fix various typos in comments and output/log messages.

update several items that are done or dead.

branches: 1.17.176;
merge ktrace-lwp.

minor updates: vax port kludge is in trace, not kill. spelling.
no 0t output in ps.

Implement ddb setting that allows all ddb output to be teed into the
kernel message buffer/log. Its off by default and can be switched on in the
kernel configuration on build time, be set as a variable in ddb and be set
using sysctl.

This adds the sysctl value
ddb.tee_msgbuf = 0
by default.

The functionality is especially added and aimed for developers who are not
blessed with a serial console and wish to keep all their ddb output in the
log. Specifying /l as a modifier to some selected commands will also put
the output in the log but not all commands provide one nor has the same
meaning for all commands.

This feature could in the future also be implemented as an ddb command but
that could lead to more bloat allthough maybe easier for non developpers to
use when mailing their backtraces from kernel crashes.

Note that point 32 (ddb output to msgbuf) is Work In Progress and is about
to be committed after more testing and cleanup/review.

Add an item to help developers without serial consoles.

Uppercase CPU, plural is CPUs.

branches: 1.11.6;
Another item on the TODO list:

Add support for cpus where sizeof(register_t) is not necessarily
the same as sizeof(void *). This is the case on SH5 using the
ILP32 ABI. On this cpu db_expr_t is, necessarily, 64-bits.
Unfortunately, in ILP32 mode, ddb will only display the low 32-bits
of any expression, including registers...

branches: 1.10.8;
ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

ddb(9)

Convenience variables (workXX) don't work, so don't document them.

branches: 1.7.2; 1.7.4;
Add x/m option functionality

branches: 1.6.2;
Update TODO list item #7.

Observe deficiency in printfs reporting symbol table problems
before we have a message buffer.

"set" now displays the name of the variable, old value, and new value,
just as "write" does.

branches: 1.3.4;
In db_command(), fix indentation and add a missing db_skip_to_eol() which
previously caused the next command following a ","-style repitition
to not be properly executed.

Fix long-standing inconsistency between manapage and code.
The "ditto" character is " (double-quote); not , (comma) as
indicated in the manpage; and not ' (single-quote) as indicated
in the man.old -style manpage from which the mdoc manpage was derived.

Some other versions of ddb use single-quote for this feature, and this
part of the code may change again in the near future, but the documentation
now affects reality now.
they

First cut at an outline of stuff to do.

s/unnsupported/unsupported/ in message.

db_get_qvalue() with size=8 on 64bit arch returns incorrect value. fixed.

changes of r1.39 was incomplete. only "examin/m" could handle 'q'.
added support 'r','x','z','d','u', and 'o' with 'q' modifier on 32bit arch.

Avoids signedness bit shift in db_get_value()

Appease UBSan.

branches: 1.23.4;
avoid leftshift of an unsigned value

Add support for 'q' qualifier for 64-bit values on examine and write.
Align data to long so that the MD parts can use aligned load/stores.
Don't use db_extend but use db_expr_t based expressions.

branches: 1.21.22; 1.21.40;
quick fix for crash(8) build failure caused by my previous commit (rev
1.20).

The problem was reported Ty Sarna and Geoff Wing. Thanks.

XXX: I made #if too messy. split db_access.c into two files for
db_{get,put}_values and db_read_{int,ptr}?

kgdb still needs db_{get,put}_value in ddb/db_access.c when SOFTWARE_SSTEP is defined in db_machdep.h.

Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.18.42; 1.18.52; 1.18.58;
Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.17.32; 1.17.70;
ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

add RCSIDs

branches: 1.15.2; 1.15.4;
constify a little

Die, register, die!

branches: 1.13.2;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.11.10;
Last loop variable type change causes undesirable non-terminating loops.
Change for-loop expressions to get the intended effect again.

Eliminate signed-unsigned comparisons.

branches: 1.9.4;
Use db_expr_t rather than int in a few strategic places to avoid
type-size problems.

Use BYTE_ORDER, not BYTE_[LM]SF.

Clean up #includes.

Add prototypes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

#if --> #ifdef

Canonicalize all #includes.

branches: 1.2.4;
add explicit rcs id

branches: 1.1.1;
Initial revision

changes of r1.39 was incomplete. only "examin/m" could handle 'q'.
added support 'r','x','z','d','u', and 'o' with 'q' modifier on 32bit arch.

branches: 1.13.64;
Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.12.42; 1.12.52; 1.12.58;
Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.11.26;
merge ktrace-lwp.

branches: 1.10.2;
make the argument of db_write_bytes const (sorry for the signature change)

branches: 1.9.16;
ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

branches: 1.8.14; 1.8.16;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.6.26;
Clean up #includes.

Add prototypes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Remove RCS logs.

add explicit rcs id

branches: 1.1.1;
Initial revision

ddb(4): Declare extern db_active if _KMEMUSER too, not just DDB.

crash(8) is built with _KMEMUSER defined but DDB not defined except
for a few specific files or on MIPS.

Seems to me crash(8) should really be built with DDB defined in all
.c files but that leads to other fallout I'm not really inclined to
diagnose right now.

ddb/db_active.h: New home for extern db_active.

This can be included unconditionally, and db_active can then be
queried unconditionally; if DDB is not in the kernel, then db_active
is a constant zero. Reduces need for #include opt_ddb.h, #ifdef DDB.

Revert introduction of db_alloc.h.

db_alloc() and friends are already provided in crash.c.
Sorry for confusing you...

Introduce db_alloc.h.

Provide db_alloc() and friends to userland, i.e., crash(8).

obsolete DB_AOUT_SYMBOLS. however, we need to leave most of the code
in db_sym.[ch] as it is used by the elf version of crash(8).

i will be cleaning up the db_sym.c code in a follow up commit to avoid
having dead code compiled.

branches: 1.41.66; 1.41.72;
TRUE -> true, FALSE -> false

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.39.4;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.37.20; 1.37.22;
merge ktrace-lwp.

branches: 1.36.2;
Add lots of const.

branches: 1.35.2;
Add a cast to make this build with -D__AUDIT__.

Teach DDB to use ksyms instead of it's own symbol table management.
Note: For a.out kernels (beware!) DDB will still use the old stuff.

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

don't need <sys/types.h> when including <sys/param.h>

add RCSIDs

branches: 1.30.2; 1.30.4;
constify a little

s/preserving ... bytes of symbol table/using ... bytes of symbol table/

"preserving" implies that the data may be copied and might otherwise
be freed, and that other data in the symbol table or sections of
the kernel might be freed, however there is no mechanism for these
things to take place, so "using" is clearer.

Add a "sifting" command to ddb (named from the Sun OpenPROM command of
the same name); it searches the symbol table(s) for all symbols matching
a given substring, and prints.

Extremely useful for when you forget that critical symbol name.

Also, with /F support (cf. "ls -F") to print a char indicating the
symbol type.

Die, register, die!

branches: 1.26.2;
Don't load the symbol table if it's corrupted.

Cast ptrdiff_t to long for printf purposes.

Don't go dereferencing pointers before verifying they are aligned.

Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.21.4;
- Make the ddb_init() interface consistent for a.out and ELF.
- Allow DDB to run-time switch for a.out or ELF symbol table support.

Merge paddr_t changes into the main branch.

branches: 1.19.2;
Change ddb_init() for a.out to provision for kernels which are not booted by
the current 'bootblock' method, like the Shark, which preloads and preformats
the symbol table right in the boot code.

