Report standard counter stats->rx_missed_errors
using hc_rx_discards_no_wqe from the hardware.
Add a dedicated workqueue to periodically run
mana_query_gf_stats every 2 seconds to get the latest
info in eth_stats and define a driver capability flag
to notify hardware of the periodic queries.
To avoid repeated failures and log flooding, the workqueue
is not rescheduled if mana_query_gf_stats fails on HWC timeout
error and the stats are reset to 0. Other errors are transient
which will not need a VF reset for recovery.
Signed-off-by: Erni Sri Satya Vennela <ernis@linux.microsoft.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Reviewed-by: Dipayaan Roy <dipayanroy@linux.microsoft.com>
Reviewed-by: Shradha Gupta <shradhagupta@linux.microsoft.com>
---
Changes in v2:
* Update commit message.
* Stop rescheduling workqueue only when HWC timeout is observed.
* Introduce new variable in mana_context for detecting HWC timeout.
* Warn once in mana_get_stat64 when HWC timeout is observed.
---
drivers/net/ethernet/microsoft/mana/mana_en.c | 46 +++++++++++++++++--
.../ethernet/microsoft/mana/mana_ethtool.c | 2 -
include/net/mana/gdma.h | 6 ++-
include/net/mana/mana.h | 6 ++-
4 files changed, 53 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c
index 009e869ef296..48df44889f05 100644
--- a/drivers/net/ethernet/microsoft/mana/mana_en.c
+++ b/drivers/net/ethernet/microsoft/mana/mana_en.c
@@ -494,6 +494,11 @@ static void mana_get_stats64(struct net_device *ndev,
netdev_stats_to_stats64(st, &ndev->stats);
+ if (apc->ac->hwc_timeout_occurred)
+ netdev_warn_once(ndev, "HWC timeout occurred\n");
+
+ st->rx_missed_errors = apc->ac->hc_stats.hc_rx_discards_no_wqe;
+
for (q = 0; q < num_queues; q++) {
rx_stats = &apc->rxqs[q]->stats;
@@ -2769,7 +2774,7 @@ int mana_config_rss(struct mana_port_context *apc, enum TRI_STATE rx,
return 0;
}
-void mana_query_gf_stats(struct mana_context *ac)
+int mana_query_gf_stats(struct mana_context *ac)
{
struct mana_query_gf_stat_resp resp = {};
struct mana_query_gf_stat_req req = {};
@@ -2812,14 +2817,14 @@ void mana_query_gf_stats(struct mana_context *ac)
sizeof(resp));
if (err) {
dev_err(dev, "Failed to query GF stats: %d\n", err);
- return;
+ return err;
}
err = mana_verify_resp_hdr(&resp.hdr, MANA_QUERY_GF_STAT,
sizeof(resp));
if (err || resp.hdr.status) {
dev_err(dev, "Failed to query GF stats: %d, 0x%x\n", err,
resp.hdr.status);
- return;
+ return err;
}
ac->hc_stats.hc_rx_discards_no_wqe = resp.rx_discards_nowqe;
@@ -2854,6 +2859,8 @@ void mana_query_gf_stats(struct mana_context *ac)
ac->hc_stats.hc_tx_mcast_pkts = resp.hc_tx_mcast_pkts;
ac->hc_stats.hc_tx_mcast_bytes = resp.hc_tx_mcast_bytes;
ac->hc_stats.hc_tx_err_gdma = resp.tx_err_gdma;
+
+ return 0;
}
void mana_query_phy_stats(struct mana_port_context *apc)
@@ -3390,6 +3397,24 @@ int mana_rdma_service_event(struct gdma_context *gc, enum gdma_service_type even
return 0;
}
+#define MANA_GF_STATS_PERIOD (2 * HZ)
+
+static void mana_gf_stats_work_handler(struct work_struct *work)
+{
+ struct mana_context *ac =
+ container_of(to_delayed_work(work), struct mana_context, gf_stats_work);
+ int err;
+
+ err = mana_query_gf_stats(ac);
+ if (err == -ETIMEDOUT) {
+ /* HWC timeout detected - reset stats and stop rescheduling */
+ ac->hwc_timeout_occurred = true;
+ memset(&ac->hc_stats, 0, sizeof(ac->hc_stats));
+ return;
+ }
+ queue_delayed_work(ac->gf_stats_wq, &ac->gf_stats_work, MANA_GF_STATS_PERIOD);
+}
+
int mana_probe(struct gdma_dev *gd, bool resuming)
{
struct gdma_context *gc = gd->gdma_context;
@@ -3478,6 +3503,15 @@ int mana_probe(struct gdma_dev *gd, bool resuming)
}
err = add_adev(gd, "eth");
+ ac->gf_stats_wq = create_singlethread_workqueue("mana_gf_stats");
+ if (!ac->gf_stats_wq) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ INIT_DELAYED_WORK(&ac->gf_stats_work, mana_gf_stats_work_handler);
+ queue_delayed_work(ac->gf_stats_wq, &ac->gf_stats_work, MANA_GF_STATS_PERIOD);
+
out:
if (err) {
mana_remove(gd, false);
@@ -3501,6 +3535,12 @@ void mana_remove(struct gdma_dev *gd, bool suspending)
int err;
int i;
+ if (ac->gf_stats_wq) {
+ cancel_delayed_work_sync(&ac->gf_stats_work);
+ destroy_workqueue(ac->gf_stats_wq);
+ ac->gf_stats_wq = NULL;
+ }
+
/* adev currently doesn't support suspending, always remove it */
if (gd->adev)
remove_adev(gd);
diff --git a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
index 3dfd96146424..99e811208683 100644
--- a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
+++ b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
@@ -213,8 +213,6 @@ static void mana_get_ethtool_stats(struct net_device *ndev,
if (!apc->port_is_up)
return;
- /* we call mana function to update stats from GDMA */
- mana_query_gf_stats(apc->ac);
/* We call this mana function to get the phy stats from GDMA and includes
* aggregate tx/rx drop counters, Per-TC(Traffic Channel) tx/rx and pause
diff --git a/include/net/mana/gdma.h b/include/net/mana/gdma.h
index 57df78cfbf82..88a81fb164a0 100644
--- a/include/net/mana/gdma.h
+++ b/include/net/mana/gdma.h
@@ -591,6 +591,9 @@ enum {
/* Driver can self reset on FPGA Reconfig EQE notification */
#define GDMA_DRV_CAP_FLAG_1_HANDLE_RECONFIG_EQE BIT(17)
+/* Driver can send HWC periodically to query stats */
+#define GDMA_DRV_CAP_FLAG_1_PERIODIC_STATS_QUERY BIT(21)
+
#define GDMA_DRV_CAP_FLAGS1 \
(GDMA_DRV_CAP_FLAG_1_EQ_SHARING_MULTI_VPORT | \
GDMA_DRV_CAP_FLAG_1_NAPI_WKDONE_FIX | \
@@ -599,7 +602,8 @@ enum {
GDMA_DRV_CAP_FLAG_1_DEV_LIST_HOLES_SUP | \
GDMA_DRV_CAP_FLAG_1_DYNAMIC_IRQ_ALLOC_SUPPORT | \
GDMA_DRV_CAP_FLAG_1_SELF_RESET_ON_EQE | \
- GDMA_DRV_CAP_FLAG_1_HANDLE_RECONFIG_EQE)
+ GDMA_DRV_CAP_FLAG_1_HANDLE_RECONFIG_EQE | \
+ GDMA_DRV_CAP_FLAG_1_PERIODIC_STATS_QUERY)
#define GDMA_DRV_CAP_FLAGS2 0
diff --git a/include/net/mana/mana.h b/include/net/mana/mana.h
index 519c4384c51f..79532490cee6 100644
--- a/include/net/mana/mana.h
+++ b/include/net/mana/mana.h
@@ -480,6 +480,10 @@ struct mana_context {
struct mana_eq *eqs;
struct dentry *mana_eqs_debugfs;
+ struct workqueue_struct *gf_stats_wq;
+ struct delayed_work gf_stats_work;
+ bool hwc_timeout_occurred;
+
struct net_device *ports[MAX_PORTS_IN_MANA_DEV];
};
@@ -577,7 +581,7 @@ u32 mana_run_xdp(struct net_device *ndev, struct mana_rxq *rxq,
struct bpf_prog *mana_xdp_get(struct mana_port_context *apc);
void mana_chn_setxdp(struct mana_port_context *apc, struct bpf_prog *prog);
int mana_bpf(struct net_device *ndev, struct netdev_bpf *bpf);
-void mana_query_gf_stats(struct mana_context *ac);
+int mana_query_gf_stats(struct mana_context *ac);
int mana_query_link_cfg(struct mana_port_context *apc);
int mana_set_bw_clamp(struct mana_port_context *apc, u32 speed,
int enable_clamping);
--
2.34.1On Wed, 29 Oct 2025 03:37:52 -0700 Erni Sri Satya Vennela wrote:
> diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c
> index 009e869ef296..48df44889f05 100644
> --- a/drivers/net/ethernet/microsoft/mana/mana_en.c
> +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c
> @@ -494,6 +494,11 @@ static void mana_get_stats64(struct net_device *ndev,
>
> netdev_stats_to_stats64(st, &ndev->stats);
>
> + if (apc->ac->hwc_timeout_occurred)
> + netdev_warn_once(ndev, "HWC timeout occurred\n");
I don't think there's much value in this print.
> +#define MANA_GF_STATS_PERIOD (2 * HZ)
> +
> +static void mana_gf_stats_work_handler(struct work_struct *work)
> +{
> + struct mana_context *ac =
> + container_of(to_delayed_work(work), struct mana_context, gf_stats_work);
> + int err;
> +
> + err = mana_query_gf_stats(ac);
> + if (err == -ETIMEDOUT) {
> + /* HWC timeout detected - reset stats and stop rescheduling */
> + ac->hwc_timeout_occurred = true;
> + memset(&ac->hc_stats, 0, sizeof(ac->hc_stats));
Not sure I've seen another device using this approach but I can't
really tell what's the best strategy. The device is unusable if it
can't provide stats..
> + return;
> + }
> + queue_delayed_work(ac->gf_stats_wq, &ac->gf_stats_work, MANA_GF_STATS_PERIOD);
> +}
> +
> int mana_probe(struct gdma_dev *gd, bool resuming)
> {
> struct gdma_context *gc = gd->gdma_context;
> @@ -3478,6 +3503,15 @@ int mana_probe(struct gdma_dev *gd, bool resuming)
> }
>
> err = add_adev(gd, "eth");
> + ac->gf_stats_wq = create_singlethread_workqueue("mana_gf_stats");
Why are you creating a workqueue? You can use system queues.
> + queue_delayed_work(ac->gf_stats_wq, &ac->gf_stats_work, MANA_GF_STATS_PERIOD);
ls wrap long lines at 80 chars.
> diff --git a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> index 3dfd96146424..99e811208683 100644
> --- a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> +++ b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> @@ -213,8 +213,6 @@ static void mana_get_ethtool_stats(struct net_device *ndev,
>
> if (!apc->port_is_up)
> return;
> - /* we call mana function to update stats from GDMA */
> - mana_query_gf_stats(apc->ac);
Why delete this? We can get fresh stats for the user in this context.
On Fri, Oct 31, 2025 at 04:17:23PM -0700, Jakub Kicinski wrote:
> On Wed, 29 Oct 2025 03:37:52 -0700 Erni Sri Satya Vennela wrote:
> > diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c
> > index 009e869ef296..48df44889f05 100644
> > --- a/drivers/net/ethernet/microsoft/mana/mana_en.c
> > +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c
> > @@ -494,6 +494,11 @@ static void mana_get_stats64(struct net_device *ndev,
> >
> > netdev_stats_to_stats64(st, &ndev->stats);
> >
> > + if (apc->ac->hwc_timeout_occurred)
> > + netdev_warn_once(ndev, "HWC timeout occurred\n");
>
> I don't think there's much value in this print.
This print is added because, if the user tries to run
"ip -s link interface" for multiple times, he should be warned about
the time out. I will add more info into the print to let the user
know that it is from this action.
>
> > +#define MANA_GF_STATS_PERIOD (2 * HZ)
> > +
> > +static void mana_gf_stats_work_handler(struct work_struct *work)
> > +{
> > + struct mana_context *ac =
> > + container_of(to_delayed_work(work), struct mana_context, gf_stats_work);
> > + int err;
> > +
> > + err = mana_query_gf_stats(ac);
> > + if (err == -ETIMEDOUT) {
> > + /* HWC timeout detected - reset stats and stop rescheduling */
> > + ac->hwc_timeout_occurred = true;
> > + memset(&ac->hc_stats, 0, sizeof(ac->hc_stats));
>
> Not sure I've seen another device using this approach but I can't
> really tell what's the best strategy. The device is unusable if it
> can't provide stats..
In the case where the HWC becomes unresponsive,
there will be a brief interval needed for the driver to recover.
During this period, if users request ethtool stats, they would
receive outdated information. To address this, we proactively
reset the stats to zero, ensuring users do not see stale data.
>
> > + return;
> > + }
> > + queue_delayed_work(ac->gf_stats_wq, &ac->gf_stats_work, MANA_GF_STATS_PERIOD);
> > +}
> > +
> > int mana_probe(struct gdma_dev *gd, bool resuming)
> > {
> > struct gdma_context *gc = gd->gdma_context;
> > @@ -3478,6 +3503,15 @@ int mana_probe(struct gdma_dev *gd, bool resuming)
> > }
> >
> > err = add_adev(gd, "eth");
> > + ac->gf_stats_wq = create_singlethread_workqueue("mana_gf_stats");
>
> Why are you creating a workqueue? You can use system queues.
Thankyou for the suggestion. I will integrate it in the next version.
>
> > + queue_delayed_work(ac->gf_stats_wq, &ac->gf_stats_work, MANA_GF_STATS_PERIOD);
>
> ls wrap long lines at 80 chars.
I will make this change for the next verison.
>
> > diff --git a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> > index 3dfd96146424..99e811208683 100644
> > --- a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> > +++ b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> > @@ -213,8 +213,6 @@ static void mana_get_ethtool_stats(struct net_device *ndev,
> >
> > if (!apc->port_is_up)
> > return;
> > - /* we call mana function to update stats from GDMA */
> > - mana_query_gf_stats(apc->ac);
>
> Why delete this? We can get fresh stats for the user in this context.
We want to prevent some user that run ethtool too frequently, like
thousands of times / sec, to overwhelm the HW channel.
© 2016 - 2025 Red Hat, Inc.