Re: IDE 2.5.69 possible bogosity...

[Bartlomiej Zolnierkiewicz]

On Thu, 22 May 2003, Ian Molton wrote:

> Hi.
>
> Im wondering if this is correct. is the test for initializing in the
> second for loop correct?

Unfortunately, yes.

> Im building an IDE driver into my kernel that calls ide_register_hw()
> twice to register its primary and secondary ports, but only the
> secondary port is recognised. the first fails, since the test in the
> first for loop fails and so does the second, so it then 'unregisters'

Too little information, your MAX_HWIFS and default io ports?

> it, despite never having been registered. somehow, this puts my drive
> INTO the hwif array, so the secondary interface registers OK, passing
> the other tests.

Where is your driver?

> a hack that allowed the primary interface to register was to register it
> twice, but that sucks.
>
> int ide_register_hw (hw_regs_t *hw, ide_hwif_t **hwifp)
> {
>         int index, retry = 1;
>         ide_hwif_t *hwif;
>
>         do {
>                 for (index = 0; index < MAX_HWIFS; ++index) {
>                         hwif = &ide_hwifs[index];
>                         if (hwif->hw.io_ports[IDE_DATA_OFFSET] ==
> hw->io_ports[IDE_DATA_OFFSET])
>                                 goto found;
>                 }
>                 for (index = 0; index < MAX_HWIFS; ++index) {
>                         hwif = &ide_hwifs[index];
>
> *** is the test for initialising (not the !initialising one) here ok?
> ***
>
>                     if ((!hwif->present && !hwif->mate && !initializing)
> ||
>                         (!hwif->hw.io_ports[IDE_DATA_OFFSET] &&
> initializing))
>                                 goto found;
>                 }
>                 for (index = 0; index < MAX_HWIFS; index++)
>                         ide_unregister(index);
>         } while (retry--);