Hello,

Please, ignore 1st patch for now.

Although it fixes the problem observed it looks like there is a better
and easier solution (many thanks to Intel Guys that explained e1000
operation in details:
http://sourceforge.net/mailarchive/forum.php?thread_name=CAGHCxhcad%3Dzx7ihX5zoDB%3DZOLGGuZty%3DBck6zSoMQ-9S3ZJo7w%40mail.gmail.com&forum_name=e1000-devel).

We'll submit the final patch soon.

Dmitry.

On Mon, Oct 15, 2012 at 6:48 PM, Dmitry Fleytman <dmi...@daynix.com> wrote:
> Following patches fix a few race conditions in E1000 code:
>
>   1st patch fixes race condition between driver shutdown and device shutdown 
> (see patch comment)
>   It also work-arounds race condition in e1000 Linux driver between RX enable 
> and RX rings init
>   (Separate patch for the second problem sent to e1000-devel/linux/kernel and 
> accepted by maintainers, see
>    
> http://sourceforge.net/mailarchive/forum.php?thread_name=1350280341.2152.12.camel%40jtkirshe-mobl&forum_name=e1000-devel)
>
>   2nd patch is pretty trivial and adds forgotten field into live migration 
> list thus fixing another race condition.
>
> Dmitry Fleytman (2):
>   Fix a race condition in E1000 device implementation:
>   Fix a race condition in E1000 device live migration.     One of
>     data-transfer related flags not in migrated fields list.
>
>  hw/e1000.c | 25 +++++++++++++++++++++++--
>  1 file changed, 23 insertions(+), 2 deletions(-)
>
> --
> 1.7.11.4
>



-- 
Dmitry Fleytman
Technology Expert and Consultant,

Daynix Computing Ltd.

Cell: +972-54-2819481

Skype: dmitry.fleytman

Reply via email to