[PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Sivaprasad Tummala]

In `l3fwd-graph` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
Cc: ndabilpuram@marvell.com
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/l3fwd-graph/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
index 92cdaa1ebe..12908acbba 100644
--- a/examples/l3fwd-graph/main.c
+++ b/examples/l3fwd-graph/main.c
@@ -49,7 +49,6 @@
 #define RX_DESC_DEFAULT 1024
 #define TX_DESC_DEFAULT 1024
 
-#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
 #define MAX_RX_QUEUE_PER_PORT 128
 
 #define MAX_RX_QUEUE_PER_LCORE 16
@@ -1076,8 +1075,6 @@ main(int argc, char **argv)
 
 		nb_rx_queue = get_port_n_rx_queues(portid);
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
 		       nb_rx_queue, n_tx_queue);
 
-- 
2.43.0

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Nithin Dabilpuram]

Acked-by: Nithin Dabilpuram <ndabilpuram@marvell.com>

On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
<sivaprasad.tummala@amd.com> wrote:
>
> In `l3fwd-graph` application, Tx queues are configured per lcore
> to enable a lockless design and achieve optimal performance.
>
> The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> introduced an artificial constraint on the number of Tx queues
> and limited core-scaling performance.
>
> This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> redundant Tx queue check, allowing Tx queues to scale directly
> with the no. of lcores.
>
> Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> Cc: ndabilpuram@marvell.com
> Cc: stable@dpdk.org
>
> Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> ---
>  examples/l3fwd-graph/main.c | 3 ---
>  1 file changed, 3 deletions(-)
>
> diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> index 92cdaa1ebe..12908acbba 100644
> --- a/examples/l3fwd-graph/main.c
> +++ b/examples/l3fwd-graph/main.c
> @@ -49,7 +49,6 @@
>  #define RX_DESC_DEFAULT 1024
>  #define TX_DESC_DEFAULT 1024
>
> -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
>  #define MAX_RX_QUEUE_PER_PORT 128
>
>  #define MAX_RX_QUEUE_PER_LCORE 16
> @@ -1076,8 +1075,6 @@ main(int argc, char **argv)
>
>                 nb_rx_queue = get_port_n_rx_queues(portid);
>                 n_tx_queue = nb_lcores;
> -               if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
> -                       n_tx_queue = MAX_TX_QUEUE_PER_PORT;
>                 printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
>                        nb_rx_queue, n_tx_queue);
>
> --
> 2.43.0
>

RE: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Konstantin Ananyev]

> On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> <sivaprasad.tummala@amd.com> wrote:
> >
> > In `l3fwd-graph` application, Tx queues are configured per lcore
> > to enable a lockless design and achieve optimal performance.
> >
> > The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> > introduced an artificial constraint on the number of Tx queues
> > and limited core-scaling performance.
> >
> > This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> > redundant Tx queue check, allowing Tx queues to scale directly
> > with the no. of lcores.
> >
> > Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> > Cc: ndabilpuram@marvell.com
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> > ---
> >  examples/l3fwd-graph/main.c | 3 ---
> >  1 file changed, 3 deletions(-)
> >
> > diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> > index 92cdaa1ebe..12908acbba 100644
> > --- a/examples/l3fwd-graph/main.c
> > +++ b/examples/l3fwd-graph/main.c
> > @@ -49,7 +49,6 @@
> >  #define RX_DESC_DEFAULT 1024
> >  #define TX_DESC_DEFAULT 1024
> >
> > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> >  #define MAX_RX_QUEUE_PER_PORT 128

AFAIK, in the mainline we actually have:
#define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE

since:
commit 88256ed85338c572d73006e4c4530a52d3b477ff
Author: Harman Kalra <hkalra@marvell.com>
Date:   Tue Jan 12 23:54:46 2021 +0530

    examples/l3fwd: remove limitation on Tx queue count

What I am missing here?

> >
> >  #define MAX_RX_QUEUE_PER_LCORE 16
> > @@ -1076,8 +1075,6 @@ main(int argc, char **argv)
> >
> >                 nb_rx_queue = get_port_n_rx_queues(portid);
> >                 n_tx_queue = nb_lcores;
> > -               if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
> > -                       n_tx_queue = MAX_TX_QUEUE_PER_PORT;
> >                 printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
> >                        nb_rx_queue, n_tx_queue);
> >
> > --
> > 2.43.0
> >

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Tummala, Sivaprasad]

[-- Attachment #1: Type: text/plain, Size: 3555 bytes --]

[AMD Official Use Only - AMD Internal Distribution Only]

Hi Konstantin,

________________________________
From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
Sent: Monday, September 22, 2025 12:19 PM
To: Nithin Dabilpuram <nithind1988@gmail.com>; Tummala, Sivaprasad <Sivaprasad.Tummala@amd.com>
Cc: jerinj@marvell.com <jerinj@marvell.com>; kirankumark@marvell.com <kirankumark@marvell.com>; ndabilpuram@marvell.com <ndabilpuram@marvell.com>; yanzhirun_163@163.com <yanzhirun_163@163.com>; david.marchand@redhat.com <david.marchand@redhat.com>; ktraynor@redhat.com <ktraynor@redhat.com>; thomas@monjalon.net <thomas@monjalon.net>; konstantin.v.ananyev@yandex.ru <konstantin.v.ananyev@yandex.ru>; bruce.richardson@intel.com <bruce.richardson@intel.com>; maxime.coquelin@redhat.com <maxime.coquelin@redhat.com>; aconole@redhat.com <aconole@redhat.com>; dev@dpdk.org <dev@dpdk.org>; stable@dpdk.org <stable@dpdk.org>
Subject: RE: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.


> > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > <sivaprasad.tummala@amd.com> wrote:
> > >
> > > In `l3fwd-graph` application, Tx queues are configured per lcore
> > > to enable a lockless design and achieve optimal performance.
> > >
> > > The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> > > introduced an artificial constraint on the number of Tx queues
> > > and limited core-scaling performance.
> > >
> > > This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> > > redundant Tx queue check, allowing Tx queues to scale directly
> > > with the no. of lcores.
> > >
> > > Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> > > Cc: ndabilpuram@marvell.com
> > > Cc: stable@dpdk.org
> > >
> > > Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> > > ---
> > >  examples/l3fwd-graph/main.c | 3 ---
> > >  1 file changed, 3 deletions(-)
> > >
> > > diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> > > index 92cdaa1ebe..12908acbba 100644
> > > --- a/examples/l3fwd-graph/main.c
> > > +++ b/examples/l3fwd-graph/main.c
> > > @@ -49,7 +49,6 @@
> > >  #define RX_DESC_DEFAULT 1024
> > >  #define TX_DESC_DEFAULT 1024
> > >
> > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > >  #define MAX_RX_QUEUE_PER_PORT 128
>
> AFAIK, in the mainline we actually have:
> #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
>
In l3fwd-graph app, this change is not available and instead we have
#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS

> since:
> commit 88256ed85338c572d73006e4c4530a52d3b477ff
> Author: Harman Kalra <hkalra@marvell.com>
> Date:   Tue Jan 12 23:54:46 2021 +0530
>
>     examples/l3fwd: remove limitation on Tx queue count
>
> What I am missing here?
This patch marked here was fixing l3fwd app and not l3fwd-graph

>
> > >
> > >  #define MAX_RX_QUEUE_PER_LCORE 16
> > > @@ -1076,8 +1075,6 @@ main(int argc, char **argv)
> > >
> > >                 nb_rx_queue = get_port_n_rx_queues(portid);
> > >                 n_tx_queue = nb_lcores;
> > > -               if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
> > > -                       n_tx_queue = MAX_TX_QUEUE_PER_PORT;
> > >                 printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
> > >                        nb_rx_queue, n_tx_queue);
> > >
> > > --
> > > 2.43.0
> > >

[-- Attachment #2: Type: text/html, Size: 6655 bytes --]

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Thomas Monjalon]

06/10/2025 10:58, Tummala, Sivaprasad:
> From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> > > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > > <sivaprasad.tummala@amd.com> wrote:
> > > >
> > > > In `l3fwd-graph` application, Tx queues are configured per lcore
> > > > to enable a lockless design and achieve optimal performance.
> > > >
> > > > The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> > > > introduced an artificial constraint on the number of Tx queues
> > > > and limited core-scaling performance.
> > > >
> > > > This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> > > > redundant Tx queue check, allowing Tx queues to scale directly
> > > > with the no. of lcores.
> > > >
> > > > Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> > > > Cc: ndabilpuram@marvell.com
> > > > Cc: stable@dpdk.org
> > > >
> > > > Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> > > > ---
> > > >  examples/l3fwd-graph/main.c | 3 ---
> > > >  1 file changed, 3 deletions(-)
> > > >
> > > > diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> > > > index 92cdaa1ebe..12908acbba 100644
> > > > --- a/examples/l3fwd-graph/main.c
> > > > +++ b/examples/l3fwd-graph/main.c
> > > > @@ -49,7 +49,6 @@
> > > >  #define RX_DESC_DEFAULT 1024
> > > >  #define TX_DESC_DEFAULT 1024
> > > >
> > > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > >  #define MAX_RX_QUEUE_PER_PORT 128
> >
> > AFAIK, in the mainline we actually have:
> > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
> >
> In l3fwd-graph app, this change is not available and instead we have
> #define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> 
> > since:
> > commit 88256ed85338c572d73006e4c4530a52d3b477ff
> > Author: Harman Kalra <hkalra@marvell.com>
> > Date:   Tue Jan 12 23:54:46 2021 +0530
> >
> >     examples/l3fwd: remove limitation on Tx queue count
> >
> > What I am missing here?
> This patch marked here was fixing l3fwd app and not l3fwd-graph

Why not applying the same change to both examples?

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Tummala, Sivaprasad]

[-- Attachment #1: Type: text/plain, Size: 3453 bytes --]

[AMD Official Use Only - AMD Internal Distribution Only]

Hi Thomas,

________________________________
From: Thomas Monjalon <thomas@monjalon.net>
Sent: Tuesday, November 18, 2025 10:39 PM
To: Konstantin Ananyev <konstantin.ananyev@huawei.com>; Tummala, Sivaprasad <Sivaprasad.Tummala@amd.com>
Cc: Nithin Dabilpuram <nithind1988@gmail.com>; stable@dpdk.org <stable@dpdk.org>; jerinj@marvell.com <jerinj@marvell.com>; kirankumark@marvell.com <kirankumark@marvell.com>; ndabilpuram@marvell.com <ndabilpuram@marvell.com>; yanzhirun_163@163.com <yanzhirun_163@163.com>; david.marchand@redhat.com <david.marchand@redhat.com>; ktraynor@redhat.com <ktraynor@redhat.com>; konstantin.v.ananyev@yandex.ru <konstantin.v.ananyev@yandex.ru>; bruce.richardson@intel.com <bruce.richardson@intel.com>; maxime.coquelin@redhat.com <maxime.coquelin@redhat.com>; aconole@redhat.com <aconole@redhat.com>; dev@dpdk.org <dev@dpdk.org>; stable@dpdk.org <stable@dpdk.org>
Subject: Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.


06/10/2025 10:58, Tummala, Sivaprasad:
> From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> > > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > > <sivaprasad.tummala@amd.com> wrote:
> > > >
> > > > In `l3fwd-graph` application, Tx queues are configured per lcore
> > > > to enable a lockless design and achieve optimal performance.
> > > >
> > > > The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> > > > introduced an artificial constraint on the number of Tx queues
> > > > and limited core-scaling performance.
> > > >
> > > > This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> > > > redundant Tx queue check, allowing Tx queues to scale directly
> > > > with the no. of lcores.
> > > >
> > > > Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> > > > Cc: ndabilpuram@marvell.com
> > > > Cc: stable@dpdk.org
> > > >
> > > > Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> > > > ---
> > > >  examples/l3fwd-graph/main.c | 3 ---
> > > >  1 file changed, 3 deletions(-)
> > > >
> > > > diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> > > > index 92cdaa1ebe..12908acbba 100644
> > > > --- a/examples/l3fwd-graph/main.c
> > > > +++ b/examples/l3fwd-graph/main.c
> > > > @@ -49,7 +49,6 @@
> > > >  #define RX_DESC_DEFAULT 1024
> > > >  #define TX_DESC_DEFAULT 1024
> > > >
> > > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > >  #define MAX_RX_QUEUE_PER_PORT 128
> >
> > AFAIK, in the mainline we actually have:
> > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
> >
> In l3fwd-graph app, this change is not available and instead we have
>> #define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
>>
>> > since:
>> > commit 88256ed85338c572d73006e4c4530a52d3b477ff
>> > Author: Harman Kalra <hkalra@marvell.com>
>> > Date:   Tue Jan 12 23:54:46 2021 +0530
>> >
>> >     examples/l3fwd: remove limitation on Tx queue count
>> >
>> > What I am missing here?
>> This patch marked here was fixing l3fwd app and not l3fwd-graph

> Why not applying the same change to both examples?
Yes, that's what the patch is intended for to fix l3fwd-graph and tx queues will scale with lcores and limited by RTE_MAX_LCORES.




[-- Attachment #2: Type: text/html, Size: 5531 bytes --]

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Thomas Monjalon]

19/11/2025 13:08, Tummala, Sivaprasad:
> From: Thomas Monjalon <thomas@monjalon.net>
> 06/10/2025 10:58, Tummala, Sivaprasad:
> > From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> > > > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > > > <sivaprasad.tummala@amd.com> wrote:
> > > > >
> > > > > In `l3fwd-graph` application, Tx queues are configured per lcore
> > > > > to enable a lockless design and achieve optimal performance.
> > > > >
> > > > > The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> > > > > introduced an artificial constraint on the number of Tx queues
> > > > > and limited core-scaling performance.
> > > > >
> > > > > This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> > > > > redundant Tx queue check, allowing Tx queues to scale directly
> > > > > with the no. of lcores.
> > > > >
> > > > > Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> > > > > Cc: ndabilpuram@marvell.com
> > > > > Cc: stable@dpdk.org
> > > > >
> > > > > Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> > > > > ---
> > > > >  examples/l3fwd-graph/main.c | 3 ---
> > > > >  1 file changed, 3 deletions(-)
> > > > >
> > > > > diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> > > > > index 92cdaa1ebe..12908acbba 100644
> > > > > --- a/examples/l3fwd-graph/main.c
> > > > > +++ b/examples/l3fwd-graph/main.c
> > > > > @@ -49,7 +49,6 @@
> > > > >  #define RX_DESC_DEFAULT 1024
> > > > >  #define TX_DESC_DEFAULT 1024
> > > > >
> > > > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > > >  #define MAX_RX_QUEUE_PER_PORT 128
> > >
> > > AFAIK, in the mainline we actually have:
> > > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
> > >
> > In l3fwd-graph app, this change is not available and instead we have
> >> #define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> >>
> >> > since:
> >> > commit 88256ed85338c572d73006e4c4530a52d3b477ff
> >> > Author: Harman Kalra <hkalra@marvell.com>
> >> > Date:   Tue Jan 12 23:54:46 2021 +0530
> >> >
> >> >     examples/l3fwd: remove limitation on Tx queue count
> >> >
> >> > What I am missing here?
> >> This patch marked here was fixing l3fwd app and not l3fwd-graph
> 
> > Why not applying the same change to both examples?
> Yes, that's what the patch is intended for to fix l3fwd-graph and tx queues will scale with lcores and limited by RTE_MAX_LCORES.

But it is not done the same way.
Here you remove MAX_TX_QUEUE_PER_PORT.
Do you want to do the same in l3fwd?

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Tummala, Sivaprasad]

[-- Attachment #1: Type: text/plain, Size: 3893 bytes --]

[AMD Official Use Only - AMD Internal Distribution Only]

Hi Thomas,

________________________________
From: Thomas Monjalon <thomas@monjalon.net>
Sent: Wednesday, November 19, 2025 6:27 PM
To: Tummala, Sivaprasad <Sivaprasad.Tummala@amd.com>
Cc: Konstantin Ananyev <konstantin.ananyev@huawei.com>; Nithin Dabilpuram <nithind1988@gmail.com>; stable@dpdk.org <stable@dpdk.org>; jerinj@marvell.com <jerinj@marvell.com>; kirankumark@marvell.com <kirankumark@marvell.com>; ndabilpuram@marvell.com <ndabilpuram@marvell.com>; yanzhirun_163@163.com <yanzhirun_163@163.com>; david.marchand@redhat.com <david.marchand@redhat.com>; ktraynor@redhat.com <ktraynor@redhat.com>; konstantin.v.ananyev@yandex.ru <konstantin.v.ananyev@yandex.ru>; bruce.richardson@intel.com <bruce.richardson@intel.com>; maxime.coquelin@redhat.com <maxime.coquelin@redhat.com>; aconole@redhat.com <aconole@redhat.com>; dev@dpdk.org <dev@dpdk.org>; stable@dpdk.org <stable@dpdk.org>
Subject: Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.


