From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=209.127.231.27 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1794884079; b=bw2Zb9GC5DEqcaBN8XuelGfvMKtlH7iDPi8lAYFm8elWSODfn7jwXQIbvCbdw0HoHxPO pJsW6zw9W9mSZZL1jOCfy27NlPRiDC8GxHs3VVLM9/clZcjFAXxMmFznWA2EiX33iLt8D MpAbYRG60rsXmd2MYcjEmWKsfVLy5OC7McYRBIyr6e5EM2w6Hj3w+/4IR0356aYHbNKPd tw5SyPdXtcaJhcnc9DN5i3XrN8ZFRDLms5nIiiyA9lOJadkX7wyGIpx6WCnI1Mm+MtVv7 8gvsw/4H7nNxuRyURzeY6LOANHkOregqaBUSXKWV0P7MX8OuiXq02s9lQqK9Sv6gnCw== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1794884079; h=DMARC-Filter:DKIM-Signature:To:Content-Transfer-Encoding: Content-Language:In-Reply-To:Cc:Message-Id:References:User-Agent: X-Lms-Return-Path:Subject:X-Original-From:From:Date:Mime-Version: Received:Content-Type; bh=B6o0YTG1q+hlC1dIwJ8MRViNtPcn8I2cwDIc51AfJlU=; b=r6CIK8QXxmWBLBGFbm5tEFBY1lTe6nE651MM+ctHug7NaNFkRSNOOl7uWBy5iePXDFP0 YNf6QaGoq9XV3fR8y7BihO1GRJxv9RNMCkQUKOiEwht4LqTiVHIO7HIOWAmrGlIrZO9EU 8mL1yYWuF+LcMccjV4NYq9hfhLYORyvQSRon2g2S5XAJ8HqNcd2gUFafpUs45UHfnN/YW VRSz7+UDocpgLtrg9J+owNTWWdhYm6o0DyUIgRbLdgJRnjfUiQ6vJA5dspfXWMU1DH5eX pGXJn95vhcvdkN9MSke1tMCltJduUsJ3EGPCaCcKu7TfDPf3otNJA9QgdTdFIiiNUuQ== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none header.from=lanxincomputing.com; spf=pass smtp.mailfrom=lanxincomputing.com; dkim=pass header.d=lanxincomputing-com.20200927.dkim.feishu.cn header.i=@lanxincomputing-com.20200927.dkim.feishu.cn header.a=rsa-sha256 header.s=s1 header.b=NAU6GYya; arc=none smtp.remote-ip=209.127.231.27 Received: from va-2-27.ptr.blmpb.com (va-2-27.ptr.blmpb.com [209.127.231.27]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D5080606A323 for ; Tue, 17 Nov 2026 02:54:37 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.2 smtp.subspace.kernel.org D5080606A323 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=lanxincomputing.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=lanxincomputing.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=lanxincomputing-com.20200927.dkim.feishu.cn header.i=@lanxincomputing-com.20200927.dkim.feishu.cn header.a=rsa-sha256 header.s=s1 header.b=NAU6GYya DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; s=s1; d=lanxincomputing-com.20200927.dkim.feishu.cn; t=1794884072; h=from:subject:mime-version:from:date:message-id:subject:to:cc: reply-to:content-type:mime-version:in-reply-to:message-id; bh=B6o0YTG1q+hlC1dIwJ8MRViNtPcn8I2cwDIc51AfJlU=; b=NAU6GYyaq4Wb0QHlXhFg7ATfwc/YgfdyT2kNCbZIUZyKk9tcTSMOUVoDiGs1MPG0IQLv4/ 8zqfZ6NmjLZcOgrYbFiwDC6W3bzyzK3v2aq4ZRRWfUfgxCLZ3ovoRB2EtUnGbIZ5I5R0fY zOugUACONXcaEweSoP3NwK7zahW9fP9lt58v6r1wcqf9PA+858ggNofyAa0IR8VfC6znSJ OQbr4iQDLyAyOP1V0r8B2SHdAz6uaKbGLk96z8yJb78yO0H3SMUYmtxjs0Gb5SzGlKqrXm kdNbaaFk/h5R0nB7A9kH5MCUoI1JUVyTMDkNYIRh+e8fBYv4qtzMvOwco6RETg== To: "Anup Patel" Content-Transfer-Encoding: quoted-printable Content-Language: en-US In-Reply-To: Cc: , , , , , , , , , , , Message-Id: References: <20260923091658.1421475-1-xiangwencheng@lanxincomputing.com> User-Agent: Mozilla Thunderbird X-Lms-Return-Path: Subject: Re: [PATCH v3] RISC-V: KVM: Serialize and atomize the IMSIC's TOPEI read-and-clear operation X-Original-From: BillXiang From: "BillXiang" Date: Tue, 17 Nov 2026 10:54:26 +0800 X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list Mime-Version: 1.0 Received: from [127.0.0.1] ([123.120.4.67]) by smtp.feishu.cn with ESMTPS; Tue, 17 Nov 2026 10:54:29 +0800 Content-Type: text/plain; charset=UTF-8 On 11/13/2026 1:36 PM, Anup Patel wrote: > On Wed, Sep 23, 2026 at 2:47=E2=80=AFPM BillXiang > wrote: >> >> The TOPEI read-and-clear must be atomic with respect to >> interrupt injection (set_bit in EIP). Otherwise an >> interrupt injected between the topei read and the >> clear_bit could be silently dropped. >> >> v3: >> - fix undefined eflags >> >> v2: >> - fix patch file curruption >> >> v1: https://lore.kernel.org/kvm-riscv/20260923070332.1376132-1-xiangwenc= heng@lanxincomputing.com/ >> >> Signed-off-by: BillXiang >> --- >> arch/riscv/kvm/aia_imsic.c | 42 +++++++++++++++++++++++++------------- >> 1 file changed, 28 insertions(+), 14 deletions(-) >> >> diff --git a/arch/riscv/kvm/aia_imsic.c b/arch/riscv/kvm/aia_imsic.c >> index c1af23e79ae0..c440088651ec 100644 >> --- a/arch/riscv/kvm/aia_imsic.c >> +++ b/arch/riscv/kvm/aia_imsic.c >> @@ -607,26 +607,25 @@ static void imsic_vsfile_cleanup(struct imsic *ims= ic) >> kvm_riscv_aia_free_hgei(old_vsfile_cpu, old_vsfile_hgei= ); >> } >> >> -static void imsic_swfile_extirq_update(struct kvm_vcpu *vcpu) >> +static void __imsic_swfile_extirq_update(struct kvm_vcpu *vcpu) >> { >> struct imsic *imsic =3D vcpu->arch.aia_context.imsic_state; >> struct imsic_mrif *mrif =3D imsic->swfile; >> - unsigned long flags; >> - >> - /* >> - * The critical section is necessary during external interrupt >> - * updates to avoid the risk of losing interrupts due to potenti= al >> - * interruptions between reading topei and updating pending stat= us. >> - */ >> - >> - raw_spin_lock_irqsave(&imsic->swfile_extirq_lock, flags); >> >> if (imsic_mrif_atomic_read(mrif, &mrif->eidelivery) && >> imsic_mrif_topei(mrif, imsic->nr_eix, imsic->nr_msis)) >> kvm_riscv_vcpu_set_interrupt(vcpu, IRQ_VS_EXT); >> else >> kvm_riscv_vcpu_unset_interrupt(vcpu, IRQ_VS_EXT); >> +} >> + >> +static void imsic_swfile_extirq_update(struct kvm_vcpu *vcpu) >> +{ >> + struct imsic *imsic =3D vcpu->arch.aia_context.imsic_state; >> + unsigned long flags; >> >> + raw_spin_lock_irqsave(&imsic->swfile_extirq_lock, flags); >> + __imsic_swfile_extirq_update(vcpu); >> raw_spin_unlock_irqrestore(&imsic->swfile_extirq_lock, flags); >> } >> >> @@ -912,12 +911,15 @@ int kvm_riscv_vcpu_aia_imsic_rmw(struct kvm_vcpu *= vcpu, unsigned long isel, >> struct imsic_mrif_eix *eix; >> int r, rc =3D KVM_INSN_CONTINUE_NEXT_SEPC; >> struct imsic *imsic =3D vcpu->arch.aia_context.imsic_state; >> + unsigned long flags; >> >> /* If IMSIC vCPU state not initialized then forward to user spa= ce */ >> if (!imsic) >> return KVM_INSN_EXIT_TO_USER_SPACE; >> >> if (isel =3D=3D KVM_RISCV_AIA_IMSIC_TOPEI) { >> + raw_spin_lock_irqsave(&imsic->swfile_extirq_lock, flags)= ; >> + >> /* Read pending and enabled interrupt with highest prio= rity */ >> topei =3D imsic_mrif_topei(imsic->swfile, imsic->nr_eix= , >> imsic->nr_msis); >> @@ -934,16 +936,20 @@ int kvm_riscv_vcpu_aia_imsic_rmw(struct kvm_vcpu *= vcpu, unsigned long isel, >> eix->eip); >> } >> } >> + if (wr_mask) >> + __imsic_swfile_extirq_update(vcpu); >=20 > The __imsic_swfile_extirq_update() also calls imsic_mrif_topei() so it is > better to pass topei value to __imsic_swfile_extirq_update().=20 Hi Anup, The topei value may have been changed before this=20 __imsic_swfile_extirq_update() is called. Would it be better to re-read=20 it in __imsic_swfile_extirq_update() as before rather than pass it in? Here is the relevant code: int kvm_riscv_vcpu_aia_imsic_rmw(struct kvm_vcpu *vcpu, unsigned long isel, unsigned long *val, unsigned long new_val, unsigned long wr_mask) { ... /* Read pending and enabled interrupt with highest priority */ topei =3D imsic_mrif_topei(imsic->swfile, imsic->nr_eix, imsic->nr_msis); if (val) *val =3D topei; /* Writes ignore value and clear top pending interrupt */ if (topei && wr_mask) { topei >>=3D TOPEI_ID_SHIFT; if (topei) { eix =3D &imsic->swfile->eix[topei / BITS_PER_TYPE(u64)]; clear_bit(topei & (BITS_PER_TYPE(u64) - 1), eix->eip); } } if (wr_mask) __imsic_swfile_extirq_update(vcpu); ... } Regards, Bill >=20 >> + >> + raw_spin_unlock_irqrestore(&imsic->swfile_extirq_lock, f= lags); >> } else { >> r =3D imsic_mrif_rmw(imsic->swfile, imsic->nr_eix, isel= , >> val, new_val, wr_mask); >> /* Forward unknown IMSIC register to user-space */ >> if (r) >> rc =3D (r =3D=3D -ENOENT) ? 0 : KVM_INSN_ILLEGA= L_TRAP; >> - } >> >> - if (wr_mask) >> - imsic_swfile_extirq_update(vcpu); >> + if (wr_mask) >> + imsic_swfile_extirq_update(vcpu); >> + } >> >> return rc; >> } >> @@ -1050,9 +1056,17 @@ int kvm_riscv_vcpu_aia_imsic_inject(struct kvm_vc= pu *vcpu, >> if (imsic->vsfile_cpu >=3D 0) { >> writel(iid, imsic->vsfile_va + IMSIC_MMIO_SETIPNUM_LE); >> } else { >> + /* >> + * Serialize the EIP set + external-irq update against t= he >> + * TOPEI read-and-clear path so that an injected interru= pt >> + * cannot be silently lost when the guest acknowledges a >> + * different (or the same) interrupt concurrently. >> + */ >> + raw_spin_lock(&imsic->swfile_extirq_lock); >> eix =3D &imsic->swfile->eix[iid / BITS_PER_TYPE(u64)]; >> set_bit(iid & (BITS_PER_TYPE(u64) - 1), eix->eip); >> - imsic_swfile_extirq_update(vcpu); >> + __imsic_swfile_extirq_update(vcpu); >> + raw_spin_unlock(&imsic->swfile_extirq_lock); >> } >> >> read_unlock_irqrestore(&imsic->vsfile_lock, flags); >> -- >> 2.53.0 >=20 > Regards, > Anup