linux-kernel - Re: [PATCH 4/7] spi: spi-geni-qcom: Add support for GPI dma

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 for Android: free password hash cracker in your pocket

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

Date:   Thu, 17 Jun 2021 11:50:55 +0530
From:   Vinod Koul <vkoul@...nel.org>
To:     Mark Brown <broonie@...nel.org>
Cc:     Bjorn Andersson <bjorn.andersson@...aro.org>,
        Wolfram Sang <wsa@...nel.org>, linux-arm-msm@...r.kernel.org,
        Andy Gross <agross@...nel.org>,
        Matthias Kaehlcke <mka@...omium.org>,
        Douglas Anderson <dianders@...omium.org>,
        Sumit Semwal <sumit.semwal@...aro.org>,
        Amit Pundir <amit.pundir@...aro.org>,
        linux-spi@...r.kernel.org, linux-i2c@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 4/7] spi: spi-geni-qcom: Add support for GPI dma

On 16-06-21, 17:32, Vinod Koul wrote:
> On 16-06-21, 12:35, Mark Brown wrote:
> > On Wed, Jun 16, 2021 at 02:20:45PM +0530, Vinod Koul wrote:
> > > But in this case, that wont work. We have a parent qup device which is
> > > the parent for both spi and dma device and needs to be used for
> > > dma-mapping! 
> > 
> > > If we allow drivers to set dma mapping device and use that, then I can
> > > reuse the core. Let me know if that is agreeable to you and I can hack
> > > this up. Maybe add a new member in spi_controller which is filled by
> > > drivers in can_dma() callback?
> > 
> > Possibly, I'd need to see the code.
> 
> Ok, let me do a prototype and share ...

So setting the dma_map_dev in the can_dma() callback does not work as we
would need device before we invoke can_dma(), so modified this to be set
earlier by driver (in driver probe, set the dma_map_dev) and use in
__spi_map_msg().

With this it works for me & I was able to get rid of driver mapping code

-- >8 --

diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index e353b7a9e54e..315f7e7545f7 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -961,11 +961,15 @@ static int __spi_map_msg(struct spi_controller *ctlr, struct spi_message *msg)
 
 	if (ctlr->dma_tx)
 		tx_dev = ctlr->dma_tx->device->dev;
+	else if (ctlr->dma_map_dev)
+		tx_dev = ctlr->dma_map_dev;
 	else
 		tx_dev = ctlr->dev.parent;
 
 	if (ctlr->dma_rx)
 		rx_dev = ctlr->dma_rx->device->dev;
+	else if (ctlr->dma_map_dev)
+		rx_dev = ctlr->dma_map_dev;
 	else
 		rx_dev = ctlr->dev.parent;
 
diff --git a/include/linux/spi/spi.h b/include/linux/spi/spi.h
index 74239d65c7fd..4d3f116f5723 100644
--- a/include/linux/spi/spi.h
+++ b/include/linux/spi/spi.h
@@ -586,6 +586,7 @@ struct spi_controller {
 	bool			(*can_dma)(struct spi_controller *ctlr,
 					   struct spi_device *spi,
 					   struct spi_transfer *xfer);
+	struct device *dma_map_dev;
 
 	/*
 	 * These hooks are for drivers that want to use the generic

-- 
~Vinod