19/11/2025 13:08, Tummala, Sivaprasad:
> From: Thomas Monjalon <thomas@monjalon.net>
> 06/10/2025 10:58, Tummala, Sivaprasad:
> > From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> > > > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > > > <sivaprasad.tummala@amd.com> wrote:
> > > > >
> > > > > In `l3fwd-graph` application, Tx queues are configured per lcore
> > > > > to enable a lockless design and achieve optimal performance.
> > > > >
> > > > > The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
> > > > > introduced an artificial constraint on the number of Tx queues
> > > > > and limited core-scaling performance.
> > > > >
> > > > > This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
> > > > > redundant Tx queue check, allowing Tx queues to scale directly
> > > > > with the no. of lcores.
> > > > >
> > > > > Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
> > > > > Cc: ndabilpuram@marvell.com
> > > > > Cc: stable@dpdk.org
> > > > >
> > > > > Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
> > > > > ---
> > > > >  examples/l3fwd-graph/main.c | 3 ---
> > > > >  1 file changed, 3 deletions(-)
> > > > >
> > > > > diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
> > > > > index 92cdaa1ebe..12908acbba 100644
> > > > > --- a/examples/l3fwd-graph/main.c
> > > > > +++ b/examples/l3fwd-graph/main.c
> > > > > @@ -49,7 +49,6 @@
> > > > >  #define RX_DESC_DEFAULT 1024
> > > > >  #define TX_DESC_DEFAULT 1024
> > > > >
> > > > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > > >  #define MAX_RX_QUEUE_PER_PORT 128
> > >
> > > AFAIK, in the mainline we actually have:
> > > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
> > >
> > In l3fwd-graph app, this change is not available and instead we have
> >> #define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> >>
> >> > since:
> >> > commit 88256ed85338c572d73006e4c4530a52d3b477ff
> >> > Author: Harman Kalra <hkalra@marvell.com>
> >> > Date:   Tue Jan 12 23:54:46 2021 +0530
> >> >
> >> >     examples/l3fwd: remove limitation on Tx queue count
> >> >
> >> > What I am missing here?
>> >> This patch marked here was fixing l3fwd app and not l3fwd-graph
>>
>> > Why not applying the same change to both examples?
>> Yes, that's what the patch is intended for to fix l3fwd-graph and tx queues will scale with lcores and limited by RTE_MAX_LCORES.

>But it is not done the same way.
> Here you remove MAX_TX_QUEUE_PER_PORT.
> Do you want to do the same in l3fwd?
Yes, it's better to fix the same in l3fwd as "MAX_TX_QUEUE_PER_PORT" is redundant.
I can submit a separate patch for l3fwd.


[-- Attachment #2: Type: text/html, Size: 6746 bytes --]

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Thomas Monjalon]

19/11/2025 14:56, Tummala, Sivaprasad:
> From: Thomas Monjalon <thomas@monjalon.net>
> 19/11/2025 13:08, Tummala, Sivaprasad:
> > From: Thomas Monjalon <thomas@monjalon.net>
> > 06/10/2025 10:58, Tummala, Sivaprasad:
> > > From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> > > > > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > > > > <sivaprasad.tummala@amd.com> wrote:
> > > > > > --- a/examples/l3fwd-graph/main.c
> > > > > > +++ b/examples/l3fwd-graph/main.c
> > > > > > @@ -49,7 +49,6 @@
> > > > > >  #define RX_DESC_DEFAULT 1024
> > > > > >  #define TX_DESC_DEFAULT 1024
> > > > > >
> > > > > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > > > >  #define MAX_RX_QUEUE_PER_PORT 128
> > > >
> > > > AFAIK, in the mainline we actually have:
> > > > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
> > > >
> > > In l3fwd-graph app, this change is not available and instead we have
> > >> #define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > >>
> > >> > since:
> > >> > commit 88256ed85338c572d73006e4c4530a52d3b477ff
> > >> > Author: Harman Kalra <hkalra@marvell.com>
> > >> > Date:   Tue Jan 12 23:54:46 2021 +0530
> > >> >
> > >> >     examples/l3fwd: remove limitation on Tx queue count
> > >> >
> > >> > What I am missing here?
> >> >> This patch marked here was fixing l3fwd app and not l3fwd-graph
> >>
> >> > Why not applying the same change to both examples?
> >> Yes, that's what the patch is intended for to fix l3fwd-graph and tx queues will scale with lcores and limited by RTE_MAX_LCORES.
> 
> >But it is not done the same way.
> > Here you remove MAX_TX_QUEUE_PER_PORT.
> > Do you want to do the same in l3fwd?
> Yes, it's better to fix the same in l3fwd as "MAX_TX_QUEUE_PER_PORT" is redundant.
> I can submit a separate patch for l3fwd.

Better to fix in a single patch.
Please check if there are similar issue in other examples.

Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

[Tummala, Sivaprasad]

[-- Attachment #1: Type: text/plain, Size: 3204 bytes --]

[AMD Official Use Only - AMD Internal Distribution Only]


________________________________
From: Thomas Monjalon <thomas@monjalon.net>
Sent: Wednesday, November 19, 2025 8:25 PM
To: Tummala, Sivaprasad <Sivaprasad.Tummala@amd.com>
Cc: Konstantin Ananyev <konstantin.ananyev@huawei.com>; Nithin Dabilpuram <nithind1988@gmail.com>; stable@dpdk.org <stable@dpdk.org>; jerinj@marvell.com <jerinj@marvell.com>; kirankumark@marvell.com <kirankumark@marvell.com>; ndabilpuram@marvell.com <ndabilpuram@marvell.com>; yanzhirun_163@163.com <yanzhirun_163@163.com>; david.marchand@redhat.com <david.marchand@redhat.com>; ktraynor@redhat.com <ktraynor@redhat.com>; konstantin.v.ananyev@yandex.ru <konstantin.v.ananyev@yandex.ru>; bruce.richardson@intel.com <bruce.richardson@intel.com>; maxime.coquelin@redhat.com <maxime.coquelin@redhat.com>; aconole@redhat.com <aconole@redhat.com>; dev@dpdk.org <dev@dpdk.org>; stable@dpdk.org <stable@dpdk.org>
Subject: Re: [PATCH] examples/l3fwd-graph: remove redundant Tx queue limit

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.

19/11/2025 14:56, Tummala, Sivaprasad:
> > From: Thomas Monjalon <thomas@monjalon.net>
> > 19/11/2025 13:08, Tummala, Sivaprasad:
> > > From: Thomas Monjalon <thomas@monjalon.net>
> > > 06/10/2025 10:58, Tummala, Sivaprasad:
> > > > From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> > > > > > On Mon, Sep 1, 2025 at 11:39 PM Sivaprasad Tummala
> > > > > > <sivaprasad.tummala@amd.com> wrote:
> > > > > > > --- a/examples/l3fwd-graph/main.c
> > > > > > > +++ b/examples/l3fwd-graph/main.c
> > > > > > > @@ -49,7 +49,6 @@
> > > > > > >  #define RX_DESC_DEFAULT 1024
> > > > > > >  #define TX_DESC_DEFAULT 1024
> > > > > > >
> > > > > > > -#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > > > > >  #define MAX_RX_QUEUE_PER_PORT 128
> > > > >
> > > > > AFAIK, in the mainline we actually have:
> > > > > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
> > > > >
> > > > In l3fwd-graph app, this change is not available and instead we have
> > > > > #define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
> > > > >
> > > > > > since:
> > > > > > commit 88256ed85338c572d73006e4c4530a52d3b477ff
> > > > > > Author: Harman Kalra <hkalra@marvell.com>
> > > > > > Date:   Tue Jan 12 23:54:46 2021 +0530
> > > > > >
> > > > > >     examples/l3fwd: remove limitation on Tx queue count
> > > > > >
> > > > > > What I am missing here?
> > > > > > This patch marked here was fixing l3fwd app and not l3fwd-graph
> > > >
> > > > > Why not applying the same change to both examples?
> > > > Yes, that's what the patch is intended for to fix l3fwd-graph and tx queues will scale with lcores and limited by RTE_MAX_LCORES.
> >
> > >But it is not done the same way.
> > > Here you remove MAX_TX_QUEUE_PER_PORT.
> > > Do you want to do the same in l3fwd?
> > Yes, it's better to fix the same in l3fwd as "MAX_TX_QUEUE_PER_PORT" is redundant.
> > I can submit a separate patch for l3fwd.
>
> Better to fix in a single patch.
> Please check if there are similar issue in other examples.
Sure, I will do the needful.


[-- Attachment #2: Type: text/html, Size: 10323 bytes --]

[PATCH v2 1/5] examples/l3fwd-graph: remove redundant Tx queue limit

[Sivaprasad Tummala]

In `l3fwd-graph` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
Cc: ndabilpuram@marvell.com
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/l3fwd-graph/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
index 92cdaa1ebe..12908acbba 100644
--- a/examples/l3fwd-graph/main.c
+++ b/examples/l3fwd-graph/main.c
@@ -49,7 +49,6 @@
 #define RX_DESC_DEFAULT 1024
 #define TX_DESC_DEFAULT 1024
 
-#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
 #define MAX_RX_QUEUE_PER_PORT 128
 
 #define MAX_RX_QUEUE_PER_LCORE 16
@@ -1076,8 +1075,6 @@ main(int argc, char **argv)
 
 		nb_rx_queue = get_port_n_rx_queues(portid);
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
 		       nb_rx_queue, n_tx_queue);
 
-- 
2.43.0

[PATCH v2 2/5] examples/l3fwd: remove redundant Tx queue limit

[Sivaprasad Tummala]

In `l3fwd` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/l3fwd/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c
index ae3b4f6439..bdf59b5399 100644
--- a/examples/l3fwd/main.c
+++ b/examples/l3fwd/main.c
@@ -49,7 +49,6 @@
 #include "l3fwd_event.h"
 #include "l3fwd_route.h"
 
-#define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
 #define MAX_RX_QUEUE_PER_PORT 128
 
 #define MAX_LCORE_PARAMS 1024
@@ -1341,8 +1340,6 @@ l3fwd_poll_resource_setup(void)
 
 		nb_rx_queue = get_port_n_rx_queues(portid);
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
 			nb_rx_queue, (unsigned)n_tx_queue );
 
-- 
2.43.0

[PATCH v2 3/5] examples/ip_fragmentation: remove redundant Tx queue limit

[Sivaprasad Tummala]

