Kernel KVM virtualization development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: JinRui <jinrui@haiwei.tech>
Cc: kvm@vger.kernel.org
Subject: Re: [PATCH v12] KVM: selftests: riscv: Add lazy V extension enablement for guests
Date: Thu, 13 Aug 2026 10:04:36 +0000	[thread overview]
Message-ID: <20260813100437.06E621F000E9@smtp.kernel.org> (raw)
In-Reply-To: <7C0512E558D1614D+20260813095615.3843757-1-jinrui@haiwei.tech>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Unaligned 32-bit memory access when decoding 32-bit instructions
--

--- Patch [1]: [PATCH v12] KVM: selftests: riscv: Add lazy V extension enablement for guests ---
commit cffe485ad3478e658853f5e74fe97f4c03d20bf6
Author: jinrui <jinrui@haiwei.tech>

KVM: selftests: riscv: Add lazy V extension enablement for guests

This commit introduces lazy enablement of the RISC-V vector (V) extension
for guest tests. It updates the exception routing to decode faulting
instructions and re-execute vector instructions when the VS status is off.

> diff --git a/tools/testing/selftests/kvm/lib/riscv/processor.c b/tools/testing/selftests/kvm/lib/riscv/processor.c
> index ded5429f34483..e677137d5e448 100644
> --- a/tools/testing/selftests/kvm/lib/riscv/processor.c
> +++ b/tools/testing/selftests/kvm/lib/riscv/processor.c

[ ... ]

> @@ -432,6 +472,37 @@ void route_exception(struct pt_regs *regs)
>  		ec = 0;
>  	}
>  
> +	/*
> +	 * Lazily enable V on the first vector instruction: if the faulting
> +	 * instruction decodes as vector while VS is off, set VS to Initial
> +	 * and re-execute it, like the kernel's riscv_v_first_use_handler().
> +	 * Genuinely illegal instructions continue to the unexpected-exception
> +	 * path.
> +	 */
> +	if (!(regs->cause & CAUSE_IRQ_FLAG) && ec == EXC_INST_ILLEGAL &&
> +	    handlers && handlers->v_available && !(regs->status & SR_VS)) {
> +		u32 insn = (u32)regs->badaddr;
> +
> +		/*
> +		 * stval is not guaranteed to hold the faulting instruction.
> +		 * Vector instructions are always 32-bit, so read a 16-bit
> +		 * halfword first and only load the full 32-bit instruction when
> +		 * it is not compressed; this avoids an unaligned or cross-page
> +		 * access on a compressed instruction.
> +		 */
> +		if (!insn) {
> +			u16 half = *(u16 *)regs->epc;
> +
> +			if ((half & 0x3) == 0x3)
> +				insn = *(u32 *)regs->epc;

[Severity: Medium]
Will this result in an unaligned 32-bit memory access on certain hardware?

If a 32-bit vector instruction is placed at a 2-byte aligned address (which
is valid when the RISC-V 'C' compressed extension is enabled), regs->epc
will only be 2-byte aligned.

Casting a potentially 2-byte aligned address to a u32 pointer is undefined
behavior in C, and the compiler typically emits a standard 32-bit word load
instruction.

On strict-alignment RISC-V hardware, won't this trap and crash the selftest
guest with a 'Load Address Misaligned' exception?

> +		}
> +
> +		if (insn_is_vector(insn)) {
> +			regs->status |= SR_VS_INITIAL;
> +			return;
> +		}
> +	}
> +

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/7C0512E558D1614D+20260813095615.3843757-1-jinrui@haiwei.tech?part=1

  reply	other threads:[~2026-08-13 10:04 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20260722073344.771230-1-jinrui@haiwei.tech>
2026-08-11  8:50 ` [PATCH v9] KVM: selftests: riscv: Add lazy V extension enablement for guests JinRui
2026-08-12 17:26   ` Anup Patel
2026-08-13  9:03   ` [PATCH v10] " JinRui
2026-08-13  9:13     ` sashiko-bot
2026-08-13  9:37     ` [PATCH v11] " JinRui
2026-08-13  9:46       ` sashiko-bot
2026-08-13  9:56       ` [PATCH v12] " JinRui
2026-08-13 10:04         ` sashiko-bot [this message]
2026-08-13 10:29         ` [PATCH v13] " JinRui

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260813100437.06E621F000E9@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=jinrui@haiwei.tech \
    --cc=kvm@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox