Re: [PATCH v1 16/20] node_device_udev: Use a worker pool for processing events and emitting nodedev event

On Fri, Apr 19, 2024 at 04:49 PM +0200, Marc Hartmayer <mhartmay(a)linux.ibm.com&gt; wrote: > Use a worker pool for processing the events (e.g. udev, mdevctl config changes) > and the initialization instead of a separate initThread and a mdevctl-thread. > This has the large advantage that we can leverage the job API and now this > thread pool is responsible to do all the "costly-work" and emitting the libvirt > nodedev events. > > Signed-off-by: Marc Hartmayer <mhartmay(a)linux.ibm.com&gt; > --- > src/node_device/node_device_udev.c | 244 +++++++++++++++++++++-------- > 1 file changed, 179 insertions(+), 65 deletions(-) > > diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c > index e4b1532dc385..67a8b5cd7132 100644 […snip…] > } > > @@ -2278,11 +2380,19 @@ nodeStateShutdownWait(void) > return 0; > > VIR_WITH_OBJECT_LOCK_GUARD(priv) { > - if (priv->initThread) > - virThreadJoin(priv->initThread); > - if (priv->udevThread) > - virThreadJoin(priv->udevThread); > + if (priv->mdevctlTimeout != -1) { > + virEventRemoveTimeout(priv->mdevctlTimeout); > + priv->mdevctlTimeout = -1; > + } > + > + if (priv->watch) { > + virEventRemoveHandle(priv->watch); > + priv->watch = -1; > + } Too many rebases… the diff should read as follows: @@ -2278,11 +2380,12 @@ nodeStateShutdownWait(void) return 0; VIR_WITH_OBJECT_LOCK_GUARD(priv) { - if (priv->initThread) - virThreadJoin(priv->initThread); if (priv->udevThread) virThreadJoin(priv->udevThread); } + + if (priv->workerPool) + virThreadPoolDrain(priv->workerPool); return 0; } […snip]