S390 Debug Feature¶
- files:
arch/s390/kernel/debug.c
arch/s390/include/asm/debug.h
Description:¶
The goal of this feature is to provide a kernel debug logging APIwhere log records can be stored efficiently in memory, where each component(e.g. device drivers) can have one separate debug log.One purpose of this is to inspect the debug logs after a production system crashin order to analyze the reason for the crash.
If the system still runs but only a subcomponent which uses dbf fails,it is possible to look at the debug logs on a live system via the Linuxdebugfs filesystem.
The debug feature may also very useful for kernel and driver development.
Design:¶
Kernel components (e.g. device drivers) can register themselves at the debugfeature with the function calldebug_register().This function initializes adebug log for the caller. For each debug log exists a number of debug areaswhere exactly one is active at one time. Each debug area consists of contiguouspages in memory. In the debug areas there are stored debug entries (log records)which are written by event- and exception-calls.
An event-call writes the specified debug entry to the active debugarea and updates the log pointer for the active area. If the endof the active debug area is reached, a wrap around is done (ring buffer)and the next debug entry will be written at the beginning of the activedebug area.
An exception-call writes the specified debug entry to the log andswitches to the next debug area. This is done in order to be surethat the records which describe the origin of the exception are notoverwritten when a wrap around for the current area occurs.
The debug areas themselves are also ordered in form of a ring buffer.When an exception is thrown in the last debug area, the following debugentries are then written again in the very first area.
There are four versions for the event- and exception-calls: One forlogging raw data, one for text, one for numbers (unsigned int and long),and one for sprintf-like formatted strings.
Each debug entry contains the following data:
Timestamp
Cpu-Number of calling task
Level of debug entry (0...6)
Return Address to caller
Flag, if entry is an exception or not
The debug logs can be inspected in a live system through entries inthe debugfs-filesystem. Under the toplevel directory “s390dbf” there isa directory for each registered component, which is named like thecorresponding component. The debugfs normally should be mounted to/sys/kernel/debug therefore the debug feature can be accessed under/sys/kernel/debug/s390dbf.
The content of the directories are files which represent different viewsto the debug log. Each component can decide which views should beused through registering them with the functiondebug_register_view().Predefined views for hex/ascii and sprintf data are provided.It is also possible to define other views. The content ofa view can be inspected simply by reading the corresponding debugfs file.
All debug logs have an actual debug level (range from 0 to 6).The default level is 3. Event and Exception functions have alevelparameter. Only debug entries with a level that is lower or equalthan the actual level are written to the log. This means, whenwriting events, high priority log entries should have a low levelvalue whereas low priority entries should have a high one.The actual debug level can be changed with the help of the debugfs-filesystemthrough writing a number string “x” to thelevel debugfs file which isprovided for every debug log. Debugging can be switched off completelyby using “-” on thelevel debugfs file.
Example:
> echo "-" > /sys/kernel/debug/s390dbf/dasd/level
It is also possible to deactivate the debug feature globally for everydebug log. You can change the behavior using 2 sysctl parameters in/proc/sys/s390dbf:
There are currently 2 possible triggers, which stop the debug featureglobally. The first possibility is to use thedebug_active sysctl. Ifset to 1 the debug feature is running. Ifdebug_active is set to 0 thedebug feature is turned off.
The second trigger which stops the debug feature is a kernel oops.That prevents the debug feature from overwriting debug information thathappened before the oops. After an oops you can reactivate the debug featureby piping 1 to/proc/sys/s390dbf/debug_active. Nevertheless, it’s notsuggested to use an oopsed kernel in a production environment.
If you want to disallow the deactivation of the debug feature, you can usethedebug_stoppable sysctl. If you setdebug_stoppable to 0 the debugfeature cannot be stopped. If the debug feature is already stopped, itwill stay deactivated.
Kernel Interfaces:¶
- booldebug_next_entry(file_private_info_t*p_info)¶
Go to the next entry
Parameters
file_private_info_t*p_infoPrivate info that is manipulated
Description
Sets the current position inp_info to the next entry. If no further entryexists the current position is set to one after the end the return valueindicates that no further entries exist.
Return
True if there are more following entries, false otherwise
- voiddebug_to_act_entry(file_private_info_t*p_info)¶
Go to the currently active entry
Parameters
file_private_info_t*p_infoPrivate info that is manipulated
Description
Sets the current position inp_info to the currently activeentry ofp_info->debug_info_snap
- booldebug_prev_entry(file_private_info_t*p_info)¶
Go to the previous entry
Parameters
file_private_info_t*p_infoPrivate info that is manipulated
Description
Sets the current position inp_info to the previous entry. If no previous entryexists the current position is set left as DEBUG_PROLOG_ENTRY and the return valueindicates that no previous entries exist.
Return
True if there are more previous entries, false otherwise
- booldebug_move_entry(file_private_info_t*p_info,boolreverse)¶
Go to next entry in either the forward or backward direction
Parameters
file_private_info_t*p_infoPrivate info that is manipulated
boolreverseIf true go to the next entry in reverse i.e. previous
Description
Sets the current position inp_info to the next (reverse == false) orprevious (reverse == true) entry.
Return
True if there are further entries in that direction,false otherwise.
- ssize_tdebug_dump(debug_info_t*id,structdebug_view*view,char*buf,size_tbuf_size,boolreverse)¶
Get a textual representation of debug info, or as much as fits
Parameters
debug_info_t*idDebug information to use
structdebug_view*viewView with which to dump the debug information
char*bufBuffer the textual debug data representation is written to
size_tbuf_sizeSize of the buffer, including the trailing ‘0’ byte
boolreverseGo backwards from the last written entry
Description
This function may be used whenever a textual representation of the debuginformation is required without using an s390dbf file.
Note
It is the callers responsibility to supply a view that is compatiblewith the debug information data.
Return
On success returns the number of bytes written to the buffer notincluding the trailing ‘0’ byte. If bug_size == 0 the function returns 0.On failure an error code less than 0 is returned.
- debug_info_t*debug_register_mode(constchar*name,intpages_per_area,intnr_areas,intbuf_size,umode_tmode,uid_tuid,gid_tgid)¶
creates and initializes debug area.
Parameters
constchar*nameName of debug log (e.g. used for debugfs entry)
intpages_per_areaNumber of pages, which will be allocated per area
intnr_areasNumber of debug areas
intbuf_sizeSize of data area in each debug entry
umode_tmodeFile mode for debugfs files. E.g. S_IRWXUGO
uid_tuidUser ID for debugfs files. Currently only 0 is supported.
gid_tgidGroup ID for debugfs files. Currently only 0 is supported.
Return
Handle for generated debug area
NULLif register failed
Description
Allocates memory for a debug log.Must not be called within an interrupt handler.
- debug_info_t*debug_register(constchar*name,intpages_per_area,intnr_areas,intbuf_size)¶
creates and initializes debug area with default file mode.
Parameters
constchar*nameName of debug log (e.g. used for debugfs entry)
intpages_per_areaNumber of pages, which will be allocated per area
intnr_areasNumber of debug areas
intbuf_sizeSize of data area in each debug entry
Return
Handle for generated debug area
NULLif register failed
Description
Allocates memory for a debug log.The debugfs file mode access permissions are read and write for user.Must not be called within an interrupt handler.
- voiddebug_register_static(debug_info_t*id,intpages_per_area,intnr_areas)¶
registers a static debug area
Parameters
debug_info_t*idHandle for static debug area
intpages_per_areaNumber of pages per area
intnr_areasNumber of debug areas
Description
Register debug_info_t defined using DEFINE_STATIC_DEBUG_INFO.
Note
- This function is called automatically via an initcall generated by
DEFINE_STATIC_DEBUG_INFO.
- voiddebug_unregister(debug_info_t*id)¶
give back debug area.
Parameters
debug_info_t*idhandle for debug log
Return
none
- voiddebug_set_level(debug_info_t*id,intnew_level)¶
Sets new actual debug level if new_level is valid.
Parameters
debug_info_t*idhandle for debug log
intnew_levelnew debug level
Return
none
- voiddebug_stop_all(void)¶
stops the debug feature if stopping is allowed.
Parameters
voidno arguments
Return
none
Description
Currently used in case of a kernel oops.
- voiddebug_set_critical(void)¶
event/exception functions try lock instead of spin.
Parameters
voidno arguments
Return
none
Description
Currently used in case of stopping all CPUs but the current one.Once in this state, functions to write a debug entry for anevent or exception no longer spin on the debug area lock,but only try to get it and fail if they do not get the lock.
- intdebug_register_view(debug_info_t*id,structdebug_view*view)¶
registers new debug view and creates debugfs dir entry
Parameters
debug_info_t*idhandle for debug log
structdebug_view*viewpointer to debug view struct
Return
0 : ok
< 0: Error
- intdebug_unregister_view(debug_info_t*id,structdebug_view*view)¶
unregisters debug view and removes debugfs dir entry
Parameters
debug_info_t*idhandle for debug log
structdebug_view*viewpointer to debug view struct
Return
0 : ok
< 0: Error
- booldebug_level_enabled(debug_info_t*id,intlevel)¶
Returns true if debug events for the specified level would be logged. Otherwise returns false.
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
Return
trueif level is less or equal to the current debug level.
- debug_entry_t*debug_event(debug_info_t*id,intlevel,void*data,intlength)¶
writes binary debug entry to active debug area (if level <= actual debug level)
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
void*datapointer to data for debug entry
intlengthlength of data in bytes
Return
Address of written debug entry
NULLif error
- debug_entry_t*debug_int_event(debug_info_t*id,intlevel,unsignedinttag)¶
writes unsigned integer debug entry to active debug area (if level <= actual debug level)
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
unsignedinttaginteger value for debug entry
Return
Address of written debug entry
NULLif error
- debug_entry_t*debug_long_event(debug_info_t*id,intlevel,unsignedlongtag)¶
writes unsigned long debug entry to active debug area (if level <= actual debug level)
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
unsignedlongtaglong integer value for debug entry
Return
Address of written debug entry
NULLif error
- debug_entry_t*debug_text_event(debug_info_t*id,intlevel,constchar*txt)¶
writes string debug entry in ascii format to active debug area (if level <= actual debug level)
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
constchar*txtstring for debug entry
Return
Address of written debug entry
NULLif error
- debug_sprintf_event¶
debug_sprintf_event(_id,_level,_fmt,...)
writes debug entry with format string and varargs (longs) to active debug area (if level $<=$ actual debug level).
Parameters
_idhandle for debug log
_leveldebug level
_fmtformat string for debug entry
...varargs used as in
sprintf()
Return
Address of written debug entry
NULLif error
Description
floats and long long datatypes cannot be used as varargs.
- debug_entry_t*debug_exception(debug_info_t*id,intlevel,void*data,intlength)¶
writes binary debug entry to active debug area (if level <= actual debug level) and switches to next debug area
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
void*datapointer to data for debug entry
intlengthlength of data in bytes
Return
Address of written debug entry
NULLif error
- debug_entry_t*debug_int_exception(debug_info_t*id,intlevel,unsignedinttag)¶
writes unsigned int debug entry to active debug area (if level <= actual debug level) and switches to next debug area
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
unsignedinttaginteger value for debug entry
Return
Address of written debug entry
NULLif error
- debug_entry_t*debug_long_exception(debug_info_t*id,intlevel,unsignedlongtag)¶
writes long debug entry to active debug area (if level <= actual debug level) and switches to next debug area
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
unsignedlongtaglong integer value for debug entry
Return
Address of written debug entry
NULLif error
- debug_entry_t*debug_text_exception(debug_info_t*id,intlevel,constchar*txt)¶
writes string debug entry in ascii format to active debug area (if level <= actual debug level) and switches to next debug area area
Parameters
debug_info_t*idhandle for debug log
intleveldebug level
constchar*txtstring for debug entry
Return
Address of written debug entry
NULLif error
- debug_sprintf_exception¶
debug_sprintf_exception(_id,_level,_fmt,...)
