From: Andre Przywara <andre.przywara@amd.com>
To: Avi Kivity <avi@redhat.com>
Cc: kvm@vger.kernel.org, "Daniel P. Berrange" <berrange@redhat.com>
Subject: [PATCH 1/3] v2: KVM-userspace: introduce -numa command line option
Date: Fri, 5 Dec 2008 14:30:46 +0100 [thread overview]
Message-ID: <49392D06.300@amd.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 612 bytes --]
The attached patch parses the command line options given at -numa and
passes it on to lower levels (namely qemu-kvm.c)
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
--
Andre Przywara
AMD-Operating System Research Center (OSRC), Dresden, Germany
Tel: +49 351 277-84917
----to satisfy European Law for business letters:
AMD Saxony Limited Liability Company & Co. KG,
Wilschdorfer Landstr. 101, 01109 Dresden, Germany
Register Court Dresden: HRA 4896, General Partner authorized
to represent: AMD Saxony LLC (Wilmington, Delaware, US)
General Manager of AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy
[-- Attachment #2: kvmnuma_cmdline.patch --]
[-- Type: text/x-patch, Size: 6360 bytes --]
commit 7f9c2380d2d4e79da30dcc219cc4274f3dd36183
Author: Andre Przywara <aprzywar@hagen.osrc.amd.com>
Date: Fri Dec 5 13:36:43 2008 +0100
introduce -numa comand line option
diff --git a/qemu/sysemu.h b/qemu/sysemu.h
index 5abda5c..40868fc 100644
--- a/qemu/sysemu.h
+++ b/qemu/sysemu.h
@@ -99,6 +99,13 @@ extern int win2k_install_hack;
extern int alt_grab;
extern int usb_enabled;
extern int smp_cpus;
+
+#define MAX_NODES 64
+extern int numnumanodes;
+extern uint64_t hostnodes[MAX_NODES];
+extern uint64_t node_mem[MAX_NODES];
+extern uint64_t node_to_cpus[MAX_NODES];
+
extern int cursor_hide;
extern int graphic_rotate;
extern int no_quit;
diff --git a/qemu/vl.c b/qemu/vl.c
index 0eea3b3..27a7cc4 100644
--- a/qemu/vl.c
+++ b/qemu/vl.c
@@ -228,6 +228,10 @@ int usb_enabled = 0;
const char *assigned_devices[MAX_DEV_ASSIGN_CMDLINE];
int assigned_devices_index;
int smp_cpus = 1;
+int numnumanodes = 0;
+uint64_t hostnodes[MAX_NODES];
+uint64_t node_mem[MAX_NODES];
+uint64_t node_to_cpus[MAX_NODES];
const char *vnc_display;
int acpi_enabled = 1;
int fd_bootchk = 1;
@@ -4099,6 +4103,10 @@ static void help(int exitcode)
"-kvm-shadow-memory megs set the amount of shadow pages to be allocated\n"
"-mem-path set the path to hugetlbfs/tmpfs mounted directory, also enables allocation of guest memory with huge pages\n"
"-option-rom rom load a file, rom, into the option ROM space\n"
+ "-numa nrnodes[,mem:size1[;size2..]][,cpu:cpu1[;cpu2..]][,pin:node1[;node2]]\n"
+ " create a multi NUMA node guest and optionally pin it to\n"
+ " to the given host nodes. If mem and cpu are omitted,\n"
+ " resources are split equally\n"
#ifdef TARGET_SPARC
"-prom-env variable=value set OpenBIOS nvram variables\n"
#endif
@@ -4196,6 +4204,7 @@ enum {
QEMU_OPTION_usb,
QEMU_OPTION_usbdevice,
QEMU_OPTION_smp,
+ QEMU_OPTION_numa,
QEMU_OPTION_vnc,
QEMU_OPTION_no_acpi,
QEMU_OPTION_curses,
@@ -4323,6 +4332,7 @@ static const QEMUOption qemu_options[] = {
{ "win2k-hack", 0, QEMU_OPTION_win2k_hack },
{ "usbdevice", HAS_ARG, QEMU_OPTION_usbdevice },
{ "smp", HAS_ARG, QEMU_OPTION_smp },
+ { "numa", HAS_ARG, QEMU_OPTION_numa},
{ "vnc", HAS_ARG, QEMU_OPTION_vnc },
#ifdef CONFIG_CURSES
{ "curses", 0, QEMU_OPTION_curses },
