Skip to contentSkip to conversation

i40e: fix for flow director counters not wrapping as expected

PATCH New Alice Michael 2 participants 20 messages Jul 14, 2017 – Aug 15, 2017

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/10 i40e: fix for flow director counters not wrapping as expected

From: Mariusz Stachura <[email protected]>

An eratta with GLQF_PCNT causes it to not wrap as expected. This can cause an error in flow director statistics. This patch resets affected counters just after reading.

Testing Hints (required if no HSD): DCR 2187

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 35 +++++++++++++++++++----------
 1 file changed, 23 insertions(+), 12 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +23 -12
index 647a7a6..43ed431 100644
@@ -600,6 +600,20 @@ static void i40e_stat_update32(struct i40e_hw *hw, u32 reg,
 }
 
 /**
+ * i40e_stat_update_and_clear32 - read and clear hw reg, update a 32 bit stat
+ * @hw: ptr to the hardware info
+ * @reg: the hw reg to read and clear
+ * @stat: ptr to the stat
+ **/
+static void i40e_stat_update_and_clear32(struct i40e_hw *hw, u32 reg, u64 *stat)
+{
+	u32 new_data = rd32(hw, reg);
+
+	wr32(hw, reg, 1); /* must write a nonzero value to clear register */
+	*stat += new_data;
+}
+
+/**
  * i40e_update_eth_stats - Update VSI-specific ethernet statistics counters.
  * @vsi: the VSI to be updated
  **/
@@ -1040,18 +1054,15 @@ static void i40e_update_pf_stats(struct i40e_pf *pf)
 			   &osd->rx_jabber, &nsd->rx_jabber);
 
 	/* FDIR stats */
-	i40e_stat_update32(hw,
-			   I40E_GLQF_PCNT(I40E_FD_ATR_STAT_IDX(pf->hw.pf_id)),
-			   pf->stat_offsets_loaded,
-			   &osd->fd_atr_match, &nsd->fd_atr_match);
-	i40e_stat_update32(hw,
-			   I40E_GLQF_PCNT(I40E_FD_SB_STAT_IDX(pf->hw.pf_id)),
-			   pf->stat_offsets_loaded,
-			   &osd->fd_sb_match, &nsd->fd_sb_match);
-	i40e_stat_update32(hw,
-		      I40E_GLQF_PCNT(I40E_FD_ATR_TUNNEL_STAT_IDX(pf->hw.pf_id)),
-		      pf->stat_offsets_loaded,
-		      &osd->fd_atr_tunnel_match, &nsd->fd_atr_tunnel_match);
+	i40e_stat_update_and_clear32(hw,
+			I40E_GLQF_PCNT(I40E_FD_ATR_STAT_IDX(hw->pf_id)),
+			&nsd->fd_atr_match);
+	i40e_stat_update_and_clear32(hw,
+			I40E_GLQF_PCNT(I40E_FD_SB_STAT_IDX(hw->pf_id)),
+			&nsd->fd_sb_match);
+	i40e_stat_update_and_clear32(hw,
+			I40E_GLQF_PCNT(I40E_FD_ATR_TUNNEL_STAT_IDX(hw->pf_id)),
+			&nsd->fd_atr_tunnel_match);
 
 	val = rd32(hw, I40E_PRTPM_EEE_STAT);
 	nsd->tx_lpi_status =
Signature · 1 line
2.9.3
#1 lore raw

PATCH 2/10 i40evf: lower message level

From: Mitch Williams <[email protected]>

We see this message regularly on VF reset or unload (which invokes a reset). It's essentially meaningless unless it's happening constantly. To prevent consternation, lower the log level to debug so it's not seen under normal circumstance.

Testing Hints (required if no HSD): Load and unload VF driver, observe dmesg in VM.

---
 drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c +1 -1
index 6c403bf..d22813e 100644
@@ -52,7 +52,7 @@ static int i40evf_send_pf_msg(struct i40evf_adapter *adapter,
 
 	err = i40e_aq_send_msg_to_pf(hw, op, 0, msg, len, NULL);
 	if (err)
-		dev_err(&adapter->pdev->dev, "Unable to send opcode %d to PF, err %s, aq_err %s\n",
+		dev_dbg(&adapter->pdev->dev, "Unable to send opcode %d to PF, err %s, aq_err %s\n",
 			op, i40evf_stat_str(hw, err),
 			i40evf_aq_str(hw, hw->aq.asq_last_status));
 	return err;
Signature · 1 line
2.9.3
#2 lore raw

PATCH 2/10 i40evf: lower message level

Quoted 17 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired-
lan at lists.osuosl.org
Cc: Williams, Mitch A <[email protected]>
Subject: [Intel-wired-lan] [next PATCH S77-V3 02/10] i40evf: lower message level

From: Mitch Williams <[email protected]>

We see this message regularly on VF reset or unload (which invokes a reset). It's essentially meaningless unless it's happening constantly.
To prevent consternation, lower the log level to debug so it's not seen under normal circumstance.

Quoted 7 lines

Testing Hints (required if no HSD): Load and unload VF driver, observe dmesg in VM.

---
 drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
#3 lore raw

PATCH 3/10 i40e: use separate state bit for miscellaneous IRQ setup

In reply to Alice Michael

From: Jacob Keller <[email protected]>

We currently (mis)use the __I40E_RECOVERY_PENDING bit to determine when we should actually request a new IRQ in i40e_setup_misc_vector().

This led to a design mistake where we open-coded the re-setup of the miscellaneous vector in i40e_resume() instead of using the function provided. If we did not open-code this and instead tried to use the i40e_setup_misc_vector() function, it would lead to never reallocating the IRQ.

This would lead to a second i40e_suspend() call failing to free the vector due to a NULL pointer dereference.

A future patch is going to re-work how the i40e_suspend() and i40e_resume() flows work to clear all IRQ vectors, which would require us to use i40e_setup_misc_vector() directly. Since during this time the __I40E_RECOVERY_PENDING bit is set, we'll never re-allocate the vector.

Rather than leaving the open-coded setup in i40e_resume() lets just fix the problem properly in i40e_setup_misc_vector().

Introduce a new state bit which indicates when the IRQ has been assigned, which will be set when i40e_setup_misc_vector is first called. This ultimately resolves the issue of re-requesting the vector, without overloading the __I40E_RECOVERY_PENDING state. This ensures that the suspend/resume cycle can use the setup function instead of open-coding the re-request during resume.

Additionally, since the only callers of i40e_stop_misc_vector also want to free it, move this code directly into the function to avoid duplication. Due to the new functionality, rename it to i40e_free_misc_vector().

This lets us drop the extra calls to free and re-enable the vector during i40e_suspend() and i40e_resume(). We don't need to call i40e_setup_misc_Vector() in i40e_resume() because it gets called by the i40e_rebuild() call.

---
 drivers/net/ethernet/intel/i40e/i40e.h      |  1 +
 drivers/net/ethernet/intel/i40e/i40e_main.c | 39 +++++++++++------------------
 2 files changed, 15 insertions(+), 25 deletions(-)
drivers/net/ethernet/intel/i40e/i40e.h +1 -0
index 5f75e67..2f1f65e 100644
@@ -136,6 +136,7 @@ enum i40e_state_t {
 	__I40E_MDD_EVENT_PENDING,
 	__I40E_VFLR_EVENT_PENDING,
 	__I40E_RESET_RECOVERY_PENDING,
+	__I40E_MISC_IRQ_REQUESTED,
 	__I40E_RESET_INTR_RECEIVED,
 	__I40E_REINIT_REQUESTED,
 	__I40E_PF_RESET_REQUESTED,
drivers/net/ethernet/intel/i40e/i40e_main.c +14 -25
index 43ed431..c41815f 100644
@@ -3605,14 +3605,20 @@ static int i40e_vsi_enable_irq(struct i40e_vsi *vsi)
 }
 
 /**
- * i40e_stop_misc_vector - Stop the vector that handles non-queue events
+ * i40e_free_misc_vector - Free the vector that handles non-queue events
  * @pf: board private structure
  **/
-static void i40e_stop_misc_vector(struct i40e_pf *pf)
+static void i40e_free_misc_vector(struct i40e_pf *pf)
 {
 	/* Disable ICR 0 */
 	wr32(&pf->hw, I40E_PFINT_ICR0_ENA, 0);
 	i40e_flush(&pf->hw);
+
+	if (pf->flags & I40E_FLAG_MSIX_ENABLED && pf->msix_entries) {
+		synchronize_irq(pf->msix_entries[0].vector);
+		free_irq(pf->msix_entries[0].vector, pf);
+		clear_bit(__I40E_MISC_IRQ_REQUESTED, pf->state);
+	}
 }
 
 /**
@@ -4467,11 +4473,7 @@ static void i40e_clear_interrupt_scheme(struct i40e_pf *pf)
 {
 	int i;
 
-	i40e_stop_misc_vector(pf);
-	if (pf->flags & I40E_FLAG_MSIX_ENABLED && pf->msix_entries) {
-		synchronize_irq(pf->msix_entries[0].vector);
-		free_irq(pf->msix_entries[0].vector, pf);
-	}
+	i40e_free_misc_vector(pf);
 
 	i40e_put_lump(pf->irq_pile, pf->iwarp_base_vector,
 		      I40E_IWARP_IRQ_PILE_ID);
@@ -8351,13 +8353,12 @@ static int i40e_setup_misc_vector(struct i40e_pf *pf)
 	struct i40e_hw *hw = &pf->hw;
 	int err = 0;
 
-	/* Only request the irq if this is the first time through, and
-	 * not when we're rebuilding after a Reset
-	 */
-	if (!test_bit(__I40E_RESET_RECOVERY_PENDING, pf->state)) {
+	/* Only request the IRQ once, the first time through. */
+	if (!test_and_set_bit(__I40E_MISC_IRQ_REQUESTED, pf->state)) {
 		err = request_irq(pf->msix_entries[0].vector,
 				  i40e_intr, 0, pf->int_name, pf);
 		if (err) {
+			clear_bit(__I40E_MISC_IRQ_REQUESTED, pf->state);
 			dev_info(&pf->pdev->dev,
 				 "request_irq for %s failed: %d\n",
 				 pf->int_name, err);
@@ -12055,11 +12056,8 @@ static int i40e_suspend(struct pci_dev *pdev, pm_message_t state)
 	wr32(hw, I40E_PFPM_APM, (pf->wol_en ? I40E_PFPM_APM_APME_MASK : 0));
 	wr32(hw, I40E_PFPM_WUFC, (pf->wol_en ? I40E_PFPM_WUFC_MAG_MASK : 0));
 
-	i40e_stop_misc_vector(pf);
-	if (pf->msix_entries) {
-		synchronize_irq(pf->msix_entries[0].vector);
-		free_irq(pf->msix_entries[0].vector, pf);
-	}
+	i40e_free_misc_vector(pf);
+
 	retval = pci_save_state(pdev);
 	if (retval)
 		return retval;
@@ -12099,15 +12097,6 @@ static int i40e_resume(struct pci_dev *pdev)
 	/* handling the reset will rebuild the device state */
 	if (test_and_clear_bit(__I40E_SUSPENDED, pf->state)) {
 		clear_bit(__I40E_DOWN, pf->state);
-		if (pf->msix_entries) {
-			err = request_irq(pf->msix_entries[0].vector,
-					  i40e_intr, 0, pf->int_name, pf);
-			if (err) {
-				dev_err(&pf->pdev->dev,
-					"request_irq for %s failed: %d\n",
-					pf->int_name, err);
-			}
-		}
 		i40e_reset_and_rebuild(pf, false, false);
 	}
 
Signature · 1 line
2.9.3
#4 lore raw

PATCH 3/10 i40e: use separate state bit for miscellaneous IRQ setup

Quoted 47 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 03/10] i40e: use separate state bit for miscellaneous IRQ setup

From: Jacob Keller <[email protected]>

We currently (mis)use the __I40E_RECOVERY_PENDING bit to determine when we should actually request a new IRQ in i40e_setup_misc_vector().

This led to a design mistake where we open-coded the re-setup of the miscellaneous vector in i40e_resume() instead of using the function provided. If we did not open-code this and instead tried to use the i40e_setup_misc_vector() function, it would lead to never reallocating the IRQ.

This would lead to a second i40e_suspend() call failing to free the vector due to a NULL pointer dereference.

A future patch is going to re-work how the i40e_suspend() and i40e_resume() flows work to clear all IRQ vectors, which would require us to use i40e_setup_misc_vector() directly. Since during this time the __I40E_RECOVERY_PENDING bit is set, we'll never re-allocate the vector.

Rather than leaving the open-coded setup in i40e_resume() lets just fix the problem properly in i40e_setup_misc_vector().

Introduce a new state bit which indicates when the IRQ has been assigned, which will be set when i40e_setup_misc_vector is first called. This ultimately resolves the issue of re-requesting the vector, without overloading the __I40E_RECOVERY_PENDING state. This ensures that the suspend/resume cycle can use the setup function instead of open-coding the re-request during resume.

Additionally, since the only callers of i40e_stop_misc_vector also want to free it, move this code directly into the function to avoid duplication. Due to the new functionality, rename it to i40e_free_misc_vector().

This lets us drop the extra calls to free and re-enable the vector during i40e_suspend() and i40e_resume(). We don't need to call i40e_setup_misc_Vector() in i40e_resume() because it gets called by the i40e_rebuild() call.

Quoted 6 lines
---
 drivers/net/ethernet/intel/i40e/i40e.h      |  1 +
 drivers/net/ethernet/intel/i40e/i40e_main.c | 39 +++++++++++---------------
---
 2 files changed, 15 insertions(+), 25 deletions(-)
#5 lore raw

PATCH 4/10 i40e: use newer generic PM support instead of legacy PM callbacks

In reply to Alice Michael

From: Jacob Keller <[email protected]>

Stop using the old legacy PM support, since we now have stable support for the newer generic PM callbacks.

This has several advantages. First, we no longer have to manage our own pci_save_state() and power changes, as it's preferred to have the PCI stack do this. Second, these routines get called for both hibernate and suspend to ram, so we can have the driver properly handle all the suspend/resume flows that it needs to.

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 54 +++++++++--------------------
 1 file changed, 17 insertions(+), 37 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +17 -37
index c41815f..b47cc14 100644
@@ -12036,14 +12036,14 @@ static void i40e_shutdown(struct pci_dev *pdev)
 
 #ifdef CONFIG_PM
 /**
- * i40e_suspend - PCI callback for moving to D3
- * @pdev: PCI device information struct
+ * i40e_suspend - PM callback for moving to D3
+ * @dev: generic device information structure
  **/
-static int i40e_suspend(struct pci_dev *pdev, pm_message_t state)
+static int i40e_suspend(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct i40e_pf *pf = pci_get_drvdata(pdev);
 	struct i40e_hw *hw = &pf->hw;
-	int retval = 0;
 
 	set_bit(__I40E_SUSPENDED, pf->state);
 	set_bit(__I40E_DOWN, pf->state);
@@ -12058,41 +12058,17 @@ static int i40e_suspend(struct pci_dev *pdev, pm_message_t state)
 
 	i40e_free_misc_vector(pf);
 
-	retval = pci_save_state(pdev);
-	if (retval)
-		return retval;
-
-	pci_wake_from_d3(pdev, pf->wol_en);
-	pci_set_power_state(pdev, PCI_D3hot);
-
-	return retval;
+	return 0;
 }
 
 /**
- * i40e_resume - PCI callback for waking up from D3
- * @pdev: PCI device information struct
+ * i40e_resume - PM callback for waking up from D3
+ * @dev: generic device information structure
  **/
-static int i40e_resume(struct pci_dev *pdev)
+static int i40e_resume(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct i40e_pf *pf = pci_get_drvdata(pdev);
-	u32 err;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	/* pci_restore_state() clears dev->state_saves, so
-	 * call pci_save_state() again to restore it.
-	 */
-	pci_save_state(pdev);
-
-	err = pci_enable_device_mem(pdev);
-	if (err) {
-		dev_err(&pdev->dev, "Cannot enable PCI device from suspend\n");
-		return err;
-	}
-	pci_set_master(pdev);
-
-	/* no wakeup events while running */
-	pci_wake_from_d3(pdev, false);
 
 	/* handling the reset will rebuild the device state */
 	if (test_and_clear_bit(__I40E_SUSPENDED, pf->state)) {
@@ -12103,22 +12079,26 @@ static int i40e_resume(struct pci_dev *pdev)
 	return 0;
 }
 
-#endif
+#endif /* CONFIG_PM */
+
 static const struct pci_error_handlers i40e_err_handler = {
 	.error_detected = i40e_pci_error_detected,
 	.slot_reset = i40e_pci_error_slot_reset,
 	.resume = i40e_pci_error_resume,
 };
 
+static SIMPLE_DEV_PM_OPS(i40e_pm_ops, i40e_suspend, i40e_resume);
+
 static struct pci_driver i40e_driver = {
 	.name     = i40e_driver_name,
 	.id_table = i40e_pci_tbl,
 	.probe    = i40e_probe,
 	.remove   = i40e_remove,
 #ifdef CONFIG_PM
-	.suspend  = i40e_suspend,
-	.resume   = i40e_resume,
-#endif
+	.driver   = {
+		.pm = &i40e_pm_ops,
+	},
+#endif /* CONFIG_PM */
 	.shutdown = i40e_shutdown,
 	.err_handler = &i40e_err_handler,
 	.sriov_configure = i40e_pci_sriov_configure,
Signature · 1 line
2.9.3
#6 lore raw

PATCH 4/10 i40e: use newer generic PM support instead of legacy PM callbacks

Quoted 20 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 04/10] i40e: use newer generic PM support instead of legacy PM callbacks

From: Jacob Keller <[email protected]>

Stop using the old legacy PM support, since we now have stable support for the newer generic PM callbacks.

This has several advantages. First, we no longer have to manage our own pci_save_state() and power changes, as it's preferred to have the PCI stack do this. Second, these routines get called for both hibernate and suspend to ram, so we can have the driver properly handle all the suspend/resume flows that it needs to.

Quoted 5 lines
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 54 +++++++++------------------
--
 1 file changed, 17 insertions(+), 37 deletions(-)
#7 lore raw

PATCH 5/10 i40e: don't clear suspended state until we finish resuming

In reply to Alice Michael

From: Jacob Keller <[email protected]>

When handling suspend and resume callbacks we want to make sure that (a) we don't suspend again if we're already suspended and (b) we don't resume again if we're already resuming. Lets make sure we test_and_set the __I40E_SUSPENDED bit in i40e_suspend which ensures that a suspend call when already suspended will exit early. Additionally, if __I40E_SUSPENDED is not set when we begin resuming, exit early as well.

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +13 -6
index b47cc14..11c9565 100644
@@ -12045,7 +12045,10 @@ static int i40e_suspend(struct device *dev)
 	struct i40e_pf *pf = pci_get_drvdata(pdev);
 	struct i40e_hw *hw = &pf->hw;
 
-	set_bit(__I40E_SUSPENDED, pf->state);
+	/* If we're already suspended, then there is nothing to do */
+	if (test_and_set_bit(__I40E_SUSPENDED, pf->state))
+		return 0;
+
 	set_bit(__I40E_DOWN, pf->state);
 
 	if (pf->wol_en && (pf->hw_features & I40E_HW_WOL_MC_MAGIC_PKT_WAKE))
@@ -12070,11 +12073,15 @@ static int i40e_resume(struct device *dev)
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct i40e_pf *pf = pci_get_drvdata(pdev);
 
-	/* handling the reset will rebuild the device state */
-	if (test_and_clear_bit(__I40E_SUSPENDED, pf->state)) {
-		clear_bit(__I40E_DOWN, pf->state);
-		i40e_reset_and_rebuild(pf, false, false);
-	}
+	/* If we're not suspended, then there is nothing to do */
+	if (!test_bit(__I40E_SUSPENDED, pf->state))
+		return 0;
+
+	clear_bit(__I40E_DOWN, pf->state);
+	i40e_reset_and_rebuild(pf, false, false);
+
+	/* Clear suspended state last after everything is recovered */
+	clear_bit(__I40E_SUSPENDED, pf->state);
 
 	return 0;
 }
Signature · 1 line
2.9.3
#8 lore raw

PATCH 5/10 i40e: don't clear suspended state until we finish resuming

Quoted 18 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired- lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 05/10] i40e: don't clear suspended state until we finish resuming

From: Jacob Keller <[email protected]>

When handling suspend and resume callbacks we want to make sure that (a) we don't suspend again if we're already suspended and (b) we don't resume again if we're already resuming. Lets make sure we test_and_set the __I40E_SUSPENDED bit in i40e_suspend which ensures that a suspend call when already suspended will exit early. Additionally, if __I40E_SUSPENDED is not set when we begin resuming, exit early as well.

Quoted 4 lines
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)
#9 lore raw

PATCH 6/10 i40e: prevent service task from running while we're suspended

In reply to Alice Michael

From: Jacob Keller <[email protected]>

Although the service task does check the suspended status before running, it might already be part way through running when we go to suspend. Lets ensure that the service task is stopped and will not be restarted again until we finish resuming. This ensures that service task code does not cause strange interactions with the suspend/resume handlers.

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 8 ++++++++
 1 file changed, 8 insertions(+)
drivers/net/ethernet/intel/i40e/i40e_main.c +8 -0
index 11c9565..0d2e652 100644
@@ -12051,6 +12051,10 @@ static int i40e_suspend(struct device *dev)
 
 	set_bit(__I40E_DOWN, pf->state);
 
+	/* Ensure service task will not be running */
+	del_timer_sync(&pf->service_timer);
+	cancel_work_sync(&pf->service_task);
+
 	if (pf->wol_en && (pf->hw_features & I40E_HW_WOL_MC_MAGIC_PKT_WAKE))
 		i40e_enable_mc_magic_wake(pf);
 
@@ -12083,6 +12087,10 @@ static int i40e_resume(struct device *dev)
 	/* Clear suspended state last after everything is recovered */
 	clear_bit(__I40E_SUSPENDED, pf->state);
 
+	/* Restart the service task */
+	mod_timer(&pf->service_timer,
+		  round_jiffies(jiffies + pf->service_timer_period));
+
 	return 0;
 }
 
Signature · 1 line
2.9.3
#10 lore raw

PATCH 6/10 i40e: prevent service task from running while we're suspended

Quoted 17 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 06/10] i40e: prevent service task from running while we're suspended

From: Jacob Keller <[email protected]>

Although the service task does check the suspended status before running, it might already be part way through running when we go to suspend. Lets ensure that the service task is stopped and will not be restarted again until we finish resuming. This ensures that service task code does not cause strange interactions with the suspend/resume handlers.

Quoted 4 lines
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 8 ++++++++
 1 file changed, 8 insertions(+)
#11 lore raw

PATCH 7/10 i40e: shutdown all IRQs and disable MSI-X when suspended

In reply to Alice Michael

From: Jacob Keller <[email protected]>

On some platforms with a large number of CPUs, we will allocate many IRQ vectors. When hibernating, the system will attempt to migrate all of the vectors back to CPU0 when shutting down all the other CPUs. It is possible that we have so many vectors that it cannot re-assign them to CPU0. This is even more likely if we have many devices installed in one platform.

The end result is failure to hibernate, as it is not possible to shutdown the CPUs. We can avoid this by disabling MSI-X and clearing our interrupt scheme when the device is suspended. A more ideal solution would be some method for the stack to properly handle this for all drivers, rather than on a case-by-case basis for each driver to fix itself.

However, until this more ideal solution exists, we can do our part and shutdown our IRQs during suspend, which should allow systems with a large number of CPUs to safely suspend or hibernate.

It may be worth investigating if we should shut down even further when we suspend as it may make the path cleaner, but this was the minimum fix for the hibernation issue mentioned here.

Testing-hints:
This affects systems with a large number of CPUs, and with multiple devices enabled. Without this change, those platforms are unable to hibernate at all.

