All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Gibson <david@gibson.dropbear.id.au>
To: peter.maydell@linaro.org
Cc: danielhb413@gmail.com, qemu-devel@nongnu.org, groug@kaod.org,
	qemu-ppc@nongnu.org, "Cédric Le Goater" <clg@kaod.org>,
	bauerman@linux.ibm.com,
	"David Gibson" <david@gibson.dropbear.id.au>
Subject: [PULL 10/33] spapr/xive: Allocate IPIs independently from the other sources
Date: Tue,  8 Sep 2020 15:19:30 +1000	[thread overview]
Message-ID: <20200908051953.1616885-11-david@gibson.dropbear.id.au> (raw)
In-Reply-To: <20200908051953.1616885-1-david@gibson.dropbear.id.au>

From: Cédric Le Goater <clg@kaod.org>

The vCPU IPIs are now allocated in kvmppc_xive_cpu_connect() when the
vCPU connects to the KVM device and not when all the sources are reset
in kvmppc_xive_source_reset()

This requires extra care for hotplug vCPUs and VM restore.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20200820134547.2355743-4-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/intc/spapr_xive_kvm.c | 47 +++++++++++++++++++++++++++++++++++-----
 1 file changed, 42 insertions(+), 5 deletions(-)

diff --git a/hw/intc/spapr_xive_kvm.c b/hw/intc/spapr_xive_kvm.c
index 3e80ea0ce9..507637d51e 100644
--- a/hw/intc/spapr_xive_kvm.c
+++ b/hw/intc/spapr_xive_kvm.c
@@ -146,6 +146,15 @@ int kvmppc_xive_cpu_synchronize_state(XiveTCTX *tctx, Error **errp)
     return s.ret;
 }
 
+static int kvmppc_xive_reset_ipi(SpaprXive *xive, CPUState *cs, Error **errp)
+{
+    unsigned long ipi = kvm_arch_vcpu_id(cs);
+    uint64_t state = 0;
+
+    return kvm_device_access(xive->fd, KVM_DEV_XIVE_GRP_SOURCE, ipi,
+                              &state, true, errp);
+}
+
 int kvmppc_xive_cpu_connect(XiveTCTX *tctx, Error **errp)
 {
     ERRP_GUARD();
@@ -175,6 +184,12 @@ int kvmppc_xive_cpu_connect(XiveTCTX *tctx, Error **errp)
         return ret;
     }
 
+    /* Create/reset the vCPU IPI */
+    ret = kvmppc_xive_reset_ipi(xive, tctx->cs, errp);
+    if (ret < 0) {
+        return ret;
+    }
+
     kvm_cpu_enable(tctx->cs);
     return 0;
 }
@@ -234,6 +249,12 @@ int kvmppc_xive_source_reset_one(XiveSource *xsrc, int srcno, Error **errp)
 
     assert(xive->fd != -1);
 
+    /*
+     * The vCPU IPIs are now allocated in kvmppc_xive_cpu_connect()
+     * and not with all sources in kvmppc_xive_source_reset()
+     */
+    assert(srcno >= SPAPR_XIRQ_BASE);
+
     if (xive_source_irq_is_lsi(xsrc, srcno)) {
         state |= KVM_XIVE_LEVEL_SENSITIVE;
         if (xsrc->status[srcno] & XIVE_STATUS_ASSERTED) {
@@ -245,12 +266,28 @@ int kvmppc_xive_source_reset_one(XiveSource *xsrc, int srcno, Error **errp)
                              true, errp);
 }
 
+/*
+ * To be valid, a source must have been claimed by the machine (valid
+ * entry in the EAS table) and if it is a vCPU IPI, the vCPU should
+ * have been enabled, which means the IPI has been allocated in
+ * kvmppc_xive_cpu_connect().
+ */
+static bool xive_source_is_valid(SpaprXive *xive, int i)
+{
+    return xive_eas_is_valid(&xive->eat[i]) &&
+        (i >= SPAPR_XIRQ_BASE || kvm_cpu_is_enabled(i));
+}
+
 static int kvmppc_xive_source_reset(XiveSource *xsrc, Error **errp)
 {
     SpaprXive *xive = SPAPR_XIVE(xsrc->xive);
     int i;
 
-    for (i = 0; i < xsrc->nr_irqs; i++) {
+    /*
+     * Skip the vCPU IPIs. These are created/reset when the vCPUs are
+     * connected in kvmppc_xive_cpu_connect()
+     */
+    for (i = SPAPR_XIRQ_BASE; i < xsrc->nr_irqs; i++) {
         int ret;
 
         if (!xive_eas_is_valid(&xive->eat[i])) {
@@ -332,7 +369,7 @@ static void kvmppc_xive_source_get_state(XiveSource *xsrc)
     for (i = 0; i < xsrc->nr_irqs; i++) {
         uint8_t pq;
 
-        if (!xive_eas_is_valid(&xive->eat[i])) {
+        if (!xive_source_is_valid(xive, i)) {
             continue;
         }
 
@@ -515,7 +552,7 @@ static void kvmppc_xive_change_state_handler(void *opaque, int running,
             uint8_t pq;
             uint8_t old_pq;
 
-            if (!xive_eas_is_valid(&xive->eat[i])) {
+            if (!xive_source_is_valid(xive, i)) {
                 continue;
             }
 
@@ -543,7 +580,7 @@ static void kvmppc_xive_change_state_handler(void *opaque, int running,
     for (i = 0; i < xsrc->nr_irqs; i++) {
         uint8_t pq;
 
-        if (!xive_eas_is_valid(&xive->eat[i])) {
+        if (!xive_source_is_valid(xive, i)) {
             continue;
         }
 
@@ -658,7 +695,7 @@ int kvmppc_xive_post_load(SpaprXive *xive, int version_id)
 
     /* Restore the EAT */
     for (i = 0; i < xive->nr_irqs; i++) {
-        if (!xive_eas_is_valid(&xive->eat[i])) {
+        if (!xive_source_is_valid(xive, i)) {
             continue;
         }
 
-- 
2.26.2



  parent reply	other threads:[~2020-09-08  5:27 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-08  5:19 [PULL 00/33] ppc-for-5.2 queue 20200908 David Gibson
2020-09-08  5:19 ` [PULL 01/33] adb: Correct class size on TYPE_ADB_DEVICE David Gibson
2020-09-08  5:19 ` [PULL 02/33] ppc/pnv: Fix TypeInfo of PnvLpcController abstract class David Gibson
2020-09-08  5:19 ` [PULL 03/33] spapr: Remove unnecessary DRC type-checker macros David Gibson
2020-09-08  5:19 ` [PULL 04/33] spapr/xive: Add a 'hv-prio' property to represent the KVM escalation priority David Gibson
2020-09-08  5:19 ` [PULL 05/33] ppc/pnv: Add a HIOMAP erase command David Gibson
2020-09-08  5:19 ` [PULL 06/33] spapr_vscsi: do not allow device hotplug David Gibson
2020-09-08  5:19 ` [PULL 07/33] spapr/xive: Use the xics flag to check for XIVE-only IRQ backends David Gibson
2020-09-08  5:19 ` [PULL 08/33] spapr/xive: Modify kvm_cpu_is_enabled() interface David Gibson
2020-09-08  5:19 ` [PULL 09/33] spapr/xive: Use kvmppc_xive_source_reset() in post_load David Gibson
2020-09-08  5:19 ` David Gibson [this message]
2020-09-08  5:19 ` [PULL 11/33] spapr/xive: Allocate vCPU IPIs from the vCPU contexts David Gibson
2020-09-08  5:19 ` [PULL 12/33] ppc/spapr_nvdimm: use g_autofree in spapr_nvdimm_validate_opts() David Gibson
2020-09-08  5:19 ` [PULL 13/33] spapr, spapr_nvdimm: fold NVDIMM validation in the same place David Gibson
2020-09-08  5:19 ` [PULL 14/33] ppc/spapr_nvdimm: do not enable support with 'nvdimm=off' David Gibson
2020-09-08  5:19 ` [PULL 15/33] target/arm: Move start-powered-off property to generic CPUState David Gibson
2020-09-08  5:19 ` [PULL 16/33] target/arm: Move setting of CPU halted state to generic code David Gibson
2020-09-08  5:19 ` [PULL 17/33] ppc/spapr: Use start-powered-off CPUState property David Gibson
2020-09-08  5:19 ` [PULL 18/33] ppc/e500: " David Gibson
2020-09-08  5:19 ` [PULL 19/33] mips/cps: " David Gibson
2020-10-07  9:28   ` Alex Bennée
2020-10-07 10:36     ` Philippe Mathieu-Daudé
2020-10-07 10:47       ` Alex Bennée
2020-10-07 11:36         ` Philippe Mathieu-Daudé
2020-09-08  5:19 ` [PULL 20/33] sparc/sun4m: Don't set cs->halted = 0 in main_cpu_reset() David Gibson
2020-09-08  5:19 ` [PULL 21/33] sparc/sun4m: Use start-powered-off CPUState property David Gibson
2020-09-08  5:19 ` [PULL 22/33] target/s390x: " David Gibson
2020-09-08  5:19 ` [PULL 23/33] hw/ppc/ppc4xx_pci: Use ARRAY_SIZE() instead of magic value David Gibson
2020-09-08  5:19 ` [PULL 24/33] hw/ppc/ppc4xx_pci: Replace pointless warning by assert() David Gibson
2020-09-08  5:19 ` [PULL 25/33] ppc: introducing spapr_numa.c NUMA code helper David Gibson
2020-09-08  5:19 ` [PULL 26/33] ppc/spapr_nvdimm: turn spapr_dt_nvdimm() static David Gibson
2020-09-08  5:19 ` [PULL 27/33] spapr: introduce SpaprMachineState::numa_assoc_array David Gibson
2020-09-08  5:19 ` [PULL 28/33] spapr, spapr_numa: handle vcpu ibm,associativity David Gibson
2020-09-08  5:19 ` [PULL 29/33] spapr, spapr_numa: move lookup-arrays handling to spapr_numa.c David Gibson
2020-09-08  5:19 ` [PULL 30/33] spapr_numa: move NVLink2 associativity " David Gibson
2020-09-08  5:19 ` [PULL 31/33] spapr: move h_home_node_associativity " David Gibson
2020-09-08  5:19 ` [PULL 32/33] spapr_numa: create a vcpu associativity helper David Gibson
2020-09-08  5:19 ` [PULL 33/33] spapr_numa: use spapr_numa_get_vcpu_assoc() in home_node hcall David Gibson
2020-09-08 16:23 ` [PULL 00/33] ppc-for-5.2 queue 20200908 Peter Maydell

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=20200908051953.1616885-11-david@gibson.dropbear.id.au \
    --to=david@gibson.dropbear.id.au \
    --cc=bauerman@linux.ibm.com \
    --cc=clg@kaod.org \
    --cc=danielhb413@gmail.com \
    --cc=groug@kaod.org \
    --cc=peter.maydell@linaro.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.