writes debug entry with format string and varargs (longs) to active debug area (if level <= actual debug level) and switches to next debug area.
Parameters
_idhandle for debug log
_leveldebug level
_fmtformat string for debug entry
...varargs used as in
sprintf()
Return
Address of written debug entry
NULLif error
Description
floats and long long datatypes cannot be used as varargs.
- DEFINE_STATIC_DEBUG_INFO¶
DEFINE_STATIC_DEBUG_INFO(var,name,pages,nr_areas,buf_size,view)
Define static debug_info_t
Parameters
varName of debug_info_t variable
nameName of debug log (e.g. used for debugfs entry)
pagesNumber of pages per area
nr_areasNumber of debug areas
buf_sizeSize of data area in each debug entry
viewPointer to debug view struct
Description
Define a static debug_info_t for early tracing. The associated debugfs logis automatically registered with the specified debug view.
Important: Users of this macro must not call any of thedebug_register/_unregister() functions for this debug_info_t!
Note
Tracing will start with a fixed number of initial pages and areas.The debug area will be changed to use the specified numbers duringarch_initcall.
Predefined views:¶
externstructdebug_viewdebug_hex_ascii_view;externstructdebug_viewdebug_sprintf_view;
Examples¶
/* * hex_ascii-view Example */#include<linux/init.h>#include<asm/debug.h>staticdebug_info_t*debug_info;staticintinit(void){/* register 4 debug areas with one page each and 4 byte data field */debug_info=debug_register("test",1,4,4);debug_register_view(debug_info,&debug_hex_ascii_view);debug_text_event(debug_info,4,"one ");debug_int_exception(debug_info,4,4711);debug_event(debug_info,3,&debug_info,4);return0;}staticvoidcleanup(void){debug_unregister(debug_info);}module_init(init);module_exit(cleanup);
/* * sprintf-view Example */#include<linux/init.h>#include<asm/debug.h>staticdebug_info_t*debug_info;staticintinit(void){/* register 4 debug areas with one page each and data field for *//* format string pointer + 2 varargs (= 3 * sizeof(long)) */debug_info=debug_register("test",1,4,sizeof(long)*3);debug_register_view(debug_info,&debug_sprintf_view);debug_sprintf_event(debug_info,2,"first event in %s:%i\n",__FILE__,__LINE__);debug_sprintf_exception(debug_info,1,"pointer to debug info: %p\n",&debug_info);return0;}staticvoidcleanup(void){debug_unregister(debug_info);}module_init(init);module_exit(cleanup);
Debugfs Interface¶
Views to the debug logs can be investigated through reading the correspondingdebugfs-files:
Example:
> ls /sys/kernel/debug/s390dbf/dasdflush hex_ascii level pages> cat /sys/kernel/debug/s390dbf/dasd/hex_ascii | sort -k2,2 -s00 00974733272:680099 2 - 02 0006ad7e 07 ea 4a 90 | ....00 00974733272:682210 2 - 02 0006ade6 46 52 45 45 | FREE00 00974733272:682213 2 - 02 0006adf6 07 ea 4a 90 | ....00 00974733272:682281 1 * 02 0006ab08 41 4c 4c 43 | EXCP01 00974733272:682284 2 - 02 0006ab16 45 43 4b 44 | ECKD01 00974733272:682287 2 - 02 0006ab28 00 00 00 04 | ....01 00974733272:682289 2 - 02 0006ab3e 00 00 00 20 | ...01 00974733272:682297 2 - 02 0006ad7e 07 ea 4a 90 | ....01 00974733272:684384 2 - 00 0006ade6 46 52 45 45 | FREE01 00974733272:684388 2 - 00 0006adf6 07 ea 4a 90 | ....
