public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Nathan Chancellor <natechancellor@gmail.com>
To: Nick Desaulniers <ndesaulniers@google.com>
Cc: mika.westerberg@linux.intel.com, hartleys@visionengravers.com,
	broonie@kernel.org, linux-spi@vger.kernel.org,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v3] spi: spi-ep93xx: Use dma_data_direction for ep93xx_spi_dma_{finish,prepare}
Date: Fri, 5 Oct 2018 13:46:19 -0700	[thread overview]
Message-ID: <20181005204619.GA19733@flashbox> (raw)
In-Reply-To: <CAKwvOd=yJoQNuVjz3=iDOcaDzwqfPO2qKjZw9OS7CL9nU9oD7g@mail.gmail.com>

On Fri, Oct 05, 2018 at 01:41:19PM -0700, Nick Desaulniers wrote:
> Nathan,
> Thanks for your patience and continued support working towards an
> optimal solution to fix this warning. Some thoughts below.
> 
> On Fri, Oct 5, 2018 at 12:29 PM Nathan Chancellor
> <natechancellor@gmail.com> wrote:
> >
> > Clang warns when one enumerated type is implicitly converted to another.
> >
> > drivers/spi/spi-ep93xx.c:342:62: warning: implicit conversion from
> > enumeration type 'enum dma_transfer_direction' to different enumeration
> > type 'enum dma_data_direction' [-Wenum-conversion]
> >         nents = dma_map_sg(chan->device->dev, sgt->sgl, sgt->nents, dir);
> >                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~
> > ./include/linux/dma-mapping.h:428:58: note: expanded from macro
> > 'dma_map_sg'
> > #define dma_map_sg(d, s, n, r) dma_map_sg_attrs(d, s, n, r, 0)
> >                                ~~~~~~~~~~~~~~~~          ^
> > drivers/spi/spi-ep93xx.c:348:57: warning: implicit conversion from
> > enumeration type 'enum dma_transfer_direction' to different enumeration
> > type 'enum dma_data_direction' [-Wenum-conversion]
> >                 dma_unmap_sg(chan->device->dev, sgt->sgl, sgt->nents, dir);
> >                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~
> > ./include/linux/dma-mapping.h:429:62: note: expanded from macro
> > 'dma_unmap_sg'
> > #define dma_unmap_sg(d, s, n, r) dma_unmap_sg_attrs(d, s, n, r, 0)
> >                                  ~~~~~~~~~~~~~~~~~~          ^
> > drivers/spi/spi-ep93xx.c:377:56: warning: implicit conversion from
> > enumeration type 'enum dma_transfer_direction' to different enumeration
> > type 'enum dma_data_direction' [-Wenum-conversion]
> >         dma_unmap_sg(chan->device->dev, sgt->sgl, sgt->nents, dir);
> >         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~
> > ./include/linux/dma-mapping.h:429:62: note: expanded from macro
> > 'dma_unmap_sg'
> > #define dma_unmap_sg(d, s, n, r) dma_unmap_sg_attrs(d, s, n, r, 0)
> >                                  ~~~~~~~~~~~~~~~~~~          ^
> > 3 warnings generated.
> >
> > dma_{,un}map_sg expect an enum of type dma_data_direction but this
> > driver uses dma_transfer_direction for everything. Convert the driver to
> > use dma_data_direction for these two functions.
> >
> > There are two places that strictly require an enum of type
> > dma_transfer_direction: the direction member in struct dma_slave_config
> > and the direction parameter in dmaengine_prep_slave_sg. To avoid using
> > an explicit cast, add a simple function, ep93xx_dma_data_to_trans_dir,
> > to safely map between the two types because they are not 1 to 1 in
> > meaning.
> >
> > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
