fic when enabling vector
lengths.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
Reviewed-by: Richard Henderson
Tested-by: Masayoshi Mizuma
---
docs/arm-cpu-features.rst | 19 ---
target/arm/cpu.c | 3 +++
target/arm/cpu.h
with Richard Henderson, who reworked
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
Reviewed-by: Richard Henderson
Tested-by: Masayoshi M
doesn't want to select a
target they can pass -1 for the target which indicates they want
to use the preferred target and have it passed back like before.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
Tested-by: Masayoshi Mizuma
Reviewed-by: Beata Mich
e other SVE CPU properties in later patches.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
Tested-by: Masayoshi Mizuma
Reviewed-by: Beata Michalska
---
target/arm/cpu64.c | 17 ++---
target/arm/kvm.c | 5 +
target/arm/kvm
Now that Arm CPUs have advertised features lets add tests to ensure
we maintain their expected availability with and without KVM.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
---
tests/Makefile.include | 5 +-
tests/arm-cpu-features.c | 240 +++
2
d
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
Tested-by: Masayoshi Mizuma
Reviewed-by: Beata Mich
me
cpu_max_get/set_sve_vq to cpu_max_get/set_sve_max_vq in order for them
to follow the typical *_get/set_ pattern.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
Tested-by: Masayoshi Mizuma
Reviewed-by: Beata Michalska
---
target/arm/cpu.c
e KVM SVE qtest
- Renamed sve to sve everywhere
- Renamed power-of-2 to power-of-two everywhere
- Picked up some more tags from Richard
Thanks!
drew
Andrew Jones (9):
target/arm/monitor: Introduce qmp_query_cpu_model_expansion
tests: arm: Introduce cpu feature tests
target/a
These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
swabbing is different than it is for fpsmid because the vector format
is a little-endian stream of words.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
Tested-by: Masayoshi Mizuma
e list of advertised
features, e.g. 'vfp' and 'neon'. The only requirement is that we can
detect invalid configurations and emit failures at QMP query time.
For 'vfp' and 'neon' this will require some refactoring to share a
validation function between the QM
On Tue, Oct 22, 2019 at 03:49:51PM +0200, Andrew Jones wrote:
> On Tue, Oct 22, 2019 at 11:29:05AM +0100, Peter Maydell wrote:
> > On Mon, 21 Oct 2019 at 17:18, Peter Maydell
> > wrote:
> > >
> > > On Mon, 21 Oct 2019 at 17:12, Andrew Jones wrote:
> > &g
On Mon, Oct 21, 2019 at 04:07:14PM +0100, Beata Michalska wrote:
> Indeed, the patch got bit messed-up. Apologies for that as well.
> I have been testing manually but I did try the test you have provided
> and yes it fails - there is a slight problem with the case when qdict_in
> is empty,but this
On Tue, Oct 22, 2019 at 11:29:05AM +0100, Peter Maydell wrote:
> On Mon, 21 Oct 2019 at 17:18, Peter Maydell wrote:
> >
> > On Mon, 21 Oct 2019 at 17:12, Andrew Jones wrote:
> > >
> > > On Mon, Oct 21, 2019 at 04:43:22PM +0100, Peter Maydell wrote:
> > &
On Mon, Oct 21, 2019 at 04:43:22PM +0100, Peter Maydell wrote:
> On Mon, 21 Oct 2019 at 15:23, Andrew Jones wrote:
> > Peter, would you mind running your test on the kvm32 machine with this
> > change before I send a v7?
>
> Still fails:
>
> pm215@pm-ct:~/qemu/build/a
On Mon, Oct 21, 2019 at 02:54:30PM +0200, Philippe Mathieu-Daudé wrote:
> On 10/21/19 2:36 PM, Peter Maydell wrote:
> > On Wed, 16 Oct 2019 at 09:54, Andrew Jones wrote:
> > >
> > > Since Linux kernel v5.2-rc1 KVM has support for enabling SVE in guests.
