From 23211b0200eac11c64fe9dadadb5ec5ef1deabc6 Mon Sep 17 00:00:00 2001 From: Alexandru Gagniuc Date: Sun, 9 Jun 2013 16:06:07 -0500 Subject: VIA VX900: Add minimal ramstage needed to boot linux This is the minimal code needed to get past ramstage, load SeaBIOS, jump to GRUB2, and boot linux (or load memtest). See individual source files for the status of each individual component. Change-Id: Ib7d5d7593c945f18af2c2fc5e0ae689ba66131a2 Signed-off-by: Alexandru Gagniuc Reviewed-on: http://review.coreboot.org/3419 Tested-by: build bot (Jenkins) Reviewed-by: Ronald G. Minnich --- src/northbridge/via/vx900/Makefile.inc | 12 ++ src/northbridge/via/vx900/chip.h | 52 +++++ src/northbridge/via/vx900/chrome9hd.c | 368 ++++++++++++++++++++++++++++++++ src/northbridge/via/vx900/lpc.c | 223 +++++++++++++++++++ src/northbridge/via/vx900/northbridge.c | 341 +++++++++++++++++++++++++++++ src/northbridge/via/vx900/pcie.c | 123 +++++++++++ src/northbridge/via/vx900/sata.c | 280 ++++++++++++++++++++++++ src/northbridge/via/vx900/traf_ctrl.c | 145 +++++++++++++ src/northbridge/via/vx900/vx900.h | 1 + 9 files changed, 1545 insertions(+) create mode 100644 src/northbridge/via/vx900/chip.h create mode 100644 src/northbridge/via/vx900/chrome9hd.c create mode 100644 src/northbridge/via/vx900/lpc.c create mode 100644 src/northbridge/via/vx900/northbridge.c create mode 100644 src/northbridge/via/vx900/pcie.c create mode 100644 src/northbridge/via/vx900/sata.c create mode 100644 src/northbridge/via/vx900/traf_ctrl.c (limited to 'src/northbridge') diff --git a/src/northbridge/via/vx900/Makefile.inc b/src/northbridge/via/vx900/Makefile.inc index 1586c87751..13d7de65ed 100644 --- a/src/northbridge/via/vx900/Makefile.inc +++ b/src/northbridge/via/vx900/Makefile.inc @@ -31,6 +31,18 @@ romstage-y += ./../../../southbridge/via/common/early_smbus_wait_until_ready.c romstage-y += ./../../../drivers/pc80/udelay_io.c romstage-$(CONFIG_COLLECT_TIMESTAMPS) += ./../../../lib/cbmem.c +ramstage-y += pci_util.c +ramstage-y += pcie.c +ramstage-y += northbridge.c +ramstage-y += chrome9hd.c +ramstage-y += traf_ctrl.c +ramstage-y += sata.c +ramstage-y += lpc.c + +# The buildsystem only includes this file if CONFIG_VGA is selected. +# We need to do some VGA I/O before the VGA can be initialized. We can make good +# use of some of the functions there, so include them unconditionally +ramstage-y += ./../../../drivers/pc80/vga/vga_io.c chipset_bootblock_inc += $(src)/northbridge/via/vx900/romstrap.inc diff --git a/src/northbridge/via/vx900/chip.h b/src/northbridge/via/vx900/chip.h new file mode 100644 index 0000000000..6334a8e123 --- /dev/null +++ b/src/northbridge/via/vx900/chip.h @@ -0,0 +1,52 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2012 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +struct northbridge_via_vx900_config { + /** + * \brief PCIe Lane[3:0] Function Select + * + * PCIe Lane3~Lane0 (PEXTX[3:0]P/VCC) can be used by the integrated + * graphic controller to output its display data. The PCIe lanes will + * be used to output DisplayPort data. + */ + u8 assign_pex_to_dp; + + /** + * \brief Lane Width for Root Port 1 + * + * Two PCIe lanes are used for Root port 1. Root port 2 is disabled. + */ + u8 pcie_port1_2_lane_wide; + + /** + * \brief PIRQ line to which to route the external interrupt + * + * The VX900 features an external interrupt which can be routed to any + * of the PIRQA->PIRQH lines. Usually, on-board devices are connected + * to the external interrupt. In some vendor BIOS's pirq table, this + * appears as link 9. + * + * Setting this line only affects the behavior of the integrated PIC. It + * has no effect on the IOAPIC. + * + * The value of this register must be a literal upper-case character + * from 'A' to 'H'. + */ + char ext_int_route_to_pirq; +}; diff --git a/src/northbridge/via/vx900/chrome9hd.c b/src/northbridge/via/vx900/chrome9hd.c new file mode 100644 index 0000000000..633328e164 --- /dev/null +++ b/src/northbridge/via/vx900/chrome9hd.c @@ -0,0 +1,368 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2012 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +#include +#include +#include +#include +#include +#include + +#include "vx900.h" + +#define CHROME_9_HD_MIN_FB_SIZE 8 +#define CHROME_9_HD_MAX_FB_SIZE 512 + +/** + * @file chrome9hd.c + * + * \brief Initialization for Chrome9HD integrated graphics adapters + * + * This takes care of the initialization we need to do before calling the VGA + * BIOS. The device is not documented in the VX900 datasheet. + * + * The device is documented in: + * Open Graphics Programming Manual + * Chrome9GraphicsHD Processor + * VX900 Series System Processor + * Part I: Graphics Core / 2D + * + * This document was released by VIA to the Xorg project, and is available at: + * + * + * STATUS: + * We do the minimal initialization described in VIA documents. Running the VGA + * option ROM does not get us a usable display. We configure the framebuffer and + * the IGP is able to use it. GRUB2 and linux are capable of getting a usable + * text console, which uses the monitor's native resolution (even 1920x1080). + * The graphical console (linux) does not work properly. + * @TODO + * 1. Figure out what sequence we need to do to get the VGA BIOS running + * properly. Use the code provided by VIA and compare their sequence to ours, + * fill in any missing steps, etc. + * 2. Make BAR2 and the framebuffer use the same memory space. This is a feature + * called "Direct framebuffer access" which allows us to save memory space by + * setting BAR2 of the VGA to the location in memory of the framebuffer. This + * reduces the amount of PCI MMIO space we need below 4G, and is especially + * useful considering we only have 8GB (33 bits) of memory-mapped space. + */ + +/* Helper to determine the framebuffer size */ +u32 chrome9hd_fb_size(void) +{ + static u32 fb_size = 0; + u8 reg8, ranksize; + u32 size_mb, tom_mb, max_size_mb; + int i; + /* We do some PCI and CMOS IO to find our value, so if we've already + * found it, save some time */ + if (fb_size != 0) + return fb_size; + /* FIXME: read fb_size from CMOS, but until that is implemented, start + * from 512MB */ + size_mb = 512; + + /* The minimum framebuffer size is 8MB. */ + size_mb = max(size_mb, CHROME_9_HD_MIN_FB_SIZE); + + const device_t mcu = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_MEMCTRL, + 0); + /* + * We have two limitations on the maximum framebuffer size: + * 1) (Sanity) No more that 1/4 of system RAM + * 2) (Hardware limitation) No larger than DRAM in last rank + * Check both of these limitations and apply them to our framebuffer */ + tom_mb = (pci_read_config16(mcu, 0x88) & 0x07ff) << (24 - 20); + max_size_mb = tom_mb >> 2; + if (size_mb > max_size_mb) { + printk(BIOS_ALERT, "The framebuffer size of of %dMB is larger" + " than 1/4 of available memory.