qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: David Gibson <david@gibson.dropbear.id.au>
To: "Cédric Le Goater" <clg@kaod.org>
Cc: qemu-devel@nongnu.org, qemu-ppc@nongnu.org,
	Greg Kurz <groug@kaod.org>,
	Gustavo Romero <gromero@linux.ibm.com>
Subject: Re: [PATCH 3/8] spapr/xive: Query the characteristics of a source in KVM
Date: Thu, 20 Aug 2020 11:33:33 +1000	[thread overview]
Message-ID: <20200820013333.GG271315@yekko.fritz.box> (raw)
In-Reply-To: <20200819130843.2230799-4-clg@kaod.org>

[-- Attachment #1: Type: text/plain, Size: 4714 bytes --]

On Wed, Aug 19, 2020 at 03:08:38PM +0200, Cédric Le Goater wrote:
> When running a guest with a kernel IRQ chip enabled, the XIVE
> characteristics of the interrupts are advertised to the guest in the
> H_INT_GET_SOURCE_INFO hcall. These characteristics depend on the
> underlying HW interrupts but today, QEMU simply advertises its own
> without checking what the host supports. It is not a problem for the
> moment, but POWER10 will (re)add support for StoreEOI and we need a
> way to in sync with the host.
> 
> The KVM_DEV_XIVE_GRP_SOURCE_INFO command lets QEMU query the XIVE
> characteristics of the underlying HW interrupts and override any
> previous setting done by QEMU. This allows the fallback mode, when the
> XIVE device is emulated by QEMU, to use its own custom settings on the
> sources but makes sure that we don't let a guest run with features
> incompatible with KVM.
> 
> It only applies to the StoreEOI feature for the moment.

Urgh.  This means that the source characteristics can change across a
migration, that's kind of a problem.

> Signed-off-by: Cédric Le Goater <clg@kaod.org>
> ---
>  include/hw/ppc/spapr_xive.h |  2 ++
>  hw/intc/spapr_xive.c        | 20 ++++++++++++++++++++
>  hw/intc/spapr_xive_kvm.c    | 26 ++++++++++++++++++++++++++
>  3 files changed, 48 insertions(+)
> 
> diff --git a/include/hw/ppc/spapr_xive.h b/include/hw/ppc/spapr_xive.h
> index 1dddcbcb9cdd..3f325723ea74 100644
> --- a/include/hw/ppc/spapr_xive.h
> +++ b/include/hw/ppc/spapr_xive.h
> @@ -84,6 +84,8 @@ void kvmppc_xive_disconnect(SpaprInterruptController *intc);
>  void kvmppc_xive_reset(SpaprXive *xive, Error **errp);
>  int kvmppc_xive_set_source_config(SpaprXive *xive, uint32_t lisn, XiveEAS *eas,
>                                    Error **errp);
> +int kvmppc_xive_get_source_info(SpaprXive *xive, uint32_t lisn, uint64_t *flags,
> +                                 Error **errp);
>  void kvmppc_xive_sync_source(SpaprXive *xive, uint32_t lisn, Error **errp);
>  uint64_t kvmppc_xive_esb_rw(XiveSource *xsrc, int srcno, uint32_t offset,
>                              uint64_t data, bool write);
> diff --git a/hw/intc/spapr_xive.c b/hw/intc/spapr_xive.c
> index 1fa09f287ac0..943b9958a68b 100644
> --- a/hw/intc/spapr_xive.c
> +++ b/hw/intc/spapr_xive.c
> @@ -932,6 +932,26 @@ static target_ulong h_int_get_source_info(PowerPCCPU *cpu,
>          args[0] |= SPAPR_XIVE_SRC_STORE_EOI;
>      }
>  
> +    if (kvm_irqchip_in_kernel()) {
> +        Error *local_err = NULL;
> +        uint64_t flags = 0;
> +
> +        kvmppc_xive_get_source_info(xive, lisn, &flags, &local_err);
> +        if (local_err) {
> +            error_report_err(local_err);
> +            return H_HARDWARE;
> +        }
> +
> +        /*
> +         * Override QEMU settings with KVM values
> +         */
> +        if (flags & XIVE_SRC_STORE_EOI) {
> +            args[0] |= SPAPR_XIVE_SRC_STORE_EOI;
> +        } else {
> +            args[0] &= ~SPAPR_XIVE_SRC_STORE_EOI;
> +        }
> +    }
> +
>      /*
>       * Force the use of the H_INT_ESB hcall in case of an LSI
>       * interrupt. This is necessary under KVM to re-trigger the
> diff --git a/hw/intc/spapr_xive_kvm.c b/hw/intc/spapr_xive_kvm.c
> index e8667ce5f621..90f4509e6959 100644
> --- a/hw/intc/spapr_xive_kvm.c
> +++ b/hw/intc/spapr_xive_kvm.c
> @@ -217,6 +217,32 @@ int kvmppc_xive_set_source_config(SpaprXive *xive, uint32_t lisn, XiveEAS *eas,
>                               &kvm_src, true, errp);
>  }
>  
> +int kvmppc_xive_get_source_info(SpaprXive *xive, uint32_t lisn, uint64_t *flags,
> +                                 Error **errp)
> +{
> +    struct kvm_ppc_xive_src kvm_src = { 0 };
> +    int ret;
> +
> +    /*
> +     * Check that KVM supports the new attribute to query source
> +     * characteristics.
> +     */
> +    if (!kvm_device_check_attr(xive->fd, KVM_DEV_XIVE_GRP_SOURCE_INFO, 0)) {
> +        return 0;
> +    }
> +
> +    ret = kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_SOURCE_INFO, lisn,
> +                            &kvm_src, false, errp);
> +    if (ret < 0) {
> +        return ret;
> +    }
> +
> +    if (kvm_src.flags & KVM_XIVE_SOURCE_FLAG_STORE_EOI) {
> +        *flags |= XIVE_SRC_STORE_EOI;
> +    }
> +    return 0;
> +}
> +
>  void kvmppc_xive_sync_source(SpaprXive *xive, uint32_t lisn, Error **errp)
>  {
>      kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_SOURCE_SYNC, lisn,

