From: Zefir Kurtisi <zefir.kurtisi@neratec.com>
To: Nick Kossifidis <mickflemm@gmail.com>
Cc: linux-wireless@vger.kernel.org, ath9k-devel@venema.h4ckr.net
Subject: Re: [PATCH] ath9k: spectral - simplify max_index calculation
Date: Thu, 18 Jun 2015 12:36:08 +0200 [thread overview]
Message-ID: <55829F18.9050807@neratec.com> (raw)
In-Reply-To: <CAFtRNNxvBC8SD3P5u=bEZPofA-A8YfHYuZnfA4Y07W7Wx4keAQ@mail.gmail.com>
On 06/18/2015 10:43 AM, Nick Kossifidis wrote:
> max_index is a 6bit signed integer in both cases (sorry for the 5bit
> typo in the comments), so the current function handles it correctly
> for both HT20 and dynamic HT20/40 modes (I've tested it extensively).
> Also you don't handle the negative indices we get from the hardware
> (you just remove the sign). Have you tested this and if you did on
> which hardware did you do the test ?
>
> 2015-06-16 11:21 GMT+02:00 Zefir Kurtisi <zefir.kurtisi@neratec.com>:
>> [...]
>> +/* return the max magnitude from the all/upper/lower bins
>> + *
>> + * in HT20: 6-bit signed number of range -28 to +27
>> + * in HT40: 6-bit unsigned number of range 0 to +63
>> + * (upper sub-channel index 0 is DC)
>> + *
>> + * Correct interpretation of the value has to be done at caller
>> + */
The comment above is taken from developer NDA documents and should be correct.
With that, in HT40 mode the max_index value has to be taken as is, while in HT20
it needs to be shifted to the unsigned range.
I used the proposed method with the chirp detector for FFTs provided for long
radar pulses on an AR9590 (patch posted the same day). Max bin index is used there
the same way as with spectral, but now I realize my mistake: for chirp detection,
the relative max_index is sufficient, while for spectral the absolute value is needed.
Toggling the MSB in HT20 shifts the signed values by 32 and leaves the index with
an offset of 4, therefore the correct operation should be:
ht20_max_index_absolute = (ht20_max_index ^ 0x20) - 4
The following code snipped proves the correctness of the operation:
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
const char *char_to_bin(char x)
{
int z;
static char b[7];
b[0] = '\0';
for (z = 32; z > 0; z >>= 1)
strcat(b, ((x & z) == z) ? "1" : "0");
return b;
}
int main (void)
{
char s;
for (s = -28; s < 28; s++)
printf("%3d: %s\n", s, char_to_bin((s ^ 0x20) - 4));
return 0;
}
I'll provide a v2 patch for you to test. My intention with this is not to fix
something that is already working, but to share the related functions between
spectral and chirp detector.
Thanks,
Zefir
WARNING: multiple messages have this Message-ID (diff)
From: Zefir Kurtisi <zefir.kurtisi@neratec.com>
To: ath9k-devel@lists.ath9k.org
Subject: [ath9k-devel] [PATCH] ath9k: spectral - simplify max_index calculation
Date: Thu, 18 Jun 2015 12:36:08 +0200 [thread overview]
Message-ID: <55829F18.9050807@neratec.com> (raw)
In-Reply-To: <CAFtRNNxvBC8SD3P5u=bEZPofA-A8YfHYuZnfA4Y07W7Wx4keAQ@mail.gmail.com>
On 06/18/2015 10:43 AM, Nick Kossifidis wrote:
> max_index is a 6bit signed integer in both cases (sorry for the 5bit
> typo in the comments), so the current function handles it correctly
> for both HT20 and dynamic HT20/40 modes (I've tested it extensively).
> Also you don't handle the negative indices we get from the hardware
> (you just remove the sign). Have you tested this and if you did on
> which hardware did you do the test ?
>
> 2015-06-16 11:21 GMT+02:00 Zefir Kurtisi <zefir.kurtisi@neratec.com>:
>> [...]
>> +/* return the max magnitude from the all/upper/lower bins
>> + *
>> + * in HT20: 6-bit signed number of range -28 to +27
>> + * in HT40: 6-bit unsigned number of range 0 to +63
>> + * (upper sub-channel index 0 is DC)
>> + *
>> + * Correct interpretation of the value has to be done at caller
>> + */
The comment above is taken from developer NDA documents and should be correct.
With that, in HT40 mode the max_index value has to be taken as is, while in HT20
it needs to be shifted to the unsigned range.
I used the proposed method with the chirp detector for FFTs provided for long
radar pulses on an AR9590 (patch posted the same day). Max bin index is used there
the same way as with spectral, but now I realize my mistake: for chirp detection,
the relative max_index is sufficient, while for spectral the absolute value is needed.
Toggling the MSB in HT20 shifts the signed values by 32 and leaves the index with
an offset of 4, therefore the correct operation should be:
ht20_max_index_absolute = (ht20_max_index ^ 0x20) - 4
The following code snipped proves the correctness of the operation:
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
const char *char_to_bin(char x)
{
int z;
static char b[7];
b[0] = '\0';
for (z = 32; z > 0; z >>= 1)
strcat(b, ((x & z) == z) ? "1" : "0");
return b;
}
int main (void)
{
char s;
for (s = -28; s < 28; s++)
printf("%3d: %s\n", s, char_to_bin((s ^ 0x20) - 4));
return 0;
}
I'll provide a v2 patch for you to test. My intention with this is not to fix
something that is already working, but to share the related functions between
spectral and chirp detector.
Thanks,
Zefir
next prev parent reply other threads:[~2015-06-18 10:36 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-06-16 9:21 [PATCH] ath9k: spectral - simplify max_index calculation Zefir Kurtisi
2015-06-16 9:21 ` [ath9k-devel] " Zefir Kurtisi
2015-06-18 8:43 ` Nick Kossifidis
2015-06-18 8:43 ` [ath9k-devel] " Nick Kossifidis
2015-06-18 10:36 ` Zefir Kurtisi [this message]
2015-06-18 10:36 ` Zefir Kurtisi
2015-06-18 14:13 ` Nick Kossifidis
2015-06-18 14:13 ` [ath9k-devel] " Nick Kossifidis
2015-06-18 15:11 ` Zefir Kurtisi
2015-06-18 15:11 ` [ath9k-devel] " Zefir Kurtisi
2015-06-18 15:59 ` Nick Kossifidis
2015-06-18 15:59 ` [ath9k-devel] " Nick Kossifidis
2015-06-18 15:34 ` Nick Kossifidis
2015-06-18 15:34 ` [ath9k-devel] " Nick Kossifidis
2015-06-18 15:46 ` Nick Kossifidis
2015-06-18 15:46 ` [ath9k-devel] " Nick Kossifidis
2015-06-18 16:40 ` Zefir Kurtisi
2015-06-18 16:40 ` [ath9k-devel] " Zefir Kurtisi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55829F18.9050807@neratec.com \
--to=zefir.kurtisi@neratec.com \
--cc=ath9k-devel@venema.h4ckr.net \
--cc=linux-wireless@vger.kernel.org \
--cc=mickflemm@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.