Hi all, I attached another patch around ath5k_hw_reset().
If call ath5k_hw_reset from ath5k_init(), should pass false for argument
change_channel. Otherwise, ath5k will be missed the initialize of some
registers. for example: ath5k_hw_write_initvals() skip a lots of register.
Sincerely,
Forrest
----- Original Message -----
From: "Forrest Zhang" <forr...@263.net>
To: "Bob Copeland" <m...@bobcopeland.com>; <ath5k-devel@lists.ath5k.org>
Cc: "Nick Kossifidis" <mickfl...@gmail.com>
Sent: Wednesday, May 13, 2009 1:32 PM
Subject: Re: RFC phy timing register bug
> Hi all,
>
> The patch is created by me (forr...@hifulltech.com) or (forr...@263.net).
>
> When I test ath5k of linux2.6.30-rc1, I found it cannot communicate with
> other device on 5G band.
>
> After comparing the register settings of ath5k with madwifi and atheros
> driver, I found two major differences:
> AR5K_PHY_PLL
> AR5K_PHY_TIMING_3
>
> The first one is very simple, just write wrong value on 5413/5414。
>
> Compare the codes around AR5K_PHY_TIMING_3, I found new ath5k driver assign
> wrong coef_exp.
>
> Although we have no document about AR5K_PHY_TIMING_3. but I believe that we
> have to keep the original algorithm. and make hardware happiness..
>
> Sincerely,
> Forrest
> ----- Original Message -----
> From: "Bob Copeland" <m...@bobcopeland.com>
> To: <ath5k-devel@lists.ath5k.org>
> Cc: "Nick Kossifidis" <mickfl...@gmail.com>; <forr...@263.net>
> Sent: Tuesday, May 12, 2009 10:46 PM
> Subject: RFC phy timing register bug
>
>
>> Hi all,
>>
>> Someone pointed out to me a change in the code with the reset rework,
>> due to my suggestion to use get_bitmask_order() which actually returns
>> highest bit set+1. Nick, can you comment whether the patch below does
>> the right thing and that the original code was correct?
>>
>> [Forrest, is this idea via you, and may I attribute you as the patch
>> author like so?
>>
>> From: Forrest Zhang <your preferred email>]
>>
>> Here's a C program I used to confirm the problem (f1 is original code,
>> f2 is get_bitmask_order with generic fls(). Output is:
>>
>> 1: 0 1
>> 0: 0 0
>> 32767: 14 15
>> 65535: 15 16
>> 65536: 16 17
>> 1048576: 20 21
>>
>> --- start ---
>> #include <inttypes.h>
>> #include <stdio.h>
>>
>> static int fls(int x)
>> {
>> int r = 32;
>>
>> if (!x)
>> return 0;
>> if (!(x & 0xffff0000u)) {
>> x <<= 16;
>> r -= 16;
>> }
>> if (!(x & 0xff000000u)) {
>> x <<= 8;
>> r -= 8;
>> }
>> if (!(x & 0xf0000000u)) {
>> x <<= 4;
>> r -= 4;
>> }
>> if (!(x & 0xc0000000u)) {
>> x <<= 2;
>> r -= 2;
>> }
>> if (!(x & 0x80000000u)) {
>> x <<= 1;
>> r -= 1;
>> }
>> return r;
>> }
>>
>> int f1(int c)
>> {
>> int bit;
>> for (bit = 31; bit > 0; bit--)
>> if ((c >> bit) & 1)
>> break;
>>
>> return bit;
>> }
>>
>> int f2(int c)
>> {
>> return fls(c);
>> }
>>
>> int main()
>> {
>> int i;
>> int nums[] = { 1, 0, 32767, 65535, 65536, 1048576 };
>>
>> for (i=0; i < sizeof(nums)/sizeof(int); i++)
>> printf("%d: %d %d\n", nums[i], f1(nums[i]), f2(nums[i]));
>>
>> return 0;
>> }
>> --- end ---
>>
>> From d668403b04487a90378fc4ec7cd1ebe1e4b7bfe1 Mon Sep 17 00:00:00 2001
>> From: Bob Copeland <m...@bobcopeland.com>
>> Date: Tue, 12 May 2009 09:40:37 -0400
>> Subject: [PATCH] ath5k: fix off-by-one when computing OFDM delta slope
>>
>> get_bitmask_order() actually returns the highest bit set plus one,
>> whereas the previous code wanted the highest bit set. Also change
>> type of coef_exp to int to deal with degenerate case of
>> coef_exp == 0.
>> ---
>> drivers/net/wireless/ath/ath5k/reset.c | 8 ++++----
>> 1 files changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/net/wireless/ath/ath5k/reset.c
>> b/drivers/net/wireless/ath/ath5k/reset.c
>> index 4a2d93e..cc75e77 100644
>> --- a/drivers/net/wireless/ath/ath5k/reset.c
>> +++ b/drivers/net/wireless/ath/ath5k/reset.c
>> @@ -51,8 +51,8 @@ static inline int ath5k_hw_write_ofdm_timings(struct
>> ath5k_hw *ah,
>> struct ieee80211_channel *channel)
>> {
>> /* Get exponent and mantissa and set it */
>> - u32 coef_scaled, coef_exp, coef_man,
>> - ds_coef_exp, ds_coef_man, clock;
>> + u32 coef_scaled, coef_man, ds_coef_exp, ds_coef_man, clock;
>> + int coef_exp;
>>
>> if (!(ah->ah_version == AR5K_AR5212) ||
>> !(channel->hw_value & CHANNEL_OFDM))
>> @@ -69,10 +69,10 @@ static inline int ath5k_hw_write_ofdm_timings(struct
>> ath5k_hw *ah,
>>
>> /* Get exponent
>> * ALGO: coef_exp = 14 - highest set bit position */
>> - coef_exp = get_bitmask_order(coef_scaled);
>> + coef_exp = get_bitmask_order(coef_scaled) - 1;
>>
>> /* Doesn't make sense if it's zero*/
>> - if (!coef_exp)
>> + if (coef_exp <= 0)
>> return -EINVAL;
>>
>> /* Note: we've shifted coef_scaled by 24 */
>> --
>> 1.6.0.6
>>
>>
>> --
>> Bob Copeland %% www.bobcopeland.com
>>
>>
>>
ath5k_hw_reset.patch
Description: Binary data
_______________________________________________ ath5k-devel mailing list ath5k-devel@lists.ath5k.org https://lists.ath5k.org/mailman/listinfo/ath5k-devel