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

  reply	other threads:[~2020-08-20  1:44 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-19 13:08 [PATCH 0/8] spapr/xive: Activate StoreEOI in P10 compat guests Cédric Le Goater
2020-08-19 13:08 ` [PATCH 1/8] spapr/xive: Add a 'hv-prio' property to represent the KVM escalation priority Cédric Le Goater
2020-08-20  0:58   ` David Gibson
2020-08-19 13:08 ` [PATCH 2/8] linux-headers: Update for KVM_DEV_XIVE_GRP_SOURCE_INFO Cédric Le Goater
2020-08-20  0:58   ` David Gibson
2020-08-20  6:49     ` Cédric Le Goater
2020-08-19 13:08 ` [PATCH 3/8] spapr/xive: Query the characteristics of a source in KVM Cédric Le Goater
2020-08-20  1:33   ` David Gibson [this message]
2020-08-20  6:58     ` Cédric Le Goater
2020-08-19 13:08 ` [PATCH 4/8] spapr/xive: Activate StoreEOI for POWER10 only Cédric Le Goater
2020-08-19 13:08 ` [PATCH 5/8] spapr/xive: Enforce the load-after-store ordering Cédric Le Goater
2020-08-19 13:08 ` [PATCH 6/8] spapr/xive: Activate StoreEOI by default Cédric Le Goater
2020-08-19 13:08 ` [PATCH 7/8] spapr/xive: Use the xics flag to check for XIVE-only IRQ backends Cédric Le Goater
2020-08-20  1:36   ` David Gibson
2020-08-20  6:59     ` Cédric Le Goater
2020-08-19 13:08 ` [PATCH 8/8] spapr/xive: Introduce a XIVE StoreEOI IRQ backend Cédric Le Goater

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=20200820013333.GG271315@yekko.fritz.box \
    --to=david@gibson.dropbear.id.au \
    --cc=clg@kaod.org \
    --cc=gromero@linux.ibm.com \
    --cc=groug@kaod.org \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.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).