See section about predefined views for explanation of the above output!
Changing the debug level¶
Example:
> cat /sys/kernel/debug/s390dbf/dasd/level3> echo "5" > /sys/kernel/debug/s390dbf/dasd/level> cat /sys/kernel/debug/s390dbf/dasd/level5
Flushing debug areas¶
Debug areas can be flushed with piping the number of the desiredarea (0...n) to the debugfs file “flush”. When using “-” all debug areasare flushed.
Examples:
Flush debug area 0:
> echo "0" > /sys/kernel/debug/s390dbf/dasd/flush
Flush all debug areas:
> echo "-" > /sys/kernel/debug/s390dbf/dasd/flush
Changing the size of debug areas¶
To resize a debug area, write the desired page count to the “pages” file.Existing data is preserved if it fits; otherwise, oldest entries are dropped.
Example:
Define 4 pages for the debug areas of debug feature “dasd”:
> echo "4" > /sys/kernel/debug/s390dbf/dasd/pages
Stopping the debug feature¶
Example:
Check if stopping is allowed:
> cat /proc/sys/s390dbf/debug_stoppable
Stop debug feature:
> echo 0 > /proc/sys/s390dbf/debug_active
crash Interface¶
Thecrash tool since v5.1.0 has a built-in commands390dbf to display all the debug logs or export them to the file system.With this tool it is possibleto investigate the debug logs on a live system and with a memory dump aftera system crash.
Investigating raw memory¶
One last possibility to investigate the debug logs at a livesystem and after a system crash is to look at the raw memoryunder VM or at the Service Element.It is possible to find the anchor of the debug-logs throughthedebug_area_first symbol in the System map. Then one hasto follow the correct pointers of the data-structures definedin debug.h and find the debug-areas in memory.Normally modules which use the debug feature will also havea global variable with the pointer to the debug-logs. Followingthis pointer it will also be possible to find the debug logs inmemory.
For this method it is recommended to use ‘16 * x + 4’ byte (x = 0..n)for the length of the data field indebug_register() inorder to see the debug entries well formatted.
Predefined Views¶
There are two predefined views: hex_ascii and sprintf.The hex_ascii view shows the data field in hex and ascii representation(e.g.45434b44|ECKD).
The sprintf view formats the debug entries in the same way as the sprintffunction would do. The sprintf event/exception functions write to thedebug entry a pointer to the format string (size = sizeof(long))and for each vararg a long value. So e.g. for a debug entry with a formatstring plus two varargs one would need to allocate a (3 * sizeof(long))byte data area in thedebug_register() function.
- IMPORTANT:
Using “%s” in sprintf event functions is dangerous. You can onlyuse “%s” in the sprintf event functions, if the memory for the passed stringis available as long as the debug feature exists. The reason behind this isthat due to performance considerations only a pointer to the string is storedin the debug feature. If you log a string that is freed afterwards, you willget an OOPS when inspecting the debug feature, because then the debug featurewill access the already freed memory.
- NOTE:
If using the sprintf view do NOT use other event/exception functionsthan the sprintf-event and -exception functions.
