All of lore.kernel.org
 help / color / mirror / Atom feed
From: Wei Liu <wei.liu2@citrix.com>
To: xen-devel@lists.xen.org
Cc: Wei Liu <wei.liu2@citrix.com>,
	ian.campbell@citrix.com, andrew.cooper3@citrix.com,
	dario.faggioli@citrix.com, ian.jackson@eu.citrix.com,
	JBeulich@suse.com, ufimtseva@gmail.com
Subject: [PATCH v5 17/24] libxl: build, check and pass vNUMA info to Xen for HVM guest
Date: Thu, 12 Feb 2015 19:44:47 +0000	[thread overview]
Message-ID: <1423770294-9779-18-git-send-email-wei.liu2@citrix.com> (raw)
In-Reply-To: <1423770294-9779-1-git-send-email-wei.liu2@citrix.com>

Transform user supplied vNUMA configuration into libxl internal
representations then libxc representations. Check validity along the
line.

Libxc has more involvement in building vmemranges in HVM case compared
to PV case. The building of vmemranges is placed after xc_hvm_build
returns, because it relies on memory hole information provided by
xc_hvm_build.

Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Cc: Ian Campbell <ian.campbell@citrix.com>
Cc: Ian Jackson <ian.jackson@eu.citrix.com>
Cc: Dario Faggioli <dario.faggioli@citrix.com>
Cc: Elena Ufimtseva <ufimtseva@gmail.com>
---
Changes in v5:
1. Check vnode 0 is large enough to accommodate video ram.

Changes in v4:
1. Adapt to new interface.
2. Rename some variables.
3. Use GCREALLOC_ARRAY.

Changes in v3:
1. Rewrite commit log.
---
 tools/libxc/xc_hvm_build_x86.c |  2 +-
 tools/libxl/libxl_create.c     |  9 +++++++
 tools/libxl/libxl_dom.c        | 37 ++++++++++++++++++++++++++++
 tools/libxl/libxl_internal.h   |  5 ++++
 tools/libxl/libxl_vnuma.c      | 56 ++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 108 insertions(+), 1 deletion(-)

diff --git a/tools/libxc/xc_hvm_build_x86.c b/tools/libxc/xc_hvm_build_x86.c
index a2a3777..bd12e30 100644
--- a/tools/libxc/xc_hvm_build_x86.c
+++ b/tools/libxc/xc_hvm_build_x86.c
@@ -407,7 +407,7 @@ static int setup_guest(xc_interface *xch,
             new_memflags |= XENMEMF_exact_node_request;
         }
 
-        end_pages = args->vmemranges[i].end >> PAGE_SHIFT;
+        end_pages = args->vmemranges[vmemid].end >> PAGE_SHIFT;
         /*
          * Consider vga hole belongs to the vmemrange that covers
          * 0xA0000-0xC0000. Note that 0x00000-0xA0000 is populated just
diff --git a/tools/libxl/libxl_create.c b/tools/libxl/libxl_create.c
index 98687bd..af04248 100644
--- a/tools/libxl/libxl_create.c
+++ b/tools/libxl/libxl_create.c
@@ -853,6 +853,15 @@ static void initiate_domain_create(libxl__egc *egc,
         goto error_out;
     }
 
+    /* Disallow PoD and vNUMA to be enabled at the same time because PoD
+     * pool is not vNUMA-aware yet.
+     */
+    if (pod_enabled && d_config->b_info.num_vnuma_nodes) {
+        ret = ERROR_INVAL;
+        LOG(ERROR, "Cannot enable PoD and vNUMA at the same time");
+        goto error_out;
+    }
+
     ret = libxl__domain_create_info_setdefault(gc, &d_config->c_info);
     if (ret) goto error_out;
 
diff --git a/tools/libxl/libxl_dom.c b/tools/libxl/libxl_dom.c
index 1ff0704..b2c9daf 100644
--- a/tools/libxl/libxl_dom.c
+++ b/tools/libxl/libxl_dom.c
@@ -893,12 +893,49 @@ int libxl__build_hvm(libxl__gc *gc, uint32_t domid,
         goto out;
     }
 
