subject:"\[PATCH\] virtio\: add VIRTQUEUE

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

2022-11-01 Thread Vladimir Sementsov-Ogievskiy


ping

On 9/19/22 22:48, Vladimir Sementsov-Ogievskiy wrote:

For now we only log the vhost device error, when virtqueue is actually
stopped. Let's add a QAPI event, which makes possible:

  - collect statistics of such errors
  - make immediate actions: take coredums or do some other debugging

The event could be reused for some other virtqueue problems (not only
for vhost devices) in future. For this it gets a generic name and
structure.

Signed-off-by: Vladimir Sementsov-Ogievskiy


--
Best regards,
Vladimir

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

2022-10-12 Thread Vladimir Sementsov-Ogievskiy


ping

On 9/19/22 22:48, Vladimir Sementsov-Ogievskiy wrote:

For now we only log the vhost device error, when virtqueue is actually
stopped. Let's add a QAPI event, which makes possible:

  - collect statistics of such errors
  - make immediate actions: take coredums or do some other debugging

The event could be reused for some other virtqueue problems (not only
for vhost devices) in future. For this it gets a generic name and
structure.

Signed-off-by: Vladimir Sementsov-Ogievskiy 


--
Best regards,
Vladimir

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

2022-09-20 Thread Roman Kagan

On Tue, Sep 20, 2022 at 06:10:08PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> On 9/20/22 17:47, Markus Armbruster wrote:
> > Vladimir Sementsov-Ogievskiy  writes:
> > 
> > > For now we only log the vhost device error, when virtqueue is actually
> > > stopped. Let's add a QAPI event, which makes possible:
> > > 
> > >   - collect statistics of such errors
> > >   - make immediate actions: take coredums or do some other debugging

+ inform the user through a management API or UI, so that (s)he can
  react somehow, e.g. reset the device driver in the guest or even build
  up some automation to do so

Note that basically every inconsistency discovered during virtqueue
processing results in a silent virtqueue stop.  The guest then just sees
the requests getting stuck somewhere in the device for no visible
reason.  This event provides a means to inform the management layer of
this situation in a timely fashion.

> > 
> > Core dumps, I presume.
> > 
> > Is QMP the right tool for the job?  Or could a trace point do?
> 
> Management tool already can collect QMP events. So, if we want to
> forward some QMP events to other subsystems (to immediately inform
> support team, or to update some statistics) it's simple to realize for
> QMP events. But I'm not sure how to do it for trace-events.. Scanning
> trace logs is not convenient.

Right.  Trace points are a debugging tool: when you expect the problem
to reproduce, you activate them and watch the logs.  On the contrary,
QMP events can trigger some logic in the management layer and provide
for some recovery action.

> > > +##
> > > +# @VIRTQUEUE_ERROR:
> > > +#
> > > +# Emitted when a device virtqueue fails in runtime.
> > > +#
> > > +# @device: the device's ID if it has one
> > > +# @path: the device's QOM path
> > > +# @virtqueue: virtqueue index
> > > +# @error: error identifier
> > > +# @description: human readable description
> > > +#
> > > +# Since: 7.2
> > > +##
> > > +{ 'event': 'VIRTQUEUE_ERROR',
> > > + 'data': { '*device': 'str', 'path': 'str', 'virtqueue': 'int',
> > > +'error': 'VirtqueueError', 'description': 'str'} }
> > 
> > Can the guest trigger the event?
> 
> Yes, but as I understand, only once per virtqueue.

Right, in the sense that every relevant dataplane implementation would
stop the virtqueue on such an error, so in order to trigger a new one
the driver would need to reset the device first.  I guess rate-limiting
is unnecessary here.

Thanks,
Roman.

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

2022-09-20 Thread Vladimir Sementsov-Ogievskiy


On 9/20/22 17:47, Markus Armbruster wrote:

Vladimir Sementsov-Ogievskiy  writes:


For now we only log the vhost device error, when virtqueue is actually
stopped. Let's add a QAPI event, which makes possible:

  - collect statistics of such errors
  - make immediate actions: take coredums or do some other debugging


Core dumps, I presume.