The format of the hex_ascii and sprintf view is as follows:
Number of area
Timestamp (formatted as seconds and microseconds since 00:00:00 CoordinatedUniversal Time (UTC), January 1, 1970)
level of debug entry
Exception flag (* = Exception)
Cpu-Number of calling task
Return Address to caller
data field
A typical line of the hex_ascii view will look like the following (first lineis only for explanation and will not be displayed when ‘cating’ the view):
area time level exception cpu caller data (hex + ascii)--------------------------------------------------------------------------00 00964419409:440690 1 - 00 88023fe
Defining views¶
Views are specified with the ‘debug_view’ structure. There are definedcallback functions which are used for reading and writing the debugfs files:
structdebug_view{charname[DEBUG_MAX_PROCF_LEN];debug_prolog_proc_t*prolog_proc;debug_header_proc_t*header_proc;debug_format_proc_t*format_proc;debug_input_proc_t*input_proc;void*private_data;};
where:
typedefint(debug_header_proc_t)(debug_info_t*id,structdebug_view*view,intarea,debug_entry_t*entry,char*out_buf);typedefint(debug_format_proc_t)(debug_info_t*id,structdebug_view*view,char*out_buf,constchar*in_buf);typedefint(debug_prolog_proc_t)(debug_info_t*id,structdebug_view*view,char*out_buf);typedefint(debug_input_proc_t)(debug_info_t*id,structdebug_view*view,structfile*file,constchar*user_buf,size_tin_buf_size,loff_t*offset);
The “private_data” member can be used as pointer to view specific data.It is not used by the debug feature itself.
The output when reading a debugfs file is structured like this:
"prolog_proc output""header_proc output 1" "format_proc output 1""header_proc output 2" "format_proc output 2""header_proc output 3" "format_proc output 3"...
When a view is read from the debugfs, the Debug Feature calls the‘prolog_proc’ once for writing the prolog.Then ‘header_proc’ and ‘format_proc’ are called for eachexisting debug entry.
The input_proc can be used to implement functionality when it is written tothe view (e.g. like withecho"0">/sys/kernel/debug/s390dbf/dasd/level).
For header_proc there can be used the default functiondebug_dflt_header_fn() which is defined in debug.h.and which produces the same header output as the predefined views.E.g:
00 00964419409:440761 2 - 00 88023ec
In order to see how to use the callback functions check the implementationof the default views!
Example:
#include<asm/debug.h>#define UNKNOWNSTR "data: %08x"constchar*messages[]={"This error...........\n","That error...........\n","Problem..............\n","Something went wrong.\n","Everything ok........\n",NULL};staticintdebug_test_format_fn(debug_info_t*id,structdebug_view*view,char*out_buf,constchar*in_buf){inti,rc=0;if(id->buf_size>=4){intmsg_nr=*((int*)in_buf);if(msg_nr<sizeof(messages)/sizeof(char*)-1)rc+=sprintf(out_buf,"%s",messages[msg_nr]);elserc+=sprintf(out_buf,UNKNOWNSTR,msg_nr);}returnrc;}structdebug_viewdebug_test_view={"myview",/* name of view */NULL,/* no prolog */&debug_dflt_header_fn,/* default header for each entry */&debug_test_format_fn,/* our own format function */NULL,/* no input function */NULL/* no private data */};
test:¶
debug_info_t*debug_info;inti;...debug_info=debug_register("test",0,4,4);debug_register_view(debug_info,&debug_test_view);for(i=0;i<10;i++)debug_int_event(debug_info,1,i);
> cat /sys/kernel/debug/s390dbf/test/myview00 00964419734:611402 1 - 00 88042ca This error...........00 00964419734:611405 1 - 00 88042ca That error...........00 00964419734:611408 1 - 00 88042ca Problem..............00 00964419734:611411 1 - 00 88042ca Something went wrong.00 00964419734:611414 1 - 00 88042ca Everything ok........00 00964419734:611417 1 - 00 88042ca data: 0000000500 00964419734:611419 1 - 00 88042ca data: 0000000600 00964419734:611422 1 - 00 88042ca data: 0000000700 00964419734:611425 1 - 00 88042ca data: 0000000800 00964419734:611428 1 - 00 88042ca data: 00000009