[Qemu-devel] [PATCH 0/4] Introduce global config and default devices

[Qemu-devel] [PATCH 0/4] Introduce global config and default devices

[Anthony Liguori]

This series introduces global config files stored in /etc/qemu.  There is both
a common config (qemu.conf) and a per-target config (target-<TARGET_NAME>.conf).

To demonstrate what can be done with global config, I've also made some
enhancements to the default device code that allows many of the builtin qemu
defaults to be overridden.  For instance, the following config file would always
disable the monitor on stdio and use virtio-net with tun/tap instead of
slirp.

/etc/qemu/qemu.conf:

[default]
  monitor = "stdio"

[net]
 type = "nic"
 default = "on"
 model = "virtio"

[net]
 type = "tap"
 default = "on"

[Qemu-devel] [PATCH 1/4] Support --confdir in configure to specify path to configuration files

[Anthony Liguori]

The default value is ${prefix}/etc/qemu.  --confdir can be used to override the
default to an absolute path.  The expectation is that when installed to /usr,
 --confdir=/etc/qemu will be used.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
---
 configure |   10 ++++++++++
 1 files changed, 10 insertions(+), 0 deletions(-)

diff --git a/configure b/configure
index 5631bbb..e86b27e 100755
--- a/configure
+++ b/configure
@@ -32,6 +32,7 @@ cpu=""
 prefix=""
 interp_prefix="/usr/gnemul/qemu-%M"
 static="no"
+confdir=""
 sparc_cpu=""
 cross_prefix=""
 cc="gcc"
@@ -453,6 +454,8 @@ for opt do
   ;;
   --static) static="yes"
   ;;
+  --confdir) confdir="$optarg"
+  ;;
   --disable-sdl) sdl="no"
   ;;
   --enable-sdl) sdl="yes"
@@ -686,6 +689,7 @@ echo "  --extra-ldflags=LDFLAGS  append extra linker flags LDFLAGS"
 echo "  --make=MAKE              use specified make [$make]"
 echo "  --install=INSTALL        use specified install [$install]"
 echo "  --static                 enable static build [$static]"
+echo "  --confdir=PATH           install config in PATH"
 echo "  --enable-debug-tcg       enable TCG debugging"
 echo "  --disable-debug-tcg      disable TCG debugging (default)"
 echo "  --enable-debug           enable common debug build options"
@@ -1828,6 +1832,7 @@ if test "$mingw32" = "yes" ; then
   fi
   mansuffix=""
   datasuffix=""
+  confsuffix=""
   docsuffix=""
   binsuffix=""
 else
@@ -1838,6 +1843,9 @@ else
   datasuffix="/share/qemu"
   docsuffix="/share/doc/qemu"
   binsuffix="/bin"
+  if test -z "$confdir" ; then
+      confdir="${prefix}/etc/qemu"
+  fi
 fi
 
 echo "Install prefix    $prefix"
@@ -1914,6 +1922,7 @@ printf " '%s'" "$0" "$@" >> $config_host_mak
 echo >> $config_host_mak
 
 echo "CONFIG_QEMU_SHAREDIR=\"$prefix$datasuffix\"" >> $config_host_mak
+echo "CONFIG_QEMU_CONFDIR=\"$confdir\"" >> $config_host_mak
 
 case "$cpu" in
   i386|x86_64|alpha|cris|hppa|ia64|m68k|microblaze|mips|mips64|ppc|ppc64|s390|s390x|sparc|sparc64)
@@ -2159,6 +2168,7 @@ echo "prefix=$prefix" >> $config_host_mak
 echo "bindir=\${prefix}$binsuffix" >> $config_host_mak
 echo "mandir=\${prefix}$mansuffix" >> $config_host_mak
 echo "datadir=\${prefix}$datasuffix" >> $config_host_mak
+echo "confdir=$confdir" >> $config_host_mak
 echo "docdir=\${prefix}$docsuffix" >> $config_host_mak
 echo "MAKE=$make" >> $config_host_mak
 echo "INSTALL=$install" >> $config_host_mak
-- 
1.6.5.2

[Qemu-devel] [PATCH 2/4] Load global config files by default

[Anthony Liguori]

A new option, -nodefconfig is introduced to prevent loading from the default
config location.  Otherwise, two configuration files will be searched for,
qemu.conf and target-<TARGET_NAME>.conf.

