linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: "Cédric Le Goater" <clg@kaod.org>
To: Paul Mackerras <paulus@ozlabs.org>,
	linuxppc-dev@ozlabs.org, kvm@vger.kernel.org
Cc: kvm-ppc@vger.kernel.org, David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [PATCH v2 2/3] KVM: PPC: Book3S HV: Don't push XIVE context when not using XIVE device
Date: Tue, 13 Aug 2019 14:18:34 +0200	[thread overview]
Message-ID: <d8435ea8-cfa5-9a3b-b081-b5541b6052b3@kaod.org> (raw)
In-Reply-To: <20190813100100.GC9567@blackberry>

On 13/08/2019 12:01, Paul Mackerras wrote:
> At present, when running a guest on POWER9 using HV KVM but not using
> an in-kernel interrupt controller (XICS or XIVE), for example if QEMU
> is run with the kernel_irqchip=off option, the guest entry code goes
> ahead and tries to load the guest context into the XIVE hardware, even
> though no context has been set up.
> 
> To fix this, we check that the "CAM word" is non-zero before pushing
> it to the hardware.  The CAM word is initialized to a non-zero value
> in kvmppc_xive_connect_vcpu() and kvmppc_xive_native_connect_vcpu(),
> and is now cleared in kvmppc_xive_{,native_}cleanup_vcpu.

If a "CAM word" is defined, it means the vCPU (VP) was enabled at the
XIVE HW level. So this is the criteria to consider that a vCPU needs
to update (push) its XIVE thread interrupt context when scheduled
to run.


Reviewed-by: Cédric Le Goater <clg@kaod.org>

Thanks,

C.

> 
> Cc: stable@vger.kernel.org # v4.11+
> Reported-by: Cédric Le Goater <clg@kaod.org>
> Fixes: 5af50993850a ("KVM: PPC: Book3S HV: Native usage of the XIVE interrupt controller")
> Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
> ---
>  arch/powerpc/kvm/book3s_hv_rmhandlers.S |  2 ++
>  arch/powerpc/kvm/book3s_xive.c          | 11 ++++++++++-
>  arch/powerpc/kvm/book3s_xive_native.c   |  3 +++
>  3 files changed, 15 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
> index 2e7e788..07181d0 100644
> --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
> +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
> @@ -942,6 +942,8 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300)
>  	ld	r11, VCPU_XIVE_SAVED_STATE(r4)
>  	li	r9, TM_QW1_OS
>  	lwz	r8, VCPU_XIVE_CAM_WORD(r4)
> +	cmpwi	r8, 0
> +	beq	no_xive
>  	li	r7, TM_QW1_OS + TM_WORD2
>  	mfmsr	r0
>  	andi.	r0, r0, MSR_DR		/* in real mode? */
> diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c
> index 09f838a..586867e 100644
> --- a/arch/powerpc/kvm/book3s_xive.c
> +++ b/arch/powerpc/kvm/book3s_xive.c
> @@ -67,8 +67,14 @@ void kvmppc_xive_push_vcpu(struct kvm_vcpu *vcpu)
>  	void __iomem *tima = local_paca->kvm_hstate.xive_tima_virt;
>  	u64 pq;
>  
> -	if (!tima)
> +	/*
> +	 * Nothing to do if the platform doesn't have a XIVE
> +	 * or this vCPU doesn't have its own XIVE context
> +	 * (e.g. because it's not using an in-kernel interrupt controller).
> +	 */
> +	if (!tima || !vcpu->arch.xive_cam_word)
>  		return;
> +
>  	eieio();
>  	__raw_writeq(vcpu->arch.xive_saved_state.w01, tima + TM_QW1_OS);
>  	__raw_writel(vcpu->arch.xive_cam_word, tima + TM_QW1_OS + TM_WORD2);
> @@ -1146,6 +1152,9 @@ void kvmppc_xive_cleanup_vcpu(struct kvm_vcpu *vcpu)
>  	/* Disable the VP */
>  	xive_native_disable_vp(xc->vp_id);
>  
> +	/* Clear the cam word so guest entry won't try to push context */
> +	vcpu->arch.xive_cam_word = 0;
> +
>  	/* Free the queues */
>  	for (i = 0; i < KVMPPC_XIVE_Q_COUNT; i++) {
>  		struct xive_q *q = &xc->queues[i];
> diff --git a/arch/powerpc/kvm/book3s_xive_native.c b/arch/powerpc/kvm/book3s_xive_native.c
> index 368427f..11b91b4 100644
> --- a/arch/powerpc/kvm/book3s_xive_native.c
> +++ b/arch/powerpc/kvm/book3s_xive_native.c
> @@ -81,6 +81,9 @@ void kvmppc_xive_native_cleanup_vcpu(struct kvm_vcpu *vcpu)
>  	/* Disable the VP */
>  	xive_native_disable_vp(xc->vp_id);
>  
> +	/* Clear the cam word so guest entry won't try to push context */
> +	vcpu->arch.xive_cam_word = 0;
> +
>  	/* Free the queues */
>  	for (i = 0; i < KVMPPC_XIVE_Q_COUNT; i++) {
>  		kvmppc_xive_native_cleanup_queue(vcpu, i);
> 


  reply	other threads:[~2019-08-13 12:23 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-13  9:58 [PATCH v2 0/3] powerpc/xive: Fix race condition leading to host crashes and hangs Paul Mackerras
2019-08-13 10:01 ` [PATCH v2 2/3] KVM: PPC: Book3S HV: Don't push XIVE context when not using XIVE device Paul Mackerras
2019-08-13 12:18   ` Cédric Le Goater [this message]
2019-08-22 10:46   ` Michael Ellerman
2019-08-13 10:03 ` [PATCH v2 1/3] KVM: PPC: Book3S HV: Fix race in re-enabling XIVE escalation interrupts Paul Mackerras
2019-08-14  4:46   ` Jordan Niethe
2019-08-14  6:05     ` Paul Mackerras
2019-08-22 10:46   ` Michael Ellerman
2019-08-13 10:06 ` [PATCH v2 3/3] powerpc/xive: Implement get_irqchip_state method for XIVE to fix shutdown race Paul Mackerras
2019-08-22 10:46   ` Michael Ellerman

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=d8435ea8-cfa5-9a3b-b081-b5541b6052b3@kaod.org \
    --to=clg@kaod.org \
    --cc=david@gibson.dropbear.id.au \
    --cc=kvm-ppc@vger.kernel.org \
    --cc=kvm@vger.kernel.org \
    --cc=linuxppc-dev@ozlabs.org \
    --cc=paulus@ozlabs.org \
    /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;
as well as URLs for NNTP newsgroup(s).