[PATCH 1/2] mac80211: do not wip out old supported rates

[PATCH 1/2] mac80211: do not wip out old supported rates

[Stanislaw Gruszka]

Use old supported rates, if AP do not provide supported rates
information element in a new managment frame.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
---
 net/mac80211/scan.c |   21 +++++++++++----------
 1 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/net/mac80211/scan.c b/net/mac80211/scan.c
index 85507bd..151d933 100644
--- a/net/mac80211/scan.c
+++ b/net/mac80211/scan.c
@@ -83,7 +83,7 @@ ieee80211_bss_info_update(struct ieee80211_local *local,
 {
 	struct cfg80211_bss *cbss;
 	struct ieee80211_bss *bss;
-	int clen;
+	int clen, srlen;
 	s32 signal = 0;
 
 	if (local->hw.flags & IEEE80211_HW_SIGNAL_DBM)
@@ -112,23 +112,24 @@ ieee80211_bss_info_update(struct ieee80211_local *local,
 		bss->dtim_period = tim_ie->dtim_period;
 	}
 
-	bss->supp_rates_len = 0;
+	/* replace old supported rates if we get new values */
+	srlen = 0;
 	if (elems->supp_rates) {
-		clen = IEEE80211_MAX_SUPP_RATES - bss->supp_rates_len;
+		clen = IEEE80211_MAX_SUPP_RATES;
 		if (clen > elems->supp_rates_len)
 			clen = elems->supp_rates_len;
-		memcpy(&bss->supp_rates[bss->supp_rates_len], elems->supp_rates,
-		       clen);
-		bss->supp_rates_len += clen;
+		memcpy(bss->supp_rates, elems->supp_rates, clen);
+		srlen += clen;
 	}
 	if (elems->ext_supp_rates) {
-		clen = IEEE80211_MAX_SUPP_RATES - bss->supp_rates_len;
+		clen = IEEE80211_MAX_SUPP_RATES - srlen;
 		if (clen > elems->ext_supp_rates_len)
 			clen = elems->ext_supp_rates_len;
-		memcpy(&bss->supp_rates[bss->supp_rates_len],
-		       elems->ext_supp_rates, clen);
-		bss->supp_rates_len += clen;
+		memcpy(bss->supp_rates + srlen, elems->ext_supp_rates, clen);
+		srlen += clen;
 	}
+	if (srlen)
+		bss->supp_rates_len = srlen;
 
 	bss->wmm_used = elems->wmm_param || elems->wmm_info;
 	bss->uapsd_supported = is_uapsd_supported(elems);
-- 
1.6.2.5

[PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Stanislaw Gruszka]

If AP do not provide us supported rates before assiociation, send
all rates we are supporting instead of empty information element.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
---
 net/mac80211/work.c |   23 ++++++++++++++---------
 1 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/net/mac80211/work.c b/net/mac80211/work.c
index 15e1ba9..f8f5c9a 100644
--- a/net/mac80211/work.c
+++ b/net/mac80211/work.c
@@ -213,15 +213,20 @@ static void ieee80211_send_assoc(struct ieee80211_sub_if_data *sdata,
 
 	sband = local->hw.wiphy->bands[wk->chan->band];
 
-	/*
-	 * Get all rates supported by the device and the AP as
-	 * some APs don't like getting a superset of their rates
-	 * in the association request (e.g. D-Link DAP 1353 in
-	 * b-only mode)...
-	 */
-	rates_len = ieee80211_compatible_rates(wk->assoc.supp_rates,
-					       wk->assoc.supp_rates_len,
-					       sband, &rates);
+	if (wk->assoc.supp_rates_len) {
+		/*
+		 * Get all rates supported by the device and the AP as
+		 * some APs don't like getting a superset of their rates
+		 * in the association request (e.g. D-Link DAP 1353 in
+		 * b-only mode)...
+		 */
+		rates_len = ieee80211_compatible_rates(wk->assoc.supp_rates,
+						       wk->assoc.supp_rates_len,
+						       sband, &rates);
+	} else {
+		rates = ~0;
+		rates_len = sband->n_bitrates;
+	}
 
 	skb = alloc_skb(local->hw.extra_tx_headroom +
 			sizeof(*mgmt) + /* bit too much but doesn't matter */
-- 
1.6.2.5

Re: [PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Stanislaw Gruszka]

On Wed, Apr 28, 2010 at 03:17:04PM +0200, Stanislaw Gruszka wrote:
> If AP do not provide us supported rates before assiociation, send
> all rates we are supporting instead of empty information element.