The ordering is a little troubling.  Command line options are parsed before
loading the default configs which means that the command line configs will be
loaded before the default config.  The effect is that the default config will
override -readconfig directives.

It's unclear the best way to handle this.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
---
 qemu-options.hx |    7 +++++++
 vl.c            |   23 +++++++++++++++++++++++
 2 files changed, 30 insertions(+), 0 deletions(-)

diff --git a/qemu-options.hx b/qemu-options.hx
index ee60d8a..57f453d 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -1965,6 +1965,13 @@ STEXI
 @item -writeconfig @var{file}
 Write device configuration to @var{file}.
 ETEXI
+DEF("nodefconfig", 0, QEMU_OPTION_nodefconfig,
+    "-nodefconfig\n"
+    "                do not load default config file\n")
+STEXI
+@item -nodefconfig
+Do not load default config files.
+ETEXI
 
 HXCOMM This is the last statement. Insert new options before this line!
 STEXI
diff --git a/vl.c b/vl.c
index e070ec9..cf12ab0 100644
--- a/vl.c
+++ b/vl.c
@@ -4690,6 +4690,7 @@ int main(int argc, char **argv, char **envp)
 #endif
     CPUState *env;
     int show_vnc_port = 0;
+    int defconfig = 1;
 
     init_clocks();
 
@@ -5457,6 +5458,9 @@ int main(int argc, char **argv, char **envp)
                     fclose(fp);
                     break;
                 }
+            case QEMU_OPTION_nodefconfig:
+                defconfig=0;
+                break;
             }
         }
     }
@@ -5471,6 +5475,25 @@ int main(int argc, char **argv, char **envp)
         data_dir = CONFIG_QEMU_SHAREDIR;
     }
 
+    if (defconfig) {
+        FILE *fp;
+        fp = fopen(CONFIG_QEMU_CONFDIR "/qemu.conf", "r");
+        if (fp) {
+            if (qemu_config_parse(fp) != 0) {
+                exit(1);
+            }
+            fclose(fp);
+        }
+
+        fp = fopen(CONFIG_QEMU_CONFDIR "/target-" TARGET_ARCH ".conf", "r");
+        if (fp) {
+            if (qemu_config_parse(fp) != 0) {
+                exit(1);
+            }
+            fclose(fp);
+        }
+    }
+
     /*
      * Default to max_cpus = smp_cpus, in case the user doesn't
      * specify a max_cpus value.
-- 
1.6.5.2

[Qemu-devel] [PATCH 3/4] Add -defaults option to allow default devices to be overridden

[Anthony Liguori]

This option can be used to toggle whether each default device is enabled or
disabled.  For character devices, the default backend can also be overridden.

For devices, we'll have to take a different approach to changing the defaults
which will be covered in the next patch.

N.B. I took special care with -nographic.  Now -nographic pretty clearly acts
as a mechanism to override the default backend devices.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
---
 qemu-config.c   |   45 +++++++++++++++++++++++++++++++++
 qemu-config.h   |    1 +
 qemu-options.hx |    7 +++++
 vl.c            |   75 +++++++++++++++++++++++++++++++++++++++++--------------
 4 files changed, 109 insertions(+), 19 deletions(-)

diff --git a/qemu-config.c b/qemu-config.c
index c3203c8..82ca399 100644
--- a/qemu-config.c
+++ b/qemu-config.c
@@ -242,6 +242,50 @@ QemuOptsList qemu_mon_opts = {
     },
 };
 
+QemuOptsList qemu_default_opts = {
+    .name = "default",
+    .head = QTAILQ_HEAD_INITIALIZER(qemu_default_opts.head),
+    .desc = {
+        {
+            .name = "serial",
+            .type = QEMU_OPT_STRING,
+        },
+        {
+            .name = "parallel",
+            .type = QEMU_OPT_STRING,
+        },
+        {
+            .name = "virtcon",
+            .type = QEMU_OPT_STRING,
+        },
+        {
+            .name = "monitor",
+            .type = QEMU_OPT_STRING,
+        },
+        {
+            .name = "vga",
+            .type = QEMU_OPT_BOOL,
+        },
+        {
+            .name = "net",
+            .type = QEMU_OPT_BOOL,
+        },
+        {
+            .name = "floppy",
+            .type = QEMU_OPT_BOOL,
+        },
+        {
+            .name = "cdrom",
+            .type = QEMU_OPT_BOOL,
+        },
+        {
+            .name = "sdcard",
+            .type = QEMU_OPT_BOOL,
+        },
+        { /* end if list */ }
+    },
+};
+
 static QemuOptsList *lists[] = {
     &qemu_drive_opts,
     &qemu_chardev_opts,
@@ -250,6 +294,7 @@ static QemuOptsList *lists[] = {
     &qemu_net_opts,
     &qemu_rtc_opts,
     &qemu_global_opts,
+    &qemu_default_opts,
     &qemu_mon_opts,
     NULL,
 };
diff --git a/qemu-config.h b/qemu-config.h
index dd89ae4..14ed67b 100644
--- a/qemu-config.h
+++ b/qemu-config.h
@@ -9,6 +9,7 @@ extern QemuOptsList qemu_net_opts;
 extern QemuOptsList qemu_rtc_opts;
 extern QemuOptsList qemu_global_opts;
 extern QemuOptsList qemu_mon_opts;
+extern QemuOptsList qemu_default_opts;
 
 int qemu_set_option(const char *str);
 int qemu_global_option(const char *str);
diff --git a/qemu-options.hx b/qemu-options.hx
index 57f453d..e81ecb5 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -1919,6 +1919,13 @@ STEXI
 Don't create default devices.
 ETEXI
 
+DEF("default", HAS_ARG, QEMU_OPTION_default, \
+    "-default arg    specify default devices\n")
+STEXI
+@item -defaults
+Override builtin default devices
+ETEXI
+
 #ifndef _WIN32
 DEF("chroot", HAS_ARG, QEMU_OPTION_chroot, \
     "-chroot dir     chroot to dir just before starting the VM\n")
diff --git a/vl.c b/vl.c
index cf12ab0..6a3529e 100644
--- a/vl.c
+++ b/vl.c
@@ -280,6 +280,11 @@ static int default_floppy = 1;
 static int default_cdrom = 1;
 static int default_sdcard = 1;
 
+static const char *default_serial_opt = "vc:80Cx24C";
+static const char *default_parallel_opt = "vc:80Cx24C";
+static const char *default_monitor_opt = "vc:80Cx24C";
+static const char *default_virtcon_opt = "vc:80Cx24C";
+
 static struct {
     const char *driver;
     int *flag;
@@ -4658,6 +4663,32 @@ static int debugcon_parse(const char *devname)
     return 0;
 }
 
+static int default_opt_init(QemuOpts *opts, void *dummy)
+{
+    const char *opt;
+
+    if ((opt = qemu_opt_get(opts, "serial"))) {
+        default_serial_opt = opt;
+    }
+    if ((opt = qemu_opt_get(opts, "parallel"))) {
+        default_parallel_opt = opt;
+    }
+    if ((opt = qemu_opt_get(opts, "virtcon"))) {
+        default_virtcon_opt = opt;
+    }
+    if ((opt = qemu_opt_get(opts, "monitor"))) {
+        default_monitor_opt = opt;
+    }
+
+    default_vga = qemu_opt_get_bool(opts, "vga", default_vga);
+    default_net = qemu_opt_get_bool(opts, "net", default_net);
+    default_floppy = qemu_opt_get_bool(opts, "floppy", default_floppy);
+    default_cdrom = qemu_opt_get_bool(opts, "cdrom", default_cdrom);
+    default_sdcard = qemu_opt_get_bool(opts, "sdcard", default_sdcard);
+
+    return 0;
+}
+
 int main(int argc, char **argv, char **envp)
 {
     const char *gdbstub_dev = NULL;
@@ -5409,6 +5440,13 @@ int main(int argc, char **argv, char **envp)
                 default_cdrom = 0;
                 default_sdcard = 0;
                 break;
+            case QEMU_OPTION_default:
+                opts = qemu_opts_parse(&qemu_default_opts, optarg, NULL);
+                if (!opts) {
+                    fprintf(stderr, "parse error: %s\n", optarg);
+                    exit(1);
+                }
+                break;
 #ifndef _WIN32
             case QEMU_OPTION_chroot:
                 chroot_dir = optarg;
@@ -5509,6 +5547,7 @@ int main(int argc, char **argv, char **envp)
         exit(1);
     }
 
+    qemu_opts_foreach(&qemu_default_opts, default_opt_init, NULL, 0);
     qemu_opts_foreach(&qemu_device_opts, default_driver_check, NULL, 0);
     qemu_opts_foreach(&qemu_global_opts, default_driver_check, NULL, 0);
 
@@ -5535,30 +5574,28 @@ int main(int argc, char **argv, char **envp)
     }
 
     if (display_type == DT_NOGRAPHIC) {
-        if (default_parallel)
-            add_device_config(DEV_PARALLEL, "null");
+        default_parallel_opt = "null";
+
         if (default_serial && default_monitor) {
-            add_device_config(DEV_SERIAL, "mon:stdio");
+            default_serial_opt = "mon:stdio";
         } else if (default_virtcon && default_monitor) {
-            add_device_config(DEV_VIRTCON, "mon:stdio");
+            default_virtcon_opt = "mon:stdio";
         } else {
-            if (default_serial)
-                add_device_config(DEV_SERIAL, "stdio");
-            if (default_virtcon)
-                add_device_config(DEV_VIRTCON, "stdio");
-            if (default_monitor)
-                monitor_parse("stdio", "readline");
+            default_serial_opt = "stdio";
+            default_virtcon_opt = "stdio";
+            default_monitor_opt = "stdio";
         }
-    } else {
-        if (default_serial)
-            add_device_config(DEV_SERIAL, "vc:80Cx24C");
-        if (default_parallel)
-            add_device_config(DEV_PARALLEL, "vc:80Cx24C");
-        if (default_monitor)
-            monitor_parse("vc:80Cx24C", "readline");
-        if (default_virtcon)
-            add_device_config(DEV_VIRTCON, "vc:80Cx24C");
     }
+
+    if (default_serial)
+        add_device_config(DEV_SERIAL, default_serial_opt);
+    if (default_parallel)
+        add_device_config(DEV_PARALLEL, default_parallel_opt);
+    if (default_monitor)
+        monitor_parse(default_monitor_opt, "readline");
+    if (default_virtcon)
+        add_device_config(DEV_VIRTCON, default_virtcon_opt);
+
     if (default_vga)
         vga_interface_type = VGA_CIRRUS;
 
-- 
1.6.5.2

[Qemu-devel] [PATCH 4/4] Allow default network type to be overridden

[Anthony Liguori]

Introduce a default option to the network device which specifies that this is
a default network device.  This approach should generalize to any other device.

The meaning of a default device is as follows: a default device is added to a
machine IIF defaults aren't disable (via -default or -nodefaults) and a
non-default device of this type hasn't been added.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
---
 net.c |   42 +++++++++++++++++++++++++++++++++++++-----
 1 files changed, 37 insertions(+), 5 deletions(-)

diff --git a/net.c b/net.c
index 6ef93e6..1aa3c77 100644
--- a/net.c
+++ b/net.c
@@ -833,6 +833,10 @@ static int net_init_nic(QemuOpts *opts,
         .name = "name",                            \
         .type = QEMU_OPT_STRING,                   \
         .help = "identifier for monitor commands", \
+     }, {                                          \
+        .name = "default",                         \
+        .type = QEMU_OPT_BOOL,                     \
+        .help = "act as default network device",   \
      }
 
 typedef int (*net_client_init_func)(QemuOpts *opts,
@@ -1056,6 +1060,15 @@ int net_client_init(Monitor *mon, QemuOpts *opts, int is_netdev)
     const char *type;
     int i;
 
