[PATCH] (1/23) sk98: no explicit module ref counting needed

[Stephen Hemminger <shemminger@xxxxxxxx>]

Network driver's shouldn't be doing module reference counting
on 2.6.  This is a bad replication of 2.4 behaviour

Signed-off-by: Stephen Hemminger <shemminger@xxxxxxxx>

diff -Nru a/drivers/net/sk98lin/skge.c b/drivers/net/sk98lin/skge.c
--- a/drivers/net/sk98lin/skge.c        2004-11-03 14:07:09 -08:00
+++ b/drivers/net/sk98lin/skge.c        2004-11-03 14:07:09 -08:00
@@ -1166,10 +1166,6 @@
        }
 #endif
 
-       if (!try_module_get(THIS_MODULE)) {
-               return (-1);    /* increase of usage count not possible */
-       }
-
        /* Set blink mode */
        if ((pAC->PciDev->vendor == 0x1186) || (pAC->PciDev->vendor == 0x11ab ))
                pAC->GIni.GILedBlinkCtrl = OEM_CONFIG_VALUE;
@@ -1177,7 +1173,6 @@
        if (pAC->BoardLevel == SK_INIT_DATA) {
                /* level 1 init common modules here */
                if (SkGeInit(pAC, pAC->IoBase, SK_INIT_IO) != 0) {
-                       module_put(THIS_MODULE); /* decrease usage count */
                        printk("%s: HWInit (1) failed.\n", 
pAC->dev[pNet->PortNr]->name);
                        return (-1);
                }
@@ -1193,7 +1188,6 @@
        if (pAC->BoardLevel != SK_INIT_RUN) {
                /* tschilling: Level 2 init modules here, check return value. */
                if (SkGeInit(pAC, pAC->IoBase, SK_INIT_RUN) != 0) {
-                       module_put(THIS_MODULE); /* decrease usage count */
                        printk("%s: HWInit (2) failed.\n", 
pAC->dev[pNet->PortNr]->name);
                        return (-1);
                }
@@ -1285,7 +1279,6 @@
 #ifdef SK_DIAG_SUPPORT
        if (pAC->DiagModeActive == DIAG_ACTIVE) {
                if (pAC->DiagFlowCtrl == SK_FALSE) {
-                       module_put(THIS_MODULE);
                        /* 
                        ** notify that the interface which has been closed
                        ** by operator interaction must not be started up 
@@ -1376,7 +1369,6 @@
        pAC->MaxPorts--;
        pNet->Up = 0;
 
-       module_put(THIS_MODULE);
        return (0);
 } /* SkGeClose */
 
@@ -4681,20 +4673,11 @@
 
        dev = pAC->dev[devNbr];
 
-       /*
-       ** Function SkGeClose() uses MOD_DEC_USE_COUNT (2.2/2.4)
-       ** or module_put() (2.6) to decrease the number of users for
-       ** a device, but if a device is to be put under control of 
-       ** the DIAG, that count is OK already and does not need to 
-       ** be adapted! Hence the opposite MOD_INC_USE_COUNT or 
-       ** try_module_get() needs to be used again to correct that.
+       /* On Linux 2.6 the network driver does NOT mess with reference
+       ** counts.  The driver MUST be able to be unloaded at any time
+       ** due to the possibility of hotplug.
        */
-       if (!try_module_get(THIS_MODULE)) {
-               return (-1);
-       }
-
        if (SkGeClose(dev) != 0) {
-               module_put(THIS_MODULE);
                return (-1);
        }
        return (0);
@@ -4723,17 +4706,6 @@
 
        if (SkGeOpen(dev) != 0) {
                return (-1);
-       } else {
-               /*
-               ** Function SkGeOpen() uses MOD_INC_USE_COUNT (2.2/2.4) 
-               ** or try_module_get() (2.6) to increase the number of 
-               ** users for a device, but if a device was just under 
-               ** control of the DIAG, that count is OK already and 
-               ** does not need to be adapted! Hence the opposite 
-               ** MOD_DEC_USE_COUNT or module_put() needs to be used 
-               ** again to correct that.
-               */
-               module_put(THIS_MODULE);
        }
 
        /*