> > >
On Thu, Oct 17, 2019 at 05:17:59PM -0400, Masayoshi Mizuma wrote:
> Hi Drew,
>
> Thank you for posting the patches, they seems to work well
> because the softlockup is gone and the timestamp jump of
> dmesg and ftrace record also disappeared after the guest
> is virsh resume'ed.
>
> Let me add co
When dumping a guest with dump-guest-memory also dump the SVE
registers if they are in use.
Signed-off-by: Andrew Jones
---
v3:
- Pulled sve_bswap64 out of kvm64.c and reused it here
- Changed fpsr_offset and sve_size to only align to a
16 byte boundary from the note payload offset, not
On Wed, Oct 16, 2019 at 04:16:57PM +0100, Beata Michalska wrote:
> On Wed, 16 Oct 2019 at 14:50, Andrew Jones wrote:
> >
> > On Wed, Oct 16, 2019 at 02:24:50PM +0100, Beata Michalska wrote:
> > > On Tue, 15 Oct 2019 at 12:56, Beata Michalska
> > > wrote:
> >
Beata Michalska noticed this missing visit_free() while reviewing
arm's implementation of qmp_query_cpu_model_expansion(), which is
modeled off this s390x implementation.
Signed-off-by: Andrew Jones
---
target/s390x/cpu_models.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/
kvm-adjvtime is a KVM specific CPU property and a first of its kind.
To accommodate it we also add kvm_arm_add_vcpu_properties() and a
KVM specific CPU properties description to the CPU features document.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
docs/arm-cpu-features.rst
If we know what the default value should be then we can test for
that as well as the feature existence.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
tests/arm-cpu-features.c | 44
1 file changed, 35 insertions(+), 9 deletions(-)
diff
When acceleration like KVM is in use it's necessary to use the host's
counter frequency when converting ticks to or from time units.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
include/qemu/timer.h | 16
1 file changed, 16 insertions(+)
diff --git
even though we don't use those registers either.
This change was originally posted as a hunk of a different, never
merged patch from Bijan Mottahedeh.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
target/arm/kvm64.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/
patch only provides the implementation. A subsequent patch
will provide the CPU property allowing the feature to be enabled.
Reported-by: Bijan Mottahedeh
Suggested-by: Bijan Mottahedeh
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
target/arm/cpu.h | 3 +++
target/arm/kvm.
same as kvm64.c. So what could go wrong? Test results would be
appreciated.
[1] https://lists.gnu.org/archive/html/qemu-devel/2018-11/msg05713.html
[2] https://patchew.org/QEMU/20191001125845.8793-1-drjo...@redhat.com/
Thanks,
drew
Andrew Jones (5):
target/arm/kvm64: kvm64 cpus have timer
On Wed, Oct 16, 2019 at 02:24:50PM +0100, Beata Michalska wrote:
> On Tue, 15 Oct 2019 at 12:56, Beata Michalska
> wrote:
> >
> > On Tue, 15 Oct 2019 at 11:56, Andrew Jones wrote:
> > >
> > > On Tue, Oct 15, 2019 at 10:59:16AM +0100, Beata Michalska wrote:
On Wed, Oct 16, 2019 at 02:26:05PM +0200, Philippe Mathieu-Daudé wrote:
> On 10/16/19 2:21 PM, Andrew Jones wrote:
> > On Wed, Oct 16, 2019 at 02:05:24PM +0200, Philippe Mathieu-Daudé wrote:
> > > > +static void test_query_cpu_model_expansion
On Wed, Oct 16, 2019 at 02:05:24PM +0200, Philippe Mathieu-Daudé wrote:
> > +static void test_query_cpu_model_expansion_kvm(const void *data)
> > +{
> > +QTestState *qts;
> > +
> > +qts = qtest_init(MACHINE "-accel kvm -cpu host");
> > +
> > +assert_has_feature(qts, "host", "pmu");
>
>
On Thu, Oct 10, 2019 at 01:33:02PM -0400, Richard Henderson wrote:
> On 10/10/19 2:16 AM, Andrew Jones wrote:
> >> It might be best to avoid the ifdef altogether:
> >>
> >> for (i = 0; i < 32; ++i) {
> >> uint64_t *d = (uint64_t *)&buf[sv
fic when enabling vector
lengths.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
Reviewed-by: Richard Henderson
---
docs/arm-cpu-features.rst | 19 ---
target/arm/cpu.c | 3 +++
target/arm/cpu.h | 2 ++
target/arm/cpu64.c| 33 +---
with Richard Henderson, who reworked
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
Reviewed-by: Richard Henderson
---
docs/ar
e other SVE CPU properties in later patches.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/cpu64.c | 17 ++---
target/arm/kvm.c | 5 +
target/arm/kvm64.c | 20 +++-
target/arm/kvm
d
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
docs/arm-cpu-features.rst
These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
swabbing is different than it is for fpsmid because the vector format
is a little-endian stream of words.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/kvm64.c | 183
e list of advertised
features, e.g. 'vfp' and 'neon'. The only requirement is that we can
detect invalid configurations and emit failures at QMP query time.
