On Sat, Sep 17, 2005 at 10:25:23PM +0400, Manu Abraham wrote:
[skip]
> static int __devinit mantis_pci_probe(struct pci_dev *pdev,
> const struct pci_device_id *mantis_pci_table)
> {
> u8 revision, latency;
> struct mantis_pci *mantis;
>
> mantis = (struct mantis_pci *)
> kmalloc(sizeof (struct mantis_pci), GFP_KERNEL);
> if (mantis == NULL) {
> dprintk(verbose, MANTIS_ERROR, 1, "Out of memory");
> return -ENOMEM;
> }
> if (pci_enable_device(pdev)) {
> dprintk(verbose, MANTIS_ERROR, 1, "Mantis PCI enable failed");
> goto err;
> }
> dprintk(verbose, MANTIS_ERROR, 1, "<2:>IRQ=%d", pdev->irq);
>
> dprintk(verbose, MANTIS_ERROR, 1, "Got a device");
> if (request_irq(pdev->irq, mantis_pci_irq, SA_SHIRQ | SA_INTERRUPT,
> DRIVER_NAME, mantis) < 0) {
Some code is obviously missing here...
> dprintk(verbose, MANTIS_DEBUG, 1, "We got an IRQ");
> return 0;
>
> err:
> dprintk(verbose, MANTIS_DEBUG, 1, "<freak out>");
> kfree(mantis);
> return -ENODEV;
> }
>
>
>
> static void __devexit mantis_pci_remove(struct pci_dev *pdev)
> {
> free_irq(pdev->irq, pdev);
Here is the next problem - you must give free_irq() the same pointer
that you have passed to request_irq(). So you need a way to get from
a struct pci_dev for your device to the corresponding struct
mantis_pci. This is done as follows:
1) In your mantis_pci_probe(), when you have initialized the device
successfully, put the pointer to you data structure into the PCI
device structure:
pci_set_drvdata(pdev, mantis);
2) In mantis_pci_remove() (and later in other PCI driver functions,
like suspend/resume) get this pointer back:
struct mantis_pci *mantis = pci_get_drvdata(pdev);
Then use that pointer where you need it (e.g., in free_irq()).
3) mantis_pci_remove() should also clear out the pointer to the driver
data:
pci_set_drvdata(pdev, NULL);
> pci_disable_device(pdev);
> }
Attachment:
pgpIilEMs4e4r.pgp
Description: PGP signature
- Follow-Ups:
- Re: free free irq and Oops on cat /proc/interrupts (2)
- From: Manu Abraham <[email protected]>
- Re: free free irq and Oops on cat /proc/interrupts (2)
- References:
- free free irq and Oops on cat /proc/interrupts (2)
- From: Manu Abraham <[email protected]>
- Re: free free irq and Oops on cat /proc/interrupts (2)
- From: Sergey Vlasov <[email protected]>
- Re: free free irq and Oops on cat /proc/interrupts (2)
- From: Manu Abraham <[email protected]>
- free free irq and Oops on cat /proc/interrupts (2)
- Prev by Date: Re: Eradic disk access during reads
- Next by Date: Re: Why don't we separate menuconfig from the kernel?
- Previous by thread: Re: free free irq and Oops on cat /proc/interrupts (2)
- Next by thread: Re: free free irq and Oops on cat /proc/interrupts (2)
- Index(es):
