It's 1ms, not 1 second. It's a workaround, just to provide a fast and dirty fix for someone who needs this.
From: "Stephen Hemminger"<step...@networkplumber.org> > Date: Tue, Oct 8, 2024, 01:54 > Subject: [External] Re: [PATCH] mlx5: fix race at mlx5_dev_close > To: "hepeng"<hepeng.0...@bytedance.com> > Cc: <dev@dpdk.org> > On Thu, 11 Apr 2024 14:17:40 +0800 > hepeng <hepeng.0...@bytedance.com> wrote: > > > > From: "hepeng.0320" <hepeng.0...@bytedance.com> > > > > mlx5_dev_close currently will set priv->sh->port[priv->dev_port - > > 1].nl_ih_port_id to RTE_MAX_ETHPORTS to avoid mlx5_dev_interrupt_nl_cb > > to use the port's dev_private, because later the rte_eth_dev_close > > will free the dev_private and set the pointer to NULL. > > > > However, since mlx5_dev_interrupt_nl_cb is running in another thread, > > I think the race still exists. So perhaps an easy fix is to wait for > > 1ms to avoid this race. > > > > Signed-off-by: hepeng.0320 <hepeng.0...@bytedance.com> > > > Not the pest way to handle this. Adding a one second delay on shutdown > hurts some availability scenarios. Looks like mlx5 needs a more coordinated > shutdown to be safe; adding big delays is not the correct fix. > > > > > >
