[PATCH -next V7 0/7] riscv: Optimize function trace

Guo Ren guoren at kernel.org
Thu Feb 9 01:11:24 PST 2023


On Thu, Feb 9, 2023 at 5:00 PM David Laight <David.Laight at aculab.com> wrote:
>
> From: Guo Ren
> > Sent: 09 February 2023 01:51
> ...
> > Yours cost one more instruction, right?
> >          addr-12  auipc
> >          addr-8    jalr
> >          addr-4    // Literal (32-bits)
> >          addr+0   nop or jmp addr-n // one more?
> >          addr+4   function_code
>
> Yes, it is 4 bytes larger but there is one less
> instruction executed (only one nop) when ftrace is disabled.
> That probably matters more than anything in the ftrace
> 'prologue' code.
I've got your point, thx. I would consider your advice and make the tradeoff.

>
> I also suspect that you can use a 32bit integer as
> a table index in 64bit mode to save a word there.
Yes, good idea. I've asked Mark Rutland. Let's wait for his feedback.
It also helps arm64.

>
>         David
>
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)



-- 
Best Regards
 Guo Ren



More information about the linux-riscv mailing list