From: Felix Fietkau <nbd@openwrt.org>
To: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Cc: "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>,
Luis Rodriguez <Luis.Rodriguez@Atheros.com>,
"linville@tuxdriver.com" <linville@tuxdriver.com>
Subject: Re: [PATCH 2/4] ath9k_hw: merge codepaths that access the cycle counter registers
Date: Mon, 04 Oct 2010 13:36:43 +0200 [thread overview]
Message-ID: <4CA9BC4B.4010503@openwrt.org> (raw)
In-Reply-To: <20101004063446.GA16078@vasanth-laptop>
On 2010-10-04 8:34 AM, Vasanthakumar Thiagarajan wrote:
>> + /* freeze counters */
>> + REG_WRITE(ah, AR_MIBC, AR_MIBC_FMC);
>> +
>> + ah->cc.cycles = REG_READ(ah, AR_CCCNT);
>> + if (ah->cc.cycles < cc.cycles) {
>> + clear = true;
>> + goto skip;
>> }
>>
>> - cycles = cc;
>> - rx_frame = rf;
>> - rx_clear = rc;
>> - tx_frame = tf;
>> + ah->cc.rx_clear = REG_READ(ah, AR_RCCNT);
>> + ah->cc.rx_frame = REG_READ(ah, AR_RFCNT);
>> + ah->cc.tx_frame = REG_READ(ah, AR_TFCNT);
>> +
>> + /* prevent wraparound */
>> + if (ah->cc.cycles & BIT(31))
>> + clear = true;
>
> This does not look right, previous if should take care of
> any wrap around.
This is not for correcting an existing wraparound. This is for making
sure that a wraparound never occurs.
>> +
>> +#define CC_DELTA(_field, _reg) ah->cc_delta._field += ah->cc._field - cc._field
> _reg is not used.
>
>>
>> +skip:
>> + if (clear) {
>> + REG_WRITE(ah, AR_CCCNT, 0);
>> + REG_WRITE(ah, AR_RFCNT, 0);
>> + REG_WRITE(ah, AR_RCCNT, 0);
>> + REG_WRITE(ah, AR_TFCNT, 0);
>
> should be able to do with single write in AR_MIBC.
No, the clear bit in AR_MIBC does not clear these counters. I tested that.
>> + /* unfreeze counters */
>> + REG_WRITE(ah, AR_MIBC, 0);
>
> Please configure the relevant bit to unfreeze the counters.
What do you mean?
- Felix
next prev parent reply other threads:[~2010-10-04 11:36 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-10-03 17:07 [PATCH 1/4] ath9k_hw: clean up calibration flags Felix Fietkau
2010-10-03 17:07 ` [PATCH 2/4] ath9k_hw: merge codepaths that access the cycle counter registers Felix Fietkau
2010-10-03 17:07 ` [PATCH 3/4] ath9k_hw: clean up register write buffering Felix Fietkau
2010-10-03 17:07 ` [PATCH 4/4] ath9k_hw: fix regression in ANI listen time calculation Felix Fietkau
2010-10-03 17:24 ` Luis R. Rodriguez
2010-10-03 19:08 ` Felix Fietkau
2010-10-03 19:10 ` Felix Fietkau
2010-10-04 8:56 ` [PATCH 3/4] ath9k_hw: clean up register write buffering Rajkumar Manoharan
2010-10-04 11:53 ` Felix Fietkau
2010-10-04 17:22 ` [PATCH v2 " Felix Fietkau
2010-10-05 4:38 ` Senthil Balasubramanian
2010-10-05 10:03 ` [PATCH v3 " Felix Fietkau
2010-10-06 20:32 ` John W. Linville
2010-10-04 6:34 ` [PATCH 2/4] ath9k_hw: merge codepaths that access the cycle counter registers Vasanthakumar Thiagarajan
2010-10-04 8:17 ` Adrian Chadd
2010-10-04 11:36 ` Felix Fietkau [this message]
2010-10-04 12:24 ` Vasanthakumar Thiagarajan
2010-10-04 12:51 ` Felix Fietkau
2010-10-05 9:59 ` Bruno Randolf
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=4CA9BC4B.4010503@openwrt.org \
--to=nbd@openwrt.org \
--cc=Luis.Rodriguez@Atheros.com \
--cc=linux-wireless@vger.kernel.org \
--cc=linville@tuxdriver.com \
--cc=vasanth@atheros.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).