[PATCH] net: mvpp2: fix XDP rx queues registering

The registration of XDP queue information is incorrect because the RX queue id we use is invalid.
When port->id == 0 it appears to works as expected yet it's no longer the case when port->id != 0.

This is due to the fact that the value we use (rxq->id) is not the per-port queue index which
XDP expects -- it's a global index which should not be used in this case. Instead we shall use
rxq->logic_rxq which is the correct, per-port value.

Signed-off-by: Louis Amas <redacted>
Signed-off-by: Emmanuel Deloget <redacted>
---

As we were trying to capture packets using XDP on our mv8040-powered
MacchiatoBin, we experienced an issue related to rx queue numbering.

Before I get to the problem itself, a bit of setup:

* the Macchiato has several ports, all of them handled using the mvpp2
  ethernet driver. We are able to reproduce our issue on any device whose
  port->id != 0 (we used eth2 for our tests). When port->id == 0 (for
  example on eth0) everything works as expected ;

* we use xdp-tutorial for our tests ; more specifically, we used the
  advanced03-AF_XDP tutorial as it provides a simple testbed. We modified
  the kernel to simplify it:

        SEC("xdp_sock")
        int xdp_sock_prog(struct xdp_md *ctx)
        {
                int index = ctx->rx_queue_index;

                /* A set entry here means that the correspnding queue_id
                 * has an active AF_XDP socket bound to it. */
                if (bpf_map_lookup_elem(&xsks_map, &index))
                        return bpf_redirect_map(&xsks_map, index, 0);

                return XDP_PASS;
        }

* we tested kernel 5.10 (out target) and 5.15 (for reference) ; both kernels
  exhibits the same symptoms ; I expect kernel 5.9 (the first linux kernel
  with XDP support in the mvpp2 driver) to exhibit the same problem.

The normal invocation of this program would be:

        ./af_xdp_user -d ETHDEV

We should then capture packets on this interface. When ETHDEV is eth0
(port->id == 0) everything works as expcted ; when using ETHDEV == eth2
we fail to capture anything.

We investigated the issue and found that XDP rx queues (setup as
struct xdp_rxq_info by the mvpp2 driver) for this device were wrong. XDP
expected them to be numbered in [0..3] but we found numbers in [32..35].

The reason for this lies in mvpp2_main.c at lines 2962 and 2966 which are
of the form (symbolic notation, close to actual code, function
mvpp2_rxq_init()):

        err = xdp_rxq_info_reg(&rxq->some_xdp_rxqinfo, port->dev, rxq->id, 0);

The rxq->id value we pass to this function is incorrect - it's a virtual queue
id which is computed as (symbolic notation, not actual code):

        rxq->id = port->id * max_rxq_count + queue_id

In our case, max_rxq_count == 32 and port->id == 1 for eth2, meaning our
rxq->id are in the range [32..35] (for 4 queues).

We tried to force the rx queue id on the XDP side by using:

        ./af_xdp_user -d eth2 -Q 32

But that failed -- as expected, because we should not have more than 4
rx queues.

The computing of rxq->id is valid, but the use of rxq->id in this context is
not. What we really want here is the rx queue id for this port, and this value
is stored in rxq->logic_rxq -- as hinted by the code in mvpp2_rxq_init().
Replacing rxq->id by this value in the two xdp_rxq_info_reg() calls fixed the
issue and allowed us to use XDP on all the Macchiato ethernet ports.

drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
index 587def69a6f7..f0ea377341c6 100644
--- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
+++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c

@@ -2959,11 +2959,11 @@ static int mvpp2_rxq_init(struct mvpp2_port *port,
        mvpp2_rxq_status_update(port, rxq->id, 0, rxq->size);

        if (priv->percpu_pools) {
-               err = xdp_rxq_info_reg(&rxq->xdp_rxq_short, port->dev, rxq->id, 0);
+               err = xdp_rxq_info_reg(&rxq->xdp_rxq_short, port->dev, rxq->logic_rxq, 0);
                if (err < 0)
                        goto err_free_dma;

-               err = xdp_rxq_info_reg(&rxq->xdp_rxq_long, port->dev, rxq->id, 0);
+               err = xdp_rxq_info_reg(&rxq->xdp_rxq_long, port->dev, rxq->logic_rxq, 0);
                if (err < 0)
                        goto err_unregister_rxq_short;

--

2.25.1


[eho.link event] <https://www.linkedin.com/company/eho.link/>

Hi,


wt., 9 lis 2021 o 11:31 Louis Amas [off-list ref] napisał(a):

Thank you for the patch and the detailed explanation. I think "Fixes:"
tag should be added to the commit message:
Fixes: b27db2274ba8 ("mvpp2: use page_pool allocator")

Other than that - LGTM, so you can add my:
Reviewed-by: Marcin Wojtas <redacted>

Best regards,
Marcin

Hello, 

From: Marcin Wojtas <redacted>

wt., 9 lis 2021 o 11:31 Louis Amas [off-list ref] napisał(a):

Thank you for the patch and the detailed explanation. I think "Fixes:"
tag should be added to the commit message:
Fixes: b27db2274ba8 ("mvpp2: use page_pool allocator")

Other than that - LGTM, so you can add my:
Reviewed-by: Marcin Wojtas <redacted>

Best regards,
Marcin

Thank you :)

We'll send an augmented version tomorrow along with a small rework of the commit message. 

Best regards, 

-- Emmanuel Deloget

The registration of XDP queue information is incorrect because the
RX queue id we use is invalid. When port->id == 0 it appears to works
as expected yet it's no longer the case when port->id != 0.

When we register the XDP rx queue information (using
xdp_rxq_info_reg() in function mvpp2_rxq_init()) we tell them to use
rxq->id as the queue id. This value iscomputed as:
rxq->id = port->id * max_rxq_count + queue_id

where max_rxq_count depends on the device version. In the MB case,
this value is 32, meaning that rx queues on eth2 are numbered from
32 to 35 - there are four of them.

Clearly, this is not the per-port queue id that XDP is expecting:
it wants a value in the range [0..3]. It shall directly use queue_id
which is stored in rxq->logic_rxq -- so let's use that value instead.

This is consistent with the remaining part of the code in
mvpp2_rxq_init().

Fixes: b27db2274ba8 ("mvpp2: use page_pool allocator")
Signed-off-by: Louis Amas <redacted>
Signed-off-by: Emmanuel Deloget <redacted>
Reviewed-by: Marcin Wojtas <redacted>
---
This is a repost of [1]. The patch itself is not changed, but the
commit message has been enhanced using part of the explaination in
order to make it clearer (hopefully) and to incorporate the
reviewed-by tag from Marcin.

v1: original patch
v2: revamped commit description (no change in the patch itself)

[1] https://lore.kernel.org/bpf/20211109103101.92382-1-louis.amas@eho.link/

 drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
index 587def69a6f7..f0ea377341c6 100644
--- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
+++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c

@@ -2959,11 +2959,11 @@ static int mvpp2_rxq_init(struct mvpp2_port *port,
        mvpp2_rxq_status_update(port, rxq->id, 0, rxq->size);

        if (priv->percpu_pools) {
-               err = xdp_rxq_info_reg(&rxq->xdp_rxq_short, port->dev, rxq->id, 0);
+               err = xdp_rxq_info_reg(&rxq->xdp_rxq_short, port->dev, rxq->logic_rxq, 0);
                if (err < 0)
                        goto err_free_dma;

-               err = xdp_rxq_info_reg(&rxq->xdp_rxq_long, port->dev, rxq->id, 0);
+               err = xdp_rxq_info_reg(&rxq->xdp_rxq_long, port->dev, rxq->logic_rxq, 0);
                if (err < 0)
                        goto err_unregister_rxq_short;

--

2.25.1


[eho.link event] <https://www.linkedin.com/company/eho.link/>

Hello,

On 10/11/2021 15:41, Louis Amas wrote:

Is there any update on this patch ? Without it, XDP only partially work 
on a MACCHIATOBin (read: it works on some ports, not on others, as 
described in our analysis sent together with the original patch).

Best regards,

-- Emmanuel Deloget

On Mon, Dec 06, 2021 at 04:37:20PM +0100, Emmanuel Deloget wrote:

Hello,

On 10/11/2021 15:41, Louis Amas wrote:

quoted

The registration of XDP queue information is incorrect because the
RX queue id we use is invalid. When port->id == 0 it appears to works
as expected yet it's no longer the case when port->id != 0.