For 'vfp' and 'neon' this will require some refactoring to share a
validation function between the QM
me
cpu_max_get/set_sve_vq to cpu_max_get/set_sve_max_vq in order for them
to follow the typical *_get/set_ pattern.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/cpu.c | 3 ++-
target/arm/cpu64.c | 52 ++
doesn't want to select a
target they can pass -1 for the target which indicates they want
to use the preferred target and have it passed back like before.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/kvm.c | 20 +++-
t
Now that Arm CPUs have advertised features lets add tests to ensure
we maintain their expected availability with and without KVM.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
---
tests/Makefile.include | 5 +-
tests/arm-cpu-features.c | 242 +++
2
7; cpu SVE default. It was still off by default.
- Cleaned up #ifdef's for sve_bswap64()
- Removed redundant KVM_CAP_ARM_SVE extension check in
kvm_arm_sve_get_vls()
- Improved the KVM SVE qtest
- Renamed sve to sve everywhere
- Renamed power-of-2 to power-of-two everywhere
- P
On Wed, Oct 09, 2019 at 03:01:15PM +0100, Beata Michalska wrote:
> On Tue, 1 Oct 2019 at 14:04, Andrew Jones wrote:
> >
> > Introduce cpu properties to give fine control over SVE vector lengths.
> > We introduce a property for each valid length up to the current
> >
On Tue, Oct 15, 2019 at 10:59:16AM +0100, Beata Michalska wrote:
> On Tue, 1 Oct 2019 at 14:04, Andrew Jones wrote:
> > +
> > +obj = object_new(object_class_get_name(oc));
> > +
> > +if (qdict_in) {
> > +Visitor *visitor;
> > +Error
On Wed, Oct 09, 2019 at 08:50:02PM -0400, Richard Henderson wrote:
> On 10/7/19 1:06 PM, Andrew Jones wrote:
> > +static void kvm_arm_vm_state_change(void *opaque, int running, RunState
> > state)
> > +{
> > +CPUState *cs = opaque;
> > +ARMCPU *cpu
On Wed, Oct 09, 2019 at 08:39:21PM -0400, Richard Henderson wrote:
> On 10/4/19 8:03 AM, Andrew Jones wrote:
> > +#ifdef TARGET_AARCH64
> > +static off_t sve_zreg_offset(uint32_t vq, int n)
> > +{
> > +off_t off = sizeof(struct aarch64_user_sve_header);
> > +
kvm-adjvtime is a KVM specific CPU property and a first of its kind.
To accommodate it we also add kvm_arm_add_vcpu_properties() and a
KVM specific CPU properties description to the CPU features document.
Signed-off-by: Andrew Jones
---
docs/arm-cpu-features.rst | 27
patch only provides the implementation. A subsequent patch
will provide the CPU property allowing the feature to be enabled.
Reported-by: Bijan Mottahedeh
Suggested-by: Bijan Mottahedeh
Signed-off-by: Andrew Jones
---
target/arm/cpu.h | 3 +++
target/arm/kvm.c | 20
If we know what the default value should be then we can test for
that as well as the feature existence.
Signed-off-by: Andrew Jones
---
tests/arm-cpu-features.c | 44
1 file changed, 35 insertions(+), 9 deletions(-)
diff --git a/tests/arm-cpu-features.c
When acceleration like KVM is in use it's necessary to use the host's
counter frequency when converting ticks to or from time units.
Signed-off-by: Andrew Jones
---
include/qemu/timer.h | 16
1 file changed, 16 insertions(+)
diff --git a/include/qemu/timer.h b/in
even though we don't use those registers either.
This change was originally posted as a hunk of a different, never
merged patch from Bijan Mottahedeh.
Signed-off-by: Andrew Jones
---
target/arm/kvm64.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/arm/kvm64.c b/target/arm/kvm64.
3.html
[2] https://patchew.org/QEMU/20191001125845.8793-1-drjo...@redhat.com/
Thanks,
drew
Andrew Jones (5):
target/arm/kvm64: kvm64 cpus have timer registers
timer: arm: Introduce functions to get the host cntfrq
target/arm/kvm: Implement cpu feature kvm-adjvtime
tests/arm-cpu-features:
On Mon, Oct 07, 2019 at 09:35:35AM +0100, Alex Bennée wrote:
>
> Andrew Jones writes:
>
> > Introduce cpu properties to give fine control over SVE vector lengths.
> > We introduce a property for each valid length up to the current
> > maximum supported, which is 20
When dumping a guest with dump-guest-memory also dump the SVE
registers if they are in use.
Signed-off-by: Andrew Jones
---
include/elf.h | 2 +
target/arm/arch_dump.c | 135 -
2 files changed, 135 insertions(+), 2 deletions(-)
diff --git a
On Fri, Oct 04, 2019 at 11:46:09AM +0200, Andrew Jones wrote:
> When dumping a guest with dump-guest-memory also dump the SVE
> registers if they are in use.
>
> Signed-off-by: Andrew Jones
> ---
> include/elf.h | 2 +
> target/
When dumping a guest with dump-guest-memory also dump the SVE
registers if they are in use.
Signed-off-by: Andrew Jones
---
include/elf.h | 2 +
target/arm/arch_dump.c | 133 -
2 files changed, 133 insertions(+), 2 deletions(-)
diff --git a
On Thu, Oct 03, 2019 at 11:55:36PM +, Zhang, Lei wrote:
> Hi Jones,
>
> Thanks for your patch.
>
> I have tested the v5 patch.
>
> All the test is passed, except one problem.
> The problem was reported
> in [[Qemu-devel] [RFC QEMU v2 0/2] arm/virt: Account for guest pause time]
> https://lis
On Tue, Oct 01, 2019 at 03:52:46PM +0200, Andrew Jones wrote:
> On Tue, Oct 01, 2019 at 02:58:41PM +0200, Andrew Jones wrote:
> > These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
> > swabbing is different than it is for fpsmid because the vector format
> &g
On Tue, Oct 01, 2019 at 02:58:41PM +0200, Andrew Jones wrote:
> These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
> swabbing is different than it is for fpsmid because the vector format
> is a little-endian stream of words.
>
> Signed-off-by: Andrew Jones
Hi E
e other SVE CPU properties in later patches.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/cpu64.c | 17 ++---
target/arm/kvm.c | 5 +
target/arm/kvm64.c | 20 +++-
target/arm/kvm
with Richard Henderson, who reworked
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
---
docs/arm-cpu-features.rst | 45 +++---
targe
doesn't want to select a
target they can pass -1 for the target which indicates they want
to use the preferred target and have it passed back like before.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/kvm.c | 20 +++-
t
fic when enabling vector
lengths.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
---
docs/arm-cpu-features.rst | 19 ---
target/arm/cpu.c | 3 +++
target/arm/cpu.h | 2 ++
target/arm/cpu64.c| 33 +
target/arm/kvm6
These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
swabbing is different than it is for fpsmid because the vector format
is a little-endian stream of words.
Signed-off-by: Andrew Jones
---
target/arm/kvm64.c | 183 ++---
1 file changed
e list of advertised
features, e.g. 'vfp' and 'neon'. The only requirement is that we can
detect invalid configurations and emit failures at QMP query time.
For 'vfp' and 'neon' this will require some refactoring to share a
validation function between the QM
d
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
---
docs/arm-cpu-features.rst | 168 +++--
include/qemu/bitops.h | 1 +
targe
me
cpu_max_get/set_sve_vq to cpu_max_get/set_sve_max_vq in order for them
to follow the typical *_get/set_ pattern.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/cpu.c | 3 ++-
target/arm/cpu64.c | 52 ++
Now that Arm CPUs have advertised features lets add tests to ensure
we maintain their expected availability with and without KVM.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
---
tests/Makefile.include | 5 +-
tests/arm-cpu-features.c | 242 +++
2
VE qtest
- Renamed sve to sve everywhere
- Renamed power-of-2 to power-of-two everywhere
- Picked up some more tags from Richard
Thanks!
drew
Andrew Jones (9):
target/arm/monitor: Introduce qmp_query_cpu_model_expansion
tests: arm: Introduce cpu feature tests
target/arm: Allow SVE t
On Tue, Sep 24, 2019 at 01:31:01PM +0200, Andrew Jones wrote:
> These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
> swabbing is different than it is for fpsmid because the vector format
> is a little-endian stream of words.
>
> Signed-off-by: Andrew Jones
> Re
On Tue, Sep 24, 2019 at 01:31:01PM +0200, Andrew Jones wrote:
> These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
> swabbing is different than it is for fpsmid because the vector format
> is a little-endian stream of words.
>
> Signed-off-by: Andrew Jones
> Re
On Fri, Sep 27, 2019 at 12:50:55AM +0100, Alex Bennée wrote:
>
> Richard Henderson writes:
>
> > On 9/24/19 4:31 AM, Andrew Jones wrote:
>
> >
> >> +#if __SIZEOF_LONG__ == 8
> >> +#define BIT(n) (1UL << (n))
> >> +#else
> >> +#
On Thu, Sep 26, 2019 at 12:07:54PM -0700, Richard Henderson wrote:
> On 9/24/19 4:31 AM, Andrew Jones wrote:
> > +static uint32_t sve_zcr_get_valid_len(ARMCPU *cpu, uint32_t start_len)
> > +{
> > +uint32_t start_vq = (start_len & 0xf) + 1;
> > +
> > +r
On Thu, Sep 26, 2019 at 12:01:36PM +0200, Auger Eric wrote:
>
>
> On 9/26/19 10:41 AM, Andrew Jones wrote:
> > On Thu, Sep 26, 2019 at 08:52:55AM +0200, Auger Eric wrote:
> >> Hi Drew,
> >>
> >> On 9/24/19 1:31 PM, Andrew Jones wrote:
> >>&g
On Thu, Sep 26, 2019 at 11:34:45AM +0200, Auger Eric wrote:
>
>
> On 9/26/19 10:21 AM, Andrew Jones wrote:
> > On Wed, Sep 25, 2019 at 03:53:42PM +0200, Auger Eric wrote:
> >> Hi Drew,
> >>
> >> On 9/24/19 1:31 PM, Andrew Jones wrote:
> >>>
On Thu, Sep 26, 2019 at 09:07:36AM +0200, Auger Eric wrote:
> Hi Drew,
>
> On 9/24/19 1:31 PM, Andrew Jones wrote:
> > Allow cpu 'host' to enable SVE when it's available, unless the
> > user chooses to disable it with the added 'sve=off' cpu property
On Thu, Sep 26, 2019 at 08:52:55AM +0200, Auger Eric wrote:
> Hi Drew,
>
> On 9/24/19 1:31 PM, Andrew Jones wrote:
> > Extend the SVE vq map initialization and validation with KVM's
> > supported vector lengths when KVM is enabled. In order to determine
> > and sel
On Wed, Sep 25, 2019 at 03:53:42PM +0200, Auger Eric wrote:
> Hi Drew,
>
> On 9/24/19 1:31 PM, Andrew Jones wrote:
> > Introduce cpu properties to give fine control over SVE vector lengths.
> > We introduce a property for each valid length up to the current
> > maxim
On Tue, Sep 24, 2019 at 01:31:00PM +0200, Andrew Jones wrote:
> +SVE CPU Property Parsing Semantics
> +--
> +
> + 1) If SVE is disabled (`sve=off`), then which SVE vector lengths
> + are enabled or disabled is irrelevant to the guest, as the ent
These are the SVE equivalents to kvm_arch_get/put_fpsimd. Note, the
swabbing is different than it is for fpsmid because the vector format
is a little-endian stream of words.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
target/arm/kvm64.c | 137
e list of advertised
features, e.g. 'vfp' and 'neon'. The only requirement is that we can
detect invalid configurations and emit failures at QMP query time.
For 'vfp' and 'neon' this will require some refactoring to share a
validation function between the QM
with Richard Henderson, who reworked
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
---
docs/arm-cpu-features.rst | 36 +---
target/arm/cpu64.c
Now that Arm CPUs have advertised features lets add tests to ensure
we maintain their expected availability with and without KVM.
Signed-off-by: Andrew Jones
Reviewed-by: Eric Auger
---
tests/Makefile.include | 5 +-
tests/arm-cpu-features.c | 242 +++
2
fic when enabling vector
lengths.
Signed-off-by: Andrew Jones
---
docs/arm-cpu-features.rst | 19 ---
target/arm/cpu.c | 3 +++
target/arm/cpu.h | 2 ++
target/arm/cpu64.c| 33 +
target/arm/kvm64.c| 14 +
doesn't want to select a
target they can pass -1 for the target which indicates they want
to use the preferred target and have it passed back like before.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/kvm.c | 20 +++-
t
e other SVE CPU properties in later patches.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
target/arm/cpu64.c | 17 ++---
target/arm/kvm.c | 5 +
target/arm/kvm64.c | 20 +++-
target/arm/kvm_arm.h | 27 +++
d
the target/arm/cpu64.c changes in order to push all the validation and
auto-enabling/disabling steps into the finalizer, resulting in a nice
LOC reduction.
Signed-off-by: Andrew Jones
---
docs/arm-cpu-features.rst | 157 +--
target/arm/cpu.c | 19
targe
sion check in
kvm_arm_sve_get_vls()
- Improved the KVM SVE qtest
- Renamed sve to sve everywhere
- Renamed power-of-2 to power-of-two everywhere
- Picked up some more tags from Richard
Thanks!
drew
Andrew Jones (9):
target/arm/monitor: Introduce qmp_query_cpu_model_expansion
tes
me
cpu_max_get/set_sve_vq to cpu_max_get/set_sve_max_vq in order for them
to follow the typical *_get/set_ pattern.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
---
target/arm/cpu.c | 3 ++-
target/arm/cpu64.c | 42 ++--
target/
!kvm_check_extension(s, KVM_CAP_ARM_IRQ_LINE_LAYOUT_2)) {
> +error_report("Using more than 256 vcpus requires a host kernel "
> + "with KVM_CAP_ARM_IRQ_LINE_LAYOUT_2");
> +ret = -EINVAL;
> +}
> +
> +return ret;
> }
>
> unsigned long kvm_arch_vcpu_id(CPUState *cpu)
> --
> 2.20.1
>
>
Reviewed-by: Andrew Jones
uint64_t address, uint32_t data, PCIDevice *dev)
> {
> diff --git a/target/arm/kvm_arm.h b/target/arm/kvm_arm.h
> index b3106c8600..b4e19457a0 100644
> --- a/target/arm/kvm_arm.h
> +++ b/target/arm/kvm_arm.h
> @@ -253,6 +253,7 @@ int kvm_arm_vgic_probe(void);
>
> void kvm_arm_pmu_set_irq(CPUState *cs, int irq);
> void kvm_arm_pmu_init(CPUState *cs);
> +int kvm_arm_set_irq(int cpu, int irqtype, int irq, int level);
>
> #else
>
> --
> 2.20.1
>
>
Reviewed-by: Andrew Jones
On Fri, Sep 13, 2019 at 10:36:12AM +0200, Eric Auger wrote:
> memory_region_register_iommu_notifier now returns an error
> in case of failure. Assert in such a case.
