[PATCH] mac80211: remove HW_SIGNAL_DB

[PATCH] mac80211: remove HW_SIGNAL_DB

[Johannes Berg]

Giving the signal in dB isn't much more useful to userspace
than giving the signal in unspecified units. This removes
some radiotap information for zd1211 (the only driver using
this flag), but it helps a lot for getting cfg80211-based
scanning which won't support dB, and zd1211 being dB is a
little fishy anyway.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Cc: Bruno Randolf <bruno@thinktube.com>
---
 drivers/net/wireless/zd1211rw/zd_mac.c |    2 +-
 include/net/mac80211.h                 |   22 ++++++++--------------
 net/mac80211/main.c                    |    1 -
 net/mac80211/rx.c                      |   11 +----------
 net/mac80211/wext.c                    |    3 +--
 5 files changed, 11 insertions(+), 28 deletions(-)

--- wireless-testing.orig/include/net/mac80211.h	2009-01-30 10:57:26.000000000 +0100
+++ wireless-testing/include/net/mac80211.h	2009-01-30 10:57:49.000000000 +0100
@@ -860,11 +860,6 @@ enum ieee80211_tkip_key_type {
  *	expect values between 0 and @max_signal.
  *	If possible please provide dB or dBm instead.
  *
- * @IEEE80211_HW_SIGNAL_DB:
- *	Hardware gives signal values in dB, decibel difference from an
- *	arbitrary, fixed reference. We expect values between 0 and @max_signal.
- *	If possible please provide dBm instead.
- *
  * @IEEE80211_HW_SIGNAL_DBM:
  *	Hardware gives signal values in dBm, decibel difference from
  *	one milliwatt. This is the preferred method since it is standardized
@@ -900,15 +895,14 @@ enum ieee80211_hw_flags {
 	IEEE80211_HW_2GHZ_SHORT_SLOT_INCAPABLE		= 1<<3,
 	IEEE80211_HW_2GHZ_SHORT_PREAMBLE_INCAPABLE	= 1<<4,
 	IEEE80211_HW_SIGNAL_UNSPEC			= 1<<5,
-	IEEE80211_HW_SIGNAL_DB				= 1<<6,
-	IEEE80211_HW_SIGNAL_DBM				= 1<<7,
-	IEEE80211_HW_NOISE_DBM				= 1<<8,
-	IEEE80211_HW_SPECTRUM_MGMT			= 1<<9,
-	IEEE80211_HW_AMPDU_AGGREGATION			= 1<<10,
-	IEEE80211_HW_SUPPORTS_PS			= 1<<11,
-	IEEE80211_HW_PS_NULLFUNC_STACK			= 1<<12,
-	IEEE80211_HW_SUPPORTS_DYNAMIC_PS		= 1<<13,
-	IEEE80211_HW_MFP_CAPABLE			= 1<<14,
+	IEEE80211_HW_SIGNAL_DBM				= 1<<6,
+	IEEE80211_HW_NOISE_DBM				= 1<<7,
+	IEEE80211_HW_SPECTRUM_MGMT			= 1<<8,
+	IEEE80211_HW_AMPDU_AGGREGATION			= 1<<9,
+	IEEE80211_HW_SUPPORTS_PS			= 1<<10,
+	IEEE80211_HW_PS_NULLFUNC_STACK			= 1<<11,
+	IEEE80211_HW_SUPPORTS_DYNAMIC_PS		= 1<<12,
+	IEEE80211_HW_MFP_CAPABLE			= 1<<13,
 };
 
 /**
--- wireless-testing.orig/drivers/net/wireless/zd1211rw/zd_mac.c	2009-01-30 10:58:08.000000000 +0100
+++ wireless-testing/drivers/net/wireless/zd1211rw/zd_mac.c	2009-01-30 10:58:15.000000000 +0100
@@ -967,7 +967,7 @@ struct ieee80211_hw *zd_mac_alloc_hw(str
 	hw->wiphy->bands[IEEE80211_BAND_2GHZ] = &mac->band;
 
 	hw->flags = IEEE80211_HW_RX_INCLUDES_FCS |
-		    IEEE80211_HW_SIGNAL_DB;
+		    IEEE80211_HW_SIGNAL_UNSPEC;
 
 	hw->wiphy->interface_modes =
 		BIT(NL80211_IFTYPE_MESH_POINT) |
--- wireless-testing.orig/net/mac80211/main.c	2009-01-30 11:04:18.000000000 +0100
+++ wireless-testing/net/mac80211/main.c	2009-01-30 11:04:28.000000000 +0100
@@ -888,7 +888,6 @@ int ieee80211_register_hw(struct ieee802
 	local->hw.conf.listen_interval = local->hw.max_listen_interval;
 
 	local->wstats_flags |= local->hw.flags & (IEEE80211_HW_SIGNAL_UNSPEC |
-						  IEEE80211_HW_SIGNAL_DB |
 						  IEEE80211_HW_SIGNAL_DBM) ?
 			       IW_QUAL_QUAL_UPDATED : IW_QUAL_QUAL_INVALID;
 	local->wstats_flags |= local->hw.flags & IEEE80211_HW_NOISE_DBM ?
--- wireless-testing.orig/net/mac80211/rx.c	2009-01-30 11:07:35.000000000 +0100
+++ wireless-testing/net/mac80211/rx.c	2009-01-30 11:08:06.000000000 +0100
@@ -86,8 +86,7 @@ ieee80211_rx_radiotap_len(struct ieee802
 