When we register the XDP rx queue information (using
xdp_rxq_info_reg() in function mvpp2_rxq_init()) we tell them to use
rxq->id as the queue id. This value iscomputed as:
rxq->id = port->id * max_rxq_count + queue_id

where max_rxq_count depends on the device version. In the MB case,
this value is 32, meaning that rx queues on eth2 are numbered from
32 to 35 - there are four of them.

Clearly, this is not the per-port queue id that XDP is expecting:
it wants a value in the range [0..3]. It shall directly use queue_id
which is stored in rxq->logic_rxq -- so let's use that value instead.

This is consistent with the remaining part of the code in
mvpp2_rxq_init().

Fixes: b27db2274ba8 ("mvpp2: use page_pool allocator")
Signed-off-by: Louis Amas <redacted>
Signed-off-by: Emmanuel Deloget <redacted>
Reviewed-by: Marcin Wojtas <redacted>
---
This is a repost of [1]. The patch itself is not changed, but the
commit message has been enhanced using part of the explaination in
order to make it clearer (hopefully) and to incorporate the
reviewed-by tag from Marcin.

v1: original patch
v2: revamped commit description (no change in the patch itself)

[1] https://lore.kernel.org/bpf/20211109103101.92382-1-louis.amas@eho.link/

  drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
index 587def69a6f7..f0ea377341c6 100644
--- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
+++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c

@@ -2959,11 +2959,11 @@ static int mvpp2_rxq_init(struct mvpp2_port *port,
  	mvpp2_rxq_status_update(port, rxq->id, 0, rxq->size);
  	if (priv->percpu_pools) {
-		err = xdp_rxq_info_reg(&rxq->xdp_rxq_short, port->dev, rxq->id, 0);
+		err = xdp_rxq_info_reg(&rxq->xdp_rxq_short, port->dev, rxq->logic_rxq, 0);
  		if (err < 0)
  			goto err_free_dma;
-		err = xdp_rxq_info_reg(&rxq->xdp_rxq_long, port->dev, rxq->id, 0);
+		err = xdp_rxq_info_reg(&rxq->xdp_rxq_long, port->dev, rxq->logic_rxq, 0);
  		if (err < 0)
  			goto err_unregister_rxq_short;

Is there any update on this patch ? Without it, XDP only partially work on a
MACCHIATOBin (read: it works on some ports, not on others, as described in
our analysis sent together with the original patch).

Hi,

I suspect if you *didn't* thread your updated patch to your previous
submission, then it would end up with a separate entry in
patchwork.kernel.org, and the netdev maintainers will notice that the
patch is ready for inclusion, having been reviewed by Marcin.

Thanks.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

On Mon, 6 Dec 2021 15:42:47 +0000 Russell King (Oracle) wrote:

On Mon, Dec 06, 2021 at 04:37:20PM +0100, Emmanuel Deloget wrote:

quoted

On 10/11/2021 15:41, Louis Amas wrote:  

quoted

The registration of XDP queue information is incorrect because the
RX queue id we use is invalid. When port->id == 0 it appears to works
as expected yet it's no longer the case when port->id != 0.

When we register the XDP rx queue information (using
xdp_rxq_info_reg() in function mvpp2_rxq_init()) we tell them to use
rxq->id as the queue id. This value iscomputed as:
rxq->id = port->id * max_rxq_count + queue_id

where max_rxq_count depends on the device version. In the MB case,
this value is 32, meaning that rx queues on eth2 are numbered from
32 to 35 - there are four of them.

Clearly, this is not the per-port queue id that XDP is expecting:
it wants a value in the range [0..3]. It shall directly use queue_id
which is stored in rxq->logic_rxq -- so let's use that value instead.

This is consistent with the remaining part of the code in
mvpp2_rxq_init().

quoted

Is there any update on this patch ? Without it, XDP only partially work on a
MACCHIATOBin (read: it works on some ports, not on others, as described in
our analysis sent together with the original patch).  

I suspect if you *didn't* thread your updated patch to your previous
submission, then it would end up with a separate entry in
patchwork.kernel.org,

Indeed, it's easier to keep track of patches which weren't posted 
as a reply in a thread, at least for me.

and the netdev maintainers will notice that the
patch is ready for inclusion, having been reviewed by Marcin.

In this case I _think_ it was dropped because it didn't apply.

Please rebase on top of net/master and repost if the changes is still
needed.

Jakub Kicinski wrote:

On Mon, 6 Dec 2021 15:42:47 +0000 Russell King (Oracle) wrote:

quoted

On Mon, Dec 06, 2021 at 04:37:20PM +0100, Emmanuel Deloget wrote:

quoted

On 10/11/2021 15:41, Louis Amas wrote:  

quoted

The registration of XDP queue information is incorrect because the
RX queue id we use is invalid. When port->id == 0 it appears to works
as expected yet it's no longer the case when port->id != 0.

When we register the XDP rx queue information (using
xdp_rxq_info_reg() in function mvpp2_rxq_init()) we tell them to use
rxq->id as the queue id. This value iscomputed as:
rxq->id = port->id * max_rxq_count + queue_id

where max_rxq_count depends on the device version. In the MB case,
this value is 32, meaning that rx queues on eth2 are numbered from
32 to 35 - there are four of them.

Clearly, this is not the per-port queue id that XDP is expecting:
it wants a value in the range [0..3]. It shall directly use queue_id
which is stored in rxq->logic_rxq -- so let's use that value instead.

This is consistent with the remaining part of the code in
mvpp2_rxq_init().

quoted

quoted

Is there any update on this patch ? Without it, XDP only partially work on a
MACCHIATOBin (read: it works on some ports, not on others, as described in
our analysis sent together with the original patch).  

I suspect if you *didn't* thread your updated patch to your previous
submission, then it would end up with a separate entry in
patchwork.kernel.org,

Indeed, it's easier to keep track of patches which weren't posted 
as a reply in a thread, at least for me.

quoted

and the netdev maintainers will notice that the
patch is ready for inclusion, having been reviewed by Marcin.

In this case I _think_ it was dropped because it didn't apply.

Please rebase on top of net/master and repost if the changes is still
needed.

Also I would add the detailed description to the actual commit not below
the "--" lines. Capturing that in the log will be useful for future
reference if we ever hit similar issue here or elsewhere.

Otherwise for patch,

Acked-by: John Fastabend <john.fastabend@gmail.com>

John,

On 06/12/2021 17:16, John Fastabend wrote:

Jakub Kicinski wrote:

quoted

On Mon, 6 Dec 2021 15:42:47 +0000 Russell King (Oracle) wrote:

quoted

On Mon, Dec 06, 2021 at 04:37:20PM +0100, Emmanuel Deloget wrote:

quoted

On 10/11/2021 15:41, Louis Amas wrote:

quoted

The registration of XDP queue information is incorrect because the
RX queue id we use is invalid. When port->id == 0 it appears to works
as expected yet it's no longer the case when port->id != 0.

When we register the XDP rx queue information (using
xdp_rxq_info_reg() in function mvpp2_rxq_init()) we tell them to use
rxq->id as the queue id. This value iscomputed as:
rxq->id = port->id * max_rxq_count + queue_id

where max_rxq_count depends on the device version. In the MB case,
this value is 32, meaning that rx queues on eth2 are numbered from
32 to 35 - there are four of them.

Clearly, this is not the per-port queue id that XDP is expecting:
it wants a value in the range [0..3]. It shall directly use queue_id
which is stored in rxq->logic_rxq -- so let's use that value instead.

This is consistent with the remaining part of the code in
mvpp2_rxq_init().

quoted

quoted

Is there any update on this patch ? Without it, XDP only partially work on a
MACCHIATOBin (read: it works on some ports, not on others, as described in
our analysis sent together with the original patch).

I suspect if you *didn't* thread your updated patch to your previous
submission, then it would end up with a separate entry in
patchwork.kernel.org,

Indeed, it's easier to keep track of patches which weren't posted
as a reply in a thread, at least for me.

quoted

and the netdev maintainers will notice that the
patch is ready for inclusion, having been reviewed by Marcin.

In this case I _think_ it was dropped because it didn't apply.

Please rebase on top of net/master and repost if the changes is still
needed.

Also I would add the detailed description to the actual commit not below
the "--" lines. Capturing that in the log will be useful for future
reference if we ever hit similar issue here or elsewhere.

Otherwise for patch,

Acked-by: John Fastabend <john.fastabend@gmail.com>

Ouch. We failed to get this email before resending the patch.

My bad - and sorry for the inconvenience.

I'll see with Louis to change again the commit message and add your 
Acked-by.

Best regards,

-- Emmanuel Deloget