[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4124798a-cda5-47fe-a67b-e84d72f3ecf8@kernel.org>
Date: Mon, 1 Jul 2024 13:32:08 +0300
From: Roger Quadros <rogerq@...nel.org>
To: Simon Horman <horms@...nel.org>
Cc: "David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>, Siddharth Vadapalli <s-vadapalli@...com>,
Julien Panis <jpanis@...libre.com>, Andrew Lunn <andrew@...n.ch>,
srk@...com, vigneshr@...com, danishanwar@...com, pekka Varis
<p-varis@...com>, netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-omap@...r.kernel.org
Subject: Re: [PATCH net-next v2 6/7] net: ethernet: ti: cpsw_ale: add helper
to setup classifier defaults
On 01/07/2024 10:35, Simon Horman wrote:
> On Fri, Jun 28, 2024 at 03:01:55PM +0300, Roger Quadros wrote:
>> Default behaviour is to have 8 classifiers to map 8 DSCP/PCP
>> priorities to N receive threads (flows). N depends on number of
>> RX channels enabled for the port.
>>
>> Signed-off-by: Roger Quadros <rogerq@...nel.org>
>> ---
>> drivers/net/ethernet/ti/cpsw_ale.c | 57 ++++++++++++++++++++++++++++++++++++++
>> drivers/net/ethernet/ti/cpsw_ale.h | 1 +
>> 2 files changed, 58 insertions(+)
>>
>> diff --git a/drivers/net/ethernet/ti/cpsw_ale.c b/drivers/net/ethernet/ti/cpsw_ale.c
>> index 75a17184d34c..51da527388df 100644
>> --- a/drivers/net/ethernet/ti/cpsw_ale.c
>> +++ b/drivers/net/ethernet/ti/cpsw_ale.c
>> @@ -1650,3 +1650,60 @@ static void cpsw_ale_policer_thread_idx_enable(struct cpsw_ale *ale, u32 idx,
>> regmap_field_write(ale->fields[ALE_THREAD_VALUE], thread_id);
>> regmap_field_write(ale->fields[ALE_THREAD_ENABLE], enable ? 1 : 0);
>> }
>> +
>> +/* Disable all policer entries and thread mappings */
>> +static void cpsw_ale_policer_reset(struct cpsw_ale *ale)
>> +{
>> + int i;
>> +
>> + for (i = 0; i < ale->params.num_policers ; i++) {
>> + cpsw_ale_policer_read_idx(ale, i);
>> + regmap_field_write(ale->fields[POL_PORT_MEN], 0);
>> + regmap_field_write(ale->fields[POL_PRI_MEN], 0);
>> + regmap_field_write(ale->fields[POL_OUI_MEN], 0);
>> + regmap_field_write(ale->fields[POL_DST_MEN], 0);
>> + regmap_field_write(ale->fields[POL_SRC_MEN], 0);
>> + regmap_field_write(ale->fields[POL_OVLAN_MEN], 0);
>> + regmap_field_write(ale->fields[POL_IVLAN_MEN], 0);
>> + regmap_field_write(ale->fields[POL_ETHERTYPE_MEN], 0);
>> + regmap_field_write(ale->fields[POL_IPSRC_MEN], 0);
>> + regmap_field_write(ale->fields[POL_IPDST_MEN], 0);
>> + regmap_field_write(ale->fields[POL_EN], 0);
>> + regmap_field_write(ale->fields[POL_RED_DROP_EN], 0);
>> + regmap_field_write(ale->fields[POL_YELLOW_DROP_EN], 0);
>> + regmap_field_write(ale->fields[POL_PRIORITY_THREAD_EN], 0);
>> +
>> + cpsw_ale_policer_thread_idx_enable(ale, i, 0, 0);
>> + }
>> +}
>> +
>> +/* Default classifer is to map 8 user priorities to N receive channels */
>> +void cpsw_ale_classifier_setup_default(struct cpsw_ale *ale, int num_rx_ch)
>> +{
>> + int pri, idx;
>> + int pri_thread_map[8][9] = { { 0, 0, 0, 0, 0, 0, 0, 0, },
>> + { 0, 0, 0, 0, 1, 1, 1, 1, },
>> + { 0, 0, 0, 0, 1, 1, 2, 2, },
>> + { 1, 0, 0, 1, 2, 2, 3, 3, },
>> + { 1, 0, 0, 1, 2, 3, 4, 4, },
>> + { 1, 0, 0, 2, 3, 4, 5, 5, },
>> + { 1, 0, 0, 2, 3, 4, 5, 6, },
>> + { 2, 0, 1, 3, 4, 5, 6, 7, } };
>
> Hi Roger,
>
> Perhaps it is obvious, but I'm wondering if it is appropriate
> to add a comment explaining the layout of the table, especially
> the latter rows where the mapping of priority to receive channel
> seems somewhat non-trivial.
Sure. I took the table straight off from the All new switch book. [1]
Priorities 3 to 7 are straight forward. Priorities 0 to 2 are listed like so in
decreasing order of priority
0 (default) Best Effort
2 Spare (undefined)
1 (lowest) Background
[1] Table 13-2 IEEE 802.1p Recommended Priority Mappings to Class of Service.
>
>> +
>> + cpsw_ale_policer_reset(ale);
>> +
>> + /* use first 8 classifiers to map 8 (DSCP/PCP) priorities to channels */
>> + for (pri = 0; pri < 8; pri++) {
>> + idx = pri;
>> +
>> + /* Classifier 'idx' match on priority 'pri' */
>> + cpsw_ale_policer_read_idx(ale, idx);
>> + regmap_field_write(ale->fields[POL_PRI_VAL], pri);
>> + regmap_field_write(ale->fields[POL_PRI_MEN], 1);
>> + cpsw_ale_policer_write_idx(ale, idx);
>> +
>> + /* Map Classifier 'idx' to thread provided by the map */
>> + cpsw_ale_policer_thread_idx_enable(ale, idx,
>> + pri_thread_map[num_rx_ch - 1][pri], 1);
>> + }
>> +}
>> diff --git a/drivers/net/ethernet/ti/cpsw_ale.h b/drivers/net/ethernet/ti/cpsw_ale.h
>> index 2cb76acc6d16..1e4e9a3dd234 100644
>> --- a/drivers/net/ethernet/ti/cpsw_ale.h
>> +++ b/drivers/net/ethernet/ti/cpsw_ale.h
>> @@ -193,5 +193,6 @@ int cpsw_ale_vlan_add_modify(struct cpsw_ale *ale, u16 vid, int port_mask,
>> int cpsw_ale_vlan_del_modify(struct cpsw_ale *ale, u16 vid, int port_mask);
>> void cpsw_ale_set_unreg_mcast(struct cpsw_ale *ale, int unreg_mcast_mask,
>> bool add);
>> +void cpsw_ale_classifier_setup_default(struct cpsw_ale *ale, int num_rx_ch);
>>
>> #endif
>>
>> --
>> 2.34.1
>>
>>
--
cheers,
-roger
Powered by blists - more mailing lists