Clean up usage of printf() vs db_printf(): The diagnostic messages
when symbol tables are loaded don't belong to the user interaction,
their output lines should not be counted by ddb's builtin pager.

A few small changes to make it easier to support non-a.out symbol table
formats in DDB:
- Reverse the sense of the symbol format #ifdef... i.e. instead of
"#ifndef DB_NO_AOUT", use "#ifdef DB_AOUT_SYMBOLS". This is pulled
in from <machine/db_machdep.h>.
- Change the signature of X_db_sym_init(). Instead of int * and char *,
it now takes void *'s for the start and end of the symbol table.
- In X_db_sym_init(), check that the pointer to the beginning of
the symbol table is aligned.
- X_db_symbol_values() now takes a "db_symtab_t *symtab" argument.

backout previous kprintf changes

printf -> kprintf, sprintf -> ksprintf

Eliminate dependency on /usr/include/*

DDB prototype changes.

Add prototypes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Fix test for empty db_symtab on big-endian machines.

Canonicalize all #includes.

Canonicalize all #includes.

branches: 1.7.2;
Changes for boot loading of symbols.

Add RCS identifiers, remove some completely useless RCS logs and patchkit
headers, and a few other insignificant changes.

Remove read from file. Merge changes from mach DDB.

Print symbolic args and line no's in stack traces.

add explicit rcs id

make SYMTAB_SPACE user-definable, as an "option"

branches: 1.1.1;
Initial revision

obsolete DB_AOUT_SYMBOLS. however, we need to leave most of the code
in db_sym.[ch] as it is used by the elf version of crash(8).

i will be cleaning up the db_sym.c code in a follow up commit to avoid
having dead code compiled.

branches: 1.3.98; 1.3.104;
merge ktrace-lwp.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.1.64;
Eliminate dependency on /usr/include/*

sys: Split struct device into a private device_impl.h.

Include this only inside autoconf itself, and a few files that abuse
autoconf in ways I can't confidently make easy fixes for.

XXX kernel ABI change requires bump -- no more use of struct device
internals allowed, previously done by some drivers

enums might be unsigned.

Add command to print device list.

Add command to print device list.

TRUE -> true, FALSE -> false

Fix lossage from boolean_t -> bool and updated x86 bus_dma.

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.23.4;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.21.8; 1.21.10;
Now that BKPT_SET takes two arguments, supply the second argument.

branches: 1.20.4; 1.20.6; 1.20.8; 1.20.10; 1.20.12;
merge ktrace-lwp.

branches: 1.19.2;
more constification fallout

branches: 1.18.2;
Add a BKPT_ADDR() macro which gives MD code a chance to munge a
breakpoint address before it's used. Currently a no-op on all but sh5.

This is useful on sh5, for example, to mask off the instruction
type encoding in the bottom two address bits, and makes it possible
to do "db> break $rXX" instead of manually munging the address.

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

add RCSIDs

branches: 1.15.2;
replace vm_map{,_entry}_t with struct vm_map{,_entry} *.

branches: 1.14.2;
remove/move more mach vm header files:

<vm/pglist.h> -> <uvm/uvm_pglist.h>
<vm/vm_inherit.h> -> <uvm/uvm_inherit.h>
<vm/vm_kern.h> -> into <uvm/uvm_extern.h>
<vm/vm_object.h> -> nothing
<vm/vm_pager.h> -> into <uvm/uvm_pager.h>

also includes a bunch of <vm/vm_page.h> include removals (due to redudancy
with <vm/vm.h>), and a scattering of other similar headers.

Interface change: db_printsym() takes a third argument, pr, a function
pointer indicating how to print the symbol. This allows db_printsym()
to called in places where db_printf() is not an appropriate output
function.

While straightforward, apologies in advance if I've introduced any minor
syntax errors; I was unable to test compilation this on all the affected
platforms.

Die, register, die!

branches: 1.11.2;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.9.6;
Merge paddr_t changes into the main branch.

branches: 1.8.2;
If `SOFTWARE_SSTEP' is defined then kgdb_stub.c needs the single-step
emulation in db_run.c:

- re-arrange `db_*_single_step()' and `db_*_temp_breakpoint()' -
which are only used in the `SOFTWARE_SSTEP' case - so these
are available to the KGDB code.
- make other code in db_run.c conditional on `DDB'.

Fix db_printf format specifications.

DDB prototype changes.

Add prototypes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Canonicalize all #includes.

branches: 1.2.4;
add explicit rcs id

branches: 1.1.1;
Initial revision

Fix lossage from boolean_t -> bool and updated x86 bus_dma.

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.18.26;
merge ktrace-lwp.

branches: 1.17.2;
more constification fallout

branches: 1.16.16;
ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

branches: 1.15.2;
replace vm_map{,_entry}_t with struct vm_map{,_entry} *.

branches: 1.14.2;
remove include of <vm/vm.h>

branches: 1.13.2;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.11.6;
Merge paddr_t changes into the main branch.

branches: 1.10.2;
If `SOFTWARE_SSTEP' is defined then kgdb_stub.c needs the single-step
emulation in db_run.c:

- re-arrange `db_*_single_step()' and `db_*_temp_breakpoint()' -
which are only used in the `SOFTWARE_SSTEP' case - so these
are available to the KGDB code.
- make other code in db_run.c conditional on `DDB'.

In the db_breakpoint_t, use a db_expr_t to store the original instruction,
not an int.

DDB prototype changes.

Clean up #includes.

Add prototypes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Remove RCS logs.

Rearrange some #includes.

add explicit rcs id

branches: 1.1.1;
Initial revision

Convert all cnpollc callers to using true/false for consistency.

KNF a comment.

Back out the following revisions on behalf of core:

sys/sys/lwp.h: revision 1.228
sys/sys/pipe.h: revision 1.40
sys/kern/uipc_socket.c: revision 1.306
sys/kern/kern_sleepq.c: revision 1.84
sys/rump/librump/rumpkern/locks_up.c: revision 1.13
sys/kern/sys_pipe.c: revision 1.165
usr.bin/fstat/fstat.c: revision 1.119
sys/rump/librump/rumpkern/locks.c: revision 1.87
sys/ddb/db_xxx.c: revision 1.78
sys/ddb/db_command.c: revision 1.187
sys/sys/condvar.h: revision 1.18
sys/ddb/db_interface.h: revision 1.42
sys/sys/socketvar.h: revision 1.166
sys/kern/uipc_syscalls.c: revision 1.209
sys/kern/kern_condvar.c: revision 1.60

Add cv_fdrestart() [...]
Use cv_fdrestart() to implement fo_restart.
Simplify/streamline pipes a little bit [...]

This changes have caused regressions and need to be debugged.
The cv_fdrestart() addition needs more discussion.

ddb: Don't restrict `show all tstiles' to wmesg `tstile'.

This is no longer the wmesg, as of last week -- now it's the name of
the underlying syncobj, like `mutex' or `rwlock'.

ddb: Constify db_syncobj_owner argument.

No need for it to be writable, and the actual type of the struct lwp
member is const.

Simplify/streamline pipes a little bit:

- Allocate only one struct pipe not two (no need to be bidirectional here).
- Then use f_flag (FREAD/FWRITE) to figure out what to do in the fileops.
- Never wake the other side or acquire long-term (I/O) lock unless needed.
- Whenever possible, defer wakeups until after locks have been released.
- Do some things locklessly in pipe_ioctl() and pipe_poll().

