Copy-and-paste error on the page for the "antenna noise" field?

Copy-and-paste error on the page for the "antenna noise" field?

[Guy Harris]

The page for the "antenna noise" field, at

	http://www.radiotap.org/fields/Antenna%20noise.html

describes it as

	RF noise power at the antenna. This field contains a single signed 8-bit value, which indicates the RF signal power at the antenna, in decibels difference from 1mW.

It looks as if this was copied and pasted from the "antenna signal" page, with the first instance of "signal" replaced by noise, but the second instance left as "signal".

Re: Copy-and-paste error on the page for the "antenna noise" field?

[Johannes Berg]

On Wed, 2023-01-11 at 19:16 -0800, Guy Harris wrote:
> The page for the "antenna noise" field, at
> 
> 	http://www.radiotap.org/fields/Antenna%20noise.html
> 
> describes it as
> 
> 	RF noise power at the antenna. This field contains a single
> signed 8-bit value, which indicates the RF signal power at the
> antenna, in decibels difference from 1mW.
> 
> It looks as if this was copied and pasted from the "antenna signal"
> page, with the first instance of "signal" replaced by noise, but the
> second instance left as "signal".

Hm. Yeah, seems like? This is really old :-)

You could argue that noise is also some kind of "RF signal" but yeah, it
seems to make sense to change that.

Any other opinions?

johannes

Re: Copy-and-paste error on the page for the "antenna noise" field?

[Guy Harris]

On Jan 11, 2023, at 11:46 PM, Johannes Berg <johannes@sipsolutions.net> wrote:

> Hm. Yeah, seems like? This is really old :-)

Yeah, I only noticed it when looking at, and copying and pasting text from, the ""antenna noise" page, to provide a comment on

	https://ask.wireshark.org/question/30268/how-to-get-noise-level-in-80211-headers-in-wireshark/

I infer, from the Wireshark dissection of the radiotap header, that their adapter, or the driver for it, provides only "antenna signal" information and no "antenna noise" information.

Re: Copy-and-paste error on the page for the "antenna noise" field?

[Johannes Berg]

On Thu, 2023-01-12 at 00:15 -0800, Guy Harris wrote:
> On Jan 11, 2023, at 11:46 PM, Johannes Berg <johannes@sipsolutions.net> wrote:
> 
> > Hm. Yeah, seems like? This is really old :-)
> 
> Yeah, I only noticed it when looking at, and copying and pasting text
> from, the ""antenna noise" page, to provide a comment on

Right.

> 	https://ask.wireshark.org/question/30268/how-to-get-noise-level-in-80211-headers-in-wireshark/
> 
> I infer, from the Wireshark dissection of the radiotap header, that
> their adapter, or the driver for it, provides only "antenna signal"
> information and no "antenna noise" information.

I don't know what adapter they're using to capture, but it _looks_ like
mac80211 (with the per-antenna signal encoded that way), and then quite
possibly Intel (WFA uses it), and then yes - we don't provide noise
information in the Intel NICs.

johannes

Re: Copy-and-paste error on the page for the "antenna noise" field?

[Guy Harris]

On Jan 12, 2023, at 12:30 AM, Johannes Berg <johannes@sipsolutions.net> wrote:

> I don't know what adapter they're using to capture, but it _looks_ like
> mac80211 (with the per-antenna signal encoded that way),

Should that encoding scheme be documented, so as to encourage its use in other OS's 802.11 frameworks?

(My long to-do list includes adding support for per-antenna values in Wireshark's "radio information adaptation layer - the "802.11 radio information"  and supporting radiotap's encoding scheme as well as the schemes used by non-pcap/pcapng capture files.  That layer is intended to provide radio information in a metadata-format-independent form, with fewer metadata implementation details poking through, and additional information derived from the raw metadata.)

Re: Copy-and-paste error on the page for the "antenna noise" field?

[Johannes Berg]

On Thu, 2023-01-12 at 00:58 -0800, Guy Harris wrote:
> On Jan 12, 2023, at 12:30 AM, Johannes Berg <johannes@sipsolutions.net> wrote:
> 
> > I don't know what adapter they're using to capture, but it _looks_ like
> > mac80211 (with the per-antenna signal encoded that way),
> 
> Should that encoding scheme be documented, so as to encourage its use
> in other OS's 802.11 frameworks?

I don't know.

For the record, the encoding is to use a separate radiotap namespace for
each antenna, and then put an antenna signal and antenna field into it,
with the signal value and antenna number, respectively.

It's honestly not the greatest encoding, but it's sometimes useful to
have that data.

I've been hoping to replace namespaces with TLVs, but obviously in this
case that doesn't work since the namespace holds the two pieces of
information together ...

johannes