Skip to contentSkip to conversation

MCS field - STBC and Ness

DISCUSSION Simon Barber 4 participants 12 messages May 11, 2012 – Nov 17, 2012 9 messages not in the archive

lore
Share & tools
lore
#0 lore raw

Hi all,

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

For 802.11n I need to know the number of HT-DLTF and HT-ELTF training symbols that are sent in the header - I have everything else needed to calculate the duration. Given the MCS and the STBC and the Ness (Number of extension spatial streams) I can calculate the number of training symbols.

Both STBC and Ness are encoded in the HT-SIG transmitted at the start of each HT frame.

The current proposal (or has it not yet been accepted?) for STBC allocates just 1 bit - but the STBC field can have values or 0, 1 or 2.

Ness can be 0-3, so the remaining 2 bits could be allocated to Ness, but this does not solve the problem of how to encode the modes with 2 spatial streams and 2 additional streams for STBC.

Does anyone know what current hardware supports - STBC values or 0,1,2? Ness values of 0-3?

Any suggestions for how to convey these

Simon

#2 lore raw
In reply to a message not in the archive

Hello,
Atheros families 9002 and 9003 support only one STBC stream. I think
I have seen a Ralink chipset
which supports 2 stream at least on receive. I am not sure about transmiter.
On Atheros you get STBC flag and two bits Ness field for received packets but In the patch there is is only flag reading but you could add Ness yourself it's only two bits left from the flag.

Myself I have tested only one stream but I did it only by looking whether other party is setting rx descriptor bit. I don't know how it looks "in the air". I trust manufacturer;o)

It's true that only one bit was allocated in wireshark extension. Should have been two but it's a bit waste. My proposal wasn't accepted but in the end I think it's ok becasue I have managed to get all info in vendor namesapes where I can do what I want.

Br,
Wojtek

Quoted 16 lines

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

For 802.11n I need to know the number of HT-DLTF and HT-ELTF training symbols that are sent in the header - I have everything else needed to calculate the duration. Given the MCS and the STBC and the Ness (Number of extension spatial streams) I can calculate the number of training symbols.

Both STBC and Ness are encoded in the HT-SIG transmitted at the start of each HT frame.

The current proposal (or has it not yet been accepted?) for STBC allocates just 1 bit - but the STBC field can have values or 0, 1 or 2.

Ness can be 0-3, so the remaining 2 bits could be allocated to Ness, but this does not solve the problem of how to encode the modes with 2 spatial streams and 2 additional streams for STBC.

Does anyone know what current hardware supports - STBC values or 0,1,2? Ness values of 0-3?

Any suggestions for how to convey these

Simon

#4 lore raw
In reply to a message not in the archive

One possible solution would be to adopt your proposal to add a single bit STBC, but also add a second bit to the 'knows' byte - STBC2. If this second bit is set, then it indicates STBC=2. This would leave one remaining bit in the 'knows' byte, to indicate the presence of Ness information, and 2 bits in the 'flags' byte for 2 bits of Ness information.

Copying the MCS definition from the radiotap.org website:

The known field indicates which information is known:

known definition
0x01 bandwidth
0x02 MCS index known (in mcs part of the field)
0x04 guard interval
0x08 HT format
0x10 FEC type
0x20 STBC - STBC is known
0x40 STBC2 - 2 STBC streams are present
0x80 Ness - Number of extra spatial streams is known

The flags field is any combination of the following:

flag definition
0x03 bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04 guard interval - 0: long GI, 1: short GI
0x08 HT format - 0: mixed, 1: greenfield
0x10 FEC type - 0: BCC, 1: LDPC
0x20 STBC - 0: no STBC streams, 1: 1 or 2 STBC streams (see STBC2)
0xc0 Ness - Number of extra spatial streams.

Simon

Quoted 49 lines · Wojciech Dubowik

Hello,
Atheros families 9002 and 9003 support only one STBC stream. I think I have seen a Ralink chipset
which supports 2 stream at least on receive. I am not sure about transmiter.
On Atheros you get STBC flag and two bits Ness field for received packets but In the patch there is is only flag reading but you could add Ness yourself it's only two bits left from the flag.

Myself I have tested only one stream but I did it only by looking whether other party is setting rx descriptor bit. I don't know how it looks "in the air". I trust manufacturer;o)

It's true that only one bit was allocated in wireshark extension. Should have been two but it's a bit waste. My proposal wasn't accepted but in the end I think it's ok becasue I have managed to get all info in vendor namesapes where I can do what I want.

