diff --git a/src/mainboard/prodrive/atlas/romstage_fsp_params.c b/src/mainboard/prodrive/atlas/romstage_fsp_params.c index ab47dd0e27..7d3d9b51f3 100644 --- a/src/mainboard/prodrive/atlas/romstage_fsp_params.c +++ b/src/mainboard/prodrive/atlas/romstage_fsp_params.c @@ -72,30 +72,4 @@ void mainboard_memory_init_params(FSPM_UPD *memupd) mcfg->CpuPcieRpClockReqMsgEnable[0] = 0; mcfg->CpuPcieRpClockReqMsgEnable[1] = 0; mcfg->CpuPcieRpClockReqMsgEnable[2] = 0; - - // shared clock - mcfg->PcieClkSrcUsage[0] = 0x80; - mcfg->PcieClkSrcClkReq[0] = 0xFF; - // i225 - mcfg->PcieClkSrcUsage[1] = 9; // RP 10 - mcfg->PcieClkSrcClkReq[1] = 1; - - // FIX Apparently Rootports don't like the idea of not having a clksrc and clkreq - // attached to it. For example if we set PcieClkSrcClkReq[1] above to 0xFF (unused) - // it will not come back out of L1. You can easily test this on windows by trying to - // update the i225 driver in device manager or use setpci in Linux to set Device in D3. - // The same applies to all other rootports no matter which devices are connected to it. - // Therefore we put each rootport (that does not have a clkreq, clksrc) - // to a not connected (not routed out) clksrc and clkreq. That seems to be a current FSP Bug. - // workaround and will be removed as soon as FSP is fixed. - mcfg->PcieClkSrcUsage[2] = 4; // Rootport 5 - mcfg->PcieClkSrcClkReq[2] = 0; - mcfg->PcieClkSrcUsage[3] = 5; // Rootport 6 - mcfg->PcieClkSrcClkReq[3] = 0; - mcfg->PcieClkSrcUsage[4] = 8; // Rootport 9 - mcfg->PcieClkSrcClkReq[4] = 0; - mcfg->PcieClkSrcUsage[5] = 6; // Rootport 7 - mcfg->PcieClkSrcClkReq[5] = 0; - mcfg->PcieClkSrcUsage[6] = 7; // Rootport 8 - mcfg->PcieClkSrcClkReq[6] = 0; }