linux-kernel - Re: [PATCH] RISC-V: KVM: Fix SBI_FWFT_POINTER_MASKING

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAAhSdy1Gg+k5U8WoFGkJvv+0TSg9mrSVCj4JTHay29uUc1_exA@mail.gmail.com>
Date: Tue, 16 Sep 2025 11:08:18 +0530
From: Anup Patel <anup@...infault.org>
To: Samuel Holland <samuel.holland@...ive.com>
Cc: Atish Patra <atish.patra@...ux.dev>, kvm-riscv@...ts.infradead.org, 
	Albert Ou <aou@...s.berkeley.edu>, Alexandre Ghiti <alex@...ti.fr>, 
	Palmer Dabbelt <palmer@...belt.com>, Paul Walmsley <paul.walmsley@...ive.com>, kvm@...r.kernel.org, 
	linux-kernel@...r.kernel.org, linux-riscv@...ts.infradead.org
Subject: Re: [PATCH] RISC-V: KVM: Fix SBI_FWFT_POINTER_MASKING_PMLEN algorithm

On Mon, Sep 15, 2025 at 11:04 AM Samuel Holland
<samuel.holland@...ive.com> wrote:
>
> The implementation of SBI_FWFT_POINTER_MASKING_PMLEN from commit
> aa04d131b88b ("RISC-V: KVM: Add support for SBI_FWFT_POINTER_MASKING_PMLEN")
> was based on a draft of the SBI 3.0 specification, and is not compliant
> with the ratified version.
>
> Update the algorithm to be compliant. Specifically, do not fall back to
> a pointer masking mode with a larger PMLEN if the mode with the
> requested PMLEN is unsupported by the hardware.
>
> Fixes: aa04d131b88b ("RISC-V: KVM: Add support for SBI_FWFT_POINTER_MASKING_PMLEN")
> Signed-off-by: Samuel Holland <samuel.holland@...ive.com>
> ---
> I saw that the RFC version of this patch already made it into
> riscv_kvm_queue, but it needs an update for ratified SBI 3.0. Feel free
> to squash this into the original commit, or I can send a replacement v2
> patch if you prefer.

Since this is fixing a commit in riscv_kvm_queue, I have squashed this
patch into the respective commit along with Drew's Reviewed-by.

Thanks,
Anup

>
>  arch/riscv/kvm/vcpu_sbi_fwft.c | 17 +++++++++++++----
>  1 file changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/arch/riscv/kvm/vcpu_sbi_fwft.c b/arch/riscv/kvm/vcpu_sbi_fwft.c
> index cacb3d4410a54..62cc9c3d57599 100644
> --- a/arch/riscv/kvm/vcpu_sbi_fwft.c
> +++ b/arch/riscv/kvm/vcpu_sbi_fwft.c
> @@ -160,14 +160,23 @@ static long kvm_sbi_fwft_set_pointer_masking_pmlen(struct kvm_vcpu *vcpu,
>         struct kvm_sbi_fwft *fwft = vcpu_to_fwft(vcpu);
>         unsigned long pmm;
>
> -       if (value == 0)
> +       switch (value) {
> +       case 0:
>                 pmm = ENVCFG_PMM_PMLEN_0;
> -       else if (value <= 7 && fwft->have_vs_pmlen_7)
> +               break;
> +       case 7:
> +               if (!fwft->have_vs_pmlen_7)
> +                       return SBI_ERR_INVALID_PARAM;
>                 pmm = ENVCFG_PMM_PMLEN_7;
> -       else if (value <= 16 && fwft->have_vs_pmlen_16)
> +               break;
> +       case 16:
> +               if (!fwft->have_vs_pmlen_16)
> +                       return SBI_ERR_INVALID_PARAM;
>                 pmm = ENVCFG_PMM_PMLEN_16;
> -       else
> +               break;
> +       default:
>                 return SBI_ERR_INVALID_PARAM;
> +       }
>
>         vcpu->arch.cfg.henvcfg &= ~ENVCFG_PMM;
>         vcpu->arch.cfg.henvcfg |= pmm;
> --
> 2.47.2
>
> base-commit: 7835b892d1d9f52fb61537757aa446fb44984215
> branch: up/kvm-fwft-pmlen