[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID:
<SA1PR12MB67987D5966B2D69688015A88C9A8A@SA1PR12MB6798.namprd12.prod.outlook.com>
Date: Thu, 18 Dec 2025 19:06:44 +0000
From: "Gupta, Suraj" <Suraj.Gupta2@....com>
To: Tomi Valkeinen <tomi.valkeinen@...asonboard.com>, Vinod Koul
<vkoul@...nel.org>, "Simek, Michal" <michal.simek@....com>, "Sagar, Vishal"
<vishal.sagar@....com>
CC: "dmaengine@...r.kernel.org" <dmaengine@...r.kernel.org>,
"linux-arm-kernel@...ts.infradead.org"
<linux-arm-kernel@...ts.infradead.org>, "linux-kernel@...r.kernel.org"
<linux-kernel@...r.kernel.org>, Laurent Pinchart
<laurent.pinchart@...asonboard.com>
Subject: RE: [PATCH] dmaengine: xilinx_dma: Add support for residue on direct
AXIDMA S2MM
[AMD Official Use Only - AMD Internal Distribution Only]
> -----Original Message-----
> From: Tomi Valkeinen <tomi.valkeinen@...asonboard.com>
> Sent: Thursday, December 18, 2025 2:10 PM
> To: Vinod Koul <vkoul@...nel.org>; Simek, Michal <michal.simek@....com>;
> Sagar, Vishal <vishal.sagar@....com>
> Cc: dmaengine@...r.kernel.org; linux-arm-kernel@...ts.infradead.org; linux-
> kernel@...r.kernel.org; Laurent Pinchart
> <laurent.pinchart@...asonboard.com>; Tomi Valkeinen
> <tomi.valkeinen@...asonboard.com>
> Subject: [PATCH] dmaengine: xilinx_dma: Add support for residue on direct
> AXIDMA S2MM
>
> Caution: This message originated from an External Source. Use proper caution
> when opening attachments, clicking links, or responding.
>
>
> AXIDMA IP supports reporting the amount of bytes transferred on the S2MM
> channel in direct mode (i.e. non-SG), but the driver does not. Thus the
> driver always reports that all of the buffer was filled.
>
> Add xilinx_dma_get_residue_axidma_direct_s2mm() which gets the residue
> amount for direct AXIDMA for S2MM direction.
>
> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@...asonboard.com>
Looks fine to me.
Reviewed-by: Suraj Gupta <suraj.gupta2@....com>
> ---
> drivers/dma/xilinx/xilinx_dma.c | 21 +++++++++++++++++++++
> 1 file changed, 21 insertions(+)
>
> diff --git a/drivers/dma/xilinx/xilinx_dma.c b/drivers/dma/xilinx/xilinx_dma.c
> index fabff602065f..64b3fba4e44f 100644
> --- a/drivers/dma/xilinx/xilinx_dma.c
> +++ b/drivers/dma/xilinx/xilinx_dma.c
> @@ -1021,6 +1021,24 @@ static u32 xilinx_dma_get_residue(struct
> xilinx_dma_chan *chan,
> return residue;
> }
>
> +static u32
> +xilinx_dma_get_residue_axidma_direct_s2mm(struct xilinx_dma_chan *chan,
> + struct xilinx_dma_tx_descriptor *desc)
> +{
> + struct xilinx_axidma_tx_segment *seg;
> + struct xilinx_axidma_desc_hw *hw;
> + u32 finished_len;
> +
> + finished_len = dma_ctrl_read(chan, XILINX_DMA_REG_BTT);
> +
> + seg = list_first_entry(&desc->segments, struct xilinx_axidma_tx_segment,
> + node);
> +
> + hw = &seg->hw;
> +
> + return hw->control - finished_len;
> +}
> +
> /**
> * xilinx_dma_chan_handle_cyclic - Cyclic dma callback
> * @chan: Driver specific dma channel
> @@ -1732,6 +1750,9 @@ static void xilinx_dma_complete_descriptor(struct
> xilinx_dma_chan *chan)
> if (chan->has_sg && chan->xdev->dma_config->dmatype !=
> XDMA_TYPE_VDMA)
> desc->residue = xilinx_dma_get_residue(chan, desc);
> + else if (chan->xdev->dma_config->dmatype == XDMA_TYPE_AXIDMA
> &&
> + chan->direction == DMA_DEV_TO_MEM && !chan->has_sg)
> + desc->residue =
> xilinx_dma_get_residue_axidma_direct_s2mm(chan, desc);
> else
> desc->residue = 0;
> desc->err = chan->err;
>
> ---
> base-commit: 7d0a66e4bb9081d75c82ec4957c50034cb0ea449
> change-id: 20251218-xilinx-dma-residue-fix-f6e4f97ffe61
>
> Best regards,
> --
> Tomi Valkeinen <tomi.valkeinen@...asonboard.com>
>
Powered by blists - more mailing lists