CVS log for src/usr.bin/config/gram.y

Up to [cvs.NetBSD.org] / src / usr.bin / config

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

With config(1) as it exists today, a kernel Makefile is able to implement
logic on kernel options so long as those options are not defflag'd or
defparam'd.  This works because such options are automatally added to the
IDENT var in the kernel Makefile as a preprocessor define, and the Makefile
can then do an operation like:

.if !empty(IDENT:M-DSOMECOOLCPUOPTION)
CFLAGS+= -mcpu=somecoolcpu
.endif

Unfortunately, this precludes making it possible to generate a compile-time
dependency on SOMECOOLCPUOPTION, or having SOMECOOLCPUOPTION imply another
kernel config option using the normal config(1) option dependency function.

Fix this by introducing a new option description keyword: mkflagvar.  This
keyword marks an already defflag'd option as wanting a kernel Makefile var
defined if that option is selected in the kernel config file.  So:

defflag opt_coolcpu.h SOMECOOLCPUOPTION ANOTHERCOOLCPUOPTION
mkflagvar SOMECOOLCPUOPTION ANOTHERCOOLCPUOPTION

will cause:

KERNEL_OPT_SOMECOOLCPUOPTION="1"
KERNEL_OPT_ANOTHERCOOLCPUOPTION="1"

...to be emitted into the kernel Makefile if those options are, in fact,
selected with "options ..." in the kernel config file, thus allowing for
a compile-time dependency on the option to be generated in addition to
Makefile logic, which now looks like:

.if !empty(KERNEL_OPT_SOMECOOLCPUOPTION)
CFLAGS+= -mcpu=somecoolcpu
.endif

Add debug output for ENDDEFS.

This makes it more clear in the debug output where config switched
from definitions to selections.

Merge changes from current as of 20200406

Pull up the following, requested by christos in ticket #776:

	usr.bin/config/TODO                             up to 1.32
	usr.bin/config/defs.h                           up to 1.105
	usr.bin/config/files.c                          up to 1.37
	usr.bin/config/gram.y                           up to 1.55
	usr.bin/config/main.c                           up to 1.99
	usr.bin/config/scan.l                           up to 1.30
	usr.bin/config/sem.c                            up to 1.84
	usr.bin/config/util.c                           up to 1.21

config(1): fix issue with incorrect handling of elifdef statements.

Keep track where more objects are declared so that we can print where things
have been redefined.

Accept "-no" as a "no" that does not cause errors if the object deleted
does not exist.

Pull up following revision(s) (requested by mlelstv in ticket #1164):
	usr.bin/config/sem.c: revision 1.74
	usr.bin/config/config.5: revision 1.35
	usr.bin/config/gram.y: revision 1.53
Extend syntax of config phrase, a quoted string instead of a device name
is passed as root specification string. This can be used to specify a
wedge by name.

Extend syntax of config phrase, a quoted string instead of a device name
is passed as root specification string. This can be used to specify a
wedge by name.

Introduce a new syntax, "buildprefix", to specify prefix of files under kernel
build subdirectory.  This is not used now that everything is built at the
top of kernel build directory.  It will become mandatory for source/object
files put out of kernel source tree to specify corresponding build subdirectory.
Only ``no dots''  relative path is accepted as "buildprefix".

Merge struct objects into struct files.  Keep per-suffix file lists.

Restore the old behavior of "maxusers" to allow the value to be overriden.
Problem reported by John D. Baker.

Convert MAXUSERS, take 2.

Define MAXUSERS in opt_param.h and include it from sys/param.h, for more
accurate dependency.  Don't define -DMAXUSERS and add it to ${CPPFLAGS}.
config(5)'s "maxusers" syntax doesn't change.

(The version just before revert was OK, but I had something bad in my obj
dirs.)

Revert "Retire optional `rule' argument of `file' command".  It is still used
in m68k ports.

Retire optional `rule' argument of `file' command.

This has existed since Rev. 1.1 of Torek's config(8) [1], but probably has
never been used by anyone.

(I can not imagine how a `file' like look like if `rule' is specified.)

[1] http://cvsweb.netbsd.org/bsdweb.cgi/src/usr.sbin/config.new/Attic/files.c?rev=1.1

Pull up following revision(s) (requested by mrg in ticket #572):
	usr.bin/config/Makefile: up to 1.10
	usr.bin/config/TODO: up to 1.14
	usr.bin/config/config.1: up to 1.17
	usr.bin/config/config.5: up to 1.25
	usr.bin/config/defs.h: up to 1.64
	usr.bin/config/files.c: up to 1.18
	usr.bin/config/gram.y: up to 1.46
	usr.bin/config/hash.c: up to 1.11
	usr.bin/config/lint.c: up to 1.15
	usr.bin/config/main.c: up to 1.74
	usr.bin/config/mkdevsw.c: up to 1.12
	usr.bin/config/mkheaders.c: up to 1.26
	usr.bin/config/mkioconf.c: up to 1.28
	usr.bin/config/mkmakefile.c: up to 1.37
	usr.bin/config/mkswap.c: up to 1.8
	usr.bin/config/pack.c: up to 1.9
	usr.bin/config/scan.l: up to 1.22
	usr.bin/config/sem.c: up to 1.71
	usr.bin/config/sem.h: up to 1.19
	usr.bin/config/util.c: up to 1.19
sync config(1) with HEAD.

makedev returns ~uint64_t, but val is int64_t. Cast for -Wconversion.

config(1), config(5): Introduce "select"

o Introduce a new selection directive "select" to select an attribute (as a
  module) and its dependencies.
o Support "no select" too.
o Stop abusing "options" to select an attribute.
o Bump config(1) version.

- add RCSID's
- fix -Wconversion (except scanner)

Allow dependency on undefined attributes, so that attribute definitions can
be written out-of-order, like:

	# b is not defined yet
	define	a: b
	file	a.c	a

	# b is defined here
	define	b
	fine	b.c

Split plain/interface/devclass attribute initializers.

Sort definition/selection rules to clarify which is which.

The word "configuration" is too ambiguous; use "selection" meaning that
user selects some of provided "definitions".

Rebase to HEAD as of a few days ago.

Rebase.

implement "no ident".

sync with head.

for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.

this commit was splitted into small chunks to avoid
a limitation of cvs.  ("Protocol error: too many arguments")

ioconf/pseudo-root have been in here unchanged for 3+ years and their use
has spread from rump kernels into dozens of kernel modules.  Let's pretend
we don't want to see config yell about them being experimental every time.

sync with head

Give option definitions their own data structure instead of using nvlists.
(and using messy hacks to make up for nvlists not holding quite the right
things)

Move locator lists to their own data structure. This can use more tidying;
it is not clear to me at the moment what the "string" and "num" values
pushed around in locator lists are supposed to actually mean.

More naming improvements.

Create a struct condexpr type to hold condition expressions, instead
of abusing struct nvlist to make trees.

(These are the a|b and a&b constructs.)

Rename a few more things; try to use the same terminology as config(5).

Create a new type struct attrlist to manage lists of attributes,
instead of using struct nvlist.

(struct nvlist holds lists (or trees!) of semi-arbitrary stuff with no
structure and almost no type safety; it should go away.)

Strengthen the logic that wraps allocations going on the parser stack.
(This allows cleaning up when a syntax error occurs, because yacc just
dumps the parser stack on the floor.)

The new logic can handle arbitrary data types instead of being limited
to struct nvlist.

Write out the expression grammar for dependency constraints instead of
using precedence rules. Precedence rules (besides being a hack) often
lead to unexpected/unwanted behavior in corner cases.

Sort productions in grammar. No functional change.

Simplify some things. Rename some productions for clarity.

A bit more of previous.

Reformat according to (my) basic standards for yacc grammars.
No functional change.

Comments.

For the simple cases, augment device-major with information on how
a driver expects /dev/node -> minor mappings to go and include that
information in devsw_conv.
(no, I didn't plow through all the MD majors files)

print ioconf WARNING in caps too

Add a highly experimental pseudo-root feature to be used in conjuction
with the also-experimental ioconf keyword.  pseudo-root allows to
specify a root at any point in the device tree without having
attachments from the actual root.

For example, instead of having a config file like this:
mainbus0 at root
bus_a? at mainbus0
bus_b? at bus_a*
device7 at bus_b?

You can have one like this:
pseudo-root bus_b*
device7 at bus_b?

This will produce the relevant ioconf.c glue for device number 7
only instead of the whole 9 yards from root.  Perhaps needless to
say, this can be used to generate snippets of config glue for
modules and, let's not deny that my main motivation for doing this,
rump components.

This is part 2/3 of my modifications to config (the last part is
autogenerating source file lists and component Makefiles).

No strong objection from cube (after a little pressuring ;), but
like he said, the implementation will most likely need some more
tweaking and may not work correctly under all pseudo-root uses yet.

Remove fs_foo.h support from deffs now that nothing uses it.

Introduce experimental support for ioconf-only configuration files.
This is done by giving the "ioconf" keyword in the config file.
As a result, config produces only ioconf.c and locators.h.  Currently,
only "monolithic" configurations with the device path starting from
root are supported.  Eventually, the goal is to support a local
root in the input file from any point along the device tree using
files.* in our kernel tree.  This will make autogenerating the
config glue for device modules possible instead of having to write
it by hand like is currently required (yes, it sounds simple to
implement, but ...).

reviewed by cube.
(the next part will demand major discussions with you, so prepare ;)

Sync with HEAD.

Third (and last) commit. See http://mail-index.netbsd.org/source-changes/2009/05/13/msg221222.html

Actually allow a full expression for the condition for "makeoptions" in
files.*

Patch from Yorick Hardy.

rename nv_int -> nv_num and make it long long, so that dev_t will fit when
we change it.

Sync w/ -current. 34 merge conflicts to follow.

sync with head.

Disable use of pseudo-devices defined "defpseudo" with interface
attributes, but provide compatibility for older source trees (according
to the "version" in conf/files and other files.*.
reviewed by cube

Seperate pseudo-devs which don't use device_t et al. (the classical
ones, mostly network pseudo interfaces) and those which use autoconf
to attach instances dynamically, or to attach child devices.
Use a new keyword "defpseudodev" for the latter.
Issue a warning if interface attributes are used with "old" pseudodevs,
but let it work as before for now. This will be removed soon.
approved by cube

sync with HEAD

Don't defattr machinearch if it's NULL.  Pointed out by
Juergen Hannken-Illjes.

Make the "machine" statement define attributes for `machine', `arch' and
all the `subarch' given as parameters.  This is useful in combination with
ifdef...

- rename {x,}error -> cfg{x,}error and {x,}warn -> cfg{x,}warn so that we
  can use err(3) and warn(3) to display errors instead of using inconsistent
  fprintf's.
- remove emalloc/estrdup etc. We have them in libutil.

Handle defopt lists properly, as ':=' syntax introduce an additional
element...  For that, introduce nvcat(nv1, nv2).

Add new syntaxic sugar to provide a value in lint (i.e., config -L)
configurations for defparams.  It will be used _only_ by config -L.

This will greatly help automating tests with lint configurations,
hopefully.

Add support for specifying a default value on a 'defparam' line.
If a default is specified then the option is always defined in the
corresponding .h file.
Particularly useful for parameters where the default action isn't the
same a defining the value zero, given that the current use of #if defined(opt)
is open to problems wehere the relevant .h file isn't actually included, so
requiring an option to always have a value makes sense.
Also included (but commented out) is code that adds a global symbol to
the object file to stop objects compiled with different values for the
same option being linked together - I'm not quite happy with it yet!

Allow deffs to take dependencies like defflag and defopt

Now that this is part of tools, we don't need the extra build glue.

Sync with head.

D'uh.  Forgot the file name argument in the "obsolete" statements.

Introduce two new statements:

    obsolete defflag <option> [, <option> [, ...]]
    obsolete defparam <option> [, <option> [, ...]]

The two statements actually do the same thing (there could be only one),
but it makes things less cryptic that way.  The optional ": deps" part of
a 'defflag' or 'defparam' still has to be dropped when it gets obsoleted.

When the user has 'options OBSOLETE_OPTION' in his configuration file, it
is ignored (that is, opt_*.h files are generated as if it wasn't there),
and the user gets a warning about it.

Bump version to 20060525.

When someone first uses that syntax in the tree, a "version 20060525"
statement should be added before its occurrence, preferably at the top
of sys/conf/files.

Suggested by Matt Thomas a few months ago.

Add 'no config <name>'.

Introduce versioning to config(1).  This will allow us to provide a way to
error out in a bit more friendly way when the user is trying to use
config(1) on a too old or too recent source tree.

To achieve that, introduce the "version NUMBER" statement which can be use
about anywhere in the config files.  Also, use two defines, CONFIG_VERSION
(which is the actual version of binary), and CONFIG_MINVERSION, which is
the minimum version the binary supports.

Allowing a range of versions serves several purposes:  first it allows me
to introduce the versioning without requiring it to be used right away in
the kernel tree, which means it will be possible to introduce new features
of config(1) rather progressively in the future.  E.g., using 'no pci' in
a config file could only require the new version in that config file, so
that the rest remains compatible.

In the end, an actual bump of the main config system (i.e., in conf/files)
will only be required when e.g., ioconf.c semantics change.

(Mostly-)silently accepted on tech-kern.  Error messages turned into
correct and meaningful English thanks to Tracy and Perry.

Introduce two new statements:

  no device at <attachment>

     <attachment> can take two forms:  either numbered/wildcarded, in which
     case only exactly matching instances will be removed, or plain (with
     no number or wildcard), in which case all matching instances will be
     removed.

     When <attachment> is a plain interface attribute, all instances using
     that attribute (either directly or through an explicit device) will be
     removed.

     E.g.:
             auich* at pci? dev ? function ?
             audio0 at audiobus?
             audio1 at auich?
             audio* at auich0

             no device at auich0     -> removes audio*
             no device at auich?     -> removes audio1
             no device at auich      -> removes audio1 _and_ audio*
             no device at audiobus?  -> removes audio0
             no device at audiobus   -> removes audio0, audio1 and audio*

  no <device>

     As in the previous case, <device> can either be numbered/starred, in
     which case all exactly matching instances are removed, or plain, in
     which case all instances of the device are removed.

    E.g.: (continuing previous example)

             no audio* -> removes 'audio* at auich0'
             no audio  -> removes all audio instances

Move package/cinclude/include handling from the parser to the scanner,
where it belongs. This has the side effect of fixing PR toolchain/30903.

config and genassym are not sysadmin tools, they are development tools.
As such, they don't belong in /usr/sbin, but rather /usr/bin.  Move them
there.