linux-kernel - Re: [PATCH] dmaengine: ti: k3-udma: Prioritize CSI RX traffic as RT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <ed0bb435-c7a0-490e-a1b4-3de05fdf4434@gmail.com>
Date: Tue, 3 Sep 2024 21:06:54 +0300
From: Péter Ujfalusi <peter.ujfalusi@...il.com>
To: Jai Luthra <j-luthra@...com>, Vinod Koul <vkoul@...nel.org>,
 Vignesh Raghavendra <vigneshr@...com>
Cc: dmaengine@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] dmaengine: ti: k3-udma: Prioritize CSI RX traffic as RT



On 02/09/2024 23:24, Péter Ujfalusi wrote:
> Hi,
> 
> On 27/08/2024 15:43, Jai Luthra wrote:
>> From: Vignesh Raghavendra <vigneshr@...com>
>>
>> Mark BCDMA CSI RX as real-time traffic using OrderID 8/15.
>> This ensures CSI traffic takes dedicated RT path towards DDR ensuring
>> proper priority when under competing system load.
>>
>> Signed-off-by: Vignesh Raghavendra <vigneshr@...com>
>> Signed-off-by: Jai Luthra <j-luthra@...com>
>> ---
>>  drivers/dma/ti/k3-udma.c | 9 +++++++++
>>  1 file changed, 9 insertions(+)
>>
>> diff --git a/drivers/dma/ti/k3-udma.c b/drivers/dma/ti/k3-udma.c
>> index 406ee199c2ac..74cdb9ec07c3 100644
>> --- a/drivers/dma/ti/k3-udma.c
>> +++ b/drivers/dma/ti/k3-udma.c
>> @@ -135,6 +135,7 @@ struct udma_match_data {
>>  	u32 statictr_z_mask;
>>  	u8 burst_size[3];
>>  	struct udma_soc_data *soc_data;
>> +	u8 order_id;
> 
> I would add a new property to the BCDM in DT, like ti,order_id to be
> configurable by device and boards if needed.

Or should the order_id be per channel configurable? Audio is handled by
generic BCDMA and might need higher priority to avoid starvation.

But how this affect other none navss DMAs, like display?

> Static 8 and 15 in code is not too nice and begs for a question why 8
> here and why 15 there...
> 
> Even if the 'defaults' in code are these magic ones, it is still better
> to have means to adjust it without the need to recompile the kernel.
> 
>>  };
>>  
>>  struct udma_soc_data {
>> @@ -2110,6 +2111,7 @@ static int udma_tisci_rx_channel_config(struct udma_chan *uc)
>>  static int bcdma_tisci_rx_channel_config(struct udma_chan *uc)
>>  {
>>  	struct udma_dev *ud = uc->ud;
>> +	const struct udma_match_data *match_data = ud->match_data;
>>  	struct udma_tisci_rm *tisci_rm = &ud->tisci_rm;
>>  	const struct ti_sci_rm_udmap_ops *tisci_ops = tisci_rm->tisci_udmap_ops;
>>  	struct udma_rchan *rchan = uc->rchan;
>> @@ -2120,6 +2122,11 @@ static int bcdma_tisci_rx_channel_config(struct udma_chan *uc)
>>  	req_rx.nav_id = tisci_rm->tisci_dev_id;
>>  	req_rx.index = rchan->id;
>>  
>> +	if (match_data->order_id) {
>> +		req_rx.valid_params |= TI_SCI_MSG_VALUE_RM_UDMAP_CH_ORDER_ID_VALID;
>> +		req_rx.rx_orderid = match_data->order_id;
>> +	}
>> +
>>  	ret = tisci_ops->rx_ch_cfg(tisci_rm->tisci, &req_rx);
>>  	if (ret)
>>  		dev_err(ud->dev, "rchan%d cfg failed %d\n", rchan->id, ret);
>> @@ -4332,6 +4339,7 @@ static struct udma_match_data am62a_bcdma_csirx_data = {
>>  		0, /* No UH Channels */
>>  	},
>>  	.soc_data = &am62a_dmss_csi_soc_data,
>> +	.order_id = 8,
>>  };
>>  
>>  static struct udma_match_data am64_bcdma_data = {
>> @@ -4370,6 +4378,7 @@ static struct udma_match_data j721s2_bcdma_csi_data = {
>>  		0, /* No UH Channels */
>>  	},
>>  	.soc_data = &j721s2_bcdma_csi_soc_data,
>> +	.order_id = 15,
>>  };
>>  
>>  static const struct of_device_id udma_of_match[] = {
>>
>> ---
>> base-commit: 6f923748057a4f6aa187e0d5b22990d633a48d12
>> change-id: 20240827-csi_rt-fc6bff701f81
>>
>> Best regards,
> 

-- 
Péter