wireless extensions should default to Y
diff mbox

Message ID 20150107101903.GA6563@amd
State Rejected
Headers show

Commit Message

Pavel Machek Jan. 7, 2015, 10:19 a.m. UTC
Do we need following patch on top of
24a0aa212ee2dbe44360288684478d76a8e20a0a revert?

I updated kernel today, and (probably because extensions were not
selectable before), the default choice was "N", which is wrong:
oldconfig should result in compatible choices being made, for example
to help bisect.

Signed-off-by: Pavel Machek <pavel@ucw.cz>

Comments

Johannes Berg Jan. 7, 2015, 10:41 a.m. UTC | #1
On Wed, 2015-01-07 at 11:19 +0100, Pavel Machek wrote:
> Do we need following patch on top of
> 24a0aa212ee2dbe44360288684478d76a8e20a0a revert?
> 
> I updated kernel today, and (probably because extensions were not
> selectable before), the default choice was "N", which is wrong:
> oldconfig should result in compatible choices being made, for example
> to help bisect.

I don't believe we need this. It has been defaulting to N for a long
time, it's just that it got thrown out of your config due to building
inbetween the patch and its revert. Had you built only before and after
that wouldn't be an issue.

If we default to Y now we send the wrong signal. If you really need to
bisect something wext related you have far bigger issues I'd think, the
code hasn't changed *forever*.

johannes

--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Pavel Machek Jan. 7, 2015, 11:03 a.m. UTC | #2
On Wed 2015-01-07 11:41:55, Johannes Berg wrote:
> On Wed, 2015-01-07 at 11:19 +0100, Pavel Machek wrote:
> > Do we need following patch on top of
> > 24a0aa212ee2dbe44360288684478d76a8e20a0a revert?
> > 
> > I updated kernel today, and (probably because extensions were not
> > selectable before), the default choice was "N", which is wrong:
> > oldconfig should result in compatible choices being made, for example
> > to help bisect.
> 
> I don't believe we need this. It has been defaulting to N for a long
> time, it's just that it got thrown out of your config due to building
> inbetween the patch and its revert. Had you built only before and after
> that wouldn't be an issue.

Well, I clearly hit the issue. If someone had _not_ build in between,
the "default y" does not change anything for him, as he will not be
asked thequestion. If someone starts config from scratch, Y is the
safe answer.

> If we default to Y now we send the wrong signal. If you really need to
> bisect something wext related you have far bigger issues I'd think, the
> code hasn't changed *forever*.

I am woried about bisecting something unrelated, and then wext coming
and breaking the bisect. But you are right that it will break the
bisect, anyway...

                                                                        Pavel
Johannes Berg Jan. 7, 2015, 11:06 a.m. UTC | #3
On Wed, 2015-01-07 at 12:03 +0100, Pavel Machek wrote:
> On Wed 2015-01-07 11:41:55, Johannes Berg wrote:
> > On Wed, 2015-01-07 at 11:19 +0100, Pavel Machek wrote:
> > > Do we need following patch on top of
> > > 24a0aa212ee2dbe44360288684478d76a8e20a0a revert?
> > > 
> > > I updated kernel today, and (probably because extensions were not
> > > selectable before), the default choice was "N", which is wrong:
> > > oldconfig should result in compatible choices being made, for example
> > > to help bisect.
> > 
> > I don't believe we need this. It has been defaulting to N for a long
> > time, it's just that it got thrown out of your config due to building
> > inbetween the patch and its revert. Had you built only before and after
> > that wouldn't be an issue.
> 
> Well, I clearly hit the issue. If someone had _not_ build in between,
> the "default y" does not change anything for him, as he will not be
> asked thequestion. If someone starts config from scratch, Y is the
> safe answer.

We've long stated that if really starting from scratch it's the wrong
thing to do and had a default N, so I don't really know why we'd revert
that only to make a short period of bisect slightly happier?

I really don't want to merge this patch and encourage more people to
build with wext enabled.

johannes

--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch
diff mbox

diff --git a/net/wireless/Kconfig b/net/wireless/Kconfig
index 29c8675..8951dba 100644
--- a/net/wireless/Kconfig
+++ b/net/wireless/Kconfig
@@ -177,6 +177,7 @@  config CFG80211_INTERNAL_REGDB
 config CFG80211_WEXT
 	bool "cfg80211 wireless extensions compatibility"
 	depends on CFG80211
+	default y
 	select WEXT_CORE
 	help
 	  Enable this option if you need old userspace for wireless