rework of "fm10k: prevent RCU issues during AER events
Subsystems (2)
PATCH v2 1/1 rework of "fm10k: prevent RCU issues during AER events
This is a rework of the original patch 308f3d5aad0e ("fm10k: prevent RCU issues during AER events", 2016-03-09), and the following email is a direct rework which replaces it in the queue. Thus, the previously mentioned version should be dropped from the queue.
Patch interdiff for review:
drivers/net/ethernet/intel/fm10k/fm10k_pci.c +6 -1
@@ -2353,11 +2353,13 @@ static void fm10k_io_resume(struct pci_dev *pdev)/* reset statistics starting values */ hw->mac.ops.rebind_hw_stats (hw, &interface->stats); +rtnl_lock (); + err =fm10k_init_queueing_scheme (interface);if (err) {dev_err (&interface->pdev->dev,"init_queueing_scheme failed: %d\n" , err); -return ; + goto unlock; }/* reassociate interrupts */
@@ -2374,6 +2376,9 @@ static void fm10k_io_resume(struct pci_dev *pdev)if (!err)netif_device_attach (netdev); + +unlock: +rtnl_unlock (); }static const struct pci_error_handlers fm10k_err_handler = {
Jacob Keller (1):
fm10k: prevent RCU issues during AER events
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
Signature · 1 line
2.8.0.rc1.177.g5628860
PATCH v2 1/1 fm10k: prevent RCU issues during AER events
During an AER action response, we were calling fm10k_close without holding the rtnl_lock() which could lead to possible RCU warnings being produced due to 64bit stat updates among other causes. Similarly, we need rtnl_lock() around fm10k_open during fm10k_io_resume. Follow the same pattern elsewhere in the driver and protect the entire open/close sequence.
---
Notes:
- v2
* add rtnl_lock() calls during fm10k_io_resume as well
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/net/ethernet/intel/fm10k/fm10k_pci.c +10 -1
@@ -2271,6 +2271,8 @@ static pci_ers_result_t fm10k_io_error_detected(struct pci_dev *pdev,if (state == pci_channel_io_perm_failure)return PCI_ERS_RESULT_DISCONNECT ; +rtnl_lock (); +if (netif_running (netdev))fm10k_close (netdev);
@@ -2279,6 +2281,8 @@ static pci_ers_result_t fm10k_io_error_detected(struct pci_dev *pdev,/* free interrupts */ fm10k_clear_queueing_scheme (interface); +rtnl_unlock (); +pci_disable_device (pdev);/* Request a slot reset. */
@@ -2349,11 +2353,13 @@ static void fm10k_io_resume(struct pci_dev *pdev)/* reset statistics starting values */ hw->mac.ops.rebind_hw_stats (hw, &interface->stats); +rtnl_lock (); + err =fm10k_init_queueing_scheme (interface);if (err) {dev_err (&interface->pdev->dev,"init_queueing_scheme failed: %d\n" , err); -return ; + goto unlock; }/* reassociate interrupts */
@@ -2370,6 +2376,9 @@ static void fm10k_io_resume(struct pci_dev *pdev)if (!err)netif_device_attach (netdev); + +unlock: +rtnl_unlock (); }static const struct pci_error_handlers fm10k_err_handler = {
Signature · 1 line
2.8.0.rc1.177.g5628860
Singh
PATCH v2 1/1 fm10k: prevent RCU issues during AER events
-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Jacob Keller
Sent: Friday, March 11, 2016 9:53 AM
To: Intel Wired LAN <[email protected]>
Subject: [Intel-wired-lan] [PATCH v2] fm10k: prevent RCU issues during AER events
During an AER action response, we were calling fm10k_close without holding the rtnl_lock() which could lead to possible RCU warnings being produced due to 64bit stat updates among other causes. Similarly, we need rtnl_lock() around fm10k_open during fm10k_io_resume. Follow the same pattern elsewhere in the driver and protect the entire open/close sequence.
---