Re: [PATCH net-next v4 13/15] ixd: add reset checks and initialize the mailbox

[Larysa Zaremba <larysa.zaremba@intel.com>]

Below is an answer to concerns from Sashiko:
> --- Patch [13]: [PATCH net-next v4 13/15] ixd: add reset checks and initialize the mailbox ---
> commit 7fa1c35636d0bae10b336e706045a6dfd122058d
> Author: Larysa Zaremba <larysa.zaremba@intel.com>
> ixd: add reset checks and initialize the mailbox
> This commit adds hardware reset triggering and mailbox initialization
> for the ixd driver, including a delayed work task to verify the reset
> completion during probe.
> [ ... ]
> > diff --git a/drivers/net/ethernet/intel/ixd/ixd_main.c b/drivers/net/ethernet/intel/ixd/ixd_main.c
> > index 75ee53152e614..35d0ff5c8d983 100644
> > --- a/drivers/net/ethernet/intel/ixd/ixd_main.c
> > +++ b/drivers/net/ethernet/intel/ixd/ixd_main.c
> [ ... ]
> > @@ -16,7 +17,15 @@ static void ixd_remove(struct pci_dev *pdev)
> >  {
> >  	struct ixd_adapter *adapter = pci_get_drvdata(pdev);
> >  
> > -	libie_pci_unmap_all_mmio_regions(&adapter->hw);
> > +	/* Do not mix removal with (re)initialization */
> > +	cancel_delayed_work_sync(&adapter->init_task.init_work);
> > +	/* Leave the device clean on exit */
> > +	if (adapter->xnm)
> > +		libie_ctlq_xn_shutdown(adapter->xnm);
> > +	ixd_trigger_reset(adapter);
> > +	ixd_deinit_dflt_mbx(adapter);
> Could this cause memory corruption or IOMMU faults?
> In ixd_remove(), ixd_trigger_reset() issues a reset but does not wait
> for it to complete. Immediately after, ixd_deinit_dflt_mbx() is called,
> which unmaps and frees the DMA rings.
> If the hardware is still actively bus mastering before the reset fully
> takes effect, could it perform DMA into memory that has already been
> unmapped? Does this path need to wait for the reset to complete using
> ixd_check_reset_complete() before freeing the mailbox resources?
> 
No, ixd_check_reset_complete() is not needed here, triggering is enough
> 
> > +
> > +	libie_pci_unmap_all_mmio_regions(&adapter->cp_ctx.mmio_info);
> >  }
>