\n" + " Limiting framebuffer to %dMB\n", size_mb, max_size_mb); + size_mb = max_size_mb; + } + + /* Now handle limitation #2 + * Look at the ending address of the memory ranks, from last to first, + * until we find one that is not zero. That is our last rank, and its + * size is the limit of our framebuffer. */ + /* FIXME: This has a bug. If we remap memory above 4G, we consider the + * memory hole as part of our RAM. Thus if we install 3G, with a TOLM of + * 2.5G, our TOM will be at 5G and we'll assume we have 5G RAM instead + * of the actual 3.5G */ + for (i = VX900_MAX_MEM_RANKS - 1; i > -1; i--) { + reg8 = pci_read_config8(mcu, 0x40 + i); + if (reg8 == 0) + continue; + /* We've reached the last populated rank */ + ranksize = reg8 - pci_read_config8(mcu, 0x48 + i); + max_size_mb = ranksize << 6; + /* That's it. We got what we needed. */ + break; + }; + if (size_mb > max_size_mb) { + printk(BIOS_ALERT, "The framebuffer size of %dMB is larger" + " than size of the last DRAM rank.\n" + " Limiting framebuffer to %dMB\n", size_mb, max_size_mb); + size_mb = max_size_mb; + } + + /* Now round the framebuffer size to the closest power of 2 */ + u8 fb_pow = 0; + while (size_mb >> fb_pow) + fb_pow++; + fb_pow--; + size_mb = (1 << fb_pow); + /* We store the framebuffer size in bytes, for simplicity */ + fb_size = size_mb << 20; + return fb_size; +} + +/** + * @defgroup vx900_int15 + * + * \brief INT15 helpers for Chrome9HD IGP + * + * The following are helpers for INT15 handlers for the VGA BIOS. The full set + * of INT15 callbacks is described in + * + * VIA/S3Graphics + * Video BIOS External Interface Specification for Chrome9 Series IGP + * VX900 Series + * + * This document is only available under NDA, however, the callbacks are very + * similar to other VIA/Intel IGP callbacks. + * + * Callback 0x5f18 is the most important one. It informs the VGA BIOS of the + * RAM speed and framebuffer size. The other callbacks seem to be optional. + * @{ + */ + +/** + * \brief Get X86_BL value for VGA INT15 function 5f18 + * + * Int15 5f18 lets the VGA BIOS know the framebuffer size and the memory speed. + * This handler is very important. If it is not implemented, the VGA BIOS will + * not work correctly. + * + * To use, just call this from the 15f18 handler, and place the return value in + * X86_BL + * + * @code{.c} + * case 0x5f18: + * X86_BX = vx900_int15_get_5f18_bl(); + * res = 0; + * break; + * @endcode + * + */ +u8 vx900_int15_get_5f18_bl(void) +{ + u8 reg8, ret; + device_t dev; + /* + * BL Bit[7:4] + * Memory Data Rate (not to be confused with fCLK) + * 0000: 66MHz + * 0001: 100MHz + * 0010: 133MHz + * 0011: 200MHz ( DDR200 ) + * 0100: 266MHz ( DDR266 ) + * 0101: 333MHz ( DDR333 ) + * 0110: 400MHz ( DDR400 ) + * 0111: 533MHz ( DDR I/II 533) + * 1000: 667MHz ( DDR I/II 667) + * 1001: 800MHz ( DDR3 800) + * 1010: 1066MHz ( DDR3 1066) + * 1011: 1333MHz ( DDR3 1333) + * Bit[3:0] + * N: Frame Buffer Size 2^N MB + */ + dev = dev_find_slot(0, PCI_DEVFN(0, 3)); + reg8 = pci_read_config8(dev, 0xa1); + ret = (u32) ((reg8 & 0x70) >> 4) + 2; + reg8 = pci_read_config8(dev, 0x90); + reg8 = ((reg8 & 0x07) + 3) << 4; + ret |= (u32) reg8; + + return ret; +} +/** @} */ + +static void chrome9hd_set_sid_vid(u16 vendor, u16 device) +{ + vga_sr_write(0x36, vendor >> 8); /* SVID high byte */ + vga_sr_write(0x35, vendor & 0xff); /* SVID low byte */ + vga_sr_write(0x38, device >> 8); /* SID high byte */ + vga_sr_write(0x37, device & 0xff); /* SID low byte */ +} + +static void chrome9hd_handle_uma(device_t dev) +{ + /* Mirror mirror, shiny glass, tell me that is not my ass */ + u32 fb_size = chrome9hd_fb_size() >> 20; + + //uma_resource(dev, 0x18, uma_memory_base>>10, uma_memory_size>>10); + + printk(BIOS_DEBUG, "UMA base 0x%.8llx (%lluMB)\n", uma_memory_base, + uma_memory_base >> 20); + printk(BIOS_DEBUG, "UMA size 0x%.8llx (%lluMB)\n", uma_memory_size, + uma_memory_size >> 20); + u8 fb_pow = 0; + while (fb_size >> fb_pow) + fb_pow++; + fb_pow--; + + /* Step 6 - Let MCU know the framebuffer size */ + device_t mcu = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_MEMCTRL, 0); + pci_mod_config8(mcu, 0xa1, 7 << 4, (fb_pow - 2) << 4); + + /* Step 7 - Let GFX know the framebuffer size (through PCI and IOCTL) + * The size we set here affects the behavior of BAR2, and the amount of + * MMIO space it requests. The default is 512MB, so if we don't set this + * before reading the resources, we could waste space below 4G */ + pci_write_config8(dev, 0xb2, ((0xff << (fb_pow - 2)) & ~(1 << 7))); + vga_sr_write(0x68, (0xff << (fb_pow - 1))); + /* And also that the framebuffer is in the system, RAM */ + pci_mod_config8(dev, 0xb0, 0, 1 << 0); +} + +/** + * \brief Initialization sequence before running the VGA BIOS + * + * This is the initialization sequence described in: + * + * BIOS Porting Guide + * VX900 Series + * All-in-One System Processor + * + * This document is only available under NDA. + */ +static void chrome9hd_biosguide_init_seq(device_t dev) +{ + device_t mcu = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_MEMCTRL, 0); + device_t host = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_HOST_BR, 0); + + /* Step 1 - Enable VGA controller */ + /* FIXME: This is the VGA hole @ 640k-768k, and the vga port io + * We need the port IO, but can we disable the memory hole? */ + pci_mod_config8(mcu, 0xa4, 0, (1 << 7)); /* VGA memory hole */ + + /* Step 2 - Forward MDA cycles to GFX */ + pci_mod_config8(host, 0x4e, 0, (1 << 1)); + + /* Step 3 - Enable GFX I/O space */ + pci_mod_config8(dev, PCI_COMMAND, 0, PCI_COMMAND_IO); + + /* Step 4 - Enable video subsystem */ + vga_enable_mask((1 << 0), (1 << 0)); + + /* FIXME: VGA IO Address Select. 