>
> Signed-off-by: Eric Auger
> ---
> exec.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/exec.c b/
q_line_layout_mismatch(MachineState *ms, int vcpus);
> +
> /**
> * kvm_arm_sync_mpstate_to_kvm
> * @cpu: ARMCPU
> @@ -281,6 +291,11 @@ static inline int
> kvm_arm_get_max_vm_ipa_size(MachineState *ms)
> return -ENOENT;
> }
>
> +static inline bool kvm_arm_irq_line_layout_mismatch(MachineState *ms, int
> vcpus)
> +{
> +return false;
> +}
> +
> static inline int kvm_arm_vgic_probe(void)
> {
> return 0;
> --
> 2.20.1
>
>
Reviewed-by: Andrew Jones
On Wed, Sep 11, 2019 at 05:51:24PM +0200, Eric Auger wrote:
> Host kernels that expose the KVM_CAP_ARM_IRQ_LINE_LAYOUT_2 capability
> allow injection of interrupts along with vcpu ids larger than 255.
> Let's encode the vpcu id on 12 bits according to the upgraded KVM_IRQ_LINE
> ABI when needed.
>
On Fri, Aug 09, 2019 at 06:31:12PM -0700, Richard Henderson wrote:
> As promised, an experiment in unifying the checks.
>
> I believe that I've tested for all of the same conditions, modulo the
> timing at which they are emitted. As before, only one error is reported,
> so if multiple errors exis
On Tue, Aug 20, 2019 at 06:08:04AM +, Zhang, Lei wrote:
> Hi Andrew,
>
> I have tested your patch on kernel- 5.2.7 + QEMU (4.0.94 + patch).
Thanks for the testing! I guess it's time for me to get back to this
series and spin a v4 (so we can test some more :-)
>
> This patch series works fin
On Thu, Aug 15, 2019 at 09:31:35AM +0100, Peter Maydell wrote:
> On Fri, 2 Aug 2019 at 13:25, Andrew Jones wrote:
> >
> > Since Linux kernel v5.2-rc1 KVM has support for enabling SVE in guests.
> > This series provides the QEMU bits for that enablement. First, we
&g
On Thu, Aug 08, 2019 at 11:37:01AM -0700, Richard Henderson wrote:
> On 8/8/19 1:50 AM, Andrew Jones wrote:
> > I'm not sure. Of course I'd need to experiment with it to be sure, but
> > I'm reluctant to go through that exercise, because I believe that a
> > def
On Wed, Aug 07, 2019 at 08:22:07AM -0700, Richard Henderson wrote:
> On 8/6/19 5:21 AM, Andrew Jones wrote:
> > That's a reasonable suggestion. I do like having self-contained
> > validation, self-contained, but when cross-dependencies arise, then
> > it does make sense t
On Fri, Aug 02, 2019 at 11:07:54AM -0700, Richard Henderson wrote:
> On 8/2/19 5:25 AM, Andrew Jones wrote:
> > +/*
> > + * SVE registers are encoded in KVM's memory in an endianness-invariant
> > format.
> > + * The byte at offset i from the start of the in-mem
On Fri, Aug 02, 2019 at 06:28:51PM -0700, Richard Henderson wrote:
> On 8/2/19 9:27 AM, Richard Henderson wrote:
> > On 8/2/19 5:25 AM, Andrew Jones wrote:
> >> Note, certainly more features may be added to the list of
> >> advertised features, e.g. 'vfp' and
e other SVE CPU properties in later patches.
Signed-off-by: Andrew Jones
---
target/arm/cpu64.c | 17 ++---
target/arm/kvm.c | 5 +
target/arm/kvm64.c | 20 +++-
target/arm/kvm_arm.h | 27 +++
tests/arm-cpu-features.c
doesn't want to select a
target they can pass -1 for the target which indicates they want
to use the preferred target and have it passed back like before.
Signed-off-by: Andrew Jones
Reviewed-by: Richard Henderson
Reviewed-by: Eric Auger
---
target/arm/kvm.c | 20 +++-
t
1101 - 1200 of 2468 matches
Mail list logo