+    /* Do not create default network devices if no defaults */
+    if (!default_net) {
+        const char *opt = qemu_opt_get(opts, "default");
+
+        if (opt && strcmp(opt, "on") == 0) {
+            return 0;
+        }
+    }
+
     type = qemu_opt_get(opts, "type");
 
     if (!is_netdev) {
@@ -1320,10 +1333,22 @@ static int net_init_netdev(QemuOpts *opts, void *dummy)
 int net_init_clients(void)
 {
     if (default_net) {
-        /* if no clients, we use a default config */
-        qemu_opts_set(&qemu_net_opts, NULL, "type", "nic");
+        QemuOpts *opts;
+
+        opts = qemu_opts_create(&qemu_net_opts, NULL, 1);
+        if (opts == NULL) {
+            return -1;
+        }
+        qemu_opt_set(opts, "type", "nic");
+        qemu_opt_set(opts, "default", "on");
+
 #ifdef CONFIG_SLIRP
-        qemu_opts_set(&qemu_net_opts, NULL, "type", "user");
+        opts = qemu_opts_create(&qemu_net_opts, NULL, 1);
+        if (opts == NULL) {
+            return -1;
+        }
+        qemu_opt_set(opts, "type", "user");
+        qemu_opt_set(opts, "default", "on");
 #endif
     }
 
@@ -1344,6 +1369,8 @@ int net_init_clients(void)
 
 int net_client_parse(QemuOptsList *opts_list, const char *optarg)
 {
+    QemuOpts *opts;
+    const char *opt;
 #if defined(CONFIG_SLIRP)
     int ret;
     if (net_slirp_parse_legacy(opts_list, optarg, &ret)) {
@@ -1351,10 +1378,15 @@ int net_client_parse(QemuOptsList *opts_list, const char *optarg)
     }
 #endif
 
-    if (!qemu_opts_parse(opts_list, optarg, "type")) {
+    opts = qemu_opts_parse(opts_list, optarg, "type");
+    if (opts == NULL) {
         return -1;
     }
 
-    default_net = 0;
+    opt = qemu_opt_get(opts, "default");
+    if (!opt || strcmp(opt, "off") == 0) {
+        default_net = 0;
+    }
+
     return 0;
 }
-- 
1.6.5.2

Re: [Qemu-devel] [PATCH 3/4] Add -defaults option to allow default devices to be overridden

[Markus Armbruster]

Anthony Liguori <aliguori@us.ibm.com> writes:

> This option can be used to toggle whether each default device is enabled or
> disabled.  For character devices, the default backend can also be overridden.
>
> For devices, we'll have to take a different approach to changing the defaults
> which will be covered in the next patch.
>
> N.B. I took special care with -nographic.  Now -nographic pretty clearly acts
> as a mechanism to override the default backend devices.
>
> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> ---
>  qemu-config.c   |   45 +++++++++++++++++++++++++++++++++
>  qemu-config.h   |    1 +
>  qemu-options.hx |    7 +++++
>  vl.c            |   75 +++++++++++++++++++++++++++++++++++++++++--------------
>  4 files changed, 109 insertions(+), 19 deletions(-)
>
> diff --git a/qemu-config.c b/qemu-config.c
> index c3203c8..82ca399 100644
> --- a/qemu-config.c
> +++ b/qemu-config.c
> @@ -242,6 +242,50 @@ QemuOptsList qemu_mon_opts = {
>      },
>  };
>  
> +QemuOptsList qemu_default_opts = {
> +    .name = "default",
> +    .head = QTAILQ_HEAD_INITIALIZER(qemu_default_opts.head),
> +    .desc = {
> +        {
> +            .name = "serial",
> +            .type = QEMU_OPT_STRING,
> +        },
[...]
> diff --git a/qemu-options.hx b/qemu-options.hx
> index 57f453d..e81ecb5 100644
> --- a/qemu-options.hx
> +++ b/qemu-options.hx
> @@ -1919,6 +1919,13 @@ STEXI
>  Don't create default devices.
>  ETEXI
>  
> +DEF("default", HAS_ARG, QEMU_OPTION_default, \
> +    "-default arg    specify default devices\n")

Isn't this too terse?