Br,
Wojtek

Quoted 27 lines

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

For 802.11n I need to know the number of HT-DLTF and HT-ELTF training symbols that are sent in the header - I have everything else needed to calculate the duration. Given the MCS and the STBC and the Ness (Number of extension spatial streams) I can calculate the number of training symbols.

Both STBC and Ness are encoded in the HT-SIG transmitted at the start of each HT frame.

The current proposal (or has it not yet been accepted?) for STBC allocates just 1 bit - but the STBC field can have values or 0, 1 or 2.

Ness can be 0-3, so the remaining 2 bits could be allocated to Ness, but this does not solve the problem of how to encode the modes with 2 spatial streams and 2 additional streams for STBC.

Does anyone know what current hardware supports - STBC values or 0,1,2? Ness values of 0-3?

Any suggestions for how to convey these

Simon

#6 lore raw
In reply to a message not in the archive

OK - perhaps a simpler/better suggestion would be for the STBC2 bit to just be the second bit of STBC information.

Simon

Quoted 85 lines · Simon Barber

One possible solution would be to adopt your proposal to add a single bit STBC, but also add a second bit to the 'knows' byte - STBC2. If this second bit is set, then it indicates STBC=2. This would leave one remaining bit in the 'knows' byte, to indicate the presence of Ness information, and 2 bits in the 'flags' byte for 2 bits of Ness information.

Copying the MCS definition from the radiotap.org website:

The known field indicates which information is known:

known definition
0x01 bandwidth
0x02 MCS index known (in mcs part of the field) 0x04 guard interval
0x08 HT format
0x10 FEC type
0x20 STBC - STBC is known
0x40 STBC2 - 2 STBC streams are present
0x80 Ness - Number of extra spatial streams is known

The flags field is any combination of the following:

flag definition
0x03 bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04 guard interval - 0: long GI, 1: short GI
0x08 HT format - 0: mixed, 1: greenfield
0x10 FEC type - 0: BCC, 1: LDPC
0x20 STBC - 0: no STBC streams, 1: 1 or 2 STBC streams (see STBC2) 0xc0 Ness - Number of extra spatial streams.

Simon

Quoted 49 lines · Wojciech Dubowik

Hello,
Atheros families 9002 and 9003 support only one STBC stream. I think I have seen a Ralink chipset
which supports 2 stream at least on receive. I am not sure about transmiter.
On Atheros you get STBC flag and two bits Ness field for received packets but In the patch there is is only flag reading but you could add Ness yourself it's only two bits left from the flag.

Myself I have tested only one stream but I did it only by looking whether other party is setting rx descriptor bit. I don't know how it looks "in the air". I trust manufacturer;o)

It's true that only one bit was allocated in wireshark extension. Should have been two but it's a bit waste. My proposal wasn't accepted but in the end I think it's ok becasue I have managed to get all info in vendor namesapes where I can do what I want.

Br,
Wojtek

Quoted 27 lines

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

For 802.11n I need to know the number of HT-DLTF and HT-ELTF training symbols that are sent in the header - I have everything else needed to calculate the duration. Given the MCS and the STBC and the Ness (Number of extension spatial streams) I can calculate the number of training symbols.

Both STBC and Ness are encoded in the HT-SIG transmitted at the start of each HT frame.

The current proposal (or has it not yet been accepted?) for STBC allocates just 1 bit - but the STBC field can have values or 0, 1 or 2.

Ness can be 0-3, so the remaining 2 bits could be allocated to Ness, but this does not solve the problem of how to encode the modes with 2 spatial streams and 2 additional streams for STBC.

Does anyone know what current hardware supports - STBC values or 0,1,2? Ness values of 0-3?

Any suggestions for how to convey these

Simon

#8 lore raw
In reply to a message not in the archive
Quoted 2 lines · Simon Barber

OK - perhaps a simpler/better suggestion would be for the STBC2 bit to just be the second bit of STBC information.

Seems reasonable to me, then we'll have filled all the bits anyway.

johannes

Quoted 90 lines

Simon

Quoted 85 lines · Simon Barber

One possible solution would be to adopt your proposal to add a single bit STBC, but also add a second bit to the 'knows' byte - STBC2. If this second bit is set, then it indicates STBC=2. This would leave one remaining bit in the 'knows' byte, to indicate the presence of Ness information, and 2 bits in the 'flags' byte for 2 bits of Ness information.

