Re: [PATCH] wireless: update regulatory sample code

Re: [PATCH] wireless: update regulatory sample code

[Luis R. Rodriguez]

On Wed, Oct 22, 2008 at 11:55:52AM -0700, Johannes Berg wrote:
> Update the same code for the recent changes. I did this as a
> separate patch to avoid removing a lot of sample code only
> to add it back again in the next patch.
> 
> Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
> ---
>  Documentation/networking/regulatory.txt |   19 ++++++++-----------
>  1 file changed, 8 insertions(+), 11 deletions(-)
> 
> --- everything.orig/Documentation/networking/regulatory.txt     2008-10-22 20:51:50.000000000 +0200
> +++ everything/Documentation/networking/regulatory.txt  2008-10-22 20:54:12.000000000 +0200
> @@ -131,7 +131,7 @@ are expected to do this during initializ
> 
>         r = zd_reg2alpha2(mac->regdomain, alpha2);
>         if (!r)
> -               regulatory_hint(hw->wiphy, alpha2, NULL);
> +               regulatory_hint(alpha2);
> 
>  Example code - drivers providing a built in regulatory domain:

Should we mention how currently this should be avoided unless
your regulatory definition is also intersected with hw capabilities
and that the same channels are the ones that are registered to cfg80211?

  Luis

Re: [PATCH] wireless: update regulatory sample code

[Luis R. Rodriguez]

On Wed, Oct 22, 2008 at 12:41:28PM -0700, Johannes Berg wrote:
> On Wed, 2008-10-22 at 05:35 -0700, Luis R. Rodriguez wrote:
> > On Wed, Oct 22, 2008 at 11:55:52AM -0700, Johannes Berg wrote:
> > > Update the same code for the recent changes. I did this as a
> > > separate patch to avoid removing a lot of sample code only
> > > to add it back again in the next patch.
> > > 
> > > Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
> > > ---
> > >  Documentation/networking/regulatory.txt |   19 ++++++++-----------
> > >  1 file changed, 8 insertions(+), 11 deletions(-)
> > > 
> > > --- everything.orig/Documentation/networking/regulatory.txt     2008-10-22 20:51:50.000000000 +0200
> > > +++ everything/Documentation/networking/regulatory.txt  2008-10-22 20:54:12.000000000 +0200
> > > @@ -131,7 +131,7 @@ are expected to do this during initializ
> > > 
> > >         r = zd_reg2alpha2(mac->regdomain, alpha2);
> > >         if (!r)
> > > -               regulatory_hint(hw->wiphy, alpha2, NULL);
> > > +               regulatory_hint(alpha2);
> > > 
> > >  Example code - drivers providing a built in regulatory domain:
> > 
> > Should we mention how currently this should be avoided unless
> > your regulatory definition is also intersected with hw capabilities
> > and that the same channels are the ones that are registered to cfg80211?

Well rd hint for when we provide a structure, that code right now
disregards when initiator == USER, and this works well only for
when the driver sending the rd struct hint also only registered
to cfg80211 the channels its legally capable of. If a driver
which does not use reg_notifier() and registers *all* of its channels
uses this then the USER set regdom will be used and *all* channels
will be left intact.

  Luis

[PATCH] wireless: update regulatory sample code

[Johannes Berg]

Update the same code for the recent changes. I did this as a
separate patch to avoid removing a lot of sample code only
to add it back again in the next patch.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
---
 Documentation/networking/regulatory.txt |   19 ++++++++-----------
 1 file changed, 8 insertions(+), 11 deletions(-)

--- everything.orig/Documentation/networking/regulatory.txt	2008-10-22 20:51:50.000000000 +0200
+++ everything/Documentation/networking/regulatory.txt	2008-10-22 20:54:12.000000000 +0200
@@ -131,7 +131,7 @@ are expected to do this during initializ
 
 	r = zd_reg2alpha2(mac->regdomain, alpha2);
 	if (!r)
-		regulatory_hint(hw->wiphy, alpha2, NULL);
+		regulatory_hint(alpha2);
 
 Example code - drivers providing a built in regulatory domain:
 --------------------------------------------------------------
@@ -178,17 +178,14 @@ Then in some part of your code after you
 
 	rd = kzalloc(size_of_regd, GFP_KERNEL);
 	if (!rd)
-	return -ENOMEM;
+		return -ENOMEM;
 
 	memcpy(rd, &mydriver_jp_regdom, sizeof(struct ieee80211_regdomain));
 
-	for (i=0; i < num_rules; i++) {
-		memcpy(&rd->reg_rules[i], &mydriver_jp_regdom.reg_rules[i],
-			sizeof(struct ieee80211_reg_rule));
-	}
-	r = regulatory_hint(hw->wiphy, NULL, rd);
-	if (r) {
-		kfree(rd);
-		return r;
-	}
+	for (i=0; i < num_rules; i++)
+		memcpy(&rd->reg_rules[i],
+		       &mydriver_jp_regdom.reg_rules[i],
+		       sizeof(struct ieee80211_reg_rule));
 
+	regulatory_struct_hint(rd, BIT(IEEE80211_BAND_2GHZ) |
+				   BIT(IEEE80211_BAND_5GHZ));

Re: [PATCH] wireless: update regulatory sample code

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 1281 bytes --]

On Wed, 2008-10-22 at 05:35 -0700, Luis R. Rodriguez wrote:
> On Wed, Oct 22, 2008 at 11:55:52AM -0700, Johannes Berg wrote:
> > Update the same code for the recent changes. I did this as a
> > separate patch to avoid removing a lot of sample code only
> > to add it back again in the next patch.
> > 
> > Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
> > ---
> >  Documentation/networking/regulatory.txt |   19 ++++++++-----------
> >  1 file changed, 8 insertions(+), 11 deletions(-)
> > 
> > --- everything.orig/Documentation/networking/regulatory.txt     2008-10-22 20:51:50.000000000 +0200
> > +++ everything/Documentation/networking/regulatory.txt  2008-10-22 20:54:12.000000000 +0200
> > @@ -131,7 +131,7 @@ are expected to do this during initializ
> > 
> >         r = zd_reg2alpha2(mac->regdomain, alpha2);
> >         if (!r)
> > -               regulatory_hint(hw->wiphy, alpha2, NULL);
> > +               regulatory_hint(alpha2);
> > 
> >  Example code - drivers providing a built in regulatory domain:
> 
> Should we mention how currently this should be avoided unless
> your regulatory definition is also intersected with hw capabilities
> and that the same channels are the ones that are registered to cfg80211?

?

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] wireless: update regulatory sample code

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 995 bytes --]

On Wed, 2008-10-22 at 07:40 -0700, Luis R. Rodriguez wrote:

> > > Should we mention how currently this should be avoided unless
> > > your regulatory definition is also intersected with hw capabilities
> > > and that the same channels are the ones that are registered to cfg80211?
> 
> Well rd hint for when we provide a structure, that code right now
> disregards when initiator == USER, and this works well only for
> when the driver sending the rd struct hint also only registered
> to cfg80211 the channels its legally capable of. If a driver
> which does not use reg_notifier() and registers *all* of its channels
> uses this then the USER set regdom will be used and *all* channels
> will be left intact.

No, only those that are available from CRDA will be used. And this is
totally irrelevant in the case where this applies anyway because the
Intel drivers register only the channels from their eeprom, the same set
they hint at. Yet it's still unrelated.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]