Hi,
On 12/12/23 08:03, Irina Ryapolova wrote:
[Changes since v1]
used satp_mode.map instead of satp_mode.supported
[Original cover]
The SATP register is an SXLEN-bit read/write WARL register. It means that CSR
fields are only defined
for a subset of bit encodings, but allow any value to be wri
On 2023/12/21 19:38, Alex Bennée wrote:
We can only request a list of registers once the vCPU has been
initialised so the user needs to use either call the get function on
vCPU initialisation or during the translation phase.
We don't expose the reg number to the plugin instead hiding it behind
a
On 12/12/23 08:03, Irina Ryapolova wrote:
Added xATP_MODE validation for vsatp/hgatp CSRs.
The xATP register is an SXLEN-bit read/write WARL register, so
the legal value must be returned (See riscv-privileged-20211203,
SATP/VSATP/HGATP CSRs).
Signed-off-by: Irina Ryapolova
---
Reviewed-by
BALATON Zoltan writes:
> On Wed, 20 Dec 2023, Fabiano Rosas wrote:
>> Warner Losh writes:
>>
>>> On Tue, Dec 19, 2023, 1:55 PM Peter Maydell
>>> wrote:
>>>
On Tue, 19 Dec 2023 at 19:40, Fabiano Rosas wrote:
>
> Dave Blanchard writes:
>
>> Hello all, can you please help me
Am 20.12.2023 um 14:47 hat Stefan Hajnoczi geschrieben:
> Add the iothread-vq-mapping parameter to assign virtqueues to IOThreads.
> Store the vq:AioContext mapping in the new struct
> VirtIOBlockDataPlane->vq_aio_context[] field and refactor the code to
> use the per-vq AioContext instead of the B
On 12/20/23 12:32, Ivan Klokov wrote:
This patch changes behavior on writing RW=01 to pmpcfg with MML=0.
RWX filed is form of collective WARL with the combination of
pmpcfg.RW=01 remains reserved for future standard use.
According to definition of WARL writing the CSR has no other side
effect
Signed-off-by: Paolo Bonzini
---
tcg/i386/tcg-target.c.inc | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/tcg/i386/tcg-target.c.inc b/tcg/i386/tcg-target.c.inc
index a83f8aab304..4e9f372d4fd 100644
--- a/tcg/i386/tcg-target.c.inc
+++ b/tcg/i386/tcg-target.c.inc
@@ -1332,
We used to set default page_size_mask to qemu_target_page_mask() but
with VFIO assignment it makes more sense to use the actual host page mask
instead.
So from now on qemu_real_host_page_mask() will be used as a default.
To be able to migrate older code, we increase the vmstat version_id
to 3 and
This series of patches adds LCOFI delegation from HS-mode to VS-mode.
This possibility must be implemented, as in the AIA spec in section 6.3.2
it is indicated in table 6.1 that in the case when the hideleg bit is set,
the corresponding vsip bit is an alias to the corresponding sip bit, also
for e
target/i386: Eip error in x86_64-softmmu
When closing PCREL, qemu-system-x86_64 run into error.
Eip modification here leads to the result.
Fixes: b5e0d5d22fbf("target/i386: Fix 32-bit wrapping of pc/eip computation")
Changes to be committed:
modified: target/i386/tcg/translate.c
---
In the AIA specification in the paragraph "Virtual interrupts for VS level"
it is indicated for interrupts 13-63: if the bit in hideleg is enabled,
then the corresponding vsip and vsie bits are aliases to sip and sie
Signed-off-by: Vadim Shakirov
---
target/riscv/csr.c | 36 +
In AIA spec in 5.1 chapter says that "... the AIA categorizes the counter
overflow interrupt (code 13) as a local interrupt. It is assumed furthermore
that any future definitions for reserved interrupt numbers 14 and 15 will
also be local interrupts" and than LCOFI belongs to LOCAL_INTERRUPTS
Sign
On Thursday, December 21, 2023 7:54 PM, Li, Xiaoyao wrote:
> On 12/21/2023 6:36 PM, Wang, Wei W wrote:
> > No need to specifically check for KVM_MEMORY_ATTRIBUTE_PRIVATE there.