Copying the MCS definition from the radiotap.org website:

The known field indicates which information is known:

known definition
0x01 bandwidth
0x02 MCS index known (in mcs part of the field) 0x04 guard interval
0x08 HT format
0x10 FEC type
0x20 STBC - STBC is known
0x40 STBC2 - 2 STBC streams are present
0x80 Ness - Number of extra spatial streams is known

The flags field is any combination of the following:

flag definition
0x03 bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04 guard interval - 0: long GI, 1: short GI
0x08 HT format - 0: mixed, 1: greenfield
0x10 FEC type - 0: BCC, 1: LDPC
0x20 STBC - 0: no STBC streams, 1: 1 or 2 STBC streams (see STBC2) 0xc0 Ness - Number of extra spatial streams.

Simon

Quoted 49 lines · Wojciech Dubowik

Hello,
Atheros families 9002 and 9003 support only one STBC stream. I think I have seen a Ralink chipset
which supports 2 stream at least on receive. I am not sure about transmiter.
On Atheros you get STBC flag and two bits Ness field for received packets but In the patch there is is only flag reading but you could add Ness yourself it's only two bits left from the flag.

Myself I have tested only one stream but I did it only by looking whether other party is setting rx descriptor bit. I don't know how it looks "in the air". I trust manufacturer;o)

It's true that only one bit was allocated in wireshark extension. Should have been two but it's a bit waste. My proposal wasn't accepted but in the end I think it's ok becasue I have managed to get all info in vendor namesapes where I can do what I want.

Br,
Wojtek

Quoted 27 lines

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

For 802.11n I need to know the number of HT-DLTF and HT-ELTF training symbols that are sent in the header - I have everything else needed to calculate the duration. Given the MCS and the STBC and the Ness (Number of extension spatial streams) I can calculate the number of training symbols.

Both STBC and Ness are encoded in the HT-SIG transmitted at the start of each HT frame.

The current proposal (or has it not yet been accepted?) for STBC allocates just 1 bit - but the STBC field can have values or 0, 1 or 2.

Ness can be 0-3, so the remaining 2 bits could be allocated to Ness, but this does not solve the problem of how to encode the modes with 2 spatial streams and 2 additional streams for STBC.

Does anyone know what current hardware supports - STBC values or 0,1,2? Ness values of 0-3?

Any suggestions for how to convey these

Simon

#10 lore raw
In reply to a message not in the archive

Another slight tweak - I think STBC is more commonly used than Ness (as witnessed by Wojciech's interest), so give STBC the simpler encoding. I believe Ness is used for beamforming soundings.

Spec would look like this:

known definition

0x01    bandwidth
0x02    MCS index known (in mcs part of the field)
0x04    guard interval
0x08    HT format
0x10    FEC type
0x20    STBC
0x40    Ness - Number of extension spatial streams is known
0x80    Ness2 - bit 1 (MSB) of Number of extension spatial streams.

The flags field is any combination of the following:

flag    definition
0x03    bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04    guard interval - 0: long GI, 1: short GI
0x08    HT format - 0: mixed, 1: greenfield
0x10    FEC type - 0: BCC, 1: LDPC
0x60    STBC - 0, 1 or 2: number of STBC streams
0x80    Ness - bit 0 (LSB) of Number of extension spatial streams.

What should I do to advance this as an approved radiotap extension?

Simon

Quoted 100 lines · Johannes Berg
Quoted 2 lines · Simon Barber

OK - perhaps a simpler/better suggestion would be for the STBC2 bit to just be the second bit of STBC information.

Seems reasonable to me, then we'll have filled all the bits anyway.

johannes

Quoted 90 lines

Simon

Quoted 85 lines · Simon Barber

One possible solution would be to adopt your proposal to add a single bit STBC, but also add a second bit to the 'knows' byte - STBC2. If this second bit is set, then it indicates STBC=2. This would leave one remaining bit in the 'knows' byte, to indicate the presence of Ness information, and 2 bits in the 'flags' byte for 2 bits of Ness information.

Copying the MCS definition from the radiotap.org website:

The known field indicates which information is known:

known definition
0x01 bandwidth
0x02 MCS index known (in mcs part of the field) 0x04 guard interval
0x08 HT format
0x10 FEC type
0x20 STBC - STBC is known
0x40 STBC2 - 2 STBC streams are present
0x80 Ness - Number of extra spatial streams is known

The flags field is any combination of the following:

flag definition
0x03 bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04 guard interval - 0: long GI, 1: short GI
0x08 HT format - 0: mixed, 1: greenfield
0x10 FEC type - 0: BCC, 1: LDPC
0x20 STBC - 0: no STBC streams, 1: 1 or 2 STBC streams (see STBC2) 0xc0 Ness - Number of extra spatial streams.

