On 31/01/2019 09:15, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2019-01-31 08:47:37)
+ /*
+ * Some future proofing on the types since the uAPI is wider than the
+ * current internal implementation.
+ */
+ if (WARN_ON((fls(user->slice_mask) >
+ sizeof(context->slice_mask) * BITS_PER_BYTE) ||
+ (fls(user->subslice_mask) >
+ sizeof(context->subslice_mask) * BITS_PER_BYTE) ||
+ overflows_type(user->min_eus_per_subslice,
+ context->min_eus_per_subslice) ||
+ overflows_type(user->max_eus_per_subslice,
+ context->max_eus_per_subslice)))
fls(0) is not well-defined, so
if (order_base_2(user->slice_mask) > BITS_PER_TYPE(context->slice_mask))
return -EINVAL;
etc
Don't WARN_ON() for user controlled variables.
Looks I was thinking it is not user controllable eg. was supposed to be
the reminder of a too narrow internal type. But that is obviously not
true.. Also not sure why I did not use overflows_type for all.
Thanks,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