---
 drivers/net/ethernet/intel/i40e/i40e_main.c     | 68 ++++++++++++++++++++++++-
 drivers/net/ethernet/intel/i40evf/i40evf_main.c |  2 +-
 2 files changed, 68 insertions(+), 2 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_main.c +67 -1
index 0d2e652..1ad014d 100644
@@ -8340,6 +8340,57 @@ static int i40e_init_interrupt_scheme(struct i40e_pf *pf)
 	return 0;
 }
 
+#ifdef CONFIG_PM
+/**
+ * i40e_restore_interrupt_scheme - Restore the interrupt scheme
+ * @pf: private board data structure
+ *
+ * Restore the interrupt scheme that was cleared when we suspended the
+ * device. This should be called during resume to re-allocate the q_vectors
+ * and reacquire IRQs.
+ */
+static int i40e_restore_interrupt_scheme(struct i40e_pf *pf)
+{
+	int err, i;
+
+	/* We cleared the MSI and MSI-X flags when disabling the old interrupt
+	 * scheme. We need to re-enabled them here in order to attempt to
+	 * re-acquire the MSI or MSI-X vectors
+	 */
+	pf->flags |= (I40E_FLAG_MSIX_ENABLED | I40E_FLAG_MSI_ENABLED);
+
+	err = i40e_init_interrupt_scheme(pf);
+	if (err)
+		return err;
+
+	/* Now that we've re-acquired IRQs, we need to remap the vectors and
+	 * rings together again.
+	 */
+	for (i = 0; i < pf->num_alloc_vsi; i++) {
+		if (pf->vsi[i]) {
+			err = i40e_vsi_alloc_q_vectors(pf->vsi[i]);
+			if (err)
+				goto err_unwind;
+			i40e_vsi_map_rings_to_vectors(pf->vsi[i]);
+		}
+	}
+
+	err = i40e_setup_misc_vector(pf);
+	if (err)
+		goto err_unwind;
+
+	return 0;
+
+err_unwind:
+	while (i--) {
+		if (pf->vsi[i])
+			i40e_vsi_free_q_vectors(pf->vsi[i]);
+	}
+
+	return err;
+}
+#endif /* CONFIG_PM */
+
 /**
  * i40e_setup_misc_vector - Setup the misc vector to handle non queue events
  * @pf: board private structure
@@ -12063,7 +12114,12 @@ static int i40e_suspend(struct device *dev)
 	wr32(hw, I40E_PFPM_APM, (pf->wol_en ? I40E_PFPM_APM_APME_MASK : 0));
 	wr32(hw, I40E_PFPM_WUFC, (pf->wol_en ? I40E_PFPM_WUFC_MAG_MASK : 0));
 
-	i40e_free_misc_vector(pf);
+	/* Clear the interrupt scheme and release our IRQs so that the system
+	 * can safely hibernate even when there are a large number of CPUs.
+	 * Otherwise hibernation might fail when mapping all the vectors back
+	 * to CPU0.
+	 */
+	i40e_clear_interrupt_scheme(pf);
 
 	return 0;
 }
@@ -12076,11 +12132,21 @@ static int i40e_resume(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct i40e_pf *pf = pci_get_drvdata(pdev);
+	int err;
 
 	/* If we're not suspended, then there is nothing to do */
 	if (!test_bit(__I40E_SUSPENDED, pf->state))
 		return 0;
 
+	/* We cleared the interrupt scheme when we suspended, so we need to
+	 * restore it now to resume device functionality.
+	 */
+	err = i40e_restore_interrupt_scheme(pf);
+	if (err) {
+		dev_err(&pdev->dev, "Cannot restore interrupt scheme: %d\n",
+			err);
+	}
+
 	clear_bit(__I40E_DOWN, pf->state);
 	i40e_reset_and_rebuild(pf, false, false);
 
drivers/net/ethernet/intel/i40evf/i40evf_main.c +1 -1
index 29953c9..d8e17c2 100644
@@ -46,7 +46,7 @@ static const char i40evf_driver_string[] =
 
 #define DRV_VERSION_MAJOR 3
 #define DRV_VERSION_MINOR 0
-#define DRV_VERSION_BUILD 0
+#define DRV_VERSION_BUILD 1
 #define DRV_VERSION __stringify(DRV_VERSION_MAJOR) "." \
 	     __stringify(DRV_VERSION_MINOR) "." \
 	     __stringify(DRV_VERSION_BUILD) \
Signature · 1 line
2.9.3
#12 lore raw

PATCH 7/10 i40e: shutdown all IRQs and disable MSI-X when suspended

Quoted 36 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 07/10] i40e: shutdown all IRQs and disable MSI-X when suspended

From: Jacob Keller <[email protected]>

On some platforms with a large number of CPUs, we will allocate many IRQ vectors. When hibernating, the system will attempt to migrate all of the vectors back to CPU0 when shutting down all the other CPUs. It is possible that we have so many vectors that it cannot re-assign them to CPU0. This is even more likely if we have many devices installed in one platform.

The end result is failure to hibernate, as it is not possible to shutdown the CPUs. We can avoid this by disabling MSI-X and clearing our interrupt scheme when the device is suspended. A more ideal solution would be some method for the stack to properly handle this for all drivers, rather than on a case-by- case basis for each driver to fix itself.

However, until this more ideal solution exists, we can do our part and shutdown our IRQs during suspend, which should allow systems with a large number of CPUs to safely suspend or hibernate.

It may be worth investigating if we should shut down even further when we suspend as it may make the path cleaner, but this was the minimum fix for the hibernation issue mentioned here.

Testing-hints:
This affects systems with a large number of CPUs, and with multiple devices enabled. Without this change, those platforms are unable to hibernate at all.

Quoted 6 lines
---
 drivers/net/ethernet/intel/i40e/i40e_main.c     | 68
++++++++++++++++++++++++-
 drivers/net/ethernet/intel/i40evf/i40evf_main.c |  2 +-
 2 files changed, 68 insertions(+), 2 deletions(-)
#13 lore raw

PATCH 8/10 i40evf: fix ring to vector mapping

In reply to Alice Michael

From: Alan Brady <[email protected]>

The current implementation for mapping queues to vectors is broken because it attempts to map each tx and rx ring to its own vector, however we use combined queues so we should actually be mapping the tx/rx rings together on one vector.