> +STEXI
> +@item -defaults
> +Override builtin default devices
> +ETEXI

This *is* too terse :)

Oh, and it's -default (sans 's').  Same typo in subject.

While we're talking about naming: isn't -default a bit too generic a
name for something that manipulates devices?  Not sure we care, as
-nodefaults is much worse, already.

[...]

[Qemu-devel] Re: [PATCH 2/4] Load global config files by default

[Gerd Hoffmann]

On 01/21/10 19:48, Anthony Liguori wrote:
> The ordering is a little troubling.  Command line options are parsed before
> loading the default configs which means that the command line configs will be
> loaded before the default config.  The effect is that the default config will
> override -readconfig directives.
 >
> It's unclear the best way to handle this.

Maybe we need two passes over the command line options?

-confdir and -nodefconfig switches would be handled in the first pass, 
then read the global config files (unless disabled), then the second pass.

I think we could also make the default device handling less messy that 
way by checking -nographic and -nodefaults in the first pass too. 
Having the machine type (-M) read in the first pass could allow some 
cleanups too.

cheers,
   Gerd

[Qemu-devel] Re: [PATCH 4/4] Allow default network type to be overridden

[Gerd Hoffmann]

On 01/21/10 19:48, Anthony Liguori wrote:
> Introduce a default option to the network device which specifies that this is
> a default network device.  This approach should generalize to any other device.

> The meaning of a default device is as follows: a default device is added to a
> machine IIF defaults aren't disable (via -default or -nodefaults) and a
> non-default device of this type hasn't been added.

I don't like the idea to have two different mechanisms for configuring 
defaults ([default] section + default = "on").  I'd suggest to pick one 
and apply it everythere.

cheers,
   Gerd

[Qemu-devel] Re: [PATCH 1/4] Support --confdir in configure to specify path to configuration files

[Paolo Bonzini]

On 01/21/2010 07:48 PM, Anthony Liguori wrote:
> --confdir can be used to override the
> default to an absolute path.  The expectation is that when installed to /usr,
>   --confdir=/etc/qemu will be used.

Why not --sysconfdir?  There are already several differences between 
Autoconf options and QEMU configure options, however unlike the others 
this one would be used by basically all distros.

Paolo

[Qemu-devel] Re: [PATCH 1/4] Support --confdir in configure to specify path to configuration files

[Anthony Liguori]

On 01/22/2010 06:43 AM, Paolo Bonzini wrote:
> On 01/21/2010 07:48 PM, Anthony Liguori wrote:
>> --confdir can be used to override the
>> default to an absolute path.  The expectation is that when installed 
>> to /usr,
>>   --confdir=/etc/qemu will be used.
>
> Why not --sysconfdir?  There are already several differences between 
> Autoconf options and QEMU configure options, however unlike the others 
> this one would be used by basically all distros.

Wasn't 100% sure on the proper naming.  sysconfdir works for me.

Regards,

Anthony Liguori

> Paolo

[Qemu-devel] Re: [PATCH 4/4] Allow default network type to be overridden

[Anthony Liguori]

On 01/22/2010 05:00 AM, Gerd Hoffmann wrote:
> On 01/21/10 19:48, Anthony Liguori wrote:
>> Introduce a default option to the network device which specifies that 
>> this is
>> a default network device.  This approach should generalize to any 
>> other device.
>
>> The meaning of a default device is as follows: a default device is 
>> added to a
>> machine IIF defaults aren't disable (via -default or -nodefaults) and a
>> non-default device of this type hasn't been added.
>
> I don't like the idea to have two different mechanisms for configuring 
> defaults ([default] section + default = "on").  I'd suggest to pick 
> one and apply it everythere.