> > I'm suggesting below:
> >
> > diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c index
> > 2d9a2455de..63ba74b221
On Thu, Dec 21, 2023 at 11:45:36AM +0100, Kevin Wolf wrote:
> Am 21.12.2023 um 02:49 hat Stefan Hajnoczi geschrieben:
> > NBDClient has a number of fields that are accessed by both the export
> > AioContext and the main loop thread. When the AioContext lock is removed
> > these fields will need ano
Am 19. Dezember 2023 00:09:23 UTC schrieb BALATON Zoltan :
>On Mon, 18 Dec 2023, Bernhard Beschow wrote:
>> Implement isa_fdc_set_{enabled,iobase} in order to implement relocation and
>> toggling of SuperI/O functions in the VIA south bridges without breaking
>> encapsulation.
>
>You may want to
On Thu, Dec 21, 2023 at 08:23:15AM +0100, Paolo Bonzini wrote:
> On 12/21/23 02:49, Stefan Hajnoczi wrote:
> > The NBD clients list is currently accessed from both the export
> > AioContext and the main loop thread. When the AioContext lock is removed
> > there will be nothing protecting the client
Am 19. Dezember 2023 00:11:37 UTC schrieb BALATON Zoltan :
>On Mon, 18 Dec 2023, Bernhard Beschow wrote:
>> This is a preparation for implementing relocation and toggling of SuperI/O
>> functions in the VT8231 device model. Upon reset, all SuperI/O functions
>> will be
>> deactivated, so in cas
When running current kernel on s390x in tcg mode *on s390x hw*, the following
is generated when loading crypto selftest module (it gets loaded automatically):
[ 10.546690] alg: skcipher: chacha20-s390 encryption test failed (wrong result) on test
vector 1, cfg="in-place (one sglist)"
[ 10.54
We reserved type names ending with 'Kind' because a simple union
'SomeSimpleUnion' generated both a struct type SomeSimpleUnion and an
enum type SomeSimpleUnionKind. Gone since commit 4e99f4b12c0 (qapi:
Drop simple unions). The commit neglected to update the documentation
not to reserve type name
ParallelState::portio_list isn't used inside ParallelState context but only
inside ISAParallelState context, so move it there.
Signed-off-by: Bernhard Beschow
Reviewed-by: BALATON Zoltan
---
include/hw/char/parallel-isa.h | 2 ++
include/hw/char/parallel.h | 2 --
hw/char/parallel.c
FDCtrl::portio_list isn't used inside FDCtrl context but only inside
FDCtrlISABus context, so move it there.
Signed-off-by: Bernhard Beschow
Reviewed-by: BALATON Zoltan
---
hw/block/fdc-internal.h | 2 --
hw/block/fdc-isa.c | 4 +++-
2 files changed, 3 insertions(+), 3 deletions(-)
diff -
portio_list_add_1() creates a MemoryRegionPortioList instance which holds a
MemoryRegion `mr` and an array of MemoryRegionPortio elements named `ports`.
Each element in the array gets assigned the same value for its .base attribute.
The same value also ends up as the .addr attribute of `mr` due to
This series implements relocation of the SuperI/O functions of the VIA south
bridges which resolves some FIXME's. It is part of my via-apollo-pro-133t
branch [1] which is an extension of bringing the VIA south bridges to the PC
machine [2]. This branch is able to run some real-world X86 BIOSes in t
This is a preparation for implementing relocation and toggling of SuperI/O
functions in the VT8231 device model. Upon reset, all SuperI/O functions will be
deactivated, so in case if no -bios is given, let the machine configure those
functions the same way Pegasos II firmware would do.
Signed-off-
Some SuperI/O devices such as the VIA south bridges or the PC87312 controller
are able to relocate their SuperI/O functions. Add a convenience function for
implementing this in the VIA south bridges.
This convenience function relies on previous simplifications in exec/ioport
which avoids some dupl
The VIA south bridges are able to relocate and toggle (enable or disable) their
SuperI/O functions. So far this is hardcoded such that all functions are always
enabled and are located at fixed addresses.
Some PC BIOSes seem to probe for I/O occupancy before activating such a function
and issue an
The real SuperI/O chips emulated by QEMU allow for relocating and enabling or
disabling their SuperI/O functions via software. So far this is not implemented.
Prepare for that by adding isa_parallel_set_{enabled,iobase}.
Signed-off-by: Bernhard Beschow
---
include/hw/char/parallel-isa.h | 3 +++
FDCtrl::iomem isn't used inside FDCtrl context but only inside FDCtrlSysBus
context, so move it there.
Signed-off-by: Bernhard Beschow
Reviewed-by: BALATON Zoltan
---
hw/block/fdc-internal.h | 2 --
hw/block/fdc-sysbus.c | 6 --
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git
Some SuperI/O devices such as the VIA south bridges or the PC87312 controller
allow to enable or disable their SuperI/O functions. Add a convenience function
for implementing this in the VIA south bridges.
The naming of the functions is inspired by its memory_region_set_enabled()
pendant.
Signed-
The real SuperI/O chips emulated by QEMU allow for relocating and enabling or
disabling their SuperI/O functions via software. So far this is not implemented.
Prepare for that by adding isa_serial_set_{enabled,iobase}.
Signed-off-by: Bernhard Beschow
---
include/hw/char/serial.h | 2 ++
hw/char
The real SuperI/O chips emulated by QEMU allow for relocating and enabling or
disabling their SuperI/O functions via software. So far this is not implemented.
Prepare for that by adding isa_fdc_set_{enabled,iobase}.
Signed-off-by: Bernhard Beschow
---
include/hw/block/fdc.h | 3 +++
hw/block/fd
On 20/12/2023 21:51, Richard Henderson wrote:
On 12/20/23 22:03, Michael Brown wrote:
For the default case, I think it would make sense to unconditionally
truncate the address to 32 bits if paging is disabled. (I am not sure
why the original commit 33dfdb5 included a test for long mode, since
Signed-off-by: Stefan Hajnoczi
---
migration/block.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/migration/block.c b/migration/block.c
index 2bcfcbfdf6..6ec6a1d6e6 100644
--- a/migration/block.c
+++ b/migration/block.c
@@ -311,10 +311,17 @@ static int mig_save_device_bulk(QEMUFile
v2:
- Drop useless if (client->quiesing) check [Kevin]
- run qio_channel_read_wake() in export AioContext to avoid race [Kevin]
- Introduce nbd_client_put_nonzero() optimization [Paolo]
- Reach goto label disconnect with client->lock taken [Paolo]
- Add doc comment explaining nbd_co_receive_request
qemu-iotests 051 fails on my machine so the change to 051.pc.out made by
the above commit appears to be incorrect, at least against the current
QEMU source tree. Revert it so that 051 passes again.
Signed-off-by: Stefan Hajnoczi
---
tests/qemu-iotests/051.pc.out | 4 ++--
1 file changed, 2 inser
nbd_trip() processes a single NBD request from start to finish and holds
an NBDClient reference throughout. NBDRequest does not outlive the scope
of nbd_trip(). Therefore it is unnecessary to ref/unref NBDClient for
each NBDRequest.
Removing these nbd_client_get()/nbd_client_put() calls will make
NBDClient has a number of fields that are accessed by both the export
AioContext and the main loop thread. When the AioContext lock is removed
these fields will need another form of protection.
Add NBDClient->lock and protect fields that are accessed by both
threads. Also add assertions where poss
Signed-off-by: Stefan Hajnoczi
Reviewed-by: Paolo Bonzini
---
hw/scsi/scsi-bus.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/hw/scsi/scsi-bus.c b/hw/scsi/scsi-bus.c
index df68a44b6a..5b08cbf60a 100644
--- a/hw/scsi/scsi-bus.c
+++ b/hw/scsi/scsi-bus.c
@@ -127,7 +127,8 @@
The NBD clients list is currently accessed from both the export
AioContext and the main loop thread. When the AioContext lock is removed
there will be nothing protecting the clients list.
Adding a lock around the clients list is tricky because NBDClient
structs are refcounted and may be freed from
On Thu, Dec 21, 2023 at 01:39:48PM +0100, Kevin Wolf wrote:
> Am 20.12.2023 um 14:47 hat Stefan Hajnoczi geschrieben:
> > qdev_alias_all_properties() aliases a DeviceState's qdev properties onto
> > an Object. This is used for VirtioPCIProxy types so that --device
> > virtio-blk-pci has properties
The address translation logic in get_physical_address() will currently
truncate physical addresses to 32 bits unless long mode is enabled.
This is incorrect when using physical address extensions (PAE) outside
of long mode, with the result that a 32-bit operating system using PAE
to access memory a
[ ... ]
+static void iommufd_backend_init(Object *obj)
+{
+IOMMUFDBackend *be = IOMMUFD_BACKEND(obj);
+
+be->fd = -1;
+be->users = 0;
+be->owned = true;
+qemu_mutex_init(&be->lock);> +}
+
+static void iommufd_backend_finalize(Object *obj)
+{
+IOMMUFDBackend *be = IOMMUFD
On 12/21/23 08:36, Vadim Shakirov wrote:
In the AIA specification in the paragraph "Virtual interrupts for VS level"
it is indicated for interrupts 13-63: if the bit in hideleg is enabled,
then the corresponding vsip and vsie bits are aliases to sip and sie
Signed-off-by: Vadim Shakirov
---
On 12/21/23 08:36, Vadim Shakirov wrote:
In AIA spec in 5.1 chapter says that "... the AIA categorizes the counter
overflow interrupt (code 13) as a local interrupt. It is assumed furthermore
that any future definitions for reserved interrupt numbers 14 and 15 will
also be local interrupts" an
I am currently focusing in kernel development, so I will probably not be
of much help in reviewing general Live Migration changes.
For above reason I am removing my Reviewer status from Migration and RDMA
Migration.
Signed-off-by: Leonardo Bras
---
MAINTAINERS | 2 --
1 file changed, 2 deletion
From: Zhao Liu
As commit 3e015d815b3f ("use g_path_get_basename instead of basename")
said, g_path_get_dirname() should be preferred over dirname() since
the former is a portable utility function that has the advantage of not
modifing the string argument.
Replace dirname() with g_path_get_dirnam
From: Zhao Liu
As commit 3e015d815b3f ("use g_path_get_basename instead of basename")
said, g_path_get_basename() and g_path_get_dirname() should be preferred
over basename() and dirname(), since g_path_get_basename() and
g_path_get_dirname() are portable utility functions that have the
advantage
From: Zhao Liu
g_path_get_basename() is a portable utility function that has the
advantage of not modifing the string argument, so it should be
preferred over basename().
And also to avoid potential compile breakage with the Musl C library
similar to [1], replace basename() with g_path_get_basen
Hi Cédric,
On 12/21/23 17:00, Cédric Le Goater wrote:
> [ ... ]
>
>
>> +static void iommufd_backend_init(Object *obj)
>> +{
>> + IOMMUFDBackend *be = IOMMUFD_BACKEND(obj);
>> +
>> + be->fd = -1;
>> + be->users = 0;
>> + be->owned = true;
>> + qemu_mutex_init(&be->lock);> +}
>> +
>>
CONFIG_ALL is tricky to use and was ported over to Meson from the
recursive processing of Makefile variables. Meson sourcesets
however have all_sources() and all_dependencies() methods that
remove the need for it.
Signed-off-by: Paolo Bonzini
---
docs/devel/build-system.rst | 15 ---
Since the option is of boolean type, the default value should be a boolean
rather than a string.
Signed-off-by: Paolo Bonzini
---
meson_options.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/meson_options.txt b/meson_options.txt
index c9baeda6395..0a99a059ec8 100644
---
Signed-off-by: Paolo Bonzini
---
meson.build | 90 ++---
1 file changed, 45 insertions(+), 45 deletions(-)
diff --git a/meson.build b/meson.build
index b8c64c5ecf7..1a1e006ba13 100644
--- a/meson.build
+++ b/meson.build
@@ -19,21 +19,8 @@ ss = impo
Signed-off-by: Paolo Bonzini
---
Makefile | 1 +
configure | 2 +-
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/Makefile b/Makefile
index 676a4a54f48..8f369903356 100644
--- a/Makefile
+++ b/Makefile
@@ -202,6 +202,7 @@ clean: recurse-clean
! -path ./roms/edk2/Ar
CONFIG_DARWIN, CONFIG_LINUX and CONFIG_BSD are used in some rules, but
only CONFIG_LINUX has substantial use. Convert them all to if...endif.
Signed-off-by: Paolo Bonzini
---
accel/tcg/meson.build | 4 +++-
backends/meson.build | 4 +++-
block/meson.build | 4 +++-
fsdev/meson.build
config_targetos is now empty and can be removed; its use in sourcesets
that do not involve target-specific files can be replaced with an empty
dictionary.
In fact, at this point *all* sourcesets that do not involve
target-specific files are just glorified mutable arrays. Enforce that
they never t
For consistency with other OSes, use if...endif for rules that are
target-independent.
Signed-off-by: Paolo Bonzini
---
backends/meson.build | 6 --
block/meson.build| 7 +--
chardev/meson.build | 26 ++
hw/usb/meson.build | 4 +++-
meson.build
Do not use a subshell to hide the shadowing of $config_host_mak.
Signed-off-by: Paolo Bonzini
---
configure | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/configure b/configure
index 6ea8f53310e..c634948051b 100755
--- a/configure
+++ b/configure
@@ -1671,10 +1671,9
These are cleanups that we can do with a completed conversion
and a global view of meson.build, including:
1) removal of CONFIG_ALL
2) regrouping of root meson.build into better-defined sections
3) renaming of targetos into host_os
4) cleaning up of probe_target_compiler
Plus a bunch of small
Signed-off-by: Paolo Bonzini
---
meson.build | 27 ++-
1 file changed, 14 insertions(+), 13 deletions(-)
diff --git a/meson.build b/meson.build
index f105462c181..57821069e91 100644
--- a/meson.build
+++ b/meson.build
@@ -40,7 +40,6 @@ qemu_moddir = get_option('libdir') /
While a simple lexicographic comparison usually works, it is less
robust than a more specific algorithm designed to compare versions.
Signed-off-by: Paolo Bonzini
---
qga/meson.build | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/qga/meson.build b/qga/meson.build
index ff7a8
And move away dependencies that are not subprojects anymore.
Signed-off-by: Paolo Bonzini
---
meson.build | 72 ++---
1 file changed, 36 insertions(+), 36 deletions(-)
diff --git a/meson.build b/meson.build
index e777f507ce3..f8d61a46ad1 100644
--
Signed-off-by: Paolo Bonzini
---
meson.build | 66 ++---
1 file changed, 33 insertions(+), 33 deletions(-)
diff --git a/meson.build b/meson.build
index 57821069e91..6cf70cbfb89 100644
--- a/meson.build
+++ b/meson.build
@@ -3221,39 +3221,6 @@ forea
Signed-off-by: Paolo Bonzini
---
tests/bench/meson.build | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/tests/bench/meson.build b/tests/bench/meson.build
index 3c799dbd983..7e76338a52d 100644
--- a/tests/bench/meson.build
+++ b/tests/bench/meson.build
@@ -3,9 +3,9 @@ qh
config_all now lists only accelerators, rename it to indicate its actual
content.
Signed-off-by: Paolo Bonzini
---
hw/mips/meson.build | 2 +-
meson.build | 16
target/arm/meson.build | 2 +-
target/mips/meson.build | 2 +-
tests/fp/meson.build| 2 +-
te
Signed-off-by: Paolo Bonzini
---
meson.build | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/meson.build b/meson.build
index 6c77d9687de..de33e323d43 100644
--- a/meson.build
+++ b/meson.build
@@ -1759,12 +1759,12 @@ if not get_option('smartcard').auto() or have_system
Keep it together with the other compiler modes, and before dependencies.
Signed-off-by: Paolo Bonzini
---
meson.build | 80 ++---
1 file changed, 40 insertions(+), 40 deletions(-)
diff --git a/meson.build b/meson.build
index f8d61a46ad1..f105462c1
This variable is about the host OS, not the target. It is used a lot
more since the Meson conversion, but the original sin dates back to 2003.
Time to fix it.
Signed-off-by: Paolo Bonzini
---
accel/tcg/meson.build| 2 +-
backends/meson.build | 4 +-
block/mes
Signed-off-by: Paolo Bonzini
---
meson.build | 97 +++--
1 file changed, 50 insertions(+), 47 deletions(-)
diff --git a/meson.build b/meson.build
index f7e7deeabdc..e777f507ce3 100644
--- a/meson.build
+++ b/meson.build
@@ -602,53 +602,6 @@ have_vh
Not entirely possible for option that depend on compiler or dependency checks,
but it's a start.
Signed-off-by: Paolo Bonzini
---
meson.build | 52
1 file changed, 28 insertions(+), 24 deletions(-)
diff --git a/meson.build b/meson.build
index
Signed-off-by: Paolo Bonzini
---
meson.build | 24 ++--
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/meson.build b/meson.build
index b5a85c4c255..b8c64c5ecf7 100644
--- a/meson.build
+++ b/meson.build
@@ -9,6 +9,10 @@ add_test_setup('thorough', env: ['G_TEST_
Queued, thanks.
Paolo
Remove assignments that match the default, and group the
targets for debian-legacy-test-cross and debian-all-test-cross
into a single arm.
Signed-off-by: Paolo Bonzini
---
configure | 86 +--
1 file changed, 27 insertions(+), 59 deletions(-)
d
Am 21.12.2023 um 16:35 hat Stefan Hajnoczi geschrieben:
> NBDClient has a number of fields that are accessed by both the export
> AioContext and the main loop thread. When the AioContext lock is removed
> these fields will need another form of protection.
>
> Add NBDClient->lock and protect fields
This is now expected by rtd so I've expanded using their example as
22.04 is one of our supported platforms. I tried to work out if there
was an easy way to re-generate a requirements.txt from our
pythondeps.toml but in the end went for the easier solution.
Signed-off-by: Alex Bennée
Cc: qemu-sta
It will hold properties shared among all vhost_vdpa instances associated
with of the same device. For example, we just need one iova_tree or one
memory listener for the entire device.
Next patches will register the vhost_vdpa memory listener at the
beginning of the VM migration at the destination
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
It is always VHOST_IOTLB_MSG_V2. We can always make it back per
vhost_dev if needed.
This change makes easier for vhost_vdpa_map and unmap not to depend on
vhost_vdpa but only in VhostVDPAShared.
Signed-off-by: Eugenio Pérez
Acked-by: Jason Wang
---
include/hw/virtio/vhost-vdpa.h | 1 -
hw/vir
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
The memory listener functions can call these too. Make vdpa_iommu work
with VhostVDPAShared.
Signed-off-by: Eugenio Pérez
Acked-by: Jason Wang
---
include/hw/virtio/vhost-vdpa.h | 2 +-
hw/virtio/vhost-vdpa.c | 16
2 files changed, 9 insertions(+), 9 deletions(-)
dif
By the end of this series dma_map and dma_unmap functions don't have the
vdpa device for tracing. Movinge trace function to shared member one.
Print it also in the vdpa initialization so log reader can relate them.
Signed-off-by: Eugenio Pérez
Acked-by: Jason Wang
---
hw/virtio/vhost-vdpa.c |
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
The callers only have the shared information by the end of this series.
Start converting this functions.
Signed-off-by: Eugenio Pérez
Acked-by: Jason Wang
---
include/hw/virtio/vhost-vdpa.h | 4 +--
hw/virtio/vhost-vdpa.c | 50 +-
net/vhost-vdpa.c
Current memory operations like pinning may take a lot of time at the
destination. Currently they are done after the source of the migration is
stopped, and before the workload is resumed at the destination. This is a
period where neigher traffic can flow, nor the VM workload can continue
(downtim
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
Next patches will register the vhost_vdpa memory listener while the VM
is migrating at the destination, so we can map the memory to the device
before stopping the VM at the source. The main goal is to reduce the
downtime.
However, the destination QEMU is unaware of which vhost_vdpa device will
re
Every property in riscv_cpu_options[] will be migrated to
riscv_cpu_properties[]. This will make their default values init
earlier, allowing cpu_init() functions to overwrite them. We'll also
implement common getters and setters that both accelerators will use,
allowing them to share validations th
Hi,
Rob reported a few days ago about vendor properties overrides that we're
having with riscv_cpu_options [1]. At that time I was doing changes in
the 'vlen' and 'elen' properties due to (yet to be sent) changes I'm
making in Vector code. I decided to extend the work I was doing and it
resulted i
They aren't being used.
Signed-off-by: Daniel Henrique Barboza
---
target/riscv/cpu_cfg.h | 2 --
1 file changed, 2 deletions(-)
diff --git a/target/riscv/cpu_cfg.h b/target/riscv/cpu_cfg.h
index f4605fb190..c67a8731d3 100644
--- a/target/riscv/cpu_cfg.h
+++ b/target/riscv/cpu_cfg.h
@@ -136,8 +
Do the same we did with 'cbom_blocksize' in the previous patch.
Remove the now unused kvm_cpu_set_cbomz_blksize() setter.
Signed-off-by: Daniel Henrique Barboza
---
target/riscv/cpu.c | 45 +-
target/riscv/kvm/kvm-cpu.c | 28
'priv_spec' and 'vext_spec' are two string options used as a fancy way
of setting integers in the CPU state (cpu->env.priv_ver and
cpu->env.vext_ver). It requires us to deal with string parsing and to
store them in cpu_cfg.
We must support these string options, but we don't need to store them.
We
Commit 7f0bdfb5bfc ("target/riscv/cpu.c: remove cfg setup from
riscv_cpu_init()") already did some of the work by making some
cpu_init() functions to explictly enable their own 'mmu' default.
The generic CPUs didn't get update by that commit, so they are still
relying on the defaults set by the 'm
The same rework did in 'priv_spec' is done for 'vext_spec'. This time is
simpler, since we only accept one value ("v1.0") and we'll always have
env->vext_ver set to VEXT_VERSION_1_00_0, thus we don't need helpers to
convert string to 'vext_ver' back and forth like we needed for
'priv_spec'.
Signed
Turning 'vlen' into a class property will allow its default value to be
overwritten by cpu_init() later on, solving the issue we have now where
CPU specific settings are getting overwritten by the default.
For 'vlen', 'elen' and the blocksize options we need a way of tracking
if the user set a val
After adding a KVM finalize() implementation, turn cbom_blocksize into a
class property. Follow the same design we used with 'vlen' and 'elen'.
The duplicated 'cbom_blocksize' KVM property can be removed from
kvm_riscv_add_cpu_user_properties().
Signed-off-by: Daniel Henrique Barboza
---
target
Do the same thing we did with 'vlen' in the previous patch with 'elen'.
Signed-off-by: Daniel Henrique Barboza
---
target/riscv/cpu.c | 52 --
target/riscv/tcg/tcg-cpu.c | 5
2 files changed, 50 insertions(+), 7 deletions(-)
diff --git a/target/
To turn cbom_blocksize and cboz_blocksize into class properties we need
KVM specific changes.
KVM is creating its own version of these options with a customized
setter() that prevents users from picking an invalid value during init()
time. This comes at the cost of duplicating each option that KVM
We'll use this function in target/riscv/cpu.c to implement setters that
won't allow vendor CPU options to be changed.
Signed-off-by: Daniel Henrique Barboza
---
target/riscv/cpu.c | 5 +
target/riscv/cpu.h | 1 +
target/riscv/tcg/tcg-cpu.c | 5 -
3 files changed, 6 insert
101 - 200 of 273 matches
Mail list logo