In `ip_fragmentation` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/ip_fragmentation/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/ip_fragmentation/main.c b/examples/ip_fragmentation/main.c
index 1f84102844..9be2f72983 100644
--- a/examples/ip_fragmentation/main.c
+++ b/examples/ip_fragmentation/main.c
@@ -134,7 +134,6 @@ struct rx_queue {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 struct __rte_cache_aligned lcore_queue_conf {
 	uint16_t n_rx_queue;
 	uint16_t tx_queue_id[RTE_MAX_ETHPORTS];
@@ -948,8 +947,6 @@ main(int argc, char **argv)
 		fflush(stdout);
 
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		ret = rte_eth_dev_configure(portid, 1, (uint16_t)n_tx_queue,
 					    &local_port_conf);
 		if (ret < 0) {
-- 
2.43.0

[PATCH v2 4/5] examples/ip_reassembly: remove redundant Tx queue limit

[Sivaprasad Tummala]

In `ip_reassembly` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: cc8f4d020c0b ("examples/ip_reassembly: initial import")
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/ip_reassembly/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/ip_reassembly/main.c b/examples/ip_reassembly/main.c
index 17ae76d4ba..141a55462f 100644
--- a/examples/ip_reassembly/main.c
+++ b/examples/ip_reassembly/main.c
@@ -145,7 +145,6 @@ struct tx_lcore_stat {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 #define MAX_RX_QUEUE_PER_PORT 128
 
 struct __rte_cache_aligned lcore_queue_conf {
@@ -1091,8 +1090,6 @@ main(int argc, char **argv)
 		fflush(stdout);
 
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		if (dev_info.tx_offload_capa & RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE)
 			local_port_conf.txmode.offloads |=
 				RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE;
-- 
2.43.0

[PATCH v2 5/5] examples/ipv4_multicast: remove redundant Tx queue limit

[Sivaprasad Tummala]

In `ipv4_multicast` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/ipv4_multicast/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/ipv4_multicast/main.c b/examples/ipv4_multicast/main.c
index 1eed645d02..b698db21de 100644
--- a/examples/ipv4_multicast/main.c
+++ b/examples/ipv4_multicast/main.c
@@ -97,7 +97,6 @@ struct mbuf_table {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 struct __rte_cache_aligned lcore_queue_conf {
 	uint64_t tx_tsc;
 	uint16_t n_rx_queue;
@@ -736,8 +735,6 @@ main(int argc, char **argv)
 		fflush(stdout);
 
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 
 		ret = rte_eth_dev_configure(portid, 1, (uint16_t)n_tx_queue,
 					    &local_port_conf);
-- 
2.43.0

[PATCH v3 0/6] examples: remove MAX_TX_QUEUE_PER_PORT

[Stephen Hemminger]

This is an old set of patches that removes the use of MAX_TX_QUEUE_PER_PORT
in all the examples.

Sivaprasad Tummala (5):
  examples/l3fwd-graph: remove redundant Tx queue limit
  examples/l3fwd: remove redundant Tx queue limit
  examples/ip_fragmentation: remove redundant Tx queue limit
  examples/ip_reassembly: remove redundant Tx queue limit
  examples/ipv4_multicast: remove redundant Tx queue limit

Stephen Hemminger (1):
  examples: remove unused define

 examples/ip_fragmentation/main.c      | 3 ---
 examples/ip_reassembly/main.c         | 3 ---
 examples/ipv4_multicast/main.c        | 3 ---
 examples/l2fwd-crypto/main.c          | 1 -
 examples/l2fwd-event/l2fwd_common.h   | 1 -
 examples/l2fwd-jobstats/main.c        | 1 -
 examples/l2fwd-keepalive/main.c       | 1 -
 examples/l2fwd-macsec/main.c          | 1 -
 examples/l2fwd/main.c                 | 2 +-
 examples/l3fwd-graph/main.c           | 3 ---
 examples/l3fwd-power/main.c           | 1 -
 examples/l3fwd/main.c                 | 3 ---
 examples/link_status_interrupt/main.c | 1 -
 13 files changed, 1 insertion(+), 23 deletions(-)

-- 
2.53.0

[PATCH v3 1/6] examples/l3fwd-graph: remove redundant Tx queue limit

[Stephen Hemminger]

From: Sivaprasad Tummala <sivaprasad.tummala@amd.com>

In `l3fwd-graph` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: 08bd1a174461 ("examples/l3fwd-graph: add graph-based l3fwd skeleton")
Cc: ndabilpuram@marvell.com
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/l3fwd-graph/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/l3fwd-graph/main.c b/examples/l3fwd-graph/main.c
index 01c65b0abd..5f89286dce 100644
--- a/examples/l3fwd-graph/main.c
+++ b/examples/l3fwd-graph/main.c
@@ -49,7 +49,6 @@
 #define RX_DESC_DEFAULT 1024
 #define TX_DESC_DEFAULT 1024
 
-#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
 #define MAX_RX_QUEUE_PER_PORT 128
 
 #define MAX_RX_QUEUE_PER_LCORE 16
@@ -1076,8 +1075,6 @@ main(int argc, char **argv)
 
 		nb_rx_queue = get_port_n_rx_queues(portid);
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
 		       nb_rx_queue, n_tx_queue);
 
-- 
2.53.0

[PATCH v3 2/6] examples/l3fwd: remove redundant Tx queue limit

[Stephen Hemminger]

From: Sivaprasad Tummala <sivaprasad.tummala@amd.com>

In `l3fwd` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_LCORE`
which is dead code.

