Thread (18 messages) 18 messages, 3 authors, 2014-05-28

[net-next 05/15] ixgbe: extract PTP clock device from ptp_init

flat view
STALE4528d SUPERSEDED: 1 (0M)

From: Jeff Kirsher <hidden>
Date: 2014-05-27 09:12:39
Subsystem: intel ethernet drivers, networking drivers, the rest · Maintainers: Tony Nguyen, Przemek Kitszel, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds

Revision v1 of 192 in this series; 1 review trailer.

Revisions (192)
  1. v1
  2. v2 [diff vs current]
  3. v1
  4. v1
  5. v2 [diff vs current]
  6. v1
  7. v1
  8. v1
  9. v1
  10. v1
  11. v1
  12. v1
  13. v1
  14. v2 [diff vs current]
  15. v1
  16. v3 [diff vs current]
  17. v1
  18. v1
  19. v1
  20. v2 [diff vs current]
  21. v2 [diff vs current]
  22. v1
  23. v2 [diff vs current]
  24. v1
  25. v2 [diff vs current]
  26. v3 [diff vs current]
  27. v1
  28. v1
  29. v1
  30. v2 [diff vs current]
  31. v1
  32. v1
  33. v1
  34. v2 [diff vs current]
  35. v1
  36. v1
  37. v1
  38. v1
  39. v1
  40. v2 [diff vs current]
  41. v1
  42. v1
  43. v1
  44. v1
  45. v1
  46. v1
  47. v1
  48. v1
  49. v1
  50. v1
  51. v1
  52. v1
  53. v1
  54. v1
  55. v2 [diff vs current]
  56. v1
  57. v1
  58. v1
  59. v1
  60. v1
  61. v1
  62. v1
  63. v1
  64. v1
  65. v1
  66. v1
  67. v1
  68. v1
  69. v1
  70. v1
  71. v1
  72. v1
  73. v1
  74. v1
  75. v1
  76. v1
  77. v1
  78. v1
  79. v1
  80. v1
  81. v1
  82. v1
  83. v2 [diff vs current]
  84. v1
  85. v1
  86. v1
  87. v1
  88. v1
  89. v2 [diff vs current]
  90. v1
  91. v1
  92. v1
  93. v1
  94. v2 [diff vs current]
  95. v1
  96. v2 [diff vs current]
  97. v1
  98. v1
  99. v1
  100. v1
  101. v1
  102. v1
  103. v1
  104. v1
  105. v1
  106. v2 [diff vs current]
  107. v1
  108. v1
  109. v2 [diff vs current]
  110. v1
  111. v1
  112. v1
  113. v1
  114. v1
  115. v1
  116. v2 [diff vs current]
  117. v1
  118. v2 [diff vs current]
  119. v2 [diff vs current]
  120. v1
  121. v2 [diff vs current]
  122. v3 [diff vs current]
  123. v1
  124. v4 [diff vs current]
  125. v1
  126. v5 [diff vs current]
  127. v6 [diff vs current]
  128. v7 [diff vs current]
  129. v1
  130. v1
  131. v2 [diff vs current]
  132. v1
  133. v1
  134. v1
  135. v2 [diff vs current]
  136. v1
  137. v1
  138. v1
  139. v1
  140. v2 [diff vs current]
  141. v1
  142. v1
  143. v1
  144. v1
  145. v2 [diff vs current]
  146. v1
  147. v1
  148. v1
  149. v2 [diff vs current]
  150. v3 [diff vs current]
  151. v4 [diff vs current]
  152. v1
  153. v1
  154. v2 [diff vs current]
  155. v5 [diff vs current]
  156. v1
  157. v1
  158. v1
  159. v2 [diff vs current]
  160. v1
  161. v1
  162. v2 [diff vs current]
  163. v3 [diff vs current]
  164. v1
  165. v2 [diff vs current]
  166. v1
  167. v1
  168. v2 [diff vs current]
  169. v1
  170. v2 [diff vs current]
  171. v1
  172. v2 [diff vs current]
  173. v3 [diff vs current]
  174. v1
  175. v1
  176. v1
  177. v2 [diff vs current]
  178. v1
  179. v1
  180. v1
  181. v1
  182. v1
  183. v1
  184. v2 [diff vs current]
  185. v1
  186. v1
  187. v1
  188. v2 [diff vs current]
  189. v1 current
  190. v1
  191. v1
  192. v2 [diff vs current]
From: Jacob Keller <jacob.e.keller@intel.com>

In order to properly handle a suspend/resume cycle, we cannot destroy
the PTP clock device. As part of this, we should only re-create the
device on first initialization. After a resume, when ixgbe_ptp_init is
called, we won't create a new clock, and we will use the old clock
device. To that end, this patch extracts the clock creation out of
ptp_init, and only calls it if we don't already have a ptp_clock
pointer.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <redacted>
Signed-off-by: Jeff Kirsher <redacted>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c | 56 +++++++++++++++++++++++-----
 1 file changed, 47 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
index 9d329f5..9df7e57 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
@@ -851,16 +851,23 @@ void ixgbe_ptp_reset(struct ixgbe_adapter *adapter)
 }
 
 /**
- * ixgbe_ptp_init
+ * ixgbe_ptp_create_clock
  * @adapter: the ixgbe private adapter structure
  *
- * This function performs the required steps for enabling ptp
- * support. If ptp support has already been loaded it simply calls the
- * cyclecounter init routine and exits.
+ * This function performs setup of the user entry point function table and
+ * initializes the PTP clock device, which is used to access the clock-like
+ * features of the PTP core. It will be called by ixgbe_ptp_init, only if
+ * there isn't already a clock device (such as after a suspend/resume cycle,
+ * where the clock device wasn't destroyed).
  */
-void ixgbe_ptp_init(struct ixgbe_adapter *adapter)
+static int ixgbe_ptp_create_clock(struct ixgbe_adapter *adapter)
 {
 	struct net_device *netdev = adapter->netdev;
+	long err;
+
+	/* do nothing if we already have a clock device */
+	if (!IS_ERR_OR_NULL(adapter->ptp_clock))
+		return 0;
 
 	switch (adapter->hw.mac.type) {
 	case ixgbe_mac_X540:
@@ -897,22 +904,53 @@ void ixgbe_ptp_init(struct ixgbe_adapter *adapter)
 		break;
 	default:
 		adapter->ptp_clock = NULL;
-		return;
+		return -EOPNOTSUPP;
 	}
 
-	spin_lock_init(&adapter->tmreg_lock);
-	INIT_WORK(&adapter->ptp_tx_work, ixgbe_ptp_tx_hwtstamp_work);
-
 	adapter->ptp_clock = ptp_clock_register(&adapter->ptp_caps,
 						&adapter->pdev->dev);
 	if (IS_ERR(adapter->ptp_clock)) {
+		err = PTR_ERR(adapter->ptp_clock);
 		adapter->ptp_clock = NULL;
 		e_dev_err("ptp_clock_register failed\n");
+		return err;
 	} else
 		e_dev_info("registered PHC device on %s\n", netdev->name);
 
+	/* set default timestamp mode to disabled here. We do this in
+	 * create_clock instead of init, because we don't want to override the
+	 * previous settings during a resume cycle.
+	 */
 	adapter->tstamp_config.rx_filter = HWTSTAMP_FILTER_NONE;
 	adapter->tstamp_config.tx_type = HWTSTAMP_TX_OFF;
+
+	return 0;
+}
+
+/**
+ * ixgbe_ptp_init
+ * @adapter: the ixgbe private adapter structure
+ *
+ * This function performs the required steps for enabling PTP
+ * support. If PTP support has already been loaded it simply calls the
+ * cyclecounter init routine and exits.
+ */
+void ixgbe_ptp_init(struct ixgbe_adapter *adapter)
+{
+	/* initialize the spin lock first since we can't control when a user
+	 * will call the entry functions once we have initialized the clock
+	 * device
+	 */
+	spin_lock_init(&adapter->tmreg_lock);
+
+	/* obtain a PTP device, or re-use an existing device */
+	if (ixgbe_ptp_create_clock(adapter))
+		return;
+
+	/* we have a clock so we can initialize work now */
+	INIT_WORK(&adapter->ptp_tx_work, ixgbe_ptp_tx_hwtstamp_work);
+
+	/* reset the PTP related hardware bits */
 	ixgbe_ptp_reset(adapter);
 
 	/* enter the IXGBE_PTP_RUNNING state */
-- 
1.9.3
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help