3 source "lib/Kconfig.debug"
5 config DEBUG_STACKOVERFLOW
6 bool "Check for stack overflows"
7 depends on DEBUG_KERNEL
9 This option will cause messages to be printed if free stack space
10 drops below a certain limit.
12 config DEBUG_STACK_USAGE
13 bool "Enable stack utilization instrumentation"
14 depends on DEBUG_KERNEL
16 Enables the display of the minimum amount of free stack which each
17 task has ever had available in the sysrq-T output.
19 This option will slow down process creation somewhat.
25 tristate "KGDB: for test case in expect"
28 This is a kgdb test case for automated testing.
31 bool "Verbose fault messages"
35 When a program crashes due to an exception, or the kernel detects
36 an internal error, the kernel can print a not so brief message
37 explaining what the problem was. This debugging information is
38 useful to developers and kernel hackers when tracking down problems,
39 but mostly meaningless to other people. This is always helpful for
40 debugging but serves no purpose on a production system.
41 Most people should say N here.
44 bool "Generate Blackfin MMR tree"
47 Create a tree of Blackfin MMRs via the debugfs tree. If
48 you enable this, you will find all MMRs laid out in the
49 /sys/kernel/debug/blackfin/ directory where you can read/write
50 MMRs directly from userspace. This is obviously just a debug
54 bool "Hardware error interrupt debugging"
55 depends on DEBUG_KERNEL
57 When enabled, the hardware error interrupt is never disabled, and
58 will happen immediately when an error condition occurs. This comes
59 at a slight cost in code size, but is necessary if you are getting
60 hardware error interrupts and need to know where they are coming
63 config DEBUG_DOUBLEFAULT
64 bool "Debug Double Faults"
67 If an exception is caused while executing code within the exception
68 handler, the NMI handler, the reset vector, or in emulator mode,
69 a double fault occurs. On the Blackfin, this is a unrecoverable
70 event. You have two options:
71 - RESET exactly when double fault occurs. The excepting
72 instruction address is stored in RETX, where the next kernel
73 boot will print it out.
74 - Print debug message. This is much more error prone, although
75 easier to handle. It is error prone since:
76 - The excepting instruction is not committed.
77 - All writebacks from the instruction are prevented.
78 - The generated exception is not taken.
79 - The EXCAUSE field is updated with an unrecoverable event
80 The only way to check this is to see if EXCAUSE contains the
81 unrecoverable event value at every exception return. By selecting
82 this option, you are skipping over the faulting instruction, and
83 hoping things stay together enough to print out a debug message.
85 This does add a little kernel code, but is the only method to debug
86 double faults - if unsure say "Y"
89 prompt "Double Fault Failure Method"
90 default DEBUG_DOUBLEFAULT_PRINT
91 depends on DEBUG_DOUBLEFAULT
93 config DEBUG_DOUBLEFAULT_PRINT
96 config DEBUG_DOUBLEFAULT_RESET
101 config DEBUG_ICACHE_CHECK
102 bool "Check Instruction cache coherency"
103 depends on DEBUG_KERNEL
104 depends on DEBUG_HWERR
106 Say Y here if you are getting weird unexplained errors. This will
107 ensure that icache is what SDRAM says it should be by doing a
108 byte wise comparison between SDRAM and instruction cache. This
109 also relocates the irq_panic() function to L1 memory, (which is
112 config DEBUG_HUNT_FOR_ZERO
113 bool "Catch NULL pointer reads/writes"
116 Say Y here to catch reads/writes to anywhere in the memory range
117 from 0x0000 - 0x0FFF (the first 4k) of memory. This is useful in
118 catching common programming errors such as NULL pointer dereferences.
120 Misbehaving applications will be killed (generate a SEGV) while the
121 kernel will trigger a panic.
123 Enabling this option will take up an extra entry in CPLB table.
124 Otherwise, there is no extra overhead.
126 config DEBUG_BFIN_HWTRACE_ON
127 bool "Turn on Blackfin's Hardware Trace"
130 All Blackfins include a Trace Unit which stores a history of the last
131 16 changes in program flow taken by the program sequencer. The history
132 allows the user to recreate the program sequencer’s recent path. This
133 can be handy when an application dies - we print out the execution
134 path of how it got to the offending instruction.