3B5 or 3D5? */ + vga_misc_mask((1 << 0), (1 << 0)); + + /* Step 5 - Unlock accessing of IO space */ + vga_sr_write(0x10, 0x01); + + chrome9hd_handle_uma(dev); + + /* Step 8 - Enable memory base register on the GFX */ + if (uma_memory_base == 0) + die("uma_memory_base not set. Abandon ship!\n"); + printk(BIOS_DEBUG, "UMA base 0x%.10llx (%lluMB)\n", uma_memory_base, + uma_memory_base >> 20); + vga_sr_write(0x6d, (uma_memory_base >> 21) & 0xff); /* base 28:21 */ + vga_sr_write(0x6e, (uma_memory_base >> 29) & 0xff); /* base 36:29 */ + vga_sr_write(0x6f, 0x00); /* base 43:37 */ + + /* Step 9 - Set SID/VID */ + chrome9hd_set_sid_vid(0x1106, 0x7122); + +} + +static void chrome9hd_init(device_t dev) +{ + print_debug("======================================================\n"); + print_debug("== Chrome9 HD INIT\n"); + print_debug("======================================================\n"); + + chrome9hd_biosguide_init_seq(dev); + + /* Prime PLL FIXME: bad comment */ + vga_sr_mask(0x3c, 1 << 2, 1 << 2); + + /* FIXME: recheck; VGA IO Address Select. 3B5 or 3D5? */ + vga_misc_mask(1 << 0, 1 << 0); + + /* FIXME: recheck; Enable Base VGA 16 Bits Decode */ + ////pci_mod_config8(host, 0x4e, 0, 1<<4); + + u32 fb_address = pci_read_config32(dev, PCI_BASE_ADDRESS_2); + fb_address &= ~0x0F; + if (!fb_address) { + printk(BIOS_WARNING, "Chrome9HD: No FB BAR assigned!\n"); + return; + } + + printk(BIOS_INFO, "Chrome: Using %dMB Framebuffer at 0x%08X.\n", + 256, fb_address); + + printk(BIOS_DEBUG, "Initializing VGA...\n"); + + pci_dev_init(dev); + + printk(BIOS_DEBUG, "Enable VGA console\n"); + + dump_pci_device(dev); +} + +static void chrome9hd_enable(device_t dev) +{ + device_t mcu = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_MEMCTRL, 0); + /* FIXME: here? -=- ACLK 250Mhz */ + pci_mod_config8(mcu, 0xbb, 0, 0x01); +} + +static void chrome9hd_disable(device_t dev) +{ + device_t mcu = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_MEMCTRL, 0); + /* Disable GFX - This step effectively renders the GFX inert + * It won't even show up as a PCI device during enumeration */ + pci_mod_config8(mcu, 0xa1, 1 << 7, 0); +} + +static struct device_operations chrome9hd_operations = { + .read_resources = pci_dev_read_resources, + .set_resources = pci_dev_set_resources, + .enable_resources = pci_dev_enable_resources, + .init = chrome9hd_init, + .disable = chrome9hd_disable, + .enable = chrome9hd_enable, + .ops_pci = 0, +}; + +static const struct pci_driver chrome9hd_driver __pci_driver = { + .ops = &chrome9hd_operations, + .vendor = PCI_VENDOR_ID_VIA, + .device = PCI_DEVICE_ID_VIA_VX900_VGA, +}; diff --git a/src/northbridge/via/vx900/lpc.c b/src/northbridge/via/vx900/lpc.c new file mode 100644 index 0000000000..ac5e4c840e --- /dev/null +++ b/src/northbridge/via/vx900/lpc.c @@ -0,0 +1,223 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2012-2013 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include "vx900.h" +#include "chip.h" + +/** + * @file lpc.c + * + * STATUS: + * We do a fair bit of setup, and most of it seems to work fairly well. There + * are still a few FIXME items here and there, but overall, this code hasn't + * been touched much from its initial 2012 version to 2013, when it was revived. + * + * We do the IOAPIC setup with the assumption that it is declared in the + * mainboard's devicetree.cb. We cannot use the IOAPIC however. The interrupts + * do not make it to the CPU. This issue is still under investigation. + * + * We also route PIRQs with CONFIG_PIRQ_ROUTE. This is currently the only way to + * get interrupts working. + * + * On the VX900, the keyboard can be connected directly to the chipset + * (referenced as "internal keyboard" in the documents). As long as that is the + * case (not connected to the superIO), and we disable the superIO keyboard LDN, + * it will work, but perhaps this should be more configurable. + */ + +static void vx900_lpc_misc_stuff(device_t dev) +{ + char extint; + u8 val; + struct northbridge_via_vx900_config *nb = (void *)dev->chip_info; + + /* GPIO 11,10 to SATALED [1,0] */ + pci_mod_config8(dev, 0xe4, 0, 1 << 0); + + /* Route the external interrupt line */ + extint = nb->ext_int_route_to_pirq; + if (extint < 'A' || extint > 'H') { + printk(BIOS_WARNING, "Invalid PIRQ%c for external interrupt\n", + extint); + } else { + printk(BIOS_INFO, "Routing external interrupt to PIRQ%c\n", + extint); + val = extint - 'A'; + val |= (1 << 3); /* bit3 enables the external int */ + pci_mod_config8(dev, 0x55, 0xf, val); + + } +} + +static void vx900_lpc_dma_setup(device_t dev) +{ + /* These are the steps recommended by VIA in order to get DMA running */ + + /* Enable Positive South Module PCI Cycle Decoding */ + /* FIXME: Setting this seems to hang our system */ + //pci_mod_config8(dev, 0x58, 0, 1<<4); + /* Positive decoding for ROM + APIC + On-board IO ports */ + pci_mod_config8(dev, 0x6c, 0, (1 << 2) | (1 << 3) | (1 << 7)); + /* Enable DMA channels. BIOS guide recommends DMA channel 2 off */ + pci_write_config8(dev, 0x53, 0xfb); + /* Disable PCI/DMA Memory Cycles Output to PCI Bus */ + pci_mod_config8(dev, 0x5b, (1 << 5), 0); + /* DMA bandwidth control - Improved bandwidth */ + pci_write_config8(dev, 0x53, 0xff); + /* ISA Positive Decoding control */ + pci_write_config8(dev, 0x6d, 0xdf); + pci_write_config8(dev, 0x6e, 0x98); + pci_write_config8(dev, 0x6f, 0x30); +} + +/** + *\brief VX900: Set up the south module IOAPIC (for the ISA/LPC bus) + * + * Enable the IOAPIC in the south module, and properly set it up. + * \n + * This is the hardware specific initialization for the IOAPIC, and complements + * the setup done by the generic IOAPIC driver. In order for the IOAPIC to work + * properly, it _must_ be declared in devicetree.cb . + * \n + * We are assuming this is called before the drivers/generic/ioapic code, + * which should be the case if devicetree.cb is set up properly. + */ +static void vx900_lpc_ioapic_setup(device_t dev) +{ + /* Find the IOAPIC, and make sure it's set up correctly in devicetree.cb + * If it's not, then the generic ioapic driver will not set it up + * correctly, and the MP table will not be correctly generated */ + device_t ioapic; + for (ioapic = dev->next; ioapic; ioapic = ioapic->next) { + if (ioapic->path.type == DEVICE_PATH_IOAPIC) + break; + } + + /* You did put an IOAPIC in devicetree.cb, didn't you? */ + if (ioapic == 0) { + /* We don't have enough info to set up the IOAPIC */ + printk(BIOS_ERR, "ERROR: South module IOAPIC not found. " + "Check your devicetree.cb\n"); + return; + } + + /* Found an IOAPIC, now we need to make sure it's the right one */ + ioapic_config_t *config = (ioapic_config_t *) ioapic->chip_info; + if (!config->have_isa_interrupts) { + /* Umh, is this the right IOAPIC ? */ + printk(BIOS_ERR, "ERROR: South module IOAPIC not carrying ISA " + "interrupts. Check your devicetree.cb\n"); + printk(BIOS_ERR, "Will not initialize this IOAPIC.\n"); + return; + } + + /* The base address of this IOAPIC _must_ be at 0xfec00000. + * Don't move this value to a #define, as people might think it's + * configurable. It is not. */ + const u32 base = config->base; + if (base != 0xfec00000) { + printk(BIOS_ERR, "ERROR: South module IOAPIC base should be at " + "0xfec00000\n but we found it at 0x%.8x\n", base); + return; + } + + print_debug("VX900 LPC: Setting up the south module IOAPIC.\n"); + /* Enable IOAPIC + * So much work for one line of code. Talk about bloat :) + * The 8259 PIC should still work even if the IOAPIC is enabled, so + * there's no crime in enabling the IOAPIC here. */ + pci_mod_config8(dev, 0x58, 0, 1 << 6); +} + +static void vx900_lpc_interrupt_stuff(device_t dev) +{ + /* Enable setting trigger mode through 0x4d0, and 0x4d1 ports + * And enable I/O recovery time */ + pci_mod_config8(dev, 0x40, 0, (1 << 2) | (1 << 6)); + /* Set serial IRQ frame width to 6 PCI cycles (recommended by VIA) + * And enable serial IRQ */ + pci_mod_config8(dev, 0x52, 3 << 0, (1 << 3) | (1 << 0)); + + /* Disable IRQ12 storm FIXME: bad comment */ + pci_mod_config8(dev, 0x51, (1 << 2), 0); + + pci_write_config8(dev, 0x4c, (1 << 6)); + + /* FIXME: Do we really need this? SeaBIOS/linux runs fine without it. + * Is this something the payload/OS should do, or is it safe for us to + * do it? */ + /* Get the IRQs up and running */ + setup_i8259(); + + vx900_lpc_dma_setup(dev); + + /* The IOAPIC is special, and we treat it separately */ + vx900_lpc_ioapic_setup(dev); +} + +static void vx900_lpc_init(device_t dev) +{ + vx900_lpc_interrupt_stuff(dev); + vx900_lpc_misc_stuff(dev); + dump_pci_device(dev); +} + +static struct device_operations vx900_lpc_ops = { + .read_resources = pci_dev_read_resources, + .set_resources = pci_dev_set_resources, + .enable_resources = pci_dev_enable_resources, + .init = vx900_lpc_init, + .scan_bus = scan_static_bus, +}; + +static const struct pci_driver lpc_driver __pci_driver = { + .ops = &vx900_lpc_ops, + .vendor = PCI_VENDOR_ID_VIA, + .device = PCI_DEVICE_ID_VIA_VX900_LPC, +}; + +#if CONFIG_PIRQ_ROUTE +void pirq_assign_irqs(const u8 * pirq) +{ + device_t lpc; + + lpc = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_LPC, 0); + + /* Take care of INTA -> INTD */ + pci_mod_config8(lpc, 0x55, (0xf << 4), pirq[0] << 4); + pci_write_config8(lpc, 0x56, pirq[1] | (pirq[2] << 4)); + pci_write_config8(lpc, 0x57, pirq[3] << 4); + + /* Enable INTE -> INTH to be on separate IRQs */ + pci_mod_config8(lpc, 0x46, 0, 1 << 4); + /* Now do INTE -> INTH */ + pci_write_config8(lpc, 0x44, pirq[4] | (pirq[5] << 4)); + pci_write_config8(lpc, 0x45, pirq[6] | (pirq[7] << 4)); +} +#endif diff --git a/src/northbridge/via/vx900/northbridge.c b/src/northbridge/via/vx900/northbridge.c new file mode 100644 index 0000000000..fe42493dbb --- /dev/null +++ b/src/northbridge/via/vx900/northbridge.c @@ -0,0 +1,341 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2012 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +#include "vx900.h" +#include "chip.h" + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#define RAM_4GB (((u64)1) << 32) + +/** + * @file northbridge.c + * + * STATUS: Pretty good + * One thing that needs to be thoroughly tested is the remap above 4G logic. + * Unfortunately, while we cannot initialize odd ranks, our testing + * possibilities are somewhat limited. A point of failure that is not covered is + * when the amount of RAM and PCI config space added up exceeds 8GB. The + * remapping mechanism will overflow, the effects of which are unknown. + */ + +void hard_reset(void) +{ + outb((1 << 2) | (1 << 1), 0xcf9); +} + +static u64 vx900_get_top_of_ram(device_t mcu) +{ + u16 reg16; + /* The last valid DRAM address is computed by the MCU + * One issue might be if we have a hole in the rank mappings, so that + * virtual ranks are not mapped successively in the linear address space + * (Ex: rank 0 mapped 0-1G, rank 1 mapped 2G-3G) + * We don't do this awkward mapping in RAM init, so we don't worry about + * it here, but it is something to keep in mind if having RAM issues */ + reg16 = pci_read_config16(mcu, 0x88) & 0x07ff; + return (u64) reg16 << 24; +} + +/* + * This guy is meant to go away, but for now, leave it in so that we can see + * if the logic to remap RAM above 4G has errors. + */ +static void killme_debug_4g_remap_reg(u32 reg32) +{ + if (reg32 & (1 << 0)) + print_debug("Mem remapping enabled\n"); + u64 remapstart = (reg32 >> 2) & 0x3ff; + u64 remapend = (reg32 >> 14) & 0x3ff; + remapstart <<= 26; + remapend <<= 26; + printk(BIOS_DEBUG, "Remapstart %lld(MB) \n", remapstart >> 20); + printk(BIOS_DEBUG, "Remapend %lld(MB) \n", remapend >> 20); +} + +/** + * \brief Remap low memory colliding with PCI MMIO space, above 4G + * + * @param mcu The memory controller + * @param tolm Top of low memory. + * + * @return The new top of memory. + */ +static u64 vx900_remap_above_4g(device_t mcu, u32 tolm) +{ + size_t i; + u8 reg8, start8, end8, start, end; + u16 reg16; + u32 reg32; + u64 