[dpdk-stable] [PATCH] net/virtio: fix read isr as MSI is enabled

[Tan, Jianfeng]

Sorry, please just NACK this patch. (I'll do that in patchwork)

I was sharing this half-baked patch with Zhiyong and forget to remove "CC".
Zhiyong and I are still working on this bug. Will send it out later.

Thanks,
Jianfeng

> -----Original Message-----
> From: Yuanhan Liu [mailto:yuanhan.liu at linux.intel.com]
> Sent: Thursday, April 6, 2017 1:19 PM
> To: Tan, Jianfeng
> Cc: Yang, Zhiyong; stable at dpdk.org; Stephen Hemminger
> Subject: Re: [dpdk-stable] [PATCH] net/virtio: fix read isr as MSI is enabled
> 
> On Tue, Mar 28, 2017 at 06:46:51AM +0000, Jianfeng Tan wrote:
> > We should not always check isr to know if link status is changed.
> 
> Did this patch fix a real issue?
> 
> > Here is how driver should handle interrupt quoted from virtio spec,
> >
> >     If MSI-X capability is disabled:
> >       – Read the ISR Status field, which will reset it to zero.
> >       – If the lower bit is set: look through the used rings of all
> >       virtqueues for the device, to see if any progress has been made
> >       by the device which requires servicing.
> >       – If the second lower bit is set: re-examine the configuration
> >       space to see what changed.
> >     If MSI-X capability is enabled:
> >       – Look through the used rings of all virtqueues mapped to that
> >       MSI-X vector for the device, to see if any progress has been
> >       made by the device which requires servicing.
> >       – If the MSI-X vector is equal to config_msix_vector, re-examine
> >       the configuration space to see what changed
> >
> > Fixes: 8d6d7e5cb3b1 ("virtio: support link state interrupt")
> > CC: stable at dpdk.org
> >
> > CC: Stephen Hemminger <stephen at networkplumber.org>
> > Signed-off-by: Jianfeng Tan <jianfeng.tan at intel.com>
> > ---
> >  drivers/net/virtio/virtio_ethdev.c | 22 +++++++++++++---------
> >  1 file changed, 13 insertions(+), 9 deletions(-)
> >
> > diff --git a/drivers/net/virtio/virtio_ethdev.c
> b/drivers/net/virtio/virtio_ethdev.c
> > index 3cf4102..cb30d11 100644
> > --- a/drivers/net/virtio/virtio_ethdev.c
> > +++ b/drivers/net/virtio/virtio_ethdev.c
> > @@ -36,6 +36,7 @@
> >  #include <stdio.h>
> >  #include <errno.h>
> >  #include <unistd.h>
> > +#include <stdbool.h>
> >
> >  #include <rte_ethdev.h>
> >  #include <rte_memcpy.h>
> > @@ -1197,20 +1198,23 @@ virtio_interrupt_handler(struct rte_intr_handle
> *handle,
> >  	struct rte_eth_dev *dev = param;
> >  	struct virtio_hw *hw = dev->data->dev_private;
> >  	uint8_t isr;
> > +	bool check_config;
> >
> > -	/* Read interrupt status which clears interrupt */
> > -	isr = vtpci_isr(hw);
> > -	PMD_DRV_LOG(INFO, "interrupt status = %#x", isr);
> > +	/* isr is used only when msix is not enabled */
> > +	if (!hw->modern && !hw->use_msix) {
> 
> The comment doesn't quite match the code: the hw->modern part is missing.
> 
> > +		isr = vtpci_isr(hw); /* Read to clears interrupt */
> > +		PMD_DRV_LOG(INFO, "interrupt status = %#x", isr);
> > +		check_config = (isr & VIRTIO_PCI_ISR_CONFIG) ? true : false;
> 
> The "? :" is unnecessary.
> 
> 	--yliu
> 
> > +	} else {
> > +		check_config = true;
> > +	}
> >
> >  	if (rte_intr_enable(handle) < 0)
> >  		PMD_DRV_LOG(ERR, "interrupt enable failed");
> >
> > -	if (isr & VIRTIO_PCI_ISR_CONFIG) {
> > -		if (virtio_dev_link_update(dev, 0) == 0)
> > -			_rte_eth_dev_callback_process(dev,
> > -
> RTE_ETH_EVENT_INTR_LSC, NULL);
> > -	}
> > -
> > +	if (check_config && virtio_dev_link_update(dev, 0) == 0)
> > +		_rte_eth_dev_callback_process(dev,
> > +					      RTE_ETH_EVENT_INTR_LSC, NULL);
> >  }
> >
> >  static void
> > --
> > 2.7.4