Also in the current implementation, in the case where we have more queues than vectors, we attempt to group the queues together into 'chunks' and map each 'chunk' of queues to a vector. Chunking them together would be more ideal if, and only if, we only had RSS because of the way the hashing algorithm works but in the case of a future patch that enables VF ADq, round robin assignment is better and still works with RSS.

This patch resolves both those issues and simplifies the code needed to accomplish this. Instead of treating the case where we have more queues than vectors as special, if we notice our vector index is greater than vectors, reset the vector index to zero and continue mapping. This should ensure that in both cases, whether we have enough vectors for each queue or not, the queues get appropriately mapped.

Testing-hints:
Make sure traffic continues to be received and features like RSS continue to work as expected.

---
 drivers/net/ethernet/intel/i40evf/i40evf_main.c | 48 ++++++-------------------
 1 file changed, 10 insertions(+), 38 deletions(-)
drivers/net/ethernet/intel/i40evf/i40evf_main.c +10 -38
index d8e17c2..551fba6 100644
@@ -432,52 +432,24 @@ i40evf_map_vector_to_txq(struct i40evf_adapter *adapter, int v_idx, int t_idx)
  **/
 static int i40evf_map_rings_to_vectors(struct i40evf_adapter *adapter)
 {
+	int rings_remaining = adapter->num_active_queues;
+	int ridx = 0, vidx = 0;
 	int q_vectors;
-	int v_start = 0;
-	int rxr_idx = 0, txr_idx = 0;
-	int rxr_remaining = adapter->num_active_queues;
-	int txr_remaining = adapter->num_active_queues;
-	int i, j;
-	int rqpv, tqpv;
 	int err = 0;
 
 	q_vectors = adapter->num_msix_vectors - NONQ_VECS;
 
-	/* The ideal configuration...
-	 * We have enough vectors to map one per queue.
-	 */
-	if (q_vectors >= (rxr_remaining * 2)) {
-		for (; rxr_idx < rxr_remaining; v_start++, rxr_idx++)
-			i40evf_map_vector_to_rxq(adapter, v_start, rxr_idx);
+	for (; ridx < rings_remaining; ridx++) {
+		i40evf_map_vector_to_rxq(adapter, vidx, ridx);
+		i40evf_map_vector_to_txq(adapter, vidx, ridx);
 
-		for (; txr_idx < txr_remaining; v_start++, txr_idx++)
-			i40evf_map_vector_to_txq(adapter, v_start, txr_idx);
-		goto out;
-	}
-
-	/* If we don't have enough vectors for a 1-to-1
-	 * mapping, we'll have to group them so there are
-	 * multiple queues per vector.
-	 * Re-adjusting *qpv takes care of the remainder.
-	 */
-	for (i = v_start; i < q_vectors; i++) {
-		rqpv = DIV_ROUND_UP(rxr_remaining, q_vectors - i);
-		for (j = 0; j < rqpv; j++) {
-			i40evf_map_vector_to_rxq(adapter, i, rxr_idx);
-			rxr_idx++;
-			rxr_remaining--;
-		}
-	}
-	for (i = v_start; i < q_vectors; i++) {
-		tqpv = DIV_ROUND_UP(txr_remaining, q_vectors - i);
-		for (j = 0; j < tqpv; j++) {
-			i40evf_map_vector_to_txq(adapter, i, txr_idx);
-			txr_idx++;
-			txr_remaining--;
-		}
+		/* In the case where we have more queues than vectors, continue
+		 * round-robin on vectors until all queues are mapped.
+		 */
+		if (++vidx >= q_vectors)
+			vidx = 0;
 	}
 
-out:
 	adapter->aq_required |= I40EVF_FLAG_AQ_MAP_VECTORS;
 
 	return err;
Signature · 1 line
2.9.3
#14 lore raw

PATCH 8/10 i40evf: fix ring to vector mapping

Quoted 34 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired-
lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 08/10] i40evf: fix ring to vector mapping

From: Alan Brady <[email protected]>

The current implementation for mapping queues to vectors is broken because it attempts to map each tx and rx ring to its own vector, however we use combined queues so we should actually be mapping the tx/rx rings together on one vector.

Also in the current implementation, in the case where we have more queues than vectors, we attempt to group the queues together into 'chunks' and map each 'chunk' of queues to a vector. Chunking them together would be more ideal if, and only if, we only had RSS because of the way the hashing algorithm works but in the case of a future patch that enables VF ADq, round robin assignment is better and still works with RSS.

This patch resolves both those issues and simplifies the code needed to accomplish this. Instead of treating the case where we have more queues than vectors as special, if we notice our vector index is greater than vectors, reset the vector index to zero and continue mapping. This should ensure that in both cases, whether we have enough vectors for each queue or not, the queues get appropriately mapped.

Testing-hints:
Make sure traffic continues to be received and features like RSS continue to work as expected.

Quoted 5 lines
---
 drivers/net/ethernet/intel/i40evf/i40evf_main.c | 48 ++++++------------------
-
 1 file changed, 10 insertions(+), 38 deletions(-)
#15 lore raw

PATCH 9/10 i40e: Enable VF to negotiate number of allocated queues

In reply to Alice Michael

From: Alan Brady <[email protected]>

Currently the PF allocates a default number of queues for each VF and cannot be changed. This patch enables the VF to request a different number of queues allocated to it. This patch also adds a new virtchnl op and capability flag to facilitate this negotiation.

After the PF receives a request message, it will set a requested number of queues for that VF. Then when the VF resets, its VSI will get a new number of queues allocated to it.

This is a best effort request and since we only allocate a guaranteed default number, if the VF tries to ask for more than the guaranteed number, there may not be enough in HW to accommodate it unless other queues for other VFs are freed. It should also be noted decreasing the number queues allocated to a VF to below the default will NOT enable the allocation of more than 32 VFs per PF and will not free queues guaranteed to each VF by default.

---
 drivers/net/ethernet/intel/i40e/i40e.h             |  1 +
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 75 ++++++++++++++++++++++
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h |  1 +
 include/linux/avf/virtchnl.h                       | 20 ++++++
 4 files changed, 97 insertions(+)
drivers/net/ethernet/intel/i40e/i40e.h +1 -0
index 2f1f65e..890508f 100644
@@ -77,6 +77,7 @@
 #define i40e_default_queues_per_vmdq(pf) \
 		(((pf)->hw_features & I40E_HW_RSS_AQ_CAPABLE) ? 4 : 1)
 #define I40E_DEFAULT_QUEUES_PER_VF	4
+#define I40E_MAX_VF_QUEUES		16
 #define I40E_DEFAULT_QUEUES_PER_TC	1 /* should be a power of 2 */
 #define i40e_pf_get_max_q_per_tc(pf) \
 		(((pf)->hw_features & I40E_HW_128_QP_RSS_CAPABLE) ? 128 : 64)
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +75 -0

111 lines not on this page · raw message

drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h +1 -0
index 1f4b0c5..5111d05 100644
@@ -97,6 +97,7 @@ struct i40e_vf {
 	u16 lan_vsi_id;		/* ID as used by firmware */
 
 	u8 num_queue_pairs;	/* num of qps assigned to VF vsis */
+	u8 num_req_queues;	/* num of requested qps */
 	u64 num_mdd_events;	/* num of mdd events detected */
 	/* num of continuous malformed or invalid msgs detected */
 	u64 num_invalid_msgs;
include/linux/avf/virtchnl.h +20 -0
index c893b95..35375a0 100644
@@ -133,6 +133,7 @@ enum virtchnl_ops {
 	VIRTCHNL_OP_CONFIG_RSS_LUT = 24,
 	VIRTCHNL_OP_GET_RSS_HENA_CAPS = 25,
 	VIRTCHNL_OP_SET_RSS_HENA = 26,
+	VIRTCHNL_OP_REQUEST_QUEUES = 29,
 };
 
 /* This macro is used to generate a compilation error if a structure
@@ -233,6 +234,7 @@ VIRTCHNL_CHECK_STRUCT_LEN(16, virtchnl_vsi_resource);
 #define VIRTCHNL_VF_OFFLOAD_RSS_AQ		0x00000008
 #define VIRTCHNL_VF_OFFLOAD_RSS_REG		0x00000010
 #define VIRTCHNL_VF_OFFLOAD_WB_ON_ITR		0x00000020
+#define VIRTCHNL_VF_OFFLOAD_REQ_QUEUES		0x00000040
 #define VIRTCHNL_VF_OFFLOAD_VLAN		0x00010000
 #define VIRTCHNL_VF_OFFLOAD_RX_POLLING		0x00020000
 #define VIRTCHNL_VF_OFFLOAD_RSS_PCTYPE_V2	0x00040000
@@ -323,6 +325,21 @@ struct virtchnl_vsi_queue_config_info {
 	struct virtchnl_queue_pair_info qpair[1];
 };
 
+/* VIRTCHNL_OP_REQUEST_QUEUES
+ * VF sends this message to request the PF to allocate additional queues to
+ * this VF.  Each VF gets a guaranteed number of queues on init but asking for
+ * additional queues must be negotiated.  This is a best effort request as it
+ * is possible the PF does not have enough queues left to support the request.
+ * If the PF cannot support the number requested it will respond with the
+ * maximum number it is able to support; otherwise it will respond with the
+ * number requested.
+ */
+
+/* VF resource request */
+struct virtchnl_vf_res_request {
+	u16 num_queue_pairs;
+};
+
 VIRTCHNL_CHECK_STRUCT_LEN(72, virtchnl_vsi_queue_config_info);
 
 /* VIRTCHNL_OP_CONFIG_IRQ_MAP
@@ -686,6 +703,9 @@ virtchnl_vc_validate_vf_msg(struct virtchnl_version_info *ver, u32 v_opcode,
 	case VIRTCHNL_OP_SET_RSS_HENA:
 		valid_len = sizeof(struct virtchnl_rss_hena);
 		break;
+	case VIRTCHNL_OP_REQUEST_QUEUES:
+		valid_len = sizeof(struct virtchnl_vf_res_request);
+		break;
 	/* These are always errors coming from the VF. */
 	case VIRTCHNL_OP_EVENT:
 	case VIRTCHNL_OP_UNKNOWN:
Signature · 1 line
2.9.3
#16 lore raw

PATCH 9/10 i40e: Enable VF to negotiate number of allocated queues

Quoted 28 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired- lan at lists.osuosl.org
Subject: [Intel-wired-lan] [next PATCH S77-V3 09/10] i40e: Enable VF to negotiate number of allocated queues

From: Alan Brady <[email protected]>

Currently the PF allocates a default number of queues for each VF and cannot be changed. This patch enables the VF to request a different number of queues allocated to it. This patch also adds a new virtchnl op and capability flag to facilitate this negotiation.

After the PF receives a request message, it will set a requested number of queues for that VF. Then when the VF resets, its VSI will get a new number of queues allocated to it.

This is a best effort request and since we only allocate a guaranteed default number, if the VF tries to ask for more than the guaranteed number, there may not be enough in HW to accommodate it unless other queues for other VFs are freed. It should also be noted decreasing the number queues allocated to a VF to below the default will NOT enable the allocation of more than 32 VFs per PF and will not free queues guaranteed to each VF by default.

Quoted 8 lines
---
 drivers/net/ethernet/intel/i40e/i40e.h             |  1 +
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 75
++++++++++++++++++++++
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h |  1 +
 include/linux/avf/virtchnl.h                       | 20 ++++++
 4 files changed, 97 insertions(+)
#17 lore raw

PATCH 10/10 i40e: refactor FW version checking

In reply to Alice Michael

From: Mitch Williams <[email protected]>

The i40e driver now supports two different devices with two different firmware versions. So be smart about how we handle these. Move the FW version macros to the appropriate header file, and add a convenience macro that checks the version based on the device. Then use this macro to check whether or not the driver can use the new link info API.