 	if (status->flag & RX_FLAG_TSFT)
 		len += 8;
-	if (local->hw.flags & IEEE80211_HW_SIGNAL_DB ||
-	    local->hw.flags & IEEE80211_HW_SIGNAL_DBM)
+	if (local->hw.flags & IEEE80211_HW_SIGNAL_DBM)
 		len += 1;
 	if (local->hw.flags & IEEE80211_HW_NOISE_DBM)
 		len += 1;
@@ -199,14 +198,6 @@ ieee80211_add_rx_radiotap_header(struct 
 	*pos = status->antenna;
 	pos++;
 
-	/* IEEE80211_RADIOTAP_DB_ANTSIGNAL */
-	if (local->hw.flags & IEEE80211_HW_SIGNAL_DB) {
-		*pos = status->signal;
-		rthdr->it_present |=
-			cpu_to_le32(1 << IEEE80211_RADIOTAP_DB_ANTSIGNAL);
-		pos++;
-	}
-
 	/* IEEE80211_RADIOTAP_DB_ANTNOISE is not used */
 
 	/* IEEE80211_RADIOTAP_RX_FLAGS */
--- wireless-testing.orig/net/mac80211/wext.c	2009-01-30 11:06:02.000000000 +0100
+++ wireless-testing/net/mac80211/wext.c	2009-01-30 11:06:09.000000000 +0100
@@ -173,8 +173,7 @@ static int ieee80211_ioctl_giwrange(stru
 	range->num_encoding_sizes = 2;
 	range->max_encoding_tokens = NUM_DEFAULT_KEYS;
 
-	if (local->hw.flags & IEEE80211_HW_SIGNAL_UNSPEC ||
-	    local->hw.flags & IEEE80211_HW_SIGNAL_DB)
+	if (local->hw.flags & IEEE80211_HW_SIGNAL_UNSPEC)
 		range->max_qual.level = local->hw.max_signal;
 	else if  (local->hw.flags & IEEE80211_HW_SIGNAL_DBM)
 		range->max_qual.level = -110;

Re: [PATCH] mac80211: remove HW_SIGNAL_DB

[Luis R. Rodriguez]

On Fri, Jan 30, 2009 at 02:13:06AM -0800, Johannes Berg wrote:
> Giving the signal in dB isn't much more useful to userspace
> than giving the signal in unspecified units. This removes
> some radiotap information for zd1211 (the only driver using
> this flag), but it helps a lot for getting cfg80211-based
> scanning which won't support dB, and zd1211 being dB is a
> little fishy anyway.
> 
> Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
> Cc: Bruno Randolf <bruno@thinktube.com>

We should be able to convert zd1211rw to use dBm anyway, we shall
release a new "vendor driver" for it which I think has some bits for this.
Just need to get around to it.

  Luis

Re: [PATCH] mac80211: remove HW_SIGNAL_DB

[Gábor Stefanik]

On Fri, Jan 30, 2009 at 5:50 PM, Luis R. Rodriguez
<lrodriguez@atheros.com> wrote:
> On Fri, Jan 30, 2009 at 02:13:06AM -0800, Johannes Berg wrote:
>> Giving the signal in dB isn't much more useful to userspace
>> than giving the signal in unspecified units. This removes
>> some radiotap information for zd1211 (the only driver using
>> this flag), but it helps a lot for getting cfg80211-based
>> scanning which won't support dB, and zd1211 being dB is a
>> little fishy anyway.
>>
>> Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
>> Cc: Bruno Randolf <bruno@thinktube.com>
>
> We should be able to convert zd1211rw to use dBm anyway, we shall
> release a new "vendor driver" for it which I think has some bits for =
this.
> Just need to get around to it.
>
>  Luis
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wirel=
ess" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

While you are at it, please also add the ability to disable in-HW
sequence number generation on a per-packet basis. This is required to
allow my proposed IEEE80211_F_TX_NOSEQ radiotap flag to work. (Right
now, even if mac80211's ieee80211_tx_h_sequence is bypassed, injected
packets will still have their sequence numbers overwritten, preventing
proper injection of fragments, and by grepping through the code of
zd1211rw, there appears to be no code to disable HW sequence number
generation. Potentially this might mean that mac80211-generated
sequence numbers also get overwritten by HW-generated ones!)

--G=E1bor
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] mac80211: remove HW_SIGNAL_DB

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 303 bytes --]

On Fri, 2009-01-30 at 08:50 -0800, Luis R. Rodriguez wrote:

> We should be able to convert zd1211rw to use dBm anyway, we shall
> release a new "vendor driver" for it which I think has some bits for this.
> Just need to get around to it.

Yeah but we can start that from UNSPEC too.

johannes

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

Re: [PATCH] mac80211: remove HW_SIGNAL_DB

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 221 bytes --]

On Fri, 2009-01-30 at 18:50 +0100, Gábor Stefanik wrote:

> While you are at it, please also add the ability to disable in-HW
> sequence number generation on a per-packet basis.

Probably not possible.

johannes

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