Is QMP the right tool for the job?  Or could a trace point do?


Management tool already can collect QMP events. So, if we want to forward some 
QMP events to other subsystems (to immediately inform support team, or to 
update some statistics) it's simple to realize for QMP events. But I'm not sure 
how to do it for trace-events.. Scanning trace logs is not convenient.

Another benefit of QMP events is that they are objects, with fields, which is 
better for machine processing than textual trace-events.




The event could be reused for some other virtqueue problems (not only
for vhost devices) in future. For this it gets a generic name and
structure.

Signed-off-by: Vladimir Sementsov-Ogievskiy 
---
  hw/virtio/vhost.c | 12 +---
  qapi/qdev.json| 25 +
  2 files changed, 34 insertions(+), 3 deletions(-)

diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index f758f177bb..caa81f2ace 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -15,6 +15,7 @@
  
  #include "qemu/osdep.h"

  #include "qapi/error.h"
+#include "qapi/qapi-events-qdev.h"
  #include "hw/virtio/vhost.h"
  #include "qemu/atomic.h"
  #include "qemu/range.h"


Only tangentially related to this patch, but here goes anyway:

/* enabled until disconnected backend stabilizes */
#define _VHOST_DEBUG 1

This is from 2016.  Has it stabilized?


Hmm, I don't know:) It works for us)

But anyway, I agree that error/debug reporting here needs an update. I don't 
think that dropping the messages is good. Some should be converted to 
errp-reporting, some to warnings or assertions..



#ifdef _VHOST_DEBUG
#define VHOST_OPS_DEBUG(retval, fmt, ...) \
do { \
error_report(fmt ": %s (%d)", ## __VA_ARGS__, \
 strerror(-retval), -retval); \

error_report() is for errors the user can do something about, not for
debug messages.

} while (0)
#else
#define VHOST_OPS_DEBUG(retval, fmt, ...) \
do { } while (0)
#endif


@@ -1287,11 +1288,16 @@ static void 
vhost_virtqueue_error_notifier(EventNotifier *n)
  struct vhost_virtqueue *vq = container_of(n, struct vhost_virtqueue,
error_notifier);
  struct vhost_dev *dev = vq->dev;
-int index = vq - dev->vqs;
  
  if (event_notifier_test_and_clear(n) && dev->vdev) {

-VHOST_OPS_DEBUG(-EINVAL,  "vhost vring error in virtqueue %d",
-dev->vq_index + index);
+int ind = vq - dev->vqs + dev->vq_index;
+DeviceState *ds = &dev->vdev->parent_obj;
+
+VHOST_OPS_DEBUG(-EINVAL,  "vhost vring error in virtqueue %d", ind);
+qapi_event_send_virtqueue_error(!!ds->id, ds->id, ds->canonical_path,
+ind, VIRTQUEUE_ERROR_VHOST_VRING_ERR,
+"vhost reported failure through vring "
+"error fd");


Do we still need VHOST_OPS_DEBUG() here?


I think, this should be decided separately from this patch. Here I keep current 
behavior and add new event.




  }
  }
  
diff --git a/qapi/qdev.json b/qapi/qdev.json

index 2708fb4e99..b7c2669c2c 100644
--- a/qapi/qdev.json
+++ b/qapi/qdev.json
@@ -158,3 +158,28 @@
  ##
  { 'event': 'DEVICE_UNPLUG_GUEST_ERROR',
'data': { '*device': 'str', 'path': 'str' } }
+
+##
+# @VirtqueueError:
+#
+# Since: 7.2
+##
+{ 'enum': 'VirtqueueError',
+  'data': [ 'vhost-vring-err' ] }
+
+##
+# @VIRTQUEUE_ERROR:
+#
+# Emitted when a device virtqueue fails in runtime.
+#
+# @device: the device's ID if it has one
+# @path: the device's QOM path
+# @virtqueue: virtqueue index
+# @error: error identifier
+# @description: human readable description
+#
+# Since: 7.2
+##
+{ 'event': 'VIRTQUEUE_ERROR',
+ 'data': { '*device': 'str', 'path': 'str', 'virtqueue': 'int',
+'error': 'VirtqueueError', 'description': 'str'} }


Can the guest trigger the event?


Yes, but as I understand, only once per virtqueue.



If yes, it needs to be rate-limited.



This may be needed if VIRTQUEUE_ERROR will be shared with other errors. Still 
adding it now will not hurt I think.

--
Best regards,
Vladimir

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

2022-09-20 Thread Markus Armbruster

Vladimir Sementsov-Ogievskiy  writes:

> For now we only log the vhost device error, when virtqueue is actually
> stopped. Let's add a QAPI event, which makes possible:
>
>  - collect statistics of such errors
>  - make immediate actions: take coredums or do some other debugging

Core dumps, I presume.

Is QMP the right tool for the job?  Or could a trace point do?

> The event could be reused for some other virtqueue problems (not only
> for vhost devices) in future. For this it gets a generic name and
> structure.
>
> Signed-off-by: Vladimir Sementsov-Ogievskiy 
> ---
>  hw/virtio/vhost.c | 12 +---
>  qapi/qdev.json| 25 +
>  2 files changed, 34 insertions(+), 3 deletions(-)
>
> diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
> index f758f177bb..caa81f2ace 100644
> --- a/hw/virtio/vhost.c
> +++ b/hw/virtio/vhost.c
> @@ -15,6 +15,7 @@
>  
>  #include "qemu/osdep.h"
>  #include "qapi/error.h"
> +#include "qapi/qapi-events-qdev.h"
>  #include "hw/virtio/vhost.h"
>  #include "qemu/atomic.h"
>  #include "qemu/range.h"

Only tangentially related to this patch, but here goes anyway:

   /* enabled until disconnected backend stabilizes */
   #define _VHOST_DEBUG 1

This is from 2016.  Has it stabilized?

   #ifdef _VHOST_DEBUG
   #define VHOST_OPS_DEBUG(retval, fmt, ...) \
   do { \
   error_report(fmt ": %s (%d)", ## __VA_ARGS__, \
strerror(-retval), -retval); \

error_report() is for errors the user can do something about, not for
debug messages.

   } while (0)
   #else
   #define VHOST_OPS_DEBUG(retval, fmt, ...) \
   do { } while (0)
   #endif

> @@ -1287,11 +1288,16 @@ static void 
> vhost_virtqueue_error_notifier(EventNotifier *n)
>  struct vhost_virtqueue *vq = container_of(n, struct vhost_virtqueue,
>error_notifier);
>  struct vhost_dev *dev = vq->dev;
> -int index = vq - dev->vqs;
>  
>  if (event_notifier_test_and_clear(n) && dev->vdev) {
> -VHOST_OPS_DEBUG(-EINVAL,  "vhost vring error in virtqueue %d",
> -dev->vq_index + index);
> +int ind = vq - dev->vqs + dev->vq_index;
> +DeviceState *ds = &dev->vdev->parent_obj;
> +
> +VHOST_OPS_DEBUG(-EINVAL,  "vhost vring error in virtqueue %d", ind);
> +qapi_event_send_virtqueue_error(!!ds->id, ds->id, ds->canonical_path,
> +ind, VIRTQUEUE_ERROR_VHOST_VRING_ERR,
> +"vhost reported failure through 
> vring "
> +"error fd");

Do we still need VHOST_OPS_DEBUG() here?

>  }
>  }
>  
> diff --git a/qapi/qdev.json b/qapi/qdev.json
> index 2708fb4e99..b7c2669c2c 100644
> --- a/qapi/qdev.json
> +++ b/qapi/qdev.json
> @@ -158,3 +158,28 @@
>  ##
>  { 'event': 'DEVICE_UNPLUG_GUEST_ERROR',
>'data': { '*device': 'str', 'path': 'str' } }
> +
> +##
> +# @VirtqueueError:
> +#
> +# Since: 7.2
> +##
> +{ 'enum': 'VirtqueueError',
> +  'data': [ 'vhost-vring-err' ] }
> +
> +##
> +# @VIRTQUEUE_ERROR:
> +#
> +# Emitted when a device virtqueue fails in runtime.
> +#
> +# @device: the device's ID if it has one
> +# @path: the device's QOM path
> +# @virtqueue: virtqueue index
> +# @error: error identifier
> +# @description: human readable description
> +#
> +# Since: 7.2
> +##
> +{ 'event': 'VIRTQUEUE_ERROR',
> + 'data': { '*device': 'str', 'path': 'str', 'virtqueue': 'int',
> +'error': 'VirtqueueError', 'description': 'str'} }

Can the guest trigger the event?

If yes, it needs to be rate-limited.

[PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

2022-09-19 Thread Vladimir Sementsov-Ogievskiy

For now we only log the vhost device error, when virtqueue is actually
stopped. Let's add a QAPI event, which makes possible:

 - collect statistics of such errors
 - make immediate actions: take coredums or do some other debugging

The event could be reused for some other virtqueue problems (not only
for vhost devices) in future. For this it gets a generic name and
structure.

Signed-off-by: Vladimir Sementsov-Ogievskiy 
---
 hw/virtio/vhost.c | 12 +---
 qapi/qdev.json| 25 +
 2 files changed, 34 insertions(+), 3 deletions(-)

diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index f758f177bb..caa81f2ace 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -15,6 +15,7 @@
 
 #include "qemu/osdep.h"
 #include "qapi/error.h"
+#include "qapi/qapi-events-qdev.h"
 #include "hw/virtio/vhost.h"
 #include "qemu/atomic.h"
 #include "qemu/range.h"
@@ -1287,11 +1288,16 @@ static void 
vhost_virtqueue_error_notifier(EventNotifier *n)
 struct vhost_virtqueue *vq = container_of(n, struct vhost_virtqueue,
   error_notifier);
 struct vhost_dev *dev = vq->dev;
-int index = vq - dev->vqs;
 
 if (event_notifier_test_and_clear(n) && dev->vdev) {
-VHOST_OPS_DEBUG(-EINVAL,  "vhost vring error in virtqueue %d",
-dev->vq_index + index);
+int ind = vq - dev->vqs + dev->vq_index;
+DeviceState *ds = &dev->vdev->parent_obj;
+
+VHOST_OPS_DEBUG(-EINVAL,  "vhost vring error in virtqueue %d", ind);
+qapi_event_send_virtqueue_error(!!ds->id, ds->id, ds->canonical_path,
+ind, VIRTQUEUE_ERROR_VHOST_VRING_ERR,
+"vhost reported failure through vring "
+"error fd");
 }
 }
 
diff --git a/qapi/qdev.json b/qapi/qdev.json
index 2708fb4e99..b7c2669c2c 100644
--- a/qapi/qdev.json
+++ b/qapi/qdev.json
@@ -158,3 +158,28 @@
 ##
 { 'event': 'DEVICE_UNPLUG_GUEST_ERROR',
   'data': { '*device': 'str', 'path': 'str' } }
+
+##
+# @VirtqueueError:
+#
+# Since: 7.2
+##
+{ 'enum': 'VirtqueueError',
+  'data': [ 'vhost-vring-err' ] }
+
+##
+# @VIRTQUEUE_ERROR:
+#
+# Emitted when a device virtqueue fails in runtime.
+#
+# @device: the device's ID if it has one
+# @path: the device's QOM path
+# @virtqueue: virtqueue index
+# @error: error identifier
+# @description: human readable description
+#
+# Since: 7.2
+##
+{ 'event': 'VIRTQUEUE_ERROR',
+ 'data': { '*device': 'str', 'path': 'str', 'virtqueue': 'int',
+'error': 'VirtqueueError', 'description': 'str'} }
-- 
2.25.1

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

Re: [PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

[PATCH] virtio: add VIRTQUEUE_ERROR QAPI event

6 matches

Site Navigation

Mail list logo

Footer information