Thread (17 messages) flat view 17 messages, 5 authors, 2016-07-18

[PATCH 7/8] uio: bind uio_dmem_genirq via OF

From: Anup Patel <hidden>
Date: 2016-07-15 10:47:53
Also in: lkml

On Fri, Jul 15, 2016 at 3:15 PM, Russell King - ARM Linux
[off-list ref] wrote:
On Fri, Jul 15, 2016 at 02:34:02PM +0530, Anup Patel wrote:
quoted
+static int uio_dmem_genirq_alloc_platdata(struct platform_device *pdev)
+{
+     struct uio_dmem_genirq_pdata pdata;
+     u32 dma_bits, regions;
+     u32 sizes[MAX_UIO_MAPS];
+     int ret;
+
+     memset(&pdata, 0, sizeof(pdata));
+
+     ret = of_property_read_u32(pdev->dev.of_node,
+                                uio_of_dma_bits_prop, &dma_bits);
+     if (ret) {
+             dev_err(&pdev->dev,
+                     "Missing property %s\n", uio_of_dma_bits_prop);
+             return ret;
+     }
+     if (dma_bits > 64)
+             dma_bits = 64;
+
+     dma_set_coherent_mask(&pdev->dev, DMA_BIT_MASK(dma_bits));
You really need to check the return value from this: this function
negotiates with the architecture, and if 64-bit DMA is not supported,
then the call will fail and you as a driver are expected to fall back
to 32-bit DMA only.

In that case, you're expected to call the same function with a 32-bit
mask, and if that fails, you're supposed to then decide that DMA is
not possible.
Thanks for pointing out.

I will fix this as-per your suggestion in next revision.

Regards,
Anup
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help