Re: [PATCH v2 SPI for-next 3/3] spi: mchp-pci1xxxx: DMA Write Support for copying data from SPI Buf

From: Mark Brown <broonie@kernel.org>
To: Thangaraj Samynathan <thangaraj.s@microchip.com>
Cc: linux-spi@vger.kernel.org, linux-kernel@vger.kernel.org,
	UNGLinuxDriver@microchip.com,
	kumaravel.thiagarajan@microchip.com,
	tharunkumar.pasumarthi@microchip.com
Subject: Re: [PATCH v2 SPI for-next 3/3] spi: mchp-pci1xxxx: DMA Write Support for copying data from SPI Buf
Date: Tue, 6 Feb 2024 12:46:32 +0000	[thread overview]
Message-ID: <ZcIqKP6cu+KThPBg@finisterre.sirena.org.uk> (raw)
In-Reply-To: <20240206034118.748801-4-thangaraj.s@microchip.com>

[-- Attachment #1: Type: text/plain, Size: 868 bytes --]

On Tue, Feb 06, 2024 at 09:11:18AM +0530, Thangaraj Samynathan wrote:

>  	if (regval & SPI_INTR) {
> -		rx_buf = p->rx_buf;
> -		memcpy_fromio(rx_buf + p->bytes_recvd, p->parent->reg_base +
> -				      SPI_MST_RSP_BUF_OFFSET(p->hw_inst), p->tx_sgl_len);
> -		p->bytes_recvd += p->tx_sgl_len;

Oh, so this is actually bidirectional - please merge this into the
previous patch, as noted on the prior patch there are issues with doing
unidirectional DMA.

As a general thing since this is a driver for the SPI controller it
would be *much* easier to tie the terms "read" and "write" to the SPI
bus read and write.  It's pretty hard to follow what's going on as
things stand since it is not clear if a given bit of code is talking
about the SPI controller or the DMA controller, and if when talking
about DMA it's talking about DMA to/from the device or to/from memory.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]