DPDK-dev Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] eal/linux: revert eventfd leak fix for MSI-X
@ 2026-11-18 15:29 David Marchand
  2026-11-19 10:55 ` David Marchand
  2026-11-19 13:11 ` Maxime Coquelin
  0 siblings, 2 replies; 5+ messages in thread
From: David Marchand @ 2026-11-18 15:29 UTC (permalink / raw)
  To: dev; +Cc: iboukris, maxime.coquelin, thomas, Harman Kalra, Anatoly Burakov

The event FDs leak fix looks correct on the principle but it seems it
was too soon to take it, as it broke net/virtio and may have broken
other drivers.

Revert this fix until we get a proper fix and acknowledgement that
other drivers are fine with the change.

Fixes: 3ba6b689893f ("eal/linux: fix eventfd leaks when enabling interrupts")

Signed-off-by: David Marchand <david.marchand@redhat.com>
---
 lib/eal/linux/eal_interrupts.c | 46 +++++++++-------------------------
 1 file changed, 12 insertions(+), 34 deletions(-)

diff --git a/lib/eal/linux/eal_interrupts.c b/lib/eal/linux/eal_interrupts.c
index dfe1660dd0..c63956cde5 100644
--- a/lib/eal/linux/eal_interrupts.c
+++ b/lib/eal/linux/eal_interrupts.c
@@ -1556,44 +1556,31 @@ RTE_EXPORT_INTERNAL_SYMBOL(rte_intr_efd_enable)
 int
 rte_intr_efd_enable(struct rte_intr_handle *intr_handle, uint32_t nb_efd)
 {
+	uint32_t i;
+	int fd;
 	uint32_t n = RTE_MIN(nb_efd, (uint32_t)RTE_MAX_RXTX_INTR_VEC_ID);
-	int ret;
 
 	assert(nb_efd != 0);
 
 	if (rte_intr_type_get(intr_handle) == RTE_INTR_HANDLE_VFIO_MSIX) {
-		if (rte_intr_nb_efd_get(intr_handle) != 0) {
-			EAL_LOG(ERR, "Interrupt eventfds are already enabled");
-			return -EBUSY;
-		}
-
-		for (uint32_t i = 0; i < n; i++) {
-			int fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
-
+		for (i = 0; i < n; i++) {
+			fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
 			if (fd < 0) {
-				ret = -errno;
 				EAL_LOG(ERR,
 					"can't setup eventfd, error %i (%s)",
 					errno, strerror(errno));
-				goto error;
+				return -errno;
 			}
 
-			/* Record each eventfd as it is created so that the
-			 * error path can close exactly those that exist.
-			 */
-			if (rte_intr_efds_index_set(intr_handle, i, fd) ||
-					rte_intr_nb_efd_set(intr_handle, i + 1)) {
-				ret = -rte_errno;
-				close(fd);
-				rte_intr_efds_index_set(intr_handle, i, -1);
-				goto error;
-			}
+			if (rte_intr_efds_index_set(intr_handle, i, fd))
+				return -rte_errno;
 		}
 
-		if (rte_intr_max_intr_set(intr_handle, NB_OTHER_INTR + n)) {
-			ret = -rte_errno;
-			goto error;
-		}
+		if (rte_intr_nb_efd_set(intr_handle, n))
+			return -rte_errno;
+
+		if (rte_intr_max_intr_set(intr_handle, NB_OTHER_INTR + n))
+			return -rte_errno;
 	} else if (rte_intr_type_get(intr_handle) == RTE_INTR_HANDLE_VDEV) {
 		/* only check, initialization would be done in vdev driver.*/
 		if ((uint64_t)rte_intr_efd_counter_size_get(intr_handle) >
@@ -1611,15 +1598,6 @@ rte_intr_efd_enable(struct rte_intr_handle *intr_handle, uint32_t nb_efd)
 	}
 
 	return 0;
-
-error:
-	for (int i = 0; i < rte_intr_nb_efd_get(intr_handle); i++) {
-		close(rte_intr_efds_index_get(intr_handle, i));
-		rte_intr_efds_index_set(intr_handle, i, -1);
-	}
-	rte_intr_nb_efd_set(intr_handle, 0);
-
-	return ret;
 }
 
 RTE_EXPORT_INTERNAL_SYMBOL(rte_intr_efd_disable)
-- 
2.56.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH] eal/linux: revert eventfd leak fix for MSI-X
  2026-11-18 15:29 [PATCH] eal/linux: revert eventfd leak fix for MSI-X David Marchand
@ 2026-11-19 10:55 ` David Marchand
  2026-11-19 13:34   ` David Marchand
  2026-11-19 13:11 ` Maxime Coquelin
  1 sibling, 1 reply; 5+ messages in thread
From: David Marchand @ 2026-11-19 10:55 UTC (permalink / raw)
  To: dev; +Cc: iboukris, maxime.coquelin, thomas, Harman Kalra, Anatoly Burakov