Some notable results:

- -30% latency on a 486DX2/66 doing 1 byte ping-pong within a single process.
- 2.5x less lock contention during "make cleandir" of src on a 48 CPU machine.
- 1.5x bandwith with 1kB messages on the same 48 CPU machine (8kB: same b/w).

Add some simple DDB show commands: condvar, selinfo, sleepq

ddb: Teach `show all tstiles' to show the type of each lock.

ddb: Cast pointer to uintptr_t first before db_expr_t.

ddb: New `show all tstiles' command.

Shows who's waiting for which locks and what the owner is up to.

XXX pullup-10

ddb: PR kern/57435 - fix duplicate "show mount"

Bolivar cannot carry double. One of these commands should have been
"show mounts", cf. page/pages and a few others. While here fix the
function name to be db_show_all_mounts to conform to the pattern.

branches: 1.181.4;
Remove 'a' modifier from "show vmem". It was removed in sbur_vmem.c rev. 1.55.

ddb: guard invocation of db_cmd_on_enter properly.

db_command_loop - do not ignore the return value from setjmp used to
guard db_cmd_on_enter. We do not want to re-execute the enter command
if it fails. Note that "fails" includes e.g. aborting long output
from the enter command with "q" at the --db more-- prompt, which is
quite likely as the default enter command is "bt".

While here, don't even bother with the whole song and dance if the
enter command is not set.

Add a "show kqueue <addr>" command.

fix mainly same typos as in my previous commit but outside sys/dev/dm.

fix typos in words "pointer" and s/fram /frame/

introduce DDB_END_CMD and replace more than 20 copies of the same
list of NULLs and 0. idea from rillig@.

all touched ports built, several booted.

branches: 1.175.2;
Remove "also;"

Add a "show fdt" ddb command

Sort the db_show_cmd entries

Put the 'added from all sub cmds' comment above the four commands that
were added as part of the work in r1.98

Trailing whitespace

Handle "no count" properly in db_kernhist_print_cmd

branches: 1.169.2;
Make the db_kernelonly() macro visible to all files and re-use it.
Register access is kernel only.

print a message for the commands crash does not support.
Allow "show lock"

- Introduce a new global kernel variable "shutting_down" to indicate that
the system is shutting down or rebooting.
- Set this global in a new function called kern_reboot(), which is currently
just a basic wrapper around cpu_reboot().
- Call kern_reboot() instead of cpu_reboot() almost everywhere; a few
places remain where it's still called directly, but those are in early
pre-main() machdep locations.

Eventually, all of the various cpu_reboot() functions should be re-factored
and common functionality moved to kern_reboot(), but that's for another day.

Redo the page allocator to perform better, especially on multi-core and
multi-socket systems. Proposed on tech-kern. While here:

- add rudimentary NUMA support - needs more work.
- remove now unused "listq" from vm_page.

Teach ddb(4) about printing the full mount list.

db_command - don't hide local static variable in the middle of other
local variable definitions. While here, get rid of the alignment of
variable names.

db_command - make setting have_addr more clear.
Don't set it to false that it's already initialized to.

db_command - make sure count is always initialized.

fix "show kernhist".

set addr = 0 if we don't have_addr and avoid using random garbage in addr.

branches: 1.160.4;

When this file is used when not building the kernel (eg: /usr/sbin/crash)
make cnpollc() go away.

call cnpollc(1) and cnpollc(0) around cngetc().

(christos has a good idea to add a function that does all 3,
and we should switch all the callers in this sequence to use
it (and fix the MD ones missing it still). not all can, as
eg, line-grabbing functions can use cngetsn(), which only
calls cnpollc() twice.)

call spl0() before cpu_reboot(), so that there's a chance that:
- interrupts can work afterwards
- this also means if IO stalls, serial break might work again.

this mimics how reboot(2) ends up calling cpu_reboot().

always call lockdebug_dismiss() from DDB -- there are always some
minimal lockdebug checks in place, even without LOCKDEBUG.

adjust lockdebug_abort() to ignore problems after ld_panic is set
so that there's a chance of this working.

this fixes ddb 'reboot' on softiron od1000.

extend the ddb "show kernhist" command. the synopsis is now:

show kernhist[/i] [addr[,count]]

the i modifier means display info instead of entries, and the count
restricts to the last count entries.

the count option is only supported for specified histories. it
could be extended but requires more logic than the single history,
as multiple histories are merged during display.

add a lockdebug_dismiss() function for DDB to use that tells
lockdebug to avoid asserts. use it for the ddb "reboot"
command, so that it doesn't matter how ddb was entered, the
reboot not trigger mutex checks and not work.

Add "show socket" command written by Hiroki SUENAGA. It prints usage of
system's socket buffers.

branches: 1.153.2;
ddb: rename "show lockstat" to "show lockstats" to avoid conflicting with lockstat(8)

Requested by mrg@

Don't pass a unset address to lockdebug_lock_print

Add a new command, show all locks, which shows information of active locks

The command shows information of all active (i.e., being held) locks that are
tracked through either of LWPs or CPUs by the LOCKDEBUG facility. The /t
modifier additionally shows a backtrace for each LWP additionally. This
feature is useful for debugging especially to analyze deadlocks.

The command is useful only if LOCKDEBUG is enabled.

Add a new command, show lockstat, which shows statistics of locks

Currently the command shows the number of allocated locks.

The command is useful only if LOCKDEBUG is enabled.

branches: 1.149.2;
Add command to print device list.

branches: 1.148.8;
Add ddb command to find a vnode by the address of its lock.
This makes it much easier to convert lockstat traces into understandable
data.

branches: 1.147.2;
ddb: rename show arptab to show routes

show arptab command of ddb is now inappropriate because it actually dumps
routes but arp entries aren't routes anymore. So rename it to show routes
and move the code from if_arp.c to route.c.

ok christos@

Improve ddb(4) show kernhist

1) really prints all the histories merged together (rather than just the
"first" when no argument specified
2) dumps a single history when an argument is given, e.g.
"show kernhist usbhist"
3) uses db_printf correctly

only call the on-enter commands if we're not recursing.
avoids having a fault in the on-enter commands repeating.

we can use show callout now from crash

branches: 1.143.6;
fix userland build, by moving all the unused stuff inside ifdef _KERNEL

branches: 1.142.2;
remove the new command not the old one.

remove previous, dmesg just works.

Add "show dmesg" that prints the contents of the message buffer.

Add "show panic" that shows the current panic string. two ports had it, and
it could be easily made MI.

branches: 1.138.2;
Remove MALLOC_DEBUG and MALLOCLOG, which is dead code after malloc(9) move
to kmem(9). Note: kmem(9) has debugging facilities under DEBUG/DIAGNOSTIC.
However, expensive kmguard and debug_freecheck have to be enabled manually.

branches: 1.137.4;
set panicstr to "reboot forced via kernel debugger" when rebooting
from ddb, similar to how sync from ddb works.

this avoids various asserts and other issues during ddb reboot.

branches: 1.136.4;
move and rename the uvm history code out of uvm_stat to "kernhist".

rename "UVMHIST" option to enable the uvm histories.

TODO:
- make UVMHIST properly depend upon KERNHIST
- enable dynamic registration of histories. this is mostly just
allocating something in a bitmap, and is only for viewing multiple
histories in a merged form.


tested on amd64 and sparc64.

Add:
usage: show proc [/a] [/p] address|pid
/a == argument is an address of any lwp
/p == argument is a pid [default]
From: Vladimir Kirillov proger at wilab dot org dot ua

branches: 1.134.2;
fix extern/static inconsistency

branches: 1.133.2; 1.133.4;
Add #ifdef MQUEUE for db_show_mqueue_cmd().