@@ -4714,6 +4724,79 @@ static void termsig_setup(void)
#endif
+#define PARSE_FLAG_BITMASK 1
+#define PARSE_FLAG_SUFFIX 2
+
+static int parse_to_array (const char *arg, uint64_t *array,
+ char delim, int maxentries, int flags)
+{
+const char *s;
+char *end;
+int num;
+unsigned long long int val,endval;
+
+ num=0;
+ for (s=arg; s!=NULL; s = strchr (end, delim))
+ {
+ if (*s == delim) s++;
+ val = strtoull (s, &end, 10);
+ if (end == s) {num++; continue;}
+ if (num >= maxentries) break;
+ switch (*end)
+ {
+ case 'g':
+ case 'G':
+ if (flags & PARSE_FLAG_SUFFIX) val *= 1024;
+ case 'm':
+ case 'M':
+ if (flags & PARSE_FLAG_SUFFIX) val *= 1024;
+ case 'k':
+ case 'K':
+ if (flags & PARSE_FLAG_SUFFIX) val *= 1024;
+ break;
+ case '-':
+ if (!(flags & PARSE_FLAG_BITMASK)) break;
+ s = end + 1;
+ endval = strtoull (s, &end, 10);
+ val = (1 << (endval + 1)) - (1 << val);
+ break;
+ default:
+ if (flags & PARSE_FLAG_BITMASK) val = 1 << val;
+ break;
+ }
+ array[num++] = val;
+ }
+ return num;
+}
+
+static int parse_numa_args (const char *opt, uint64_t *hostnodes,
+ uint64_t *mems, uint64_t *cpus, int maxentries)
+{
+const char *s;
+char *arg, *val, *end;
+int num;
+
+ arg = strdup(opt);
+ s = strtok(arg, ",");
+ if (s == NULL) {free (arg); return -1;}
+ num = strtol (s, &end, 10);
+ if (s == end) {free (arg); return -1;}
+ while ((s=strtok(NULL, ","))!=NULL) {
+ if ((val = strchr (s, ':'))) {
+ *val++ = 0;
+ if (!strcmp (s, "mem")) {
+ parse_to_array (val, mems, ';', maxentries, PARSE_FLAG_SUFFIX);
+ } else if (!strcmp (s, "cpu")) {
+ parse_to_array (val, cpus, ';', maxentries, PARSE_FLAG_BITMASK);
+ } else if (!strcmp (s, "pin")) {
+ parse_to_array (val, hostnodes, ';', maxentries, 0);
+ }
+ }
+ }
+ free (arg);
+ return num;
+}
+
int main(int argc, char **argv)
{
#ifdef CONFIG_GDBSTUB
@@ -4812,6 +4895,12 @@ int main(int argc, char **argv)
parallel_devices[i] = NULL;
parallel_device_index = 0;
+ for(i = 0; i < MAX_NODES; i++) {
+ hostnodes[i] = (uint64_t)-1;
+ node_to_cpus[i] = 0;
+ node_mem[i] = 0;
+ }
+
usb_devices_index = 0;
assigned_devices_index = 0;
@@ -5293,6 +5382,14 @@ int main(int argc, char **argv)
exit(1);
}
break;
+ case QEMU_OPTION_numa:
+ numnumanodes = parse_numa_args (optarg,
+ hostnodes, node_mem, node_to_cpus, MAX_NODES);
+ if (numnumanodes < 0) {
+ fprintf(stderr, "Invalid number of NUMA nodes\n");
+ exit(1);
+ }
+ break;
case QEMU_OPTION_vnc:
vnc_display = optarg;
break;
@@ -5445,6 +5542,24 @@ int main(int argc, char **argv)
monitor_device = "stdio";
}
+ if (numnumanodes > 0) {
+ int i;
+
+ if (numnumanodes > smp_cpus)
+ numnumanodes = smp_cpus;
+
+ for (i = 0; i < numnumanodes; i++) if (node_mem[i] != 0) break;
+ if (i == numnumanodes) {
+ for (i = 0; i < numnumanodes; i++)
+ node_mem[i] = (ram_size / numnumanodes) & ~((1 << 20UL) - 1);
+ }
+ for (i = 0; i < numnumanodes; i++) if (node_to_cpus[i] != 0) break;
+ if (i == numnumanodes) {
+ for (i = 0; i < smp_cpus; i++)
+ node_to_cpus[i % numnumanodes] |= 1<<i;
+ }
+ }
+
#ifndef _WIN32
if (daemonize) {
pid_t pid;
reply other threads:[~2008-12-05 13:30 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=49392D06.300@amd.com \
--to=andre.przywara@amd.com \
--cc=avi@redhat.com \
--cc=berrange@redhat.com \
--cc=kvm@vger.kernel.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.