On Mon, Mar 30, 2026 at 10:59:48AM +0200, Marek Szyprowski wrote:
> On 30.03.2026 10:40, Maxime Ripard wrote:
> > Commit 3a236f6a5cf2 ("dma: contiguous: Turn heap registration logic
> > around") didn't remove one last call to dma_heap_cma_register_heap()
> > that it removed, thus breaking the build.
> >
> > That last call is in dma_contiguous_reserve(), to handle the
> > registration of the default CMA region heap instance if it's declared in
> > the device tree.
> >
> > However, the default CMA region instance is already handled by
> > retrieving it through dev_get_cma_area() in the CMA heap driver, so the
> > call to dma_heap_cma_register_heap() wasn't actually needed.
> >
> > Let's remove this call, the now unused function definition, its now
> > empty header, and all includes of this header.
> >
> > Fixes: 3a236f6a5cf2 ("dma: contiguous: Turn heap registration logic around")
> > Reported-by: Mark Brown <broonie@kernel.org>
> > Closes: https://lore.kernel.org/linux-next/acbjaDJ1a-YQC64d@sirena.co.uk/
> > Signed-off-by: Maxime Ripard <mripard@kernel.org>
> > ---
> >  drivers/dma-buf/heaps/cma_heap.c  |  1 -
> >  include/linux/dma-buf/heaps/cma.h | 16 ----------------
> >  kernel/dma/contiguous.c           |  5 -----
> >  3 files changed, 22 deletions(-)
> >
> > diff --git a/drivers/dma-buf/heaps/cma_heap.c b/drivers/dma-buf/heaps/cma_heap.c
> > index f8a3d87f3ccee9630383ba28502eb40b10671cc2..cc517ac68a0bec0788abcb338c03f530d169013b 100644
> > --- a/drivers/dma-buf/heaps/cma_heap.c
> > +++ b/drivers/dma-buf/heaps/cma_heap.c
> > @@ -12,11 +12,10 @@
> >  
> >  #define pr_fmt(fmt) "cma_heap: " fmt
> >  
> >  #include <linux/cma.h>
> >  #include <linux/dma-buf.h>
> > -#include <linux/dma-buf/heaps/cma.h>
> >  #include <linux/dma-heap.h>
> >  #include <linux/dma-map-ops.h>
> >  #include <linux/err.h>
> >  #include <linux/highmem.h>
> >  #include <linux/io.h>
> > diff --git a/include/linux/dma-buf/heaps/cma.h b/include/linux/dma-buf/heaps/cma.h
> > deleted file mode 100644
> > index e751479e21e703e24a5f799b4a7fc8bd0df3c1c4..0000000000000000000000000000000000000000
> > --- a/include/linux/dma-buf/heaps/cma.h
> > +++ /dev/null
> > @@ -1,16 +0,0 @@
> > -/* SPDX-License-Identifier: GPL-2.0 */
> > -#ifndef DMA_BUF_HEAP_CMA_H_
> > -#define DMA_BUF_HEAP_CMA_H_
> > -
> > -struct cma;
> > -
> > -#ifdef CONFIG_DMABUF_HEAPS_CMA
> > -int dma_heap_cma_register_heap(struct cma *cma);
> > -#else
> > -static inline int dma_heap_cma_register_heap(struct cma *cma)
> > -{
> > -	return 0;
> > -}
> > -#endif // CONFIG_DMABUF_HEAPS_CMA
> > -
> > -#endif // DMA_BUF_HEAP_CMA_H_
> > diff --git a/kernel/dma/contiguous.c b/kernel/dma/contiguous.c
> > index ad50512d71d3088a73e4b1ac02d6e6122374888e..9fe001c712339f8388d3f40cca3dfff3f707fcbf 100644
> > --- a/kernel/dma/contiguous.c
> > +++ b/kernel/dma/contiguous.c
> > @@ -40,11 +40,10 @@
> >  #include <asm/page.h>
> >  
> >  #include <linux/memblock.h>
> >  #include <linux/err.h>
> >  #include <linux/sizes.h>
> > -#include <linux/dma-buf/heaps/cma.h>
> >  #include <linux/dma-map-ops.h>
> >  #include <linux/cma.h>
> >  #include <linux/nospec.h>
> >  
> >  #ifdef CONFIG_CMA_SIZE_MBYTES
> > @@ -270,14 +269,10 @@ void __init dma_contiguous_reserve(phys_addr_t limit)
> >  						  selected_limit,
> >  						  &dma_contiguous_default_area,
> >  						  fixed);
> >  		if (ret)
> >  			return;
> > -
> > -		ret = dma_heap_cma_register_heap(dma_contiguous_default_area);
> > -		if (ret)
> > -			pr_warn("Couldn't register default CMA heap.");
>
> After this change no dma-buf heap for the default CMA area is created if it has 
> not been specified in device-tree. This might be especially a problem for the
> non-dt systems.

I don't think that's the case? My understanding is that
dma_contiguous_reserve() is called by the arch code, and will create
that region only if it hasn't be set by the DT (that excerpt above is
run only if !dma_contiguous_default_area).

However, for the DT case, dma_contiguous_default_area will be set by
rmem_cma_setup() a bit below if linux,cma-default is set in the DT.

dma_contiguous_reserved() is called (on arm64 at least) through
bootmem_init(), called as part of setup_arch().

rmem_cma_setup() is called through RESERVEDMEM_OF_DECLARE, so through
__reserved_mem_init_node(), so, if we consider only the public
functions, through:

  * fdt_scan_reserved_mem_reg_nodes(), called by
    unflatten_device_tree(), called right before bootmem_init() in
    setup_arch();

  * or fdt_scan_reserved_mem() and then
    early_init_fdt_scan_reserved_mem(), called in arm64_memblock_init(),
    itself called in setup_arch() earlier than both
    unflatten_device_tree(), and bootmem_init().

Thus, the DT case will run first and set dma_contiguous_default_area if
relevant on that platform, and if it's not set, the non-DT case will set
it up. Either way, dma_contiguous_default_area will be set.

The CMA heap runs much later using a regular module_init. It will
retrieve dma_contiguous_default_area through dev_get_cma_area(), and
will create a heap instance for the default area.

Am I misunderstanding something?
Maxime