Double POR timings (tMOD, tMRD, and tZQ) for more margin per lab

Change-Id: Ia1e575d93370372ae6bcb393258bb6f95e324874 Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/41639 Tested-by: Jenkins Server <pfd-jenkins+hostboot@us.ibm.com> Tested-by: Hostboot CI <hostboot-ci+hostboot@us.ibm.com> Reviewed-by: Louis Stermole <stermole@us.ibm.com> Reviewed-by: JACOB L. HARVEY <jlharvey@us.ibm.com> Reviewed-by: Jennifer A. Stofer <stofer@us.ibm.com> Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/41641 Tested-by: FSP CI Jenkins <fsp-CI-jenkins+hostboot@us.ibm.com> Tested-by: Jenkins OP Build CI <op-jenkins+hostboot@us.ibm.com> Reviewed-by: Daniel M. Crowell <dcrowell@us.ibm.com>
open-power · Jun 14, 2017 · 7074cfc · 7074cfc
1 parent 0116e1f
commit 7074cfc
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 9 deletions.
diff --git a/src/import/chips/p9/procedures/hwp/memory/lib/dimm/ddr4/mrs_load_ddr4.C b/src/import/chips/p9/procedures/hwp/memory/lib/dimm/ddr4/mrs_load_ddr4.C
@@ -84,17 +84,21 @@ fapi2::ReturnCode mrs_load( const fapi2::Target<TARGET_TYPE_DIMM>& i_target,
 {
     FAPI_INF("ddr4::mrs_load %s", mss::c_str(i_target));
 
+    // Doubling tMRD delay to increase margin per lab request
+    const size_t DOUBLE_TMRD = 2 * mss::tmrd();
+    const size_t DOUBLE_TMOD = 2 * mss::tmod(i_target);
+
     static const std::vector< mrs_data<TARGET_TYPE_MCBIST> > MRS_DATA =
     {
         // JEDEC ordering of MRS per DDR4 power on sequence
-        {  3, mrs03, mrs03_decode, mss::tmrd()  },
-        {  6, mrs06, mrs06_decode, mss::tmrd()  },
-        {  5, mrs05, mrs05_decode, mss::tmrd()  },
-        {  4, mrs04, mrs04_decode, mss::tmrd()  },
-        {  2, mrs02, mrs02_decode, mss::tmrd()  },
-        {  1, mrs01, mrs01_decode, mss::tmrd()  },
+        {  3, mrs03, mrs03_decode, DOUBLE_TMRD  },
+        {  6, mrs06, mrs06_decode, DOUBLE_TMRD  },
+        {  5, mrs05, mrs05_decode, DOUBLE_TMRD  },
+        {  4, mrs04, mrs04_decode, DOUBLE_TMRD  },
+        {  2, mrs02, mrs02_decode, DOUBLE_TMRD  },
+        {  1, mrs01, mrs01_decode, DOUBLE_TMRD  },
         // We need to wait tmod before zqcl, a non-mrs command
-        {  0, mrs00, mrs00_decode, mss::tmod(i_target) },
+        {  0, mrs00, mrs00_decode, DOUBLE_TMOD },
     };
 
     std::vector< uint64_t > l_ranks;

diff --git a/src/import/chips/p9/procedures/hwp/memory/lib/dimm/ddr4/zqcal.C b/src/import/chips/p9/procedures/hwp/memory/lib/dimm/ddr4/zqcal.C
@@ -69,8 +69,12 @@ fapi2::ReturnCode setup_dram_zqcal( const fapi2::Target<fapi2::TARGET_TYPE_DIMM>
     // Note: this isn't general - assumes Nimbus via MCBIST instruction here BRS
     l_inst = ccs::zqcl_command<TARGET_TYPE_MCBIST>(i_target, i_rank);
 
-    l_inst.arr1.insertFromRight<MCBIST_CCS_INST_ARR1_00_IDLES,
-                                MCBIST_CCS_INST_ARR1_00_IDLES_LEN>(tDLLK + mss::tzqinit());
+    // Doubling tZQ to better margin per lab request
+    {
+        const size_t DELAY = 2 * (tDLLK + mss::tzqinit());
+        l_inst.arr1.insertFromRight<MCBIST_CCS_INST_ARR1_00_IDLES,
+                                    MCBIST_CCS_INST_ARR1_00_IDLES_LEN>(DELAY);
+    }
 
     // There's nothing to decode here.
     FAPI_INF("ZQCL 0x%016llx:0x%016llx %s:rank %d",