6:54 a.m.
From: Bastien Nocera <hadess(a)hadess.net>
As noted in https://www.kernel.org/doc/Documentation/rfkill.txt
rfkill-input is deprecated, and rfkill keys are handled in user-space.
---
config-generic | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/config-generic b/config-generic
index 055d5d2..cbe4571 100644
--- a/config-generic
+++ b/config-generic
@@ -1381,7 +1381,7 @@ CONFIG_LWTUNNEL=y
CONFIG_RFKILL=m
CONFIG_RFKILL_GPIO=m
-CONFIG_RFKILL_INPUT=y
+CONFIG_RFKILL_INPUT=n
CONFIG_ETHERNET=y
--
2.5.0
1:21 p.m.
On Sun, Jan 10, 2016 at 7:54 AM, Bastien Nocera <bnocera(a)redhat.com> wrote:
From: Bastien Nocera <hadess(a)hadess.net>
As noted in https://www.kernel.org/doc/Documentation/rfkill.txt
rfkill-input is deprecated, and rfkill keys are handled in user-space.
That may be what the document says, but the Kconfig at least seems to
imply something else.
---
config-generic | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/config-generic b/config-generic
index 055d5d2..cbe4571 100644
--- a/config-generic
+++ b/config-generic
@@ -1381,7 +1381,7 @@ CONFIG_LWTUNNEL=y
CONFIG_RFKILL=m
CONFIG_RFKILL_GPIO=m
-CONFIG_RFKILL_INPUT=y
+CONFIG_RFKILL_INPUT=n
There's a couple things wrong here. First, it should be:
# CONFIG_RFKILL_INPUT is not set
I can't remember if Kconfig accepts =n, but I don't think it does for
whatever reason.
The second thing wrong here is that you can set this in
config-generic, but it doesn't matter because the Kconfig section for
this is:
config RFKILL_INPUT
bool "RF switch input support" if EXPERT
depends on RFKILL
depends on INPUT = y || RFKILL = INPUT
default y if !EXPERT
which basically means that unless we have CONFIG_EXPERT set, you can't
override it from the default. We aren't going to set CONFIG_EXPERT
because it brings in a whole bunch of other random options we don't
want to deal with.
Perhaps you could email upstream and see if they can remove that
"default y if !EXPERT" ? Also, did you actually test a typical Fedora
setup with your change in place?
josh
2:18 a.m.
Hi,
On 11-01-16 20:21, Josh Boyer wrote:
On Sun, Jan 10, 2016 at 7:54 AM, Bastien Nocera
<bnocera(a)redhat.com> wrote:
> From: Bastien Nocera <hadess(a)hadess.net>
>
> As noted in https://www.kernel.org/doc/Documentation/rfkill.txt
> rfkill-input is deprecated, and rfkill keys are handled in user-space.
That may be what the document says, but the Kconfig at least seems to
imply something else.
> ---
> config-generic | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/config-generic b/config-generic
> index 055d5d2..cbe4571 100644
> --- a/config-generic
> +++ b/config-generic
> @@ -1381,7 +1381,7 @@ CONFIG_LWTUNNEL=y
>
> CONFIG_RFKILL=m
> CONFIG_RFKILL_GPIO=m
> -CONFIG_RFKILL_INPUT=y
> +CONFIG_RFKILL_INPUT=n
There's a couple things wrong here. First, it should be:
# CONFIG_RFKILL_INPUT is not set
I can't remember if Kconfig accepts =n, but I don't think it does for
whatever reason.
The second thing wrong here is that you can set this in
config-generic, but it doesn't matter because the Kconfig section for
this is:
config RFKILL_INPUT
bool "RF switch input support" if EXPERT
depends on RFKILL
depends on INPUT = y || RFKILL = INPUT
default y if !EXPERT
which basically means that unless we have CONFIG_EXPERT set, you can't
override it from the default.
AFAIK Kconfig, that is not true, adding
# CONFIG_RFKILL_INPUT is not set
Should still override the default, a default is just that a default, not
a hard select. Making a default conditional on EXPERT is a bit weird,
it just means that if EXPERT is set, there is no default.
Regards,
Hans
5:21 a.m.
On Tue, Jan 12, 2016 at 3:18 AM, Hans de Goede <hdegoede(a)redhat.com> wrote:
Hi,
On 11-01-16 20:21, Josh Boyer wrote:
>
> On Sun, Jan 10, 2016 at 7:54 AM, Bastien Nocera <bnocera(a)redhat.com>
> wrote:
>>
>> From: Bastien Nocera <hadess(a)hadess.net>
>>
>> As noted in https://www.kernel.org/doc/Documentation/rfkill.txt
>> rfkill-input is deprecated, and rfkill keys are handled in user-space.
>
>
> That may be what the document says, but the Kconfig at least seems to
> imply something else.
>
>> ---
>> config-generic | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/config-generic b/config-generic
>> index 055d5d2..cbe4571 100644
>> --- a/config-generic
>> +++ b/config-generic
>> @@ -1381,7 +1381,7 @@ CONFIG_LWTUNNEL=y
>>
>> CONFIG_RFKILL=m
>> CONFIG_RFKILL_GPIO=m
>> -CONFIG_RFKILL_INPUT=y
>> +CONFIG_RFKILL_INPUT=n
>
>
> There's a couple things wrong here. First, it should be:
>
> # CONFIG_RFKILL_INPUT is not set
>
> I can't remember if Kconfig accepts =n, but I don't think it does for
> whatever reason.
>
> The second thing wrong here is that you can set this in
> config-generic, but it doesn't matter because the Kconfig section for
> this is:
>
> config RFKILL_INPUT
> bool "RF switch input support" if EXPERT
> depends on RFKILL
> depends on INPUT = y || RFKILL = INPUT
> default y if !EXPERT
>
> which basically means that unless we have CONFIG_EXPERT set, you can't
> override it from the default.
AFAIK Kconfig, that is not true, adding
# CONFIG_RFKILL_INPUT is not set
Should still override the default, a default is just that a default, not
a hard select. Making a default conditional on EXPERT is a bit weird,
it just means that if EXPERT is set, there is no default.
I tried exactly that. It is still set. So either I'm correct, or
there is something that is doing a select on it. I don't care if I'm
correct or not, but what I keep being told is not working and I cannot
find anywhere that selects that option.
[jwboyer@vader kernel]$ grep CONFIG_RFKILL_INPUT config-*
config-generic:# CONFIG_RFKILL_INPUT is not set
[jwboyer@vader kernel]$ fedpkg prep
<snip>
[jwboyer@vader kernel]$ grep CONFIG_RFKILL_INPUT
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-*
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-aarch64.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-armv7hl.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-armv7hl-lpae.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686-debug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686-PAE.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686-PAEdebug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64-debug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64le.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64le-debug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64p7.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-s390x.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-x86_64.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-x86_64-debug.config:CONFIG_RFKILL_INPUT=y
[jwboyer@vader kernel]$
Seriously, please stop guessing at things and test changes before
sending them or telling others they work.
josh
6:22 a.m.
----- Original Message -----
<snip>
> AFAIK Kconfig, that is not true, adding
>
> # CONFIG_RFKILL_INPUT is not set
>
> Should still override the default, a default is just that a default, not
> a hard select. Making a default conditional on EXPERT is a bit weird,
> it just means that if EXPERT is set, there is no default.
I tried exactly that. It is still set. So either I'm correct, or
there is something that is doing a select on it. I don't care if I'm
correct or not, but what I keep being told is not working and I cannot
find anywhere that selects that option.
[jwboyer@vader kernel]$ grep CONFIG_RFKILL_INPUT config-*
config-generic:# CONFIG_RFKILL_INPUT is not set
[jwboyer@vader kernel]$ fedpkg prep
<snip>
[jwboyer@vader kernel]$ grep CONFIG_RFKILL_INPUT
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-*
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-aarch64.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-armv7hl.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-armv7hl-lpae.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686-debug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686-PAE.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-i686-PAEdebug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64-debug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64le.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64le-debug.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-ppc64p7.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-s390x.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-x86_64.config:CONFIG_RFKILL_INPUT=y
kernel-4.4.fc24/linux-4.4.0-1.fc24.x86_64/configs/kernel-4.4.0-x86_64-debug.config:CONFIG_RFKILL_INPUT=y
[jwboyer@vader kernel]$
Seriously, please stop guessing at things and test changes before
sending them or telling others they work.
I hope this isn't directed at me, but whatever.
I've sent a mail to linux-wireless asking why it wasn't possible to disable it
without
enabling CONFIG_EXPERT:
https://marc.info/?l=linux-wireless&m=145260077927073&w=2
Cheers
3025
days inactive
3027
days old
kernel@lists.fedoraproject.org
4 comments
3 participants
participants (3)
-
Bastien Nocera -
Hans de Goede -
Josh Boyer