XXX: Wrong. DDB should be dynamic, modular, etc.

extend 'show event' to take /i /t and /m modifiers, to select interrupt,
trap or misc event types. you can mix them with /f as well, to show all
including zero events for traps and misc, "show event/ftm"

Make 'show event', 'dmesg' work with crash(8).
XXX dmesg fails exactly the same way as /sbin/dmesg.

Avoid "warning: cast from pointer to integer of different size"
on 32-bit sparc64 kernels.

Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

Do not set last_count to zero after every db_command function call it is
static variable. This should fix PR 40712.

- fix vmem unittest. rename VMEM_DEBUG so that it won't be abused again.
- reimplement vmem sanity checks with less code duplication.
- reimplement ddb vmem-related commands in a more consistent ways.
remove automatic whatis.

branches: 1.126.2;
Fix my bug introduced in rev 1.98 of db_command.c where command prefix
not ambiguous in one table was choosed even if it was ambiguous in another
table in same list. Also apply more KNF rules.

Patch was supplied by uwe@.

ddb: new 'show vmem' and 'show all vmems' commands.
Useful to inspect vmem(9) structures.

Make DDB shows right register values on 32-bit sparc64 kernels.
- define db_expr_t as int64_t.
- %gsr is 32-bit, so use db_sparc_intop.
- sync arguments with function prototype.
- ansify.

Add some basic DDB stuff for modules.

more <sys/buf.h> police

+ifdef AIO

XXX ddb needs a way to register commands at runtime!

branches: 1.120.2; 1.120.4;
Replace more intptr_t to uintptr_t when casting VA.

- Fix message queue permissions problems.
- Rake into account umask when creating mqueue.
- Restore DDB command, which was accidentally lost (hi martin).
- Misc.

Make "show uvmhist" available to all arches (not just sparc*) in ddb.

Add a show files command that prints the vnodes associated with a
given struct proc and, if LOCKDEBUG is built in, the lock status of
the uvmobject.

Whitespace clean up.

branches: 1.115.2; 1.115.4; 1.115.6;
Remove clause 3 and 4 from TNF licenses

branches: 1.114.6; 1.114.8; 1.114.10;
Fix help string for "w" - it is "write" not "watch".

When doing a reboot from ddb, set panicstr, so that later LOCKDEBUG checks
show the same mercy as with a real panic.

add ddb "whatis" command. inspired from solaris ::whatis dcmd.

branches: 1.111.2; 1.111.4; 1.111.6;
Add a missing "else". Should finally fix PR kern/37110.

branches: 1.110.2;
Pacify some gcc uninitialized warnings at -O3

machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.108.2;
db_cmd_search: When we compare names we'd better stop at the nul byte.
Don't flip-flop between found and ambiguous. (hi, martin!)

Fix command name matching, avoid jumps to null function pointers and
some minor cosmetics.

Do another setjmp right before entering the command loop, we do not want
to execute the on-enter commands again on errors.

Some cleanup:
- When registering command tables, make sure the builtin commands are
already registered
- Make the command table entry structure private
- Do not bother to store the number of commands in a table, we can quickly
calc that if needed.

No need to db_init_commands() early any more - it will happen on first
entry to ddb.

Special case the initialization of builtin command tables to avoid using
malloc. Fixes PR kern/37042.
Allow "mach" commands registered by LKMs even if there is no builtin
mach command for the arch.

Minor clenaup - db_cmd_loop_done is a bool nowadays, so use true/false
instead of 0/1.
Also make db_init_commands() non-static.

Simplify the "tables are already setup" test.
Print a message on unknown commands (as we used to do before) and get rid
of the now unused macro previously used for that.

Fix build for archs w/o DB_MACHINE_COMMANDS.
Patch from KIYOHARA Takashi.

branches: 1.99.2;
Add missing argument - fixes compilation if KGDB is defined.

Add a new option DDB_VERBOSE_HELP that adds online help to ddb.
From Adam Hamsik.
Minor modifications by me, all bugs are probably mine.

Implementation of POSIX message queues.

Reviewed by: <ad>, <tech-kern>

branches: 1.96.2; 1.96.6; 1.96.8;
Import of POSIX Asynchronous I/O.
Seems to be quite stable. Some work still left to do.

Please note, that syscalls are not yet MP-safe, because
of the file and vnode subsystems.

Reviewed by: <tech-kern>, <ad>

branches: 1.95.4; 1.95.6;
TRUE -> true, FALSE -> false

Fix lossage from boolean_t -> bool and updated x86 bus_dma.

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.92.2;
Merge newlock2 to head.

__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.89.2; 1.89.4;
ANSIfy db_error.

integrate kauth.

Move declaration up ahead of active code in the function, so that
this can be built again with gcc 2.95.3, as used by vax.

quell GCC 4.1 uninitialised variable warnings.

XXX: we should audit the tree for which old ones are no longer needed
after getting the older compilers out of the tree..

Coverity CID 1094: Protect against NULL deref.

branches: 1.84.2; 1.84.4; 1.84.6;
Add a "show all pages" command to DDB which prints one line per physical
page in the system. Useful for getting some idea where all your memory's
gone, at least on a sufficiently small system.

branches: 1.83.2; 1.83.4;
add ddb "sh mbuf" command.

branches: 1.82.2;
add "show all pools" command for ddb.

make DDB_COMMANDONENTER sysctl. (ddb.commandonenter)
suggested by Christos Zoulas.

add a new option DDB_COMMANDONENTER.
which can be useful for eg. broken console input.

branches: 1.79.2; 1.79.8;
more constification fallout

Add lots of const.

move decl of db_machine_command_table from db_interface.h to db_command.c
to avoid an array of incomplete type. found by gcc4.

branches: 1.76.4;
That's not supposed to be there. Sorry.

branches: 1.75.2;
Tango on sysctl_createv() and flags. The flags have all been renamed,
and sysctl_createv() now uses more arguments.

add "show mount" ddb command

add "b" as a synonym for "break", since someone added "bt"
and made "b" an ambiguous prefix.

Add a "show sched_qs" command to dump the run queues. Format is:
1 pid.lid (p_comm)
pid.lid (p_comm)
...

If a queue has procs, but no sched_whichqs bit, it shows:
!1 pid.lid (p_comm)

branches: 1.71.2;
switch to use strlcpy() from libkern.

(1) Don't print the message buffer (via ddb's dmesg command) if the
message buffer has not yet been set up, mimicking code from the top of
the sysctl routine for retrieving the message buffer.

(2) Add a /l modifier to the trace command. This makes it print the
backtrace using printf() instead of db_printf(), which has the nice
side-effect of also putting it into the message buffer. A kernel with
ddb in it but disabled (ie, ddb.onpanic set to zero) will print a
backtrace (which ends up in the message buffer) before dumping (or
not, depending on the value of kern.dump_on_panic) and rebooting, but
if ddb is not disabled, the backtrace is not printed, and there's no
way to get it to display a backtrace such that you can retrieve it
after the dump. The backtrace printed by gdb is sometimes a little
different.

(3) Documentation for the above.

Add a rudimentary 'trap to kgdb' command if KGDB is defined. This allows
systems with only one serial port for console, etc., to have at least some
basic access to remote-kgdb-over-serial, even if it is a bit awkward.

Add a synonym for trace - 'bt' and document it

Remove variable that is only assigned too but not referenced.