Please note these two patches together with "QoS fixes":
- mac80211: explicitly disable/enable QoS
- iwlwifi: manage QoS by mac stack
resolve:
https://bugzilla.redhat.com/show_bug.cgi?id=558002

"QoS fixes" alone resolve:
https://bugzilla.redhat.com/show_bug.cgi?id=539878

So I think we want all 4 patches in stable.

John,if possible push patches to Linus tree. When they lend there,
I will post backported patches to stable ML. I have them already
prepared ant tested for 2.6.32 :)

Thanks
Stanislaw

Re: [PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Johannes Berg]

On Wed, 2010-04-28 at 15:17 +0200, Stanislaw Gruszka wrote:
> If AP do not provide us supported rates before assiociation, send
> all rates we are supporting instead of empty information element.
> 
> Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
> ---
>  net/mac80211/work.c |   23 ++++++++++++++---------
>  1 files changed, 14 insertions(+), 9 deletions(-)
> 
> diff --git a/net/mac80211/work.c b/net/mac80211/work.c
> index 15e1ba9..f8f5c9a 100644
> --- a/net/mac80211/work.c
> +++ b/net/mac80211/work.c
> @@ -213,15 +213,20 @@ static void ieee80211_send_assoc(struct ieee80211_sub_if_data *sdata,
>  
>  	sband = local->hw.wiphy->bands[wk->chan->band];
>  
> -	/*
> -	 * Get all rates supported by the device and the AP as
> -	 * some APs don't like getting a superset of their rates
> -	 * in the association request (e.g. D-Link DAP 1353 in
> -	 * b-only mode)...
> -	 */
> -	rates_len = ieee80211_compatible_rates(wk->assoc.supp_rates,
> -					       wk->assoc.supp_rates_len,
> -					       sband, &rates);
> +	if (wk->assoc.supp_rates_len) {
> +		/*
> +		 * Get all rates supported by the device and the AP as
> +		 * some APs don't like getting a superset of their rates
> +		 * in the association request (e.g. D-Link DAP 1353 in
> +		 * b-only mode)...
> +		 */
> +		rates_len = ieee80211_compatible_rates(wk->assoc.supp_rates,
> +						       wk->assoc.supp_rates_len,
> +						       sband, &rates);
> +	} else {
> +		rates = ~0;
> +		rates_len = sband->n_bitrates;
> +	}

Could you add a comment about what AP this is, and that it's a
workaround please? The AP is surely supposed to send this information!

johannes

Re: [PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Stanislaw Gruszka]

On Wed, 28 Apr 2010 15:39:24 +0200
Johannes Berg <johannes@sipsolutions.net> wrote:

> Could you add a comment about what AP this is

Bug reporter don't know what AP it is. He fail to connect
to wireless network at school.
https://bugzilla.redhat.com/show_bug.cgi?id=558002

>, and that it's a
> workaround please?

I don't think this is workaround. For me this is right behavior
and fix for workaround introduced in commit
36d16ae73becc5978fe22866e9ab66b509211afe "mac80211: fix association
with some APs"

> The AP is surely supposed to send this information!

The same as we are, no matter what :-)

Cheers
Stanislaw

Re: [PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Johannes Berg]

On Wed, 2010-04-28 at 15:57 +0200, Stanislaw Gruszka wrote:
> On Wed, 28 Apr 2010 15:39:24 +0200
> Johannes Berg <johannes@sipsolutions.net> wrote:
> 
> > Could you add a comment about what AP this is
> 
> Bug reporter don't know what AP it is. He fail to connect
> to wireless network at school.
> https://bugzilla.redhat.com/show_bug.cgi?id=558002

Ah bugger.

