From: Elena Ufimtseva <ufimtseva@gmail.com>
To: xen-devel@lists.xen.org
Cc: keir@xen.org, Elena Ufimtseva <ufimtseva@gmail.com>,
stefano.stabellini@eu.citrix.com, george.dunlap@eu.citrix.com,
msw@linux.com, dario.faggioli@citrix.com, lccycc123@gmail.com,
JBeulich@suse.com
Subject: [PATCH 1/7] xen: vNUMA support for PV guests
Date: Wed, 16 Oct 2013 18:37:52 -0400 [thread overview]
Message-ID: <1381963072-7703-1-git-send-email-ufimtseva@gmail.com> (raw)
Defines XENMEM subop hypercall for PV vNUMA
enabled guests and data structures that provide vNUMA
topology information from per-domain vnuma topology
build info.
Signed-off-by: Elena Ufimtseva <ufimtseva@gmail.com>
---
Changes since RFC v2:
- fixed code style;
- the memory copying in hypercall happens in one go for arrays;
- fixed error codes logic;
---
xen/common/domain.c | 10 ++++++
xen/common/domctl.c | 72 +++++++++++++++++++++++++++++++++++++++++++
xen/common/memory.c | 41 ++++++++++++++++++++++++
xen/include/public/domctl.h | 14 +++++++++
xen/include/public/memory.h | 8 +++++
xen/include/xen/domain.h | 11 +++++++
xen/include/xen/sched.h | 1 +
xen/include/xen/vnuma.h | 18 +++++++++++
8 files changed, 175 insertions(+)
create mode 100644 xen/include/xen/vnuma.h
diff --git a/xen/common/domain.c b/xen/common/domain.c
index 5999779..d084292 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -539,6 +539,7 @@ int domain_kill(struct domain *d)
tmem_destroy(d->tmem);
domain_set_outstanding_pages(d, 0);
d->tmem = NULL;
+ domain_vnuma_destroy(&d->vnuma);
/* fallthrough */
case DOMDYING_dying:
rc = domain_relinquish_resources(d);
@@ -1287,6 +1288,15 @@ int continue_hypercall_on_cpu(
return 0;
}
+void domain_vnuma_destroy(struct domain_vnuma_info *v)
+{
+ v->nr_vnodes = 0;
+ xfree(v->vnuma_memblks);
+ xfree(v->vcpu_to_vnode);
+ xfree(v->vdistance);
+ xfree(v->vnode_to_pnode);
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/common/domctl.c b/xen/common/domctl.c
index 870eef1..042e2d2 100644
--- a/xen/common/domctl.c
+++ b/xen/common/domctl.c
@@ -29,6 +29,7 @@
#include <asm/page.h>
#include <public/domctl.h>
#include <xsm/xsm.h>
+#include <xen/vnuma.h>
static DEFINE_SPINLOCK(domctl_lock);
DEFINE_SPINLOCK(vcpu_alloc_lock);
@@ -871,6 +872,77 @@ long do_domctl(XEN_GUEST_HANDLE_PARAM(xen_domctl_t) u_domctl)
}
break;
+ case XEN_DOMCTL_setvnumainfo:
+ {
+ unsigned int i, j, nr_vnodes, dist_size;
+ unsigned int dist;
+
+ ret = -EFAULT;
+ dist = i = j = 0;
+
+ nr_vnodes = op->u.vnuma.nr_vnodes;
+ if ( nr_vnodes == 0 )
+ goto err_dom;
+ dist_size = nr_vnodes * nr_vnodes;
+
+ d->vnuma.vdistance = xmalloc_bytes(
+ sizeof(*d->vnuma.vdistance) * dist_size);
+ d->vnuma.vnuma_memblks = xmalloc_bytes(
+ sizeof(*d->vnuma.vnuma_memblks) * nr_vnodes);
+ d->vnuma.vcpu_to_vnode = xmalloc_bytes(
+ sizeof(*d->vnuma.vcpu_to_vnode) * d->max_vcpus);
+ d->vnuma.vnode_to_pnode = xmalloc_bytes(
+ sizeof(*d->vnuma.vnode_to_pnode) * nr_vnodes);
+
+ if ( d->vnuma.vdistance == NULL || d->vnuma.vnuma_memblks == NULL ||
+ d->vnuma.vcpu_to_vnode == NULL ||
+ d->vnuma.vnode_to_pnode == NULL )
+ {
+ ret = -ENOMEM;
+ goto err_dom;
+ }
+
+ d->vnuma.nr_vnodes = nr_vnodes;
+ if ( !guest_handle_is_null(op->u.vnuma.vdistance) )
+ if ( unlikely(copy_from_guest(d->vnuma.vdistance,
+ op->u.vnuma.vdistance,
+ dist_size)) )
+ goto err_dom;
+
+ if ( !guest_handle_is_null(op->u.vnuma.vnuma_memblks) )
+ if ( unlikely(copy_from_guest(d->vnuma.vnuma_memblks,
+ op->u.vnuma.vnuma_memblks,
+ nr_vnodes)))
+ goto err_dom;
+
+ if ( !guest_handle_is_null(op->u.vnuma.vcpu_to_vnode) )
+ if ( unlikely(copy_from_guest(d->vnuma.vcpu_to_vnode,
+ op->u.vnuma.vcpu_to_vnode,
+ d->max_vcpus)) )
+ goto err_dom;
+
+ if ( !guest_handle_is_null(op->u.vnuma.vnode_to_pnode) )
+ {
+ if ( unlikely(copy_from_guest(d->vnuma.vnode_to_pnode,
+ op->u.vnuma.vnode_to_pnode,
+ nr_vnodes)) )
+ goto err_dom;
+
+ }
+ else
+ for( i = 0; i < nr_vnodes; i++ )
+ d->vnuma.vnode_to_pnode[i] = NUMA_NO_NODE;
+ ret = 0;
+err_dom:
+ if (ret != 0) {
+ d->vnuma.nr_vnodes = 0;
+ xfree( d->vnuma.vdistance );
+ xfree( d->vnuma.vnuma_memblks );
+ xfree( d->vnuma.vnode_to_pnode );
+ }
+ }
+ break;
+
default:
ret = arch_do_domctl(op, d, u_domctl);
break;
diff --git a/xen/common/memory.c b/xen/common/memory.c
index 50b740f..e8915e4 100644
--- a/xen/common/memory.c
+++ b/xen/common/memory.c
@@ -28,6 +28,7 @@
#include <public/memory.h>
#include <xsm/xsm.h>
#include <xen/trace.h>
+#include <xen/vnuma.h>
struct memop_args {
/* INPUT */
@@ -732,7 +733,47 @@ long do_memory_op(unsigned long cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
rcu_unlock_domain(d);
break;
+ case XENMEM_get_vnuma_info:
+ {
+ struct vnuma_topology_info mtopology;
+ struct domain *d;
+ unsigned int max_vcpus, nr_vnodes = 0;
+ rc = -EINVAL;
+
+ if( copy_from_guest(&mtopology, arg, 1) )
+ return -EFAULT;
+ if ( (d = rcu_lock_domain_by_any_id(mtopology.domid)) == NULL )
+ return -ESRCH;
+
+ nr_vnodes = d->vnuma.nr_vnodes;
+ max_vcpus = d->max_vcpus;
+
+ if ((nr_vnodes == 0) || (nr_vnodes > max_vcpus))
+ goto vnumaout;
+
+ mtopology.nr_vnodes = nr_vnodes;
+
+ if (copy_to_guest(arg , &mtopology, 1) != 0)
+ goto vnumaout;
+
+ if (copy_to_guest(mtopology.vnuma_memblks,
+ d->vnuma.vnuma_memblks,
+ nr_vnodes) != 0)
+ goto vnumaout;
+ if (copy_to_guest(mtopology.vdistance,
+ d->vnuma.vdistance,
+ nr_vnodes * nr_vnodes) != 0)
+ goto vnumaout;
+ if (copy_to_guest(mtopology.vcpu_to_vnode,
+ d->vnuma.vcpu_to_vnode,
+ max_vcpus) != 0)
+ goto vnumaout;
+ rc = 0;
+vnumaout:
+ rcu_unlock_domain(d);
+ break;
+ }
default:
rc = arch_memory_op(op, arg);
break;
diff --git a/xen/include/public/domctl.h b/xen/include/public/domctl.h
index d4e479f..1adab25 100644
--- a/xen/include/public/domctl.h
+++ b/xen/include/public/domctl.h
@@ -35,6 +35,7 @@
#include "xen.h"
#include "grant_table.h"
#include "hvm/save.h"
+#include "memory.h"
#define XEN_DOMCTL_INTERFACE_VERSION 0x00000009
@@ -863,6 +864,17 @@ struct xen_domctl_set_max_evtchn {
typedef struct xen_domctl_set_max_evtchn xen_domctl_set_max_evtchn_t;
DEFINE_XEN_GUEST_HANDLE(xen_domctl_set_max_evtchn_t);
+struct xen_domctl_vnuma {
+ uint16_t nr_vnodes;
+ XEN_GUEST_HANDLE_64(uint) vdistance;
+ XEN_GUEST_HANDLE_64(uint) vcpu_to_vnode;
+ XEN_GUEST_HANDLE_64(uint) vnode_to_pnode;
+ XEN_GUEST_HANDLE_64(vnuma_memblk_t) vnuma_memblks;
+};
+
+typedef struct xen_domctl_vnuma xen_domctl_vnuma_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_vnuma_t);
+
struct xen_domctl {
uint32_t cmd;
#define XEN_DOMCTL_createdomain 1
@@ -932,6 +944,7 @@ struct xen_domctl {
#define XEN_DOMCTL_setnodeaffinity 68
#define XEN_DOMCTL_getnodeaffinity 69
#define XEN_DOMCTL_set_max_evtchn 70
+#define XEN_DOMCTL_setvnumainfo 71
#define XEN_DOMCTL_gdbsx_guestmemio 1000
#define XEN_DOMCTL_gdbsx_pausevcpu 1001
#define XEN_DOMCTL_gdbsx_unpausevcpu 1002
@@ -992,6 +1005,7 @@ struct xen_domctl {
struct xen_domctl_set_broken_page_p2m set_broken_page_p2m;
struct xen_domctl_gdbsx_pauseunp_vcpu gdbsx_pauseunp_vcpu;
struct xen_domctl_gdbsx_domstatus gdbsx_domstatus;
+ struct xen_domctl_vnuma vnuma;
uint8_t pad[128];
} u;
};
diff --git a/xen/include/public/memory.h b/xen/include/public/memory.h
index 7a26dee..93a17e9 100644
--- a/xen/include/public/memory.h
+++ b/xen/include/public/memory.h
@@ -459,6 +459,14 @@ DEFINE_XEN_GUEST_HANDLE(xen_mem_sharing_op_t);
* The zero value is appropiate.
*/
+struct vnuma_memblk {
+ uint64_t start, end;
+};
+typedef struct vnuma_memblk vnuma_memblk_t;
+DEFINE_XEN_GUEST_HANDLE(vnuma_memblk_t);
+
+#define XENMEM_get_vnuma_info 25
+
#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
#endif /* __XEN_PUBLIC_MEMORY_H__ */
diff --git a/xen/include/xen/domain.h b/xen/include/xen/domain.h
index a057069..c9d53e3 100644
--- a/xen/include/xen/domain.h
+++ b/xen/include/xen/domain.h
@@ -4,6 +4,7 @@
#include <public/xen.h>
#include <asm/domain.h>
+#include <public/memory.h>
typedef union {
struct vcpu_guest_context *nat;
@@ -89,4 +90,14 @@ extern unsigned int xen_processor_pmbits;
extern bool_t opt_dom0_vcpus_pin;
+struct domain_vnuma_info {
+ uint16_t nr_vnodes;
+ uint *vdistance;
+ uint *vcpu_to_vnode;
+ uint *vnode_to_pnode;
+ vnuma_memblk_t *vnuma_memblks;
+};
+
+void domain_vnuma_destroy(struct domain_vnuma_info *v);
+
#endif /* __XEN_DOMAIN_H__ */
diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h
index 25bf637..6693b35 100644
--- a/xen/include/xen/sched.h
+++ b/xen/include/xen/sched.h
@@ -408,6 +408,7 @@ struct domain
nodemask_t node_affinity;
unsigned int last_alloc_node;
spinlock_t node_affinity_lock;
+ struct domain_vnuma_info vnuma;
};
struct domain_setup_info
diff --git a/xen/include/xen/vnuma.h b/xen/include/xen/vnuma.h
new file mode 100644
index 0000000..2637476
--- /dev/null
+++ b/xen/include/xen/vnuma.h
@@ -0,0 +1,18 @@
+#ifndef _VNUMA_H
+#define _VNUMA_H
+#include <public/memory.h>
+
+/* DEFINE_XEN_GUEST_HANDLE(vnuma_memblk_t); */
+
+struct vnuma_topology_info {
+ domid_t domid;
+ uint16_t nr_vnodes;
+ uint32_t _pad;
+ XEN_GUEST_HANDLE_64(uint) vdistance;
+ XEN_GUEST_HANDLE_64(uint) vcpu_to_vnode;
+ XEN_GUEST_HANDLE_64(vnuma_memblk_t) vnuma_memblks;
+};
+typedef struct vnuma_topology_info vnuma_topology_info_t;
+DEFINE_XEN_GUEST_HANDLE(vnuma_topology_info_t);
+
+#endif
--
1.7.10.4
next reply other threads:[~2013-10-16 22:37 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-10-16 22:37 Elena Ufimtseva [this message]
2013-10-17 9:05 ` [PATCH 1/7] xen: vNUMA support for PV guests Jan Beulich
2013-10-22 14:31 ` Li Yechen
2013-11-21 10:00 ` Li Yechen
2013-11-22 18:29 ` Dario Faggioli
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=1381963072-7703-1-git-send-email-ufimtseva@gmail.com \
--to=ufimtseva@gmail.com \
--cc=JBeulich@suse.com \
--cc=dario.faggioli@citrix.com \
--cc=george.dunlap@eu.citrix.com \
--cc=keir@xen.org \
--cc=lccycc123@gmail.com \
--cc=msw@linux.com \
--cc=stefano.stabellini@eu.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).