From: David Gibson <david@gibson.dropbear.id.au>
To: "Cédric Le Goater" <clg@kaod.org>
Cc: Satheesh Rajendran <sathnaga@linux.vnet.ibm.com>,
qemu-ppc@nongnu.org, qemu-devel@nongnu.org,
Greg Kurz <groug@kaod.org>
Subject: Re: [Qemu-devel] [PATCH v2 2/2] spapr/irq: Only claim VALID interrupts at the KVM level
Date: Mon, 16 Sep 2019 10:44:32 +1000 [thread overview]
Message-ID: <20190916004432.GG2104@umbus.fritz.box> (raw)
In-Reply-To: <20190911133937.2716-3-clg@kaod.org>
[-- Attachment #1: Type: text/plain, Size: 4594 bytes --]
On Wed, Sep 11, 2019 at 03:39:37PM +0200, Cédric Le Goater wrote:
> A typical pseries VM with 16 vCPUs, one disk, one network adapater
> uses less than 100 interrupts but the whole IRQ number space of the
> QEMU machine is allocated at reset time and it is 8K wide. This is
> wasting a considerable amount of interrupt numbers in the global IRQ
> space which has 1M interrupts per socket on a POWER9.
>
> To optimise the HW resources, only request at the KVM level interrupts
> which have been claimed by the guest. This will help to increase the
> maximum number of VMs per system and also help supporting nested guests
> using the XIVE interrupt mode.
>
> Signed-off-by: Cédric Le Goater <clg@kaod.org>
Applied to ppc-for-4.2, thanks.
> ---
> hw/intc/spapr_xive_kvm.c | 29 ++++++++++++++++++++++++++---
> hw/intc/xics_kvm.c | 8 ++++++++
> 2 files changed, 34 insertions(+), 3 deletions(-)
>
> diff --git a/hw/intc/spapr_xive_kvm.c b/hw/intc/spapr_xive_kvm.c
> index 17af4d19f54e..71b88d7797bc 100644
> --- a/hw/intc/spapr_xive_kvm.c
> +++ b/hw/intc/spapr_xive_kvm.c
> @@ -255,11 +255,16 @@ void kvmppc_xive_source_reset_one(XiveSource *xsrc, int srcno, Error **errp)
>
> static void kvmppc_xive_source_reset(XiveSource *xsrc, Error **errp)
> {
> + SpaprXive *xive = SPAPR_XIVE(xsrc->xive);
> int i;
>
> for (i = 0; i < xsrc->nr_irqs; i++) {
> Error *local_err = NULL;
>
> + if (!xive_eas_is_valid(&xive->eat[i])) {
> + continue;
> + }
> +
> kvmppc_xive_source_reset_one(xsrc, i, &local_err);
> if (local_err) {
> error_propagate(errp, local_err);
> @@ -328,11 +333,18 @@ uint64_t kvmppc_xive_esb_rw(XiveSource *xsrc, int srcno, uint32_t offset,
>
> static void kvmppc_xive_source_get_state(XiveSource *xsrc)
> {
> + SpaprXive *xive = SPAPR_XIVE(xsrc->xive);
> int i;
>
> for (i = 0; i < xsrc->nr_irqs; i++) {
> + uint8_t pq;
> +
> + if (!xive_eas_is_valid(&xive->eat[i])) {
> + continue;
> + }
> +
> /* Perform a load without side effect to retrieve the PQ bits */
> - uint8_t pq = xive_esb_read(xsrc, i, XIVE_ESB_GET);
> + pq = xive_esb_read(xsrc, i, XIVE_ESB_GET);
>
> /* and save PQ locally */
> xive_source_esb_set(xsrc, i, pq);
> @@ -521,9 +533,14 @@ static void kvmppc_xive_change_state_handler(void *opaque, int running,
> */
> if (running) {
> for (i = 0; i < xsrc->nr_irqs; i++) {
> - uint8_t pq = xive_source_esb_get(xsrc, i);
> + uint8_t pq;
> uint8_t old_pq;
>
> + if (!xive_eas_is_valid(&xive->eat[i])) {
> + continue;
> + }
> +
> + pq = xive_source_esb_get(xsrc, i);
> old_pq = xive_esb_read(xsrc, i, XIVE_ESB_SET_PQ_00 + (pq << 8));
>
> /*
> @@ -545,7 +562,13 @@ static void kvmppc_xive_change_state_handler(void *opaque, int running,
> * migration is in progress.
> */
> for (i = 0; i < xsrc->nr_irqs; i++) {
> - uint8_t pq = xive_esb_read(xsrc, i, XIVE_ESB_GET);
> + uint8_t pq;
> +
> + if (!xive_eas_is_valid(&xive->eat[i])) {
> + continue;
> + }
> +
> + pq = xive_esb_read(xsrc, i, XIVE_ESB_GET);
>
> /*
> * PQ is set to PENDING to possibly catch a triggered
> diff --git a/hw/intc/xics_kvm.c b/hw/intc/xics_kvm.c
> index a4d2e876cc5f..ba90d6dc966c 100644
> --- a/hw/intc/xics_kvm.c
> +++ b/hw/intc/xics_kvm.c
> @@ -190,6 +190,10 @@ void ics_get_kvm_state(ICSState *ics)
> for (i = 0; i < ics->nr_irqs; i++) {
> ICSIRQState *irq = &ics->irqs[i];
>
> + if (ics_irq_free(ics, i)) {
> + continue;
> + }
> +
> kvm_device_access(kernel_xics_fd, KVM_DEV_XICS_GRP_SOURCES,
> i + ics->offset, &state, false, &error_fatal);
>
> @@ -301,6 +305,10 @@ int ics_set_kvm_state(ICSState *ics, Error **errp)
> Error *local_err = NULL;
> int ret;
>
> + if (ics_irq_free(ics, i)) {
> + continue;
> + }
> +
> ret = ics_set_kvm_state_one(ics, i, &local_err);
> if (ret < 0) {
> error_propagate(errp, local_err);
--
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 --]
next prev parent reply other threads:[~2019-09-16 5:04 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-11 13:39 [Qemu-devel] [PATCH v2 0/2] spapr: reduce the number of IRQ Cédric Le Goater
2019-09-11 13:39 ` [Qemu-devel] [PATCH v2 1/2] spapr/irq: Introduce an ics_irq_free() helper Cédric Le Goater
2019-09-11 14:09 ` Greg Kurz
2019-09-16 0:38 ` David Gibson
2019-09-11 13:39 ` [Qemu-devel] [PATCH v2 2/2] spapr/irq: Only claim VALID interrupts at the KVM level Cédric Le Goater
2019-09-11 14:14 ` Greg Kurz
2019-09-16 0:44 ` David Gibson [this message]
2019-09-25 15:41 ` Greg Kurz
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=20190916004432.GG2104@umbus.fritz.box \
--to=david@gibson.dropbear.id.au \
--cc=clg@kaod.org \
--cc=groug@kaod.org \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@nongnu.org \
--cc=sathnaga@linux.vnet.ibm.com \
/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).