tor, newtor, chunk; + + /* + * The remapping mechanism works like this: + * + * - Choose the top of low memory. + * This becomes the "remap from" + * - Choose a chunk above 4G where to remap. + * This becomes "remap to" + * - Choose a chunk above 4G where to end the remapping. + * This becomes "remap until" + * + * This remaps a "chunk" of memory where we want to. + * sizeof(chunk) = until - to; + * + * Therefore the memory region from "from" to " from + sizeof(chunk)" + * becomes accessible at "to" to "until" + */ + if (tolm >= vx900_get_top_of_ram(mcu)) { + print_debug("Nothing to remap\n"); + } + + /* This is how the Vendor BIOS. Keep it for comparison for now */ + killme_debug_4g_remap_reg(0x00180141); + /* We can remap with a granularity of 64MB, so align tolm */ + tolm &= ~((64 * MiB) - 1); + + /* The "start remapping from where ?" register */ + reg16 = ((tolm >> 20) & 0xfff) << 4; + pci_mod_config16(mcu, 0x84, 0xfff0, reg16); + + /* Find the chunk size */ + tor = vx900_get_top_of_ram(mcu); + printk(BIOS_DEBUG, "Top of RAM %lldMB\n", tor >> 20); + + if (tor < RAM_4GB) { + chunk = tor - tolm; + newtor = RAM_4GB + chunk; + } else { + chunk = (RAM_4GB - tolm); + newtor = tor + chunk; + } + printk(BIOS_DEBUG, "New top of RAM %lldMB\n", newtor >> 20); + + reg8 = tolm >> 26; + /* Which rank does the PCI TOLM fall on? */ + for (i = 0; i < VX900_MAX_MEM_RANKS; i++) { + end8 = pci_read_config8(mcu, 0x40 + i); + if (reg8 > end8) + continue; + start8 = pci_read_config8(mcu, 0x48 + i); + if (reg8 <= start8) + continue; + printk(BIOS_DEBUG, "Address %x falls on rank %ld\n", tolm, i); + break; + } + + for (; i < VX900_MAX_MEM_RANKS; i++) { + start = pci_read_config8(mcu, 0x48 + i); + end = pci_read_config8(mcu, 0x40 + i); + + if (end == 0) { + printk(BIOS_DEBUG, "Huh? rank %ld empty?\n", i); + continue; + } + + if (end < (tolm >> 26)) { + printk(BIOS_DEBUG, "Huh? rank %ld don't need remap?\n", + i); + continue; + } + + printk(BIOS_DEBUG, "Physical rank %u is mapped to\n" + " Start address: 0x%.10llx (%dMB)\n" + " End address: 0x%.10llx (%dMB)\n", + (int)i, + ((u64) start << 26), (start << (26 - 20)), + ((u64) end << 26), (end << (26 - 20))); + + if (end < (RAM_4GB >> 26)) + end = (RAM_4GB >> 26); + + if (end >= (tolm >> 26)) + end += chunk >> 26; + + if (start > (tolm >> 26)) + start += chunk >> 26; + + pci_write_config8(mcu, 0x48 + i, start); + pci_write_config8(mcu, 0x40 + i, end); + + printk(BIOS_DEBUG, "ReMapped Physical rank %u, to\n" + " Start address: 0x%.10llx (%dMB)\n" + " End address: 0x%.10llx (%dMB)\n", + (int)i, + ((u64) start << 26), (start << (26 - 20)), + ((u64) end << 26), (end << (26 - 20))); + } + + /* The "remap to where?" register */ + reg32 = ((max(tor, RAM_4GB) >> 26) & 0x3ff) << 2; + /* The "remap until where?" register */ + reg32 |= ((newtor >> 26) & 0x3ff) << 14; + /* Now enable the goodies */ + reg32 |= (1 << 0); + pci_write_config32(mcu, 0xf8, reg32); + printk(BIOS_DEBUG, "Wrote remap map %x\n", reg32); + killme_debug_4g_remap_reg(reg32); + + printk(BIOS_DEBUG, "New top of memory is at %lldMB\n", newtor >> 20); + return newtor; +} + +static void vx900_set_resources(device_t dev) +{ + u32 pci_tolm, tomk, vx900_tolm, full_tolmk, fbufk, tolmk; + + print_debug("========================================" + "========================================\n"); + print_debug("============= VX900 memory sizing & Co. " + "========================================\n"); + print_debug("========================================" + "========================================\n"); + + int idx = 10; + const device_t mcu = dev_find_device(PCI_VENDOR_ID_VIA, + PCI_DEVICE_ID_VIA_VX900_MEMCTRL, + 0); + if (!mcu) { + die("Something is terribly wrong.\n" + " We tried locating the MCU on the PCI bus, " + "but couldn't find it. Halting.\n"); + } + + /* How much low adrress space do we have? */ + pci_tolm = find_pci_tolm(dev->link_list); + printk(BIOS_SPEW, "Found PCI tolm at %.8x\n", pci_tolm); + printk(BIOS_SPEW, "Found PCI tolm at %dMB\n", pci_tolm >> 20); + + /* Figure out the total amount of RAM */ + tomk = vx900_get_top_of_ram(mcu) >> 10; + printk(BIOS_SPEW, "Found top of memory at %dMB\n", tomk >> 10); + + /* Do the same for top of low RAM */ + vx900_tolm = (pci_read_config16(mcu, 0x84) & 0xfff0) >> 4; + full_tolmk = vx900_tolm << (20 - 10); + /* Remap above 4G if needed */ + full_tolmk = min(full_tolmk, pci_tolm >> 10); + printk(BIOS_SPEW, "Found top of low memory at %dMB\n", + full_tolmk >> 10); + + /* What about the framebuffer for the integrated GPU? */ + fbufk = chrome9hd_fb_size() >> 10; + printk(BIOS_SPEW, "Integrated graphics buffer: %dMB\n", fbufk >> 10); + + /* Can't use the framebuffer as system RAM, sorry */ + tolmk = min(full_tolmk, tomk); + tolmk -= fbufk; + ram_resource(dev, idx++, 0, 640); + printk(BIOS_SPEW, "System ram left: %dMB\n", tolmk >> 10); + /* FIXME: how can we avoid leaving this hole? + * Leave a hole for VGA, 0xa0000 - 0xc0000 ?? */ + /* TODO: VGA Memory hole can be disabled in SNMIC. Upper 64k of ROM seem + * to be always mapped to the top of 1M, but this can be overcome with + * some smart positive/subtractive resource decoding */ + ram_resource(dev, idx++, 768, (tolmk - 768)); + uma_memory_size = fbufk << 10; + uma_memory_base = tolmk << 10; + + printk(BIOS_DEBUG, "UMA @ %lldMB + %lldMB\n", uma_memory_base >> 20, + uma_memory_size >> 20); + /* FIXME: How do we handle remapping above 4G? */ + u64 tor = vx900_remap_above_4g(mcu, pci_tolm); + ram_resource(dev, idx++, RAM_4GB >> 10, (tor - RAM_4GB) >> 10); + + /* Leave some space for ACPI, PIRQ and MP tables */ + high_tables_base = (tolmk << 10) - HIGH_MEMORY_SIZE; + high_tables_size = HIGH_MEMORY_SIZE; + printk(BIOS_DEBUG, "high_tables_base: %08llx, size %lld\n", + high_tables_base, high_tables_size); + + print_debug("======================================================\n"); + assign_resources(dev->link_list); +} + +static void vx900_read_resources(device_t dev) +{ + /* Our fixed resources start at 0 */ + int idx = 0; + /* Reserve our ROM mapped space */ + struct resource *res; + res = new_resource(dev, idx++); + res->size = CONFIG_ROM_SIZE; + res->base = 0xffffffff - (res->size - 1); + res->flags = IORESOURCE_MEM | IORESOURCE_ASSIGNED | IORESOURCE_FIXED; + + /* Now do the same for our MMCONF + * We always run with MMCONF enabled. We need to access the extended + * config space when configuring PCI-Express links */ + res = new_resource(dev, idx++); + res->size = 256 * MiB; + res->base = CONFIG_MMCONF_BASE_ADDRESS; + res->flags = IORESOURCE_MEM | IORESOURCE_ASSIGNED | IORESOURCE_FIXED; + + pci_domain_read_resources(dev); +} + +static struct device_operations pci_domain_ops = { + .read_resources = vx900_read_resources, + .set_resources = vx900_set_resources, + .enable_resources = NULL, + .init = NULL, + .scan_bus = pci_domain_scan_bus, + /* We always run with MMCONF enabled. */ + .ops_pci_bus = &pci_ops_mmconf, +}; + +static void cpu_bus_init(device_t dev) +{ + initialize_cpus(dev->link_list); +} + +static void cpu_bus_noop(device_t dev) +{ +} + +static struct device_operations cpu_bus_ops = { + .read_resources = cpu_bus_noop, + .set_resources = cpu_bus_noop, + .enable_resources = cpu_bus_noop, + .init = cpu_bus_init, + .scan_bus = 0, +}; + +static void enable_dev(device_t dev) +{ + /* Set the operations if it is a special bus type */ + if (dev->path.type == DEVICE_PATH_DOMAIN) { + dev->ops = &pci_domain_ops; + } else if (dev->path.type == DEVICE_PATH_CPU_CLUSTER) { + dev->ops = &cpu_bus_ops; + } +} + +struct chip_operations northbridge_via_vx900_ops = { + CHIP_NAME("VIA VX900 Chipset") + .enable_dev = enable_dev, +}; diff --git a/src/northbridge/via/vx900/pcie.c b/src/northbridge/via/vx900/pcie.c new file mode 100644 index 0000000000..109e5c9ffe --- /dev/null +++ b/src/northbridge/via/vx900/pcie.c @@ -0,0 +1,123 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2013 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +#include +#include +#include +#include +#include + +#include "vx900.h" + +/** + * @file pcie.c + * + * STATUS: + * We do part of the sequence to initialize the PCIE link. The problem is that + * the reset signal for each slot is connected to a GPO pin, but We don't know + * which GPO pin. We need to figure out which GPIO pin is hooked to which slot, + * and have a mechanism to specify this per-mainboard (devicetree.cb). + * + * There is currently no timeout detection mechanism for when a link comes up. + * If the link never comes up, we hang. + */ + +static void vx900_pcie_link_init(device_t dev) +{ + u8 reg8; + u32 reg32; + + u8 fn = dev->path.pci.devfn & 0x07; + + /* Step 1 : Check for presence of PCIE device */ + reg8 = pci_read_config8(dev, 0x5a); + + if (reg8 & (1 << 6)) + printk(BIOS_DEBUG, "Card detected in PEX%i\n", fn); + else + return; + + /* Step 2: Wait for device to enter L0 state */ + /* FIXME: implement timeout detection */ + while (0x8a != pci_read_config8(dev, 0x1c3)) ; + + /* Step 3: Clear PCIe error status, then check for failures */ + pci_write_config32(dev, 0x104, 0xffffffff); + reg32 = pci_read_config32(dev, 0x104); + if (0 != reg32) { + printk(BIOS_DEBUG, "PEX init error. flags 0x%.8x\n", reg32); + return; + } + + pci_write_config32(dev, 0x110, 0xffffffff); + reg32 = pci_read_config32(dev, 0x110); + if (0 != reg32) + printk(BIOS_DEBUG, "PEX errors. flags 0x%.8x\n", reg32); + + pci_write_config8(dev, 0xa4, 0xff); + if (pci_read_config8(dev, 0x4a) & (1 << 3)) + print_debug("Unsupported request detected.\n"); + + pci_write_config8(dev, 0x15a, 0xff); + if (pci_read_config8(dev, 0x15a) & (1 << 1)) + print_debug("Negotiation pending.\n"); + + /* Step 4: Read vendor ID */ + /* FIXME: Do we want to run through the whole sequence and delay boot + * by several seconds if the device does not respond properly the first + * time? */ +} + +static void vx900_pex_dev_set_resources(device_t dev) +{ + assign_resources(dev->link_list); +} + +static void vx900_pex_init(device_t dev) +{ + /* FIXME: For some reason, PEX0 hangs on init. Find issue, fix it. */ + if ((dev->path.pci.devfn & 0x7) == 0) + return; + + vx900_pcie_link_init(dev); +} + +static struct device_operations vx900_pex_ops = { + .read_resources = pci_bus_read_resources, + .set_resources = vx900_pex_dev_set_resources, + .enable_resources = pci_bus_enable_resources, + .init = vx900_pex_init, + .scan_bus = pciexp_scan_bridge, + .reset_bus = pci_bus_reset, +}; + +static const unsigned short pci_device_ids[] = { + PCI_DEVICE_ID_VIA_VX900_PEX1, + PCI_DEVICE_ID_VIA_VX900_PEX2, + PCI_DEVICE_ID_VIA_VX900_PEX3, + PCI_DEVICE_ID_VIA_VX900_PEX4, + 0, +}; + +static const struct pci_driver pex_driver __pci_driver = { + .ops = &vx900_pex_ops, + .vendor = PCI_VENDOR_ID_VIA, + .devices = pci_device_ids, + +}; diff --git a/src/northbridge/via/vx900/sata.c b/src/northbridge/via/vx900/sata.c new file mode 100644 index 0000000000..63295e56c3 --- /dev/null +++ b/src/northbridge/via/vx900/sata.c @@ -0,0 +1,280 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2012 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +#include +#include +#include + +#include "vx900.h" + +/** + * @file sata.c + * + * STATUS: Pretty good + * The only issue is the SATA EPHY configuration. We do not know if it is board + * specific or not. Otherwise, the SATA controller works without issues. + */ + +static void vx900_print_sata_errors(u32 flags) +{ + /* Status flags */ + printk(BIOS_DEBUG, "\tPhyRdy %s\n", + (flags & (1 << 16)) ? "changed" : "not changed"); + printk(BIOS_DEBUG, "\tCOMWAKE %s\n", + (flags & (1 << 16)) ? "detected" : "not detected"); + printk(BIOS_DEBUG, "\tExchange as determined by COMINIT %s\n", + (flags & (1 << 26)) ? "occured" : "not occured"); + printk(BIOS_DEBUG, "\tPort selector presence %s\n", + (flags & (1 << 27)) ? "detected" : "not detected"); + /* Errors */ + if (flags & (1 << 0)) + print_debug("\tRecovered data integrity ERROR\n"); + if (flags & (1 << 1)) + print_debug("\tRecovered data communication ERROR\n"); + if (flags & (1 << 8)) + print_debug("\tNon-recovered Transient Data Integrity ERROR\n"); + if (flags & (1 << 9)) + print_debug("\tNon-recovered Persistent Communication or" + "\tData Integrity ERROR\n"); + if (flags & (1 << 10)) + print_debug("\tProtocol ERROR\n"); + if (flags & (1 << 11)) + print_debug("\tInternal ERROR\n"); + if (flags & (1 << 17)) + print_debug("\tPHY Internal ERROR\n"); + if (flags & (1 << 19)) + print_debug("\t10B to 8B Decode ERROR\n"); + if (flags & (1 << 20)) + print_debug("\tDisparity ERROR\n"); + if (flags & (1 << 21)) + print_debug("\tCRC ERROR\n"); + if (flags & (1 << 22)) + print_debug("\tHandshake ERROR\n"); + if (flags & (1 << 23)) + print_debug("\tLink Sequence ERROR\n"); + if (flags & (1 << 24)) + print_debug("\tTransport State Transition ERROR\n"); + if (flags & (1 << 25)) + print_debug("\tUNRECOGNIZED FIS type\n"); +} + +static void vx900_dbg_sata_errors(device_t dev) +{ + /* Port 0 */ + if (pci_read_config8(dev, 0xa0) & (1 << 0)) { + print_debug("Device detected in SATA port 0.