6:04 a.m.
qemu now supports gic-version option for the virt machine. This patchset
allows to use it in libvirt.
v1 => v2:
- Added capability flag
Pavel Fedin (3):
qemu: Introduce QEMU_CAPS_MACH_VIRT_GIC_VERSION capability
qemu: Add support for gic-version machine option
qemu: Add test cases for gic-version option
src/qemu/qemu_capabilities.c | 5 +++
src/qemu/qemu_capabilities.h | 1 +
src/qemu/qemu_command.c | 43 +++++++++++++++-------
.../qemuxml2argv-aarch64-gicv3.args | 6 +++
.../qemuxml2argv-aarch64-gicv3.xml | 26 +++++++++++++
tests/qemuxml2argvtest.c | 5 +++
6 files changed, 73 insertions(+), 13 deletions(-)
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.args
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.xml
--
2.1.4
6:04 a.m.
New subject: [libvirt] [PATCH v2 1/3] qemu: Introduce QEMU_CAPS_MACH_VIRT_GIC_VERSION capability
Unfortunately qemu currently doesn't offer introspection for machine types,
so we have to rely on version number, similar to QEMU_CAPS_MACHINE_USB_OPT.
Signed-off-by: Pavel Fedin <p.fedin(a)samsung.com>
---
src/qemu/qemu_capabilities.c | 5 +++++
src/qemu/qemu_capabilities.h | 1 +
2 files changed, 6 insertions(+)
diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c
index eb2edf5..98d306c 100644
--- a/src/qemu/qemu_capabilities.c
+++ b/src/qemu/qemu_capabilities.c
@@ -298,6 +298,7 @@ VIR_ENUM_IMPL(virQEMUCaps, QEMU_CAPS_LAST,
"rtl8139",
"e1000",
"virtio-net",
+ "gic-version",
);
@@ -3404,6 +3405,10 @@ virQEMUCapsInitQMPMonitor(virQEMUCapsPtr qemuCaps,
if (qemuCaps->version >= 2004000)
virQEMUCapsSet(qemuCaps, QEMU_CAPS_VHOSTUSER_MULTIQUEUE);
+ /* Since 2.4.50 ARM virt machine supports gic-version option */
+ if (qemuCaps->version >= 2004050)
+ virQEMUCapsSet(qemuCaps, QEMU_CAPS_MACH_VIRT_GIC_VERSION);
+
if (virQEMUCapsProbeQMPCommands(qemuCaps, mon) < 0)
goto cleanup;
if (virQEMUCapsProbeQMPEvents(qemuCaps, mon) < 0)
diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h
index 2623574..a1d98d0 100644
--- a/src/qemu/qemu_capabilities.h
+++ b/src/qemu/qemu_capabilities.h
@@ -238,6 +238,7 @@ typedef enum {
QEMU_CAPS_DEVICE_RTL8139 = 196, /* -device rtl8139 */
QEMU_CAPS_DEVICE_E1000 = 197, /* -device e1000 */
QEMU_CAPS_DEVICE_VIRTIO_NET = 198, /* -device virtio-net-* */
+ QEMU_CAPS_MACH_VIRT_GIC_VERSION = 199, /* -machine virt,gic-version */
QEMU_CAPS_LAST, /* this must always be the last item */
} virQEMUCapsFlags;
--
2.1.4
6:04 a.m.
New subject: [libvirt] [PATCH v2 2/3] qemu: Add support for gic-version machine option
Support for GICv3 has been recently introduced in qemu using gic-version
option for the 'virt' machine. The option can actually take values of
'2', '3' and 'host', however, since in libvirt this is a numeric
parameter, we limit it only to 2 and 3. Value of 2 is not added to the
command line in order to keep backward compatibility with older qemu
versions.
Signed-off-by: Pavel Fedin <p.fedin(a)samsung.com>
---
src/qemu/qemu_command.c | 43 ++++++++++++++++++++++++++++++-------------
1 file changed, 30 insertions(+), 13 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index bb1835c..a47e188 100644
--- a/src/qemu/qemu_command.c
+++ b/src/qemu/qemu_command.c
@@ -7702,19 +7702,6 @@ qemuBuildCpuArgStr(virQEMUDriverPtr driver,
have_cpu = true;
}
- if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON) {
- if (def->gic_version && def->gic_version != 2) {
- virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
- _("gic version '%u' is not supported"),
- def->gic_version);
- goto cleanup;
- }
-
- /* There's no command line argument currently to turn on/off GIC. It's
- * done automatically by qemu-system-aarch64. But if this changes, lets
- * put the code here. */
- }
-
if (virBufferCheckError(&buf) < 0)
goto cleanup;
@@ -7931,6 +7918,36 @@ qemuBuildMachineArgStr(virCommandPtr cmd,
return -1;
}
+ if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON) {
+ if (def->gic_version) {
+ if (!STREQ(def->os.machine, "virt") &&
+ !STRPREFIX(def->os.machine, "virt-")) {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+ _("gic-version option is available "
+ "only for virt machine"));
+ virBufferFreeAndReset(&buf);
+ return -1;
+ }
+
+ /* 2 is the default, so we don't put it as option for
+ * backwards compatibility
+ */
+ if (def->gic_version != 2) {
+ if (virQEMUCapsGet(qemuCaps,
+ QEMU_CAPS_MACH_VIRT_GIC_VERSION)) {
+ virBufferAsprintf(&buf, ",gic-version=%d",
+ def->gic_version);
+ } else {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+ _("gic-version option is not available
"
+ "with this QEMU binary"));
+ virBufferFreeAndReset(&buf);
+ return -1;
+ }
+ }
+ }
+ }
+
virCommandAddArgBuffer(cmd, &buf);
}
--
2.1.4
9:34 a.m.
New subject: [libvirt] [PATCH v2 2/3] qemu: Add support for gic-version machine option
On Wed, Sep 30, 2015 at 02:04:10PM +0300, Pavel Fedin wrote:
Support for GICv3 has been recently introduced in qemu using
gic-version
option for the 'virt' machine. The option can actually take values of
'2', '3' and 'host', however, since in libvirt this is a numeric
parameter, we limit it only to 2 and 3. Value of 2 is not added to the
command line in order to keep backward compatibility with older qemu
versions.
Signed-off-by: Pavel Fedin <p.fedin(a)samsung.com>
---
src/qemu/qemu_command.c | 43 ++++++++++++++++++++++++++++++-------------
1 file changed, 30 insertions(+), 13 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index bb1835c..a47e188 100644
--- a/src/qemu/qemu_command.c
+++ b/src/qemu/qemu_command.c
@@ -7702,19 +7702,6 @@ qemuBuildCpuArgStr(virQEMUDriverPtr driver,
have_cpu = true;
}
- if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON) {
- if (def->gic_version && def->gic_version != 2) {
- virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
- _("gic version '%u' is not supported"),
- def->gic_version);
- goto cleanup;
- }
-
- /* There's no command line argument currently to turn on/off GIC. It's
- * done automatically by qemu-system-aarch64. But if this changes, lets
- * put the code here. */
- }
-
if (virBufferCheckError(&buf) < 0)
goto cleanup;
@@ -7931,6 +7918,36 @@ qemuBuildMachineArgStr(virCommandPtr cmd,
return -1;
}
+ if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON) {
+ if (def->gic_version) {
+ if (!STREQ(def->os.machine, "virt") &&
+ !STRPREFIX(def->os.machine, "virt-")) {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+ _("gic-version option is available "
+ "only for virt machine"));
+ virBufferFreeAndReset(&buf);
+ return -1;
+ }
Indentation's off here.
Also before this patch we would allow def->gic_version == 2 for any
machine type. I don't have a problem with this since GIC doesn't make
sense anywhere else then on ARM machines, but shouldn't we check for
the fact that the request is for ARM (in case, for example, if ppc64
gains some 'virt' machine type)? Because we have no guarantee that
it's ARM just based on the machine type.
+
+ /* 2 is the default, so we don't put it as option for
+ * backwards compatibility
+ */
+ if (def->gic_version != 2) {
+ if (virQEMUCapsGet(qemuCaps,
+ QEMU_CAPS_MACH_VIRT_GIC_VERSION)) {
+ virBufferAsprintf(&buf, ",gic-version=%d",
+ def->gic_version);
+ } else {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+ _("gic-version option is not available
"
+ "with this QEMU binary"));
+ virBufferFreeAndReset(&buf);
+ return -1;
I'd change this to:
if (gic != 2) {
if (!caps)
error;
append_cmd();
}
If you're ok with that, just let me know and I'll push it with the
following diff squashed in, right after the release:
diff --git i/src/qemu/qemu_command.c w/src/qemu/qemu_command.c
index a47e1883d976..72149bdd1eef 100644
--- i/src/qemu/qemu_command.c
+++ w/src/qemu/qemu_command.c
@@ -7918,35 +7918,36 @@ qemuBuildMachineArgStr(virCommandPtr cmd,
return -1;
}
- if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON) {
- if (def->gic_version) {
- if (!STREQ(def->os.machine, "virt") &&
- !STRPREFIX(def->os.machine, "virt-")) {
- virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+ if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON) {
+ if (def->gic_version) {
+ if ((def->os.arch != VIR_ARCH_ARMV7L &&
+ def->os.arch != VIR_ARCH_AARCH64) ||
+ (STRNEQ(def->os.machine, "virt") &&
+ !STRPREFIX(def->os.machine, "virt-"))) {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
_("gic-version option is available "
"only for virt machine"));
virBufferFreeAndReset(&buf);
return -1;
- }
+ }
- /* 2 is the default, so we don't put it as option for
- * backwards compatibility
- */
+ /* 2 is the default, so we don't put it as option for
+ * backwards compatibility
+ */
if (def->gic_version != 2) {
- if (virQEMUCapsGet(qemuCaps,
- QEMU_CAPS_MACH_VIRT_GIC_VERSION)) {
- virBufferAsprintf(&buf, ",gic-version=%d",
- def->gic_version);
- } else {
+ if (!virQEMUCapsGet(qemuCaps,
+ QEMU_CAPS_MACH_VIRT_GIC_VERSION)) {
virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
_("gic-version option is not available
"
"with this QEMU binary"));
virBufferFreeAndReset(&buf);
return -1;
}
+
+ virBufferAsprintf(&buf, ",gic-version=%d",
def->gic_version);
}
- }
- }
+ }
+ }
virCommandAddArgBuffer(cmd, &buf);
}
--
Martin
1:53 a.m.
New subject: [libvirt] [PATCH v2 2/3] qemu: Add support for gic-version machine option
Hello!
Indentation's off here.
Damn, sorry, overlooked...
Also before this patch we would allow def->gic_version == 2 for
any
machine type. I don't have a problem with this since GIC doesn't make
sense anywhere else then on ARM machines,
I'm OK with this. I used 0 for 'no version supplied' just because libvirt
originally does this.
but shouldn't we check for
the fact that the request is for ARM (in case, for example, if ppc64
gains some 'virt' machine type)? Because we have no guarantee that
it's ARM just based on the machine type.
Yes, i guess we should.
I'd change this to:
if (gic != 2) {
if (!caps)
error;
append_cmd();
}
You know, if we are talking about making changes in parser code, we could do more.
Actually, as i
said in my cover letter, qemu supports more than just 2 or 3. We can also specify
'host' for 'best
possible'. Could we accommodate this somehow too? I believe in order to do this, we
should change
parameter type from numeric to string.
And also we could add some another boolean, which would allow to disable in-kernel GIC
emulation
(kernel_irqchip=off). This works with any machine type, BTW, not only with ARM. Something
like <gic
kvm='off'/>.
I believe these changes could go as a separate patch, after we discuss details.
If you're ok with that, just let me know and I'll push it
with the
following diff squashed in, right after the release:
Yes, ACK.
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
"%s",
_("gic-version option is available "
"only for virt machine"));
Then "...only for ARM virt machine".
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
2:17 a.m.
New subject: [libvirt] [PATCH v2 2/3] qemu: Add support for gic-version machine option
On Thu, Oct 01, 2015 at 09:53:00AM +0300, Pavel Fedin wrote:
Hello!
> Indentation's off here.
Damn, sorry, overlooked...
> Also before this patch we would allow def->gic_version == 2 for any
> machine type. I don't have a problem with this since GIC doesn't make
> sense anywhere else then on ARM machines,
I'm OK with this. I used 0 for 'no version supplied' just because libvirt
originally does this.
> but shouldn't we check for
> the fact that the request is for ARM (in case, for example, if ppc64
> gains some 'virt' machine type)? Because we have no guarantee that
> it's ARM just based on the machine type.
Yes, i guess we should.
> I'd change this to:
>
> if (gic != 2) {
> if (!caps)
> error;
> append_cmd();
> }
You know, if we are talking about making changes in parser code, we could do more.
Actually, as i
said in my cover letter, qemu supports more than just 2 or 3. We can also specify
'host' for 'best
possible'. Could we accommodate this somehow too? I believe in order to do this, we
should change
parameter type from numeric to string.
Yes, we can certainly do that.
And also we could add some another boolean, which would allow to
disable in-kernel GIC emulation
(kernel_irqchip=off). This works with any machine type, BTW, not only with ARM. Something
like <gic
kvm='off'/>.
I don't know what that is. Is that only GIC related? Where could I
find more details? If that makes sense for anything, we can sure do
that, the only reason why I would be against this, which I can come up
now, is if there is nobody using it.
I believe these changes could go as a separate patch, after we
discuss details.
Yeah, sure.
> If you're ok with that, just let me know and I'll push it
with the
> following diff squashed in, right after the release:
Yes, ACK.
> + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
> _("gic-version option is available "
> "only for virt machine"));
Then "...only for ARM virt machine".
Oh, good catch, I'll fix that, thanks.
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
3:01 a.m.
New subject: [libvirt] [PATCH v2 2/3] qemu: Add support for gic-version machine option
Hello!
> And also we could add some another boolean, which would allow to
disable in-kernel GIC
emulation
>(kernel_irqchip=off). This works with any machine type, BTW, not only with ARM.
Something
like <gic
>kvm='off'/>.
>
I don't know what that is. Is that only GIC related?
No, not only GIC. It applies to any possible irqchip. When set to off, this option
disables using
KVM acceleration for the irqchip, and qemu's software emulation is used instead.
Where could I find more details?
Something like "aarch64-system-qemu -machine virt,?" i guess. Or similar
command line for different
arch and different machine. It is implemented as machine option, but applies to any
machine, not
only to virt, and not only for ARM.
The option is also mentioned on http://wiki.qemu.org/KVM, but information is outdated (it
defaults
to on since long ago).
If that makes sense for anything, we can sure do
that, the only reason why I would be against this, which I can come up
now, is if there is nobody using it.
Yes, this option was bitrot in the ARM kernel, but i have recently fixed it
(https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/?...
945f1765cd48eb0da). I understand that the primary use of this on e.g. x86 would be for
testing
purposes only, but for ARM it can be more useful, because you cannot run everything on
everything.
For example, if you want GICv2 system, but your host is pure GICv3 without backwards
compatibility,
you could want to use this option, so that you use software-emulated GICv2. Of course this
would not
give you the top performance, but this is still usable, i tested it. Perhaps other
systems, like
PowerPC, can also take advantage of it (from KVM code i see there are several different
irqchips for
PPC, and i'm in doubt that any of these irqchips can emulate everything else).
There is one more complication with this option - CP15 timer currently will not work,
because
virtual timer accesses cannot be intercepted by the hypervisor. I am considering fixing
this in the
future, there are at least two possible solutions. But, still, you can run, for example,
vexpress
guest with a little bit patched device tree or reconfigured kernel (arch-timer disabled),
because
this hardware has another, memory-mapped timer, which can be used with software emulation.
And it
works.
Well, if you think that this is not ready for the wide public yet, we can leave out this
part.
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
6:04 a.m.
New subject: [libvirt] [PATCH v2 3/3] qemu: Add test cases for gic-version option
These tests make sure that we can use this option only when the capability is
set.
Signed-off-by: Pavel Fedin <p.fedin(a)samsung.com>
---
.../qemuxml2argv-aarch64-gicv3.args | 6 +++++
.../qemuxml2argv-aarch64-gicv3.xml | 26 ++++++++++++++++++++++
tests/qemuxml2argvtest.c | 5 +++++
3 files changed, 37 insertions(+)
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.args
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.xml
diff --git a/tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.args
b/tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.args
new file mode 100644
index 0000000..d697669
--- /dev/null
+++ b/tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.args
@@ -0,0 +1,6 @@
+LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test QEMU_AUDIO_DRV=none \
+/usr/bin/qemu-system-aarch64 -S -machine virt,accel=tcg,gic-version=3 -cpu cortex-a53 -m
1024 -smp 1 \
+-nographic -nodefaults -monitor unix:/tmp/test-monitor,server,nowait -boot c \
+-kernel /aarch64.kernel -initrd /aarch64.initrd -append console=ttyAMA0 -usb \
+-net nic,macaddr=52:54:00:09:a4:37,vlan=0,model=virtio,name=net0 \
+-net user,vlan=0,name=hostnet0
diff --git a/tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.xml
b/tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.xml
new file mode 100644
index 0000000..f8d63c3
--- /dev/null
+++ b/tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.xml
@@ -0,0 +1,26 @@
+<domain type="qemu">
+ <name>aarch64test</name>
+ <uuid>6ba410c5-1e5c-4d57-bee7-2228e7ffa32f</uuid>
+ <memory>1048576</memory>
+ <currentMemory>1048576</currentMemory>
+ <vcpu>1</vcpu>
+ <features>
+ <acpi/>
+ <gic version='3'/>
+ </features>
+ <cpu match='exact'>
+ <model>cortex-a53</model>
+ </cpu>
+ <os>
+ <type arch="aarch64" machine="virt">hvm</type>
+ <kernel>/aarch64.kernel</kernel>
+ <initrd>/aarch64.initrd</initrd>
+ <cmdline>console=ttyAMA0</cmdline>
+ </os>
+ <devices>
+ <emulator>/usr/bin/qemu-system-aarch64</emulator>
+ <interface type='user'>
+ <mac address='52:54:00:09:a4:37'/>
+ </interface>
+ </devices>
+</domain>
diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c
index ae67779..e6bf2f1 100644
--- a/tests/qemuxml2argvtest.c
+++ b/tests/qemuxml2argvtest.c
@@ -1644,6 +1644,11 @@ mymain(void)
QEMU_CAPS_CPU_HOST, QEMU_CAPS_KVM);
DO_TEST("aarch64-gic", QEMU_CAPS_DEVICE, QEMU_CAPS_DRIVE,
QEMU_CAPS_KVM);
+ DO_TEST("aarch64-gicv3", QEMU_CAPS_DEVICE, QEMU_CAPS_DRIVE,
+ QEMU_CAPS_KVM, QEMU_CAPS_MACHINE_OPT,
+ QEMU_CAPS_MACH_VIRT_GIC_VERSION);
+ DO_TEST_FAILURE("aarch64-gicv3", QEMU_CAPS_DEVICE, QEMU_CAPS_DRIVE,
+ QEMU_CAPS_KVM, QEMU_CAPS_MACHINE_OPT);
driver.caps->host.cpu->arch = VIR_ARCH_AARCH64;
DO_TEST("aarch64-kvm-32-on-64", QEMU_CAPS_DEVICE, QEMU_CAPS_DRIVE,
--
2.1.4
9:35 a.m.
On Wed, Sep 30, 2015 at 02:04:08PM +0300, Pavel Fedin wrote:
qemu now supports gic-version option for the virt machine. This
patchset
allows to use it in libvirt.
v1 => v2:
- Added capability flag
ACK series after release with my diff squashed in in case the original
author agrees.
Pavel Fedin (3):
qemu: Introduce QEMU_CAPS_MACH_VIRT_GIC_VERSION capability
qemu: Add support for gic-version machine option
qemu: Add test cases for gic-version option
src/qemu/qemu_capabilities.c | 5 +++
src/qemu/qemu_capabilities.h | 1 +
src/qemu/qemu_command.c | 43 +++++++++++++++-------
.../qemuxml2argv-aarch64-gicv3.args | 6 +++
.../qemuxml2argv-aarch64-gicv3.xml | 26 +++++++++++++
tests/qemuxml2argvtest.c | 5 +++
6 files changed, 73 insertions(+), 13 deletions(-)
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.args
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-gicv3.xml
--
2.1.4
12:48 p.m.
