xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Juergen Gross <jgross@suse.com>
To: xen-devel@lists.xen.org
Cc: Juergen Gross <jgross@suse.com>, Wei Liu <wei.liu2@citrix.com>,
	Ian Jackson <ian.jackson@eu.citrix.com>,
	Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Subject: [PATCH v4 3/3] libxl: add force option for xl vcpu-pin
Date: Thu, 10 Mar 2016 06:27:14 +0100	[thread overview]
Message-ID: <1457587634-22819-4-git-send-email-jgross@suse.com> (raw)
In-Reply-To: <1457587634-22819-1-git-send-email-jgross@suse.com>

In order to be able to undo a vcpu pin override in case of a kernel
driver error add a flag "-f" to the "xl vcpu-pin" command forcing the
hypervisor to undo the override.

Cc: Ian Jackson <ian.jackson@eu.citrix.com>
Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Cc: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
---
V4: - change "force" variable to bool as suggested by Dario Faggioli
    - add #define to libxl indicating availability of
      libxl_set_vcpuaffinity_force() as requested by Wei Liu
    - add force option to xl man page as requested by Dario Faggioli
---
 docs/man/xl.pod.1         |  7 ++++++-
 tools/libxl/libxl.c       | 31 +++++++++++++++++++++++++------
 tools/libxl/libxl.h       | 10 ++++++++++
 tools/libxl/xl_cmdimpl.c  | 26 +++++++++++++++++++++++---
 tools/libxl/xl_cmdtable.c |  3 ++-
 5 files changed, 66 insertions(+), 11 deletions(-)

diff --git a/docs/man/xl.pod.1 b/docs/man/xl.pod.1
index 4279c7c..aedcaea 100644
--- a/docs/man/xl.pod.1
+++ b/docs/man/xl.pod.1
@@ -689,7 +689,7 @@ after B<vcpu-set>, go to B<SEE ALSO> section for information.
 Lists VCPU information for a specific domain.  If no domain is
 specified, VCPU information for all domains will be provided.
 
-=item B<vcpu-pin> I<domain-id> I<vcpu> I<cpus hard> I<cpus soft>
+=item B<vcpu-pin> [I<-f|--force>] I<domain-id> I<vcpu> I<cpus hard> I<cpus soft>
 
 Set hard and soft affinity for a I<vcpu> of <domain-id>. Normally VCPUs
 can float between available CPUs whenever Xen deems a different run state
@@ -716,6 +716,11 @@ leaving its hard affinity untouched. On the othe hand:
 will set both hard and soft affinity, the former to pCPUs 3 and 4, the
 latter to pCPUs 6,7,8, and 9.
 
+Specifying I<-f> or I<--force> will remove a temporary pinning done by the
+operating system (normally this should be done by the operating system).
+In case a temporary pinning is active for a vcpu the affinity of this vcpu
+can't be changed without this option.
+
 =item B<vm-list>
 
 Prints information about guests. This list excludes information about
diff --git a/tools/libxl/libxl.c b/tools/libxl/libxl.c
index 4cdc169..53f0100 100644
--- a/tools/libxl/libxl.c
+++ b/tools/libxl/libxl.c
@@ -5301,18 +5301,20 @@ err:
     return NULL;
 }
 
-int libxl_set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid, uint32_t vcpuid,
-                           const libxl_bitmap *cpumap_hard,
-                           const libxl_bitmap *cpumap_soft)
+static int libxl__set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid,
+                                   uint32_t vcpuid,
+                                   const libxl_bitmap *cpumap_hard,
+                                   const libxl_bitmap *cpumap_soft,
+                                   unsigned flags)
 {
     GC_INIT(ctx);
     libxl_bitmap hard, soft;
-    int rc, flags = 0;
+    int rc;
 
     libxl_bitmap_init(&hard);
     libxl_bitmap_init(&soft);
 
-    if (!cpumap_hard && !cpumap_soft) {
+    if (!cpumap_hard && !cpumap_soft && !flags) {
         rc = ERROR_INVAL;
         goto out;
     }
@@ -5327,7 +5329,7 @@ int libxl_set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid, uint32_t vcpuid,
             goto out;
 
         libxl__bitmap_copy_best_effort(gc, &hard, cpumap_hard);
-        flags = XEN_VCPUAFFINITY_HARD;
+        flags |= XEN_VCPUAFFINITY_HARD;
     }
     if (cpumap_soft) {
         rc = libxl_cpu_bitmap_alloc(ctx, &soft, 0);
@@ -5378,6 +5380,23 @@ int libxl_set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid, uint32_t vcpuid,
     return rc;
 }
 
+int libxl_set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid, uint32_t vcpuid,
+                           const libxl_bitmap *cpumap_hard,
+                           const libxl_bitmap *cpumap_soft)
+{
+    return libxl__set_vcpuaffinity(ctx, domid, vcpuid, cpumap_hard,
+                                   cpumap_soft, 0);
+}
+
+int libxl_set_vcpuaffinity_force(libxl_ctx *ctx, uint32_t domid,
+                                 uint32_t vcpuid,
+                                 const libxl_bitmap *cpumap_hard,
+                                 const libxl_bitmap *cpumap_soft)
+{
+    return libxl__set_vcpuaffinity(ctx, domid, vcpuid, cpumap_hard,
+                                   cpumap_soft, XEN_VCPUAFFINITY_FORCE);
+}
+
 int libxl_set_vcpuaffinity_all(libxl_ctx *ctx, uint32_t domid,
                                unsigned int max_vcpus,
                                const libxl_bitmap *cpumap_hard,
diff --git a/tools/libxl/libxl.h b/tools/libxl/libxl.h
index f9e3ef5..d587e0c 100644
--- a/tools/libxl/libxl.h
+++ b/tools/libxl/libxl.h
@@ -148,6 +148,12 @@
 #define LIBXL_HAVE_VCPUINFO_SOFT_AFFINITY 1
 
 /*
+ * LIBXL_HAVE_SET_VCPUAFFINITY_FORCE indicates that the
+ * libxl_set_vcpuaffinity_force() library call is available.
+ */
+#define LIBXL_HAVE_SET_VCPUAFFINITY_FORCE 1
+
+/*
  * LIBXL_HAVE_DEVICE_DISK_DIRECT_IO_SAFE indicates that a
  * 'direct_io_safe' field (of boolean type) is present in
  * libxl_device_disk.
@@ -1715,6 +1721,10 @@ int libxl_get_physinfo(libxl_ctx *ctx, libxl_physinfo *physinfo);
 int libxl_set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid, uint32_t vcpuid,
                            const libxl_bitmap *cpumap_hard,
                            const libxl_bitmap *cpumap_soft);
+int libxl_set_vcpuaffinity_force(libxl_ctx *ctx, uint32_t domid,
+                                 uint32_t vcpuid,
+                                 const libxl_bitmap *cpumap_hard,
+                                 const libxl_bitmap *cpumap_soft);
 int libxl_set_vcpuaffinity_all(libxl_ctx *ctx, uint32_t domid,
                                unsigned int max_vcpus,
                                const libxl_bitmap *cpumap_hard,
diff --git a/tools/libxl/xl_cmdimpl.c b/tools/libxl/xl_cmdimpl.c
index 411473d..51fa487 100644
--- a/tools/libxl/xl_cmdimpl.c
+++ b/tools/libxl/xl_cmdimpl.c
@@ -5344,6 +5344,10 @@ int main_vcpulist(int argc, char **argv)
 
 int main_vcpupin(int argc, char **argv)
 {
+    static struct option opts[] = {
+        {"force", 0, 0, 'f'},
+        COMMON_LONG_OPTS
+    };
     libxl_vcpuinfo *vcpuinfo;
     libxl_bitmap cpumap_hard, cpumap_soft;;
     libxl_bitmap *soft = &cpumap_soft, *hard = &cpumap_hard;
@@ -5356,12 +5360,17 @@ int main_vcpupin(int argc, char **argv)
     const char *vcpu, *hard_str, *soft_str;
     char *endptr;
     int opt, nb_cpu, nb_vcpu, rc = EXIT_FAILURE;
+    bool force = false;
 
     libxl_bitmap_init(&cpumap_hard);
     libxl_bitmap_init(&cpumap_soft);
 
-    SWITCH_FOREACH_OPT(opt, "", NULL, "vcpu-pin", 3) {
-        /* No options */
+    SWITCH_FOREACH_OPT(opt, "f", opts, "vcpu-pin", 3) {
+    case 'f':
+        force = true;
+        break;
+    default:
+        break;
     }
 
     domid = find_domain(argv[optind]);
@@ -5376,6 +5385,10 @@ int main_vcpupin(int argc, char **argv)
             fprintf(stderr, "Error: Invalid argument %s as VCPU.\n", vcpu);
             goto out;
         }
+        if (force) {
+            fprintf(stderr, "Error: --force and 'all' as VCPU not allowed.\n");
+            goto out;
+        }
         vcpuid = -1;
     }
 
@@ -5437,7 +5450,14 @@ int main_vcpupin(int argc, char **argv)
         goto out;
     }
 
-    if (vcpuid != -1) {
+    if (force) {
+        if (libxl_set_vcpuaffinity_force(ctx, domid, vcpuid, hard, soft)) {
+            fprintf(stderr, "Could not set affinity for vcpu `%ld'.\n",
+                    vcpuid);
+            goto out;
+        }
+    }
+    else if (vcpuid != -1) {
         if (libxl_set_vcpuaffinity(ctx, domid, vcpuid, hard, soft)) {
             fprintf(stderr, "Could not set affinity for vcpu `%ld'.\n",
                     vcpuid);
diff --git a/tools/libxl/xl_cmdtable.c b/tools/libxl/xl_cmdtable.c
index fdc1ac6..7cc0401 100644
--- a/tools/libxl/xl_cmdtable.c
+++ b/tools/libxl/xl_cmdtable.c
@@ -222,7 +222,8 @@ struct cmd_spec cmd_table[] = {
     { "vcpu-pin",
       &main_vcpupin, 1, 1,
       "Set which CPUs a VCPU can use",
-      "<Domain> <VCPU|all> <Hard affinity|-|all> <Soft affinity|-|all>",
+      "[option] <Domain> <VCPU|all> <Hard affinity|-|all> <Soft affinity|-|all>",
+      "-f, --force        undo an override pinning done by the kernel",
     },
     { "vcpu-set",
       &main_vcpuset, 0, 1,
-- 
2.6.2


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

  parent reply	other threads:[~2016-03-10  5:27 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-10  5:27 [PATCH v4 0/3] add hypercall option to temporarily pin a vcpu Juergen Gross
2016-03-10  5:27 ` [PATCH v4 1/3] libxc: do some retries in xc_cpupool_removecpu() for EBUSY case Juergen Gross
2016-03-10 10:26   ` Dario Faggioli
2016-03-10 17:16   ` Wei Liu
2016-04-12 13:02   ` Olaf Hering
2016-04-12 13:45     ` Juergen Gross
2016-04-12 13:50       ` Olaf Hering
2016-04-12 13:57       ` Wei Liu
2016-04-14 16:04   ` Ian Jackson
2016-03-10  5:27 ` [PATCH v4 2/3] libxl: print message how to recover from xl cpupool-cpu-remove errors Juergen Gross
2016-04-14 16:06   ` Ian Jackson
2016-04-14 17:10     ` Dario Faggioli
2016-04-14 17:22       ` Ian Jackson
2016-04-14 17:27         ` Dario Faggioli
2016-03-10  5:27 ` Juergen Gross [this message]
2016-03-10 10:25   ` [PATCH v4 3/3] libxl: add force option for xl vcpu-pin Dario Faggioli
2016-03-10 17:17   ` Wei Liu
2016-04-14 16:10   ` Ian Jackson
2016-04-14 17:18     ` Dario Faggioli
2016-03-22  7:36 ` [PATCH v4 0/3] add hypercall option to temporarily pin a vcpu Juergen Gross
2016-03-24 14:03   ` Wei Liu
2016-03-24 17:58     ` Juergen Gross
2016-03-24 19:41       ` Wei Liu
2016-03-30 10:28         ` Juergen Gross
2016-03-30 14:23           ` Konrad Rzeszutek Wilk
2016-04-14 16:11             ` Ian Jackson
2016-04-14 19:15               ` Konrad Rzeszutek Wilk

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=1457587634-22819-4-git-send-email-jgross@suse.com \
    --to=jgross@suse.com \
    --cc=ian.jackson@eu.citrix.com \
    --cc=stefano.stabellini@eu.citrix.com \
    --cc=wei.liu2@citrix.com \
    --cc=xen-devel@lists.xen.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).