• Michel Thierry's avatar
    drm/i915: Finish gen6/7 dynamic page table allocation · 4933d519
    Michel Thierry authored
    This patch continues on the idea from "Track GEN6 page table usage".
    From here on, in the steady state, PDEs are all pointing to the scratch
    page table (as recommended in the spec). When an object is allocated in
    the VA range, the code will determine if we need to allocate a page for
    the page table. Similarly when the object is destroyed, we will remove,
    and free the page table pointing the PDE back to the scratch page.
    
    Following patches will work to unify the code a bit as we bring in GEN8
    support. GEN6 and GEN8 are different enough that I had a hard time to
    get to this point with as much common code as I do.
    
    The aliasing PPGTT must pre-allocate all of the page tables. There are a
    few reasons for this. Two trivial ones: aliasing ppgtt goes through the
    ggtt paths, so it's hard to maintain, we currently do not restore the
    default context (assuming the previous force reload is indeed
    necessary). Most importantly though, the only way (it seems from
    empirical evidence) to invalidate the CS TLBs on non-render ring is to
    either use ring sync (which requires actually stopping the rings in
    order to synchronize when the sync completes vs. where you are in
    execution), or to reload DCLV.  Since without full PPGTT we do not ever
    reload the DCLV register, there is no good way to achieve this. The
    simplest solution is just to not support dynamic page table
    creation/destruction in the aliasing PPGTT.
    
    We could always reload DCLV, but this seems like quite a bit of excess
    overhead only to save at most 2MB-4k of memory for the aliasing PPGTT
    page tables.
    
    v2: Make the page table bitmap declared inside the function (Chris)
    Simplify the way scratching address space works.
    Move the alloc/teardown tracepoints up a level in the call stack so that
    both all implementations get the trace.
    
    v3: Updated trace event to spit out a name
    
    v4: Aliasing ppgtt is now initialized differently (in setup global gtt)
    
    v5: Rebase to latest code. Also removed unnecessary aliasing ppgtt check
    for trace, as it is no longer possible after the PPGTT cleanup patch series
    of a couple of months ago (Daniel).
    
    v6: Implement changes from code review (Daniel):
     - allocate/teardown_va_range calls added.
     - Add a scratch page allocation helper (only need the address).
     - Move trace events to a new patch.
     - Use updated mark_tlbs_dirty.
     - Moved pt preallocation for aliasing ppgtt into gen6_ppgtt_init.
    
    v7: teardown_va_range removed (Daniel).
        In init, gen6_ppgtt_clear_range call is only needed for aliasing ppgtt.
    
    v8: Rebase after s/page_tables/page_table/.
    
    v9: Remove unnecessary scratch flag in page_table struct, future patches
    can just compare against ppgtt->scratch_pt, and alloc_pt_scratch becomes
    redundant. Initialize scratch_pt and pt. (Mika)
    
    v10: Clean up aliasing ppgtt init error path and prevent leaking the
    ppgtt obj when init fails. (Mika)
    Updated commit author. (Daniel)
    
    Cc: Mika Kuoppala <mika.kuoppala@intel.com>
    Signed-off-by: default avatarBen Widawsky <ben@bwidawsk.net>
    Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v4+)
    Reviewed-by: default avatarMika Kuoppala <mika.kuoppala@intel.com>
    Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
    4933d519
i915_gem_gtt.h 13.7 KB