On 09/30/2015 07:04 AM, Pavel Fedin wrote:
qemu now supports gic-version option for the virt machine. This
patchset
allows to use it in libvirt.
This isn't specifically related to your patch, but I'm curious: Can you
explain what this setting actually does? When would apps like virt-manager
want to enable it?
- Cole
2 a.m.
Hello!
Can you explain what this setting actually does?
Of course. It allows to choose GIC version used by ARM 'virt' machine. By default
it's 2, but you
can now set it to 3.
Actually, initially i proposed another machine type, like 'virt-v3', IMHO this
would be more
logical because this is kind of different architecture. But Peter Maydell disagreed and
insisted on
option. So we have an option.
When would apps like virt-manager want to enable it?
When choosing 'virt' machine type for ARM architecture.
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
4:26 a.m.
On Thu, Oct 01, 2015 at 10:00:13AM +0300, Pavel Fedin wrote:
Hello!
> Can you explain what this setting actually does?
Of course. It allows to choose GIC version used by ARM 'virt' machine. By default
it's 2, but you
can now set it to 3.
Actually, initially i proposed another machine type, like 'virt-v3', IMHO this
would be more
logical because this is kind of different architecture. But Peter Maydell disagreed and
insisted on
option. So we have an option.
I think Cole wanted to know more like how it can be useful for end
users. Since I see GIC as something similar to APIC, I'm guessing the
usefulness won't be directly visible, but should help. About what
exactly it does or what are the differences between, that I cannot
help with, unfortunately.
> When would apps like virt-manager want to enable it?
When choosing 'virt' machine type for ARM architecture.
Unconditionally? Doesn't that pose a problem, e.g. for requesting
version that is not supported or breaking something that expects lower
version?
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
6:31 a.m.
Hello!
I think Cole wanted to know more like how it can be useful for end
users. Since I see GIC as something similar to APIC, I'm guessing the
usefulness won't be directly visible, but should help. About what
exactly it does or what are the differences between, that I cannot
help with, unfortunately.
Ah, i assumed you know ARM architecture, sorry.
So, GIC for ARM is really what is APIC for x86. It's an interrupt controller. There
are several
generations of this controller, up to 3 as of now (there is spec for v4, but i don't
know any
real-life implementation, additionally it's fully compatible with v3). v1 is really
obsolete, and
AFAIK v2 is fully compatible with it, so for us the only important ones are v2 and v3.
v2 is what was included into the majority of ARM32 systems. Some ARM64 systems use it
too, but it
has a number of limitations. The largest one is that it cannot support more than 16 CPUs.
In order
to overcome this, GICv3 was developed. GICv3 is significantly different from GICv2, so it
requires
own implementation, and own kernel driver.
The next thing concerns virtualization, and here comes our qemu support. GICv3 can
include
backwards v2 compatibility, but only as an option. And, for example, on my board
there's no such
option. Therefore, in order to be able to use KVM GIC acceleration, you have to emulate
GICv3 guest
on such a host. You simply will be unable to run GICv2 guest under KVM in this case.
So, "GIC type" setting should be thought of like a "board subtype".
This is different hardware,
with different software support.
> When choosing 'virt' machine type for ARM architecture.
>
Unconditionally? Doesn't that pose a problem, e.g. for requesting
version that is not supported
It shouldn't. If you specify version unsupported by qemu, qemu will complain. I
guess, the same
with any other option.
or breaking something that expects lower version?
qemu's policy is to default to 2. libvirt's policy is NOT to specify gic-version
property unless
value different from 2 is explicitly specified in the XML. I guess, virt-manager could
also default
to '2' (or 'Hypervisor's default', which would be the same).
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
3221
days inactive
3222
days old
12 comments
3 participants
participants (3)
-
Cole Robinson -
Martin Kletzander -
Pavel Fedin