Messages in this thread Patch in this message | | | Date | Mon, 26 Aug 2013 12:43:06 +0200 (CEST) | From | Guennadi Liakhovetski <> | Subject | Re: [PATCH 0/3] DMA: shdma: several stylistic improvements and support for new SoCs |
| |
Hi Vinod,
Thanks for applying the patches!
On Sun, 25 Aug 2013, Vinod Koul wrote:
> On Tue, Jul 02, 2013 at 05:45:49PM +0200, Guennadi Liakhovetski wrote: > > Hi > > > > The first two patches in this small series improve driver internals a bit > > by using preferred APIs, the 3rd patch adds support for new DMAC versions, > > present on AG5, APE6, H2. > Applied, thanks > > Although the patch 1 didnt apply for me :( I have manually remove a hunk, I dont > which tree yoy genertaed this agaisnt!
I'm not sure why it didn't apply for you. AFAICS, you mean this hunk:
diff --git a/drivers/dma/sh/shdma.c b/drivers/dma/sh/shdma.c index c7950be..3083d62 100644 --- a/drivers/dma/sh/shdma.c +++ b/drivers/dma/sh/shdma.c
[snip]
@@ -909,17 +881,12 @@ static int sh_dmae_remove(struct platform_device *pdev) { struct sh_dmae_device *shdev = platform_get_drvdata(pdev); struct dma_device *dma_dev = &shdev->shdma_dev.dma_dev; - struct resource *res; - int errirq = platform_get_irq(pdev, 0); dma_async_device_unregister(dma_dev); /* Is a NOP if this controller isn't registered with of-dma */ of_dma_controller_free(pdev->dev.of_node); - if (errirq > 0) - free_irq(errirq, shdev); - spin_lock_irq(&sh_dmae_lock); list_del_rcu(&shdev->node); spin_unlock_irq(&sh_dmae_lock); It didn't apply, because it is based on top of "DMA: shdma: also support single top-level DMAC DT nodes" which has been dropped for now in favour of keeping a multiplexer node compulsory for shdma for now, sorry. But just dropping it isn't right too. With managed error IRQ allocation calling free_irq() on clean up path is wrong. I'll post an incremental patch asap.
Thanks Guennadi --- Guennadi Liakhovetski, Ph.D. Freelance Open-Source Software Developer http://www.open-technology.de/
| |