6.12-stable review patch. If anyone has any objections, please let me know.
------------------
From: Hariprasad Kelam hkelam@marvell.com
[ Upstream commit 479c58016099d19686e36f6c50f912360839a7fa ]
QOS is designed to create a new send queue whenever a class is created, ensuring proper shaping and scheduling. However, when multiple send queues are created and deleted in a loop, SMMU errors are observed.
This patch addresses the issue by performing an data cache sync during the teardown of QOS send queues.
Fixes: ab6dddd2a669 ("octeontx2-pf: qos send queues management") Signed-off-by: Hariprasad Kelam hkelam@marvell.com Reviewed-by: Simon Horman horms@kernel.org Link: https://patch.msgid.link/20250522094742.1498295-1-hkelam@marvell.com Signed-off-by: Paolo Abeni pabeni@redhat.com Signed-off-by: Sasha Levin sashal@kernel.org --- .../ethernet/marvell/octeontx2/nic/qos_sq.c | 22 +++++++++++++++++++ 1 file changed, 22 insertions(+)
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/qos_sq.c b/drivers/net/ethernet/marvell/octeontx2/nic/qos_sq.c index 9d887bfc31089..ac9345644068e 100644 --- a/drivers/net/ethernet/marvell/octeontx2/nic/qos_sq.c +++ b/drivers/net/ethernet/marvell/octeontx2/nic/qos_sq.c @@ -256,6 +256,26 @@ int otx2_qos_enable_sq(struct otx2_nic *pfvf, int qidx) return err; }
+static int otx2_qos_nix_npa_ndc_sync(struct otx2_nic *pfvf) +{ + struct ndc_sync_op *req; + int rc; + + mutex_lock(&pfvf->mbox.lock); + + req = otx2_mbox_alloc_msg_ndc_sync_op(&pfvf->mbox); + if (!req) { + mutex_unlock(&pfvf->mbox.lock); + return -ENOMEM; + } + + req->nix_lf_tx_sync = true; + req->npa_lf_sync = true; + rc = otx2_sync_mbox_msg(&pfvf->mbox); + mutex_unlock(&pfvf->mbox.lock); + return rc; +} + void otx2_qos_disable_sq(struct otx2_nic *pfvf, int qidx) { struct otx2_qset *qset = &pfvf->qset; @@ -285,6 +305,8 @@ void otx2_qos_disable_sq(struct otx2_nic *pfvf, int qidx)
otx2_qos_sqb_flush(pfvf, sq_idx); otx2_smq_flush(pfvf, otx2_get_smq_idx(pfvf, sq_idx)); + /* NIX/NPA NDC sync */ + otx2_qos_nix_npa_ndc_sync(pfvf); otx2_cleanup_tx_cqes(pfvf, cq);
mutex_lock(&pfvf->mbox.lock);