> >, and that it's a
> > workaround please?
> 
> I don't think this is workaround. For me this is right behavior
> and fix for workaround introduced in commit
> 36d16ae73becc5978fe22866e9ab66b509211afe "mac80211: fix association
> with some APs"

Oh yeah so it's a workaround for a workaround. Oh well. Can we still
have a comment about it in the else branch?

johannes

[PATCH 2/2 v2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Stanislaw Gruszka]

If AP do not provide us supported rates before assiociation, send
all rates we are supporting instead of empty information element.

v1 -> v2: Add comment.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
---
 net/mac80211/work.c |   28 +++++++++++++++++++---------
 1 files changed, 19 insertions(+), 9 deletions(-)

diff --git a/net/mac80211/work.c b/net/mac80211/work.c
index 15e1ba9..949c2d1 100644
--- a/net/mac80211/work.c
+++ b/net/mac80211/work.c
@@ -213,15 +213,25 @@ static void ieee80211_send_assoc(struct ieee80211_sub_if_data *sdata,
 
 	sband = local->hw.wiphy->bands[wk->chan->band];
 
-	/*
-	 * Get all rates supported by the device and the AP as
-	 * some APs don't like getting a superset of their rates
-	 * in the association request (e.g. D-Link DAP 1353 in
-	 * b-only mode)...
-	 */
-	rates_len = ieee80211_compatible_rates(wk->assoc.supp_rates,
-					       wk->assoc.supp_rates_len,
-					       sband, &rates);
+	if (wk->assoc.supp_rates_len) {
+		/*
+		 * Get all rates supported by the device and the AP as
+		 * some APs don't like getting a superset of their rates
+		 * in the association request (e.g. D-Link DAP 1353 in
+		 * b-only mode)...
+		 */
+		rates_len = ieee80211_compatible_rates(wk->assoc.supp_rates,
+						       wk->assoc.supp_rates_len,
+						       sband, &rates);
+	} else {
+		/*
+		 * In case AP not provide any supported rates information
+		 * before association, we send information element(s) with
+		 * all rates that we support.
+		 */
+		rates = ~0;
+		rates_len = sband->n_bitrates;
+	}
 
 	skb = alloc_skb(local->hw.extra_tx_headroom +
 			sizeof(*mgmt) + /* bit too much but doesn't matter */
-- 
1.6.2.5

Re: [PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[John W. Linville]

On Wed, Apr 28, 2010 at 03:33:37PM +0200, Stanislaw Gruszka wrote:
> On Wed, Apr 28, 2010 at 03:17:04PM +0200, Stanislaw Gruszka wrote:
> > If AP do not provide us supported rates before assiociation, send
> > all rates we are supporting instead of empty information element.
> 
> Please note these two patches together with "QoS fixes":
> - mac80211: explicitly disable/enable QoS
> - iwlwifi: manage QoS by mac stack
> resolve:
> https://bugzilla.redhat.com/show_bug.cgi?id=558002
> 
> "QoS fixes" alone resolve:
> https://bugzilla.redhat.com/show_bug.cgi?id=539878
> 
> So I think we want all 4 patches in stable.
> 
> John,if possible push patches to Linus tree. When they lend there,
> I will post backported patches to stable ML. I have them already
> prepared ant tested for 2.6.32 :)

It is fairly late in the 2.6.34 cycle, and it isn't obvious to me
that these fix any sort of regression.  Am I wrong?

John
-- 
John W. Linville		Someday the world will need a hero, and you
linville@tuxdriver.com			might be all we have.  Be ready.

Re: [PATCH 2/2] mac80211: fix supported rates IE if AP doesn't give us it's rates

[Stanislaw Gruszka]

On Wed, 28 Apr 2010 16:33:04 -0400
"John W. Linville" <linville@tuxdriver.com> wrote:

> > Please note these two patches together with "QoS fixes":
> > - mac80211: explicitly disable/enable QoS
> > - iwlwifi: manage QoS by mac stack
> > resolve:
> > https://bugzilla.redhat.com/show_bug.cgi?id=558002
> > 
> > "QoS fixes" alone resolve:
> > https://bugzilla.redhat.com/show_bug.cgi?id=539878
> > 
> > So I think we want all 4 patches in stable.
> > 
> > John,if possible push patches to Linus tree. When they lend there,
> > I will post backported patches to stable ML. I have them already
> > prepared ant tested for 2.6.32 :)
> 
> It is fairly late in the 2.6.34 cycle, and it isn't obvious to me
> that these fix any sort of regression.  Am I wrong?

If you count as regression bugs introduced in 2.6.34 development
cycle, these are not regression. Bugs are Fedora 12 regression over
Fedora 11, so most likely these bugs were introduced in 2.6.32
or 2.6.31 kernel. I have no strong filings to fix them fast, since
we have them unfixed for long time and not hear many complains. We
may fix them for Fedora now, and post to -stable after some more testing
and 2.6.34 release.

Cheers
Stanislaw