> > ---
> >
> > v1 -> v2:
> >
> > * Fix escaped hash symbols for '#define' lines (\# -> #)
> >
> > v2 -> v3:
> >
> > * Instead of changing the dir parameter in the prepare and finish
> >   functions, convert the driver to use dma_data_direction for everywhere
> >   that makes sense. Add a helper function to convert between the two
> >   types as there is still a couple of locations that need a transfer
> >   enum (and no such function exists). Alternatively, the function could
> >   just convert from transfer -> data and be used in dma_{,un}map_sg and
> >   the rest of the driver be left alone. Hopefully I understood what Mika
> >   wanted!
> >
> >  drivers/spi/spi-ep93xx.c | 37 ++++++++++++++++++++++++++-----------
> >  1 file changed, 26 insertions(+), 11 deletions(-)
> >
> > diff --git a/drivers/spi/spi-ep93xx.c b/drivers/spi/spi-ep93xx.c
> > index f1526757aaf6..667da7964d63 100644
> > --- a/drivers/spi/spi-ep93xx.c
> > +++ b/drivers/spi/spi-ep93xx.c
> > @@ -246,6 +246,19 @@ static int ep93xx_spi_read_write(struct spi_master *master)
> >         return -EINPROGRESS;
> >  }
> >
> > +static enum dma_transfer_direction
> > +ep93xx_dma_data_to_trans_dir(enum dma_data_direction dir)
> > +{
> > +       switch (dir) {
> > +       case DMA_TO_DEVICE:
> > +               return DMA_MEM_TO_DEV;
> > +       case DMA_FROM_DEVICE:
> > +               return DMA_DEV_TO_MEM;
> > +       default:
> > +               return DMA_TRANS_NONE;
> > +       }
> > +}
> > +
> >  /**
> >   * ep93xx_spi_dma_prepare() - prepares a DMA transfer
> >   * @master: SPI master
> > @@ -257,7 +270,7 @@ static int ep93xx_spi_read_write(struct spi_master *master)
> >   */
> >  static struct dma_async_tx_descriptor *
> >  ep93xx_spi_dma_prepare(struct spi_master *master,
> > -                      enum dma_transfer_direction dir)
> > +                      enum dma_data_direction dir)
> >  {
> >         struct ep93xx_spi *espi = spi_master_get_devdata(master);
> >         struct spi_transfer *xfer = master->cur_msg->state;
> > @@ -277,9 +290,9 @@ ep93xx_spi_dma_prepare(struct spi_master *master,
> >                 buswidth = DMA_SLAVE_BUSWIDTH_1_BYTE;
> >
> >         memset(&conf, 0, sizeof(conf));
> > -       conf.direction = dir;
> > +       conf.direction = ep93xx_dma_data_to_trans_dir(dir);
> >
> > -       if (dir == DMA_DEV_TO_MEM) {
> > +       if (dir == DMA_FROM_DEVICE) {
> >                 chan = espi->dma_rx;
> >                 buf = xfer->rx_buf;
> >                 sgt = &espi->rx_sgt;
> > @@ -343,7 +356,9 @@ ep93xx_spi_dma_prepare(struct spi_master *master,
> >         if (!nents)
> >                 return ERR_PTR(-ENOMEM);
> >
> > -       txd = dmaengine_prep_slave_sg(chan, sgt->sgl, nents, dir, DMA_CTRL_ACK);
> > +       txd = dmaengine_prep_slave_sg(chan, sgt->sgl, nents,
> > +                                     ep93xx_dma_data_to_trans_dir(dir),
> 
> So we've already calculated this once already in this function, and
> the value of `dir` hasn't changed since.  While the compiler might be
> able to optimize out this common sub expression (CSE or GVN I think
> are the compiler passes that could optimize this out), I think it
> would be better to guarantee that this precalculated value is always
> used by replacing the rematerialization of the value
> (`ep93xx_dma_data_to_trans_dir(dir)`) with `conf.direction`.
> 

Hi Nick,

That's a good point, I will change that for v4 (I'll wait for Mika and
Hartley to chime in before sending it out).

Thanks for the review,
Nathan

> > +                                     DMA_CTRL_ACK);
> >         if (!txd) {
> >                 dma_unmap_sg(chan->device->dev, sgt->sgl, sgt->nents, dir);
> >                 return ERR_PTR(-ENOMEM);
> > @@ -360,13 +375,13 @@ ep93xx_spi_dma_prepare(struct spi_master *master,
> >   * unmapped.
> >   */
> >  static void ep93xx_spi_dma_finish(struct spi_master *master,
> > -                                 enum dma_transfer_direction dir)
> > +                                 enum dma_data_direction dir)
> >  {
> >         struct ep93xx_spi *espi = spi_master_get_devdata(master);
> >         struct dma_chan *chan;
> >         struct sg_table *sgt;
> >
> > -       if (dir == DMA_DEV_TO_MEM) {
> > +       if (dir == DMA_FROM_DEVICE) {
> >                 chan = espi->dma_rx;
> >                 sgt = &espi->rx_sgt;
> >         } else {
> > @@ -381,8 +396,8 @@ static void ep93xx_spi_dma_callback(void *callback_param)
> >  {
> >         struct spi_master *master = callback_param;
> >
> > -       ep93xx_spi_dma_finish(master, DMA_MEM_TO_DEV);
> > -       ep93xx_spi_dma_finish(master, DMA_DEV_TO_MEM);
> > +       ep93xx_spi_dma_finish(master, DMA_TO_DEVICE);
> > +       ep93xx_spi_dma_finish(master, DMA_FROM_DEVICE);
> >
> >         spi_finalize_current_transfer(master);
> >  }
> > @@ -392,15 +407,15 @@ static int ep93xx_spi_dma_transfer(struct spi_master *master)
> >         struct ep93xx_spi *espi = spi_master_get_devdata(master);
> >         struct dma_async_tx_descriptor *rxd, *txd;
> >
> > -       rxd = ep93xx_spi_dma_prepare(master, DMA_DEV_TO_MEM);
> > +       rxd = ep93xx_spi_dma_prepare(master, DMA_FROM_DEVICE);
> >         if (IS_ERR(rxd)) {
> >                 dev_err(&master->dev, "DMA RX failed: %ld\n", PTR_ERR(rxd));
> >                 return PTR_ERR(rxd);
> >         }
> >
> > -       txd = ep93xx_spi_dma_prepare(master, DMA_MEM_TO_DEV);
> > +       txd = ep93xx_spi_dma_prepare(master, DMA_TO_DEVICE);
> >         if (IS_ERR(txd)) {
> > -               ep93xx_spi_dma_finish(master, DMA_DEV_TO_MEM);
> > +               ep93xx_spi_dma_finish(master, DMA_FROM_DEVICE);
> >                 dev_err(&master->dev, "DMA TX failed: %ld\n", PTR_ERR(txd));
> >                 return PTR_ERR(txd);
> >         }
> > --
> > 2.19.0
> >
> 
> 
> -- 
> Thanks,
> ~Nick Desaulniers

  reply	other threads:[~2018-10-05 20:46 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-04  2:39 [PATCH] spi: spi-ep93xx: Change dir type in ep93xx_spi_dma_{finish,prepare} Nathan Chancellor
2018-10-04  2:40 ` [PATCH v2] " Nathan Chancellor
2018-10-04 21:32   ` Nick Desaulniers
2018-10-04 21:47     ` Nathan Chancellor
2018-10-04 23:12       ` Nick Desaulniers
2018-10-05 19:25   ` [PATCH v3] spi: spi-ep93xx: Use dma_data_direction for ep93xx_spi_dma_{finish,prepare} Nathan Chancellor
2018-10-05 20:41     ` Nick Desaulniers
2018-10-05 20:46       ` Nathan Chancellor [this message]
2018-10-08  7:52     ` Mika Westerberg
2018-10-08 18:08     ` [PATCH v4] " Nathan Chancellor
2018-10-08 18:14       ` Nick Desaulniers
2018-10-08 18:53       ` Mika Westerberg
2018-10-08 23:34       ` Applied "spi: spi-ep93xx: Use dma_data_direction for ep93xx_spi_dma_{finish,prepare}" to the spi tree Mark Brown
2018-10-04 10:32 ` [PATCH] spi: spi-ep93xx: Change dir type in ep93xx_spi_dma_{finish,prepare} Mark Brown
2018-10-04 16:04   ` Nathan Chancellor
2018-10-04 16:41     ` Nick Desaulniers
2018-10-05 11:15       ` Mark Brown
2018-10-05  8:28   ` Mika Westerberg
2018-10-05  8:52     ` Nathan Chancellor

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181005204619.GA19733@flashbox \
    --to=natechancellor@gmail.com \
    --cc=broonie@kernel.org \
    --cc=hartleys@visionengravers.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-spi@vger.kernel.org \
    --cc=mika.westerberg@linux.intel.com \
    --cc=ndesaulniers@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox