From: Tudor Ambarus <tudor.ambarus@xxxxxxxxxxxxx> commit e6af9b05bec63cd4d1de2a33968cd0be2a91282a upstream. Cyclic channels must too call issue_pending in order to start a transfer. Start the transfer in issue_pending regardless of the type of channel. This wrongly worked before, because in the past the transfer was started at tx_submit level when only a desc in the transfer list. Fixes: e1f7c9eee707 ("dmaengine: at_xdmac: creation of the atmel eXtended DMA Controller driver") Change-Id: If1bf3e13329cebb9904ae40620f6cf2b7f06fe9f Signed-off-by: Tudor Ambarus <tudor.ambarus@xxxxxxxxxxxxx> Link: https://lore.kernel.org/r/20211215110115.191749-3-tudor.ambarus@xxxxxxxxxxxxx Signed-off-by: Vinod Koul <vkoul@xxxxxxxxxx> Signed-off-by: Mickael GARDET <m.gardet@xxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/dma/at_xdmac.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) --- a/drivers/dma/at_xdmac.c +++ b/drivers/dma/at_xdmac.c @@ -1726,11 +1726,13 @@ static irqreturn_t at_xdmac_interrupt(in static void at_xdmac_issue_pending(struct dma_chan *chan) { struct at_xdmac_chan *atchan = to_at_xdmac_chan(chan); + unsigned long flags; dev_dbg(chan2dev(&atchan->chan), "%s\n", __func__); - if (!at_xdmac_chan_is_cyclic(atchan)) - at_xdmac_advance_work(atchan); + spin_lock_irqsave(&atchan->lock, flags); + at_xdmac_advance_work(atchan); + spin_unlock_irqrestore(&atchan->lock, flags); return; }