Skip to contentSkip to conversation

xen/vNUMA: Per-domain vNUMA initialization.

RFC PATCH In review Elena Ufimtseva 3 participants 3 messages Sep 13, 2013 – Sep 16, 2013

lore
Share & tools
lore
#0 lore raw

RFC PATCH v2 2/7 xen/vNUMA: Per-domain vNUMA initialization.

Per-domain vNUMA topology initialization.
domctl hypercall is used to set vNUMA topology
per domU during domain build time.

---
Changes since v1:

* uses hypercall_bounce_pre/post to initialize pointers before issuing
domctl hypercall;
* fixed type int to unsigned int;

---
 tools/libxc/xc_dom.h    |    8 +++++++
 tools/libxc/xc_domain.c |   57 +++++++++++++++++++++++++++++++++++++++++++++++
 tools/libxc/xenctrl.h   |    9 ++++++++
 3 files changed, 74 insertions(+)
tools/libxc/xc_dom.h +8 -0
index 86e23ee..790f145 100644
@@ -114,6 +114,14 @@ struct xc_dom_image {
     struct xc_dom_phys *phys_pages;
     int realmodearea_log;
 
+    /* vNUMA topology and memory allocation structure
+     * Defines the way to allocate XEN
+     * memory from phys NUMA nodes by providing mask
+     * vnuma_to_pnuma */
+    uint16_t nr_vnodes;
+    uint64_t *vmemsizes;
+    unsigned int *vnode_to_pnode;
+
     /* malloc memory pool */
     struct xc_dom_mem *memblocks;
 
tools/libxc/xc_domain.c +57 -0

63 lines not on this page · raw message

tools/libxc/xenctrl.h +9 -0
index f2cebaf..41f1233 100644
@@ -1083,6 +1083,15 @@ int xc_domain_set_memmap_limit(xc_interface *xch,
                                uint32_t domid,
                                unsigned long map_limitkb);
 
+int xc_domain_setvnodes(xc_interface *xch,
+                        uint32_t domid,
+                        uint16_t nr_vnodes,
+                        uint16_t nr_vcpus,
+                        vnuma_memblk_t *vmemareas,
+                        unsigned int *vdistance,
+                        unsigned int *vcpu_to_vnode,
+                        unsigned int *vnode_to_pnode);
+
 #if defined(__i386__) || defined(__x86_64__)
 /*
  * PC BIOS standard E820 types and structure.
Signature · 1 line
1.7.10.4
#1 lore raw

BTW, Elena, you got Matt's address wrong in most of the patch e-mails. This one is fine, but all the others have <sw_AT_linux.com>, instead of msw, which then bounces! :-P

Also, the threading. As I'm sure you've seen, patches are usually sent as replies to the cover letter. Actually, I'm quite sure it was like that in your previous submission, while it is not this time.

Not sure what went wrong, and no big deal... just keep an eye on it, ok? :-P

Dario

Signature · 4 lines
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xen.org/xen-devel

#2 lore raw
In reply to Elena Ufimtseva
Quoted 1 line · Elena Ufimtseva

Per-domain vNUMA topology initialization.

The one-line summary should have more hints about what code this touches, so the right people notice it; for example:
vNUMA: libxc plumbing to set guest vnodes

Also, you should cc the tools maintainers, Ian Campbell and Ian Jackson.

Quoted 27 lines
tools/libxc/xc_domain.c +19 -0
index 3257e2a..8837fcc 100644
@@ -1629,6 +1629,63 @@ int xc_domain_set_virq_handler(xc_interface *xch, uint32_t domid, int virq)
     return do_domctl(xch, &domctl);
 }
 
+/* Informs XEN that domain is vNUMA aware */
+int xc_domain_setvnodes(xc_interface *xch,
+                        uint32_t domid,
+                        uint16_t nr_vnodes,
+                        uint16_t nr_vcpus,
+                        vnuma_memblk_t *vmemblks,
+                        unsigned int *vdistance,
+                        unsigned int *vcpu_to_vnode,
+                        unsigned int *vnode_to_pnode)
+{
+    int rc;
+    DECLARE_DOMCTL;
+    DECLARE_HYPERCALL_BOUNCE(vmemblks, sizeof(*vmemblks) * nr_vnodes,
+                                    XC_HYPERCALL_BUFFER_BOUNCE_IN);
+    DECLARE_HYPERCALL_BOUNCE(vdistance, sizeof(*vdistance) * nr_vnodes * nr_vnodes,
+                                    XC_HYPERCALL_BUFFER_BOUNCE_IN);
+    DECLARE_HYPERCALL_BOUNCE(vcpu_to_vnode, sizeof(*vcpu_to_vnode) * nr_vcpus,
+                                    XC_HYPERCALL_BUFFER_BOUNCE_IN);
+    DECLARE_HYPERCALL_BOUNCE(vnode_to_pnode, sizeof(*vnode_to_pnode) * nr_vnodes,

tools/libxc/xc_domain.c int xc_domain_set_virq_handler(xc_interface *xch, uint32_t domid, int virq) · line 1660

Quoted 10 lines
tools/libxc/xc_domain.c +10 -0
+                                    XC_HYPERCALL_BUFFER_BOUNCE_IN);
+
+    if ( vdistance == NULL || vcpu_to_vnode == NULL || vmemblks == NULL )
+    /* vnode_to_pnode can be null on non-NUMA machines */
+    {
+            PERROR("Parameters are wrong XEN_DOMCTL_setvnumainfo\n");
+            return -EINVAL;
+    }
+
+    rc = -EINVAL;

You might as well be consistent here and set this above, and then do "return rc" rather than "return -EINVAL".

-George