\n"); + u32 flags = pci_read_config32(dev, 0xa8); + vx900_print_sata_errors(flags); + }; + /* Port 1 */ + if (pci_read_config8(dev, 0xa1) & (1 << 0)) { + print_debug("Device detected in SATA port 1.\n"); + u32 flags = pci_read_config32(dev, 0xac); + vx900_print_sata_errors(flags); + }; +} + +typedef u8 sata_phy_config[64]; + +static sata_phy_config reference_ephy = { + 0x80, 0xb8, 0xf0, 0xfe, 0x40, 0x7e, 0xf6, 0xdd, + 0x1a, 0x22, 0xa0, 0x10, 0x02, 0xa9, 0x7c, 0x7e, + 0x00, 0x00, 0x00, 0x00, 0x40, 0x30, 0x84, 0x8c, + 0x75, 0x26, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x20, 0x40, 0xd0, 0x41, 0x40, 0x00, 0x00, 0x08, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x20, 0x40, 0x50, 0x41, 0x40, 0x00, 0x00, 0x00, + 0x00, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, +}; + +static u32 sata_phy_read32(device_t dev, u8 index) +{ + /* The SATA PHY control registers are accessed by a funny index/value + * scheme. Each byte (0,1,2,3) has its own 4-bit index */ + index = (index >> 2) & 0xf; + u16 i16 = index | (index << 4) | (index << 8) | (index << 12); + /* The index */ + pci_write_config16(dev, 0x68, i16); + /* The value */ + return pci_read_config32(dev, 0x64); +} + +static void sata_phy_write32(device_t dev, u8 index, u32 val) +{ + /* The SATA PHY control registers are accessed by a funny index/value + * scheme. Each byte (0,1,2,3) has its own 4-bit index */ + index = (index >> 2) & 0xf; + u16 i16 = index | (index << 4) | (index << 8) | (index << 12); + /* The index */ + pci_write_config16(dev, 0x68, i16); + /* The value */ + pci_write_config32(dev, 0x64, val); +} + +static void vx900_sata_read_phy_config(device_t dev, sata_phy_config cfg) +{ + size_t i; + u32 *data = (u32 *) cfg; + for (i = 0; i < (sizeof(sata_phy_config)) >> 2; i++) { + data[i] = sata_phy_read32(dev, i << 2); + } +} + +static void vx900_sata_write_phy_config(device_t dev, sata_phy_config cfg) +{ + size_t i; + u32 *data = (u32 *) cfg; + for (i = 0; i < (sizeof(sata_phy_config)) >> 2; i++) { + sata_phy_write32(dev, i << 2, data[i]); + } +} + +static void vx900_sata_dump_phy_config(sata_phy_config cfg) +{ + print_debug("SATA PHY config:\n"); + int i; + for (i = 0; i < sizeof(sata_phy_config); i++) { + unsigned char val; + if ((i & 0x0f) == 0) { + print_debug_hex8(i); + print_debug_char(':'); + } + val = cfg[i]; + if ((i & 7) == 0) + print_debug(" |"); + print_debug_char(' '); + print_debug_hex8(val); + if ((i & 0x0f) == 0x0f) { + print_debug("\n"); + } + } +} + +/** + * \brief VX900: Place the onboard SATA controller in Native IDE mode + * + * AHCI mode requires a sub-class of 0x06, and Interface of 0x0 + * SATA mode requires a sub-class of 0x06, and Interface of 0x00 + * Unfortunately, setting the class to SATA, will prevent us from modyfing the + * interface register to an AHCI/SATA compliant value. Thus, payloads or OS may + * not properly identify this as a SATA controller. + * We could set the class code to 0x04, which would cause the interface register + * to become 0x00, which represents a RAID controller. Unfortunately, when we do + * this, SeaBIOS will skip this as a storage device, and we will not be able to + * boot. + * Our only option is to operate in IDE mode. We choose native IDE so that we + * can freely assign an IRQ, and are not forced to use IRQ14 + */ +static void vx900_native_ide_mode(device_t dev) +{ + /* Disable subclass write protect */ + pci_mod_config8(dev, 0x45, 1 << 7, 0); + /* Change the device class to IDE */ + pci_write_config16(dev, PCI_CLASS_DEVICE, PCI_CLASS_STORAGE_IDE); + /* Re-enable subclass write protect */ + pci_mod_config8(dev, 0x45, 0, 1 << 7); + /* Put it in native IDE mode */ + pci_write_config8(dev, PCI_CLASS_PROG, 0x8f); +} + +static void vx900_sata_init(device_t dev) +{ + /* Enable SATA primary channel IO access */ + pci_mod_config8(dev, 0x40, 0, 1 << 1); + /* Just SATA, so it makes sense to be in native SATA mode */ + vx900_native_ide_mode(dev); + + /* TP Layer Idle at least 20us before the Following Command */ + pci_mod_config8(dev, 0x53, 0, 1 << 7); + /* Resend COMRESET When Recovering SATA Gen2 Device Error */ + pci_mod_config8(dev, 0x62, 1 << 1, 1 << 7); + + /* Fix "PMP Device Can’t Detect HDD Normally" (VIA Porting Guide) + * SATA device detection will not work unless we clear these bits. + * Without doing this, SeaBIOS (and potentially other payloads) will + * timeout when detecting SATA devices */ + pci_mod_config8(dev, 0x89, (1 << 3) | (1 << 6), 0); + + /* 12.7 Two Software Resets May Affect the System + * When the software does the second reset before the first reset + * finishes, it may cause the system hang. It would do one software + * reset and check the BSY bit of one port only, and the BSY bit of + * other port would be 1, then it does another software reset + * immediately and causes the system hang. + * This is because the first software reset doesn’t finish, and the + * state machine of the host controller conflicts, it can’t finish the + * second one anymore. The BSY bit of slave port would be always 1 after + * the second software reset issues. BIOS should set the following + * bit to avoid this issue. */ + pci_mod_config8(dev, 0x80, 0, 1 << 6); + + /* We need to set the EPHY values before doing anything with the link */ + sata_phy_config ephy; + vx900_sata_read_phy_config(dev, ephy); + if (1) { + vx900_sata_dump_phy_config(ephy); + vx900_sata_write_phy_config(dev, reference_ephy); + } else { + /* Enable TX and RX driving resistance */ + /* TX - 50 Ohm */ + ephy[1] &= ~(0x1f << 3); + ephy[1] |= (1 << 7) | (8 << 3); + /* RX - 50 Ohm */ + ephy[2] &= ~(0x1f << 3); + ephy[2] |= (1 << 7) | (8 << 3); + vx900_sata_write_phy_config(dev, ephy); + } + + vx900_sata_read_phy_config(dev, ephy); + vx900_sata_dump_phy_config(ephy); + + /* Clear error flags */ + pci_write_config32(dev, 0xa8, 0xffffffff); + pci_write_config32(dev, 0xac, 0xffffffff); + + /* Start OOB link negotiation sequence */ + pci_mod_config8(dev, 0xb9, 0, 3 << 4); + + /* FIXME: From now on, we are just doing DEBUG stuff + * Wait until PHY communication is enabled */ + u32 wloops = 0; + while (!