The barrier() was intended to reduce possibility of racing with the interrupt handler, but driver's code evolved significantly and today's driver enables interrupt only when it waits for completion notification. Hence barrier() has no good use anymore, let's remove it. Reviewed-by: Michał Mirosław <mirq-linux@xxxxxxxxxxxx> Signed-off-by: Dmitry Osipenko <digetx@xxxxxxxxx> --- drivers/i2c/busses/i2c-tegra.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) diff --git a/drivers/i2c/busses/i2c-tegra.c b/drivers/i2c/busses/i2c-tegra.c index 29e096422b95..2393f52fc584 100644 --- a/drivers/i2c/busses/i2c-tegra.c +++ b/drivers/i2c/busses/i2c-tegra.c @@ -795,18 +795,17 @@ static int tegra_i2c_fill_tx_fifo(struct tegra_i2c_dev *i2c_dev) words_to_transfer = tx_fifo_avail; /* - * Update state before writing to FIFO. If this casues us - * to finish writing all bytes (AKA buf_remaining goes to 0) we - * have a potential for an interrupt (PACKET_XFER_COMPLETE is - * not maskable). We need to make sure that the isr sees - * buf_remaining as 0 and doesn't call us back re-entrantly. + * Update state before writing to FIFO. Note that this may + * cause us to finish writing all bytes (AKA buf_remaining + * goes to 0), hence we have a potential for an interrupt + * (PACKET_XFER_COMPLETE is not maskable), but GIC interrupt + * is disabled at this point. */ buf_remaining -= words_to_transfer * BYTES_PER_FIFO_WORD; tx_fifo_avail -= words_to_transfer; i2c_dev->msg_buf_remaining = buf_remaining; i2c_dev->msg_buf = buf + words_to_transfer * BYTES_PER_FIFO_WORD; - barrier(); i2c_writesl(i2c_dev, buf, I2C_TX_FIFO, words_to_transfer); @@ -827,10 +826,8 @@ static int tegra_i2c_fill_tx_fifo(struct tegra_i2c_dev *i2c_dev) memcpy(&val, buf, buf_remaining); val = le32_to_cpu(val); - /* Again update before writing to FIFO to make sure isr sees. */ i2c_dev->msg_buf_remaining = 0; i2c_dev->msg_buf = NULL; - barrier(); i2c_writel(i2c_dev, val, I2C_TX_FIFO); } -- 2.27.0
