> Triggered by the framebuffer endian issues we have with stdvga I've
> started to check where we stand with virtio-gpu and whenever we have to
> fix something in the virtio protocol before setting in stone with the
> upstream merge.
Let me start by saying its not that I don't care about endiannes
On 09/29/14 06:25, Stefano Stabellini wrote:
On Mon, 29 Sep 2014, Stefano Stabellini wrote:
On Fri, 26 Sep 2014, Don Slutz wrote:
This adds synchronisation of the vcpu registers
between Xen and QEMU.
Signed-off-by: Don Slutz
[...]
diff --git a/xen-hvm.c b/xen-hvm.c
index 05e522c..e1274bb 1
On Mon, Sep 29, 2014 at 11:31:04AM +0200, Gerd Hoffmann wrote:
> From: Benjamin Herrenschmidt
>
> Provide different functions for converting from an LE vs a BE
> framebuffer. We cannot rely on the simple cases always being
> shared surfaces since cirrus will need to always shadow for
> cursor emu
On Mon, Sep 29, 2014 at 11:31:11AM +0200, Gerd Hoffmann wrote:
> From: Benjamin Herrenschmidt
>
> Include the endian state in the migration stream as an optional
> subsection which we only include when the endian isn't the default,
> thus enabling backward compatibility of the common case.
>
> S
> > > >
> > > > Hm...
> > > >
> > > > void virtio_instance_init_common(Object *proxydev,
> > > > DeviceState *vdev,
> > > > size_t vdevsize,
> > > > const char *vdevname)
> > > > {
> > > > object_
> Subject: Re: [Qemu-devel] [PATCH v3 0/7] add description field in
> ObjectProperty and PropertyInfo struct
>
> Hi,
>
> Am 27.09.2014 um 07:13 schrieb arei.gong...@huawei.com:
> > Gonglei (7):
> > qom: add error handler for object_property_print()
> > qom: add error handler for object alias
On 09/29/14 07:53, Alexander Graf wrote:
On 29.09.14 13:10, Paolo Bonzini wrote:
Il 29/09/2014 10:12, Alexander Graf ha scritto:
Could you instead plug into the existing cpu_synchronize_registers()
framework and just implement register synchronization for the Xen side,
just like it's been done
On 09/29/14 09:14, Paolo Bonzini wrote:
Il 29/09/2014 14:57, Alexander Graf ha scritto:
Yeah, that would be possible. You do not even need synchronize_state,
it seems to me that it introduces more complication for little gain.
Well, it makes all accels behave the same and keep information alw
QEMU currently allows the number of VCPUs to not be a multiple of the
number of threads per socket, but the smbios socket count calculation
introduced by commit c97294ec1b9e36887e119589d456557d72ab37b5 doesn't
take that into account, triggering an assertion. e.g.:
$ ./x86_64-softmmu/qemu-system-
Hi,
> Subject: Re: [PATCH v3 4/7] qom: add description field in ObjectProperty
> struct
>
> Il 27/09/2014 07:13, arei.gong...@huawei.com ha scritto:
> > +void object_property_set_description(Object *obj, const char *name,
> > + const char *description, Error
>
>
> On 09/26/2014 11:14 PM, arei.gong...@huawei.com wrote:
> > From: Gonglei
> >
> > Add a new "description" field to DevicePropertyInfo.
> > The descriptions can serve as documentation in the code,
> > and they can be used to provide better help. For example:
> >
> > $./qemu-system-x86_64 -devic
Signed-off-by: Jan Vesely
---
v2: add usb_mouse_properties
use macros for bmAttributes
v3: rebase
v4: add compat property
rebase
hw/i386/pc_piix.c | 4
hw/usb/dev-hid.c | 71 ++-
2 files changed, 74 insertions(+), 1 deletion(-)
Signed-off-by: Jan Vesely
---
v2: rebase
v3: add compat property
rebase
hw/i386/pc_piix.c | 4
hw/usb/dev-hid.c | 66 ++-
2 files changed, 69 insertions(+), 1 deletion(-)
diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
index 6a17
Signed-off-by: Jan Vesely
---
v2: rebase
v3: fixup error handling
hw/usb/dev-hid.c | 41 +
1 file changed, 21 insertions(+), 20 deletions(-)
diff --git a/hw/usb/dev-hid.c b/hw/usb/dev-hid.c
index 467ec86..5d37d63 100644
--- a/hw/usb/dev-hid.c
+++ b/hw/usb
On Mon, 09/29 12:56, Paolo Bonzini wrote:
> Il 29/09/2014 12:11, Paolo Bonzini ha scritto:
> > Il 28/09/2014 03:48, Fam Zheng ha scritto:
> >> +virtio_scsi_complete_req(req);
> >> +} else {
> >> +assert(r = -EINPROGRESS);
> >> +}
> >> }
> >
> > = instead of == here.
> >
>
On 2014/9/29 18:01, Michael S. Tsirkin wrote:
On Sun, Sep 28, 2014 at 10:59:05AM +0800, Chen, Tiejun wrote:
On 2014/9/3 9:40, Kay, Allen M wrote:
-Original Message-
From: Chen, Tiejun
Sent: Monday, September 01, 2014 12:50 AM
To: Michael S. Tsirkin
Cc: xen-de...@lists.xensource.com;
On Mon, 09/29 12:56, Paolo Bonzini wrote:
> @@ -241,7 +235,6 @@ static int virtio_scsi_do_tmf(VirtIOSCSI *s,
> VirtIOSCSIReq *req)
> BusChild *kid;
> int target;
> int ret = 0;
> -int cancel_count;
>
> if (s->dataplane_started && bdrv_get_aio_context(d->conf.bs) != s->ctx
On Mon, 09/29 11:28, Stefan Hajnoczi wrote:
> On Mon, Sep 29, 2014 at 01:26:29PM +0800, Fam Zheng wrote:
> > +int qemu_epoll(GPollFD *fds, guint nfds, int64_t timeout)
> > +{
> > +/* A copy of last fd array, used to skip epoll_prepare when nothing
> > + * changed. */
> > +static GPollFD
From: Gonglei
Add a new "description" field to DevicePropertyInfo.
The descriptions can serve as documentation in the code,
and they can be used to provide better help. For example:
$./qemu-system-x86_64 -device virtio-blk-pci,?
Before this patch:
virtio-blk-pci.iothread=link
virtio-blk-pci.x-
From: Gonglei
This patch series based on qom-next tree:
https://github.com/afaerber/qemu-cpu/commits/qom-next
Add a description field in both ObjectProperty and PropertyInfo struct.
The descriptions can serve as documentation in the code,
and they can be used to provide better help. For example
From: Gonglei
The descriptions can serve as documentation in the code,
and they can be used to provide better help.
Cc: Paolo Bonzini
Cc: Michael S. Tsirkin
Cc: Markus Armbruster
Signed-off-by: Gonglei
Signed-off-by: Paolo Bonzini
---
hw/core/qdev-properties-system.c | 4
hw/core/qdev
From: Gonglei
The legacy_name is useless now, the better help
information provied by description field of property.
Cc: Paolo Bonzini
Cc: Michael S. Tsirkin
Cc: Markus Armbruster
Signed-off-by: Gonglei
---
hw/core/qdev-properties-system.c | 4
hw/core/qdev-properties.c| 6 -
From: Gonglei
Set all static qdev properties' descriptions to object property's
description.
Cc: Paolo Bonzini
Cc: Michael S. Tsirkin
Cc: Markus Armbruster
Signed-off-by: Gonglei
---
hw/core/qdev.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/hw/core/qdev.c b/hw/core/qdev.c
inde
From: Gonglei
The descriptions can serve as documentation in the code,
and they can be used to provide better help.
When we call object_property_add_alias() adding alias properties to
the source object on the target Object, set the object property's
description to the source object property's.
v2: Introduce patch 1 to allow qemu_epoll to access AioContext.
Don't pollute g_* namespace. (Stefan)
Fam Zheng (2):
main-loop: Pass AioContext into qemu_poll_ns
main-loop: Use epoll on Linux
Makefile.objs| 1 +
aio-posix.c | 2 +-
include/block/aio.h |
A new implementation for qemu_poll_ns based on epoll is introduced here
to address the slowness of g_poll and ppoll when the number of fds are
high.
On my laptop this would reduce the virtio-blk on top of null-aio
device's response time from 32 us to 29 us with few fds (~10), and 48 us
to 32 us wi
qemu_poll_ns may ultilize the information in AioContext to achieve
better performance.
Signed-off-by: Fam Zheng
---
aio-posix.c | 2 +-
include/qemu/timer.h | 3 ++-
main-loop.c | 6 --
qemu-timer.c | 2 +-
4 files changed, 8 insertions(+), 5 deletions(-)
diff --gi
v5: Squash in Paolo's suggested patch into patch 7, fixed "notifier->tmf_req"
and did some test.
v4: Call notifier_list_notify in scsi_req_complete. (Paolo)
v3: Address Paolo's comments:
- scsi_req_canceled -> scsi_req_cancel_complete
- Drop unnecessary changes in scsi_req_alloc and s
Before, scsi_req_cancel will take ownership of the canceled request and unref
it. We did this because we didn't know whether AIO CB will be called or not
during the cancelling, so we set the io_canceled flag before calling it, and
skip unref in the potentially called callbacks, which is not very ni
Now that we always called the cb in bdrv_aio_cancel, let's make scsi-generic
callbacks check io_canceled flag similarly to scsi-disk.
Signed-off-by: Fam Zheng
---
hw/scsi/scsi-generic.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/hw/scsi/scsi-generic.c b/hw/scs
The only user of this function is spapr_vscsi.c. We can convert to
scsi_req_cancel plus adding a check in vscsi_request_cancelled.
Suggested-by: Paolo Bonzini
Signed-off-by: Fam Zheng
---
hw/scsi/scsi-bus.c| 15 ---
hw/scsi/spapr_vscsi.c | 11 ---
2 files changed, 8 inse
The only two implementations are identical to each other, with nothing specific
to device: they only call bdrv_aio_cancel with the SCSIRequest.aiocb.
Let's move it to scsi-bus.
Signed-off-by: Fam Zheng
---
hw/scsi/scsi-bus.c | 4 ++--
hw/scsi/scsi-disk.c| 14 --
hw/scsi/scs
Let the aio cb do the clean up and notification job after scsi_req_cancel, in
preparation for asynchronous cancellation.
Signed-off-by: Fam Zheng
---
hw/scsi/scsi-bus.c | 14 ++
hw/scsi/scsi-disk.c| 8
hw/scsi/scsi-generic.c | 1 +
include/hw/scsi/scsi.h | 1 +
4
Devices will call this function to start an asynchronous cancellation. The
bus->info->cancel will be called after the request is canceled.
Devices will probably need to track a separate TMF request that triggers this
cancellation, and wait until the cancellation is done before completing it. So
we
For VIRTIO_SCSI_T_TMF_ABORT_TASK and VIRTIO_SCSI_T_TMF_ABORT_TASK_SET,
use scsi_req_cancel_async to start the cancellation.
Because each tmf command may cancel multiple requests, we need to use a
counter to track the number of remaining requests we still need to wait
for.
Signed-off-by: Fam Zheng
openstack review at:
https://review.openstack.org/#/c/123957/
Qemu patches at:
http://patchwork.ozlabs.org/patch/393494/ ; and
http://patchwork.ozlabs.org/patch/393495/
** Changed in: nova
Assignee: (unassigned) => Tony Breeds (o-tony)
** Changed in: qemu
Status: New => In Prog
Greg Kurz writes:
> On Mon, 29 Sep 2014 19:15:05 +0300
> "Michael S. Tsirkin" wrote:
>
> It may be possible to change the way SLOF configures the virtq so
> that it wouldn't suffer from the device being reset. We have some
> time left before 2.2. I have a tentative patch for SLOF that I'll
> be g
From: Gonglei
object_initialize() leaves the object with a refcount of 1.
object_property_add_child() adds its own reference which is dropped
again when the property is deleted.
The upshot of this is that we always have a refcount >= 1. Upon hot
unplug the virtio-serial child is not finalized!
From: Gonglei
virtio-net-pci, virtio-net-s390, and virtio-net-ccw all duplicate the
qdev properties of their VirtIONet child. This approach does not work
well with string or pointer properties since we must be careful about
leaking or double-freeing them.
Use the QOM alias property to forward pr
From: Gonglei
{virtio, vhost}-scsi-{pci, s390, ccw} all duplicate the
qdev properties of their VirtIOSCSI/VHostSCSI child.
This approach does not work well with string or pointer
properties since we must be careful about leaking or
double-freeing them.
Use the QOM alias property to forward prope
From: Gonglei
object_initialize() leaves the object with a refcount of 1.
object_property_add_child() adds its own reference which is dropped
again when the property is deleted.
The upshot of this is that we always have a refcount >= 1. Upon hot
unplug the virtio-balloon child is not finalized!
From: Gonglei
object_initialize() leaves the object with a refcount of 1.
object_property_add_child() adds its own reference which is dropped
again when the property is deleted.
The upshot of this is that we always have a refcount >= 1. Upon hot
unplug the virtio-rng child is not finalized!
Dr
From: Gonglei
For better code sharing, add a wrapper help funciton
for various virtio devices.
Cc: Cornelia Huck
Cc: Michael S. Tsirkin
Signed-off-by: Gonglei
---
hw/s390x/s390-virtio-bus.c | 42 +--
hw/s390x/virtio-ccw.c | 42 +---
From: Gonglei
object_initialize() leaves the object with a refcount of 1.
object_property_add_child() adds its own reference which is dropped
again when the property is deleted.
The upshot of this is that we always have a refcount >= 1. Upon hot
unplug the virtio-scsi/vhost-scsi child is not fi
From: Gonglei
virtio-rng-{pci, s390, ccw} all duplicate the
qdev properties of their VirtIORNG child.
This approach does not work well with string or pointer
properties since we must be careful about leaking or
double-freeing them.
Use the QOM alias property to forward property accesses to the
V
From: Gonglei
object_initialize() leaves the object with a refcount of 1.
object_property_add_child() adds its own reference which is dropped
again when the property is deleted.
The upshot of this is that we always have a refcount >= 1. Upon hot
unplug the virtio-net child is not finalized!
Dr
From: Gonglei
object_initialize() leaves the object with a refcount of 1.
object_property_add_child() adds its own reference which is
dropped again when the property is deleted.
The upshot of this is that we always have a refcount >= 1. Upon
unplug the virtio-9p child is not finalized!
Drop our
From: Gonglei
virtio-serial-{pci, s390, ccw} all duplicate the
qdev properties of their VirtIOSerial child.
This approach does not work well with string or pointer
properties since we must be careful about leaking or
double-freeing them.
Use the QOM alias property to forward property accesses to
From: Gonglei
virtio-$device-{pci, s390, ccw} all duplicate the
qdev properties of their virtio child. This approach does
not work well with string or pointer properties since we
must be careful about leaking or double-freeing them.
Use the QOM alias property to forward property accesses to the
From: Gonglei
virtio-9p-pci all duplicate the qdev properties of their
V9fsState child. This approach does not work well with
string or pointer properties since we must be careful
about leaking or double-freeing them.
Use the QOM alias property to forward property accesses to the
V9fsState child
Kevin Wolf writes:
> Am 29.09.2014 um 15:05 hat Markus Armbruster geschrieben:
>> Kevin Wolf writes:
>>
>> > Am 16.09.2014 um 20:12 hat Markus Armbruster geschrieben:
>> >> @@ -903,21 +899,19 @@ DriveInfo *drive_new(QemuOpts *all_opts,
>> >> BlockInterfaceType block_default_type)
>> >> }
>
Dear all,
this patch fixes the issues I reported (related to VFIO trace points).
Many thanks for that.
Best Regards
Eric
On 09/23/2014 12:37 PM, Stefan Hajnoczi wrote:
> Use \w for properties and trace event names since they are both drawn
> from [a-zA-Z0-9_] character sets.
>
> The .* for ma
the limit of 0xff for 16 byte CDBs is intentional to
avoid overflows on 32-bit architectures.
Signed-off-by: Peter Lieven
Reviewed-by: Ronnie Sahlberg
---
block/iscsi.c | 12 ++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/block/iscsi.c b/block/iscsi.c
index 5c
Signed-off-by: Peter Lieven
Reviewed-by: Ronnie Sahlberg
---
block.c |4
include/block/block_int.h |3 +++
2 files changed, 7 insertions(+)
diff --git a/block.c b/block.c
index c5a251c..26806e0 100644
--- a/block.c
+++ b/block.c
@@ -541,6 +541,7 @@ void bdrv_refre
John Snow writes:
> The if_max_devs table as in the past been an immutable
> default that controls the mapping of index => (bus,unit)
> for all boards and all HBAs for each interface type.
>
> Since adding this mapping information to the HBA device
> itself is currently unwieldly from the perspec
Signed-off-by: Peter Lieven
Reviewed-by: Ronnie Sahlberg
---
block.c |5 +
1 file changed, 5 insertions(+)
diff --git a/block.c b/block.c
index 26806e0..a26e6a5 100644
--- a/block.c
+++ b/block.c
@@ -4559,6 +4559,11 @@ static int multiwrite_merge(BlockDriverState *bs,
BlockRequest *req
This series adds the basics for introducing a maximum transfer length
to the block layer. Its main purpose is currently avoiding that
a multiwrite_merge exceeds the max_xfer_len of an attached iSCSI LUN.
This is a required bug fix.
Splitting up requests according to the max_transfer_length will fo
On 23.09.2014 11:51, Paolo Bonzini wrote:
Il 23/09/2014 11:46, Paolo Bonzini ha scritto:
Il 02/09/2014 11:17, Peter Lieven ha scritto:
Juan is back, I'll let him pick it through his tree.
Juan, have you picked this up?
commit db80facefa62dff42bb50c73b0f03eda5f732b49
Author: Peter Lieven
Date
> >
> > In QEMU, ARI Forwarding is enabled defualt at emulation of PCIe
> > ports. ARI Forwarding enable setting at firmware/OS Control handoff.
> > If the bit is Set when a non-ARI Device is present, the non-ARI
> > Device can respond to Configuration Space accesses under what it
> > interprets as
John Snow writes:
> This patch adds the 'units_per_default_bus' property which
> allows individual boards to declare their desired
> index => (bus,unit) mapping for their default HBA, so that
> boards such as Q35 can specify that its default if_ide HBA,
> AHCI, only accepts one unit per bus.
>
>
This patch removes all DPRINTF and replace them by trace points.
A few DPRINTF used in error cases were transformed into error_report.
Signed-off-by: Eric Auger
---
- __func__ is removed since trace point name does the same job
- HWADDR_PRIx were replaced by PRIx64
- this transformation just is
Hi,
> > 'console' is mandatory; I guess that's okay.
> >
> > Are we guaranteed that either all events are sent? Or is there a need to
>
> Events can be dropped at hardware level if the event queue is full, for
> example. Would have to modify individual drivers to return error codes,
> i supp
201 - 262 of 262 matches
Mail list logo
