Chris Wilson <ch...@chris-wilson.co.uk> writes: > On Mon, Sep 07, 2015 at 05:21:53PM +0300, Francisco Jerez wrote: >> I'm sure you had some specific practical advantage in mind? Bashing the >> rather sensitive L3 configuration registers to see if something sticks >> is a hack with potential implications. I'd prefer to avoid that unless >> there is a reason why it could be useful. > > It prevents an incorrect assumption about kernel ABI. The cost of which > is seeing whether you know how to program registers correctly from the > ring. > -Chris
All other codepaths that need to use LRI to write registers share the same can_do_pipelined_register_writes() test and make assumptions about the kernel ABI from there. Ideally there would be a kernel param userspace could use to query whether the command parser is active, but there is not, so it doesn't look like we can do better than pick some harmless register (the L3 config registers are not), write it from a batch, and infer whether the command parser is active or not from the result. If you have any evidence that the assumption about the kernel ABI made here is wrong, please share it, we should probably just blacklist the specific command parser revisions for which the assumption doesn't hold. > > -- > Chris Wilson, Intel Open Source Technology Centre
signature.asc
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
