RE: [PATCH v7 1/6] riscv: mm: dma-noncoherent: Switch using function pointers for cache management

[Biju Das]

Hi Christoph Hellwig,

> -----Original Message-----
> From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
> Sent: Tuesday, April 4, 2023 6:29 AM
> To: Prabhakar <prabhakar.csengg@xxxxxxxxx>
> Cc: Arnd Bergmann <arnd@xxxxxxxx>; Conor Dooley
> <conor.dooley@xxxxxxxxxxxxx>; Geert Uytterhoeven <geert+renesas@xxxxxxxxx>;
> Heiko Stuebner <heiko@xxxxxxxxx>; Guo Ren <guoren@xxxxxxxxxx>; Andrew Jones
> <ajones@xxxxxxxxxxxxxxxx>; Paul Walmsley <paul.walmsley@xxxxxxxxxx>; Palmer
> Dabbelt <palmer@xxxxxxxxxxx>; Albert Ou <aou@xxxxxxxxxxxxxxxxx>; Samuel
> Holland <samuel@xxxxxxxxxxxx>; linux-riscv@xxxxxxxxxxxxxxxxxxx; Rob Herring
> <robh+dt@xxxxxxxxxx>; Krzysztof Kozlowski
> <krzysztof.kozlowski+dt@xxxxxxxxxx>; devicetree@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; linux-renesas-soc@xxxxxxxxxxxxxxx; Biju Das
> <biju.das.jz@xxxxxxxxxxxxxx>; Prabhakar Mahadev Lad <prabhakar.mahadev-
> lad.rj@xxxxxxxxxxxxxx>
> Subject: Re: [PATCH v7 1/6] riscv: mm: dma-noncoherent: Switch using
> function pointers for cache management
> 
> On Thu, Mar 30, 2023 at 09:42:12PM +0100, Prabhakar wrote:
> > From: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx>
> >
> > Currently, selecting which CMOs to use on a given platform is done
> > using and ALTERNATIVE_X() macro. This was manageable when there were
> > just two CMO implementations, but now that there are more and more
> > platforms coming needing custom CMOs, the use of the ALTERNATIVE_X() macro
> is unmanageable.
> >
> > To avoid such issues this patch switches to use of function pointers
> > instead of ALTERNATIVE_X() macro for cache management (the only
> > drawback being performance over the previous approach).
> >
> > void (*clean_range)(unsigned long addr, unsigned long size); void
> > (*inv_range)(unsigned long addr, unsigned long size); void
> > (*flush_range)(unsigned long addr, unsigned long size);
> >
> 
> NAK.  Function pointers for somthing high performance as cache maintainance
> is a complete no-go.

Just a question, how does function pointer makes a performance difference compared to
ALTERNATIVE_X() macros?

On both cases, we are pushing function parameters to stack, jumping to the actual routine
And then on return pop the variables from stack. Am I missing something here?

Benchmark results by [1], shows that there is no performance degradation. I am not sure
What type of benchmarking used in this case and How accurate is this benchmark?

https://lore.kernel.org/linux-renesas-soc/40cdea465fef49a8a337b48e2a748138c66a08eb.camel@xxxxxxxxxx/T/#m093c1f3d8f7f0e15bd2909900bf137d5714c553c

Cheers,
Biju