Testing Hints (required if no HSD): Compile, load on both X710 and X722 and make sure it doesn't complain about API version on latest FW.

---
 drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h   | 10 +++++++++-
 drivers/net/ethernet/intel/i40e/i40e_common.c       |  6 ++++--
 drivers/net/ethernet/intel/i40e/i40e_main.c         |  2 +-
 drivers/net/ethernet/intel/i40evf/i40e_adminq_cmd.h | 10 +++++++++-
 4 files changed, 23 insertions(+), 5 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h +9 -1
index 5d5f422..e2a9ec8 100644
@@ -34,7 +34,15 @@
  */
 
 #define I40E_FW_API_VERSION_MAJOR	0x0001
-#define I40E_FW_API_VERSION_MINOR	0x0005
+#define I40E_FW_API_VERSION_MINOR_X722	0x0005
+#define I40E_FW_API_VERSION_MINOR_X710	0x0007
+
+#define I40E_FW_MINOR_VERSION(_h) ((_h)->mac.type == I40E_MAC_XL710 ? \
+					I40E_FW_API_VERSION_MINOR_X710 : \
+					I40E_FW_API_VERSION_MINOR_X722)
+
+/* API version 1.7 implements additional link and PHY-specific APIs  */
+#define I40E_MINOR_VER_GET_LINK_INFO_XL710 0x0007
 
 struct i40e_aq_desc {
 	__le16 flags;
drivers/net/ethernet/intel/i40e/i40e_common.c +4 -2
index 111426b..f01a54a 100644
@@ -1593,8 +1593,10 @@ i40e_status i40e_aq_get_phy_capabilities(struct i40e_hw *hw,
 		status = I40E_ERR_UNKNOWN_PHY;
 
 	if (report_init) {
-		hw->phy.phy_types = le32_to_cpu(abilities->phy_type);
-		hw->phy.phy_types |= ((u64)abilities->phy_type_ext << 32);
+		if (hw->mac.type ==  I40E_MAC_XL710 &&
+		    hw->aq.api_maj_ver == I40E_FW_API_VERSION_MAJOR &&
+		    hw->aq.api_min_ver >= I40E_MINOR_VER_GET_LINK_INFO_XL710) {
+			status = i40e_aq_get_link_info(hw, true, NULL, NULL);
 	}
 
 	return status;
drivers/net/ethernet/intel/i40e/i40e_main.c +1 -1
index 1ad014d..fb0c8c6 100644
@@ -11420,7 +11420,7 @@ static int i40e_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		 i40e_nvm_version_str(hw));
 
 	if (hw->aq.api_maj_ver == I40E_FW_API_VERSION_MAJOR &&
-	    hw->aq.api_min_ver > I40E_FW_API_VERSION_MINOR)
+	    hw->aq.api_min_ver > I40E_FW_MINOR_VERSION(hw))
 		dev_info(&pdev->dev,
 			 "The driver for the device detected a newer version of the NVM image than expected. Please install the most recent version of the network driver.\n");
 	else if (hw->aq.api_maj_ver == 1 && hw->aq.api_min_ver < 4)
drivers/net/ethernet/intel/i40evf/i40e_adminq_cmd.h +9 -1
index 83e63e5..f9f48d1 100644
@@ -34,7 +34,15 @@
  */
 
 #define I40E_FW_API_VERSION_MAJOR	0x0001
-#define I40E_FW_API_VERSION_MINOR	0x0005
+#define I40E_FW_API_VERSION_MINOR_X722	0x0005
+#define I40E_FW_API_VERSION_MINOR_X710	0x0007
+
+#define I40E_FW_MINOR_VERSION(_h) ((_h)->mac.type == I40E_MAC_XL710 ? \
+					I40E_FW_API_VERSION_MINOR_X710 : \
+					I40E_FW_API_VERSION_MINOR_X722)
+
+/* API version 1.7 implements additional link and PHY-specific APIs  */
+#define I40E_MINOR_VER_GET_LINK_INFO_XL710 0x0007
 
 struct i40e_aq_desc {
 	__le16 flags;
Signature · 1 line
2.9.3
#18 lore raw

PATCH 10/10 i40e: refactor FW version checking

Quoted 21 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired- lan at lists.osuosl.org
Cc: Williams, Mitch A <[email protected]>
Subject: [Intel-wired-lan] [next PATCH S77-V3 10/10] i40e: refactor FW version checking

From: Mitch Williams <[email protected]>

The i40e driver now supports two different devices with two different firmware versions. So be smart about how we handle these. Move the FW version macros to the appropriate header file, and add a convenience macro that checks the version based on the device. Then use this macro to check whether or not the driver can use the new link info API.

Testing Hints (required if no HSD): Compile, load on both X710 and X722 and make sure it doesn't complain about API version on latest FW.

Quoted 7 lines
---
 drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h   | 10 +++++++++-
 drivers/net/ethernet/intel/i40e/i40e_common.c       |  6 ++++--
 drivers/net/ethernet/intel/i40e/i40e_main.c         |  2 +-
 drivers/net/ethernet/intel/i40evf/i40e_adminq_cmd.h | 10 +++++++++-
 4 files changed, 23 insertions(+), 5 deletions(-)
#19 lore raw

PATCH 1/10 i40e: fix for flow director counters not wrapping as expected

In reply to Alice Michael
Quoted 16 lines

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On Behalf Of Alice Michael
Sent: Friday, July 14, 2017 6:27 AM
To: Michael, Alice <[email protected]>; intel-wired- lan at lists.osuosl.org
Cc: Stachura, Mariusz <[email protected]>
Subject: [Intel-wired-lan] [next PATCH S77-V3 01/10] i40e: fix for flow director counters not wrapping as expected

From: Mariusz Stachura <[email protected]>

An eratta with GLQF_PCNT causes it to not wrap as expected. This can cause an error in flow director statistics. This patch resets affected counters just after reading.

Quoted 7 lines

Testing Hints (required if no HSD): DCR 2187

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 35 +++++++++++++++++++--
--------
 1 file changed, 23 insertions(+), 12 deletions(-)