linux-wireless.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ryder Lee <ryder.lee@mediatek.com>
To: <greearb@candelatech.com>, <linux-wireless@vger.kernel.org>
Subject: Re: [PATCH 1/8] mt76: mt7915: cache sgi in wcid.
Date: Fri, 19 Nov 2021 01:06:25 +0800	[thread overview]
Message-ID: <c0a09707cf9bcf1205b60c7a36e5afa2a6b2f39c.camel@mediatek.com> (raw)
In-Reply-To: <20211118164549.3863-1-greearb@candelatech.com>

On Thu, 2021-11-18 at 08:45 -0800, greearb@candelatech.com wrote:
> From: Ben Greear <greearb@candelatech.com>
> 
> Explicitly cache short_gi and he_gi in wcid, don't try to store
> it in the wcid.rate object.  Slightly less confusing and less fragile
> when TXS starts parsing lots of frames.
> 
> Signed-off-by: Ben Greear <greearb@candelatech.com>
> ---
> 
> This is actually the first series, not that one I posted a few
> minutes
> ago.  txs, tx overrides and other things.  Rebased on top of 5.16
> 
>  drivers/net/wireless/mediatek/mt76/mt76.h       |  5 +++++
>  drivers/net/wireless/mediatek/mt76/mt7915/mac.c | 17 +++++++++++++
> ----
>  2 files changed, 18 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/wireless/mediatek/mt76/mt76.h
> b/drivers/net/wireless/mediatek/mt76/mt76.h
> index e2da720a91b6..7234703b3c60 100644
> --- a/drivers/net/wireless/mediatek/mt76/mt76.h
> +++ b/drivers/net/wireless/mediatek/mt76/mt76.h
> @@ -245,7 +245,12 @@ struct mt76_wcid {
>  	struct ewma_signal rssi;
>  	int inactive_count;
>  
> +	/* cached rate, updated from mac_sta_poll() and from TXS
> callback logic,
> +	 * in 7915 at least.
> +	 */
>  	struct rate_info rate;
> +	bool rate_short_gi; /* cached HT/VHT short_gi, from
> mac_sta_poll() */
> +	u8 rate_he_gi; /* cached HE GI, from mac_sta_poll() */
>  
>  	u16 idx;
>  	u8 hw_key_idx;
> diff --git a/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> b/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> index 5fcf35f2d9fb..61ade279b35d 100644
> --- a/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> +++ b/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> @@ -206,13 +206,19 @@ static void mt7915_mac_sta_poll(struct
> mt7915_dev *dev)
>  			u8 offs = 24 + 2 * bw;
>  
>  			rate->he_gi = (val & (0x3 << offs)) >> offs;
> +			msta->wcid.rate_he_gi = rate->he_gi; /* cache
> for later */
>  		} else if (rate->flags &
>  			   (RATE_INFO_FLAGS_VHT_MCS |
> RATE_INFO_FLAGS_MCS)) {
> -			if (val & BIT(12 + bw))
> +			if (val & BIT(12 + bw)) {
>  				rate->flags |=
> RATE_INFO_FLAGS_SHORT_GI;
> -			else
> +				msta->wcid.rate_short_gi = 1;
> +			}
> +			else {
>  				rate->flags &=
> ~RATE_INFO_FLAGS_SHORT_GI;
> +				msta->wcid.rate_short_gi = 0;
> +			}
>  		}
> +		/* TODO:  Deal with HT_MCS */
>  	}
>  
>  	rcu_read_unlock();
> @@ -1411,7 +1417,7 @@ mt7915_mac_add_txs_skb(struct mt7915_dev *dev,
> struct mt76_wcid *wcid, int pid,
>  			goto out;
>  
>  		rate.flags = RATE_INFO_FLAGS_MCS;
> -		if (wcid->rate.flags & RATE_INFO_FLAGS_SHORT_GI)
> +		if (wcid->rate_short_gi)
>  			rate.flags |= RATE_INFO_FLAGS_SHORT_GI;
>  		break;
>  	case MT_PHY_TYPE_VHT:
> @@ -1419,6 +1425,8 @@ mt7915_mac_add_txs_skb(struct mt7915_dev *dev,
> struct mt76_wcid *wcid, int pid,
>  			goto out;
>  
>  		rate.flags = RATE_INFO_FLAGS_VHT_MCS;
> +		if (wcid->rate_short_gi)
> +			rate.flags |= RATE_INFO_FLAGS_SHORT_GI;
>  		break;
>  	case MT_PHY_TYPE_HE_SU:
>  	case MT_PHY_TYPE_HE_EXT_SU:
> @@ -1427,11 +1435,12 @@ mt7915_mac_add_txs_skb(struct mt7915_dev
> *dev, struct mt76_wcid *wcid, int pid,
>  		if (rate.mcs > 11)
>  			goto out;
>  
> -		rate.he_gi = wcid->rate.he_gi;
> +		rate.he_gi = wcid->rate_he_gi;
>  		rate.he_dcm = FIELD_GET(MT_TX_RATE_DCM, txrate);
>  		rate.flags = RATE_INFO_FLAGS_HE_MCS;
>  		break;
>  	default:
> +		WARN_ON_ONCE(true);
>  		goto out;
>  	}

Maybe we can consider switching to use existing fixed-rate method in
debugfs? On th other hand, that's the only way to configure HE MU UL
fixed-rate.

Ryder


  parent reply	other threads:[~2021-11-18 17:06 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-18 16:45 [PATCH 1/8] mt76: mt7915: cache sgi in wcid greearb
2021-11-18 16:45 ` [PATCH 2/8] mt76: mt7915: allow processing TXS for 'NO_SKB' pkt-ids greearb
2021-11-18 16:45 ` [PATCH 3/8] mt76: mt7915: debugfs hook to enable TXS for NO_SKB pkt-ids greearb
2021-11-18 16:45 ` [PATCH 4/8] mt76: mt7915: add note about TXSFM 0x2 greearb
2021-11-18 17:02   ` Ryder Lee
2021-11-18 17:06     ` Ben Greear
2021-11-18 16:45 ` [PATCH 5/8] mt76: mt7915: txfree status to show txcount instead of latency greearb
2021-11-18 16:45 ` [PATCH 6/8] mt76: mt7915: report tx-retries greearb
2021-11-18 17:19   ` Ryder Lee
2021-11-18 17:26     ` Ben Greear
2021-11-18 16:45 ` [PATCH 7/8] mt76: mt7915: add support for tx-overrides greearb
2021-11-19  7:16   ` kernel test robot
2021-11-18 16:45 ` [PATCH 8/8] mt76: mt7915: fix SGI reporting when using tx-overrides greearb
2021-11-18 17:06 ` Ryder Lee [this message]
2021-11-18 17:10   ` [PATCH 1/8] mt76: mt7915: cache sgi in wcid Ben Greear
2021-11-18 17:20     ` Ryder Lee

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=c0a09707cf9bcf1205b60c7a36e5afa2a6b2f39c.camel@mediatek.com \
    --to=ryder.lee@mediatek.com \
    --cc=greearb@candelatech.com \
    --cc=linux-wireless@vger.kernel.org \
    /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).