Skip to contentSkip to conversation

of/platform: Initialise default DMA masks

PATCH New Robin Murphy 3 participants 5 messages Jul 27, 2018 Device tree

lore
Share & tools
lore

Robin Murphy

author
#0 lore raw

PATCH 1/1 of/platform: Initialise default DMA masks

When of_dma_configure() was first born in 591c1ee465ce ("of: configure the platform device dma parameters"), everything DMA-related was factored out of of_platform_device_create_pdata() as seemed appropriate at the time. However, now that of_dma_configure() has grown into the generic handler for processing DMA-related properties from DT for all kinds of devices, it is no longer an appropriate place to be doing OF-platform-specific business. Since there are still plenty of platform drivers not setting their own masks and depending on the bus default, let's reinstate that inialisation in the OF-platform code itself, and restore the long-standing status quo from 0589342c2794 ("of: set dma_mask to point to coherent_dma_mask")

CC Rob Herring <[email protected]>
CC Frank Rowand <[email protected]>
Signed-off-by Robin Murphy <[email protected]>

---

This shouldn't strictly depend on the changes currently queued in the dma-mapping tree, so should be OK to go through the DT tree in parallel. Ideally we'd fix all DMA-capable drivers to set their masks correctly, but in the short term everyone's going to get cross about -rc1 not booting...

Robin.

 drivers/of/platform.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/of/platform.c +3 -0
index 6925d993e1f0..7ba90c290a42 100644
@@ -185,6 +185,9 @@ static struct platform_device *of_platform_device_create_pdata(
 	if (!dev)
 		goto err_clear_flag;
 
+	dev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
+	if (!dev->dev.dma_mask)
+		dev->dev.dma_mask = &dev->dev.coherent_dma_mask;
 	dev->dev.bus = &platform_bus_type;
 	dev->dev.platform_data = platform_data;
 	of_msi_configure(&dev->dev, dev->dev.of_node);
Signature · 1 line
2.17.1.dirty
#1 lore raw

PATCH 1/1 of/platform: Initialise default DMA masks

Quoted 5 lines · Robin Murphy

This shouldn't strictly depend on the changes currently queued in the dma-mapping tree, so should be OK to go through the DT tree in parallel. Ideally we'd fix all DMA-capable drivers to set their masks correctly, but in the short term everyone's going to get cross about -rc1 not booting...

I'd prefer to pick this up through the dma-mapping tree, and in fact move it before the bus_dma_mask changes just to keep a bisectable tree.

author
#2 lore raw

PATCH 1/1 of/platform: Initialise default DMA masks

Quoted 10 lines · Christoph Hellwig
Quoted 5 lines · Robin Murphy

This shouldn't strictly depend on the changes currently queued in the dma-mapping tree, so should be OK to go through the DT tree in parallel. Ideally we'd fix all DMA-capable drivers to set their masks correctly, but in the short term everyone's going to get cross about -rc1 not booting...

I'd prefer to pick this up through the dma-mapping tree, and in fact move it before the bus_dma_mask changes just to keep a bisectable tree.

Sure, if Rob and Frank are OK with that I don't mind either way.

Thanks,
Robin.

#3 lore raw

PATCH 1/1 of/platform: Initialise default DMA masks

Quoted 9 lines · Robin Murphy
Quoted 7 lines · Christoph Hellwig
Quoted 5 lines · Robin Murphy

This shouldn't strictly depend on the changes currently queued in the dma-mapping tree, so should be OK to go through the DT tree in parallel. Ideally we'd fix all DMA-capable drivers to set their masks correctly, but in the short term everyone's going to get cross about -rc1 not booting...

Quoted 5 lines
Quoted 3 lines

I'd prefer to pick this up through the dma-mapping tree, and in fact move it before the bus_dma_mask changes just to keep a bisectable tree.

Sure, if Rob and Frank are OK with that I don't mind either way.

NP.

Acked-by Rob Herring <[email protected]>
#4 lore raw

PATCH 1/1 of/platform: Initialise default DMA masks

Quoted 14 lines · Rob Herring
Quoted 13 lines · Robin Murphy
Quoted 10 lines · Christoph Hellwig
Quoted 5 lines · Robin Murphy

This shouldn't strictly depend on the changes currently queued in the dma-mapping tree, so should be OK to go through the DT tree in parallel. Ideally we'd fix all DMA-capable drivers to set their masks correctly, but in the short term everyone's going to get cross about -rc1 not booting...

I'd prefer to pick this up through the dma-mapping tree, and in fact move it before the bus_dma_mask changes just to keep a bisectable tree.

Quoted 5 lines
Quoted 1 line

Sure, if Rob and Frank are OK with that I don't mind either way.

NP.

Acked-by Rob Herring <[email protected]>

Ok, applied so that it gets into the next linux-next.