(pci_read_config8(dev, 0xa0) & (1 << 1))) + wloops++; + printk(BIOS_SPEW, "SATA wait loops: %u\n", wloops); + + vx900_dbg_sata_errors(dev); +} + +static void vx900_sata_read_resources(device_t dev) +{ + pci_dev_read_resources(dev); +} + +static struct device_operations vga_operations = { + .read_resources = vx900_sata_read_resources, + .set_resources = pci_dev_set_resources, + .enable_resources = pci_dev_enable_resources, + .init = vx900_sata_init, +}; + +static const struct pci_driver chrome9hd_driver __pci_driver = { + .ops = &vga_operations, + .vendor = PCI_VENDOR_ID_VIA, + .device = PCI_DEVICE_ID_VIA_VX900_SATA, +}; diff --git a/src/northbridge/via/vx900/traf_ctrl.c b/src/northbridge/via/vx900/traf_ctrl.c new file mode 100644 index 0000000000..2e73ea439e --- /dev/null +++ b/src/northbridge/via/vx900/traf_ctrl.c @@ -0,0 +1,145 @@ +/* + * This file is part of the coreboot project. + * + * Copyright (C) 2012 Alexandru Gagniuc + * + * 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, either version 2 of the License, or + * (at your option) any later version. + * + * 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, see . + */ + +#include +#include +#include +#include +#include +#include + +#include "vx900.h" +#include "chip.h" + +/** + * @file traf_ctrl.c + * + * STATUS: + * The same issues with the IOAPIC pointe in lpc.c also apply here. + * + * We need to check if the current PCIE lane configuration mechanism is sane. + */ + +/** + * \brief VX900: Set up the north module IOAPIC (for PCIE and VGA) + * + * Enable the IOAPIC in the south module, and properly set it up. + * \n + * This is the hardware specific initialization for the IOAPIC, and complements + * the setup done by the generic IOAPIC driver. In order for the IOAPIC to work + * properly, it _must_ be declared in devicetree.cb . + * \n + * We are assuming this is called before the drivers/generic/ioapic code, + * which should be the case if devicetree.cb is set up properly. + */ +static void vx900_north_ioapic_setup(device_t dev) +{ + u8 base_val; + device_t ioapic; + ioapic_config_t *config; + /* Find the IOAPIC, and make sure it's set up correctly in devicetree.cb + * If it's not, then the generic ioapic driver will not set it up + * correctly, and the MP table will not be correctly generated */ + for (ioapic = dev->next; ioapic; ioapic = ioapic->next) { + if (ioapic->path.type == DEVICE_PATH_IOAPIC) + break; + } + /* You did put an IOAPIC in devicetree.cb, didn't you? */ + if (ioapic == 0) { + /* We don't have enough info to set up the IOAPIC */ + printk(BIOS_ERR, "ERROR: North module IOAPIC not found. " + "Check your devicetree.cb\n"); + return; + } + /* Found our IOAPIC, and it should not carry ISA interrupts */ + config = (ioapic_config_t *) ioapic->chip_info; + if (config->have_isa_interrupts) { + /* Umh, is this the right IOAPIC ? */ + printk(BIOS_ERR, "ERROR: North module IOAPIC should not carry " + "ISA interrupts.\n" "Check your devicetree.cb\n"); + printk(BIOS_ERR, "Will not initialize this IOAPIC.\n"); + return; + } + /* The base address of this IOAPIC _must_ + * be between 0xfec00000 and 0xfecfff00 + * be 256-byte aligned + */ + if ((config->base < 0xfec0000 || config->base > 0xfecfff00) + || ((config->base & 0xff) != 0)) { + printk(BIOS_ERR, "ERROR: North module IOAPIC base should be " + "between 0xfec00000 and 0xfecfff00\n" + "and must be aligned to a 256-byte boundary, " + "but we found it at 0x%.8x\n", config->base); + return; + } + + printk(BIOS_DEBUG, "VX900 TRAF_CTR: Setting up the north module IOAPIC " + "at 0%.8x\n", config->base); + + /* First register of the IOAPIC base */ + base_val = (config->base >> 8) & 0xff; + pci_write_config8(dev, 0x41, base_val); + /* Second register of the base. + * Bit[7] also enables the IOAPIC and bit[5] enables MSI cycles */ + base_val = (config->base >> 16) & 0xf; + pci_mod_config8(dev, 0x40, 0, base_val | (1 << 7) | (1 << 5)); +} + +/* + * Configures the PCI-express ports + * + * FIXME: triple-quadruple-check this + */ +static void vx900_pex_link_setup(device_t dev) +{ + u8 reg8; + struct northbridge_via_vx900_config *nb = (void *)dev->chip_info; + + reg8 = pci_read_config8(dev, 0xb0); + reg8 &= ~((1 << 7) | (1 << 3)); + + if (nb->assign_pex_to_dp) + reg8 |= (1 << 7); + + if (!nb->pcie_port1_2_lane_wide) + reg8 |= (1 << 3); + + pci_write_config8(dev, 0xb0, reg8); +} + +static void vx900_traf_ctr_init(device_t dev) +{ + vx900_north_ioapic_setup(dev); + vx900_pex_link_setup(dev); +} + +static struct device_operations traf_ctrl_ops = { + .read_resources = pci_dev_read_resources, + .set_resources = pci_dev_set_resources, + .enable_resources = pci_dev_enable_resources, + .init = vx900_traf_ctr_init, + /* Need this here, or the IOAPIC driver won't be called */ + .scan_bus = scan_static_bus, +}; + +static const struct pci_driver traf_ctrl_driver __pci_driver = { + .ops = &traf_ctrl_ops, + .vendor = PCI_VENDOR_ID_VIA, + .device = PCI_DEVICE_ID_VIA_VX900_TRAF, +}; diff --git a/src/northbridge/via/vx900/vx900.h b/src/northbridge/via/vx900/vx900.h index 52c895c598..b4a77c5b6b 100644 --- a/src/northbridge/via/vx900/vx900.h +++ b/src/northbridge/via/vx900/vx900.h @@ -42,6 +42,7 @@ #include u32 chrome9hd_fb_size(void); +u8 vx900_int15_get_5f18_bl(void); /* We use these throughout the code. They really belong in a generic part of * coreboot, but until bureaucracy gets them there, we still need them */ -- cgit v1.2.3