[PATCH v3 2/3] dma-mapping: replace zone_dma_bits by zone_dma_limit

[Nathan Chancellor]

On Thu, Aug 01, 2024 at 03:44:54PM +0200, Christoph Hellwig wrote:
> On Wed, Jul 31, 2024 at 06:24:24PM -0700, Nathan Chancellor wrote:
> > Unfortunately, I am not sure either... I do not see anything obviously,
> > so perhaps it could just be avoided with the __diag() infrastructure?
> > 
> > diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c
> > index 3dbc0b89d6fb..b58e7eb9c8f1 100644
> > --- a/kernel/dma/direct.c
> > +++ b/kernel/dma/direct.c
> > @@ -20,7 +20,12 @@
> >   * it for entirely different regions. In that case the arch code needs to
> >   * override the variable below for dma-direct to work properly.
> >   */
> > +__diag_push();
> > +__diag_ignore(clang, 13, "-Wconstant-conversion",
> > +	      "Clang incorrectly thinks the n == 64 case in DMA_BIT_MASK() can happen here,"
> > +	      "which would truncate with a 32-bit phys_addr_t");
> >  phys_addr_t zone_dma_limit __ro_after_init = DMA_BIT_MASK(24);
> 
> So..  The code above is clearly wrong, as DMA_BIT_MASK always returns a
> u64, and phys_addr_t can be smaller than that.  So at least in this case
> the warning seems perfectly valid and the code has issues because it is
> mixing different concepts.

Sure, that seems like a reasonable way to look at things even if the
warning itself is a false positive.

> Where do you see warnings like this upstream?

I don't see this upstream, this is from patch 2 of this series:

https://lore.kernel.org/053fa4806a2c63efcde80caca473a8b670a2701c.1722249878.git.baruch@tkos.co.il/

Cheers,
Nathan