Hello,
I am a beginner in rtnet programming.I am now modifying the normal
Linux ethernet driver into rtnet ethernet driver ,modifications are made
comparing the drivers given in the rtnet-0.9.11 driver folder and following
the steps given in the README.drvporting.Made modification to the driver
mainly based on the drivers rt_pcnet32.c,rt_natsemi.c,rt_8139too.c &
rt_via-rhine.c ect.
But my driver includes some more functions that are not defined in the
drivers above mentioned.How can i modify these function to suit the rtnet
format?
some of the functions are:
*1.netif_rx_schedule_prep(dev),
2.skb_shinfo(skb),
3.rtnl_unlock (),rtnl_lock (),rtnl_lock_interruptible () etc..*
I have some more queries -*poll() fuction is used in rtnet programming?
How can i expand the rtskb pool size?
Is data_len is used in rtnet?*
In my case the problem is with the receiving section..it is not workiing
properly..
can any one help me..
*I will post the changes in form of a patch against rtnet.git*
static void tsi109_rx_int(rtnet_device *dev)
{
tsi109_prv_data *data = dev->priv;// tsi109_prv_data *data =
(struct tsi109_prv_data *)(dev);
// Naishab
rtdm_lockctx_t context;
// A race could cause dev to already be scheduled, so it's not an
// error if that happens (and interrupts shouldn't be re-masked,
// because that can cause harmful races, if poll has already
// unmasked them but not cleared LINK_STATE_SCHED).
//
// This can happen if this code races with tsi109_poll(), which masks
// the interrupts after tsi109_irq_one() read the mask, but before
// netif_rx_schedule is called. It could also happen due to calls
// from tsi109_check_rxring().
// struct napi_struct temp_napi;
// if (netif_rx_schedule_prep(dev, &temp_napi)) {
if (netif_rx_schedule_prep((struct net_device *) dev)) {
// Mask, rather than ack, the receive interrupts. The ack
// will happen in tsi109_poll().
//Naishab spin_lock_irq(&data->intmasklock);
rtdm_lock_get_irqsave(&data->intmasklock, context);
data->regs[TSI109_EC_INTMASK] |= TSI109_INT_RXQUEUE0 |
TSI109_INT_RXTHRESH |
TSI109_INT_RXOVERRUN |
TSI109_INT_RXERROR |
TSI109_INT_RXWAIT;
mb();
//Naishab spin_unlock_irq(&data->intmasklock);
rtdm_lock_put_irqrestore(&data->intmasklock, context);
__netif_rx_schedule((struct net_device *) dev);
// __netif_rx_schedule(dev, &temp_napi);
} else {
if (!rtnetif_running(dev)) {
// This can happen if an interrupt occurs while the
// interface is being brought down, as the START
// bit is cleared before the stop function is called.
//
// In this case, the interrupts must be masked, or
// they will continue indefinitely.
//
// There's a race here if the interface is brought down
// and then up in rapid succession, as the device could
// be made running after the above check and before
// the masking below. This will only happen if the IRQ
// thread has a lower priority than the task brining
// up the interface. Fixing this race would likely
// require changes in generic code.
//Naishab spin_lock_irq(&data->intmasklock);
rtdm_lock_get_irqsave(&data->intmasklock, context);
data->regs[TSI109_EC_INTMASK] |= TSI109_INT_RXQUEUE0 |
TSI109_INT_RXTHRESH |
TSI109_INT_RXOVERRUN |
TSI109_INT_RXERROR |
TSI109_INT_RXWAIT;
mb();
//Naishab spin_unlock_irq(&data->intmasklock);
rtdm_lock_put_irqrestore(&data->intmasklock, context);
}
}
}
can anyone help me????????? thanks in
advance..please reply to my mail id...
anju amry joseph
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
RTnet-users mailing list
RTnet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rtnet-users