136 By turning this off, you may save a tiny amount of power.
139 prompt "Omit loop Tracing"
140 default DEBUG_BFIN_HWTRACE_COMPRESSION_OFF
141 depends on DEBUG_BFIN_HWTRACE_ON
143 The trace buffer can be configured to omit recording of changes in
144 program flow that match either the last entry or one of the last
145 two entries. Omitting one of these entries from the record prevents
146 the trace buffer from overflowing because of any sort of loop (for, do
147 while, etc) in the program.
149 Because zero-overhead Hardware loops are not recorded in the trace buffer,
150 this feature can be used to prevent trace overflow from loops that
151 are nested four deep.
153 config DEBUG_BFIN_HWTRACE_COMPRESSION_OFF
154 bool "Trace all Loops"
156 The trace buffer records all changes of flow
158 config DEBUG_BFIN_HWTRACE_COMPRESSION_ONE
159 bool "Compress single-level loops"
161 The trace buffer does not record single loops - helpful if trace
162 is spinning on a while or do loop.
164 config DEBUG_BFIN_HWTRACE_COMPRESSION_TWO
165 bool "Compress two-level loops"
167 The trace buffer does not record loops two levels deep. Helpful if
168 the trace is spinning in a nested loop
172 config DEBUG_BFIN_HWTRACE_COMPRESSION
174 depends on DEBUG_BFIN_HWTRACE_ON
175 default 0 if DEBUG_BFIN_HWTRACE_COMPRESSION_OFF
176 default 1 if DEBUG_BFIN_HWTRACE_COMPRESSION_ONE
177 default 2 if DEBUG_BFIN_HWTRACE_COMPRESSION_TWO
180 config DEBUG_BFIN_HWTRACE_EXPAND
181 bool "Expand Trace Buffer greater than 16 entries"
182 depends on DEBUG_BFIN_HWTRACE_ON
185 By selecting this option, every time the 16 hardware entries in
186 the Blackfin's HW Trace buffer are full, the kernel will move them
187 into a software buffer, for dumping when there is an issue. This
188 has a great impact on performance, (an interrupt every 16 change of
189 flows) and should normally be turned off, except in those nasty
192 config DEBUG_BFIN_HWTRACE_EXPAND_LEN
193 int "Size of Trace buffer (in power of 2k)"
195 depends on DEBUG_BFIN_HWTRACE_EXPAND
198 This sets the size of the software buffer that the trace information
200 0 for (2^0) 1k, or 256 entries,
201 1 for (2^1) 2k, or 512 entries,
202 2 for (2^2) 4k, or 1024 entries,
203 3 for (2^3) 8k, or 2048 entries,
204 4 for (2^4) 16k, or 4096 entries
206 config DEBUG_BFIN_NO_KERN_HWTRACE
207 bool "Turn off hwtrace in CPLB handlers"
208 depends on DEBUG_BFIN_HWTRACE_ON
211 The CPLB error handler contains a lot of flow changes which can
212 quickly fill up the hardware trace buffer. When debugging crashes,
213 the hardware trace may indicate that the problem lies in kernel
214 space when in reality an application is buggy.
216 Say Y here to disable hardware tracing in some known "jumpy" pieces
217 of code so that the trace buffer will extend further back.
222 select SERIAL_CORE_CONSOLE
224 This option enables special console drivers which allow the kernel
225 to print messages very early in the bootup process.
227 This is useful for kernel debugging when your machine crashes very
228 early before the console code is initialized. After enabling this
229 feature, you must add "earlyprintk=serial,uart0,57600" to the
230 command line (bootargs). It is safe to say Y here in all cases, as
231 all of this lives in the init section and is thrown away after the
232 kernel boots completely.
235 bool "Display the CPLB information"
237 Display the CPLB information via /proc/cplbinfo.
240 bool "Check the user pointer address"
243 Usually the pointer transfer from user space is checked to see if its
244 address is in the kernel space.
246 Say N here to disable that check to improve the performance.