2024-02-28 13:39:59

by Marcelo Tosatti

Subject: [PATCH -v2 net-next] net/core/dev.c: enable timestamp static key if CPU isolation is configured

For systems that use CPU isolation (via nohz_full), creating or destroying
a socket with timestamping (SOCK_TIMESTAMPING_RX_SOFTWARE and
SOF_TIMESTAMPING_OPT_TX_SWHW) might cause a
static key to be enabled/disabled. This in turn causes undesired
IPIs to isolated CPUs.

So enable the static key unconditionally, if CPU isolation is enabled,
thus avoiding the IPIs.

Signed-off-by: Marcelo Tosatti <[email protected]>

---
v2: mention SOF_TIMESTAMPING_OPT_TX_SWHW in the commit log (Willem de Bruijn / Paolo Abeni)

diff --git a/net/core/dev.c b/net/core/dev.c
index c588808be77f..15a32f5900e6 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -155,6 +155,7 @@
#include <net/netdev_rx_queue.h>
#include <net/page_pool/types.h>
#include <net/page_pool/helpers.h>
+#include <linux/sched/isolation.h>

#include "dev.h"
#include "net-sysfs.h"
@@ -11851,3 +11852,14 @@ static int __init net_dev_init(void)
}

subsys_initcall(net_dev_init);
+
+static int __init net_dev_late_init(void)
+{
+ /* avoid static key IPIs to isolated CPUs */
+ if (housekeeping_enabled(HK_TYPE_MISC))
+ net_enable_timestamp();
+
+ return 0;
+}
+
+late_initcall(net_dev_late_init);

2024-02-28 15:33:10

by Willem de Bruijn

Subject: Re: [PATCH -v2 net-next] net/core/dev.c: enable timestamp static key if CPU isolation is configured

Marcelo Tosatti wrote:
>
> For systems that use CPU isolation (via nohz_full), creating or destroying
> a socket with timestamping (SOCK_TIMESTAMPING_RX_SOFTWARE and
> SOF_TIMESTAMPING_OPT_TX_SWHW) might cause a

Again, SOF_TIMESTAMPING_OPT_TX_SWHW is irrelevant here.

See Documentation/networking/timestamping.rst for a definition of the
various flags.

> static key to be enabled/disabled. This in turn causes undesired
> IPIs to isolated CPUs.
>
> So enable the static key unconditionally, if CPU isolation is enabled,
> thus avoiding the IPIs.
>
> Signed-off-by: Marcelo Tosatti <[email protected]>
>
> ---
> v2: mention SOF_TIMESTAMPING_OPT_TX_SWHW in the commit log (Willem de Bruijn / Paolo Abeni)

I did not suggest that

The subject line is slightly corrupted: net-next -v2

>
> diff --git a/net/core/dev.c b/net/core/dev.c
> index c588808be77f..15a32f5900e6 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -155,6 +155,7 @@
> #include <net/netdev_rx_queue.h>
> #include <net/page_pool/types.h>
> #include <net/page_pool/helpers.h>
> +#include <linux/sched/isolation.h>
>
> #include "dev.h"
> #include "net-sysfs.h"
> @@ -11851,3 +11852,14 @@ static int __init net_dev_init(void)
> }
>
> subsys_initcall(net_dev_init);
> +
> +static int __init net_dev_late_init(void)
> +{
> + /* avoid static key IPIs to isolated CPUs */
> + if (housekeeping_enabled(HK_TYPE_MISC))
> + net_enable_timestamp();
> +
> + return 0;
> +}
> +
> +late_initcall(net_dev_late_init);
>