"make allnoconfig" exists to ease testing of minimal configurations.
Documentation/SubmitChecklist includes a note to test with allnoconfig.
This helps catch missing dependencies on common-but-not-required
functionality, which might otherwise go unnoticed.

However, allnoconfig still leaves many symbols enabled, because they're
hidden behind CONFIG_EMBEDDED or CONFIG_EXPERT.  For instance, allnoconfig
still has CONFIG_PRINTK and CONFIG_BLOCK enabled, so drivers don't
typically get build-tested with those disabled.

To address this, introduce a new Kconfig option "allnoconfig_y", used on
symbols which only exist to hide other symbols.  Set it on CONFIG_EMBEDDED
(which then selects CONFIG_EXPERT).  allnoconfig will then disable all the
symbols hidden behind those.

Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Tested-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Michal Marek <mmarek@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

When a symbol changes type from tristate to bool, and was previously set to
'm', a subsequent silentoldconfig would warn about inconsistency, such as:

    include/config/auto.conf:3014:warning: symbol value 'm' invalid for
    HOTPLUG_PCI_PCIE

Seen by Linus with the merge in aa8032b (sequence to reproduce by Michal):
    git checkout 1fe0135
    make mrproper
    make allmodconfig
    make silentoldconfig
    git checkout aa8032b
    make allmodconfig
    make silentoldconfig

Since HOTPLUG_PCI_PCIE changed from tristate to bool in aa8032b, it was
previously set to 'm' in auto.conf by the first allmodconfig+silentoldconfig,
but then was set to 'y' by the second allmodconfig. Then the second
silentoldconfig prints the warning.

The warning in this case is a spurious warning, which happens at the time
kconfig tries to detect symbols that have changed, to touch the empty
header files in include/config used for dependency-tracking by make.

Silence the warning when we read the old auto.conf file, since it is
perfectly legit that a symbol changed type since the previous call.

Thread in:
    http://marc.info/?l=linux-pci&m=137569198904000&w=2

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

This reverts commit 8357b48549e17b3e4e402c7f977b65708922e60f.

It breaks more stuff than it fixes.

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Fengguang Wu <fengguang.wu@intel.com>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Michal Marek <mmarek@suse.cz>

Currently, randconfig does randomise choice entries, unless KCONFIG_ALLCONFIG
is specified.

For example, given those two files (Thomas' test-case):

    ---8<--- Config.test.in
    config OPTIONA
        bool "Option A"

    choice
        prompt "This is a choice"

    config CHOICE_OPTIONA
        bool "Choice Option A"

    config CHOICE_OPTIONB
        bool "Choice Option B"

    endchoice

    config OPTIONB
        bool "Option B"
    ---8<--- Config.test.in

    ---8<--- config.defaults
    CONFIG_OPTIONA=y
    ---8<--- config.defaults

And running:
    ./scripts/kconfig/conf --randconfig Config.test.in

does properly randomise the two choice symbols (and the two booleans).

However, running:
    KCONFIG_ALLCONFIG=config.defaults \
    ./scripts/kconfig/conf --randconfig Config.test.in

does *not* reandomise the two choice entries, and only CHOICE_OPTIONA
will ever be selected. (OPTIONA will always be set (expected), and
OPTIONB will be be properly randomised (expected).)

This patch defers setting that a choice has a value until a symbol for
that choice is indeed set, so that choices are properly randomised when
KCONFIG_ALLCONFIG is set, but not if a symbol for that choice is set.

Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>

---
Changes v3 -> v4
  - fix previous issue where some choices would not be set, which would
    cause silentoldconfig to ask for them and was then breaking this
    workflow (as reported by Arnd and Sedat):
        KCONFIG_ALLCONFIG=foo.defconfig make randconfig
        make silentoldconfig </dev/nullo
    which I have tested (3h28min!) with:
        touch defconfig
        for(( i=0; i<10000; i++ )); do
            KCONFIG_ALLCONFIG=$(pwd)/defconfig make randconfig >/dev/null 2>&1
            make silentoldconfig </dev/null >/dev/null 2>&1 || break
        done
    which did not break at all.
  - change done in v3 (below) is already fixed by a previous patch

Changes v2 -> v3
  - ensure only one symbol is set in a choice

Changes v1 -> v2:
  - further postpone setting that a choice has a value until
    one is indeed set
  - do not print symbols that are part of an invisible choice

Because of choice-in-a-choice constructs, it can happen that not all
symbols are assigned a value during randconfig, leading in rare cases
to this situation:

    ---8<--- choice-in-choice.in
    choice
        bool "A/B/C"
    config A
        bool "A"

    config B
        bool "B"
    if B
    choice
        bool "E/F"
    config E
        bool "E"
    config F
        bool "F"
    endchoice
    endif # B

    config C
        bool "C"
    endchoice
    ---8<---

    $ ./scripts/kconfig/conf --randconfig choice-in-choice.in
    [--SNIP--]
    $ ./scripts/kconfig/conf --silentoldconfig choice-in-choice.in </dev/null
    [--SNIP--]
    A/B/C
      1. A (A)
    > 2. B (B)
      3. C (C)
    choice[1-3]: 2
      E/F
      > 1. E (E) (NEW)
        2. F (F) (NEW)
      choice[1-2]: aborted!

    Console input/output is redirected. Run 'make oldconfig' to update
    configuration.

Fix this by looping in randconfig for as long as some symbol gets assigned
a value.

Note: this was spotted with the USB EHCI Debug Device Gadget (USB_G_DBGP),
which uses this choice-in-a-choice construct, and exhibits this problem.
The example above is just a stripped-down minimalist test-case.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

Currently, randconfig may set more than one symbol in a given choice.
Given this config file:
    config A
        bool "A"
    if A
    choice
        bool "B/C/D"
    config B
        bool "B"
    config C
        bool "C"
    config D
        bool "D"
    endchoice
    endif # A

Then randconfig generates such .config files (case where A is not set is not
shown below for brevity), and where only the right-most .config is valid:
  CONFIG_A=y                  CONFIG_A=y                  CONFIG_A=y
  CONFIG_B=y                  CONFIG_B=y                  CONFIG_B=y
  CONFIG_C=y                  # CONFIG_C is not set       # CONFIG_C is not set
  # CONFIG_D is not set       CONFIG_D=y                  # CONFIG_D is not set

That is, in a randomised choice, the first symbol is always selected,
and at most one other symbol may be selected.

This is due to symbol randomised in a choice not being properly flagged
as having a value.

Fix that by flagging those symbols adequately: have a user-defined value,
and be not valid (to force recalculation of the symbol).

Note: if the choice is not conditional, then the randomisation is properly
done.

Reported-by: Matthieu CASTET <matthieu.castet@parrot.com>
Signed-off-by: Matthieu CASTET <matthieu.castet@parrot.com>
[yann.morin.1998@free.fr: independently re-done the same patch as Matthieu,
                          as pointed out by Sedat]
Cc: Arnaud Lacombe <lacombar@gmail.com>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

The defconfig and Kconfig combination below, which is based on 3.10-rc4
Kconfigs, resulted in several options getting set to "m" instead of "y".

defconfig.choice:
---8<---
CONFIG_MODULES=y
CONFIG_USB_ZERO=y
---8<---

Kconfig.choice:
---8<---
menuconfig MODULES
	bool "Enable loadable module support"

config CONFIGFS_FS
	tristate "Userspace-driven configuration filesystem"

config OCFS2_FS
        tristate "OCFS2 file system support"
        depends on CONFIGFS_FS
        select CRC32

config USB_LIBCOMPOSITE
	tristate
	select CONFIGFS_FS

choice
	tristate "USB Gadget Drivers"
	default USB_ETH

config USB_ZERO
	tristate "Gadget Zero (DEVELOPMENT)"
	select USB_LIBCOMPOSITE

config USB_ETH
	tristate "Ethernet Gadget (with CDC Ethernet support)"
	select USB_LIBCOMPOSITE

endchoice

config CRC32
        tristate "CRC32/CRC32c functions"
        default y

choice
        prompt "CRC32 implementation"
        depends on CRC32
        default CRC32_SLICEBY8

config CRC32_SLICEBY8
        bool "Slice by 8 bytes"

endchoice
---8<---

$ scripts/kconfig/conf --defconfig=defconfig.choice Kconfig.choice

would result in:

.config:
---8<---
CONFIG_MODULES=y
CONFIG_CONFIGFS_FS=m
CONFIG_USB_LIBCOMPOSITE=m
CONFIG_USB_ZERO=m
CONFIG_CRC32=y
CONFIG_CRC32_SLICEBY8=y
---8<---

when the expected result would be:

.config:
---8<---
CONFIG_MODULES=y
CONFIG_CONFIGFS_FS=y
CONFIG_USB_LIBCOMPOSITE=y
CONFIG_USB_ZERO=y
CONFIG_CRC32=y
CONFIG_CRC32_SLICEBY8=y
---8<---

Signed-off-by: Arve Hjønnevåg <arve@android.com>
[yann.morin.1998@free.fr: add the resulting .config to commit log,
                          remove unneeded USB_GADGET from the defconfig]
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>

This reverts commit 422c809f03f043d0950d8362214818e956a9daee.
It causes more harm than it solves issues.

Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Greg KH <greg@kroah.com>

Currently the odds to set each symbol is (rounded):
    booleans:   y: 50%          n: 50%
    tristates:  y: 33%  m: 33%  n: 33%

Introduce a KCONFIG_PROBABILITY environment variable to tweak the
probabilities (in percentage), as such:
    KCONFIG_PROBABILITY     y:n split           y:m:n split
    -----------------------------------------------------------------
[1] unset or empty          50  : 50            33  : 33  : 34
[2] N                        N  : 100-N         N/2 : N/2 : 100-N
    N:M                     N+M : 100-(N+M)      N  :  M  : 100-(N+M)
    N:M:L                    N  : 100-N          M  :  L  : 100-(M+L)

[1] The current behaviour is kept as default, for backward compatibility
[2] The solution initially implemented by Peter for Buildroot, see:
    http://git.buildroot.org/buildroot/commit/?id=3435c1afb5

Signed-off-by: Peter Korsgaard <jacmet@uclibc.org>
[yann.morin.1998@free.fr: add to Documentation/]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

Currently, randconfig does randomise choice entries, unless KCONFIG_ALLCONFIG
is specified.

For example, given those two files (Thomas' test-case):

    ---8<--- Config.test.in
    config OPTIONA
        bool "Option A"

    choice
        prompt "This is a choice"

    config CHOICE_OPTIONA
        bool "Choice Option A"

    config CHOICE_OPTIONB
        bool "Choice Option B"

    endchoice

    config OPTIONB
        bool "Option B"
    ---8<--- Config.test.in

    ---8<--- config.defaults
    CONFIG_OPTIONA=y
    ---8<--- config.defaults

And running:
    ./scripts/kconfig/conf --randconfig Config.test.in

does properly randomise the two choice symbols (and the two booleans).

However, running:
    KCONFIG_ALLCONFIG=config.defaults \
    ./scripts/kconfig/conf --randconfig Config.test.in

does *not* reandomise the two choice entries, and only CHOICE_OPTIONA
will ever be selected. (OPTIONA will always be set (expected), and
OPTIONB will be be properly randomised (expected).)

This patch defers setting that a choice has a value until a symbol for
that choice is indeed set, so that choices are properly randomised when
KCONFIG_ALLCONFIG is set, but not if a symbol for that choice is set.

Also, as a side-efect, this patch fixes the following case:

    ---8<---
    choice
    config OPTION_A
        bool "Option A"
    config OPTION_B
        bool "Option B"
    config OPTION_C
        bool "Option C"
    endchoice
    ---8<---

which could previously generate such .config files:

    ---8<---                            ---8<---
    CONFIG_OPTION_A=y                   CONFIG_OPTION_A=y
    CONFIG_OPTION_B=y                   # CONFIG_OPTION_B is not set
    # CONFIG_OPTION_C is not set        CONFIG_OPTION_C=y
    ---8<---                            ---8<---

Ie., the first entry in a choice is always set, plus zero or one of
the other options may be set.

This patch ensures that only one option may be set for a choice.

Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Arnaud Lacombe <lacombar@gmail.com>

---
Changes v2 -> v3
  - ensure only one symbol is set in a choice

Changes v1 -> v2:
  - further postpone setting that a choice has a value until
    one is indeed set
  - do not print symbols that are part of an invisible choice