+    if (info->num_vnuma_nodes != 0) {
+        int i;
+
+        args.nr_vmemranges = state->num_vmemranges;
+        args.vmemranges = libxl__malloc(gc, sizeof(*args.vmemranges) *
+                                        args.nr_vmemranges);
+
+        for (i = 0; i < args.nr_vmemranges; i++) {
+            args.vmemranges[i].start = state->vmemranges[i].start;
+            args.vmemranges[i].end   = state->vmemranges[i].end;
+            args.vmemranges[i].flags = state->vmemranges[i].flags;
+            args.vmemranges[i].nid   = state->vmemranges[i].nid;
+        }
+
+        /* Consider video ram belongs to vmemrange 0 -- just shrink it
+         * by the size of video ram.
+         */
+        if (((args.vmemranges[0].end - args.vmemranges[0].start) >> 10)
+            < info->video_memkb) {
+            LOG(ERROR, "vmemrange 0 too small to contain video ram");
+            goto out;
+        }
+
+        args.vmemranges[0].end -= (info->video_memkb << 10);
+    }
+
     ret = xc_hvm_build(ctx->xch, domid, &args);
     if (ret) {
         LOGEV(ERROR, ret, "hvm building failed");
         goto out;
     }
 
+    if (info->num_vnuma_nodes != 0) {
+        ret = libxl__vnuma_build_vmemrange_hvm(gc, domid, info, state, &args);
+        if (ret) {
+            LOGEV(ERROR, ret, "hvm build vmemranges failed");
+            goto out;
+        }
+        ret = libxl__vnuma_config_check(gc, info, state);
+        if (ret) goto out;
+        ret = set_vnuma_info(gc, domid, info, state);
+        if (ret) goto out;
+    }
     ret = hvm_build_set_params(ctx->xch, domid, info, state->store_port,
                                &state->store_mfn, state->console_port,
                                &state->console_mfn, state->store_domid,
diff --git a/tools/libxl/libxl_internal.h b/tools/libxl/libxl_internal.h
index 7d1e1cf..e93089a 100644
--- a/tools/libxl/libxl_internal.h
+++ b/tools/libxl/libxl_internal.h
@@ -3408,6 +3408,11 @@ int libxl__vnuma_build_vmemrange_pv(libxl__gc *gc,
                                     uint32_t domid,
                                     libxl_domain_build_info *b_info,
                                     libxl__domain_build_state *state);
+int libxl__vnuma_build_vmemrange_hvm(libxl__gc *gc,
+                                     uint32_t domid,
+                                     libxl_domain_build_info *b_info,
+                                     libxl__domain_build_state *state,
+                                     struct xc_hvm_build_args *args);
 
 _hidden int libxl__ms_vm_genid_set(libxl__gc *gc, uint32_t domid,
                                    const libxl_ms_vm_genid *id);
diff --git a/tools/libxl/libxl_vnuma.c b/tools/libxl/libxl_vnuma.c
index 3d46239..eefca38 100644
--- a/tools/libxl/libxl_vnuma.c
+++ b/tools/libxl/libxl_vnuma.c
@@ -163,6 +163,62 @@ int libxl__vnuma_build_vmemrange_pv(libxl__gc *gc,
     return libxl__arch_vnuma_build_vmemrange(gc, domid, b_info, state);
 }
 
