On Sun, Feb 04, 2024 at 06:21:06PM +0200, [email protected] wrote:
> Use `ret = dev_err_probe(...);` pattern for all messages in dspi_request_dma()
> for the sake of uniforming them. While at it, fix indentation issue reported
> by Vladimir Oltean.
When did I do that? This is v1.
>
> Signed-off-by: Andy Shevchenko <[email protected]>
> ---
> drivers/spi/spi-fsl-dspi.c | 15 +++++----------
> 1 file changed, 5 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/spi/spi-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
> index c9eae046f66c..e0832f1f10bd 100644
> --- a/drivers/spi/spi-fsl-dspi.c
> +++ b/drivers/spi/spi-fsl-dspi.c
> @@ -502,15 +502,12 @@ static int dspi_request_dma(struct fsl_dspi *dspi, phys_addr_t phy_addr)
> return -ENOMEM;
>
> dma->chan_rx = dma_request_chan(dev, "rx");
> - if (IS_ERR(dma->chan_rx)) {
> - return dev_err_probe(dev, PTR_ERR(dma->chan_rx),
> - "rx dma channel not available\n");
> - }
> + if (IS_ERR(dma->chan_rx))
> + return dev_err_probe(dev, PTR_ERR(dma->chan_rx), "rx dma channel not available\n");
>
> dma->chan_tx = dma_request_chan(dev, "tx");
> if (IS_ERR(dma->chan_tx)) {
> - ret = PTR_ERR(dma->chan_tx);
> - dev_err_probe(dev, ret, "tx dma channel not available\n");
> + ret = dev_err_probe(dev, PTR_ERR(dma->chan_tx), "tx dma channel not available\n");
> goto err_tx_channel;
> }
>
> @@ -541,16 +538,14 @@ static int dspi_request_dma(struct fsl_dspi *dspi, phys_addr_t phy_addr)
> cfg.direction = DMA_DEV_TO_MEM;
> ret = dmaengine_slave_config(dma->chan_rx, &cfg);
> if (ret) {
> - dev_err(dev, "can't configure rx dma channel\n");
> - ret = -EINVAL;
> + ret = dev_err_probe(dev, -EINVAL, "can't configure rx dma channel\n");
Passing -EINVAL to dev_err_probe() here doesn't work. It overwrites the "ret"
from dmaengine_slave_config().
int dev_err_probe(const struct device *dev, int err, const char *fmt, ...)
{
struct va_format vaf;
va_list args;
va_start(args, fmt);
vaf.fmt = fmt;
vaf.va = &args;
if (err != -EPROBE_DEFER) { // <-------- always true
dev_err(dev, "error %pe: %pV", ERR_PTR(err), &vaf);
} else {
device_set_deferred_probe_reason(dev, &vaf);
dev_dbg(dev, "error %pe: %pV", ERR_PTR(err), &vaf);
}
va_end(args);
return err;
}
> goto err_slave_config;
> }
>
> cfg.direction = DMA_MEM_TO_DEV;
> ret = dmaengine_slave_config(dma->chan_tx, &cfg);
> if (ret) {
> - dev_err(dev, "can't configure tx dma channel\n");
> - ret = -EINVAL;
> + ret = dev_err_probe(dev, -EINVAL, "can't configure tx dma channel\n");
Same here.
> goto err_slave_config;
> }
>
> --
> 2.43.0
>
>
On Sun, Feb 04, 2024 at 08:24:17PM +0200, Vladimir Oltean wrote:
> > @@ -541,16 +538,14 @@ static int dspi_request_dma(struct fsl_dspi *dspi, phys_addr_t phy_addr)
> > cfg.direction = DMA_DEV_TO_MEM;
> > ret = dmaengine_slave_config(dma->chan_rx, &cfg);
> > if (ret) {
> > - dev_err(dev, "can't configure rx dma channel\n");
> > - ret = -EINVAL;
> > + ret = dev_err_probe(dev, -EINVAL, "can't configure rx dma channel\n");
>
> Passing -EINVAL to dev_err_probe() here doesn't work. It overwrites the "ret"
> from dmaengine_slave_config().
>
> int dev_err_probe(const struct device *dev, int err, const char *fmt, ...)
> {
> struct va_format vaf;
> va_list args;
>
> va_start(args, fmt);
> vaf.fmt = fmt;
> vaf.va = &args;
>
> if (err != -EPROBE_DEFER) { // <-------- always true
> dev_err(dev, "error %pe: %pV", ERR_PTR(err), &vaf);
> } else {
> device_set_deferred_probe_reason(dev, &vaf);
> dev_dbg(dev, "error %pe: %pV", ERR_PTR(err), &vaf);
> }
>
> va_end(args);
>
> return err;
> }
Ah, the original code also ignores the dmaengine_slave_config() return
code and replaces it with -EINVAL? I wonder why that is... It doesn't
appear to be a widespread pattern. Pretty arbitrary. Could you please
make 2 patches, one which preserves the original return code and another
which uses dev_err_probe()?
On Sun, Feb 4, 2024 at 8:24 PM Vladimir Oltean <[email protected]> wrote:
> On Sun, Feb 04, 2024 at 06:21:06PM +0200, [email protected] wrote:
> > Use `ret = dev_err_probe(...);` pattern for all messages in dspi_request_dma()
> > for the sake of uniforming them. While at it, fix indentation issue reported
> > by Vladimir Oltean.
>
> When did I do that? This is v1.
In the original submission (v2 of it to be precise) by Minjie.
..
> > ret = dmaengine_slave_config(dma->chan_rx, &cfg);
> > if (ret) {
> > - dev_err(dev, "can't configure rx dma channel\n");
> > - ret = -EINVAL;
> > + ret = dev_err_probe(dev, -EINVAL, "can't configure rx dma channel\n");
>
> Passing -EINVAL to dev_err_probe() here doesn't work. It overwrites the "ret"
> from dmaengine_slave_config().
True, but this patch doesn't change the behaviour.
..
> > if (ret) {
> > - dev_err(dev, "can't configure tx dma channel\n");
> > - ret = -EINVAL;
> > + ret = dev_err_probe(dev, -EINVAL, "can't configure tx dma channel\n");
>
> Same here.
Same answer here.
> > goto err_slave_config;
> > }
--
With Best Regards,
Andy Shevchenko
On Sun, Feb 4, 2024 at 8:31 PM Vladimir Oltean <[email protected]> wrote:
> On Sun, Feb 04, 2024 at 08:24:17PM +0200, Vladimir Oltean wrote:
..
> > Passing -EINVAL to dev_err_probe() here doesn't work. It overwrites the "ret"
> > from dmaengine_slave_config().
> Ah, the original code also ignores the dmaengine_slave_config() return
> code and replaces it with -EINVAL? I wonder why that is... It doesn't
> appear to be a widespread pattern. Pretty arbitrary. Could you please
> make 2 patches, one which preserves the original return code and another
> which uses dev_err_probe()?
Sure.
--
With Best Regards,
Andy Shevchenko