Received-SPF: pass (google.com: domain of linux-kernel+bounces-99460-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
From: Subbaraya Sundeep <sbhatta@marvell.com>
To: <netdev@vger.kernel.org>, <linux-kernel@vger.kernel.org>
CC: <davem@davemloft.net>, <edumazet@google.com>, <kuba@kernel.org>,
        <pabeni@redhat.com>, <sgoutham@marvell.com>, <lcherian@marvell.com>,
        <gakula@marvell.com>, <hkelam@marvell.com>, <naveenm@marvell.com>,
        <horms@kernel.org>, Subbaraya Sundeep <sbhatta@marvell.com>
Subject: [net PATCH 5/5] octeontx2-af: Use separate handlers for interrupts
Date: Tue, 12 Mar 2024 00:46:52 +0530
Message-ID: <1710184612-29027-6-git-send-email-sbhatta@marvell.com>
In-Reply-To: <1710184612-29027-1-git-send-email-sbhatta@marvell.com>
References: <1710184612-29027-1-git-send-email-sbhatta@marvell.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain

For PF to AF interrupt vector and VF to AF vector same
interrupt handler is registered which is causing race condition.
When two interrupts are raised to two CPUs at same time
then two cores serve same event corrupting the data.

Fixes: 7304ac4567bc ("octeontx2-af: Add mailbox IRQ and msg handlers")
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 drivers/net/ethernet/marvell/octeontx2/af/rvu.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu.c
index d2a575a645d8..6a911ea0cff9 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu.c
@@ -2531,10 +2531,9 @@ static void rvu_queue_work(struct mbox_wq_info *mw, int first,
 	}
 }

-static irqreturn_t rvu_mbox_intr_handler(int irq, void *rvu_irq)
+static irqreturn_t rvu_mbox_pf_intr_handler(int irq, void *rvu_irq)
 {
 	struct rvu *rvu = (struct rvu *)rvu_irq;
-	int vfs = rvu->vfs;
 	u64 intr;

 	intr = rvu_read64(rvu, BLKADDR_RVUM, RVU_AF_PFAF_MBOX_INT);
@@ -2548,6 +2547,18 @@ static irqreturn_t rvu_mbox_intr_handler(int irq, void *rvu_irq)

 	rvu_queue_work(&rvu->afpf_wq_info, 0, rvu->hw->total_pfs, intr);

+	return IRQ_HANDLED;
+}
+
+static irqreturn_t rvu_mbox_intr_handler(int irq, void *rvu_irq)
+{
+	struct rvu *rvu = (struct rvu *)rvu_irq;
+	int vfs = rvu->vfs;
+	u64 intr;
+
+	/* Sync with mbox memory region */
+	rmb();
+
 	/* Handle VF interrupts */
 	if (vfs > 64) {
 		intr = rvupf_read64(rvu, RVU_PF_VFPF_MBOX_INTX(1));
@@ -2889,7 +2900,7 @@ static int rvu_register_interrupts(struct rvu *rvu)
 	/* Register mailbox interrupt handler */
 	sprintf(&rvu->irq_name[RVU_AF_INT_VEC_MBOX * NAME_SIZE], "RVUAF Mbox");
 	ret = request_irq(pci_irq_vector(rvu->pdev, RVU_AF_INT_VEC_MBOX),
-			  rvu_mbox_intr_handler, 0,
+			  rvu_mbox_pf_intr_handler, 0,
 			  &rvu->irq_name[RVU_AF_INT_VEC_MBOX * NAME_SIZE], rvu);
 	if (ret) {
 		dev_err(rvu->dev,
--
2.17.1