Simon

Quoted 49 lines · Wojciech Dubowik

Hello,
Atheros families 9002 and 9003 support only one STBC stream. I think I have seen a Ralink chipset
which supports 2 stream at least on receive. I am not sure about transmiter.
On Atheros you get STBC flag and two bits Ness field for received packets but In the patch there is is only flag reading but you could add Ness yourself it's only two bits left from the flag.

Myself I have tested only one stream but I did it only by looking whether other party is setting rx descriptor bit. I don't know how it looks "in the air". I trust manufacturer;o)

It's true that only one bit was allocated in wireshark extension. Should have been two but it's a bit waste. My proposal wasn't accepted but in the end I think it's ok becasue I have managed to get all info in vendor namesapes where I can do what I want.

Br,
Wojtek

Quoted 27 lines

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

For 802.11n I need to know the number of HT-DLTF and HT-ELTF training symbols that are sent in the header - I have everything else needed to calculate the duration. Given the MCS and the STBC and the Ness (Number of extension spatial streams) I can calculate the number of training symbols.

Both STBC and Ness are encoded in the HT-SIG transmitted at the start of each HT frame.

The current proposal (or has it not yet been accepted?) for STBC allocates just 1 bit - but the STBC field can have values or 0, 1 or 2.

Ness can be 0-3, so the remaining 2 bits could be allocated to Ness, but this does not solve the problem of how to encode the modes with 2 spatial streams and 2 additional streams for STBC.

Does anyone know what current hardware supports - STBC values or 0,1,2? Ness values of 0-3?

Any suggestions for how to convey these

Simon

#20 lore raw
In reply to a message not in the archive
Quoted 5 lines · Simon Barber

Hi all,

I am writing a Wireshark extension to visualise 802.11 frames on a timeline. In order to do this I need to be able to calculate the duration from start of TX to end of TX for each frame.

Just in case it is helpful, here is a report written by an undergraduate project team that I mentored back in 2004. The task I'd given them was to produce a Wi-Fi visualizer. They were not very successful in their software development, but the report sets out the goal and some design ideas.

Dave

Signature · 2 lines
David Young
[email protected]    Urbana, IL    (217) 721-9981
#12 lore raw
In reply to a message not in the archive
Quoted 26 lines · Simon Barber

Another slight tweak - I think STBC is more commonly used than Ness (as witnessed by Wojciech's interest), so give STBC the simpler encoding. I believe Ness is used for beamforming soundings.

Spec would look like this:

known definition

0x01    bandwidth
0x02    MCS index known (in mcs part of the field)
0x04    guard interval
0x08    HT format
0x10    FEC type
0x20    STBC
0x40    Ness - Number of extension spatial streams is known
0x80    Ness2 - bit 1 (MSB) of Number of extension spatial streams.

The flags field is any combination of the following:

flag    definition
0x03    bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04    guard interval - 0: long GI, 1: short GI
0x08    HT format - 0: mixed, 1: greenfield
0x10    FEC type - 0: BCC, 1: LDPC
0x60    STBC - 0, 1 or 2: number of STBC streams
0x80    Ness - bit 0 (LSB) of Number of extension spatial streams.
Quoted 1 line

What should I do to advance this as an approved radiotap extension?

I suppose you already have patches for wireshark for this? I can help you out with a patch for the Linux kernel as an example producer of this, and then I think just repost it to the list again to leave some more discussion time.

johannes

#14 lore raw
In reply to a message not in the archive

Hi Johannes & Wojciech,

My work has been on hold, but has recently come off hold. I'm getting the patches together. As I rebase my work onto the current wireshark trunk I've found a simpler implementation of STBC has been applied by Wojciech. My proposal extends the STBC from his implemented 1 bit to the full 2 bits allowed in the 11n specification, and also includes 2 bits of Ness. I've posted an ieee80211 patch, and an intel driver patch (both attached to the proposed radiotap extension page). l'll post a wireshark patch shortly to start the comment period.

Simon

Quoted 36 lines · Johannes Berg
Quoted 26 lines · Simon Barber

Another slight tweak - I think STBC is more commonly used than Ness (as witnessed by Wojciech's interest), so give STBC the simpler encoding. I believe Ness is used for beamforming soundings.

