Paul Durrant wrote:
> Garrett D'Amore wrote:
>>
>> Hmm... I seem to recall that full MSI-X interrupt support on x86 was
>> either only recently added, or scheduled to be added.
>>
>> What build are you using?
>>
>
> build 83
>
> Actually I answered my own question: I found the function:
>
> /* ARGSUSED */
> uint_t
> i_ddi_get_msix_alloc_limit(dev_info_t *dip)
> {
> uint_t msix_alloc_limit = ddi_msix_alloc_limit;
>
> #if defined(__sparc)
> if (ddi_prop_exists(DDI_DEV_T_ANY, dip, DDI_PROP_NOTPROM |
> DDI_PROP_DONTPASS, "#msix-request")) {
> msix_alloc_limit = MAX(DDI_MAX_MSIX_ALLOC,
> ddi_msix_alloc_limit);
> }
> #endif
>
> return (msix_alloc_limit);
> }
>
> ... and, guess what, ddi_msix_alloc_limit defaults to 2! At least it's
> a global and not a #define. It's a stupid default for such a high
> perf. box.
Probably the limit should be derived from the number of cores in the system.
-- Garrett
>
> Paul
>
> _______________________________________________
> driver-discuss mailing list
> driver-discuss@opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/driver-discuss
>
_______________________________________________
driver-discuss mailing list
driver-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/driver-discuss
