTEL_IMR
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900195 help
Bryan O'Donoghue28a375d2015-01-30 16:29:38 +0000196 This option enables automated sanity testing of the IMR code.
197 Some simple tests are run to verify IMR bounds checking, alignment
198 and overlapping. This option is really only useful if you are
199 debugging an IMR memory map or are modifying the IMR code and want to
200 test your changes.
201
202 If unsure say N here.
203
Ingo Molnare97131a2015-05-05 11:34:49 +0200204config X86_DEBUG_FPU
205 bool "Debug the x86 FPU code"
206 depends on DEBUG_KERNEL
207 default y
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900208 help
Ingo Molnare97131a2015-05-05 11:34:49 +0200209 If this option is enabled then there will be extra sanity
210 checks and (boot time) debug printouts added to the kernel.
211 This debugging adds some small amount of runtime overhead
212 to the kernel.
213
214 If unsure, say N.
215
Srinivas Pandruvada956079e2015-05-06 12:15:54 -0700216config PUNIT_ATOM_DEBUG
217 tristate "ATOM Punit debug driver"
Arnd Bergmannd689c642017-07-19 14:53:05 +0200218 depends on PCI
Srinivas Pandruvada956079e2015-05-06 12:15:54 -0700219 select DEBUG_FS
220 select IOSF_MBI
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900221 help
Srinivas Pandruvada956079e2015-05-06 12:15:54 -0700222 This is a debug driver, which gets the power states
223 of all Punit North Complex devices. The power states of
224 each device is exposed as part of the debugfs interface.
225 The current power state can be read from
226 /sys/kernel/debug/punit_atom/dev_power_state
227
Josh Poimboeuf81d38712017-07-25 08:54:24 -0500228choice
229 prompt "Choose kernel unwinder"
Josh Poimboeuffc72ae42017-10-13 15:02:01 -0500230 default UNWINDER_ORC if X86_64
231 default UNWINDER_FRAME_POINTER if X86_32
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900232 help
Josh Poimboeuf81d38712017-07-25 08:54:24 -0500233 This determines which method will be used for unwinding kernel stack
234 traces for panics, oopses, bugs, warnings, perf, /proc/<pid>/stack,
235 livepatch, lockdep, and more.
236
Josh Poimboeuf11af8472017-10-13 15:02:00 -0500237config UNWINDER_ORC
Josh Poimboeufee9f8fc2017-07-24 18:36:57 -0500238 bool "ORC unwinder"
Josh Poimboeuf81d38712017-07-25 08:54:24 -0500239 depends on X86_64
Josh Poimboeufee9f8fc2017-07-24 18:36:57 -0500240 select STACK_VALIDATION
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900241 help
Josh Poimboeufee9f8fc2017-07-24 18:36:57 -0500242 This option enables the ORC (Oops Rewind Capability) unwinder for
243 unwinding kernel stack traces. It uses a custom data format which is
244 a simplified version of the DWARF Call Frame Information standard.
245
246 This unwinder is more accurate across interrupt entry frames than the
Josh Poimboeufa34a7662017-07-24 18:36:58 -0500247 frame pointer unwinder. It also enables a 5-10% performance
248 improvement across the entire kernel compared to frame pointers.
Josh Poimboeufee9f8fc2017-07-24 18:36:57 -0500249
250 Enabling this option will increase the kernel's runtime memory usage
251 by roughly 2-4MB, depending on your kernel config.
252
Josh Poimboeuffc72ae42017-10-13 15:02:01 -0500253config UNWINDER_FRAME_POINTER
254 bool "Frame pointer unwinder"
255 select FRAME_POINTER
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900256 help
Josh Poimboeuffc72ae42017-10-13 15:02:01 -0500257 This option enables the frame pointer unwinder for unwinding kernel
258 stack traces.
259
260 The unwinder itself is fast and it uses less RAM than the ORC
261 unwinder, but the kernel text size will grow by ~3% and the kernel's
262 overall performance will degrade by roughly 5-10%.
263
Josh Poimboeuf11af8472017-10-13 15:02:00 -0500264config UNWINDER_GUESS
Josh Poimboeuf81d38712017-07-25 08:54:24 -0500265 bool "Guess unwinder"
266 depends on EXPERT
Andrey Ryabinin0a373d4f2017-11-30 15:35:54 +0300267 depends on !STACKDEPOT
Masahiro Yamadaa7f7f622020-06-14 01:50:22 +0900268 help
Josh Poimboeuf81d38712017-07-25 08:54:24 -0500269 This option enables the "guess" unwinder for unwinding kernel stack
270 traces. It scans the stack and reports every kernel text address it
271 finds. Some of the addresses it reports may be incorrect.
272
273 While this option often produces false positives, it can still be
274 useful in many cases. Unlike the other unwinders, it has no runtime
275 overhead.
276
277endchoice
278
279config FRAME_POINTER
Josh Poimboeuf11af8472017-10-13 15:02:00 -0500280 depends on !UNWINDER_ORC && !UNWINDER_GUESS
Josh Poimboeuf81d38712017-07-25 08:54:24 -0500281 bool