Re: 2.4.23 IDE hang on boot with two single-channel controllers

[Jeremy Jackson]

Hi,

(watch crossposting when replying all)

I'm hoping to reach the maintainer of the Linux IDE driver for the 
Compaq TriFlex controller.  I have a problem with this driver when used 
with a Compaq Armada 7730MT while docked in the base station.

The driver appears to only support one triflex controller, due to a 
missing check (that other chipset drivers have) that should prevent it 
from registering a /proc interface more than once.  The result is that 
it hangs on boot in proc_ide_create() in an infinite loop.

triflex.c:

static unsigned int __init init_chipset_triflex(struct pci_dev *dev,
                const char *name)
{
#ifdef CONFIG_PROC_FS
        ide_pci_register_host_proc(&triflex_proc);
#endif
        return 0;
}

It also appears that triflex_get_info() doesn't support more that one 
controller.

I won't go into more detail until I can establish who might care :)

Regards,

Jeremy Jackson
> --- Begin Message ---
> Hi All,
>
> Already posted this to linux-kernel.
>
> kdb shows proc_ide_create() stuck in a loop when booting on a Compaq 
> Armada 7730MT while attached to the docking station.
>
> This is a unique IDE hardware setup.  Channel ide0's controller is in 
> the laptop, while ide1 is a separate controller (pci device) in the 
> docking station and is not always present.
>
> This seems to be triggering a bug in ide-proc.c:
>
> void proc_ide_create(void)
> {
> #ifdef CONFIG_BLK_DEV_IDEPCI
>         ide_pci_host_proc_t *p = ide_pci_host_proc_list;
> #endif /* CONFIG_BLK_DEV_IDEPCI */
>
>         proc_ide_root = proc_mkdir("ide", 0);
>         if (!proc_ide_root) return;
>
>         create_proc_ide_interfaces();
>
>         create_proc_read_entry("drivers", 0, proc_ide_root,
>                                 proc_ide_read_drivers, NULL);
>
> #ifdef CONFIG_BLK_DEV_IDEPCI
>         while (p != NULL)  <------------------- INFINITE LOOP HERE
>         {
>                 if (p->name != NULL && p->set == 1 && p->get_info != NULL)
>                 {
>                         p->parent = proc_ide_root;
>                         create_proc_info_entry(p->name, 0, p->parent, 
> p->get_info);
>                         p->set = 2;
>                 }
>                 p = p->next;
>         }
> #endif /* CONFIG_BLK_DEV_IDEPCI */
> }
>
> I'm not sure if the problem is in the loop or bad data being setup 
> before it starts.
>
> Assistance fixing it would be appreciated.
>
> Regards,
>
> Jeremy Jackson
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-ide" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
> --- End Message ---