Re: [PATCH 1/9] rtlwifi: avoid race registering with mac80211

[Larry Finger <Larry.Finger@lwfinger.net>]

On 05/03/2012 12:21 PM, Joshua Roys wrote:
>
> Do you think I should have moved the
> set_bit(RTL_STATUS_INTERFACE_START, &rtlpriv->status);
> lines also? Or is the idea of the patch just wrong?

I am still thinking about the problem, but a few things are clear. There can be 
no ieee80211 callbacks until the firmware has been loaded from the file. On 
bootup when user-space code is not running, there will be a delay, which is the 
reason we switched to asynchronous fw loading. Thus, the call to 
ieee80211_register_hw() must not be done until the firmware is available.

If your analysis that the calling of rtl_init_core() is out of order, then the 
following should fix it:


Index: wireless-testing-new/drivers/net/wireless/rtlwifi/pci.c
===================================================================
--- wireless-testing-new.orig/drivers/net/wireless/rtlwifi/pci.c
+++ wireless-testing-new/drivers/net/wireless/rtlwifi/pci.c
@@ -1853,14 +1853,6 @@ int __devinit rtl_pci_probe(struct pci_d
   	/*like read eeprom and so on */
   	rtlpriv->cfg->ops->read_eeprom_info(hw);

-	if (rtlpriv->cfg->ops->init_sw_vars(hw)) {
-		RT_TRACE(rtlpriv, COMP_ERR, DBG_EMERG, "Can't init_sw_vars\n");
-		err = -ENODEV;
-		goto fail3;
-	}
-
-	rtlpriv->cfg->ops->init_sw_leds(hw);
-
   	/*aspm */
   	rtl_pci_init_aspm(hw);

@@ -1879,6 +1871,14 @@ int __devinit rtl_pci_probe(struct pci_d
   		goto fail3;
   	}

+	if (rtlpriv->cfg->ops->init_sw_vars(hw)) {
+		RT_TRACE(rtlpriv, COMP_ERR, DBG_EMERG, "Can't init_sw_vars\n");
+		err = -ENODEV;
+		goto fail3;
+	}
+
+	rtlpriv->cfg->ops->init_sw_leds(hw);
+
   	err = sysfs_create_group(&pdev->dev.kobj, &rtl_attribute_group);
   	if (err) {
   		RT_TRACE(rtlpriv, COMP_ERR, DBG_EMERG,

This one works on my system, and is being tested against the Bugzilla entry I 
noted earlier.

Larry