RE: [RFC PATCH v1] devres: align devres.data strictly only for devm_kmalloc()

[Alexey Brodkin]

Hi Marc,

We sort of expected something like that to happen at some point.
Funny enough it's been a year since my change was accepted in v4.20
and only now somebody noticed :)

Though quite a few questions below.

> Commit a66d972465d15 ("devres: Align data[] to ARCH_KMALLOC_MINALIGN")
> increased the alignment of devres.data unconditionally.
> 
> Some platforms have very strict alignment requirements for DMA-safe
> addresses, e.g. 128 bytes on arm64. There, struct devres amounts to:
> 	3 pointers + pad_to_128 + data + pad_to_256
> i.e. ~220 bytes of padding.

Could you please elaborate a bit on mentioned paddings?
I may understand the first one for 128 bytes but where does the
second one for 256 bytes come from?

> Let's enforce the alignment only for devm_kmalloc().

Ok so for devm_kmalloc() we don't change anything, right?
We still add the same padding before real data array.

> ---
> I had not been aware that dynamic allocation granularity on arm64 was
> 128 bytes. This means there's a lot of waste on small allocations.

Now probably I'm missing something but when do you expect to save something?
If those smaller allocations are done with devm_kmalloc() you aren't
saving anything.

> I suppose there's no easy solution, though.

Right! It took a while till I was able to propose something
people [almost silently] agreed with.

> ---
>  drivers/base/devres.c | 23 +++++++++++++----------
>  1 file changed, 13 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/base/devres.c b/drivers/base/devres.c
> index 0bbb328bd17f..bf39188613d9 100644
> --- a/drivers/base/devres.c
> +++ b/drivers/base/devres.c
> @@ -26,14 +26,7 @@ struct devres_node {
> 
>  struct devres {
>  	struct devres_node		node;
> -	/*
> -	 * Some archs want to perform DMA into kmalloc caches
> -	 * and need a guaranteed alignment larger than
> -	 * the alignment of a 64-bit integer.
> -	 * Thus we use ARCH_KMALLOC_MINALIGN here and get exactly the same
> -	 * buffer alignment as if it was allocated by plain kmalloc().
> -	 */
> -	u8 __aligned(ARCH_KMALLOC_MINALIGN) data[];
> +	u8				data[];
>  };
> 
>  struct devres_group {
> @@ -789,9 +782,16 @@ static void devm_kmalloc_release(struct device *dev, void *res)
>  	/* noop */
>  }
> 
> +#define DEVM_KMALLOC_PADDING_SIZE \
> +	(ARCH_KMALLOC_MINALIGN - sizeof(struct devres) % ARCH_KMALLOC_MINALIGN)

Even given your update with:
------------------------------->8--------------------------------
#define DEVM_KMALLOC_PADDING_SIZE \
  ((ARCH_KMALLOC_MINALIGN - sizeof(struct devres)) % ARCH_KMALLOC_MINALIGN)
------------------------------->8--------------------------------
I don't think I understand why do you need that "% ARCH_KMALLOC_MINALIGN" part?

>  static int devm_kmalloc_match(struct device *dev, void *res, void *data)
>  {
> -	return res == data;
> +	/*
> +	 * 'res' is dr->data (not DMA-safe)
> +	 * 'data' is the hand-aligned address from devm_kmalloc
> +	 */
> +	return res + DEVM_KMALLOC_PADDING_SIZE == data;
>  }
> 
>  /**
> @@ -811,6 +811,9 @@ void * devm_kmalloc(struct device *dev, size_t size, gfp_t gfp)
>  {
>  	struct devres *dr;
> 
> +	/* Add enough padding to provide a DMA-safe address */
> +	size += DEVM_KMALLOC_PADDING_SIZE;

This implementation gets ugly and potentially will lead to problems later
when people will start changing code here. Compared to that initially aligned by
the compiler dr->data looks much more foolproof.

>  	/* use raw alloc_dr for kmalloc caller tracing */
>  	dr = alloc_dr(devm_kmalloc_release, size, gfp, dev_to_node(dev));
>  	if (unlikely(!dr))
> @@ -822,7 +825,7 @@ void * devm_kmalloc(struct device *dev, size_t size, gfp_t gfp)
>  	 */
>  	set_node_dbginfo(&dr->node, "devm_kzalloc_release", size);
>  	devres_add(dev, dr->data);
> -	return dr->data;
> +	return dr->data + DEVM_KMALLOC_PADDING_SIZE;

Ditto. But first I'd like to understand what are you trying to really do
with your change and then we'll see if there could be any better implementation.

-Alexey