Spec would look like this:

known definition

0x01    bandwidth
0x02    MCS index known (in mcs part of the field)
0x04    guard interval
0x08    HT format
0x10    FEC type
0x20    STBC
0x40    Ness - Number of extension spatial streams is known
0x80    Ness2 - bit 1 (MSB) of Number of extension spatial streams.

The flags field is any combination of the following:

flag    definition
0x03    bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04    guard interval - 0: long GI, 1: short GI
0x08    HT format - 0: mixed, 1: greenfield
0x10    FEC type - 0: BCC, 1: LDPC
0x60    STBC - 0, 1 or 2: number of STBC streams
0x80    Ness - bit 0 (LSB) of Number of extension spatial streams.
Quoted 1 line

What should I do to advance this as an approved radiotap extension?

I suppose you already have patches for wireshark for this? I can help you out with a patch for the Linux kernel as an example producer of this, and then I think just repost it to the list again to leave some more discussion time.

johannes

#16 lore raw
In reply to a message not in the archive

OK - I have now posted 3 patches on the proposal page (see Attachments):

1. A patch that applies to the Linux kernel v3.7-rc1 to collect the new STBC and Ness parameters from a wireless driver, and add them into the MCS radiotap field.
2. A patch to the Intel wireless driver in the kernel to collect STBC and Ness information.
3. A patch to wireshark to display STBC and Ness information.

With this I believe we have everything needed to start the 3 week comment period.

Simon

Quoted 51 lines · Simon Barber

Hi Johannes & Wojciech,

My work has been on hold, but has recently come off hold. I'm getting the patches together. As I rebase my work onto the current wireshark trunk I've found a simpler implementation of STBC has been applied by Wojciech. My proposal extends the STBC from his implemented 1 bit to the full 2 bits allowed in the 11n specification, and also includes 2 bits of Ness. I've posted an ieee80211 patch, and an intel driver patch (both attached to the proposed radiotap extension page). l'll post a wireshark patch shortly to start the comment period.

Simon

Quoted 36 lines · Johannes Berg
Quoted 26 lines · Simon Barber

Another slight tweak - I think STBC is more commonly used than Ness (as witnessed by Wojciech's interest), so give STBC the simpler encoding. I believe Ness is used for beamforming soundings.

Spec would look like this:

known definition

0x01    bandwidth
0x02    MCS index known (in mcs part of the field)
0x04    guard interval
0x08    HT format
0x10    FEC type
0x20    STBC
0x40    Ness - Number of extension spatial streams is known
0x80    Ness2 - bit 1 (MSB) of Number of extension spatial streams.

The flags field is any combination of the following:

flag    definition
0x03    bandwidth - 0: 20, 1: 40, 2: 20L, 3: 20U
0x04    guard interval - 0: long GI, 1: short GI
0x08    HT format - 0: mixed, 1: greenfield
0x10    FEC type - 0: BCC, 1: LDPC
0x60    STBC - 0, 1 or 2: number of STBC streams
0x80    Ness - bit 0 (LSB) of Number of extension spatial streams.
Quoted 1 line

What should I do to advance this as an approved radiotap extension?

I suppose you already have patches for wireshark for this? I can help you out with a patch for the Linux kernel as an example producer of this, and then I think just repost it to the list again to leave some more discussion time.

johannes

#18 lore raw
In reply to a message not in the archive
Quoted 8 lines · Simon Barber

OK - I have now posted 3 patches on the proposal page (see Attachments):

1. A patch that applies to the Linux kernel v3.7-rc1 to collect the new STBC and Ness parameters from a wireless driver, and add them into the MCS radiotap field.
2. A patch to the Intel wireless driver in the kernel to collect STBC and Ness information.
3. A patch to wireshark to display STBC and Ness information.

Nice!

Quoted 2 lines

With this I believe we have everything needed to start the 3 week comment period.

Might be worthwhile to post a new thread so it gets everybody's attention?

johannes

#19 lore raw

Standardisation - adding 2 bit STBC and Ness to MCS

In reply to a message not in the archive

http://www.radiotap.org/suggested-fields/MCS%20extension%20for%20STBC%20and%20Ness

I have posted 3 patches on the proposal page (see Attachments):

1. A patch that applies to the Linux kernel v3.7-rc1 to collect the new STBC and Ness parameters from a wireless driver, and add them into the MCS radiotap field.
2. A patch to the Intel wireless driver in the kernel to collect STBC and Ness information.
3. A patch to wireshark to display STBC and Ness information.

With this I believe we have everything needed to start the 3 week comment period.

Simon Barber