From: Julien Grall <jgrall@amazon.com>

Hi all,

At the moment, Live-Update will, by default, not proceed if there are
in-flight transactions. It is possible force it by passing -F but this
will break any connection with in-flight transactions.

There are PV drivers out that may never terminate some transaction. On
host running such guest, we would need to use -F. Unfortunately, this
also risks to break well-behaving guests (and even dom0) because
Live-Update will happen as soon as the timeout is hit.

This series aims to allow to Live-Update more safely even when the option
-F is used.

The first part of the series contains a few fixes for bug found while
testing Live-Update.

Cheers,

Julien Grall (10):
  MAINTAINERS: Add myself as reviewers for tools/xenstore
  tools/xenstored: Introduce lu_get_connection() and use it
  tools/xenstore: Don't assume conn->in points to the LU request
  tools/xenstored: Limit the number of requests a connection can delay
  tools/xenstored: xenstored_core.h should include fcntl.h
  tools/xenstored: Introduce a wrapper for conn->funcs->can_{read,
    write}
  tools/xenstored: delay_request: don't assume conn->in == in
  tools/xenstored: Extend restore code to handle multiple input buffer
  tools/xenstored: Dump delayed requests
  tools/xenstored: Delay new transaction while Live-Update is pending

 MAINTAINERS                        |   1 +
 tools/xenstore/xenstored_control.c |  66 +++++++++-
 tools/xenstore/xenstored_control.h |   7 ++
 tools/xenstore/xenstored_core.c    | 196 +++++++++++++++++++++++------
 tools/xenstore/xenstored_core.h    |   8 +-
 tools/xenstore/xenstored_domain.c  |  46 +++----
 tools/xenstore/xenstored_domain.h  |   6 +-
 7 files changed, 255 insertions(+), 75 deletions(-)

-- 
2.17.1

Hi,

On 16/06/2021 16:43, Julien Grall wrote:
> From: Julien Grall <jgrall@amazon.com>
> 
> Hi all,
> 
> At the moment, Live-Update will, by default, not proceed if there are
> in-flight transactions. It is possible force it by passing -F but this
> will break any connection with in-flight transactions.
> 
> There are PV drivers out that may never terminate some transaction. On
> host running such guest, we would need to use -F. Unfortunately, this
> also risks to break well-behaving guests (and even dom0) because
> Live-Update will happen as soon as the timeout is hit.
> 
> This series aims to allow to Live-Update more safely even when the option
> -F is used.
> 
> The first part of the series contains a few fixes for bug found while
> testing Live-Update.
> 
> Cheers,
> 
> Julien Grall (10):
>    MAINTAINERS: Add myself as reviewers for tools/xenstore
>    tools/xenstored: Introduce lu_get_connection() and use it
>    tools/xenstore: Don't assume conn->in points to the LU request
>    tools/xenstored: Limit the number of requests a connection can delay
>    tools/xenstored: xenstored_core.h should include fcntl.h
>    tools/xenstored: Introduce a wrapper for conn->funcs->can_{read,
>      write}
>    tools/xenstored: delay_request: don't assume conn->in == in
>    tools/xenstored: Extend restore code to handle multiple input buffer

I have committed the first 8 patches.

>    tools/xenstored: Dump delayed requests
>    tools/xenstored: Delay new transaction while Live-Update is pending

Cheers,

-- 
Julien Grall