Re: [PATCH v6 1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Incorrect memory barrier used before writing to I/O port in `outsq()`.
--

commit 8166082f73bb766f324d8ad444fbd61adbfe4558
Author: Yunhui Cui <cuiyunhui@bytedance.com>
riscv: io: avoid null-pointer arithmetic in PIO helpers

This commit restricts port I/O string helpers to only be available when
CONFIG_HAS_IOPORT is enabled. This prevents clang from diagnosing null-pointer
arithmetic from the PCI_IOBASE expression when port I/O is not supported.

> diff --git a/arch/riscv/include/asm/io.h b/arch/riscv/include/asm/io.h
> index 09bb5f57a9d34..92d5f831f3495 100644
> --- a/arch/riscv/include/asm/io.h
> +++ b/arch/riscv/include/asm/io.h
> @@ -116,26 +118,32 @@ __io_writes_outs(writes, u32, l, __io_bw(), __io_aw())

[ ... ]

>  __io_writes_outs(writes, u64, q, __io_bw(), __io_aw())
>  #define writesq(addr, buffer, count) __writesq(addr, buffer, count)
>  
> +#ifdef CONFIG_HAS_IOPORT
>  __io_writes_outs(outs, u64, q, __io_pbr(), __io_paw())

[Severity: High]
This is a pre-existing issue, but does outsq() use the wrong memory barrier?

While outsb(), outsw(), and outsl() correctly pass __io_pbw() as their bfence
parameter, outsq() passes __io_pbr(). Because __io_pbr() expands to a read
barrier instead of a write barrier, could a weakly ordered CPU reorder
earlier memory writes after the I/O port writes?

>  #define outsq(addr, buffer, count) __outsq(PCI_IOBASE + (addr), buffer, count)
>  #endif
> +#endif

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260701122948.57584-1-cuiyunhui@bytedance.com?part=1