From: Larysa Zaremba <larysa.zaremba@intel.com>
To: intel-wired-lan@lists.osuosl.org,
Jacob Keller <jacob.e.keller@intel.com>
Cc: maciej.fijalkowski@intel.com,
Jesper Dangaard Brouer <hawk@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Larysa Zaremba <larysa.zaremba@intel.com>,
Przemek Kitszel <przemyslaw.kitszel@intel.com>,
John Fastabend <john.fastabend@gmail.com>,
Alexei Starovoitov <ast@kernel.org>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
netdev@vger.kernel.org, Jakub Kicinski <kuba@kernel.org>,
bpf@vger.kernel.org, Paolo Abeni <pabeni@redhat.com>,
Magnus Karlsson <magnus.karlsson@gmail.com>,
igor.bagnucki@intel.com, linux-kernel@vger.kernel.org
Subject: [Intel-wired-lan] [PATCH iwl-net 3/3] ice: map XDP queues to vectors in ice_vsi_map_rings_to_vectors()
Date: Wed, 15 May 2024 18:02:16 +0200 [thread overview]
Message-ID: <20240515160246.5181-4-larysa.zaremba@intel.com> (raw)
In-Reply-To: <20240515160246.5181-1-larysa.zaremba@intel.com>
ice_pf_dcb_recfg() re-maps queues to vectors with
ice_vsi_map_rings_to_vectors(), which does not restore the previous
state for XDP queues. This leads to no AF_XDP traffic after rebuild.
Map XDP queues to vectors in ice_vsi_map_rings_to_vectors().
Also, move the code around, so XDP queues are mapped independently only
through .ndo_bpf().
Fixes: 6624e780a577 ("ice: split ice_vsi_setup into smaller functions")
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
---
drivers/net/ethernet/intel/ice/ice.h | 1 +
drivers/net/ethernet/intel/ice/ice_base.c | 3 +
drivers/net/ethernet/intel/ice/ice_lib.c | 14 ++--
drivers/net/ethernet/intel/ice/ice_main.c | 96 ++++++++++++++---------
4 files changed, 68 insertions(+), 46 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index b91b2594b29d..da8c8afebc93 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -940,6 +940,7 @@ int ice_vsi_determine_xdp_res(struct ice_vsi *vsi);
int ice_prepare_xdp_rings(struct ice_vsi *vsi, struct bpf_prog *prog,
enum ice_xdp_cfg cfg_type);
int ice_destroy_xdp_rings(struct ice_vsi *vsi, enum ice_xdp_cfg cfg_type);
+void ice_map_xdp_rings(struct ice_vsi *vsi);
int
ice_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames,
u32 flags);
diff --git a/drivers/net/ethernet/intel/ice/ice_base.c b/drivers/net/ethernet/intel/ice/ice_base.c
index 687f6cb2b917..5d396c1a7731 100644
--- a/drivers/net/ethernet/intel/ice/ice_base.c
+++ b/drivers/net/ethernet/intel/ice/ice_base.c
@@ -842,6 +842,9 @@ void ice_vsi_map_rings_to_vectors(struct ice_vsi *vsi)
}
rx_rings_rem -= rx_rings_per_v;
}
+
+ if (ice_is_xdp_ena_vsi(vsi))
+ ice_map_xdp_rings(vsi);
}
/**
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index dd8b374823ee..7629b0190578 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -2274,13 +2274,6 @@ static int ice_vsi_cfg_def(struct ice_vsi *vsi)
if (ret)
goto unroll_vector_base;
- ice_vsi_map_rings_to_vectors(vsi);
-
- /* Associate q_vector rings to napi */
- ice_vsi_set_napi_queues(vsi);
-
- vsi->stat_offsets_loaded = false;
-
if (ice_is_xdp_ena_vsi(vsi)) {
ret = ice_vsi_determine_xdp_res(vsi);
if (ret)
@@ -2291,6 +2284,13 @@ static int ice_vsi_cfg_def(struct ice_vsi *vsi)
goto unroll_vector_base;
}
+ ice_vsi_map_rings_to_vectors(vsi);
+
+ /* Associate q_vector rings to napi */
+ ice_vsi_set_napi_queues(vsi);
+
+ vsi->stat_offsets_loaded = false;
+
/* ICE_VSI_CTRL does not need RSS so skip RSS processing */
if (vsi->type != ICE_VSI_CTRL)
/* Do not exit if configuring RSS had an issue, at
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 2a270aacd24a..1b61ca3a6eb6 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -2707,6 +2707,60 @@ static void ice_vsi_assign_bpf_prog(struct ice_vsi *vsi, struct bpf_prog *prog)
bpf_prog_put(old_prog);
}
+static struct ice_tx_ring *ice_xdp_ring_from_qid(struct ice_vsi *vsi, int qid)
+{
+ struct ice_q_vector *q_vector;
+ struct ice_tx_ring *ring;
+
+ if (static_key_enabled(&ice_xdp_locking_key))
+ return vsi->xdp_rings[qid % vsi->num_xdp_txq];
+
+ q_vector = vsi->rx_rings[qid]->q_vector;
+ ice_for_each_tx_ring(ring, q_vector->tx)
+ if (ice_ring_is_xdp(ring))
+ return ring;
+
+ return NULL;
+}
+
+/**
+ * ice_map_xdp_rings - Map XDP rings to interrupt vectors
+ * @vsi: the VSI with XDP rings being configured
+ *
+ * Map XDP rings to interrupt vectors and perform the configuration steps
+ * dependent on the mapping.
+ */
+void ice_map_xdp_rings(struct ice_vsi *vsi)
+{
+ int xdp_rings_rem = vsi->num_xdp_txq;
+ int v_idx, q_idx;
+
+ /* follow the logic from ice_vsi_map_rings_to_vectors */
+ ice_for_each_q_vector(vsi, v_idx) {
+ struct ice_q_vector *q_vector = vsi->q_vectors[v_idx];
+ int xdp_rings_per_v, q_id, q_base;
+
+ xdp_rings_per_v = DIV_ROUND_UP(xdp_rings_rem,
+ vsi->num_q_vectors - v_idx);
+ q_base = vsi->num_xdp_txq - xdp_rings_rem;
+
+ for (q_id = q_base; q_id < (q_base + xdp_rings_per_v); q_id++) {
+ struct ice_tx_ring *xdp_ring = vsi->xdp_rings[q_id];
+
+ xdp_ring->q_vector = q_vector;
+ xdp_ring->next = q_vector->tx.tx_ring;
+ q_vector->tx.tx_ring = xdp_ring;
+ }
+ xdp_rings_rem -= xdp_rings_per_v;
+ }
+
+ ice_for_each_rxq(vsi, q_idx) {
+ vsi->rx_rings[q_idx]->xdp_ring = ice_xdp_ring_from_qid(vsi,
+ q_idx);
+ ice_tx_xsk_pool(vsi, q_idx);
+ }
+}
+
/**
* ice_prepare_xdp_rings - Allocate, configure and setup Tx rings for XDP
* @vsi: VSI to bring up Tx rings used by XDP
@@ -2719,7 +2773,6 @@ int ice_prepare_xdp_rings(struct ice_vsi *vsi, struct bpf_prog *prog,
enum ice_xdp_cfg cfg_type)
{
u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
- int xdp_rings_rem = vsi->num_xdp_txq;
struct ice_pf *pf = vsi->back;
struct ice_qs_cfg xdp_qs_cfg = {
.qs_mutex = &pf->avail_q_mutex,
@@ -2732,8 +2785,7 @@ int ice_prepare_xdp_rings(struct ice_vsi *vsi, struct bpf_prog *prog,
.mapping_mode = ICE_VSI_MAP_CONTIG
};
struct device *dev;
- int i, v_idx;
- int status;
+ int status, i;
dev = ice_pf_to_dev(pf);
vsi->xdp_rings = devm_kcalloc(dev, vsi->num_xdp_txq,
@@ -2752,42 +2804,6 @@ int ice_prepare_xdp_rings(struct ice_vsi *vsi, struct bpf_prog *prog,
if (ice_xdp_alloc_setup_rings(vsi))
goto clear_xdp_rings;
- /* follow the logic from ice_vsi_map_rings_to_vectors */
- ice_for_each_q_vector(vsi, v_idx) {
- struct ice_q_vector *q_vector = vsi->q_vectors[v_idx];
- int xdp_rings_per_v, q_id, q_base;
-
- xdp_rings_per_v = DIV_ROUND_UP(xdp_rings_rem,
- vsi->num_q_vectors - v_idx);
- q_base = vsi->num_xdp_txq - xdp_rings_rem;
-
- for (q_id = q_base; q_id < (q_base + xdp_rings_per_v); q_id++) {
- struct ice_tx_ring *xdp_ring = vsi->xdp_rings[q_id];
-
- xdp_ring->q_vector = q_vector;
- xdp_ring->next = q_vector->tx.tx_ring;
- q_vector->tx.tx_ring = xdp_ring;
- }
- xdp_rings_rem -= xdp_rings_per_v;
- }
-
- ice_for_each_rxq(vsi, i) {
- if (static_key_enabled(&ice_xdp_locking_key)) {
- vsi->rx_rings[i]->xdp_ring = vsi->xdp_rings[i % vsi->num_xdp_txq];
- } else {
- struct ice_q_vector *q_vector = vsi->rx_rings[i]->q_vector;
- struct ice_tx_ring *ring;
-
- ice_for_each_tx_ring(ring, q_vector->tx) {
- if (ice_ring_is_xdp(ring)) {
- vsi->rx_rings[i]->xdp_ring = ring;
- break;
- }
- }
- }
- ice_tx_xsk_pool(vsi, i);
- }
-
/* omit the scheduler update if in reset path; XDP queues will be
* taken into account at the end of ice_vsi_rebuild, where
* ice_cfg_vsi_lan is being called
@@ -2795,6 +2811,8 @@ int ice_prepare_xdp_rings(struct ice_vsi *vsi, struct bpf_prog *prog,
if (cfg_type == ICE_XDP_CFG_PART)
return 0;
+ ice_map_xdp_rings(vsi);
+
/* tell the Tx scheduler that right now we have
* additional queues
*/
--
2.43.0
next prev parent reply other threads:[~2024-05-15 16:09 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-15 16:02 [Intel-wired-lan] [PATCH iwl-net 0/3] Fix AF_XDP problems after changing queue number Larysa Zaremba
2024-05-15 16:02 ` [Intel-wired-lan] [PATCH iwl-net 1/3] ice: remove af_xdp_zc_qps bitmap Larysa Zaremba
2024-05-16 8:27 ` Simon Horman
2024-05-15 16:02 ` [Intel-wired-lan] [PATCH iwl-net 2/3] ice: add flag to distinguish reset from .ndo_bpf in XDP rings config Larysa Zaremba
2024-05-16 8:27 ` Simon Horman
2024-05-16 10:00 ` Sergey Temerkhanov
2024-05-15 16:02 ` Larysa Zaremba [this message]
2024-05-16 8:27 ` [Intel-wired-lan] [PATCH iwl-net 3/3] ice: map XDP queues to vectors in ice_vsi_map_rings_to_vectors() Simon Horman
2024-05-16 11:43 ` Larysa Zaremba
2024-05-16 11:59 ` Simon Horman
2024-05-16 17:12 ` Jacob Keller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240515160246.5181-4-larysa.zaremba@intel.com \
--to=larysa.zaremba@intel.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=hawk@kernel.org \
--cc=igor.bagnucki@intel.com \
--cc=intel-wired-lan@lists.osuosl.org \
--cc=jacob.e.keller@intel.com \
--cc=john.fastabend@gmail.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=maciej.fijalkowski@intel.com \
--cc=magnus.karlsson@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=przemyslaw.kitszel@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).