Re: [PATCH 00/13] Gen7 batch buffer command parser

[Daniel Vetter <daniel@ffwll.ch>]

On Wed, Jan 29, 2014 at 01:55:01PM -0800, bradley.d.volkin@intel.com wrote:
> From: Brad Volkin <bradley.d.volkin@intel.com>
> 3) Coherency. I've found a coherency issue on VLV when reading the batch buffer
>    from the CPU during execbuffer2. Userspace writes the batch via pwrite fast
>    path before calling execbuffer2. The parser reads stale data. This works fine
>    on IVB and HSW, so I believe it's an LLC vs. non-LLC issue. I'm just unclear
>    on what the correct flushing or synchronization is for this scenario. This
>    only matters if we get PPGTT working on VLV and enable the parser there.

Hm, adopting the shmem_read clflushing didn't help for this? That would be
fairly shocking, since it means our shmem read paths are broken. Which are
used e.g. by the libva readback code for the encoded bitstream.

One thing aside: When resending the complete series (even if it's just a
subset) it's better to start a new thread. We tend to use in-reply-to only
when resending individual patches, while the review discussion is still
ongoing. That way the discussion stays together. But when there's been a
bit a longer break it's imo better to start a new thread.

Cheers, Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch