On Mon Mar 30, 2026 at 7:24 PM CEST, Marek Vasut wrote: > On 3/30/26 3:34 PM, Markus Schneider-Pargmann wrote: > > Hello Markus, > >> On Mon Mar 16, 2026 at 12:57 AM CET, Marek Vasut wrote: >>> Rework dev_phys_to_bus() and dev_bus_to_phys() to respect the size >>> of the area specified in dma-ranges DT property. The area outside >>> of ranges is remapped 1:1, while the area in the ranges is remapped >>> according to the description in the dma-ranges property. >>> >>> Adjust the test to test the area within the remapped range, not area >>> outside the remapped range, which was incorrect. >>> >> >> Should this have a Fixes tag? >> >>> Signed-off-by: Marek Vasut <[email protected]> >> >> Not a big expert, but this looks good to me: >> >> Reviewed-by: Markus Schneider-Pargmann <[email protected]> > > This was broken long enough, so this should go into next so it would get > sufficient testing before the next release. > > [...] > >>> +++ b/include/dm/device.h >>> @@ -166,8 +166,9 @@ enum { >>> * When CONFIG_DEVRES is enabled, devm_kmalloc() and >>> friends will >>> * add to this list. Memory so-allocated will be freed >>> * automatically when the device is removed / unbound >>> - * @dma_offset: Offset between the physical address space (CPU's) and the >>> - * device's bus address space >>> + * @dma_cpu: DMA physical address space (CPU's) >>> + * @dma_bus: DMA device's bus address space >>> + * @dma_size: DMA window size >>> * @iommu: IOMMU device associated with this device >>> */ >>> struct udevice { >>> @@ -196,7 +197,9 @@ struct udevice { >>> struct list_head devres_head; >>> #endif >>> #if CONFIG_IS_ENABLED(DM_DMA) >>> - ulong dma_offset; >>> + phys_addr_t dma_cpu; >>> + dma_addr_t dma_bus; >>> + u64 dma_size; >>> #endif > > [...] > >>> +++ b/include/phys2bus.h >>> @@ -21,17 +21,23 @@ static inline unsigned long bus_to_phys(unsigned long >>> bus) >>> } >>> #endif >>> >>> -#if CONFIG_IS_ENABLED(DM) >>> +#if CONFIG_IS_ENABLED(DM_DMA) >> >> This is not related to the code changes right? > > It is, see the matching hunk above and below why this is guarded by DM_DMA .
Ah thanks, yes, the macro had the check for DM_DMA, I missed that.
Best
Markus
>
>>> #include <dm/device.h>
>>>
>>> static inline dma_addr_t dev_phys_to_bus(struct udevice *dev, phys_addr_t
>>> phys)
>>> {
>>> - return phys - dev_get_dma_offset(dev);
>>> + /* If the PA is in the remap range, apply remap. */
>>> + if (phys >= dev->dma_cpu && phys < dev->dma_cpu + dev->dma_size)
>>> + phys -= dev->dma_cpu - dev->dma_bus;
>>> + return phys;
> [...]
signature.asc
Description: PGP signature
