On Thu, 2020-06-11 at 16:00 +0200, Erik Skultety wrote:
On Tue, Jun 09, 2020 at 04:43:48PM -0500, Jonathon Jongsma wrote:
> Add the ability to destroy mdev node devices via the mdevctl
> utility.
>
> Signed-off-by: Jonathon Jongsma <jjongsma(a)redhat.com>
> ---
> src/node_device/node_device_driver.c | 46
> ++++++++++++++++++++++++++++
> src/node_device/node_device_driver.h | 3 ++
> 2 files changed, 49 insertions(+)
>
> diff --git a/src/node_device/node_device_driver.c
> b/src/node_device/node_device_driver.c
> index dbc7eb4d1e..c956bb55fc 100644
> --- a/src/node_device/node_device_driver.c
> +++ b/src/node_device/node_device_driver.c
> @@ -790,6 +790,45 @@ nodeDeviceCreateXML(virConnectPtr conn,
> }
>
>
> +virCommandPtr
> +nodeDeviceGetMdevctlStopCommand(const char *uuid,
> + bool persist)
> +{
> + g_autofree char *mdevctl = virFindFileInPath(MDEVCTL);
> + const char *subcommand;
> +
> + if (!mdevctl)
> + return NULL;
> +
> + if (persist)
> + subcommand = "undefine";
"undefine" is a NOP on active mdevs IIRC, so again the helper name is
confusing.
Oh, you're right. This part was meant to plan ahead for persistent
mediated devices, but since it's not yet used (and since it doesn't
have the effect intended, as you point out), it should probably just be
removed from this patch series.
> + else
> + subcommand = "stop";
> +
> + virCommandPtr cmd = virCommandNewArgList(mdevctl,
> + subcommand,
> + "-u",
> + uuid,
> + NULL);
> +
> + return cmd;
> +}
Like I mentioned already, we could have a generic translator to the
mdevctl
subcommands.
Regards,
Erik