interrupts: do not resize event list for non MSIX

Message ID 20211029073819.29336-1-david.marchand@redhat.com (mailing list archive)
State Accepted, archived
Delegated to: David Marchand
Headers
Series interrupts: do not resize event list for non MSIX |

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/iol-broadcom-Functional success Functional Testing PASS
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-broadcom-Performance success Performance Testing PASS
ci/github-robot: build success github build: passed
ci/iol-x86_64-unit-testing success Testing PASS
ci/iol-x86_64-compile-testing success Testing PASS
ci/Intel-compilation warning apply issues
ci/iol-aarch64-unit-testing success Testing PASS
ci/iol-aarch64-compile-testing success Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/iol-intel-Functional success Functional Testing PASS

Commit Message

David Marchand Oct. 29, 2021, 7:38 a.m. UTC
  Resizing event list only makes sense in MSIX case.

Besides, event list has always been RTE_MAX_RXTX_INTR_VEC_ID large.
Let's restore this assumption for code that might rely on this property
and only enlarge the event list when necessary.

Bugzilla ID: 843, 865
Fixes: 8cb5d08db940 ("interrupts: extend event list")

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 drivers/bus/pci/linux/pci_vfio.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
  

Comments

Harman Kalra Oct. 29, 2021, 9:38 a.m. UTC | #1
> -----Original Message-----
> From: David Marchand <david.marchand@redhat.com>
> Sent: Friday, October 29, 2021 1:08 PM
> To: dev@dpdk.org
> Cc: Anatoly Burakov <anatoly.burakov@intel.com>; Dmitry Kozlyuk
> <dmitry.kozliuk@gmail.com>; Harman Kalra <hkalra@marvell.com>
> Subject: [EXT] [PATCH] interrupts: do not resize event list for non MSIX
> 
> External Email
> 
> ----------------------------------------------------------------------
> Resizing event list only makes sense in MSIX case.
> 
> Besides, event list has always been RTE_MAX_RXTX_INTR_VEC_ID large.
> Let's restore this assumption for code that might rely on this property and
> only enlarge the event list when necessary.
> 
> Bugzilla ID: 843, 865
> Fixes: 8cb5d08db940 ("interrupts: extend event list")
> 

Thanks David for fixing this.

Acked-by: Harman Kalra <hkalra@marvell.com>

Thanks
Harman

> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
>  drivers/bus/pci/linux/pci_vfio.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/bus/pci/linux/pci_vfio.c b/drivers/bus/pci/linux/pci_vfio.c
> index f622e7f8e6..edcee92556 100644
> --- a/drivers/bus/pci/linux/pci_vfio.c
> +++ b/drivers/bus/pci/linux/pci_vfio.c
> @@ -266,12 +266,6 @@ pci_vfio_setup_interrupts(struct rte_pci_device
> *dev, int vfio_dev_fd)
>  			return -1;
>  		}
> 
> -		/* Reallocate the efds and elist fields of intr_handle based
> -		 * on PCI device MSIX size.
> -		 */
> -		if (rte_intr_event_list_update(dev->intr_handle, irq.count))
> -			return -1;
> -
>  		/* if this vector cannot be used with eventfd, fail if we
> explicitly
>  		 * specified interrupt type, otherwise continue */
>  		if ((irq.flags & VFIO_IRQ_INFO_EVENTFD) == 0) { @@ -283,6
> +277,14 @@ pci_vfio_setup_interrupts(struct rte_pci_device *dev, int
> vfio_dev_fd)
>  				continue;
>  		}
> 
> +		/* Reallocate the efds and elist fields of intr_handle based
> +		 * on PCI device MSIX size.
> +		 */
> +		if (i == VFIO_PCI_MSIX_IRQ_INDEX &&
> +				(uint32_t)rte_intr_nb_intr_get(dev-
> >intr_handle) < irq.count &&
> +				rte_intr_event_list_update(dev->intr_handle,
> irq.count))
> +			return -1;
> +
>  		/* set up an eventfd for interrupts */
>  		fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
>  		if (fd < 0) {
> --
> 2.23.0
  
David Marchand Oct. 29, 2021, 12:43 p.m. UTC | #2
On Fri, Oct 29, 2021 at 11:38 AM Harman Kalra <hkalra@marvell.com> wrote:
> > Resizing event list only makes sense in MSIX case.
> >
> > Besides, event list has always been RTE_MAX_RXTX_INTR_VEC_ID large.
> > Let's restore this assumption for code that might rely on this property and
> > only enlarge the event list when necessary.
> >
> > Bugzilla ID: 843, 865
> > Fixes: 8cb5d08db940 ("interrupts: extend event list")
> >
> > Signed-off-by: David Marchand <david.marchand@redhat.com>
> Acked-by: Harman Kalra <hkalra@marvell.com>

Applied, thanks.
  

Patch

diff --git a/drivers/bus/pci/linux/pci_vfio.c b/drivers/bus/pci/linux/pci_vfio.c
index f622e7f8e6..edcee92556 100644
--- a/drivers/bus/pci/linux/pci_vfio.c
+++ b/drivers/bus/pci/linux/pci_vfio.c
@@ -266,12 +266,6 @@  pci_vfio_setup_interrupts(struct rte_pci_device *dev, int vfio_dev_fd)
 			return -1;
 		}
 
-		/* Reallocate the efds and elist fields of intr_handle based
-		 * on PCI device MSIX size.
-		 */
-		if (rte_intr_event_list_update(dev->intr_handle, irq.count))
-			return -1;
-
 		/* if this vector cannot be used with eventfd, fail if we explicitly
 		 * specified interrupt type, otherwise continue */
 		if ((irq.flags & VFIO_IRQ_INFO_EVENTFD) == 0) {
@@ -283,6 +277,14 @@  pci_vfio_setup_interrupts(struct rte_pci_device *dev, int vfio_dev_fd)
 				continue;
 		}
 
+		/* Reallocate the efds and elist fields of intr_handle based
+		 * on PCI device MSIX size.
+		 */
+		if (i == VFIO_PCI_MSIX_IRQ_INDEX &&
+				(uint32_t)rte_intr_nb_intr_get(dev->intr_handle) < irq.count &&
+				rte_intr_event_list_update(dev->intr_handle, irq.count))
+			return -1;
+
 		/* set up an eventfd for interrupts */
 		fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
 		if (fd < 0) {