spi: tegra: checking for ERR_PTR instead of NULL
Subsystems (4)
PATCH 1/1 spi: tegra: checking for ERR_PTR instead of NULL
dma_request_slave_channel() returns NULL on error and not ERR_PTRs. I've fixed this by using dma_request_slave_channel_reason() which does return ERR_PTRs.
Fixes a915d150f68d ('spi: tegra: convert to standard DMA DT bindings')
Signed-off-by Dan Carpenter <[email protected]>
drivers/spi/spi-tegra20-slink.c +2 -2
@@ -612,8 +612,8 @@ static int tegra_slink_init_dma_param(struct tegra_slink_data *tspi,int ret;struct dma_slave_config dma_sconfig; - dma_chan =dma_request_slave_channel (tspi->dev, - dma_to_memory ?"rx" :"tx" ); + dma_chan =dma_request_slave_channel_reason (tspi->dev, + dma_to_memory ?"rx" :"tx" );if (IS_ERR (dma_chan)) { ret =PTR_ERR (dma_chan);if (ret != -EPROBE_DEFER )
Stephen Warren
AppliedQuoted 6 lines · Dan Carpenter
dma_request_slave_channel()returns NULL on error and notERR_PTRs. I've fixed this by usingdma_request_slave_channel_reason()which does returnERR_PTRs.Fixes a915d150f68d ('spi: tegra: convert to standard DMA DT bindings')Signed-off-by Dan Carpenter <[email protected]>
Oops. I validate there aren't any more issues like this, at least in the commits in the series this fixes.
I've applied this to Tegra's for-3.14/dmas-resets-rework branch, since that's the only place where the commit this fixes exists.