[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <6b7dd983-c948-4c54-b221-4dbf4a2f1bee@kernel.org>
Date: Thu, 31 Oct 2024 12:45:51 +0200
From: Roger Quadros <rogerq@...nel.org>
To: Maxime Chevallier <maxime.chevallier@...tlin.com>
Cc: Andrew Lunn <andrew+netdev@...n.ch>, "David S. Miller"
<davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
Alexei Starovoitov <ast@...nel.org>, Daniel Borkmann <daniel@...earbox.net>,
Jesper Dangaard Brouer <hawk@...nel.org>,
John Fastabend <john.fastabend@...il.com>, Simon Horman <horms@...nel.org>,
Vignesh Raghavendra <vigneshr@...com>,
Siddharth Vadapalli <s-vadapalli@...com>,
Md Danish Anwar <danishanwar@...com>,
Govindarajan Sriramakrishnan <srk@...com>, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, bpf@...r.kernel.org
Subject: Re: [PATCH net v2] net: ethernet: ti: am65-cpsw: Fix multi queue Rx
on J7
Hello Maxime,
On 30/10/2024 20:17, Maxime Chevallier wrote:
> Hello Roger,
>
> On Wed, 30 Oct 2024 15:53:58 +0200
> Roger Quadros <rogerq@...nel.org> wrote:
>
>> On J7 platforms, setting up multiple RX flows was failing
>> as the RX free descriptor ring 0 is shared among all flows
>> and we did not allocate enough elements in the RX free descriptor
>> ring 0 to accommodate for all RX flows.
>>
>> This issue is not present on AM62 as separate pair of
>> rings are used for free and completion rings for each flow.
>>
>> Fix this by allocating enough elements for RX free descriptor
>> ring 0.
>>
>> However, we can no longer rely on desc_idx (descriptor based
>> offsets) to identify the pages in the respective flows as
>> free descriptor ring includes elements for all flows.
>> To solve this, introduce a new swdata data structure to store
>> flow_id and page. This can be used to identify which flow (page_pool)
>> and page the descriptor belonged to when popped out of the
>> RX rings.
>
> [...]
>
>> @@ -339,7 +339,7 @@ static int am65_cpsw_nuss_rx_push(struct am65_cpsw_common *common,
>> struct device *dev = common->dev;
>> dma_addr_t desc_dma;
>> dma_addr_t buf_dma;
>> - void *swdata;
>> + struct am65_cpsw_swdata *swdata;
>
> There's a reverse xmas-tree issue here, where variables should be
> declared from the longest line to the shortest.
Will fix.
>
> [...]
>
>> static void am65_cpsw_nuss_rx_cleanup(void *data, dma_addr_t desc_dma)
>> {
>> - struct am65_cpsw_rx_flow *flow = data;
>> + struct am65_cpsw_rx_chn *rx_chn = data;
>> struct cppi5_host_desc_t *desc_rx;
>> - struct am65_cpsw_rx_chn *rx_chn;
>> + struct am65_cpsw_swdata *swdata;
>> dma_addr_t buf_dma;
>> u32 buf_dma_len;
>> - void *page_addr;
>> - void **swdata;
>> - int desc_idx;
>> + struct page *page;
>> + u32 flow_id;
>
> Here as well
ok.
>
> [...]
>
>> rx_chn->rx_chn = k3_udma_glue_request_rx_chn(dev, "rx", &rx_cfg);
>> @@ -2455,10 +2441,12 @@ static int am65_cpsw_nuss_init_rx_chns(struct am65_cpsw_common *common)
>> flow = &rx_chn->flows[i];
>> flow->id = i;
>> flow->common = common;
>> + flow->irq = -EINVAL;
>
> I've tried to follow the code and I don't get that assignment for the
> irq field, does it really have to do with the current change or is it
> another issue that's being fixed ?
>
> Sorry if I missed the point here.
You are right. This change is unrelated to the subject.
I will split it out into another patch. It is meant to fix a problem in the error path.
>
> Thanks,
>
> Maxime
--
cheers,
-roger
Powered by blists - more mailing lists