On Wed, 18 Nov 2026 at 16:30, David Marchand <david.marchand@redhat.com> wrote:
>
> The event FDs leak fix looks correct on the principle but it seems it
> was too soon to take it, as it broke net/virtio and may have broken
> other drivers.
>
> Revert this fix until we get a proper fix and acknowledgement that
> other drivers are fine with the change.
>
> Fixes: 3ba6b689893f ("eal/linux: fix eventfd leaks when enabling interrupts")
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>

Recheck-request: github-robot: build

-- 
David Marchand


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] eal/linux: revert eventfd leak fix for MSI-X
  2026-11-18 15:29 [PATCH] eal/linux: revert eventfd leak fix for MSI-X David Marchand
  2026-11-19 10:55 ` David Marchand
@ 2026-11-19 13:11 ` Maxime Coquelin
  2026-11-19 15:01   ` Thomas Monjalon
  1 sibling, 1 reply; 5+ messages in thread
From: Maxime Coquelin @ 2026-11-19 13:11 UTC (permalink / raw)
  To: David Marchand; +Cc: dev, iboukris, thomas, Harman Kalra, Anatoly Burakov

On Wed, Nov 18, 2026 at 4:30 PM David Marchand
<david.marchand@redhat.com> wrote:
>
> The event FDs leak fix looks correct on the principle but it seems it
> was too soon to take it, as it broke net/virtio and may have broken
> other drivers.
>
> Revert this fix until we get a proper fix and acknowledgement that
> other drivers are fine with the change.
>
> Fixes: 3ba6b689893f ("eal/linux: fix eventfd leaks when enabling interrupts")
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
>  lib/eal/linux/eal_interrupts.c | 46 +++++++++-------------------------
>  1 file changed, 12 insertions(+), 34 deletions(-)
>

Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>

Thanks,
Maxime


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] eal/linux: revert eventfd leak fix for MSI-X
  2026-11-19 10:55 ` David Marchand
@ 2026-11-19 13:34   ` David Marchand
  0 siblings, 0 replies; 5+ messages in thread
From: David Marchand @ 2026-11-19 13:34 UTC (permalink / raw)
  To: dev; +Cc: iboukris, maxime.coquelin, thomas, Harman Kalra, Anatoly Burakov

On Thu, 19 Nov 2026 at 11:55, David Marchand <david.marchand@redhat.com> wrote:
>
> On Wed, 18 Nov 2026 at 16:30, David Marchand <david.marchand@redhat.com> wrote:
> >
> > The event FDs leak fix looks correct on the principle but it seems it
> > was too soon to take it, as it broke net/virtio and may have broken
> > other drivers.
> >
> > Revert this fix until we get a proper fix and acknowledgement that
> > other drivers are fine with the change.
> >
> > Fixes: 3ba6b689893f ("eal/linux: fix eventfd leaks when enabling interrupts")
> >
> > Signed-off-by: David Marchand <david.marchand@redhat.com>
>
> Recheck-request: github-robot: build

Looks like I did not use the right label.
The error on MSVC build does not make sense wrt to this patch, but trying again:
Recheck-request: github-robot


-- 
David Marchand


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] eal/linux: revert eventfd leak fix for MSI-X
  2026-11-19 13:11 ` Maxime Coquelin
@ 2026-11-19 15:01   ` Thomas Monjalon
  0 siblings, 0 replies; 5+ messages in thread
From: Thomas Monjalon @ 2026-11-19 15:01 UTC (permalink / raw)
  To: David Marchand
  Cc: dev, iboukris, Harman Kalra, Anatoly Burakov, Maxime Coquelin

19/11/2026 14:11, Maxime Coquelin:
> On Wed, Nov 18, 2026 at 4:30 PM David Marchand
> <david.marchand@redhat.com> wrote:
> >
> > The event FDs leak fix looks correct on the principle but it seems it
> > was too soon to take it, as it broke net/virtio and may have broken
> > other drivers.
> >
> > Revert this fix until we get a proper fix and acknowledgement that
> > other drivers are fine with the change.
> >
> > Fixes: 3ba6b689893f ("eal/linux: fix eventfd leaks when enabling interrupts")
> >
> > Signed-off-by: David Marchand <david.marchand@redhat.com>
> 
> Acked-by: Maxime Coquelin <maxime.coquelin@redhat.com>

Applied, thanks.




^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2026-11-19 15:01 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 15:29 [PATCH] eal/linux: revert eventfd leak fix for MSI-X David Marchand
2026-11-19 10:55 ` David Marchand
2026-11-19 13:34   ` David Marchand
2026-11-19 13:11 ` Maxime Coquelin
2026-11-19 15:01   ` Thomas Monjalon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox