Message ID | 1464621067-25727-1-git-send-email-mika.kuoppala@intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
What about skl, this also seems to need the WA until A0?
On 03/06/2016 21:31, Matthew Auld wrote: > What about skl, this also seems to need the WA until A0? SKL:A0 is pre-production stepping, it can be ignored. regards Arun > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/intel-gfx >
On 4 June 2016 at 09:54, Arun Siluvery <arun.siluvery@linux.intel.com> wrote: > On 03/06/2016 21:31, Matthew Auld wrote: >> >> What about skl, this also seems to need the WA until A0? > > > SKL:A0 is pre-production stepping, it can be ignored. > > regards > Arun > > >> _______________________________________________ >> Intel-gfx mailing list >> Intel-gfx@lists.freedesktop.org >> https://lists.freedesktop.org/mailman/listinfo/intel-gfx >> > In which case this patch is: Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Arun Siluvery <arun.siluvery@linux.intel.com> writes: > [ text/plain ] > On 03/06/2016 21:31, Matthew Auld wrote: >> What about skl, this also seems to need the WA until A0? > > SKL:A0 is pre-production stepping, it can be ignored. > Yes, the rationale here was that we have kbl:A0 around still but there are no skl:A0 in use that we care. -Mika > regards > Arun > >> _______________________________________________ >> Intel-gfx mailing list >> Intel-gfx@lists.freedesktop.org >> https://lists.freedesktop.org/mailman/listinfo/intel-gfx >>
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 5b0f271b9d91..0612b6c8ffcf 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1261,6 +1261,22 @@ static int gen9_init_indirectctx_bb(struct intel_engine_cs *engine, return ret; index = ret; + /* WaClearSlmSpaceAtContextSwitch:kbl */ + /* Actual scratch location is at 128 bytes offset */ + if (IS_KBL_REVID(engine->i915, 0, KBL_REVID_A0)) { + uint32_t scratch_addr + = engine->scratch.gtt_offset + 2*CACHELINE_BYTES; + + wa_ctx_emit(batch, index, GFX_OP_PIPE_CONTROL(6)); + wa_ctx_emit(batch, index, (PIPE_CONTROL_FLUSH_L3 | + PIPE_CONTROL_GLOBAL_GTT_IVB | + PIPE_CONTROL_CS_STALL | + PIPE_CONTROL_QW_WRITE)); + wa_ctx_emit(batch, index, scratch_addr); + wa_ctx_emit(batch, index, 0); + wa_ctx_emit(batch, index, 0); + wa_ctx_emit(batch, index, 0); + } /* Pad to end of cacheline */ while (index % CACHELINE_DWORDS) wa_ctx_emit(batch, index, MI_NOOP);
This workaround for bdw and chv, is also needed for kbl A0. References: HSD#1911519, BSID#569 Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> --- drivers/gpu/drm/i915/intel_lrc.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)