From: Shinas Rasheed <srasheed@marvell.com>
To: <netdev@vger.kernel.org>, <linux-kernel@vger.kernel.org>
Cc: <hgani@marvell.com>, <sedara@marvell.com>, <vimleshk@marvell.com>,
<thaller@redhat.com>, <wizhao@redhat.com>, <kheib@redhat.com>,
<egallen@redhat.com>, <konguyen@redhat.com>, <horms@kernel.org>,
<einstein.xue@synaxg.com>, Shinas Rasheed <srasheed@marvell.com>,
Veerasenareddy Burru <vburru@marvell.com>,
Satananda Burla <sburla@marvell.com>,
Andrew Lunn <andrew+netdev@lunn.ch>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Subject: [PATCH net v1 3/4] octeon_ep_vf: fix race conditions in ndo_get_stats64
Date: Mon, 2 Dec 2024 23:21:29 -0800 [thread overview]
Message-ID: <20241203072130.2316913-4-srasheed@marvell.com> (raw)
In-Reply-To: <20241203072130.2316913-1-srasheed@marvell.com>
ndo_get_stats64() can race with ndo_stop(), which frees input and
output queue resources. Implement device state variable to protect
against such resource usage.
Fixes: c3fad23cdc06 ("octeon_ep_vf: add support for ndo ops")
Signed-off-by: Shinas Rasheed <srasheed@marvell.com>
---
.../marvell/octeon_ep_vf/octep_vf_main.c | 29 +++++++++++++++++++
.../marvell/octeon_ep_vf/octep_vf_main.h | 9 ++++++
2 files changed, 38 insertions(+)
diff --git a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.c b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.c
index 7e6771c9cdbb..12b95fb21e64 100644
--- a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.c
+++ b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.c
@@ -497,6 +497,8 @@ static int octep_vf_open(struct net_device *netdev)
if (ret)
octep_vf_link_up(netdev);
+ set_bit(OCTEP_VF_DEV_STATE_OPEN, &oct->state);
+
return 0;
set_queues_err:
@@ -511,6 +513,11 @@ static int octep_vf_open(struct net_device *netdev)
return -1;
}
+static bool octep_vf_drv_busy(struct octep_vf_device *oct)
+{
+ return test_bit(OCTEP_VF_DEV_STATE_READ_STATS, &oct->state);
+}
+
/**
* octep_vf_stop() - stop the octeon network device.
*
@@ -525,6 +532,14 @@ static int octep_vf_stop(struct net_device *netdev)
netdev_info(netdev, "Stopping the device ...\n");
+ clear_bit(OCTEP_VF_DEV_STATE_OPEN, &oct->state);
+ /* Make sure device state open is cleared so that no more
+ * stats fetch can happen intermittently
+ */
+ smp_mb__after_atomic();
+ while (octep_vf_drv_busy(oct))
+ msleep(20);
+
/* Stop Tx from stack */
netif_carrier_off(netdev);
netif_tx_disable(netdev);
@@ -782,6 +797,16 @@ static void octep_vf_get_stats64(struct net_device *netdev,
u64 tx_packets, tx_bytes, rx_packets, rx_bytes;
int q;
+ set_bit(OCTEP_VF_DEV_STATE_READ_STATS, &oct->state);
+ /* Make sure read stats state is set, so that ndo_stop
+ * doesn't clear resources as they are read
+ */
+ smp_mb__after_atomic();
+ if (!test_bit(OCTEP_VF_DEV_STATE_OPEN, &oct->state)) {
+ clear_bit(OCTEP_VF_DEV_STATE_READ_STATS, &oct->state);
+ return;
+ }
+
tx_packets = 0;
tx_bytes = 0;
rx_packets = 0;
@@ -807,6 +832,7 @@ static void octep_vf_get_stats64(struct net_device *netdev,
stats->rx_missed_errors = oct->iface_rx_stats.dropped_pkts_fifo_full;
stats->tx_dropped = oct->iface_tx_stats.dropped;
}
+ clear_bit(OCTEP_VF_DEV_STATE_READ_STATS, &oct->state);
}
/**
@@ -1140,6 +1166,9 @@ static int octep_vf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
dev_err(&pdev->dev, "Failed to register netdev\n");
goto delete_mbox;
}
+
+ clear_bit(OCTEP_VF_DEV_STATE_OPEN, &octep_vf_dev->state);
+
dev_info(&pdev->dev, "Device probe successful\n");
return 0;
diff --git a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.h b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.h
index 5769f62545cd..692b3bf94722 100644
--- a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.h
+++ b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_main.h
@@ -213,6 +213,12 @@ struct octep_vf_fw_info {
u16 tx_ol_flags;
};
+/* Device state */
+enum octep_vf_dev_state {
+ OCTEP_VF_DEV_STATE_OPEN,
+ OCTEP_VF_DEV_STATE_READ_STATS,
+};
+
/* The Octeon device specific private data structure.
* Each Octeon device has this structure to represent all its components.
*/
@@ -282,6 +288,9 @@ struct octep_vf_device {
/* offset for iface stats */
u32 ctrl_mbox_ifstats_offset;
+ /* Device state */
+ unsigned long state;
+
/* Negotiated Mbox version */
u32 mbox_neg_ver;
--
2.25.1
next prev parent reply other threads:[~2024-12-03 7:21 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-12-03 7:21 [PATCH net v1 0/4] Fix " Shinas Rasheed
2024-12-03 7:21 ` [PATCH net v1 1/4] octeon_ep: fix " Shinas Rasheed
2024-12-04 2:23 ` Jakub Kicinski
2024-12-03 7:21 ` [PATCH net v1 2/4] octeon_ep: remove firmware stats fetch " Shinas Rasheed
2024-12-03 7:21 ` Shinas Rasheed [this message]
2024-12-03 7:21 ` [PATCH net v1 4/4] octeon_ep_vf: " Shinas Rasheed
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=20241203072130.2316913-4-srasheed@marvell.com \
--to=srasheed@marvell.com \
--cc=andrew+netdev@lunn.ch \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=egallen@redhat.com \
--cc=einstein.xue@synaxg.com \
--cc=hgani@marvell.com \
--cc=horms@kernel.org \
--cc=kheib@redhat.com \
--cc=konguyen@redhat.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sburla@marvell.com \
--cc=sedara@marvell.com \
--cc=thaller@redhat.com \
--cc=vburru@marvell.com \
--cc=vimleshk@marvell.com \
--cc=wizhao@redhat.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
all inboxes | Powered by JetHome®