Casting from a pointer, to a db_expr_t, has to go via an intptr_t.
(db_expr_t == int64_t on sh5, for example).

branches: 1.65.8;
Add a "show event" ddb command to show the event counters.

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

add RCSIDs

Add `show malloc' to dump the internal state of the MALLOC_DEBUG code.

Add an easy way to dump the message buffer from ddb.

branches: 1.60.2;
Default to kernel_map if no addr is available.

replace vm_map{,_entry}_t with struct vm_map{,_entry} *.

use _KERNEL_OPT

branches: 1.57.2;
alphabetize db_command tables

make db_command_table[] const; since there is no way to dynamically extend
or modify the table at the moment, there is no functionality lost.
See also the discussion on tech-kern (Subject: 'machine' DDB command).

Require the machine-dependant DDB commands to be in db_machine_command_table[]
and link it directly to db_command_table[] so that it's not necessary
to do this at runtime. Make db_machine_command_table[] const on all ports.
g/c now unneded stuff, like db_machine_commands_install(), db_machine_init()

Patch written by enami.

db_command_table needs to be writable at least when DB_MACHINE_COMMANDS
is defined.

constify a little

Restore the top level `callout' command.

add ddb commands "show uvmexp" and "show ncache".
the former used to be "call uvm_dump", the latter is new.

printf format safety

remove include of <vm/vm.h>

In db_command(), fix indentation and add a missing db_skip_to_eol() which
previously caused the next command following a ","-style repitition
to not be properly executed.

In the "reboot" and (badly misnamed) "sync" commands, clear db_recover
before calling cpu_reboot() (which never returns) so we can debug
faults occurring in cpu_reboot().

Only define db_show_arptab if NARP>1;
pointed out by cgd in kern/10314

Stop using DDB-specific "%n" kprintf() format specifier. This was
important as db_sym() can be called through printf() rather than
db_printf(), causing a trap as the standard %n usage is invoked.

The DDB-specific %n code should disappear from subr_prf.c soon...

Change printing of cpu number in the prompt in the MULTIPROCESSOR case

branches: 1.43.2;
Rename the machine-specific stack trace printing functions
from db_stack_trace_cmd() to db_stack_trace_print(),
and add an additional argument, a function pointer for an
output routine (i.e. printf() or db_printf()).

Add db_stack_trace_cmd() in db_command.[ch], calling
db_stack_trace_print() with db_printf() as the printer.

Move count==-1 special handling from db_stack_trace_print() [nee
db_stack_trace_cmd()] to db_stack_trace_cmd() [nascent here].

Again, I'm unable to test compilation on all affected platforms,
so advance apologies for potential brokenness.

Fix a printf format.

Add a "sifting" command to ddb (named from the Sun OpenPROM command of
the same name); it searches the symbol table(s) for all symbols matching
a given substring, and prints.

Extremely useful for when you forget that critical symbol name.

Also, with /F support (cf. "ls -F") to print a char indicating the
symbol type.

Install "show arptab" (db_show_arptab) in the ddb command tree.
Move prototype from netinet/if_inarp.h to ddb/db_interface.h.
Change function to have standard ddb parameters (though they're
ignored).

"p" should be an alias for "print", and was until "ps" was added.
Now it is, again.

From kern/9569.
Preserve ddb's idea of "count" from one command to the next if
repeating a command via hitting newline.

actually add buf- and vnode-printing commands to the table.

add ddb commands for printing vnodes and bufs.

Die, register, die!

On MP, print cpu number prior to prompt

db_cmd_list(): variable ``width'' might have be used uninitialized, initialize
it to 0

- change db_cmd_list() to list commands vertically rather than horizontally
- sort entries in the various command tables, so that the `help' lists
are easier to use. this included hacking db_machine_commands_install()
to search for the "machine" entry to change the `more cmds' pointer,
rather than assuming it was the first entry
- add a `sync' command, which is effectively `reboot 0x100'.
- remove db_help_cmd(); it was unused (and was almost a duplicate of
db_cmd_list()).
- move some extern decls to db_output.h, since they're used in more than one
place now
- rename NEXT_TAB to DB_NEXT_TAB and move to db_output.h

branches: 1.31.2; 1.31.4; 1.31.6;
Improve the pool allocator's diagnostic helpers, adding the ability to
log on a per-pool basis, reentrancy checking, and dumping various pool
information from DDB.

Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.28.2;
completely remove Mach VM support. all that is left is the all the
header files as UVM still uses (most of) these.

branches: 1.27.4;
Include required "opt_ddb.h" header file.

defopt DDB.

- add defopt's for UVM, UVMHIST and PMAP_NEW.
- remove unnecessary UVMHIST_DECL's.

initial import of the new virtual memory system, UVM, into -current.

UVM was written by chuck cranor <chuck@maria.wustl.edu>, with some
minor portions derived from the old Mach code. i provided some help
getting swap and paging working, and other bug fixes/ideas. chuck
silvers <chuq@chuq.com> also provided some other fixes.

this is the rest of the MI portion changes.

this will be KNF'd shortly. :-)

Add a 'reboot' command with a single parameter specifying the boot flags.

branches: 1.22.4;
Add a "kill" command that calls db_kill_proc().

db_expr_t became a long.

Fix db_printf format specifications.

Kernel version of longjmp() does not take the return value of setjmp() as
an argument. setjmp() always returns 1.

Bring setjmp/longjmp prototypes into scope.

eliminate setjmp.h

DDB prototype changes.

Need to include <vm/vm.h>

Fix nits pointed out by GCC 2.7.

Clean up #includes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Add "show callout".

struct command => struct db_command, and define it in db_command.h

Pointers to arrays behave oddly in C...

Push and pop db_recover for error handling, so we can set up temporary catch
points easily. (Alluded to by m68k code, but not implemented yet for some
reason.)

Canonicalize all #includes.

Call vm_map_print and vm_object_print with func ptr to db_printf. Add DDB_MACHINE_COMMANDS from Mach 3.0 version.

branches: 1.5.2;
Add RCS identifiers, remove some completely useless RCS logs and patchkit
headers, and a few other insignificant changes.

Add 'ps' command. Add -more- pager to output from Mach ddb.

add explicit rcs id

after 0.2.2 "stable" patches applied

branches: 1.1.1;
Initial revision

fix various typos in compatibility, mainly in comments.

introduce DDB_END_CMD and replace more than 20 copies of the same
list of NULLs and 0. idea from rillig@.

all touched ports built, several booted.

branches: 1.40.2;
Whitespace

Trailing whitespace

Revert introduction of db_alloc.h.

db_alloc() and friends are already provided in crash.c.
Sorry for confusing you...

Introduce db_alloc.h.

Provide db_alloc() and friends to userland, i.e., crash(8).

Make the db_kernelonly() macro visible to all files and re-use it.
Register access is kernel only.

branches: 1.35.64;
Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.34.2;
Fix my bug introduced in rev 1.98 of db_command.c where command prefix
not ambiguous in one table was choosed even if it was ambiguous in another
table in same list. Also apply more KNF rules.

Patch was supplied by uwe@.

branches: 1.33.8; 1.33.10;
Remove clause 3 and 4 from TNF licenses

branches: 1.32.2; 1.32.4;
Fix CS_MACH definition. 0x10 is spelled 0x10, not 0x16.

branches: 1.31.6;
Convert many of the uses of __attribute__ to equivalent
__packed, __unused and __dead macros from cdefs.h

branches: 1.30.4; 1.30.6; 1.30.10;
No need to db_init_commands() early any more - it will happen on first
entry to ddb.

Minor clenaup - db_cmd_loop_done is a bool nowadays, so use true/false
instead of 0/1.
Also make db_init_commands() non-static.

Add a new option DDB_VERBOSE_HELP that adds online help to ddb.
From Adam Hamsik.
Minor modifications by me, all bugs are probably mine.

branches: 1.27.4; 1.27.16; 1.27.18;
Fix lossage from boolean_t -> bool and updated x86 bus_dma.

branches: 1.26.8;
Mark db_error __attribute__((__noreturn__)).

branches: 1.25.4; 1.25.8;
make DDB_COMMANDONENTER sysctl. (ddb.commandonenter)
suggested by Christos Zoulas.

branches: 1.24.2; 1.24.8;
more constification fallout

Add lots of const.

branches: 1.22.16;
ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

Add `show malloc' to dump the internal state of the MALLOC_DEBUG code.

branches: 1.20.2; 1.20.4;
Require the machine-dependant DDB commands to be in db_machine_command_table[]
and link it directly to db_command_table[] so that it's not necessary
to do this at runtime. Make db_machine_command_table[] const on all ports.
g/c now unneded stuff, like db_machine_commands_install(), db_machine_init()

Patch written by enami.

constify a little

add ddb commands "show uvmexp" and "show ncache".
the former used to be "call uvm_dump", the latter is new.

Rename the machine-specific stack trace printing functions
from db_stack_trace_cmd() to db_stack_trace_print(),
and add an additional argument, a function pointer for an
output routine (i.e. printf() or db_printf()).

Add db_stack_trace_cmd() in db_command.[ch], calling
db_stack_trace_print() with db_printf() as the printer.

Move count==-1 special handling from db_stack_trace_print() [nee
db_stack_trace_cmd()] to db_stack_trace_cmd() [nascent here].

Again, I'm unable to test compilation on all affected platforms,
so advance apologies for potential brokenness.

Add a "sifting" command to ddb (named from the Sun OpenPROM command of
the same name); it searches the symbol table(s) for all symbols matching
a given substring, and prints.

Extremely useful for when you forget that critical symbol name.

Also, with /F support (cf. "ls -F") to print a char indicating the
symbol type.

add ddb commands for printing vnodes and bufs.

- change db_cmd_list() to list commands vertically rather than horizontally
- sort entries in the various command tables, so that the `help' lists
are easier to use. this included hacking db_machine_commands_install()
to search for the "machine" entry to change the `more cmds' pointer,
rather than assuming it was the first entry
- add a `sync' command, which is effectively `reboot 0x100'.
- remove db_help_cmd(); it was unused (and was almost a duplicate of
db_cmd_list()).
- move some extern decls to db_output.h, since they're used in more than one
place now
- rename NEXT_TAB to DB_NEXT_TAB and move to db_output.h

branches: 1.13.2; 1.13.4; 1.13.6;
Improve the pool allocator's diagnostic helpers, adding the ability to
log on a per-pool basis, reentrancy checking, and dumping various pool
information from DDB.

Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.10.4; 1.10.8;
initial import of the new virtual memory system, UVM, into -current.

UVM was written by chuck cranor <chuck@maria.wustl.edu>, with some
minor portions derived from the old Mach code. i provided some help
getting swap and paging working, and other bug fixes/ideas. chuck
silvers <chuq@chuq.com> also provided some other fixes.

this is the rest of the MI portion changes.

this will be KNF'd shortly. :-)

Add a 'reboot' command with a single parameter specifying the boot flags.

branches: 1.8.12;
DDB prototype changes.

Clean up #includes.

Add prototypes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

struct command => struct db_command, and define it in db_command.h

Remove RCS logs.

branches: 1.2.4;
add explicit rcs id

branches: 1.1.1;
Initial revision

Some cleanup:
- When registering command tables, make sure the builtin commands are
already registered
- Make the command table entry structure private
- Do not bother to store the number of commands in a table, we can quickly
calc that if needed.

branches: 1.1.2;
Add a new option DDB_VERBOSE_HELP that adds online help to ddb.
From Adam Hamsik.
Minor modifications by me, all bugs are probably mine.

Make this build.

Hi jakllsch@.

Unbreak db_cpu_*. Allows backtraces on i386 to progress beyond the
first function by not doing a NULL-deref in the debugger.

Bug located by me. Patch implementation by riastradh@. Hi rmind@.

branches: 1.5.2;
Remove cpu_queue (and thus eleminate another use of CIRCLEQ) by replacing
its uses with cpu_infos array. Extra testing by christos@.

branches: 1.4.4; 1.4.14; 1.4.18;
When looking for the next element of cpu_queue compare the pointer to the
address of the head, not to the head itself. Not sure if the cast of the
arg to db_value_name() is right, but works on i386 and compiles on sparc64.

Stack traces from ddb work again on i386.

branches: 1.3.2; 1.3.4; 1.3.6; 1.3.10; 1.3.12; 1.3.14;
include <stdbool.h> if not _KERNEL, and don't include kauth.h if _KERNEL.

now crash builds on i386.

include ddb headers *after* sys/types.h or sys/param.h.

Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.1.2; 1.1.4; 1.1.6;
Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

remove the MI Debugger() added recently. It doesn't make sense in
it's current form.
make Debugger just an alias for cpu_Debugger

rename the MD Debugger() to cpu_Debugger()
add MI Debugger() which switches to console if wscons is used prior
to calling cpu_Debugger()

remove more DB_ELFSIZE

use Elf_Sym ** instead of casting.

branches: 1.27.22;
Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.25.8; 1.25.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.24.38; 1.24.40; 1.24.42;
TRUE -> true, FALSE -> false

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.22.22; 1.22.60;
make stack trace on i386 work again.
(signed vs unsigned)

Fix signed/unsigned comparison warnings.

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

don't need <sys/types.h> when including <sys/param.h>

add RCSIDs

PR/14498: Wesley Chen: Identify loaded modules vs. kernel, because we don't
load the elf header with them.

branches: 1.16.4;
Extra sanity checking on ELF headers: Make sure the bootloader actually
loaded (or claims to have loaded) everything we need.

Don't depend on the bootloader's having loaded the .shstrtab section
in order to load the symbol table. Instead of using the sections
called ".symtab" and ".strtab", use the first SYMTAB section (the
ELF spec says there should currently only be one) and the STRTAB
section that's linked to it. I believe this is more robust, and it
certainly makes life easier for the bootloader.

branches: 1.14.2; 1.14.4;
constify a little

s/preserving ... bytes of symbol table/using ... bytes of symbol table/

"preserving" implies that the data may be copied and might otherwise
be freed, and that other data in the symbol table or sections of
the kernel might be freed, however there is no mechanism for these
things to take place, so "using" is clearer.

Detect ELF sections (string and symbol tables) by looking them
up by name (in .shstrtab) instead of guessing based on section
type (and throwing away the small one). In addition to being cleaner,
multiple symbol tables are no longer and error condition, so
booting netbsd.gdb no longer results in ddb being unable to use
*any* symbols.

branches: 1.11.4;
Add a "sifting" command to ddb (named from the Sun OpenPROM command of
the same name); it searches the symbol table(s) for all symbols matching
a given substring, and prints.

Extremely useful for when you forget that critical symbol name.

Also, with /F support (cf. "ls -F") to print a char indicating the
symbol type.

Update to match new SVR4-style definition names in <sys/exec_elf.h>.

Remove ELFNAME macros which are already defines in exec_elf.h

branches: 1.8.8; 1.8.10; 1.8.12;
Avoid arithmetic on `void *' since that's not ANSI C.

- Make the ddb_init() interface consistent for a.out and ELF.
- Allow DDB to run-time switch for a.out or ELF symbol table support.

Merge paddr_t changes into the main branch.

bzero->memset, bcopy->memcpy, bcmp->memcmp

branches: 1.4.2;
Whitespace.

Clean up usage of printf() vs db_printf(): The diagnostic messages
when symbol tables are loaded don't belong to the user interaction,
their output lines should not be counted by ddb's builtin pager.

Make sure debugging symbol sections are aligned.

branches: 1.1.2;
Add support for Elf symbol tables to DDB, written from scratch by me.
Known to work in DDB under NetBSD/alpha.

changes of r1.39 was incomplete. only "examin/m" could handle 'q'.
added support 'r','x','z','d','u', and 'o' with 'q' modifier on 32bit arch.

- examin/m displays with splitting by spaces as specified size
- add support 'q' modifier on all arch
- consider endianness

Add support for 'p' qualifier for pointer values on examine.
this shows as a pointer with symbol if possible. (e.g. "x/p $sp,10")

- add or adjust /* FALLTHROUGH */ where appropriate
- add __unreachable() after functions that can return but won't in
this case, and thus can't be marked __dead easily

branches: 1.36.18;
Add support for 'q' qualifier for 64-bit values on examine and write.
Align data to long so that the MD parts can use aligned load/stores.
Don't use db_extend but use db_expr_t based expressions.

branches: 1.35.14; 1.35.32;
Introduce DDB_EXPR_FMT and replace the logic around DB_EXPR_T_IS_QUAD.

branches: 1.34.4; 1.34.6;
Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.33.4;
more <sys/buf.h> police

branches: 1.32.38; 1.32.42; 1.32.46; 1.32.48; 1.32.52;
TRUE -> true, FALSE -> false

Fix lossage from boolean_t -> bool and updated x86 bus_dma.

branches: 1.30.4;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.28.20; 1.28.22;
merge ktrace-lwp.

branches: 1.27.2;
more constification fallout

Fix format string warnings for platforms where
sizeof(db_addr_t) > sizeof(long). e.g. ILP32 on sh5.

branches: 1.25.2;
On some platforms, sizeof(db_expr_t) > sizeof(long), so add a method
to select the right format string (at compile time) when displaying
variables of type db_expr_t.

This fixes a problem where ddb(4) would only display the low 32-bits
of registers for an ILP32 kernel on SH5, even though registers
(and db_expr_t) are always 64-bits wide.

switch to use strlcpy() from libkern.

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

add RCSIDs

branches: 1.21.2; 1.21.4;
#include <systm.h> for string fn prototypes.

Add x/m option functionality

Add two new functions, db_format_radix() and db_format_hex(), which supplant
the former %r and %z nonstandard kprintf strings.

Interface change: db_printsym() takes a third argument, pr, a function
pointer indicating how to print the symbol. This allows db_printsym()
to called in places where db_printf() is not an appropriate output
function.

While straightforward, apologies in advance if I've introduced any minor
syntax errors; I was unable to test compilation this on all the affected
platforms.

Die, register, die!

branches: 1.16.2;
New format modifier: `L' gives true `long' vs `l' or `legacy' long.

Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.13.20;
db_expr_t became a long.

fix various 64-bitness bugs.

Fix db_printf format specifications.

DDB prototype changes.

Make x/a print the address; simplify slightly.

When use default form when printing location (i.e. after step).
Use less space printing longs (so you can fit four per line).

Fix some weird formatting.

Fix data size specifiers and multiple format specifiers, like:
db> x/bxxxx ADDRESS

branches: 1.5.2;
Make "examine" (x/i) always print the address.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Canonicalize all #includes.

branches: 1.2.4;
add explicit rcs id

branches: 1.1.1;
Initial revision

Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.15.42; 1.15.52; 1.15.58;
TRUE -> true, FALSE -> false

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.13.32; 1.13.70;
ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

add RCSIDs

branches: 1.11.2; 1.11.4;
Fix compiler warning.

Fix for:

7. Numbers starting with [a-f] should work, but symbols
of the same name should have priority.

branches: 1.9.2;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.7.16;
Use db_expr_t rather than int when reading a value from the kernel
address space, to avoid type-size problems.

fix various 64-bitness bugs.

DDB prototype changes.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Canonicalize all #includes.

branches: 1.2.4;
add explicit rcs id

branches: 1.1.1;
Initial revision

tab consistency. nfci.

Remove clause 3 and 4 from TNF licenses

branches: 1.15.38; 1.15.40; 1.15.42;
Fix lossage from boolean_t -> bool and updated x86 bus_dma.

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.13.26;
merge ktrace-lwp.

branches: 1.12.2;
more constification fallout

branches: 1.11.2;
switch to use strlcpy() from libkern.

Add a rudimentary 'trap to kgdb' command if KGDB is defined. This allows
systems with only one serial port for console, etc., to have at least some
basic access to remote-kgdb-over-serial, even if it is a bit awkward.

Add db_recover

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

branches: 1.7.2; 1.7.4;
Move platform db_trap callback from arch/mips into ddb as suggested by
jhawk. This callback is used by platform code to manage things like
watchdogs that should be disabled while in ddb. Done as a callback
for processors such as mips that support lots of different systems.

branches: 1.6.10;
- Make the ddb_init() interface consistent for a.out and ELF.
- Allow DDB to run-time switch for a.out or ELF symbol table support.

Assign copyright to TNF.

Change ddb_init() for a.out to provision for kernels which are not booted by
the current 'bootblock' method, like the Shark, which preloads and preformats
the symbol table right in the boot code.

Add declaration for db_delete_line

Update prototypes for new function signatures. Note, the ddb_init()
signature is currently different for the DB_AOUT_SYMBOLS vs. the
DB_ELF_SYMBOLS cases. (XXX)

DDB prototype changes.

Trailing whitespace

Print db> prompt in db_read_line(), not db_readline().

The former is what DDB repl calls. The latter performs the actual
input so let other code use it without the unwanted db> prompt. It's
already used by ACPI (and AcpiOsWaitCommandReady supplies its own
prompt). I also use it for my uncommitted Forth scripting for DDB.

branches: 1.26.60; 1.26.64;
Define db_hist_put() only when DDB_HISTORY_SIZE != 0 so that
build successes on ports like hpcarm.

Fix indent in prevous commit.

Don't overflow DDB command history. Coyote Point changelist description:

DDB is flakey. The command history wanders past the bounds. Way
past. When it hits some boolean that indicates a.out format symbol
tables are to be used, and here is the pointer to the function, the
call thru the NULL function pointer renders the debug session entirely
unsatisfactory, outcome wise.

branches: 1.23.2; 1.23.4;
Make ddb compile and work in userspace. Mostly this is comprised of three
types of changes:

- Add a few new methods to replace stuff like p_find(), CPU_INFO_FOREACH.

- Use db_read_bytes() instead of accessing kernel structures directly,
and similar changes.

- Add ifdef _KERNEL where the above hasn't been done, and an XXX comment.

branches: 1.22.42; 1.22.52; 1.22.58;
TRUE -> true, FALSE -> false

branches: 1.21.26;
merge ktrace-lwp.

branches: 1.20.2;
Add lots of const.

branches: 1.19.6;
Dump DDB_* and SYMTAB_SPACE options to opt_ddbparam.h rather than opt_ddb.h.
These options are used in limited files but #include "opt_ddb.h" are
everywhere, and changing them caused almost full recompilation.

/*CONTCOND*/ while (0)'ed macros

ANSIfy, KNF.
Make some variables and functions static when not used outside of a module.
Make variables in headers extern.
Delete the unused db_find_watchpoint() function.

- pull in opt_kgdb.h where necessary
- replace opt_kgdb_machdep.h with opt_kgdb.h
- defparam opt_kgdb.h:
KGDB_DEV KGDB_DEVNAME KGDB_DEVADDR KGDB_DEVRATE KGDB_DEVMODE
- move from opt_ddbparam.h to opt_ddb.h:
DDB_FROMCONSOLE DDB_ONPANIC DDB_HISTORY_SIZE DDB_BREAK_CHAR SYMTAB_SPACE
- replace KGDBDEV with KGDB_DEV
- replace KGDBADDR with KGDB_DEVADDR
- replace KGDBMODE with KGDB_DEVMODE
- replace KGDBRATE with KGDB_DEVRATE
- use `9600' instead of `0x2580' for 9600 baud rate
- use correct quotes for options KGDB_DEVNAME="\"com\""
- use correct quotes for options KGDB_DEV="17*256+0"
- remove unnecessary dependancy on Makefile for kgdb_stub.o
- minor whitespace cleanup

add RCSIDs

branches: 1.14.2; 1.14.4;
Fix CTRL-T to behave properly. It should swaps the current and previous
characters, not the previous and second-to-previous.

Die, register, die!

branches: 1.12.2;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.10.8;
defopt DDB.

s/DB_HISTORY_SIZE/DDB_HISTORY_SIZE/, for consistency with other DDB #defines

PR/4504: Tero Kivinen: Add DB_HISTORY_SIZE option to enable history in ddb

DDB prototype changes.

Remove bogus definition of cnpollc().

db_error() does not return.

New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'

Canonicalize all #includes.

branches: 1.2.4;
add explicit rcs id

branches: 1.1.1;
Initial revision

ddb: New db_stack_trace_print_ra.

Takes an optional extra argument for the return address so that the
stack parser has a chance to use instruction-dependent interpretation
of the frame to find parent frames.

Implemented only on Alpha for now, which needs it. Other
architectures just ignore the extra argument.

PR port-alpha/57716

XXX pullup-10
XXX pullup-9
XXX pullup-8

Back out the following revisions on behalf of core:

sys/sys/lwp.h: revision 1.228
sys/sys/pipe.h: revision 1.40
sys/kern/uipc_socket.c: revision 1.306
sys/kern/kern_sleepq.c: revision 1.84
sys/rump/librump/rumpkern/locks_up.c: revision 1.13
sys/kern/sys_pipe.c: revision 1.165
usr.bin/fstat/fstat.c: revision 1.119
sys/rump/librump/rumpkern/locks.c: revision 1.87
sys/ddb/db_xxx.c: revision 1.78
sys/ddb/db_command.c: revision 1.187
sys/sys/condvar.h: revision 1.18
sys/ddb/db_interface.h: revision 1.42
sys/sys/socketvar.h: revision 1.166
sys/kern/uipc_syscalls.c: revision 1.209
sys/kern/kern_condvar.c: revision 1.60

Add cv_fdrestart() [...]
Use cv_fdrestart() to implement fo_restart.
Simplify/streamline pipes a little bit [...]

This changes have caused regressions and need to be debugged.
The cv_fdrestart() addition needs more discussion.

Simplify/streamline pipes a little bit:

- Allocate only one struct pipe not two (no need to be bidirectional here).
- Then use f_flag (FREAD/FWRITE) to figure out what to do in the fileops.
- Never wake the other side or acquire long-term (I/O) lock unless needed.
- Whenever possible, defer wakeups until after locks have been released.
- Do some things locklessly in pipe_ioctl() and pipe_poll().

Some notable results:

- -30% latency on a 486DX2/66 doing 1 byte ping-pong within a single process.
- 2.5x less lock contention during "make cleandir" of src on a 48 CPU machine.
- 1.5x bandwith with 1kB messages on the same 48 CPU machine (8kB: same b/w).

Add some simple DDB show commands: condvar, selinfo, sleepq

don't provide db_lstacktrace() for crash (yet?).

db_lstacktrace() can't use db_stacktrace_print and log() directly.

log() takes a 'int level' first argument, that must be supplied.
add an inline wrapper that calls vlog() with LOG_INFO, and the
supplied va_list.

(not noticed because this macro is not used anywhere in src but
i have a use in some uncommited code, that now failed to compile.)

Allow for MD code to use a different function than db_stack_trace_print()
to be used with db_stacktrace() and friends by pre-defining the
db_stacktrace_print macro.

branches: 1.37.10;
add two db_stack_print_trace frontends that emit to the system log
or the user's terminal.

kind of wrong to be called db_*() when they'll typically be used
outside of ddb itself, but db_stacktrace() itself is no different.
fix this by making the stack trace support separate from ddb?

branches: 1.36.4;
Add command to print device list.

Revert printf return value change.

[syzkaller] Cast all the printf's to (void *)
> as a result of new printf(9) declaration.

add a simple stacktrace macro

branches: 1.32.2;
ddb: rename show arptab to show routes

show arptab command of ddb is now inappropriate because it actually dumps
routes but arp entries aren't routes anymore. So rename it to show routes
and move the code from if_arp.c to route.c.

ok christos@

branches: 1.31.14;
remove previous, dmesg just works.

Add "show dmesg" that prints the contents of the message buffer.

Add "show panic" that shows the current panic string. two ports had it, and
it could be easily made MI.

branches: 1.28.6;
add __printflike where needed

branches: 1.27.4; 1.27.8;
Introduce DDB_EXPR_FMT and replace the logic around DB_EXPR_T_IS_QUAD.

Add:
usage: show proc [/a] [/p] address|pid
/a == argument is an address of any lwp
/p == argument is a pid [default]
From: Vladimir Kirillov proger at wilab dot org dot ua

branches: 1.25.4; 1.25.6;
- fix vmem unittest. rename VMEM_DEBUG so that it won't be abused again.
- reimplement vmem sanity checks with less code duplication.
- reimplement ddb vmem-related commands in a more consistent ways.
remove automatic whatis.

branches: 1.24.4;
Add some basic DDB stuff for modules.

branches: 1.23.2;
Add a show files command that prints the vnodes associated with a
given struct proc and, if LOCKDEBUG is built in, the lock status of
the uvmobject.

branches: 1.22.2; 1.22.4; 1.22.6;
Remove clause 3 and 4 from TNF licenses

branches: 1.21.20; 1.21.22; 1.21.24;
Implementation of POSIX message queues.

Reviewed by: <ad>, <tech-kern>

branches: 1.20.2; 1.20.6; 1.20.8;
Import of POSIX Asynchronous I/O.
Seems to be quite stable. Some work still left to do.

Please note, that syscalls are not yet MP-safe, because
of the file and vnode subsystems.

Reviewed by: <tech-kern>, <ad>

branches: 1.19.4; 1.19.6;
Fix lossage from boolean_t -> bool and updated x86 bus_dma.

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.17.24;
Add CONSTCOND to DB_EXPR_T_IS_QUAD.

branches: 1.16.2;
add "show all pools" command for ddb.