I agree that it's awkward.  Need to think a bit about it.

Regards,

Anthony Liguori

> cheers,
>   Gerd
>

Re: [Qemu-devel] [PATCH 3/4] Add -defaults option to allow default devices to be overridden

[Anthony Liguori]

On 01/22/2010 04:15 AM, Markus Armbruster wrote:
> Anthony Liguori<aliguori@us.ibm.com>  writes:
>
>    
>> This option can be used to toggle whether each default device is enabled or
>> disabled.  For character devices, the default backend can also be overridden.
>>
>> For devices, we'll have to take a different approach to changing the defaults
>> which will be covered in the next patch.
>>
>> N.B. I took special care with -nographic.  Now -nographic pretty clearly acts
>> as a mechanism to override the default backend devices.
>>
>> Signed-off-by: Anthony Liguori<aliguori@us.ibm.com>
>> ---
>>   qemu-config.c   |   45 +++++++++++++++++++++++++++++++++
>>   qemu-config.h   |    1 +
>>   qemu-options.hx |    7 +++++
>>   vl.c            |   75 +++++++++++++++++++++++++++++++++++++++++--------------
>>   4 files changed, 109 insertions(+), 19 deletions(-)
>>
>> diff --git a/qemu-config.c b/qemu-config.c
>> index c3203c8..82ca399 100644
>> --- a/qemu-config.c
>> +++ b/qemu-config.c
>> @@ -242,6 +242,50 @@ QemuOptsList qemu_mon_opts = {
>>       },
>>   };
>>
>> +QemuOptsList qemu_default_opts = {
>> +    .name = "default",
>> +    .head = QTAILQ_HEAD_INITIALIZER(qemu_default_opts.head),
>> +    .desc = {
>> +        {
>> +            .name = "serial",
>> +            .type = QEMU_OPT_STRING,
>> +        },
>>      
> [...]
>    
>> diff --git a/qemu-options.hx b/qemu-options.hx
>> index 57f453d..e81ecb5 100644
>> --- a/qemu-options.hx
>> +++ b/qemu-options.hx
>> @@ -1919,6 +1919,13 @@ STEXI
>>   Don't create default devices.
>>   ETEXI
>>
>> +DEF("default", HAS_ARG, QEMU_OPTION_default, \
>> +    "-default arg    specify default devices\n")
>>      
> Isn't this too terse?
>    

Yes.  Thanks for pointing that out.

>> +STEXI
>> +@item -defaults
>> +Override builtin default devices
>> +ETEXI
>>      
> This *is* too terse :)
>
> Oh, and it's -default (sans 's').  Same typo in subject.
>
> While we're talking about naming: isn't -default a bit too generic a
> name for something that manipulates devices?  Not sure we care, as
> -nodefaults is much worse, already.
>    

Absolutely.  I'm going to split out the config loading bits because they 
should be easy to commit.  How to best handle defaults could use some 
more thought.  I think this is a key bit of functionality to get right 
though because it solves a whole class of problems relating to upstream 
policy vs. downstream policy.

I very much like the idea of having a qdev property 'default' to signify 
that this is a default device.  It means we could easily allow a user to 
universally enable usb devices, etc without baking a default_usb concept 
into qemu.  Ideally, we could eliminate the whole default mess we have 
today by doing this, provided that we can implement the right semantics.

Today, those semantics are, if we specify any device of the same 
"class", do not load default devices of that class.  It's unclear how to 
specify the concept of a "class" though.  I know Gerd brought this up 
ages ago and both Paul and I were very opposed to it but I certainly 
appreciate the fact that it would simplify default device handling.

It's definitely clear that each device needs to be able to be part of 
multiple default-class's.  We could potentially introduce a qdev 
property and label every device with a class array but that's a lot of 
ugliness just to support defaults.  I still would prefer that there was 
a more discoverable way to determine the class of a device as opposed to 
explicitly labelling it.

If we need to do class tagging, I guess it's not the end of the world...

Regards,

Anthony Liguori

> [...]
>