Re: [PATCH 2/2] RISC-V: make use of variadic sbi_ecall

From: "Radim Krčmář" <rkrcmar@ventanamicro.com>
To: "David Laight" <david.laight.linux@gmail.com>
Cc: linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
	"Paul Walmsley" <paul.walmsley@sifive.com>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Alexandre Ghiti" <alex@ghiti.fr>,
	"Atish Patra" <atishp@rivosinc.com>,
	"Andrew Jones" <ajones@ventanamicro.com>,
	"Clément Léger" <cleger@rivosinc.com>,
	"Anup Patel" <apatel@ventanamicro.com>
Subject: Re: [PATCH 2/2] RISC-V: make use of variadic sbi_ecall
Date: Fri, 13 Jun 2025 19:08:17 +0200	[thread overview]
Message-ID: <DALKK6MHWD7D.3ZFXSS6YNLHL@ventanamicro.com> (raw)
In-Reply-To: <20250613165226.751541c6@pumpkin>

2025-06-13T16:52:26+01:00, David Laight <david.laight.linux@gmail.com>:
> On Fri, 13 Jun 2025 16:10:52 +0200
> Radim Krčmář <rkrcmar@ventanamicro.com> wrote:
>
> ...
>> The macro will result in the same arguments as before, and it is what
>> the sbi_ecall actually should do.
>
> Ugg...
> Are you using pre-processor 'magic' to add a pile of zeros and then
> select the first 'n' arguments?

Sadly, yes.  The zeros should have never been introduced in the code, so
I'm hiding them for the moment.
I planned to remove the useless zeros in future patches.

> That ought to be banned as error prone.

Sounds reasonable to me.  I will try to fix the insane tracepoint bloat,
and send a v2 that fills only the correct amount of registers for the
ecall instruction.

> I think the one for strncpy() removes the immediate compile error for:
> 	strncpy(dest, src, 1, 2);

I would have preferred if sbi.h had a wrapper function for each SBI
function, and we never allowed direct invocation of the ecall
instruction outside of the library.

_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv