On 11/13/2013 12:17 PM, Lucas Stach wrote:
This corrects the Tegra AC97 clock handling to be in line with the devicetree. Audio PLL is controlled by the machine driver, only AC97 host controller clock remains to be independent.
diff --git a/sound/soc/tegra/tegra20_ac97.c b/sound/soc/tegra/tegra20_ac97.c
@@ -353,7 +352,7 @@ static int tegra20_ac97_platform_probe(struct platform_device *pdev) of_dma, 2) < 0) { dev_err(&pdev->dev, "No DMA resource\n"); ret = -ENODEV;
goto err_clk_put;
goto err;
That's going to conflict (admittedly very slightly) with the patch I sent on Friday to convert to the standard DT DMA bindings[1] since that block of code is removed.
Perhaps you could remove the removal of label err_clk_put from this patch, and keep the rest? Then, I think there won't be any conflict at all. You can send the s/err_clk_put/err/ conversion later; it's logically a separate cleanup anyway, right?
[1] http://www.spinics.net/lists/arm-kernel/msg286785.html
Aside from that and Thierry's comment, Acked-by: Stephen Warren swarren@nvidia.com