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/l3fwd/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c
index df035b508c..6866811526 100644
--- a/examples/l3fwd/main.c
+++ b/examples/l3fwd/main.c
@@ -48,7 +48,6 @@
 #include "l3fwd_event.h"
 #include "l3fwd_route.h"
 
-#define MAX_TX_QUEUE_PER_PORT RTE_MAX_LCORE
 #define MAX_RX_QUEUE_PER_PORT 128
 
 #define MAX_LCORE_PARAMS 1024
@@ -1373,8 +1372,6 @@ l3fwd_poll_resource_setup(void)
 
 		nb_rx_queue = get_port_n_rx_queues(portid);
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		printf("Creating queues: nb_rxq=%d nb_txq=%u... ",
 			nb_rx_queue, (unsigned)n_tx_queue );
 
-- 
2.53.0

[PATCH v3 3/6] examples/ip_fragmentation: remove redundant Tx queue limit

[Stephen Hemminger]

From: Sivaprasad Tummala <sivaprasad.tummala@amd.com>

In `ip_fragmentation` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/ip_fragmentation/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/ip_fragmentation/main.c b/examples/ip_fragmentation/main.c
index 2180682373..132550e497 100644
--- a/examples/ip_fragmentation/main.c
+++ b/examples/ip_fragmentation/main.c
@@ -133,7 +133,6 @@ struct rx_queue {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 struct __rte_cache_aligned lcore_queue_conf {
 	uint16_t n_rx_queue;
 	uint16_t tx_queue_id[RTE_MAX_ETHPORTS];
@@ -947,8 +946,6 @@ main(int argc, char **argv)
 		fflush(stdout);
 
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		ret = rte_eth_dev_configure(portid, 1, (uint16_t)n_tx_queue,
 					    &local_port_conf);
 		if (ret < 0) {
-- 
2.53.0

[PATCH v3 4/6] examples/ip_reassembly: remove redundant Tx queue limit

[Stephen Hemminger]

From: Sivaprasad Tummala <sivaprasad.tummala@amd.com>

In `ip_reassembly` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: cc8f4d020c0b ("examples/ip_reassembly: initial import")
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/ip_reassembly/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/ip_reassembly/main.c b/examples/ip_reassembly/main.c
index 520fbea1c2..0814d47a66 100644
--- a/examples/ip_reassembly/main.c
+++ b/examples/ip_reassembly/main.c
@@ -144,7 +144,6 @@ struct tx_lcore_stat {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 #define MAX_RX_QUEUE_PER_PORT 128
 
 struct __rte_cache_aligned lcore_queue_conf {
@@ -1097,8 +1096,6 @@ main(int argc, char **argv)
 		fflush(stdout);
 
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 		if (dev_info.tx_offload_capa & RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE)
 			local_port_conf.txmode.offloads |=
 				RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE;
-- 
2.53.0

[PATCH v3 5/6] examples/ipv4_multicast: remove redundant Tx queue limit

[Stephen Hemminger]

From: Sivaprasad Tummala <sivaprasad.tummala@amd.com>

In `ipv4_multicast` application, Tx queues are configured per lcore
to enable a lockless design and achieve optimal performance.

The `MAX_TX_QUEUE_PER_PORT` macro, defined as `RTE_MAX_ETHPORTS`,
introduced an artificial constraint on the number of Tx queues
and limited core-scaling performance.

This patch removes the unused `MAX_TX_QUEUE_PER_PORT` macro and
redundant Tx queue check, allowing Tx queues to scale directly
with the no. of lcores.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org

Signed-off-by: Sivaprasad Tummala <sivaprasad.tummala@amd.com>
---
 examples/ipv4_multicast/main.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/examples/ipv4_multicast/main.c b/examples/ipv4_multicast/main.c
index bd4c3f335b..1cb621cb8b 100644
--- a/examples/ipv4_multicast/main.c
+++ b/examples/ipv4_multicast/main.c
@@ -96,7 +96,6 @@ struct mbuf_table {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 struct __rte_cache_aligned lcore_queue_conf {
 	uint64_t tx_tsc;
 	uint16_t n_rx_queue;
@@ -735,8 +734,6 @@ main(int argc, char **argv)
 		fflush(stdout);
 
 		n_tx_queue = nb_lcores;
-		if (n_tx_queue > MAX_TX_QUEUE_PER_PORT)
-			n_tx_queue = MAX_TX_QUEUE_PER_PORT;
 
 		ret = rte_eth_dev_configure(portid, 1, (uint16_t)n_tx_queue,
 					    &local_port_conf);
-- 
2.53.0

[PATCH v3 6/6] examples: remove unused define

[Stephen Hemminger]

The #define MAX_TX_QUEUE_PER_PORT was copy/pasted across
all the examples but never used.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
 examples/l2fwd-crypto/main.c          | 1 -
 examples/l2fwd-event/l2fwd_common.h   | 1 -
 examples/l2fwd-jobstats/main.c        | 1 -
 examples/l2fwd-keepalive/main.c       | 1 -
 examples/l2fwd-macsec/main.c          | 1 -
 examples/l2fwd/main.c                 | 2 +-
 examples/l3fwd-power/main.c           | 1 -
 examples/link_status_interrupt/main.c | 1 -
 8 files changed, 1 insertion(+), 8 deletions(-)

diff --git a/examples/l2fwd-crypto/main.c b/examples/l2fwd-crypto/main.c
index ff189b5fab..aab356aa6a 100644
--- a/examples/l2fwd-crypto/main.c
+++ b/examples/l2fwd-crypto/main.c
@@ -100,7 +100,6 @@ struct op_buffer {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 
 enum l2fwd_crypto_xform_chain {
 	L2FWD_CRYPTO_CIPHER_HASH,
diff --git a/examples/l2fwd-event/l2fwd_common.h b/examples/l2fwd-event/l2fwd_common.h
index f4f1c45cd1..53774e0fe6 100644
--- a/examples/l2fwd-event/l2fwd_common.h
+++ b/examples/l2fwd-event/l2fwd_common.h
@@ -43,7 +43,6 @@
 
 #define MAX_PKT_BURST 32
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 
 #define RX_DESC_DEFAULT 1024
 #define TX_DESC_DEFAULT 1024
diff --git a/examples/l2fwd-jobstats/main.c b/examples/l2fwd-jobstats/main.c
index f2c284cebb..856914e9f2 100644
--- a/examples/l2fwd-jobstats/main.c
+++ b/examples/l2fwd-jobstats/main.c
@@ -64,7 +64,6 @@ static uint32_t l2fwd_dst_ports[RTE_MAX_ETHPORTS];
 static unsigned int l2fwd_rx_queue_per_lcore = 1;
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 /* List of queues to be polled for given lcore. 8< */
 struct __rte_cache_aligned lcore_queue_conf {
 	unsigned n_rx_port;
diff --git a/examples/l2fwd-keepalive/main.c b/examples/l2fwd-keepalive/main.c
index 12ca60c3e4..fde997f778 100644
--- a/examples/l2fwd-keepalive/main.c
+++ b/examples/l2fwd-keepalive/main.c
@@ -65,7 +65,6 @@ static uint32_t l2fwd_dst_ports[RTE_MAX_ETHPORTS];
 static unsigned int l2fwd_rx_queue_per_lcore = 1;
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 struct __rte_cache_aligned lcore_queue_conf {
 	unsigned n_rx_port;
 	unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE];
diff --git a/examples/l2fwd-macsec/main.c b/examples/l2fwd-macsec/main.c
index 98763440bc..e40e5b6284 100644
--- a/examples/l2fwd-macsec/main.c
+++ b/examples/l2fwd-macsec/main.c
@@ -83,7 +83,6 @@ static uint16_t nb_port_pair_params;
 static unsigned int l2fwd_rx_queue_per_lcore = 1;
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 /* List of queues to be polled for a given lcore. 8< */
 struct __rte_cache_aligned lcore_queue_conf {
 	unsigned int n_rx_port;
diff --git a/examples/l2fwd/main.c b/examples/l2fwd/main.c
index 59ea3172ae..1c4a89ae90 100644
--- a/examples/l2fwd/main.c
+++ b/examples/l2fwd/main.c
@@ -80,7 +80,7 @@ static uint16_t nb_port_pair_params;
 static unsigned int l2fwd_rx_queue_per_lcore = 1;
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
+
 /* List of queues to be polled for a given lcore. 8< */
 struct __rte_cache_aligned lcore_queue_conf {
 	unsigned n_rx_port;
diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
index ff0e61e639..0915ed5658 100644
--- a/examples/l3fwd-power/main.c
+++ b/examples/l3fwd-power/main.c
@@ -221,7 +221,6 @@ struct __rte_cache_aligned lcore_rx_queue {
 };
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT RTE_MAX_ETHPORTS
 #define MAX_RX_QUEUE_PER_PORT 128
 
 #define MAX_RX_QUEUE_INTERRUPT_PER_PORT 16
diff --git a/examples/link_status_interrupt/main.c b/examples/link_status_interrupt/main.c
index aa33e71d7a..52bfb3c35e 100644
--- a/examples/link_status_interrupt/main.c
+++ b/examples/link_status_interrupt/main.c
@@ -63,7 +63,6 @@ static unsigned lsi_dst_ports[RTE_MAX_ETHPORTS] = {0};
 #define MAX_PKT_BURST 32
 
 #define MAX_RX_QUEUE_PER_LCORE 16
-#define MAX_TX_QUEUE_PER_PORT 16
 /* List of queues must be polled for a give lcore. 8< */
 struct __rte_cache_aligned lcore_queue_conf {
 	unsigned n_rx_port;
-- 
2.53.0

Re: [PATCH v3 0/6] examples: remove MAX_TX_QUEUE_PER_PORT

[Bruce Richardson]

On Thu, Jun 25, 2026 at 10:46:50AM -0700, Stephen Hemminger wrote:
> This is an old set of patches that removes the use of MAX_TX_QUEUE_PER_PORT
> in all the examples.
> 
> Sivaprasad Tummala (5):
>   examples/l3fwd-graph: remove redundant Tx queue limit
>   examples/l3fwd: remove redundant Tx queue limit
>   examples/ip_fragmentation: remove redundant Tx queue limit
>   examples/ip_reassembly: remove redundant Tx queue limit
>   examples/ipv4_multicast: remove redundant Tx queue limit
> 
> Stephen Hemminger (1):
>   examples: remove unused define
> 
>  examples/ip_fragmentation/main.c      | 3 ---
>  examples/ip_reassembly/main.c         | 3 ---
>  examples/ipv4_multicast/main.c        | 3 ---
>  examples/l2fwd-crypto/main.c          | 1 -
>  examples/l2fwd-event/l2fwd_common.h   | 1 -
>  examples/l2fwd-jobstats/main.c        | 1 -
>  examples/l2fwd-keepalive/main.c       | 1 -
>  examples/l2fwd-macsec/main.c          | 1 -
>  examples/l2fwd/main.c                 | 2 +-
>  examples/l3fwd-graph/main.c           | 3 ---
>  examples/l3fwd-power/main.c           | 1 -
>  examples/l3fwd/main.c                 | 3 ---
>  examples/link_status_interrupt/main.c | 1 -
>  13 files changed, 1 insertion(+), 23 deletions(-)
> 
Series-acked-by: Bruce Richardson <bruce.richardson@intel.com>