| Message ID | 20210623095942.3325-6-wsa+renesas@sang-engineering.com (mailing list archive) |
|---|---|
| State | Under Review |
| Delegated to: | Geert Uytterhoeven |
| Headers | show |
| Series | i2c: use proper DMAENGINE API for termination | expand |
Hi Wolfram, On Wed, Jun 23, 2021 at 12:01 PM Wolfram Sang <wsa+renesas@sang-engineering.com> wrote: > dmaengine_terminate_all() is deprecated in favor of explicitly saying if > it should be sync or async. Here, we want dmaengine_terminate_sync() > because there is no other synchronization code in the driver to handle > an async case. > > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> Thanks for your patch! Is this safe? The driver is not using a threaded irq, and DMA termination may be called from the interrupt handler. Have you tried triggering DMA termination, with lockdep enabled? Gr{oetje,eeting}s, Geert
> Is this safe? The driver is not using a threaded irq, and DMA termination > may be called from the interrupt handler. You are right, this will not work. Not my best day today, I overlooked it for i2c-rcar and lost the note pointing out the same issue for stm32f7 :( > Have you tried triggering DMA termination, with lockdep enabled? Nope. As the code didn't show signs of async nature, I assumed sync was desired anyhow.
diff --git a/drivers/i2c/busses/i2c-rcar.c b/drivers/i2c/busses/i2c-rcar.c index 327c092a4130..41ff327882ef 100644 --- a/drivers/i2c/busses/i2c-rcar.c +++ b/drivers/i2c/busses/i2c-rcar.c @@ -391,9 +391,9 @@ static void rcar_i2c_cleanup_dma(struct rcar_i2c_priv *priv) if (priv->dma_direction == DMA_NONE) return; else if (priv->dma_direction == DMA_FROM_DEVICE) - dmaengine_terminate_all(priv->dma_rx); + dmaengine_terminate_sync(priv->dma_rx); else if (priv->dma_direction == DMA_TO_DEVICE) - dmaengine_terminate_all(priv->dma_tx); + dmaengine_terminate_sync(priv->dma_tx); rcar_i2c_dma_unmap(priv); }
dmaengine_terminate_all() is deprecated in favor of explicitly saying if it should be sync or async. Here, we want dmaengine_terminate_sync() because there is no other synchronization code in the driver to handle an async case. Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> --- drivers/i2c/busses/i2c-rcar.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)