+/* Build vmemranges for HVM guest */
+int libxl__vnuma_build_vmemrange_hvm(libxl__gc *gc,
+                                     uint32_t domid,
+                                     libxl_domain_build_info *b_info,
+                                     libxl__domain_build_state *state,
+                                     struct xc_hvm_build_args *args)
+{
+    uint64_t hole_start, hole_end, next;
+    int nid, nr_vmemrange;
+    xen_vmemrange_t *vmemranges;
+
+    /* Derive vmemranges from vnode size and memory hole.
+     *
+     * Guest physical address space layout:
+     * [0, hole_start) [hole_start, hole_end) [hole_end, highmem_end)
+     */
+    hole_start = args->lowmem_end < args->mmio_start ?
+        args->lowmem_end : args->mmio_start;
+    hole_end = (args->mmio_start + args->mmio_size) > (1ULL << 32) ?
+        (args->mmio_start + args->mmio_size) : (1ULL << 32);
+
+    assert(state->vmemranges == NULL);
+
+    next = 0;
+    nr_vmemrange = 0;
+    vmemranges = NULL;
+    for (nid = 0; nid < b_info->num_vnuma_nodes; nid++) {
+        libxl_vnode_info *p = &b_info->vnuma_nodes[nid];
+        uint64_t remaining_bytes = p->memkb << 10;
+
+        while (remaining_bytes > 0) {
+            uint64_t count = remaining_bytes;
+
+            if (next >= hole_start && next < hole_end)
+                next = hole_end;
+            if ((next < hole_start) && (next + remaining_bytes >= hole_start))
+                count = hole_start - next;
+
+            GCREALLOC_ARRAY(vmemranges, nr_vmemrange+1);
+            vmemranges[nr_vmemrange].start = next;
+            vmemranges[nr_vmemrange].end = next + count;
+            vmemranges[nr_vmemrange].flags = 0;
+            vmemranges[nr_vmemrange].nid = nid;
+
+            nr_vmemrange++;
+            remaining_bytes -= count;
+            next += count;
+        }
+    }
+
+    state->vmemranges = vmemranges;
+    state->num_vmemranges = nr_vmemrange;
+
+    return 0;
+}
+
 /*
  * Local variables:
  * mode: C
-- 
1.9.1

  parent reply	other threads:[~2015-02-12 19:44 UTC|newest]

Thread overview: 94+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-02-12 19:44 [PATCH v5 00/24] Virtual NUMA for PV and HVM Wei Liu
2015-02-12 19:44 ` [PATCH v5 01/24] xen: dump vNUMA information with debug key "u" Wei Liu
2015-02-13 11:50   ` Andrew Cooper
2015-02-16 14:35     ` Dario Faggioli
2015-02-12 19:44 ` [PATCH v5 02/24] xen: make two memory hypercalls vNUMA-aware Wei Liu
2015-02-13 12:00   ` Andrew Cooper
2015-02-13 13:24     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 03/24] libxc: duplicate snippet to allocate p2m_host array Wei Liu
2015-02-12 19:44 ` [PATCH v5 04/24] libxc: add p2m_size to xc_dom_image Wei Liu
2015-02-16 14:46   ` Dario Faggioli
2015-02-16 14:49     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 05/24] libxc: allocate memory with vNUMA information for PV guest Wei Liu
2015-02-13 14:30   ` Andrew Cooper
2015-02-13 15:05     ` Wei Liu
2015-02-13 15:17       ` Andrew Cooper
2015-02-16 16:58   ` Dario Faggioli
2015-02-16 17:44     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 06/24] libxl: introduce vNUMA types Wei Liu
2015-02-16 14:58   ` Dario Faggioli
2015-02-16 15:17     ` Wei Liu
2015-02-16 15:56       ` Dario Faggioli
2015-02-16 16:11         ` Wei Liu
2015-02-16 16:51           ` Dario Faggioli
2015-02-16 17:38             ` Wei Liu
2015-02-17 10:42               ` Dario Faggioli
2015-02-12 19:44 ` [PATCH v5 07/24] libxl: add vmemrange to libxl__domain_build_state Wei Liu
2015-02-16 16:00   ` Dario Faggioli
2015-02-16 16:15     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 08/24] libxl: introduce libxl__vnuma_config_check Wei Liu
2015-02-13 14:15   ` Ian Jackson
2015-02-13 15:12     ` Wei Liu
2015-02-13 15:39       ` Elena Ufimtseva
2015-02-13 16:06         ` Wei Liu
2015-02-13 16:11           ` Elena Ufimtseva
2015-02-17 16:51             ` Dario Faggioli
2015-02-22 15:50               ` Wei Liu
2015-02-17 16:44       ` Dario Faggioli
2015-02-13 15:40   ` Andrew Cooper
2015-02-17 12:56     ` Wei Liu
2015-03-02 15:13       ` Ian Campbell
2015-03-02 15:25         ` Andrew Cooper
2015-03-02 16:05           ` Ian Campbell
2015-02-17 16:38   ` Dario Faggioli
2015-02-22 15:47     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 09/24] libxl: x86: factor out e820_host_sanitize Wei Liu
2015-02-13 15:42   ` Andrew Cooper
2015-02-16 17:00     ` Dario Faggioli
2015-02-12 19:44 ` [PATCH v5 10/24] libxl: functions to build vmemranges for PV guest Wei Liu
2015-02-13 15:49   ` Andrew Cooper
2015-02-17 14:08     ` Wei Liu
2015-02-17 15:28   ` Dario Faggioli
2015-02-17 15:32     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 11/24] libxl: build, check and pass vNUMA info to Xen " Wei Liu
2015-02-13 15:54   ` Andrew Cooper
2015-02-17 14:49   ` Dario Faggioli
2015-02-12 19:44 ` [PATCH v5 12/24] hvmloader: retrieve vNUMA information from hypervisor Wei Liu
2015-02-13 15:58   ` Andrew Cooper
2015-02-17 11:36   ` Jan Beulich
2015-02-17 11:42     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 13/24] hvmloader: construct SRAT Wei Liu
2015-02-13 16:07   ` Andrew Cooper
2015-02-12 19:44 ` [PATCH v5 14/24] hvmloader: construct SLIT Wei Liu
2015-02-13 16:10   ` Andrew Cooper
2015-02-12 19:44 ` [PATCH v5 15/24] libxc: indentation change to xc_hvm_build_x86.c Wei Liu
2015-02-12 19:44 ` [PATCH v5 16/24] libxc: allocate memory with vNUMA information for HVM guest Wei Liu
2015-02-13 16:22   ` Andrew Cooper
2015-02-12 19:44 ` Wei Liu [this message]
2015-02-13 14:21   ` [PATCH v5 17/24] libxl: build, check and pass vNUMA info to Xen " Ian Jackson
2015-02-13 15:18     ` Wei Liu
2015-02-17 14:26   ` Dario Faggioli
2015-02-17 14:41     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 18/24] libxl: disallow memory relocation when vNUMA is enabled Wei Liu
2015-02-13 14:17   ` Ian Jackson
2015-02-13 15:18     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 19/24] libxl: define LIBXL_HAVE_VNUMA Wei Liu
2015-02-13 14:12   ` Ian Jackson
2015-02-13 15:21     ` Wei Liu
2015-02-13 15:26       ` Ian Jackson
2015-02-13 15:27         ` Ian Jackson
2015-02-13 15:28         ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 20/24] libxlu: rework internal representation of setting Wei Liu
2015-02-13 14:24   ` Ian Jackson
2015-02-12 19:44 ` [PATCH v5 21/24] libxlu: nested list support Wei Liu
2015-02-12 19:44 ` [PATCH v5 22/24] libxlu: introduce new APIs Wei Liu
2015-02-13 14:12   ` Ian Jackson
2015-02-16 19:10     ` Wei Liu
2015-02-16 19:47       ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 23/24] xl: introduce xcalloc Wei Liu
2015-02-12 20:17   ` Andrew Cooper
2015-02-13 10:25     ` Wei Liu
2015-02-12 19:44 ` [PATCH v5 24/24] xl: vNUMA support Wei Liu
2015-02-24 16:19   ` Dario Faggioli
2015-02-24 16:31     ` Wei Liu
2015-02-24 16:44       ` 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=1423770294-9779-18-git-send-email-wei.liu2@citrix.com \
    --to=wei.liu2@citrix.com \
    --cc=JBeulich@suse.com \
    --cc=andrew.cooper3@citrix.com \
    --cc=dario.faggioli@citrix.com \
    --cc=ian.campbell@citrix.com \
    --cc=ian.jackson@eu.citrix.com \
    --cc=ufimtseva@gmail.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 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.