On 06/30/2011 08:42 AM, Rafał Miłecki wrote:
> 2011/6/30 Hauke Mehrtens <hauke@xxxxxxxxxx>:
- err = bcma_get_next_core(bus, &eromptr, core);
- if (err == -ENXIO)
+ err = bcma_get_next_core(bus, &eromptr, NULL, core_num, core);
+ if (err == -ENODEV) {
+ core_num++;
+ continue;
+ } else if (err == -ENXIO)
continue;
else if (err == -ESPIPE)
break;
else if (err < 0)
return err;
+ core->core_index = core_num++;
+ bus->nr_cores++;
+
pr_info("Core %d found: %s "
"(manuf 0x%03X, id 0x%03X, rev 0x%02X, class 0x%X)\n",
- bus->nr_cores, bcma_device_name(&core->id),
+ core->core_index, bcma_device_name(&core->id),
core->id.manuf, core->id.id, core->id.rev,
core->id.class);
- core->core_index = bus->nr_cores++;
>
> Didn't you just change core indexes (0, 1, ...) to numbers (1, 2,
> ...)? It would break enabling IRQs on PCI.
No, the cores are getting the same indexes numbers as before. While
scanning core_num is increased for every core found also for cores we
are not searching for or we already found. Then core_num will be
assigned to core->core_index. bus->nr_cores is no used any more so it
could be removed or is it needed for something else?
As you can see here [0] it finds core 0 and 3 at first and then the others.
>
> Didn't test it however yet, I'll have access to my machines tomorrow.
>
Hauke
[0] http://permalink.gmane.org/gmane.linux.kernel.wireless.general/71851
