From af786b618d5e1aafe1e5b8d9ed49b55903381a2e Mon Sep 17 00:00:00 2001 From: Scott Duplichan Date: Tue, 19 Oct 2010 04:26:17 +0000 Subject: When debug logging is enabled, a message such as '* AP 02 timed out:02010501' is sometimes logged. The reason is that the AP first sets a completion value such as 0x13, which is what function wait_cpu_state() is waiting for. Then a short time later, the AP calls function init_fidvid_ap(). This function sets a completion value of 01. When logging is off, wait_cpu_state is fast enough to see the initial completion value for each of the APs. But with logging enabled, one or more APs may go on to complete function init_fidvid_ap, which sets the completion value to 01. While mostly harmless, the timeout does increase boot time. This patch eliminates the timeout by making function wait_cpu_state recognize 01 as an additional valid AP completion value. Signed-off-by: Scott Duplichan Acked-by: Peter Stuge git-svn-id: svn://svn.coreboot.org/coreboot/trunk@5966 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1 --- src/cpu/amd/model_10xxx/fidvid.c | 1676 +++++++++++++++---------------- src/cpu/amd/model_10xxx/init_cpus.c | 1893 +++++++++++++++++------------------ 2 files changed, 1783 insertions(+), 1786 deletions(-) (limited to 'src/cpu') diff --git a/src/cpu/amd/model_10xxx/fidvid.c b/src/cpu/amd/model_10xxx/fidvid.c index 7007d4cf8f..6104f9c3f6 100644 --- a/src/cpu/amd/model_10xxx/fidvid.c +++ b/src/cpu/amd/model_10xxx/fidvid.c @@ -1,838 +1,838 @@ -/* - * This file is part of the coreboot project. - * - * Copyright (C) 2007 Advanced Micro Devices, Inc. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; version 2 of the License. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#if SET_FIDVID == 1 -#include - -#define SET_FIDVID_DEBUG 1 - -// if we are tight of CAR stack, disable it -#define SET_FIDVID_STORE_AP_APICID_AT_FIRST 1 - -static inline void print_debug_fv(const char *str, u32 val) -{ -#if SET_FIDVID_DEBUG == 1 - printk(BIOS_DEBUG, "%s%x\n", str, val); -#endif -} - -static inline void print_debug_fv_8(const char *str, u8 val) -{ -#if SET_FIDVID_DEBUG == 1 - printk(BIOS_DEBUG, "%s%02x\n", str, val); -#endif -} - -static inline void print_debug_fv_64(const char *str, u32 val, u32 val2) -{ -#if SET_FIDVID_DEBUG == 1 - printk(BIOS_DEBUG, "%s%x%x\n", str, val, val2); -#endif -} - -struct fidvid_st { - u32 common_fid; -}; - -static void enable_fid_change(u8 fid) -{ - u32 dword; - u32 nodes; - device_t dev; - int i; - - nodes = get_nodes(); - - for (i = 0; i < nodes; i++) { - dev = NODE_PCI(i, 3); - dword = pci_read_config32(dev, 0xd4); - dword &= ~0x1F; - dword |= (u32) fid & 0x1F; - dword |= 1 << 5; // enable - pci_write_config32(dev, 0xd4, dword); - printk(BIOS_DEBUG, "FID Change Node:%02x, F3xD4: %08x \n", i, - dword); - } -} - -static void recalculateVsSlamTimeSettingOnCorePre(device_t dev) -{ - u8 pviModeFlag; - u8 highVoltageVid, lowVoltageVid, bValue; - u16 minimumSlamTime; - u16 vSlamTimes[7] = { 1000, 2000, 3000, 4000, 6000, 10000, 20000 }; /* Reg settings scaled by 100 */ - u32 dtemp; - msr_t msr; - - /* This function calculates the VsSlamTime using the range of possible - * voltages instead of a hardcoded 200us. - * Note:This function is called from setFidVidRegs and setUserPs after - * programming a custom Pstate. - */ - - /* Calculate Slam Time - * Vslam = 0.4us/mV * Vp0 - (lowest out of Vpmin or Valt) - * In our case, we will scale the values by 100 to avoid - * decimals. - */ - - /* Determine if this is a PVI or SVI system */ - dtemp = pci_read_config32(dev, 0xA0); - - if (dtemp & PVI_MODE) - pviModeFlag = 1; - else - pviModeFlag = 0; - - /* Get P0's voltage */ - msr = rdmsr(0xC0010064); - highVoltageVid = (u8) ((msr.lo >> PS_CPU_VID_SHFT) & 0x7F); - - /* If SVI, we only care about CPU VID. - * If PVI, determine the higher voltage b/t NB and CPU - */ - if (pviModeFlag) { - bValue = (u8) ((msr.lo >> PS_NB_VID_SHFT) & 0x7F); - if (highVoltageVid > bValue) - highVoltageVid = bValue; - } - - /* Get Pmin's index */ - msr = rdmsr(0xC0010061); - bValue = (u8) ((msr.lo >> PS_CUR_LIM_SHFT) & BIT_MASK_3); - - /* Get Pmin's VID */ - msr = rdmsr(0xC0010064 + bValue); - lowVoltageVid = (u8) ((msr.lo >> PS_CPU_VID_SHFT) & 0x7F); - - /* If SVI, we only care about CPU VID. - * If PVI, determine the higher voltage b/t NB and CPU - */ - if (pviModeFlag) { - bValue = (u8) ((msr.lo >> PS_NB_VID_SHFT) & 0x7F); - if (lowVoltageVid > bValue) - lowVoltageVid = bValue; - } - - /* Get AltVID */ - dtemp = pci_read_config32(dev, 0xDC); - bValue = (u8) (dtemp & BIT_MASK_7); - - /* Use the VID with the lowest voltage (higher VID) */ - if (lowVoltageVid < bValue) - lowVoltageVid = bValue; - - /* If Vids are 7Dh - 7Fh, force 7Ch to keep calculations linear */ - if (lowVoltageVid > 0x7C) { - lowVoltageVid = 0x7C; - if (highVoltageVid > 0x7C) - highVoltageVid = 0x7C; - } - - bValue = (u8) (lowVoltageVid - highVoltageVid); - - /* Each Vid increment is 12.5 mV. The minimum slam time is: - * vidCodeDelta * 12.5mV * 0.4us/mV - * Scale by 100 to avoid decimals. - */ - minimumSlamTime = bValue * (125 * 4); - - /* Now round up to nearest register setting. - * Note that if we don't find a value, we - * will fall through to a value of 7 - */ - for (bValue = 0; bValue < 7; bValue++) { - if (minimumSlamTime <= vSlamTimes[bValue]) - break; - } - - /* Apply the value */ - dtemp = pci_read_config32(dev, 0xD8); - dtemp &= VSSLAM_MASK; - dtemp |= bValue; - pci_write_config32(dev, 0xd8, dtemp); -} - -static void prep_fid_change(void) -{ - u32 dword, dtemp; - u32 nodes; - device_t dev; - int i; - - /* This needs to be run before any Pstate changes are requested */ - - nodes = get_nodes(); - - for (i = 0; i < nodes; i++) { - printk(BIOS_DEBUG, "Prep FID/VID Node:%02x \n", i); - dev = NODE_PCI(i, 3); - - dword = pci_read_config32(dev, 0xd8); - dword &= VSRAMP_MASK; - dword |= VSRAMP_VALUE; - pci_write_config32(dev, 0xd8, dword); - - /* Figure out the value for VsSlamTime and program it */ - recalculateVsSlamTimeSettingOnCorePre(dev); - - /* Program fields in Clock Power/Control register0 (F3xD4) */ - /* set F3xD4 Clock Power/Timing Control 0 Register - * NbClkDidApplyAll=1b - * NbClkDid=100b - * PowerStepUp= "platform dependent" - * PowerStepDown= "platform dependent" - * LinkPllLink=01b - * ClkRampHystSel=HW default - */ - /* check platform type */ - if (!(get_platform_type() & AMD_PTYPE_SVR)) { - /* For non-server platform - * PowerStepUp=01000b - 50nS - * PowerStepDown=01000b - 50ns - */ - dword = pci_read_config32(dev, 0xd4); - dword &= CPTC0_MASK; - dword |= NB_CLKDID_ALL | NB_CLKDID | PW_STP_UP50 | PW_STP_DN50 | LNK_PLL_LOCK; /* per BKDG */ - pci_write_config32(dev, 0xd4, dword); - } else { - dword = pci_read_config32(dev, 0xd4); - dword &= CPTC0_MASK; - /* get number of cores for PowerStepUp & PowerStepDown in server - 1 core - 400nS - 0000b - 2 cores - 200nS - 0010b - 3 cores - 133nS -> 100nS - 0011b - 4 cores - 100nS - 0011b - */ - switch (get_core_num_in_bsp(i)) { - case 0: - dword |= PW_STP_UP400 | PW_STP_DN400; - break; - case 1: - case 2: - dword |= PW_STP_UP200 | PW_STP_DN200; - break; - case 3: - dword |= PW_STP_UP100 | PW_STP_DN100; - break; - default: - dword |= PW_STP_UP100 | PW_STP_DN100; - break; - } - dword |= NB_CLKDID_ALL | NB_CLKDID | LNK_PLL_LOCK; - pci_write_config32(dev, 0xd4, dword); - } - - /* check PVI/SVI */ - dword = pci_read_config32(dev, 0xA0); - if (dword & PVI_MODE) { /* PVI */ - /* set slamVidMode to 0 for PVI */ - dword &= VID_SLAM_OFF | PLLLOCK_OFF; - dword |= PLLLOCK_DFT_L; - pci_write_config32(dev, 0xA0, dword); - } else { /* SVI */ - /* set slamVidMode to 1 for SVI */ - dword &= PLLLOCK_OFF; - dword |= PLLLOCK_DFT_L | VID_SLAM_ON; - pci_write_config32(dev, 0xA0, dword); - - dtemp = dword; - - /* Program F3xD8[PwrPlanes] according F3xA0[DulaVdd] */ - dword = pci_read_config32(dev, 0xD8); - - if (dtemp & DUAL_VDD_BIT) - dword |= PWR_PLN_ON; - else - dword &= PWR_PLN_OFF; - pci_write_config32(dev, 0xD8, dword); - } - - /* Note the following settings are additional from the ported - * function setFidVidRegs() - */ - dword = pci_read_config32(dev, 0xDc); - dword |= 0x5 << 12; /* NbsynPtrAdj set to 0x5 per BKDG (needs reset) */ - pci_write_config32(dev, 0xdc, dword); - - /* Rev B settings - FIXME: support other revs. */ - dword = 0xA0E641E6; - pci_write_config32(dev, 0x84, dword); - - dword = 0xE600A681; - pci_write_config32(dev, 0x80, dword); - - dword = pci_read_config32(dev, 0x80); - printk(BIOS_DEBUG, " F3x80: %08x \n", dword); - dword = pci_read_config32(dev, 0x84); - printk(BIOS_DEBUG, " F3x84: %08x \n", dword); - dword = pci_read_config32(dev, 0xD4); - printk(BIOS_DEBUG, " F3xD4: %08x \n", dword); - dword = pci_read_config32(dev, 0xD8); - printk(BIOS_DEBUG, " F3xD8: %08x \n", dword); - dword = pci_read_config32(dev, 0xDC); - printk(BIOS_DEBUG, " F3xDC: %08x \n", dword); - - - } -} - - -static void UpdateSinglePlaneNbVid(void) -{ - u32 nbVid, cpuVid; - u8 i; - msr_t msr; - - /* copy higher voltage (lower VID) of NBVID & CPUVID to both */ - for (i = 0; i < 5; i++) { - msr = rdmsr(PS_REG_BASE + i); - nbVid = (msr.lo & PS_CPU_VID_M_ON) >> PS_CPU_VID_SHFT; - cpuVid = (msr.lo & PS_NB_VID_M_ON) >> PS_NB_VID_SHFT; - - if (nbVid != cpuVid) { - if (nbVid > cpuVid) - nbVid = cpuVid; - - msr.lo = msr.lo & PS_BOTH_VID_OFF; - msr.lo = msr.lo | (u32) ((nbVid) << PS_NB_VID_SHFT); - msr.lo = msr.lo | (u32) ((nbVid) << PS_CPU_VID_SHFT); - wrmsr(PS_REG_BASE + i, msr); - } - } -} - -static void fixPsNbVidBeforeWR(u32 newNbVid, u32 coreid) -{ - msr_t msr; - u8 startup_pstate; - - /* This function sets NbVid before the warm reset. - * Get StartupPstate from MSRC001_0071. - * Read Pstate register pionted by [StartupPstate]. - * and copy its content to P0 and P1 registers. - * Copy newNbVid to P0[NbVid]. - * transition to P1 on all cores, - * then transition to P0 on core 0. - * Wait for MSRC001_0063[CurPstate] = 000b on core 0. - */ - - msr = rdmsr(0xc0010071); - startup_pstate = (msr.hi >> (32 - 32)) & 0x07; - - /* Copy startup pstate to P1 and P0 MSRs. Set the maxvid for this node in P0. - * Then transition to P1 for corex and P0 for core0. - * These setting will be cleared by the warm reset - */ - msr = rdmsr(0xC0010064 + startup_pstate); - wrmsr(0xC0010065, msr); - wrmsr(0xC0010064, msr); - - msr.lo &= ~0xFE000000; // clear nbvid - msr.lo |= newNbVid << 25; - wrmsr(0xC0010064, msr); - - UpdateSinglePlaneNbVid(); - - // Transition to P1 for all APs and P0 for core0. - msr = rdmsr(0xC0010062); - msr.lo = (msr.lo & ~0x07) | 1; - wrmsr(0xC0010062, msr); - - // Wait for P1 to set. - do { - msr = rdmsr(0xC0010063); - } while (msr.lo != 1); - - if (coreid == 0) { - msr.lo = msr.lo & ~0x07; - wrmsr(0xC0010062, msr); - // Wait for P0 to set. - do { - msr = rdmsr(0xC0010063); - } while (msr.lo != 0); - } -} - -static void coreDelay(void) -{ - u32 saved; - u32 hi, lo, msr; - u32 cycles; - - /* delay ~40us - This seems like a hack to me... - It would be nice to have a central delay function. */ - - cycles = 8000 << 3; /* x8 (number of 1.25ns ticks) */ - - msr = 0x10; /* TSC */ - _RDMSR(msr, &lo, &hi); - saved = lo; - do { - _RDMSR(msr, &lo, &hi); - } while (lo - saved < cycles); -} - -static void transitionVid(u32 targetVid, u8 dev, u8 isNb) -{ - u32 currentVid, dtemp; - msr_t msr; - u8 vsTimecode; - u16 timeTable[8] = { 10, 20, 30, 40, 60, 100, 200, 500 }; - int vsTime; - - /* This function steps or slam the Nb VID to the target VID. - * It uses VSRampTime for [SlamVidMode]=0 ([PviMode]=1) - * or VSSlamTime for [SlamVidMode]=1 ([PviMode]=0)to time period. - */ - - /* get the current VID */ - msr = rdmsr(0xC0010071); - if (isNb) - currentVid = (msr.lo >> NB_VID_POS) & BIT_MASK_7; - else - currentVid = (msr.lo >> CPU_VID_POS) & BIT_MASK_7; - - /* Read MSRC001_0070 COFVID Control Register */ - msr = rdmsr(0xC0010070); - - /* check PVI/SPI */ - dtemp = pci_read_config32(dev, 0xA0); - if (dtemp & PVI_MODE) { /* PVI, step VID */ - if (currentVid < targetVid) { - while (currentVid < targetVid) { - currentVid++; - if (isNb) - msr.lo = (msr.lo & NB_VID_MASK_OFF) | (currentVid << NB_VID_POS); - else - msr.lo = (msr.lo & CPU_VID_MASK_OFF) | (currentVid << CPU_VID_POS); - wrmsr(0xC0010070, msr); - - /* read F3xD8[VSRampTime] */ - dtemp = pci_read_config32(dev, 0xD8); - vsTimecode = (u8) ((dtemp >> VS_RAMP_T) & 0x7); - vsTime = (int)timeTable[vsTimecode]; - do { - coreDelay(); - vsTime -= 40; - } while (vsTime > 0); - } - } else if (currentVid > targetVid) { - while (currentVid > targetVid) { - currentVid--; - if (isNb) - msr.lo = (msr.lo & NB_VID_MASK_OFF) | (currentVid << NB_VID_POS); - else - msr.lo = (msr.lo & CPU_VID_MASK_OFF) | (currentVid << CPU_VID_POS); - wrmsr(0xC0010070, msr); - - /* read F3xD8[VSRampTime] */ - dtemp = pci_read_config32(dev, 0xD8); - vsTimecode = (u8) ((dtemp >> VS_RAMP_T) & 0x7); - vsTime = (int)timeTable[vsTimecode]; - do { - coreDelay(); - vsTime -= 40; - } while (vsTime > 0); - } - } - } else { /* SVI, slam VID */ - if (isNb) - msr.lo = (msr.lo & NB_VID_MASK_OFF) | (targetVid << NB_VID_POS); - else - msr.lo = (msr.lo & CPU_VID_MASK_OFF) | (targetVid << CPU_VID_POS); - wrmsr(0xC0010070, msr); - - /* read F3xD8[VSRampTime] */ - dtemp = pci_read_config32(dev, 0xD8); - vsTimecode = (u8) ((dtemp >> VS_RAMP_T) & 0x7); - vsTime = (int)timeTable[vsTimecode]; - do { - coreDelay(); - vsTime -= 40; - } while (vsTime > 0); - } -} - - -static void init_fidvid_ap(u32 bsp_apicid, u32 apicid, u32 nodeid, u32 coreid) -{ - device_t dev; - u32 vid_max; - u32 fid_max; - u8 nb_cof_vid_update; - u8 pvimode; - u32 reg1fc; - u32 send; - u8 nodes; - u8 i; - - printk(BIOS_DEBUG, "FIDVID on AP: %02x\n", apicid); - - /* Steps 1-6 of BIOS NB COF and VID Configuration - * for SVI and Single-Plane PVI Systems. - */ - - /* If any node has nb_cof_vid_update set all nodes need an update. */ - nodes = get_nodes(); - nb_cof_vid_update = 0; - for (i = 0; i < nodes; i++) { - if (pci_read_config32(NODE_PCI(i, 3), 0x1FC) & 1) { - nb_cof_vid_update = 1; - break; - } - } - - dev = NODE_PCI(nodeid, 3); - pvimode = (pci_read_config32(dev, 0xA0) >> 8) & 1; - reg1fc = pci_read_config32(dev, 0x1FC); - - if (nb_cof_vid_update) { - if (pvimode) { - vid_max = (reg1fc >> 7) & 0x7F; - fid_max = (reg1fc >> 2) & 0x1F; - - /* write newNbVid to P-state Reg's NbVid always if NbVidUpdatedAll=1 */ - fixPsNbVidBeforeWR(vid_max, coreid); - } else { /* SVI */ - vid_max = ((reg1fc >> 7) & 0x7F) - ((reg1fc >> 17) & 0x1F); - fid_max = ((reg1fc >> 2) & 0x1F) + ((reg1fc >> 14) & 0x7); - transitionVid(vid_max, dev, IS_NB); - } - - /* fid setup is handled by the BSP at the end. */ - - } else { /* ! nb_cof_vid_update */ - /* Use max values */ - if (pvimode) - UpdateSinglePlaneNbVid(); - } - - send = (nb_cof_vid_update << 16) | (fid_max << 8); - send |= (apicid << 24); // ap apicid - - // Send signal to BSP about this AP max fid - // This also indicates this AP is ready for warm reset (if required). - lapic_write(LAPIC_MSG_REG, send | 1); -} - -static u32 calc_common_fid(u32 fid_packed, u32 fid_packed_new) -{ - u32 fidmax; - u32 fidmax_new; - - fidmax = (fid_packed >> 8) & 0xFF; - - fidmax_new = (fid_packed_new >> 8) & 0xFF; - - if (fidmax > fidmax_new) { - fidmax = fidmax_new; - } - - fid_packed &= 0xFF << 16; - fid_packed |= (fidmax << 8); - fid_packed |= fid_packed_new & (0xFF << 16); // set nb_cof_vid_update - - return fid_packed; -} - -static void init_fidvid_bsp_stage1(u32 ap_apicid, void *gp) -{ - u32 readback = 0; - u32 timeout = 1; - - struct fidvid_st *fvp = gp; - int loop; - - print_debug_fv("Wait for AP stage 1: ap_apicid = ", ap_apicid); - - loop = 100000; - while (--loop > 0) { - if (lapic_remote_read(ap_apicid, LAPIC_MSG_REG, &readback) != 0) - continue; - if ((readback & 0x3f) == 1) { - timeout = 0; - break; /* target ap is in stage 1 */ - } - } - - if (timeout) { - printk(BIOS_DEBUG, "%s: timed out reading from ap %02x\n", - __func__, ap_apicid); - return; - } - - print_debug_fv("\treadback = ", readback); - - fvp->common_fid = calc_common_fid(fvp->common_fid, readback); - - print_debug_fv("\tcommon_fid(packed) = ", fvp->common_fid); - -} - -static void updateSviPsNbVidAfterWR(u32 newNbVid) -{ - msr_t msr; - u8 i; - - /* This function copies newNbVid to NbVid bits in P-state Registers[4:0] - * for SVI mode. - */ - - for (i = 0; i < 5; i++) { - msr = rdmsr(0xC0010064 + i); - if ((msr.hi >> 31) & 1) { /* PstateEn? */ - msr.lo &= ~(0x7F << 25); - msr.lo |= (newNbVid & 0x7F) << 25; - wrmsr(0xC0010064 + i, msr); - } - } -} - - -static void fixPsNbVidAfterWR(u32 newNbVid, u8 NbVidUpdatedAll) -{ - msr_t msr; - u8 i; - u8 StartupPstate; - - /* This function copies newNbVid to NbVid bits in P-state - * Registers[4:0] if its NbDid bit=0 and PstateEn bit =1 in case of - * NbVidUpdatedAll =0 or copies copies newNbVid to NbVid bits in - * P-state Registers[4:0] if its and PstateEn bit =1 in case of - * NbVidUpdatedAll=1. Then transition to StartPstate. - */ - - /* write newNbVid to P-state Reg's NbVid if its NbDid=0 */ - for (i = 0; i < 5; i++) { - msr = rdmsr(0xC0010064 + i); - /* NbDid (bit 22 of P-state Reg) == 0 or NbVidUpdatedAll = 1 */ - if ((((msr.lo >> 22) & 1) == 0) || NbVidUpdatedAll) { - msr.lo &= ~(0x7F << 25); - msr.lo |= (newNbVid & 0x7F) << 25; - wrmsr(0xC0010064 + i, msr); - } - } - - UpdateSinglePlaneNbVid(); - - /* For each core in the system, transition all cores to StartupPstate */ - msr = rdmsr(0xC0010071); - StartupPstate = msr.hi & 0x07; - msr = rdmsr(0xC0010062); - msr.lo = StartupPstate; - wrmsr(0xC0010062, msr); - - /* Wait for StartupPstate to set. */ - do { - msr = rdmsr(0xC0010063); - } while (msr.lo != StartupPstate); -} - -static void set_p0(void) -{ - msr_t msr; - - // Transition P0 for calling core. - msr = rdmsr(0xC0010062); - msr.lo = (msr.lo & ~0x07); - wrmsr(0xC0010062, msr); - - /* Wait for P0 to set. */ - do { - msr = rdmsr(0xC0010063); - } while (msr.lo != 0); -} - -static void finalPstateChange(void) -{ - /* Enble P0 on all cores for best performance. - * Linux can slow them down later if need be. - * It is safe since they will be in C1 halt - * most of the time anyway. - */ - set_p0(); -} - -static void init_fidvid_stage2(u32 apicid, u32 nodeid) -{ - msr_t msr; - device_t dev; - u32 reg1fc; - u32 dtemp; - u32 nbvid; - u8 nb_cof_vid_update; - u8 nodes; - u8 NbVidUpdateAll; - u8 i; - u8 pvimode; - - /* After warm reset finish the fid/vid setup for all cores. */ - - /* If any node has nb_cof_vid_update set all nodes need an update. */ - nodes = get_nodes(); - nb_cof_vid_update = 0; - for (i = 0; i < nodes; i++) { - if (pci_read_config32(NODE_PCI(i, 3), 0x1FC) & 1) { - nb_cof_vid_update = 1; - break; - } - } - - dev = NODE_PCI(nodeid, 3); - pvimode = (pci_read_config32(dev, 0xA0) >> 8) & 1; - reg1fc = pci_read_config32(dev, 0x1FC); - nbvid = (reg1fc >> 7) & 0x7F; - NbVidUpdateAll = (reg1fc >> 1) & 1; - - if (nb_cof_vid_update) { - if (pvimode) { - nbvid = (reg1fc >> 7) & 0x7F; - /* write newNbVid to P-state Reg's NbVid if its NbDid=0 */ - fixPsNbVidAfterWR(nbvid, NbVidUpdateAll); - } else { /* SVI */ - nbvid = ((reg1fc >> 7) & 0x7F) - ((reg1fc >> 17) & 0x1F); - updateSviPsNbVidAfterWR(nbvid); - } - } else { /* !nb_cof_vid_update */ - if (pvimode) - UpdateSinglePlaneNbVid(); - } - dtemp = pci_read_config32(dev, 0xA0); - dtemp &= PLLLOCK_OFF; - dtemp |= PLLLOCK_DFT_L; - pci_write_config32(dev, 0xA0, dtemp); - - finalPstateChange(); - - /* Set TSC to tick at the P0 ndfid rate */ - msr = rdmsr(HWCR); - msr.lo |= 1 << 24; - wrmsr(HWCR, msr); -} - - -#if SET_FIDVID_STORE_AP_APICID_AT_FIRST == 1 -struct ap_apicid_st { - u32 num; - // it could use 256 bytes for 64 node quad core system - u8 apicid[NODE_NUMS * 4]; -}; - -static void store_ap_apicid(unsigned ap_apicid, void *gp) -{ - struct ap_apicid_st *p = gp; - - p->apicid[p->num++] = ap_apicid; - -} -#endif - - -static int init_fidvid_bsp(u32 bsp_apicid, u32 nodes) -{ -#if SET_FIDVID_STORE_AP_APICID_AT_FIRST == 1 - struct ap_apicid_st ap_apicidx; - u32 i; -#endif - struct fidvid_st fv; - device_t dev; - u32 vid_max; - u32 fid_max=0; - u8 nb_cof_vid_update; - u32 reg1fc; - u8 pvimode; - - printk(BIOS_DEBUG, "FIDVID on BSP, APIC_id: %02x\n", bsp_apicid); - /* FIXME: The first half of this function is nearly the same as - * init_fidvid_bsp() and the code could be combined. - */ - - /* Steps 1-6 of BIOS NB COF and VID Configuration - * for SVI and Single-Plane PVI Systems. - */ - - /* If any node has nb_cof_vid_update set all nodes need an update. */ - nb_cof_vid_update = 0; - for (i = 0; i < nodes; i++) { - if (pci_read_config32(NODE_PCI(i, 3), 0x1FC) & 1) { - nb_cof_vid_update = 1; - break; - } - } - - dev = NODE_PCI(0, 3); - pvimode = (pci_read_config32(dev, 0xA0) >> 8) & 1; - reg1fc = pci_read_config32(dev, 0x1FC); - - if (nb_cof_vid_update) { - if (pvimode) { - vid_max = (reg1fc >> 7) & 0x7F; - fid_max = (reg1fc >> 2) & 0x1F; - - /* write newNbVid to P-state Reg's NbVid always if NbVidUpdatedAll=1 */ - fixPsNbVidBeforeWR(vid_max, 0); - } else { /* SVI */ - vid_max = ((reg1fc >> 7) & 0x7F) - ((reg1fc >> 17) & 0x1F); - fid_max = ((reg1fc >> 2) & 0x1F) + ((reg1fc >> 14) & 0x7); - transitionVid(vid_max, dev, IS_NB); - } - - /* fid setup is handled by the BSP at the end. */ - - } else { /* ! nb_cof_vid_update */ - /* Use max values */ - if (pvimode) - UpdateSinglePlaneNbVid(); - } - - fv.common_fid = (nb_cof_vid_update << 16) | (fid_max << 8); - print_debug_fv("BSP fid = ", fv.common_fid); - -#if SET_FIDVID_STORE_AP_APICID_AT_FIRST == 1 && SET_FIDVID_CORE0_ONLY == 0 - /* For all APs (We know the APIC ID of all APs even when the APIC ID - is lifted) remote read from AP LAPIC_MSG_REG about max fid. - Then calculate the common max fid that can be used for all - APs and BSP */ - ap_apicidx.num = 0; - - for_each_ap(bsp_apicid, SET_FIDVID_CORE_RANGE, store_ap_apicid, &ap_apicidx); - - for (i = 0; i < ap_apicidx.num; i++) { - init_fidvid_bsp_stage1(ap_apicidx.apicid[i], &fv); - } -#else - for_each_ap(bsp_apicid, SET_FIDVID_CORE0_ONLY, init_fidvid_bsp_stage1, &fv); -#endif - - print_debug_fv("common_fid = ", fv.common_fid); - - if (fv.common_fid & (1 << 16)) { /* check nb_cof_vid_update */ - - // Enable the common fid and other settings. - enable_fid_change((fv.common_fid >> 8) & 0x1F); - - // nbfid change need warm reset, so reset at first - return 1; - } - - return 0; // No FID/VID changes. Don't reset -} -#endif +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2007 Advanced Micro Devices, Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; version 2 of the License. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#if SET_FIDVID == 1 +#include + +#define SET_FIDVID_DEBUG 1 + +// if we are tight of CAR stack, disable it +#define SET_FIDVID_STORE_AP_APICID_AT_FIRST 1 + +static inline void print_debug_fv(const char *str, u32 val) +{ +#if SET_FIDVID_DEBUG == 1 + printk(BIOS_DEBUG, "%s%x\n", str, val); +#endif +} + +static inline void print_debug_fv_8(const char *str, u8 val) +{ +#if SET_FIDVID_DEBUG == 1 + printk(BIOS_DEBUG, "%s%02x\n", str, val); +#endif +} + +static inline void print_debug_fv_64(const char *str, u32 val, u32 val2) +{ +#if SET_FIDVID_DEBUG == 1 + printk(BIOS_DEBUG, "%s%x%x\n", str, val, val2); +#endif +} + +struct fidvid_st { + u32 common_fid; +}; + +static void enable_fid_change(u8 fid) +{ + u32 dword; + u32 nodes; + device_t dev; + int i; + + nodes = get_nodes(); + + for (i = 0; i < nodes; i++) { + dev = NODE_PCI(i, 3); + dword = pci_read_config32(dev, 0xd4); + dword &= ~0x1F; + dword |= (u32) fid & 0x1F; + dword |= 1 << 5; // enable + pci_write_config32(dev, 0xd4, dword); + printk(BIOS_DEBUG, "FID Change Node:%02x, F3xD4: %08x \n", i, + dword); + } +} + +static void recalculateVsSlamTimeSettingOnCorePre(device_t dev) +{ + u8 pviModeFlag; + u8 highVoltageVid, lowVoltageVid, bValue; + u16 minimumSlamTime; + u16 vSlamTimes[7] = { 1000, 2000, 3000, 4000, 6000, 10000, 20000 }; /* Reg settings scaled by 100 */ + u32 dtemp; + msr_t msr; + + /* This function calculates the VsSlamTime using the range of possible + * voltages instead of a hardcoded 200us. + * Note:This function is called from setFidVidRegs and setUserPs after + * programming a custom Pstate. + */ + + /* Calculate Slam Time + * Vslam = 0.4us/mV * Vp0 - (lowest out of Vpmin or Valt) + * In our case, we will scale the values by 100 to avoid + * decimals. + */ + + /* Determine if this is a PVI or SVI system */ + dtemp = pci_read_config32(dev, 0xA0); + + if (dtemp & PVI_MODE) + pviModeFlag = 1; + else + pviModeFlag = 0; + + /* Get P0's voltage */ + msr = rdmsr(0xC0010064); + highVoltageVid = (u8) ((msr.lo >> PS_CPU_VID_SHFT) & 0x7F); + + /* If SVI, we only care about CPU VID. + * If PVI, determine the higher voltage b/t NB and CPU + */ + if (pviModeFlag) { + bValue = (u8) ((msr.lo >> PS_NB_VID_SHFT) & 0x7F); + if (highVoltageVid > bValue) + highVoltageVid = bValue; + } + + /* Get Pmin's index */ + msr = rdmsr(0xC0010061); + bValue = (u8) ((msr.lo >> PS_CUR_LIM_SHFT) & BIT_MASK_3); + + /* Get Pmin's VID */ + msr = rdmsr(0xC0010064 + bValue); + lowVoltageVid = (u8) ((msr.lo >> PS_CPU_VID_SHFT) & 0x7F); + + /* If SVI, we only care about CPU VID. + * If PVI, determine the higher voltage b/t NB and CPU + */ + if (pviModeFlag) { + bValue = (u8) ((msr.lo >> PS_NB_VID_SHFT) & 0x7F); + if (lowVoltageVid > bValue) + lowVoltageVid = bValue; + } + + /* Get AltVID */ + dtemp = pci_read_config32(dev, 0xDC); + bValue = (u8) (dtemp & BIT_MASK_7); + + /* Use the VID with the lowest voltage (higher VID) */ + if (lowVoltageVid < bValue) + lowVoltageVid = bValue; + + /* If Vids are 7Dh - 7Fh, force 7Ch to keep calculations linear */ + if (lowVoltageVid > 0x7C) { + lowVoltageVid = 0x7C; + if (highVoltageVid > 0x7C) + highVoltageVid = 0x7C; + } + + bValue = (u8) (lowVoltageVid - highVoltageVid); + + /* Each Vid increment is 12.5 mV. The minimum slam time is: + * vidCodeDelta * 12.5mV * 0.4us/mV + * Scale by 100 to avoid decimals. + */ + minimumSlamTime = bValue * (125 * 4); + + /* Now round up to nearest register setting. + * Note that if we don't find a value, we + * will fall through to a value of 7 + */ + for (bValue = 0; bValue < 7; bValue++) { + if (minimumSlamTime <= vSlamTimes[bValue]) + break; + } + + /* Apply the value */ + dtemp = pci_read_config32(dev, 0xD8); + dtemp &= VSSLAM_MASK; + dtemp |= bValue; + pci_write_config32(dev, 0xd8, dtemp); +} + +static void prep_fid_change(void) +{ + u32 dword, dtemp; + u32 nodes; + device_t dev; + int i; + + /* This needs to be run before any Pstate changes are requested */ + + nodes = get_nodes(); + + for (i = 0; i < nodes; i++) { + printk(BIOS_DEBUG, "Prep FID/VID Node:%02x \n", i); + dev = NODE_PCI(i, 3); + + dword = pci_read_config32(dev, 0xd8); + dword &= VSRAMP_MASK; + dword |= VSRAMP_VALUE; + pci_write_config32(dev, 0xd8, dword); + + /* Figure out the value for VsSlamTime and program it */ + recalculateVsSlamTimeSettingOnCorePre(dev); + + /* Program fields in Clock Power/Control register0 (F3xD4) */ + /* set F3xD4 Clock Power/Timing Control 0 Register + * NbClkDidApplyAll=1b + * NbClkDid=100b + * PowerStepUp= "platform dependent" + * PowerStepDown= "platform dependent" + * LinkPllLink=01b + * ClkRampHystSel=HW default + */ + /* check platform type */ + if (!(get_platform_type() & AMD_PTYPE_SVR)) { + /* For non-server platform + * PowerStepUp=01000b - 50nS + * PowerStepDown=01000b - 50ns + */ + dword = pci_read_config32(dev, 0xd4); + dword &= CPTC0_MASK; + dword |= NB_CLKDID_ALL | NB_CLKDID | PW_STP_UP50 | PW_STP_DN50 | LNK_PLL_LOCK; /* per BKDG */ + pci_write_config32(dev, 0xd4, dword); + } else { + dword = pci_read_config32(dev, 0xd4); + dword &= CPTC0_MASK; + /* get number of cores for PowerStepUp & PowerStepDown in server + 1 core - 400nS - 0000b + 2 cores - 200nS - 0010b + 3 cores - 133nS -> 100nS - 0011b + 4 cores - 100nS - 0011b + */ + switch (get_core_num_in_bsp(i)) { + case 0: + dword |= PW_STP_UP400 | PW_STP_DN400; + break; + case 1: + case 2: + dword |= PW_STP_UP200 | PW_STP_DN200; + break; + case 3: + dword |= PW_STP_UP100 | PW_STP_DN100; + break; + default: + dword |= PW_STP_UP100 | PW_STP_DN100; + break; + } + dword |= NB_CLKDID_ALL | NB_CLKDID | LNK_PLL_LOCK; + pci_write_config32(dev, 0xd4, dword); + } + + /* check PVI/SVI */ + dword = pci_read_config32(dev, 0xA0); + if (dword & PVI_MODE) { /* PVI */ + /* set slamVidMode to 0 for PVI */ + dword &= VID_SLAM_OFF | PLLLOCK_OFF; + dword |= PLLLOCK_DFT_L; + pci_write_config32(dev, 0xA0, dword); + } else { /* SVI */ + /* set slamVidMode to 1 for SVI */ + dword &= PLLLOCK_OFF; + dword |= PLLLOCK_DFT_L | VID_SLAM_ON; + pci_write_config32(dev, 0xA0, dword); + + dtemp = dword; + + /* Program F3xD8[PwrPlanes] according F3xA0[DulaVdd] */ + dword = pci_read_config32(dev, 0xD8); + + if (dtemp & DUAL_VDD_BIT) + dword |= PWR_PLN_ON; + else + dword &= PWR_PLN_OFF; + pci_write_config32(dev, 0xD8, dword); + } + + /* Note the following settings are additional from the ported + * function setFidVidRegs() + */ + dword = pci_read_config32(dev, 0xDc); + dword |= 0x5 << 12; /* NbsynPtrAdj set to 0x5 per BKDG (needs reset) */ + pci_write_config32(dev, 0xdc, dword); + + /* Rev B settings - FIXME: support other revs. */ + dword = 0xA0E641E6; + pci_write_config32(dev, 0x84, dword); + + dword = 0xE600A681; + pci_write_config32(dev, 0x80, dword); + + dword = pci_read_config32(dev, 0x80); + printk(BIOS_DEBUG, " F3x80: %08x \n", dword); + dword = pci_read_config32(dev, 0x84); + printk(BIOS_DEBUG, " F3x84: %08x \n", dword); + dword = pci_read_config32(dev, 0xD4); + printk(BIOS_DEBUG, " F3xD4: %08x \n", dword); + dword = pci_read_config32(dev, 0xD8); + printk(BIOS_DEBUG, " F3xD8: %08x \n", dword); + dword = pci_read_config32(dev, 0xDC); + printk(BIOS_DEBUG, " F3xDC: %08x \n", dword); + + + } +} + + +static void UpdateSinglePlaneNbVid(void) +{ + u32 nbVid, cpuVid; + u8 i; + msr_t msr; + + /* copy higher voltage (lower VID) of NBVID & CPUVID to both */ + for (i = 0; i < 5; i++) { + msr = rdmsr(PS_REG_BASE + i); + nbVid = (msr.lo & PS_CPU_VID_M_ON) >> PS_CPU_VID_SHFT; + cpuVid = (msr.lo & PS_NB_VID_M_ON) >> PS_NB_VID_SHFT; + + if (nbVid != cpuVid) { + if (nbVid > cpuVid) + nbVid = cpuVid; + + msr.lo = msr.lo & PS_BOTH_VID_OFF; + msr.lo = msr.lo | (u32) ((nbVid) << PS_NB_VID_SHFT); + msr.lo = msr.lo | (u32) ((nbVid) << PS_CPU_VID_SHFT); + wrmsr(PS_REG_BASE + i, msr); + } + } +} + +static void fixPsNbVidBeforeWR(u32 newNbVid, u32 coreid) +{ + msr_t msr; + u8 startup_pstate; + + /* This function sets NbVid before the warm reset. + * Get StartupPstate from MSRC001_0071. + * Read Pstate register pionted by [StartupPstate]. + * and copy its content to P0 and P1 registers. + * Copy newNbVid to P0[NbVid]. + * transition to P1 on all cores, + * then transition to P0 on core 0. + * Wait for MSRC001_0063[CurPstate] = 000b on core 0. + */ + + msr = rdmsr(0xc0010071); + startup_pstate = (msr.hi >> (32 - 32)) & 0x07; + + /* Copy startup pstate to P1 and P0 MSRs. Set the maxvid for this node in P0. + * Then transition to P1 for corex and P0 for core0. + * These setting will be cleared by the warm reset + */ + msr = rdmsr(0xC0010064 + startup_pstate); + wrmsr(0xC0010065, msr); + wrmsr(0xC0010064, msr); + + msr.lo &= ~0xFE000000; // clear nbvid + msr.lo |= newNbVid << 25; + wrmsr(0xC0010064, msr); + + UpdateSinglePlaneNbVid(); + + // Transition to P1 for all APs and P0 for core0. + msr = rdmsr(0xC0010062); + msr.lo = (msr.lo & ~0x07) | 1; + wrmsr(0xC0010062, msr); + + // Wait for P1 to set. + do { + msr = rdmsr(0xC0010063); + } while (msr.lo != 1); + + if (coreid == 0) { + msr.lo = msr.lo & ~0x07; + wrmsr(0xC0010062, msr); + // Wait for P0 to set. + do { + msr = rdmsr(0xC0010063); + } while (msr.lo != 0); + } +} + +static void coreDelay(void) +{ + u32 saved; + u32 hi, lo, msr; + u32 cycles; + + /* delay ~40us + This seems like a hack to me... + It would be nice to have a central delay function. */ + + cycles = 8000 << 3; /* x8 (number of 1.25ns ticks) */ + + msr = 0x10; /* TSC */ + _RDMSR(msr, &lo, &hi); + saved = lo; + do { + _RDMSR(msr, &lo, &hi); + } while (lo - saved < cycles); +} + +static void transitionVid(u32 targetVid, u8 dev, u8 isNb) +{ + u32 currentVid, dtemp; + msr_t msr; + u8 vsTimecode; + u16 timeTable[8] = { 10, 20, 30, 40, 60, 100, 200, 500 }; + int vsTime; + + /* This function steps or slam the Nb VID to the target VID. + * It uses VSRampTime for [SlamVidMode]=0 ([PviMode]=1) + * or VSSlamTime for [SlamVidMode]=1 ([PviMode]=0)to time period. + */ + + /* get the current VID */ + msr = rdmsr(0xC0010071); + if (isNb) + currentVid = (msr.lo >> NB_VID_POS) & BIT_MASK_7; + else + currentVid = (msr.lo >> CPU_VID_POS) & BIT_MASK_7; + + /* Read MSRC001_0070 COFVID Control Register */ + msr = rdmsr(0xC0010070); + + /* check PVI/SPI */ + dtemp = pci_read_config32(dev, 0xA0); + if (dtemp & PVI_MODE) { /* PVI, step VID */ + if (currentVid < targetVid) { + while (currentVid < targetVid) { + currentVid++; + if (isNb) + msr.lo = (msr.lo & NB_VID_MASK_OFF) | (currentVid << NB_VID_POS); + else + msr.lo = (msr.lo & CPU_VID_MASK_OFF) | (currentVid << CPU_VID_POS); + wrmsr(0xC0010070, msr); + + /* read F3xD8[VSRampTime] */ + dtemp = pci_read_config32(dev, 0xD8); + vsTimecode = (u8) ((dtemp >> VS_RAMP_T) & 0x7); + vsTime = (int)timeTable[vsTimecode]; + do { + coreDelay(); + vsTime -= 40; + } while (vsTime > 0); + } + } else if (currentVid > targetVid) { + while (currentVid > targetVid) { + currentVid--; + if (isNb) + msr.lo = (msr.lo & NB_VID_MASK_OFF) | (currentVid << NB_VID_POS); + else + msr.lo = (msr.lo & CPU_VID_MASK_OFF) | (currentVid << CPU_VID_POS); + wrmsr(0xC0010070, msr); + + /* read F3xD8[VSRampTime] */ + dtemp = pci_read_config32(dev, 0xD8); + vsTimecode = (u8) ((dtemp >> VS_RAMP_T) & 0x7); + vsTime = (int)timeTable[vsTimecode]; + do { + coreDelay(); + vsTime -= 40; + } while (vsTime > 0); + } + } + } else { /* SVI, slam VID */ + if (isNb) + msr.lo = (msr.lo & NB_VID_MASK_OFF) | (targetVid << NB_VID_POS); + else + msr.lo = (msr.lo & CPU_VID_MASK_OFF) | (targetVid << CPU_VID_POS); + wrmsr(0xC0010070, msr); + + /* read F3xD8[VSRampTime] */ + dtemp = pci_read_config32(dev, 0xD8); + vsTimecode = (u8) ((dtemp >> VS_RAMP_T) & 0x7); + vsTime = (int)timeTable[vsTimecode]; + do { + coreDelay(); + vsTime -= 40; + } while (vsTime > 0); + } +} + + +static void init_fidvid_ap(u32 bsp_apicid, u32 apicid, u32 nodeid, u32 coreid) +{ + device_t dev; + u32 vid_max; + u32 fid_max; + u8 nb_cof_vid_update; + u8 pvimode; + u32 reg1fc; + u32 send; + u8 nodes; + u8 i; + + printk(BIOS_DEBUG, "FIDVID on AP: %02x\n", apicid); + + /* Steps 1-6 of BIOS NB COF and VID Configuration + * for SVI and Single-Plane PVI Systems. + */ + + /* If any node has nb_cof_vid_update set all nodes need an update. */ + nodes = get_nodes(); + nb_cof_vid_update = 0; + for (i = 0; i < nodes; i++) { + if (pci_read_config32(NODE_PCI(i, 3), 0x1FC) & 1) { + nb_cof_vid_update = 1; + break; + } + } + + dev = NODE_PCI(nodeid, 3); + pvimode = (pci_read_config32(dev, 0xA0) >> 8) & 1; + reg1fc = pci_read_config32(dev, 0x1FC); + + if (nb_cof_vid_update) { + if (pvimode) { + vid_max = (reg1fc >> 7) & 0x7F; + fid_max = (reg1fc >> 2) & 0x1F; + + /* write newNbVid to P-state Reg's NbVid always if NbVidUpdatedAll=1 */ + fixPsNbVidBeforeWR(vid_max, coreid); + } else { /* SVI */ + vid_max = ((reg1fc >> 7) & 0x7F) - ((reg1fc >> 17) & 0x1F); + fid_max = ((reg1fc >> 2) & 0x1F) + ((reg1fc >> 14) & 0x7); + transitionVid(vid_max, dev, IS_NB); + } + + /* fid setup is handled by the BSP at the end. */ + + } else { /* ! nb_cof_vid_update */ + /* Use max values */ + if (pvimode) + UpdateSinglePlaneNbVid(); + } + + send = (nb_cof_vid_update << 16) | (fid_max << 8); + send |= (apicid << 24); // ap apicid + + // Send signal to BSP about this AP max fid + // This also indicates this AP is ready for warm reset (if required). + lapic_write(LAPIC_MSG_REG, send | F10_APSTATE_RESET); +} + +static u32 calc_common_fid(u32 fid_packed, u32 fid_packed_new) +{ + u32 fidmax; + u32 fidmax_new; + + fidmax = (fid_packed >> 8) & 0xFF; + + fidmax_new = (fid_packed_new >> 8) & 0xFF; + + if (fidmax > fidmax_new) { + fidmax = fidmax_new; + } + + fid_packed &= 0xFF << 16; + fid_packed |= (fidmax << 8); + fid_packed |= fid_packed_new & (0xFF << 16); // set nb_cof_vid_update + + return fid_packed; +} + +static void init_fidvid_bsp_stage1(u32 ap_apicid, void *gp) +{ + u32 readback = 0; + u32 timeout = 1; + + struct fidvid_st *fvp = gp; + int loop; + + print_debug_fv("Wait for AP stage 1: ap_apicid = ", ap_apicid); + + loop = 100000; + while (--loop > 0) { + if (lapic_remote_read(ap_apicid, LAPIC_MSG_REG, &readback) != 0) + continue; + if ((readback & 0x3f) == 1) { + timeout = 0; + break; /* target ap is in stage 1 */ + } + } + + if (timeout) { + printk(BIOS_DEBUG, "%s: timed out reading from ap %02x\n", + __func__, ap_apicid); + return; + } + + print_debug_fv("\treadback = ", readback); + + fvp->common_fid = calc_common_fid(fvp->common_fid, readback); + + print_debug_fv("\tcommon_fid(packed) = ", fvp->common_fid); + +} + +static void updateSviPsNbVidAfterWR(u32 newNbVid) +{ + msr_t msr; + u8 i; + + /* This function copies newNbVid to NbVid bits in P-state Registers[4:0] + * for SVI mode. + */ + + for (i = 0; i < 5; i++) { + msr = rdmsr(0xC0010064 + i); + if ((msr.hi >> 31) & 1) { /* PstateEn? */ + msr.lo &= ~(0x7F << 25); + msr.lo |= (newNbVid & 0x7F) << 25; + wrmsr(0xC0010064 + i, msr); + } + } +} + + +static void fixPsNbVidAfterWR(u32 newNbVid, u8 NbVidUpdatedAll) +{ + msr_t msr; + u8 i; + u8 StartupPstate; + + /* This function copies newNbVid to NbVid bits in P-state + * Registers[4:0] if its NbDid bit=0 and PstateEn bit =1 in case of + * NbVidUpdatedAll =0 or copies copies newNbVid to NbVid bits in + * P-state Registers[4:0] if its and PstateEn bit =1 in case of + * NbVidUpdatedAll=1. Then transition to StartPstate. + */ + + /* write newNbVid to P-state Reg's NbVid if its NbDid=0 */ + for (i = 0; i < 5; i++) { + msr = rdmsr(0xC0010064 + i); + /* NbDid (bit 22 of P-state Reg) == 0 or NbVidUpdatedAll = 1 */ + if ((((msr.lo >> 22) & 1) == 0) || NbVidUpdatedAll) { + msr.lo &= ~(0x7F << 25); + msr.lo |= (newNbVid & 0x7F) << 25; + wrmsr(0xC0010064 + i, msr); + } + } + + UpdateSinglePlaneNbVid(); + + /* For each core in the system, transition all cores to StartupPstate */ + msr = rdmsr(0xC0010071); + StartupPstate = msr.hi & 0x07; + msr = rdmsr(0xC0010062); + msr.lo = StartupPstate; + wrmsr(0xC0010062, msr); + + /* Wait for StartupPstate to set. */ + do { + msr = rdmsr(0xC0010063); + } while (msr.lo != StartupPstate); +} + +static void set_p0(void) +{ + msr_t msr; + + // Transition P0 for calling core. + msr = rdmsr(0xC0010062); + msr.lo = (msr.lo & ~0x07); + wrmsr(0xC0010062, msr); + + /* Wait for P0 to set. */ + do { + msr = rdmsr(0xC0010063); + } while (msr.lo != 0); +} + +static void finalPstateChange(void) +{ + /* Enble P0 on all cores for best performance. + * Linux can slow them down later if need be. + * It is safe since they will be in C1 halt + * most of the time anyway. + */ + set_p0(); +} + +static void init_fidvid_stage2(u32 apicid, u32 nodeid) +{ + msr_t msr; + device_t dev; + u32 reg1fc; + u32 dtemp; + u32 nbvid; + u8 nb_cof_vid_update; + u8 nodes; + u8 NbVidUpdateAll; + u8 i; + u8 pvimode; + + /* After warm reset finish the fid/vid setup for all cores. */ + + /* If any node has nb_cof_vid_update set all nodes need an update. */ + nodes = get_nodes(); + nb_cof_vid_update = 0; + for (i = 0; i < nodes; i++) { + if (pci_read_config32(NODE_PCI(i, 3), 0x1FC) & 1) { + nb_cof_vid_update = 1; + break; + } + } + + dev = NODE_PCI(nodeid, 3); + pvimode = (pci_read_config32(dev, 0xA0) >> 8) & 1; + reg1fc = pci_read_config32(dev, 0x1FC); + nbvid = (reg1fc >> 7) & 0x7F; + NbVidUpdateAll = (reg1fc >> 1) & 1; + + if (nb_cof_vid_update) { + if (pvimode) { + nbvid = (reg1fc >> 7) & 0x7F; + /* write newNbVid to P-state Reg's NbVid if its NbDid=0 */ + fixPsNbVidAfterWR(nbvid, NbVidUpdateAll); + } else { /* SVI */ + nbvid = ((reg1fc >> 7) & 0x7F) - ((reg1fc >> 17) & 0x1F); + updateSviPsNbVidAfterWR(nbvid); + } + } else { /* !nb_cof_vid_update */ + if (pvimode) + UpdateSinglePlaneNbVid(); + } + dtemp = pci_read_config32(dev, 0xA0); + dtemp &= PLLLOCK_OFF; + dtemp |= PLLLOCK_DFT_L; + pci_write_config32(dev, 0xA0, dtemp); + + finalPstateChange(); + + /* Set TSC to tick at the P0 ndfid rate */ + msr = rdmsr(HWCR); + msr.lo |= 1 << 24; + wrmsr(HWCR, msr); +} + + +#if SET_FIDVID_STORE_AP_APICID_AT_FIRST == 1 +struct ap_apicid_st { + u32 num; + // it could use 256 bytes for 64 node quad core system + u8 apicid[NODE_NUMS * 4]; +}; + +static void store_ap_apicid(unsigned ap_apicid, void *gp) +{ + struct ap_apicid_st *p = gp; + + p->apicid[p->num++] = ap_apicid; + +} +#endif + + +static int init_fidvid_bsp(u32 bsp_apicid, u32 nodes) +{ +#if SET_FIDVID_STORE_AP_APICID_AT_FIRST == 1 + struct ap_apicid_st ap_apicidx; + u32 i; +#endif + struct fidvid_st fv; + device_t dev; + u32 vid_max; + u32 fid_max=0; + u8 nb_cof_vid_update; + u32 reg1fc; + u8 pvimode; + + printk(BIOS_DEBUG, "FIDVID on BSP, APIC_id: %02x\n", bsp_apicid); + /* FIXME: The first half of this function is nearly the same as + * init_fidvid_bsp() and the code could be combined. + */ + + /* Steps 1-6 of BIOS NB COF and VID Configuration + * for SVI and Single-Plane PVI Systems. + */ + + /* If any node has nb_cof_vid_update set all nodes need an update. */ + nb_cof_vid_update = 0; + for (i = 0; i < nodes; i++) { + if (pci_read_config32(NODE_PCI(i, 3), 0x1FC) & 1) { + nb_cof_vid_update = 1; + break; + } + } + + dev = NODE_PCI(0, 3); + pvimode = (pci_read_config32(dev, 0xA0) >> 8) & 1; + reg1fc = pci_read_config32(dev, 0x1FC); + + if (nb_cof_vid_update) { + if (pvimode) { + vid_max = (reg1fc >> 7) & 0x7F; + fid_max = (reg1fc >> 2) & 0x1F; + + /* write newNbVid to P-state Reg's NbVid always if NbVidUpdatedAll=1 */ + fixPsNbVidBeforeWR(vid_max, 0); + } else { /* SVI */ + vid_max = ((reg1fc >> 7) & 0x7F) - ((reg1fc >> 17) & 0x1F); + fid_max = ((reg1fc >> 2) & 0x1F) + ((reg1fc >> 14) & 0x7); + transitionVid(vid_max, dev, IS_NB); + } + + /* fid setup is handled by the BSP at the end. */ + + } else { /* ! nb_cof_vid_update */ + /* Use max values */ + if (pvimode) + UpdateSinglePlaneNbVid(); + } + + fv.common_fid = (nb_cof_vid_update << 16) | (fid_max << 8); + print_debug_fv("BSP fid = ", fv.common_fid); + +#if SET_FIDVID_STORE_AP_APICID_AT_FIRST == 1 && SET_FIDVID_CORE0_ONLY == 0 + /* For all APs (We know the APIC ID of all APs even when the APIC ID + is lifted) remote read from AP LAPIC_MSG_REG about max fid. + Then calculate the common max fid that can be used for all + APs and BSP */ + ap_apicidx.num = 0; + + for_each_ap(bsp_apicid, SET_FIDVID_CORE_RANGE, store_ap_apicid, &ap_apicidx); + + for (i = 0; i < ap_apicidx.num; i++) { + init_fidvid_bsp_stage1(ap_apicidx.apicid[i], &fv); + } +#else + for_each_ap(bsp_apicid, SET_FIDVID_CORE0_ONLY, init_fidvid_bsp_stage1, &fv); +#endif + + print_debug_fv("common_fid = ", fv.common_fid); + + if (fv.common_fid & (1 << 16)) { /* check nb_cof_vid_update */ + + // Enable the common fid and other settings. + enable_fid_change((fv.common_fid >> 8) & 0x1F); + + // nbfid change need warm reset, so reset at first + return 1; + } + + return 0; // No FID/VID changes. Don't reset +} +#endif diff --git a/src/cpu/amd/model_10xxx/init_cpus.c b/src/cpu/amd/model_10xxx/init_cpus.c index e582ca8af3..9207c2e0ba 100644 --- a/src/cpu/amd/model_10xxx/init_cpus.c +++ b/src/cpu/amd/model_10xxx/init_cpus.c @@ -1,948 +1,945 @@ -/* - * This file is part of the coreboot project. - * - * Copyright (C) 2007-2008 Advanced Micro Devices, Inc. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; version 2 of the License. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#include "defaults.h" -#include -#include -#include -#include -#include -#include - -#include -#include - -//it takes the CONFIG_ENABLE_APIC_EXT_ID and CONFIG_APIC_ID_OFFSET and CONFIG_LIFT_BSP_APIC_ID -#ifndef SET_FIDVID - #define SET_FIDVID 1 -#endif - -#ifndef SET_FIDVID_CORE0_ONLY - /* MSR FIDVID_CTL and FIDVID_STATUS are shared by cores, - Need to do every AP to set common FID/VID */ - #define SET_FIDVID_CORE0_ONLY 0 -#endif - -static void prep_fid_change(void); -static void init_fidvid_stage2(u32 apicid, u32 nodeid); -void cpuSetAMDMSR(void); - -#if CONFIG_PCI_IO_CFG_EXT == 1 -static void set_EnableCf8ExtCfg(void) -{ - // set the NB_CFG[46]=1; - msr_t msr; - msr = rdmsr(NB_CFG_MSR); - // EnableCf8ExtCfg: We need that to access CONFIG_PCI_IO_CFG_EXT 4K range - msr.hi |= (1 << (46 - 32)); - wrmsr(NB_CFG_MSR, msr); -} -#else -static void set_EnableCf8ExtCfg(void) { } -#endif - - -typedef void (*process_ap_t) (u32 apicid, void *gp); - -//core_range = 0 : all cores -//core range = 1 : core 0 only -//core range = 2 : cores other than core0 - -static void for_each_ap(u32 bsp_apicid, u32 core_range, process_ap_t process_ap, - void *gp) -{ - // here assume the OS don't change our apicid - u32 ap_apicid; - - u32 nodes; - u32 siblings; - u32 disable_siblings; - u32 cores_found; - u32 nb_cfg_54; - int i, j; - u32 ApicIdCoreIdSize; - - /* get_nodes define in ht_wrapper.c */ - nodes = get_nodes(); - - if (!CONFIG_LOGICAL_CPUS || - read_option(CMOS_VSTART_multi_core, CMOS_VLEN_multi_core, 0) != 0) { // 0 means multi core - disable_siblings = 1; - } else { - disable_siblings = 0; - } - - /* Assume that all node are same stepping, otherwise we can use use - nb_cfg_54 from bsp for all nodes */ - nb_cfg_54 = read_nb_cfg_54(); - - ApicIdCoreIdSize = (cpuid_ecx(0x80000008) >> 12 & 0xf); - if (ApicIdCoreIdSize) { - siblings = ((1 << ApicIdCoreIdSize) - 1); - } else { - siblings = 3; //quad core - } - - for (i = 0; i < nodes; i++) { - cores_found = get_core_num_in_bsp(i); - - u32 jstart, jend; - - if (core_range == 2) { - jstart = 1; - } else { - jstart = 0; - } - - if (disable_siblings || (core_range == 1)) { - jend = 0; - } else { - jend = cores_found; - } - - for (j = jstart; j <= jend; j++) { - ap_apicid = - i * (nb_cfg_54 ? (siblings + 1) : 1) + - j * (nb_cfg_54 ? 1 : 64); - -#if (CONFIG_ENABLE_APIC_EXT_ID == 1) && (CONFIG_APIC_ID_OFFSET > 0) -#if CONFIG_LIFT_BSP_APIC_ID == 0 - if ((i != 0) || (j != 0)) /* except bsp */ -#endif - ap_apicid += CONFIG_APIC_ID_OFFSET; -#endif - - if (ap_apicid == bsp_apicid) - continue; - - process_ap(ap_apicid, gp); - - } - } -} - -static inline int lapic_remote_read(int apicid, int reg, u32 *pvalue) -{ - int timeout; - u32 status; - int result; - lapic_wait_icr_idle(); - lapic_write(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(apicid)); - lapic_write(LAPIC_ICR, LAPIC_DM_REMRD | (reg >> 4)); - -/* Extra busy check compared to lapic.h */ - timeout = 0; - do { - status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; - } while (status == LAPIC_ICR_BUSY && timeout++ < 1000); - - timeout = 0; - do { - status = lapic_read(LAPIC_ICR) & LAPIC_ICR_RR_MASK; - } while (status == LAPIC_ICR_RR_INPROG && timeout++ < 1000); - - result = -1; - - if (status == LAPIC_ICR_RR_VALID) { - *pvalue = lapic_read(LAPIC_RRR); - result = 0; - } - return result; -} - -/* Use the LAPIC timer count register to hold each cores init status */ -#define LAPIC_MSG_REG 0x380 - -#if SET_FIDVID == 1 -static void init_fidvid_ap(u32 bsp_apicid, u32 apicid, u32 nodeid, u32 coreid); -#endif - -static inline __attribute__ ((always_inline)) -void print_apicid_nodeid_coreid(u32 apicid, struct node_core_id id, - const char *str) -{ - printk(BIOS_DEBUG, - "%s --- { APICID = %02x NODEID = %02x COREID = %02x} ---\n", str, - apicid, id.nodeid, id.coreid); -} - -static u32 wait_cpu_state(u32 apicid, u32 state) -{ - u32 readback = 0; - u32 timeout = 1; - int loop = 4000000; - while (--loop > 0) { - if (lapic_remote_read(apicid, LAPIC_MSG_REG, &readback) != 0) - continue; - if ((readback & 0x3f) == state) { - timeout = 0; - break; //target cpu is in stage started - } - } - if (timeout) { - if (readback) { - timeout = readback; - } - } - - return timeout; -} - -static void wait_ap_started(u32 ap_apicid, void *gp) -{ - u32 timeout; - timeout = wait_cpu_state(ap_apicid, 0x13); // started - printk(BIOS_DEBUG, "* AP %02x", ap_apicid); - if (timeout) { - printk(BIOS_DEBUG, " timed out:%08x\n", timeout); - } else { - printk(BIOS_DEBUG, "started\n"); - } -} - -void wait_all_other_cores_started(u32 bsp_apicid) -{ - // all aps other than core0 - printk(BIOS_DEBUG, "started ap apicid: "); - for_each_ap(bsp_apicid, 2, wait_ap_started, (void *)0); - printk(BIOS_DEBUG, "\n"); -} - -void allow_all_aps_stop(u32 bsp_apicid) -{ - /* Called by the BSP to indicate AP can stop */ - - /* FIXME Do APs use this? */ - - // allow aps to stop use 6 bits for state - lapic_write(LAPIC_MSG_REG, (bsp_apicid << 24) | 0x14); -} - -static void enable_apic_ext_id(u32 node) -{ - u32 val; - - val = pci_read_config32(NODE_HT(node), 0x68); - val |= (HTTC_APIC_EXT_SPUR | HTTC_APIC_EXT_ID | HTTC_APIC_EXT_BRD_CST); - pci_write_config32(NODE_HT(node), 0x68, val); -} - -static void STOP_CAR_AND_CPU(void) -{ - msr_t msr; - - /* Disable L2 IC to L3 connection (Only for CAR) */ - msr = rdmsr(BU_CFG2); - msr.lo &= ~(1 << ClLinesToNbDis); - wrmsr(BU_CFG2, msr); - - disable_cache_as_ram(); // inline - /* stop all cores except node0/core0 the bsp .... */ - stop_this_cpu(); -} - -#if CONFIG_RAMINIT_SYSINFO -static u32 init_cpus(u32 cpu_init_detectedx, struct sys_info *sysinfo) -#else -static u32 init_cpus(u32 cpu_init_detectedx) -#endif -{ - u32 bsp_apicid = 0; - u32 apicid; - struct node_core_id id; - - /* - * already set early mtrr in cache_as_ram.inc - */ - - /* that is from initial apicid, we need nodeid and coreid - later */ - id = get_node_core_id_x(); - - /* NB_CFG MSR is shared between cores, so we need make sure - core0 is done at first --- use wait_all_core0_started */ - if (id.coreid == 0) { - set_apicid_cpuid_lo(); /* only set it on core0 */ - set_EnableCf8ExtCfg(); /* only set it on core0 */ -#if (CONFIG_ENABLE_APIC_EXT_ID == 1) - enable_apic_ext_id(id.nodeid); -#endif - } - - enable_lapic(); - -#if (CONFIG_ENABLE_APIC_EXT_ID == 1) && (CONFIG_APIC_ID_OFFSET > 0) - u32 initial_apicid = get_initial_apicid(); - -#if CONFIG_LIFT_BSP_APIC_ID == 0 - if (initial_apicid != 0) // other than bsp -#endif - { - /* use initial apic id to lift it */ - u32 dword = lapic_read(LAPIC_ID); - dword &= ~(0xff << 24); - dword |= - (((initial_apicid + CONFIG_APIC_ID_OFFSET) & 0xff) << 24); - - lapic_write(LAPIC_ID, dword); - } -#if CONFIG_LIFT_BSP_APIC_ID == 1 - bsp_apicid += CONFIG_APIC_ID_OFFSET; -#endif - -#endif - - /* get the apicid, it may be lifted already */ - apicid = lapicid(); - - // show our apicid, nodeid, and coreid - if (id.coreid == 0) { - if (id.nodeid != 0) //all core0 except bsp - print_apicid_nodeid_coreid(apicid, id, " core0: "); - } else { //all other cores - print_apicid_nodeid_coreid(apicid, id, " corex: "); - } - - if (cpu_init_detectedx) { - print_apicid_nodeid_coreid(apicid, id, - "\n\n\nINIT detected from "); - printk(BIOS_DEBUG, "\nIssuing SOFT_RESET...\n"); - soft_reset(); - } - - if (id.coreid == 0) { - if (!(warm_reset_detect(id.nodeid))) //FIXME: INIT is checked above but check for more resets? - distinguish_cpu_resets(id.nodeid); // Also indicates we are started - } - // Mark the core as started. - lapic_write(LAPIC_MSG_REG, (apicid << 24) | 0x13); - - if (apicid != bsp_apicid) { - /* Setup each AP's cores MSRs. - * This happens after HTinit. - * The BSP runs this code in it's own path. - */ - update_microcode(cpuid_eax(1)); - cpuSetAMDMSR(); - -#if SET_FIDVID == 1 -#if (CONFIG_LOGICAL_CPUS == 1) && (SET_FIDVID_CORE0_ONLY == 1) - // Run on all AP for proper FID/VID setup. - if (id.coreid == 0) // only need set fid for core0 -#endif - { - // check warm(bios) reset to call stage2 otherwise do stage1 - if (warm_reset_detect(id.nodeid)) { - printk(BIOS_DEBUG, - "init_fidvid_stage2 apicid: %02x\n", - apicid); - init_fidvid_stage2(apicid, id.nodeid); - } else { - printk(BIOS_DEBUG, - "init_fidvid_ap(stage1) apicid: %02x\n", - apicid); - init_fidvid_ap(bsp_apicid, apicid, id.nodeid, - id.coreid); - } - } -#endif - - /* AP is ready, configure MTRRs and go to sleep */ - set_var_mtrr(0, 0x00000000, CONFIG_RAMTOP, MTRR_TYPE_WRBACK); - - STOP_CAR_AND_CPU(); - - printk(BIOS_DEBUG, - "\nAP %02x should be halted but you are reading this....\n", - apicid); - } - - return bsp_apicid; -} - -static u32 is_core0_started(u32 nodeid) -{ - u32 htic; - device_t device; - device = NODE_PCI(nodeid, 0); - htic = pci_read_config32(device, HT_INIT_CONTROL); - htic &= HTIC_ColdR_Detect; - return htic; -} - -void wait_all_core0_started(void) -{ - /* When core0 is started, it will distingush_cpu_resets - * So wait for that to finish */ - u32 i; - u32 nodes = get_nodes(); - - printk(BIOS_DEBUG, "core0 started: "); - for (i = 1; i < nodes; i++) { // skip bsp, because it is running on bsp - while (!is_core0_started(i)) { - } - printk(BIOS_DEBUG, " %02x", i); - } - printk(BIOS_DEBUG, "\n"); -} - -#if CONFIG_MAX_PHYSICAL_CPUS > 1 -/** - * void start_node(u32 node) - * - * start the core0 in node, so it can generate HT packet to feature code. - * - * This function starts the AP nodes core0s. wait_all_core0_started() in - * romstage.c waits for all the AP to be finished before continuing - * system init. - */ -static void start_node(u8 node) -{ - u32 val; - - /* Enable routing table */ - printk(BIOS_DEBUG, "Start node %02x", node); - -#if CONFIG_NORTHBRIDGE_AMD_AMDFAM10 - /* For FAM10 support, we need to set Dram base/limit for the new node */ - pci_write_config32(NODE_MP(node), 0x44, 0); - pci_write_config32(NODE_MP(node), 0x40, 3); -#endif - - /* Allow APs to make requests (ROM fetch) */ - val = pci_read_config32(NODE_HT(node), 0x6c); - val &= ~(1 << 1); - pci_write_config32(NODE_HT(node), 0x6c, val); - - printk(BIOS_DEBUG, " done.\n"); -} - -/** - * static void setup_remote_node(u32 node) - * - * Copy the BSP Adress Map to each AP. - */ -static void setup_remote_node(u8 node) -{ - /* There registers can be used with F1x114_x Address Map at the - same time, So must set them even 32 node */ - static const u16 pci_reg[] = { - /* DRAM Base/Limits Registers */ - 0x44, 0x4c, 0x54, 0x5c, 0x64, 0x6c, 0x74, 0x7c, - 0x40, 0x48, 0x50, 0x58, 0x60, 0x68, 0x70, 0x78, - 0x144, 0x14c, 0x154, 0x15c, 0x164, 0x16c, 0x174, 0x17c, - 0x140, 0x148, 0x150, 0x158, 0x160, 0x168, 0x170, 0x178, - /* MMIO Base/Limits Registers */ - 0x84, 0x8c, 0x94, 0x9c, 0xa4, 0xac, 0xb4, 0xbc, - 0x80, 0x88, 0x90, 0x98, 0xa0, 0xa8, 0xb0, 0xb8, - /* IO Base/Limits Registers */ - 0xc4, 0xcc, 0xd4, 0xdc, - 0xc0, 0xc8, 0xd0, 0xd8, - /* Configuration Map Registers */ - 0xe0, 0xe4, 0xe8, 0xec, - }; - u16 i; - - printk(BIOS_DEBUG, "setup_remote_node: %02x", node); - - /* copy the default resource map from node 0 */ - for (i = 0; i < ARRAY_SIZE(pci_reg); i++) { - u32 value; - u16 reg; - reg = pci_reg[i]; - value = pci_read_config32(NODE_MP(0), reg); - pci_write_config32(NODE_MP(node), reg, value); - - } - printk(BIOS_DEBUG, " done\n"); -} -#endif /* CONFIG_MAX_PHYSICAL_CPUS > 1 */ - -static void AMD_Errata281(u8 node, u32 revision, u32 platform) -{ - /* Workaround for Transaction Scheduling Conflict in - * Northbridge Cross Bar. Implement XCS Token adjustment - * for ganged links. Also, perform fix up for the mixed - * revision case. - */ - - u32 reg, val; - u8 i; - u8 mixed = 0; - u8 nodes = get_nodes(); - - if (platform & AMD_PTYPE_SVR) { - /* For each node we need to check for a "broken" node */ - if (!(revision & (AMD_DR_B0 | AMD_DR_B1))) { - for (i = 0; i < nodes; i++) { - if (mctGetLogicalCPUID(i) & - (AMD_DR_B0 | AMD_DR_B1)) { - mixed = 1; - break; - } - } - } - - if ((revision & (AMD_DR_B0 | AMD_DR_B1)) || mixed) { - - /* F0X68[22:21] DsNpReqLmt0 = 01b */ - val = pci_read_config32(NODE_PCI(node, 0), 0x68); - val &= ~0x00600000; - val |= 0x00200000; - pci_write_config32(NODE_PCI(node, 0), 0x68, val); - - /* F3X6C */ - val = pci_read_config32(NODE_PCI(node, 3), 0x6C); - val &= ~0x700780F7; - val |= 0x00010094; - pci_write_config32(NODE_PCI(node, 3), 0x6C, val); - - /* F3X7C */ - val = pci_read_config32(NODE_PCI(node, 3), 0x7C); - val &= ~0x707FFF1F; - val |= 0x00144514; - pci_write_config32(NODE_PCI(node, 3), 0x7C, val); - - /* F3X144[3:0] RspTok = 0001b */ - val = pci_read_config32(NODE_PCI(node, 3), 0x144); - val &= ~0x0000000F; - val |= 0x00000001; - pci_write_config32(NODE_PCI(node, 3), 0x144, val); - - for (i = 0; i < 3; i++) { - reg = 0x148 + (i * 4); - val = pci_read_config32(NODE_PCI(node, 3), reg); - val &= ~0x000000FF; - val |= 0x000000DB; - pci_write_config32(NODE_PCI(node, 3), reg, val); - } - } - } -} - -static void AMD_Errata298(void) -{ - /* Workaround for L2 Eviction May Occur during operation to - * set Accessed or dirty bit. - */ - - msr_t msr; - u8 i; - u8 affectedRev = 0; - u8 nodes = get_nodes(); - - /* For each core we need to check for a "broken" node */ - for (i = 0; i < nodes; i++) { - if (mctGetLogicalCPUID(i) & (AMD_DR_B0 | AMD_DR_B1 | AMD_DR_B2)) { - affectedRev = 1; - break; - } - } - - if (affectedRev) { - msr = rdmsr(HWCR); - msr.lo |= 0x08; /* Set TlbCacheDis bit[3] */ - wrmsr(HWCR, msr); - - msr = rdmsr(BU_CFG); - msr.lo |= 0x02; /* Set TlbForceMemTypeUc bit[1] */ - wrmsr(BU_CFG, msr); - - msr = rdmsr(OSVW_ID_Length); - msr.lo |= 0x01; /* OS Visible Workaround - MSR */ - wrmsr(OSVW_ID_Length, msr); - - msr = rdmsr(OSVW_Status); - msr.lo |= 0x01; /* OS Visible Workaround - MSR */ - wrmsr(OSVW_Status, msr); - } - - if (!affectedRev && (mctGetLogicalCPUID(0xFF) & AMD_DR_B3)) { - msr = rdmsr(OSVW_ID_Length); - msr.lo |= 0x01; /* OS Visible Workaround - MSR */ - wrmsr(OSVW_ID_Length, msr); - - } -} - -static u32 get_platform_type(void) -{ - u32 ret = 0; - - switch (SYSTEM_TYPE) { - case 1: - ret |= AMD_PTYPE_DSK; - break; - case 2: - ret |= AMD_PTYPE_MOB; - break; - case 0: - ret |= AMD_PTYPE_SVR; - break; - default: - break; - } - - /* FIXME: add UMA support. */ - - /* All Fam10 are multi core */ - ret |= AMD_PTYPE_MC; - - return ret; -} - -static void AMD_SetupPSIVID_d(u32 platform_type, u8 node) -{ - u32 dword; - int i; - msr_t msr; - - if (platform_type & (AMD_PTYPE_MOB | AMD_PTYPE_DSK)) { - - /* The following code sets the PSIVID to the lowest support P state - * assuming that the VID for the lowest power state is below - * the VDD voltage regulator threshold. (This also assumes that there - * is a Pstate lower than P0) - */ - - for (i = 4; i >= 0; i--) { - msr = rdmsr(PS_REG_BASE + i); - /* Pstate valid? */ - if (msr.hi & PS_EN_MASK) { - dword = pci_read_config32(NODE_PCI(i, 3), 0xA0); - dword &= ~0x7F; - dword |= (msr.lo >> 9) & 0x7F; - pci_write_config32(NODE_PCI(i, 3), 0xA0, dword); - break; - } - } - } -} - -/** - * AMD_CpuFindCapability - Traverse PCI capability list to find host HT links. - * HT Phy operations are not valid on links that aren't present, so this - * prevents invalid accesses. - * - * Returns the offset of the link register. - */ -static BOOL AMD_CpuFindCapability(u8 node, u8 cap_count, u8 * offset) -{ - u32 reg; - u32 val; - - /* get start of CPU HT Host Capabilities */ - val = pci_read_config32(NODE_PCI(node, 0), 0x34); - val &= 0xFF; //reg offset of first link - - cap_count++; - - /* Traverse through the capabilities. */ - do { - reg = pci_read_config32(NODE_PCI(node, 0), val); - /* Is the capability block a HyperTransport capability block? */ - if ((reg & 0xFF) == 0x08) { - /* Is the HT capability block an HT Host Capability? */ - if ((reg & 0xE0000000) == (1 << 29)) - cap_count--; - } - - if (cap_count) - val = (reg >> 8) & 0xFF; //update reg offset - } while (cap_count && val); - - *offset = (u8) val; - - /* If requested capability found val != 0 */ - if (!cap_count) - return TRUE; - else - return FALSE; -} - -/** - * AMD_checkLinkType - Compare desired link characteristics using a logical - * link type mask. - * - * Returns the link characteristic mask. - */ -static u32 AMD_checkLinkType(u8 node, u8 link, u8 regoff) -{ - u32 val; - u32 linktype = 0; - - /* Check connect, init and coherency */ - val = pci_read_config32(NODE_PCI(node, 0), regoff + 0x18); - val &= 0x1F; - - if (val == 3) - linktype |= HTPHY_LINKTYPE_COHERENT; - - if (val == 7) - linktype |= HTPHY_LINKTYPE_NONCOHERENT; - - if (linktype) { - /* Check gen3 */ - val = pci_read_config32(NODE_PCI(node, 0), regoff + 0x08); - - if (((val >> 8) & 0x0F) > 6) - linktype |= HTPHY_LINKTYPE_HT3; - else - linktype |= HTPHY_LINKTYPE_HT1; - - /* Check ganged */ - val = pci_read_config32(NODE_PCI(node, 0), (link << 2) + 0x170); - - if (val & 1) - linktype |= HTPHY_LINKTYPE_GANGED; - else - linktype |= HTPHY_LINKTYPE_UNGANGED; - } - return linktype; -} - -/** - * AMD_SetHtPhyRegister - Use the HT link's HT Phy portal registers to update - * a phy setting for that link. - */ -static void AMD_SetHtPhyRegister(u8 node, u8 link, u8 entry) -{ - u32 phyReg; - u32 phyBase; - u32 val; - - /* Determine this link's portal */ - if (link > 3) - link -= 4; - - phyBase = ((u32) link << 3) | 0x180; - - /* Get the portal control register's initial value - * and update it to access the desired phy register - */ - phyReg = pci_read_config32(NODE_PCI(node, 4), phyBase); - - if (fam10_htphy_default[entry].htreg > 0x1FF) { - phyReg &= ~HTPHY_DIRECT_OFFSET_MASK; - phyReg |= HTPHY_DIRECT_MAP; - } else { - phyReg &= ~HTPHY_OFFSET_MASK; - } - - /* Now get the current phy register data - * LinkPhyDone = 0, LinkPhyWrite = 0 is a read - */ - phyReg |= fam10_htphy_default[entry].htreg; - pci_write_config32(NODE_PCI(node, 4), phyBase, phyReg); - - do { - val = pci_read_config32(NODE_PCI(node, 4), phyBase); - } while (!(val & HTPHY_IS_COMPLETE_MASK)); - - /* Now we have the phy register data, apply the change */ - val = pci_read_config32(NODE_PCI(node, 4), phyBase + 4); - val &= ~fam10_htphy_default[entry].mask; - val |= fam10_htphy_default[entry].data; - pci_write_config32(NODE_PCI(node, 4), phyBase + 4, val); - - /* write it through the portal to the phy - * LinkPhyDone = 0, LinkPhyWrite = 1 is a write - */ - phyReg |= HTPHY_WRITE_CMD; - pci_write_config32(NODE_PCI(node, 4), phyBase, phyReg); - - do { - val = pci_read_config32(NODE_PCI(node, 4), phyBase); - } while (!(val & HTPHY_IS_COMPLETE_MASK)); -} - -void cpuSetAMDMSR(void) -{ - /* This routine loads the CPU with default settings in fam10_msr_default - * table . It must be run after Cache-As-RAM has been enabled, and - * Hypertransport initialization has taken place. Also note - * that it is run on the current processor only, and only for the current - * processor core. - */ - msr_t msr; - u8 i; - u32 revision, platform; - - printk(BIOS_DEBUG, "cpuSetAMDMSR "); - - revision = mctGetLogicalCPUID(0xFF); - platform = get_platform_type(); - - for (i = 0; i < ARRAY_SIZE(fam10_msr_default); i++) { - if ((fam10_msr_default[i].revision & revision) && - (fam10_msr_default[i].platform & platform)) { - msr = rdmsr(fam10_msr_default[i].msr); - msr.hi &= ~fam10_msr_default[i].mask_hi; - msr.hi |= fam10_msr_default[i].data_hi; - msr.lo &= ~fam10_msr_default[i].mask_lo; - msr.lo |= fam10_msr_default[i].data_lo; - wrmsr(fam10_msr_default[i].msr, msr); - } - } - AMD_Errata298(); - - printk(BIOS_DEBUG, " done\n"); -} - -static void cpuSetAMDPCI(u8 node) -{ - /* This routine loads the CPU with default settings in fam10_pci_default - * table . It must be run after Cache-As-RAM has been enabled, and - * Hypertransport initialization has taken place. Also note - * that it is run for the first core on each node - */ - u8 i, j; - u32 revision, platform; - u32 val; - u8 offset; - - printk(BIOS_DEBUG, "cpuSetAMDPCI %02d", node); - - revision = mctGetLogicalCPUID(node); - platform = get_platform_type(); - - AMD_SetupPSIVID_d(platform, node); /* Set PSIVID offset which is not table driven */ - - for (i = 0; i < ARRAY_SIZE(fam10_pci_default); i++) { - if ((fam10_pci_default[i].revision & revision) && - (fam10_pci_default[i].platform & platform)) { - val = pci_read_config32(NODE_PCI(node, - fam10_pci_default[i]. - function), - fam10_pci_default[i].offset); - val &= ~fam10_pci_default[i].mask; - val |= fam10_pci_default[i].data; - pci_write_config32(NODE_PCI(node, - fam10_pci_default[i]. - function), - fam10_pci_default[i].offset, val); - } - } - - for (i = 0; i < ARRAY_SIZE(fam10_htphy_default); i++) { - if ((fam10_htphy_default[i].revision & revision) && - (fam10_htphy_default[i].platform & platform)) { - /* HT Phy settings either apply to both sublinks or have - * separate registers for sublink zero and one, so there - * will be two table entries. So, here we only loop - * through the sublink zeros in function zero. - */ - for (j = 0; j < 4; j++) { - if (AMD_CpuFindCapability(node, j, &offset)) { - if (AMD_checkLinkType(node, j, offset) - & fam10_htphy_default[i].linktype) { - AMD_SetHtPhyRegister(node, j, - i); - } - } else { - /* No more capabilities, - * link not present - */ - break; - } - } - } - } - - /* FIXME: add UMA support and programXbarToSriReg(); */ - - AMD_Errata281(node, revision, platform); - - /* FIXME: if the dct phy doesn't init correct it needs to reset. - if (revision & (AMD_DR_B2 | AMD_DR_B3)) - dctPhyDiag(); */ - - printk(BIOS_DEBUG, " done\n"); -} - -#ifdef UNUSED_CODE -static void cpuInitializeMCA(void) -{ - /* Clears Machine Check Architecture (MCA) registers, which power on - * containing unknown data, on currently running processor. - * This routine should only be executed on initial power on (cold boot), - * not across a warm reset because valid data is present at that time. - */ - - msr_t msr; - u32 reg; - u8 i; - - if (cpuid_edx(1) & 0x4080) { /* MCE and MCA (edx[7] and edx[14]) */ - msr = rdmsr(MCG_CAP); - if (msr.lo & MCG_CTL_P) { /* MCG_CTL_P bit is set? */ - msr.lo &= 0xFF; - msr.lo--; - msr.lo <<= 2; /* multiply the count by 4 */ - reg = MC0_STA + msr.lo; - msr.lo = msr.hi = 0; - for (i = 0; i < 4; i++) { - wrmsr(reg, msr); - reg -= 4; /* Touch status regs for each bank */ - } - } - } -} -#endif - -/** - * finalize_node_setup() - * - * Do any additional post HT init - * - */ -static void finalize_node_setup(struct sys_info *sysinfo) -{ - u8 i; - u8 nodes = get_nodes(); - u32 reg; - -#if CONFIG_RAMINIT_SYSINFO - /* read Node0 F0_0x64 bit [8:10] to find out SbLink # */ - reg = pci_read_config32(NODE_HT(0), 0x64); - sysinfo->sblk = (reg >> 8) & 7; - sysinfo->sbbusn = 0; - sysinfo->nodes = nodes; - sysinfo->sbdn = get_sbdn(sysinfo->sbbusn); -#endif - - for (i = 0; i < nodes; i++) { - cpuSetAMDPCI(i); - } - -#if SET_FIDVID == 1 - // Prep each node for FID/VID setup. - prep_fid_change(); -#endif - -#if CONFIG_MAX_PHYSICAL_CPUS > 1 - /* Skip the BSP, start at node 1 */ - for (i = 1; i < nodes; i++) { - setup_remote_node(i); - start_node(i); - } -#endif -} - -#include "fidvid.c" +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2007-2008 Advanced Micro Devices, Inc. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; version 2 of the License. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "defaults.h" +#include +#include +#include +#include +#include +#include + +#include +#include + +//it takes the CONFIG_ENABLE_APIC_EXT_ID and CONFIG_APIC_ID_OFFSET and CONFIG_LIFT_BSP_APIC_ID +#ifndef SET_FIDVID + #define SET_FIDVID 1 +#endif + +#ifndef SET_FIDVID_CORE0_ONLY + /* MSR FIDVID_CTL and FIDVID_STATUS are shared by cores, + Need to do every AP to set common FID/VID */ + #define SET_FIDVID_CORE0_ONLY 0 +#endif + +static void prep_fid_change(void); +static void init_fidvid_stage2(u32 apicid, u32 nodeid); +void cpuSetAMDMSR(void); + +#if CONFIG_PCI_IO_CFG_EXT == 1 +static void set_EnableCf8ExtCfg(void) +{ + // set the NB_CFG[46]=1; + msr_t msr; + msr = rdmsr(NB_CFG_MSR); + // EnableCf8ExtCfg: We need that to access CONFIG_PCI_IO_CFG_EXT 4K range + msr.hi |= (1 << (46 - 32)); + wrmsr(NB_CFG_MSR, msr); +} +#else +static void set_EnableCf8ExtCfg(void) { } +#endif + + +typedef void (*process_ap_t) (u32 apicid, void *gp); + +//core_range = 0 : all cores +//core range = 1 : core 0 only +//core range = 2 : cores other than core0 + +static void for_each_ap(u32 bsp_apicid, u32 core_range, process_ap_t process_ap, + void *gp) +{ + // here assume the OS don't change our apicid + u32 ap_apicid; + + u32 nodes; + u32 siblings; + u32 disable_siblings; + u32 cores_found; + u32 nb_cfg_54; + int i, j; + u32 ApicIdCoreIdSize; + + /* get_nodes define in ht_wrapper.c */ + nodes = get_nodes(); + + if (!CONFIG_LOGICAL_CPUS || + read_option(CMOS_VSTART_multi_core, CMOS_VLEN_multi_core, 0) != 0) { // 0 means multi core + disable_siblings = 1; + } else { + disable_siblings = 0; + } + + /* Assume that all node are same stepping, otherwise we can use use + nb_cfg_54 from bsp for all nodes */ + nb_cfg_54 = read_nb_cfg_54(); + + ApicIdCoreIdSize = (cpuid_ecx(0x80000008) >> 12 & 0xf); + if (ApicIdCoreIdSize) { + siblings = ((1 << ApicIdCoreIdSize) - 1); + } else { + siblings = 3; //quad core + } + + for (i = 0; i < nodes; i++) { + cores_found = get_core_num_in_bsp(i); + + u32 jstart, jend; + + if (core_range == 2) { + jstart = 1; + } else { + jstart = 0; + } + + if (disable_siblings || (core_range == 1)) { + jend = 0; + } else { + jend = cores_found; + } + + for (j = jstart; j <= jend; j++) { + ap_apicid = + i * (nb_cfg_54 ? (siblings + 1) : 1) + + j * (nb_cfg_54 ? 1 : 64); + +#if (CONFIG_ENABLE_APIC_EXT_ID == 1) && (CONFIG_APIC_ID_OFFSET > 0) +#if CONFIG_LIFT_BSP_APIC_ID == 0 + if ((i != 0) || (j != 0)) /* except bsp */ +#endif + ap_apicid += CONFIG_APIC_ID_OFFSET; +#endif + + if (ap_apicid == bsp_apicid) + continue; + + process_ap(ap_apicid, gp); + + } + } +} + +static inline int lapic_remote_read(int apicid, int reg, u32 *pvalue) +{ + int timeout; + u32 status; + int result; + lapic_wait_icr_idle(); + lapic_write(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(apicid)); + lapic_write(LAPIC_ICR, LAPIC_DM_REMRD | (reg >> 4)); + +/* Extra busy check compared to lapic.h */ + timeout = 0; + do { + status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; + } while (status == LAPIC_ICR_BUSY && timeout++ < 1000); + + timeout = 0; + do { + status = lapic_read(LAPIC_ICR) & LAPIC_ICR_RR_MASK; + } while (status == LAPIC_ICR_RR_INPROG && timeout++ < 1000); + + result = -1; + + if (status == LAPIC_ICR_RR_VALID) { + *pvalue = lapic_read(LAPIC_RRR); + result = 0; + } + return result; +} + +#if SET_FIDVID == 1 +static void init_fidvid_ap(u32 bsp_apicid, u32 apicid, u32 nodeid, u32 coreid); +#endif + +static inline __attribute__ ((always_inline)) +void print_apicid_nodeid_coreid(u32 apicid, struct node_core_id id, + const char *str) +{ + printk(BIOS_DEBUG, + "%s --- { APICID = %02x NODEID = %02x COREID = %02x} ---\n", str, + apicid, id.nodeid, id.coreid); +} + +static u32 wait_cpu_state(u32 apicid, u32 state) +{ + u32 readback = 0; + u32 timeout = 1; + int loop = 4000000; + while (--loop > 0) { + if (lapic_remote_read(apicid, LAPIC_MSG_REG, &readback) != 0) + continue; + if ((readback & 0x3f) == state || (readback & 0x3f) == F10_APSTATE_RESET) { + timeout = 0; + break; //target cpu is in stage started + } + } + if (timeout) { + if (readback) { + timeout = readback; + } + } + + return timeout; +} + +static void wait_ap_started(u32 ap_apicid, void *gp) +{ + u32 timeout; + timeout = wait_cpu_state(ap_apicid, F10_APSTATE_STARTED); + printk(BIOS_DEBUG, "* AP %02x", ap_apicid); + if (timeout) { + printk(BIOS_DEBUG, " timed out:%08x\n", timeout); + } else { + printk(BIOS_DEBUG, "started\n"); + } +} + +void wait_all_other_cores_started(u32 bsp_apicid) +{ + // all aps other than core0 + printk(BIOS_DEBUG, "started ap apicid: "); + for_each_ap(bsp_apicid, 2, wait_ap_started, (void *)0); + printk(BIOS_DEBUG, "\n"); +} + +void allow_all_aps_stop(u32 bsp_apicid) +{ + /* Called by the BSP to indicate AP can stop */ + + /* FIXME Do APs use this? */ + + // allow aps to stop use 6 bits for state + lapic_write(LAPIC_MSG_REG, (bsp_apicid << 24) | F10_APSTATE_STOPPED); +} + +static void enable_apic_ext_id(u32 node) +{ + u32 val; + + val = pci_read_config32(NODE_HT(node), 0x68); + val |= (HTTC_APIC_EXT_SPUR | HTTC_APIC_EXT_ID | HTTC_APIC_EXT_BRD_CST); + pci_write_config32(NODE_HT(node), 0x68, val); +} + +static void STOP_CAR_AND_CPU(void) +{ + msr_t msr; + + /* Disable L2 IC to L3 connection (Only for CAR) */ + msr = rdmsr(BU_CFG2); + msr.lo &= ~(1 << ClLinesToNbDis); + wrmsr(BU_CFG2, msr); + + disable_cache_as_ram(); // inline + /* stop all cores except node0/core0 the bsp .... */ + stop_this_cpu(); +} + +#if CONFIG_RAMINIT_SYSINFO +static u32 init_cpus(u32 cpu_init_detectedx, struct sys_info *sysinfo) +#else +static u32 init_cpus(u32 cpu_init_detectedx) +#endif +{ + u32 bsp_apicid = 0; + u32 apicid; + struct node_core_id id; + + /* + * already set early mtrr in cache_as_ram.inc + */ + + /* that is from initial apicid, we need nodeid and coreid + later */ + id = get_node_core_id_x(); + + /* NB_CFG MSR is shared between cores, so we need make sure + core0 is done at first --- use wait_all_core0_started */ + if (id.coreid == 0) { + set_apicid_cpuid_lo(); /* only set it on core0 */ + set_EnableCf8ExtCfg(); /* only set it on core0 */ +#if (CONFIG_ENABLE_APIC_EXT_ID == 1) + enable_apic_ext_id(id.nodeid); +#endif + } + + enable_lapic(); + +#if (CONFIG_ENABLE_APIC_EXT_ID == 1) && (CONFIG_APIC_ID_OFFSET > 0) + u32 initial_apicid = get_initial_apicid(); + +#if CONFIG_LIFT_BSP_APIC_ID == 0 + if (initial_apicid != 0) // other than bsp +#endif + { + /* use initial apic id to lift it */ + u32 dword = lapic_read(LAPIC_ID); + dword &= ~(0xff << 24); + dword |= + (((initial_apicid + CONFIG_APIC_ID_OFFSET) & 0xff) << 24); + + lapic_write(LAPIC_ID, dword); + } +#if CONFIG_LIFT_BSP_APIC_ID == 1 + bsp_apicid += CONFIG_APIC_ID_OFFSET; +#endif + +#endif + + /* get the apicid, it may be lifted already */ + apicid = lapicid(); + + // show our apicid, nodeid, and coreid + if (id.coreid == 0) { + if (id.nodeid != 0) //all core0 except bsp + print_apicid_nodeid_coreid(apicid, id, " core0: "); + } else { //all other cores + print_apicid_nodeid_coreid(apicid, id, " corex: "); + } + + if (cpu_init_detectedx) { + print_apicid_nodeid_coreid(apicid, id, + "\n\n\nINIT detected from "); + printk(BIOS_DEBUG, "\nIssuing SOFT_RESET...\n"); + soft_reset(); + } + + if (id.coreid == 0) { + if (!(warm_reset_detect(id.nodeid))) //FIXME: INIT is checked above but check for more resets? + distinguish_cpu_resets(id.nodeid); // Also indicates we are started + } + // Mark the core as started. + lapic_write(LAPIC_MSG_REG, (apicid << 24) | F10_APSTATE_STARTED); + + if (apicid != bsp_apicid) { + /* Setup each AP's cores MSRs. + * This happens after HTinit. + * The BSP runs this code in it's own path. + */ + update_microcode(cpuid_eax(1)); + cpuSetAMDMSR(); + +#if SET_FIDVID == 1 +#if (CONFIG_LOGICAL_CPUS == 1) && (SET_FIDVID_CORE0_ONLY == 1) + // Run on all AP for proper FID/VID setup. + if (id.coreid == 0) // only need set fid for core0 +#endif + { + // check warm(bios) reset to call stage2 otherwise do stage1 + if (warm_reset_detect(id.nodeid)) { + printk(BIOS_DEBUG, + "init_fidvid_stage2 apicid: %02x\n", + apicid); + init_fidvid_stage2(apicid, id.nodeid); + } else { + printk(BIOS_DEBUG, + "init_fidvid_ap(stage1) apicid: %02x\n", + apicid); + init_fidvid_ap(bsp_apicid, apicid, id.nodeid, + id.coreid); + } + } +#endif + + /* AP is ready, configure MTRRs and go to sleep */ + set_var_mtrr(0, 0x00000000, CONFIG_RAMTOP, MTRR_TYPE_WRBACK); + + STOP_CAR_AND_CPU(); + + printk(BIOS_DEBUG, + "\nAP %02x should be halted but you are reading this....\n", + apicid); + } + + return bsp_apicid; +} + +static u32 is_core0_started(u32 nodeid) +{ + u32 htic; + device_t device; + device = NODE_PCI(nodeid, 0); + htic = pci_read_config32(device, HT_INIT_CONTROL); + htic &= HTIC_ColdR_Detect; + return htic; +} + +void wait_all_core0_started(void) +{ + /* When core0 is started, it will distingush_cpu_resets + * So wait for that to finish */ + u32 i; + u32 nodes = get_nodes(); + + printk(BIOS_DEBUG, "core0 started: "); + for (i = 1; i < nodes; i++) { // skip bsp, because it is running on bsp + while (!is_core0_started(i)) { + } + printk(BIOS_DEBUG, " %02x", i); + } + printk(BIOS_DEBUG, "\n"); +} + +#if CONFIG_MAX_PHYSICAL_CPUS > 1 +/** + * void start_node(u32 node) + * + * start the core0 in node, so it can generate HT packet to feature code. + * + * This function starts the AP nodes core0s. wait_all_core0_started() in + * romstage.c waits for all the AP to be finished before continuing + * system init. + */ +static void start_node(u8 node) +{ + u32 val; + + /* Enable routing table */ + printk(BIOS_DEBUG, "Start node %02x", node); + +#if CONFIG_NORTHBRIDGE_AMD_AMDFAM10 + /* For FAM10 support, we need to set Dram base/limit for the new node */ + pci_write_config32(NODE_MP(node), 0x44, 0); + pci_write_config32(NODE_MP(node), 0x40, 3); +#endif + + /* Allow APs to make requests (ROM fetch) */ + val = pci_read_config32(NODE_HT(node), 0x6c); + val &= ~(1 << 1); + pci_write_config32(NODE_HT(node), 0x6c, val); + + printk(BIOS_DEBUG, " done.\n"); +} + +/** + * static void setup_remote_node(u32 node) + * + * Copy the BSP Adress Map to each AP. + */ +static void setup_remote_node(u8 node) +{ + /* There registers can be used with F1x114_x Address Map at the + same time, So must set them even 32 node */ + static const u16 pci_reg[] = { + /* DRAM Base/Limits Registers */ + 0x44, 0x4c, 0x54, 0x5c, 0x64, 0x6c, 0x74, 0x7c, + 0x40, 0x48, 0x50, 0x58, 0x60, 0x68, 0x70, 0x78, + 0x144, 0x14c, 0x154, 0x15c, 0x164, 0x16c, 0x174, 0x17c, + 0x140, 0x148, 0x150, 0x158, 0x160, 0x168, 0x170, 0x178, + /* MMIO Base/Limits Registers */ + 0x84, 0x8c, 0x94, 0x9c, 0xa4, 0xac, 0xb4, 0xbc, + 0x80, 0x88, 0x90, 0x98, 0xa0, 0xa8, 0xb0, 0xb8, + /* IO Base/Limits Registers */ + 0xc4, 0xcc, 0xd4, 0xdc, + 0xc0, 0xc8, 0xd0, 0xd8, + /* Configuration Map Registers */ + 0xe0, 0xe4, 0xe8, 0xec, + }; + u16 i; + + printk(BIOS_DEBUG, "setup_remote_node: %02x", node); + + /* copy the default resource map from node 0 */ + for (i = 0; i < ARRAY_SIZE(pci_reg); i++) { + u32 value; + u16 reg; + reg = pci_reg[i]; + value = pci_read_config32(NODE_MP(0), reg); + pci_write_config32(NODE_MP(node), reg, value); + + } + printk(BIOS_DEBUG, " done\n"); +} +#endif /* CONFIG_MAX_PHYSICAL_CPUS > 1 */ + +static void AMD_Errata281(u8 node, u32 revision, u32 platform) +{ + /* Workaround for Transaction Scheduling Conflict in + * Northbridge Cross Bar. Implement XCS Token adjustment + * for ganged links. Also, perform fix up for the mixed + * revision case. + */ + + u32 reg, val; + u8 i; + u8 mixed = 0; + u8 nodes = get_nodes(); + + if (platform & AMD_PTYPE_SVR) { + /* For each node we need to check for a "broken" node */ + if (!(revision & (AMD_DR_B0 | AMD_DR_B1))) { + for (i = 0; i < nodes; i++) { + if (mctGetLogicalCPUID(i) & + (AMD_DR_B0 | AMD_DR_B1)) { + mixed = 1; + break; + } + } + } + + if ((revision & (AMD_DR_B0 | AMD_DR_B1)) || mixed) { + + /* F0X68[22:21] DsNpReqLmt0 = 01b */ + val = pci_read_config32(NODE_PCI(node, 0), 0x68); + val &= ~0x00600000; + val |= 0x00200000; + pci_write_config32(NODE_PCI(node, 0), 0x68, val); + + /* F3X6C */ + val = pci_read_config32(NODE_PCI(node, 3), 0x6C); + val &= ~0x700780F7; + val |= 0x00010094; + pci_write_config32(NODE_PCI(node, 3), 0x6C, val); + + /* F3X7C */ + val = pci_read_config32(NODE_PCI(node, 3), 0x7C); + val &= ~0x707FFF1F; + val |= 0x00144514; + pci_write_config32(NODE_PCI(node, 3), 0x7C, val); + + /* F3X144[3:0] RspTok = 0001b */ + val = pci_read_config32(NODE_PCI(node, 3), 0x144); + val &= ~0x0000000F; + val |= 0x00000001; + pci_write_config32(NODE_PCI(node, 3), 0x144, val); + + for (i = 0; i < 3; i++) { + reg = 0x148 + (i * 4); + val = pci_read_config32(NODE_PCI(node, 3), reg); + val &= ~0x000000FF; + val |= 0x000000DB; + pci_write_config32(NODE_PCI(node, 3), reg, val); + } + } + } +} + +static void AMD_Errata298(void) +{ + /* Workaround for L2 Eviction May Occur during operation to + * set Accessed or dirty bit. + */ + + msr_t msr; + u8 i; + u8 affectedRev = 0; + u8 nodes = get_nodes(); + + /* For each core we need to check for a "broken" node */ + for (i = 0; i < nodes; i++) { + if (mctGetLogicalCPUID(i) & (AMD_DR_B0 | AMD_DR_B1 | AMD_DR_B2)) { + affectedRev = 1; + break; + } + } + + if (affectedRev) { + msr = rdmsr(HWCR); + msr.lo |= 0x08; /* Set TlbCacheDis bit[3] */ + wrmsr(HWCR, msr); + + msr = rdmsr(BU_CFG); + msr.lo |= 0x02; /* Set TlbForceMemTypeUc bit[1] */ + wrmsr(BU_CFG, msr); + + msr = rdmsr(OSVW_ID_Length); + msr.lo |= 0x01; /* OS Visible Workaround - MSR */ + wrmsr(OSVW_ID_Length, msr); + + msr = rdmsr(OSVW_Status); + msr.lo |= 0x01; /* OS Visible Workaround - MSR */ + wrmsr(OSVW_Status, msr); + } + + if (!affectedRev && (mctGetLogicalCPUID(0xFF) & AMD_DR_B3)) { + msr = rdmsr(OSVW_ID_Length); + msr.lo |= 0x01; /* OS Visible Workaround - MSR */ + wrmsr(OSVW_ID_Length, msr); + + } +} + +static u32 get_platform_type(void) +{ + u32 ret = 0; + + switch (SYSTEM_TYPE) { + case 1: + ret |= AMD_PTYPE_DSK; + break; + case 2: + ret |= AMD_PTYPE_MOB; + break; + case 0: + ret |= AMD_PTYPE_SVR; + break; + default: + break; + } + + /* FIXME: add UMA support. */ + + /* All Fam10 are multi core */ + ret |= AMD_PTYPE_MC; + + return ret; +} + +static void AMD_SetupPSIVID_d(u32 platform_type, u8 node) +{ + u32 dword; + int i; + msr_t msr; + + if (platform_type & (AMD_PTYPE_MOB | AMD_PTYPE_DSK)) { + + /* The following code sets the PSIVID to the lowest support P state + * assuming that the VID for the lowest power state is below + * the VDD voltage regulator threshold. (This also assumes that there + * is a Pstate lower than P0) + */ + + for (i = 4; i >= 0; i--) { + msr = rdmsr(PS_REG_BASE + i); + /* Pstate valid? */ + if (msr.hi & PS_EN_MASK) { + dword = pci_read_config32(NODE_PCI(i, 3), 0xA0); + dword &= ~0x7F; + dword |= (msr.lo >> 9) & 0x7F; + pci_write_config32(NODE_PCI(i, 3), 0xA0, dword); + break; + } + } + } +} + +/** + * AMD_CpuFindCapability - Traverse PCI capability list to find host HT links. + * HT Phy operations are not valid on links that aren't present, so this + * prevents invalid accesses. + * + * Returns the offset of the link register. + */ +static BOOL AMD_CpuFindCapability(u8 node, u8 cap_count, u8 * offset) +{ + u32 reg; + u32 val; + + /* get start of CPU HT Host Capabilities */ + val = pci_read_config32(NODE_PCI(node, 0), 0x34); + val &= 0xFF; //reg offset of first link + + cap_count++; + + /* Traverse through the capabilities. */ + do { + reg = pci_read_config32(NODE_PCI(node, 0), val); + /* Is the capability block a HyperTransport capability block? */ + if ((reg & 0xFF) == 0x08) { + /* Is the HT capability block an HT Host Capability? */ + if ((reg & 0xE0000000) == (1 << 29)) + cap_count--; + } + + if (cap_count) + val = (reg >> 8) & 0xFF; //update reg offset + } while (cap_count && val); + + *offset = (u8) val; + + /* If requested capability found val != 0 */ + if (!cap_count) + return TRUE; + else + return FALSE; +} + +/** + * AMD_checkLinkType - Compare desired link characteristics using a logical + * link type mask. + * + * Returns the link characteristic mask. + */ +static u32 AMD_checkLinkType(u8 node, u8 link, u8 regoff) +{ + u32 val; + u32 linktype = 0; + + /* Check connect, init and coherency */ + val = pci_read_config32(NODE_PCI(node, 0), regoff + 0x18); + val &= 0x1F; + + if (val == 3) + linktype |= HTPHY_LINKTYPE_COHERENT; + + if (val == 7) + linktype |= HTPHY_LINKTYPE_NONCOHERENT; + + if (linktype) { + /* Check gen3 */ + val = pci_read_config32(NODE_PCI(node, 0), regoff + 0x08); + + if (((val >> 8) & 0x0F) > 6) + linktype |= HTPHY_LINKTYPE_HT3; + else + linktype |= HTPHY_LINKTYPE_HT1; + + /* Check ganged */ + val = pci_read_config32(NODE_PCI(node, 0), (link << 2) + 0x170); + + if (val & 1) + linktype |= HTPHY_LINKTYPE_GANGED; + else + linktype |= HTPHY_LINKTYPE_UNGANGED; + } + return linktype; +} + +/** + * AMD_SetHtPhyRegister - Use the HT link's HT Phy portal registers to update + * a phy setting for that link. + */ +static void AMD_SetHtPhyRegister(u8 node, u8 link, u8 entry) +{ + u32 phyReg; + u32 phyBase; + u32 val; + + /* Determine this link's portal */ + if (link > 3) + link -= 4; + + phyBase = ((u32) link << 3) | 0x180; + + /* Get the portal control register's initial value + * and update it to access the desired phy register + */ + phyReg = pci_read_config32(NODE_PCI(node, 4), phyBase); + + if (fam10_htphy_default[entry].htreg > 0x1FF) { + phyReg &= ~HTPHY_DIRECT_OFFSET_MASK; + phyReg |= HTPHY_DIRECT_MAP; + } else { + phyReg &= ~HTPHY_OFFSET_MASK; + } + + /* Now get the current phy register data + * LinkPhyDone = 0, LinkPhyWrite = 0 is a read + */ + phyReg |= fam10_htphy_default[entry].htreg; + pci_write_config32(NODE_PCI(node, 4), phyBase, phyReg); + + do { + val = pci_read_config32(NODE_PCI(node, 4), phyBase); + } while (!(val & HTPHY_IS_COMPLETE_MASK)); + + /* Now we have the phy register data, apply the change */ + val = pci_read_config32(NODE_PCI(node, 4), phyBase + 4); + val &= ~fam10_htphy_default[entry].mask; + val |= fam10_htphy_default[entry].data; + pci_write_config32(NODE_PCI(node, 4), phyBase + 4, val); + + /* write it through the portal to the phy + * LinkPhyDone = 0, LinkPhyWrite = 1 is a write + */ + phyReg |= HTPHY_WRITE_CMD; + pci_write_config32(NODE_PCI(node, 4), phyBase, phyReg); + + do { + val = pci_read_config32(NODE_PCI(node, 4), phyBase); + } while (!(val & HTPHY_IS_COMPLETE_MASK)); +} + +void cpuSetAMDMSR(void) +{ + /* This routine loads the CPU with default settings in fam10_msr_default + * table . It must be run after Cache-As-RAM has been enabled, and + * Hypertransport initialization has taken place. Also note + * that it is run on the current processor only, and only for the current + * processor core. + */ + msr_t msr; + u8 i; + u32 revision, platform; + + printk(BIOS_DEBUG, "cpuSetAMDMSR "); + + revision = mctGetLogicalCPUID(0xFF); + platform = get_platform_type(); + + for (i = 0; i < ARRAY_SIZE(fam10_msr_default); i++) { + if ((fam10_msr_default[i].revision & revision) && + (fam10_msr_default[i].platform & platform)) { + msr = rdmsr(fam10_msr_default[i].msr); + msr.hi &= ~fam10_msr_default[i].mask_hi; + msr.hi |= fam10_msr_default[i].data_hi; + msr.lo &= ~fam10_msr_default[i].mask_lo; + msr.lo |= fam10_msr_default[i].data_lo; + wrmsr(fam10_msr_default[i].msr, msr); + } + } + AMD_Errata298(); + + printk(BIOS_DEBUG, " done\n"); +} + +static void cpuSetAMDPCI(u8 node) +{ + /* This routine loads the CPU with default settings in fam10_pci_default + * table . It must be run after Cache-As-RAM has been enabled, and + * Hypertransport initialization has taken place. Also note + * that it is run for the first core on each node + */ + u8 i, j; + u32 revision, platform; + u32 val; + u8 offset; + + printk(BIOS_DEBUG, "cpuSetAMDPCI %02d", node); + + revision = mctGetLogicalCPUID(node); + platform = get_platform_type(); + + AMD_SetupPSIVID_d(platform, node); /* Set PSIVID offset which is not table driven */ + + for (i = 0; i < ARRAY_SIZE(fam10_pci_default); i++) { + if ((fam10_pci_default[i].revision & revision) && + (fam10_pci_default[i].platform & platform)) { + val = pci_read_config32(NODE_PCI(node, + fam10_pci_default[i]. + function), + fam10_pci_default[i].offset); + val &= ~fam10_pci_default[i].mask; + val |= fam10_pci_default[i].data; + pci_write_config32(NODE_PCI(node, + fam10_pci_default[i]. + function), + fam10_pci_default[i].offset, val); + } + } + + for (i = 0; i < ARRAY_SIZE(fam10_htphy_default); i++) { + if ((fam10_htphy_default[i].revision & revision) && + (fam10_htphy_default[i].platform & platform)) { + /* HT Phy settings either apply to both sublinks or have + * separate registers for sublink zero and one, so there + * will be two table entries. So, here we only loop + * through the sublink zeros in function zero. + */ + for (j = 0; j < 4; j++) { + if (AMD_CpuFindCapability(node, j, &offset)) { + if (AMD_checkLinkType(node, j, offset) + & fam10_htphy_default[i].linktype) { + AMD_SetHtPhyRegister(node, j, + i); + } + } else { + /* No more capabilities, + * link not present + */ + break; + } + } + } + } + + /* FIXME: add UMA support and programXbarToSriReg(); */ + + AMD_Errata281(node, revision, platform); + + /* FIXME: if the dct phy doesn't init correct it needs to reset. + if (revision & (AMD_DR_B2 | AMD_DR_B3)) + dctPhyDiag(); */ + + printk(BIOS_DEBUG, " done\n"); +} + +#ifdef UNUSED_CODE +static void cpuInitializeMCA(void) +{ + /* Clears Machine Check Architecture (MCA) registers, which power on + * containing unknown data, on currently running processor. + * This routine should only be executed on initial power on (cold boot), + * not across a warm reset because valid data is present at that time. + */ + + msr_t msr; + u32 reg; + u8 i; + + if (cpuid_edx(1) & 0x4080) { /* MCE and MCA (edx[7] and edx[14]) */ + msr = rdmsr(MCG_CAP); + if (msr.lo & MCG_CTL_P) { /* MCG_CTL_P bit is set? */ + msr.lo &= 0xFF; + msr.lo--; + msr.lo <<= 2; /* multiply the count by 4 */ + reg = MC0_STA + msr.lo; + msr.lo = msr.hi = 0; + for (i = 0; i < 4; i++) { + wrmsr(reg, msr); + reg -= 4; /* Touch status regs for each bank */ + } + } + } +} +#endif + +/** + * finalize_node_setup() + * + * Do any additional post HT init + * + */ +static void finalize_node_setup(struct sys_info *sysinfo) +{ + u8 i; + u8 nodes = get_nodes(); + u32 reg; + +#if CONFIG_RAMINIT_SYSINFO + /* read Node0 F0_0x64 bit [8:10] to find out SbLink # */ + reg = pci_read_config32(NODE_HT(0), 0x64); + sysinfo->sblk = (reg >> 8) & 7; + sysinfo->sbbusn = 0; + sysinfo->nodes = nodes; + sysinfo->sbdn = get_sbdn(sysinfo->sbbusn); +#endif + + for (i = 0; i < nodes; i++) { + cpuSetAMDPCI(i); + } + +#if SET_FIDVID == 1 + // Prep each node for FID/VID setup. + prep_fid_change(); +#endif + +#if CONFIG_MAX_PHYSICAL_CPUS > 1 + /* Skip the BSP, start at node 1 */ + for (i = 1; i < nodes; i++) { + setup_remote_node(i); + start_node(i); + } +#endif +} + +#include "fidvid.c" -- cgit v1.2.3