On Fri, Jan 13, 2023 at 5:25 AM Jason Wang wrote:
>
>
> 在 2023/1/13 01:24, Eugenio Pérez 写道:
> > A vdpa net device must initialize with SVQ in order to be migratable,
> > and initialization code verifies conditions. If the device is not
> > initialized with the x-svq parameter, it will not
On Fri, Jan 13, 2023 at 5:10 AM Jason Wang wrote:
>
> On Fri, Jan 13, 2023 at 1:24 AM Eugenio Pérez wrote:
> >
> > At this moment it is only possible to migrate to a vdpa device running
> > with x-svq=on. As a protective measure, the rewind of the inflight
> > descriptors was done at the
Hi all,
qemu 7.2.0 fails to boot my usual test setup using -kernel (see
the actual script below). I've bisected this down to:
commit ffe2d2382e5f1aae1abc4081af407905ef380311
Author: Jason A. Donenfeld
Date: Wed Sep 21 11:31:34 2022 +0200
x86: re-enable rng seeding via SetupData
with
On 12/1/23 20:14, Kevin Wolf wrote:
blk_unref() can't report any errors that happen while closing the image.
For example, if qcow2 hits an -ENOSPC error while writing out dirty
bitmaps when it's closed, it prints error messages to stderr, but
'qemu-img bitmap' won't see any error return value
Drive-by comment...
Kevin Wolf writes:
> This series addresses the problem described in these bug reports:
> https://gitlab.com/qemu-project/qemu/-/issues/1330
> https://bugzilla.redhat.com/show_bug.cgi?id=2147617
>
> qcow2 can fail when writing back dirty bitmaps in qcow2_inactivate().
>
On 12/1/23 20:14, Kevin Wolf wrote:
In order to write the bitmap table to the image file, it is converted to
big endian. If the write fails, it is passed to clear_bitmap_table() to
free all of the clusters it had allocated before. However, if we don't
convert it back to native endianness first,
On Fri, Jan 13, 2023 at 4:53 AM Jason Wang wrote:
>
> On Fri, Jan 13, 2023 at 1:24 AM Eugenio Pérez wrote:
> >
> > Only create iova_tree if and when it is needed.
> >
> > The cleanup keeps being responsability of last VQ but this change allows
> > to merge both cleanup functions.
> >
> >
On Fri, Jan 13, 2023 at 12:53 AM BALATON Zoltan wrote:
> On Thu, 12 Jan 2023, Howard Spoelstra wrote:
> > On Wed, Jan 11, 2023 at 1:15 AM BALATON Zoltan
> wrote:
> >
> >> On Tue, 10 Jan 2023, Mark Cave-Ayland wrote:
> >>> On 04/01/2023 21:59, BALATON Zoltan wrote:
> >>>
> Setting emulated
On 13/01/2023 08.05, Philippe Mathieu-Daudé wrote:
On 13/1/23 02:05, Richard Henderson wrote:
On 1/11/23 08:31, Philippe Mathieu-Daudé wrote:
Implement Richard's suggestion to use __builtin_bswap().
Convert to __builtin_bswap() one patch per OS to simplify
maintainers review.
Since v2:
-
On 12/1/23 01:34, Alistair Francis wrote:
On Mon, Jan 2, 2023 at 9:55 PM Daniel Henrique Barboza
wrote:
The microchip_icicle_kit, sifive_u, spike and virt boards are now doing
the same steps when '-kernel' is used:
- execute load_kernel()
- load init_rd()
- write kernel_cmdline
Let's fold
On 13/1/23 03:47, Bin Meng wrote:
On Fri, Jan 13, 2023 at 6:37 AM Daniel Henrique Barboza
wrote:
The microchip_icicle_kit, sifive_u, spike and virt boards are now doing
the same steps when '-kernel' is used:
- execute load_kernel()
- load init_rd()
- write kernel_cmdline
Let's fold
On 11/1/23 18:21, Daniel Henrique Barboza wrote:
Follow the QEMU convention of naming MachineState pointers as 'ms' by
renaming the instance in create_fdt() where we're calling it 'mc'.
Cc: Paul Burton
Cc: Aleksandar Rikalo
Suggested-by: Philippe Mathieu-Daudé
Signed-off-by: Daniel Henrique
On 13/1/23 02:05, Richard Henderson wrote:
On 1/11/23 08:31, Philippe Mathieu-Daudé wrote:
Implement Richard's suggestion to use __builtin_bswap().
Convert to __builtin_bswap() one patch per OS to simplify
maintainers review.
Since v2:
- Rebased adapting ./configure changes to meson
Since
On Fri, Jan 13, 2023 at 4:12 AM Jason Wang wrote:
>
> On Fri, Jan 13, 2023 at 1:24 AM Eugenio Pérez wrote:
> >
> > VHOST_BACKEND_F_IOTLB_ASID is the feature bit, not the bitmask. Since
> > the device under test also provided VHOST_BACKEND_F_IOTLB_MSG_V2 and
> > VHOST_BACKEND_F_IOTLB_BATCH, this
Hi Alistair,
On Thu, Jan 12, 2023 at 8:36 AM Alistair Francis wrote:
>
> On Mon, Jan 2, 2023 at 9:55 PM Daniel Henrique Barboza
> wrote:
> >
> > The microchip_icicle_kit, sifive_u, spike and virt boards are now doing
> > the same steps when '-kernel' is used:
> >
> > - execute load_kernel()
> >
On Fri, Jan 13, 2023 at 1:25 AM Eugenio Pérez wrote:
>
> This allows net to restart the device backend to configure SVQ on it.
>
> Ideally, these changes should not be net specific. However, the vdpa net
> backend is the one with enough knowledge to configure everything because
> of some reasons:
On Fri, Jan 13, 2023 at 1:25 AM Eugenio Pérez wrote:
>
> Since some actions move to the start function instead of init, the
> device features may not be the parent vdpa device's, but the one
> returned by vhost backend. If transition to SVQ is supported, the vhost
> backend will return
On Fri, Jan 13, 2023 at 1:25 AM Eugenio Pérez wrote:
>
> This is needed for qemu to know it can suspend the device to retrieve
> its status and enable SVQ with it, so all the process is transparent to
> the guest.
>
> Signed-off-by: Eugenio Pérez
Acked-by: Jason Wang
We probably need to add
On Fri, Jan 13, 2023 at 1:25 AM Eugenio Pérez wrote:
>
> To restore the device at the destination of a live migration we send the
> commands through control virtqueue. For a device to read CVQ it must
> have received the DRIVER_OK status bit.
This probably requires the support from the parent
在 2023/1/13 01:24, Eugenio Pérez 写道:
A vdpa net device must initialize with SVQ in order to be migratable,
and initialization code verifies conditions. If the device is not
initialized with the x-svq parameter, it will not expose _F_LOG so vhost
sybsystem will block VM migration from its
On 1/12/23 6:03 PM, Michael S. Tsirkin wrote:
> On Thu, Jan 12, 2023 at 10:55:25PM +, Bernhard Beschow wrote:
>> I think the change Michael suggests is very minimalistic: Move the if
>> condition around xen_igd_reserve_slot() into the function itself and
>> always call it there unconditionally
On Fri, Jan 13, 2023 at 1:24 AM Eugenio Pérez wrote:
>
> At this moment it is only possible to migrate to a vdpa device running
> with x-svq=on. As a protective measure, the rewind of the inflight
> descriptors was done at the destination. That way if the source sent a
> virtqueue with inuse
On Fri, Jan 13, 2023 at 1:24 AM Eugenio Pérez wrote:
>
> Only create iova_tree if and when it is needed.
>
> The cleanup keeps being responsability of last VQ but this change allows
> to merge both cleanup functions.
>
> Signed-off-by: Eugenio Pérez
> ---
> net/vhost-vdpa.c | 101
On Fri, Jan 13, 2023 at 11:25 AM Zhu, Lingshan wrote:
>
>
>
> On 1/13/2023 10:31 AM, Jason Wang wrote:
> > On Fri, Jan 13, 2023 at 1:27 AM Eugenio Pérez wrote:
> >> Spuriously kick the destination device's queue so it knows in case there
> >> are new descriptors.
> >>
> >> RFC: This is somehow a
On Fri, Jan 13, 2023 at 1:24 AM Eugenio Pérez wrote:
>
> VHOST_BACKEND_F_IOTLB_ASID is the feature bit, not the bitmask. Since
> the device under test also provided VHOST_BACKEND_F_IOTLB_MSG_V2 and
> VHOST_BACKEND_F_IOTLB_BATCH, this went unnoticed.
>
> Fixes: c1a1008685 ("vdpa: always start CVQ
On Fri, Jan 13, 2023 at 6:37 AM Daniel Henrique Barboza
wrote:
>
> The microchip_icicle_kit, sifive_u, spike and virt boards are now doing
> the same steps when '-kernel' is used:
>
> - execute load_kernel()
> - load init_rd()
> - write kernel_cmdline
>
> Let's fold everything inside
On Fri, Jan 13, 2023 at 1:27 AM Eugenio Pérez wrote:
>
> Spuriously kick the destination device's queue so it knows in case there
> are new descriptors.
>
> RFC: This is somehow a gray area. The guest may have placed descriptors
> in a virtqueue but not kicked it, so it might be surprised if the
For passthrough decoder (a decoder hosted by a cxl component with only
one downstream port), its cache_mem_registers field COMMITTED
(see spec 2.0 8.2.5.12 - CXL HDM Decoder Capability Structure) will not
be set by the current Linux CXL driver. Without the fix, for a cxl
topology setup with a
Hi Daniel,
On Wed, Jan 11, 2023 at 1:03 PM Alistair Francis wrote:
>
> On Wed, Jan 11, 2023 at 6:17 AM Daniel Henrique Barboza
> wrote:
> >
> > Hi,
> >
> > I found this bug when testing my avocado changes in riscv-to-apply.next.
> > The sifive_u board, both 32 and 64 bits, stopped booting
On Thu, Jan 12, 2023 at 1:27 AM Daniel Henrique Barboza
wrote:
>
> Follow the QEMU convention of naming MachineState pointers as 'ms' by
> renaming the instance in create_fdt() where we're calling it 'mc'.
>
> Cc: Paul Burton
> Cc: Aleksandar Rikalo
> Suggested-by: Philippe Mathieu-Daudé
>
On 1/11/23 08:31, Philippe Mathieu-Daudé wrote:
Implement Richard's suggestion to use __builtin_bswap().
Convert to __builtin_bswap() one patch per OS to simplify
maintainers review.
Since v2:
- Rebased adapting ./configure changes to meson
Since v1:
- Remove the Haiku/BSD ifdef'ry (Peter)
-
On 1/12/23 11:27, Keith Busch wrote:
On Thu, Jan 12, 2023 at 06:45:55PM +0100, Klaus Jensen wrote:
On Jan 12 09:34, Keith Busch wrote:
On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
am wondering if there
On Thu, 12 Jan 2023, Philippe Mathieu-Daudé wrote:
On 12/1/23 10:57, Akihiko Odaki wrote:
The definitions for E1000_VFTA_ENTRY_SHIFT, E1000_VFTA_ENTRY_MASK, and
E1000_VFTA_ENTRY_BIT_SHIFT_MASK were copied from:
On 1/12/23 09:45, Klaus Jensen wrote:
On Jan 12 09:34, Keith Busch wrote:
On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
am wondering if there is something going on with the kernel driver (but
I certainly
On Thu, 12 Jan 2023, Howard Spoelstra wrote:
On Wed, Jan 11, 2023 at 1:15 AM BALATON Zoltan wrote:
On Tue, 10 Jan 2023, Mark Cave-Ayland wrote:
On 04/01/2023 21:59, BALATON Zoltan wrote:
Setting emulated machine type with a property called "via" is
confusing users so deprecate the "via"
On Thu, Jan 12, 2023 at 10:55:25PM +, Bernhard Beschow wrote:
> I think the change Michael suggests is very minimalistic: Move the if
> condition around xen_igd_reserve_slot() into the function itself and
> always call it there unconditionally -- basically turning three lines
> into one. Since
Am 12. Januar 2023 20:11:54 UTC schrieb Chuck Zmudzinski :
>On 1/12/23 2:18 PM, Bernhard Beschow wrote:
>>
>>
>> Am 11. Januar 2023 15:40:24 UTC schrieb Chuck Zmudzinski :
>>>On 1/10/23 3:16 AM, Michael S. Tsirkin wrote:
On Tue, Jan 10, 2023 at 02:08:34AM -0500, Chuck Zmudzinski wrote:
On Thu, Jan 12, 2023 at 05:21:30PM +, Jonathan Cameron wrote:
> On Thu, 12 Jan 2023 10:39:17 -0500
> Gregory Price wrote:
>
> > On Wed, Jan 11, 2023 at 02:24:32PM +, Jonathan Cameron via wrote:
> > > Gregory's patches were posted as part of his work on adding volatile
> > > support.
> >
On 1/12/23 07:28, Alessandro Di Federico wrote:
On Tue, 10 Jan 2023 11:56:50 -0800
Richard Henderson wrote:
However, at some point we do want to keep some target addresses in
the proper size. For instance within the softmmu tlb, where
CPUTLBEntry is either 16 or 32 bytes, depending.
So
The microchip_icicle_kit, sifive_u, spike and virt boards are now doing
the same steps when '-kernel' is used:
- execute load_kernel()
- load init_rd()
- write kernel_cmdline
Let's fold everything inside riscv_load_kernel() to avoid code
repetition. To not change the behavior of boards that
The only remaining caller is riscv_load_kernel_and_initrd() which
belongs to the same file.
Signed-off-by: Daniel Henrique Barboza
Reviewed-by: Philippe Mathieu-Daudé
Reviewed-by: Bin Meng
---
hw/riscv/boot.c | 80 -
include/hw/riscv/boot.h | 1
Hi,
These are the 2 remaining patches, patches 10 and 11, of:
"[PATCH v5 00/11] riscv: OpenSBI boot test and cleanups"
The first 9 patches are already available in riscv-to-apply.next.
The only change made was in patch 10 where we're now handling the case
where load_elf_ram_sym is padding the
On Thu, Jan 12, 2023 at 05:14:57PM -0500, Peter Xu wrote:
> On Thu, Jan 12, 2023 at 07:52:41PM +, Dr. David Alan Gilbert wrote:
> > * David Hildenbrand (da...@redhat.com) wrote:
> > > On 12.01.23 18:56, Dr. David Alan Gilbert wrote:
> > > > * David Hildenbrand (da...@redhat.com) wrote:
> > > >
On Thu, Jan 12, 2023 at 07:52:41PM +, Dr. David Alan Gilbert wrote:
> * David Hildenbrand (da...@redhat.com) wrote:
> > On 12.01.23 18:56, Dr. David Alan Gilbert wrote:
> > > * David Hildenbrand (da...@redhat.com) wrote:
> > > > For virtio-mem, we want to have the plugged/unplugged state of
On Thu, Jan 12, 2023 at 06:40:00PM +, Dr. David Alan Gilbert wrote:
> * David Hildenbrand (da...@redhat.com) wrote:
> > On 12.01.23 18:43, Dr. David Alan Gilbert wrote:
> > > * David Hildenbrand (da...@redhat.com) wrote:
> > > > ... and store it in the migration state. This is a preparation
On Thu, Jan 12, 2023 at 04:56:43PM +0100, Philippe Mathieu-Daudé wrote:
> When flex is not available, binutils sources default to the
> 'missing' script, but the current script available is not in
> the format expected by the 'configure' script:
>
> $ ./configure
> ...
>
On Thu, Jan 12, 2023 at 02:24:02PM +, Anton Kochkov wrote:
> Signed-off-by: Eitan Eliahu
> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/652
> ---
> target/tricore/translate.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/target/tricore/translate.c
On Thu, Dec 9, 2021 at 10:34 AM Or Ozeri wrote:
>
> The blk_get_max_hw_transfer API was recently added in 6.1.0.
> It allows querying an underlying block device its max transfer capability.
> This commit changes virtio-blk to use this.
>
> Signed-off-by: Or Ozeri
> ---
> hw/block/virtio-blk.c |
On Wed, Jan 11, 2023 at 1:15 AM BALATON Zoltan wrote:
> On Tue, 10 Jan 2023, Mark Cave-Ayland wrote:
> > On 04/01/2023 21:59, BALATON Zoltan wrote:
> >
> >> Setting emulated machine type with a property called "via" is
> >> confusing users so deprecate the "via" option in favour of newly added
>
On 1/12/23 2:18 PM, Bernhard Beschow wrote:
>
>
> Am 11. Januar 2023 15:40:24 UTC schrieb Chuck Zmudzinski :
>>On 1/10/23 3:16 AM, Michael S. Tsirkin wrote:
>>> On Tue, Jan 10, 2023 at 02:08:34AM -0500, Chuck Zmudzinski wrote:
Intel specifies that the Intel IGD must occupy slot 2 on the PCI
* David Hildenbrand (da...@redhat.com) wrote:
> On 12.01.23 18:56, Dr. David Alan Gilbert wrote:
> > * David Hildenbrand (da...@redhat.com) wrote:
> > > For virtio-mem, we want to have the plugged/unplugged state of memory
> > > blocks available before migrating any actual RAM content, and perform
* David Hildenbrand (da...@redhat.com) wrote:
> Ordinary memory preallocation runs when QEMU starts up and creates the
> memory backends, before processing the incoming migration stream. With
> virtio-mem, we don't know which memory blocks to preallocate before
> migration started. Now that we
* David Hildenbrand (da...@redhat.com) wrote:
> The bitmap and the size are immutable while migration is active: see
> virtio_mem_is_busy(). We can migrate this information early, before
> migrating any actual RAM content. Further, all information we need for
> sanity checks is immutable as well.
On Thu, Jan 12, 2023 at 06:45:55PM +0100, Klaus Jensen wrote:
> On Jan 12 09:34, Keith Busch wrote:
> > On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
> > >
> > > The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
> > > am wondering if there is something going
Am 11. Januar 2023 15:40:24 UTC schrieb Chuck Zmudzinski :
>On 1/10/23 3:16 AM, Michael S. Tsirkin wrote:
>> On Tue, Jan 10, 2023 at 02:08:34AM -0500, Chuck Zmudzinski wrote:
>>> Intel specifies that the Intel IGD must occupy slot 2 on the PCI bus,
>>> as noted in docs/igd-assign.txt in the
blk_unref() can't report any errors that happen while closing the image.
For example, if qcow2 hits an -ENOSPC error while writing out dirty
bitmaps when it's closed, it prints error messages to stderr, but
'qemu-img bitmap' won't see any error return value and will therefore
look successful with
This series addresses the problem described in these bug reports:
https://gitlab.com/qemu-project/qemu/-/issues/1330
https://bugzilla.redhat.com/show_bug.cgi?id=2147617
qcow2 can fail when writing back dirty bitmaps in qcow2_inactivate().
However, when the function is called through blk_unref(),
This tests that when an error happens while writing back bitmaps to the
image file in qcow2_inactivate(), 'qemu-img bitmap/commit' actually
return an error value in their exit code instead of making the operation
look successful to scripts.
Signed-off-by: Kevin Wolf
---
In order to write the bitmap table to the image file, it is converted to
big endian. If the write fails, it is passed to clear_bitmap_table() to
free all of the clusters it had allocated before. However, if we don't
convert it back to native endianness first, we'll free things at a wrong
offset.
On 1/12/23 09:45, Klaus Jensen wrote:
On Jan 12 09:34, Keith Busch wrote:
On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
am wondering if there is something going on with the kernel driver (but
I certainly
blk_unref() can't report any errors that happen while closing the image.
For example, if qcow2 hits an -ENOSPC error while writing out dirty
bitmaps when it's closed, it prints error messages to stderr, but
'qemu-img commit' won't see any error return value and will therefore
look successful with
* David Hildenbrand (da...@redhat.com) wrote:
> On 12.01.23 18:43, Dr. David Alan Gilbert wrote:
> > * David Hildenbrand (da...@redhat.com) wrote:
> > > ... and store it in the migration state. This is a preparation for
> > > storing selected vmds's already in qemu_savevm_state_setup().
> > >
> >
* David Hildenbrand (da...@redhat.com) wrote:
> On 12.01.23 17:58, Dr. David Alan Gilbert wrote:
> > * David Hildenbrand (da...@redhat.com) wrote:
> > > Let's move more code into vmstate_save(), reducing code duplication and
> > > preparing for reuse of vmstate_save() in qemu_savevm_state_setup().
* David Hildenbrand (da...@redhat.com) wrote:
> "prealloc=on" for the memory backend does not work as expected, as
> virtio-mem will simply discard all preallocated memory immediately again.
> In the best case, it's an expensive NOP. In the worst case, it's an
> unexpected allocation error.
>
>
Am 12. Januar 2023 13:32:02 UTC schrieb "Philippe Mathieu-Daudé"
:
>On 12/1/23 13:50, Philippe Mathieu-Daudé wrote:
>> On 9/1/23 18:23, Bernhard Beschow wrote:
>>> Now that the PIIX3 and PIIX4 device models are sufficiently consolidated,
>>> their implementations can be merged into one file
* David Hildenbrand (da...@redhat.com) wrote:
> Let's factor out this check, to be used in virtio-mem context next.
>
> While at it, fix a spelling error in a related comment.
>
> Signed-off-by: David Hildenbrand
Reviewed-by: Dr. David Alan Gilbert
> ---
> include/migration/misc.h | 4 +++-
Am 12. Januar 2023 16:36:30 UTC schrieb "Philippe Mathieu-Daudé"
:
>On 9/1/23 18:23, Bernhard Beschow wrote:
>> Now that the PIIX3 and PIIX4 device models are sufficiently consolidated,
>> their implementations can be merged into one file for further
>> consolidation.
>>
>> Signed-off-by:
On 12.01.23 18:56, Dr. David Alan Gilbert wrote:
* David Hildenbrand (da...@redhat.com) wrote:
For virtio-mem, we want to have the plugged/unplugged state of memory
blocks available before migrating any actual RAM content, and perform
sanity checks before touching anything on the destination.
Am 12. Januar 2023 16:31:23 UTC schrieb "Philippe Mathieu-Daudé"
:
>On 12/1/23 16:04, Philippe Mathieu-Daudé wrote:
>> On 9/1/23 18:23, Bernhard Beschow wrote:
>>> Now that the PIIX3 and PIIX4 device models are sufficiently consolidated,
>>> their implementations can be merged into one file
On Thu, Jan 12, 2023 at 05:44:31PM +, Daniel P. Berrangé wrote:
> On Thu, Jan 12, 2023 at 12:37:46PM -0500, Michael S. Tsirkin wrote:
> > On Thu, Jan 12, 2023 at 03:47:19PM +0100, Markus Armbruster wrote:
> > > "Michael S. Tsirkin" writes:
> > >
> > > > On Thu, Jan 12, 2023 at 08:51:32AM
* David Hildenbrand (da...@redhat.com) wrote:
> For virtio-mem, we want to have the plugged/unplugged state of memory
> blocks available before migrating any actual RAM content, and perform
> sanity checks before touching anything on the destination. This
> information is immutable on the
On 12.01.23 17:58, Dr. David Alan Gilbert wrote:
* David Hildenbrand (da...@redhat.com) wrote:
Let's move more code into vmstate_save(), reducing code duplication and
preparing for reuse of vmstate_save() in qemu_savevm_state_setup(). We
have to move vmstate_save() to make the compiler happy.
On Thu, Jan 12, 2023 at 03:58:56PM +, Peter Maydell wrote:
> On Thu, 12 Jan 2023 at 15:14, Daniel P. Berrangé wrote:
> >
> > On Thu, Jan 12, 2023 at 08:51:26AM -0500, Michael S. Tsirkin wrote:
> > > On Thu, Jan 12, 2023 at 12:50:05PM +0100, Markus Armbruster wrote:
> > > >
On Thu, Jan 12, 2023 at 06:45:55PM +0100, Klaus Jensen wrote:
> On Jan 12 09:34, Keith Busch wrote:
> > On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
> > >
> > > The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
> > > am wondering if there is something going
On 12.01.23 18:43, Dr. David Alan Gilbert wrote:
* David Hildenbrand (da...@redhat.com) wrote:
... and store it in the migration state. This is a preparation for
storing selected vmds's already in qemu_savevm_state_setup().
Signed-off-by: David Hildenbrand
---
migration/migration.c | 4
On Thu, Jan 12, 2023 at 12:37:46PM -0500, Michael S. Tsirkin wrote:
> On Thu, Jan 12, 2023 at 03:47:19PM +0100, Markus Armbruster wrote:
> > "Michael S. Tsirkin" writes:
> >
> > > On Thu, Jan 12, 2023 at 08:51:32AM -0500, Michael S. Tsirkin wrote:
> > >> On Thu, Jan 12, 2023 at 12:50:05PM +0100,
On Jan 12 09:34, Keith Busch wrote:
> On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
> >
> > The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
> > am wondering if there is something going on with the kernel driver (but
> > I certainly do not rule out that
* David Hildenbrand (da...@redhat.com) wrote:
> ... and store it in the migration state. This is a preparation for
> storing selected vmds's already in qemu_savevm_state_setup().
>
> Signed-off-by: David Hildenbrand
> ---
> migration/migration.c | 4
> migration/migration.h | 4
>
* David Hildenbrand (da...@redhat.com) wrote:
> Let's move more code into vmstate_save(), reducing code duplication and
> preparing for reuse of vmstate_save() in qemu_savevm_state_setup(). We
> have to move vmstate_save() to make the compiler happy.
>
> We'll now also trace from
On Thu, Jan 12, 2023 at 12:50:05PM +0100, Markus Armbruster wrote:
> docs/devel/style.rst mandates:
>
> The "qemu/osdep.h" header contains preprocessor macros that affect
> the behavior of core system headers like . It must be
> the first include so that core system headers included
It's possible to migrate vdpa net devices if they are shadowed from the
start. But to always shadow the dataplane is effectively break its host
passthrough, so its not convenient in vDPA scenarios.
This series enables dynamically switching to shadow mode only at
migration time. This allow full
On Thu, Jan 12, 2023 at 03:47:19PM +0100, Markus Armbruster wrote:
> "Michael S. Tsirkin" writes:
>
> > On Thu, Jan 12, 2023 at 08:51:32AM -0500, Michael S. Tsirkin wrote:
> >> On Thu, Jan 12, 2023 at 12:50:05PM +0100, Markus Armbruster wrote:
> >> > docs/devel/style.rst mandates:
> >> >
> >> >
Spuriously kick the destination device's queue so it knows in case there
are new descriptors.
RFC: This is somehow a gray area. The guest may have placed descriptors
in a virtqueue but not kicked it, so it might be surprised if the device
starts processing it.
However, that information is not in
Since some actions move to the start function instead of init, the
device features may not be the parent vdpa device's, but the one
returned by vhost backend. If transition to SVQ is supported, the vhost
backend will return _F_LOG_ALL to signal the device is migratable.
Add VHOST_F_LOG_ALL. HW
To restore the device at the destination of a live migration we send the
commands through control virtqueue. For a device to read CVQ it must
have received the DRIVER_OK status bit.
However this opens a window where the device could start receiving
packets in rx queue 0 before it receives the RSS
A vdpa net device must initialize with SVQ in order to be migratable,
and initialization code verifies conditions. If the device is not
initialized with the x-svq parameter, it will not expose _F_LOG so vhost
sybsystem will block VM migration from its initialization.
Next patches change this.
At this moment it is only possible to migrate to a vdpa device running
with x-svq=on. As a protective measure, the rewind of the inflight
descriptors was done at the destination. That way if the source sent a
virtqueue with inuse descriptors they are always discarded.
Since this series allows to
On Thu, Jan 12, 2023 at 06:27:04PM +0100, Nina Schoetterl-Glausch wrote:
> On Thu, 2023-01-12 at 12:10 +, Daniel P. Berrangé wrote
>
> [...]
> >
> > We already have 'query-cpus-fast' wich returns one entry for
> > each CPU. In fact why do we need to add query-topology at all.
> > Can't we
On Thu, 12 Jan 2023 12:50:05 +0100
Markus Armbruster wrote:
> docs/devel/style.rst mandates:
>
> The "qemu/osdep.h" header contains preprocessor macros that affect
> the behavior of core system headers like . It must be
> the first include so that core system headers included by
This is needed for qemu to know it can suspend the device to retrieve
its status and enable SVQ with it, so all the process is transparent to
the guest.
Signed-off-by: Eugenio Pérez
---
hw/virtio/vhost-vdpa.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
This way the device's vhost_vdpa can make the choice about exposing or
not the _F_LOG feature.
At the moment it is always false.
Signed-off-by: Eugenio Pérez
---
include/hw/virtio/vhost-vdpa.h | 1 +
hw/virtio/vhost-vdpa.c | 2 +-
2 files changed, 2 insertions(+), 1 deletion(-)
diff
On Thu, 2023-01-12 at 12:10 +, Daniel P. Berrangé wrote
[...]
>
> We already have 'query-cpus-fast' wich returns one entry for
> each CPU. In fact why do we need to add query-topology at all.
> Can't we just add book-id / drawer-id / polarity / dedicated
> to the query-cpus-fast result ?
Is
This allows net to restart the device backend to configure SVQ on it.
Ideally, these changes should not be net specific. However, the vdpa net
backend is the one with enough knowledge to configure everything because
of some reasons:
* Queues might need to be shadowed or not depending on its kind
VHOST_BACKEND_F_IOTLB_ASID is the feature bit, not the bitmask. Since
the device under test also provided VHOST_BACKEND_F_IOTLB_MSG_V2 and
VHOST_BACKEND_F_IOTLB_BATCH, this went unnoticed.
Fixes: c1a1008685 ("vdpa: always start CVQ in SVQ mode if possible")
Signed-off-by: Eugenio Pérez
---
Vhost-vdpa networking devices need to meet a few conditions to be
migratable. If SVQ is not enabled from the beginning, to suspend the
device to retrieve the vq state is the first requirement.
Expose _F_LOG only in that case.
Signed-off-by: Eugenio Pérez
---
net/vhost-vdpa.c | 22
To restore the device at the destination of a live migration we send the
commands through control virtqueue. For a device to read CVQ it must
have received the DRIVER_OK status bit.
However this opens a window where the device could start receiving
packets in rx queue 0 before it receives the RSS
This allows to passthrough or shadow the data depending on migration
state in next patches.
Signed-off-by: Eugenio Pérez
---
net/vhost-vdpa.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/net/vhost-vdpa.c b/net/vhost-vdpa.c
index 75cca497c8..631424d9c4 100644
---
Only create iova_tree if and when it is needed.
The cleanup keeps being responsability of last VQ but this change allows
to merge both cleanup functions.
Signed-off-by: Eugenio Pérez
---
net/vhost-vdpa.c | 101 +--
1 file changed, 71 insertions(+),
On Thu, 12 Jan 2023 10:39:17 -0500
Gregory Price wrote:
> On Wed, Jan 11, 2023 at 02:24:32PM +, Jonathan Cameron via wrote:
> > Gregory's patches were posted as part of his work on adding volatile
> > support.
> >
On Mon, 9 Jan 2023 at 11:53, Philippe Mathieu-Daudé wrote:
>
> Various trivial cleanups:
> - Remove unnecessary sysmem argument from pxa2xx_init()
> - Use IEC binary prefix definitions
> - Remove unreachable code calling pflash_cfi01_register()
> (another use in PPC sam460ex will be removed
1 - 100 of 357 matches
Mail list logo