October 2004 - coreboot - mail.coreboot.org

RE: Makefile changes for symbols
by Ronald G. Minnich Oct. 22, 2004

Oct. 22, 2004

On Fri, 22 Oct 2004, Dave Aubin wrote: > I don't know if this is your problem or not, but I have seen > Issues when using #if 0 or #ifdef <something no set> causing > Alignment issues. What you co do is run the file through the > Pre-procesor to generate output as what the compiler sees and then > Use this as your source. Windows .net actuality is the one I've > Seen this alignment issue with. That might be it. I think it's an actual thing a compiler supports or does not. The Xen guy was horrified that 8c didn't just support it ... "how do you write drivers" etc. Here's where I put in some obligatory old-fart reminiscence about V6 C and how it laid out structs but I'm too tired to do it. ron

1 0

RE: Freebios2 recovery progress... pci64 mem optimization
by YhLu Oct. 22, 2004

Oct. 22, 2004

If no agp, Still need gart? YH PCI: 00:18.3 94 <- [0x00f8000000 - 0x00fbffffff] mem <gart> PCI: 00:19.3 94 <- [0x00f8000000 - 0x00fbffffff] mem <gart>

3 3

RE: Freebios2 recovery progress...
by YhLu Oct. 22, 2004

Oct. 22, 2004

I agree that all onboard device should use the same subsystem id as MB's. Regards YH -----Original Message----- From: ebiederman(a)lnxi.com [mailto:ebiederman@lnxi.com] Sent: Thursday, October 21, 2004 10:46 AM To: YhLu; Li-Ta Lo Cc: LinuxBIOS Subject: Re: Freebios2 recovery progress... Li-Ta Lo <ollie(a)lanl.gov> writes: > How did you do that ? Where do you put those IDs ? YhLu <YhLu(a)tyan.com> writes: > You changes all mb config for AMD/Iwill/IBM....that took some time. > > If there is on board scsi, I want to set subsystem id, I need to put that in > Config.lb? Or more likely src/mainboarod/???/???/Options.lb src/devices/pci_device.c:pci_dev_enable_resources() calls dev->ops->pci_ops->set_subsystem() when the device dev->on_mainboard is true. In turn config.g sets that in static.c for all devices in the static tree. I believe those are the correct semantics. That the subsystem is essentially an identifier for the board the chips are sitting on. The values are come from: Look at MAINBOARD_PCI_SUBSYSTEM_VENDOR_ID and MAINBOARD_SUBSYSTEM_DEVICE_ID. So for the onboard scsi case you likely need to have a driver for the onboard scsi device that implements the set_subsystem method because there is not a generic way of doing that. Eric

5 9

RE: Makefile changes for symbols
by Rodney.Kittleson＠bench.com Oct. 22, 2004

Oct. 22, 2004

Yep. The format isn't compatible with SourcePoint. Rod -----Original Message----- From: Ronald G. Minnich [mailto:rminnich@lanl.gov] Sent: Friday, October 22, 2004 10:33 AM To: Kimball, Stephen Cc: Kittleson, Rodney C.; linuxbios(a)clustermatic.org; Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols On Fri, 22 Oct 2004 Stephen.Kimball(a)bench.com wrote: > Rod, > > I'm using Source Point 7.0.0. I can see the source files. I can > display source, disassembly or mixed. However, the disassembly > doesn't match up with the source code. There is a way to give Source > Point an offset when the symbols are loaded, but I can't seem to get > the right value. I know the address of _start and __protected_start > but these labels are not in linuxbios_c.o. I need to find the > location of a symbol in linuxbios_c.o to compute the offset. there is a file we create called linuxbios.map, did you check that out? ron

2 1

about i845gv ddr sdram init
by zhu shi song Oct. 22, 2004

Oct. 22, 2004

After working hard about two weeks, I can't init sdram successfully. I'll attach the raminit.inc please help me check what's wrong. I've contacted with intel, they wouldn't like to provide me red book. tks zhu _______________________________ Do you Yahoo!? Declare Yourself - Register online to vote today! http://vote.yahoo.com jmp intel_i845gv_out general_pci_clear_set_byte: //0003981B: 660FCF bswap edi movl %edi, %ebp //0003981E: BF2498 mov di,09824 ; //00039821: E951FF jmp 000039775 -------- (3) CALLDI(general_pci_read_byte) //00039824: F6D7 not bh notb %bh //00039826: 22C7 and al,bh andb %bh, %al //00039828: F6D7 not bh notb %bh //0003982A: 0AC3 or al,bl orb %bl, %al //0003982C: BF3198 mov di,09831 ; //0003982F: EB96 jmps 0000397C7 -------- (4) CALLDI(general_pci_write_byte) //00039831: 660FCF bswap edi movl %ebp, %edi //00039834: FFE7 jmp di RETDI gmch_clear_set_byte: //00039836: BA0000 mov dx,00000 ; movw $0x0000, %dx //00039839: EBE0 jmps 00003981B -------- (5) jmp general_pci_clear_set_byte ich4_clear_set_byte: //0003983B: BAF800 mov dx,000F8 ; movw $0x00F8, %dx //0003983E: EBDB jmps 00003981B -------- (6) jmp general_pci_clear_set_byte clear_byte: //00039840: 8AF8 mov bh,al movb %al, %bh //00039842: B300 mov bl,000 ; movb $0x00, %bl //00039844: EBD5 jmps 00003981B -------- (7) jmp general_pci_clear_set_byte set_byte: //00039846: 0FB6D8 movzx bx,al movzx %al, %bx //00039849: EBD0 jmps 00003981B -------- (8) jmp general_pci_clear_set_byte gmch_clear_byte: //0003984B: BA0000 mov dx,00000 ; movw $0x0000, %dx //0003984E: EBF0 jmps 000039840 -------- (5) jmp clear_byte ich4_clear_byte: //00039850: BAF800 mov dx,000F8 ; movw $0x00F8, %dx //00039853: EBEB jmps 000039840 -------- (6) jmp clear_byte gmch_set_byte: //00039855: BA0000 mov dx,00000 ;" " movw $0x0000, %dx //00039858: EBEC jmps 000039846 -------- (7) jmp set_byte ich4_pci_set_byte: //0003985A: BAF800 mov dx,000F8 ; movw $0x00F8, %dx //0003985D: EBE7 jmps 000039846 -------- (8) jmp set_byte pci_read_dword: movw $0xCF8,%dx andl $0xfffffffc, %eax outl %eax, %dx movw $0xCFC, %dx inl %dx, %eax RETSP pci_write_dword: movw $0xCF8,%dx andl $0xfffffffc,%eax outl %eax,%dx movw $0xCFC, %dx movl %ebx,%eax outl %eax, %dx RETSP pci_read_word: movw %ax, %dx bswapl %edx movw $0x0CF8, %dx andl $0xfffffffc, %eax outl %eax, %dx bswapl %edx movw %dx, %ax bswapl %edx andw $0x02, %ax addw %ax, %dx addw $0x04, %dx inw %dx, %ax RETSP pci_write_word: movw %ax, %dx bswapl %edx movw $0x0CF8, %dx andl $0xfffffffc, %eax outl %eax, %dx bswapl %edx movw %dx, %ax bswapl %edx andw $0x02, %ax addw %ax, %dx addw $0x04, %dx movw %bx, %ax outw %ax, %dx RETSP pci_read_byte: movw %ax, %dx bswapl %edx movw $0x0CF8, %dx andl $0xfffffffc, %eax outl %eax, %dx bswapl %edx movw %dx, %ax bswapl %edx andw $0x03, %ax addw %ax, %dx addw $0x04, %dx inb %dx, %al RETSP pci_write_byte: movw %ax, %dx bswapl %edx movw $0x0CF8, %dx andl $0xfffffffc, %eax outl %eax, %dx bswapl %edx movw %dx, %ax bswapl %edx andw $0x03, %ax addw %ax, %dx addw $0x04, %dx mov %bl, %al outb %al, %dx RETSP ram_3cb60: bswap %edx movw $0x0DFB, %bx mulw %bx movw $0x03E8, %bx divw %bx orw %dx, %dx je no_c incw %ax no_c: movzx %ax, %ebx movw $0x0800, %dx movw $0x0001, %ax outw %ax, %dx addb $0x08, %dl inl %dx, %eax andl $0x00FFFFFF, %eax addl %eax, %ebx subb $0x08, %dl 1: testl $0xFF000000, %ebx je test_out in_loop: inw %dx, %ax andw $0x01, %ax je in_loop outw %ax, %dx subl $0x01000000, %ebx jmp 1b test_out: addb $0x08, %dl 2: inl %dx, %eax andl $0x00FFFFFF, %eax cmpl %eax, %ebx jb 2b bswap %edx RETSP #define DEBUG_RAM_CONFIG 3 /* DDR DIMM Mode register Definitions */ #define BURST_2 (1<<0) #define BURST_4 (2<<0) #define BURST_8 (3<<0) #define BURST_SEQUENTIAL (0<<3) #define BURST_INTERLEAVED (1<<3) #define CAS_2_0 (1 << 5) #define CAS_2_5 00 #define MODE_NORM (0 << 7) #define MODE_DLL_RESET (2 << 7) #define MODE_TEST (1 << 7) #define BURST_LENGTH BURST_4 #define BURST_TYPE BURST_INTERLEAVED #define CAS_LATENCY CAS_2_5 #define MRS_VALUE (MODE_NORM | CAS_LATENCY | BURST_TYPE | BURST_LENGTH) #define EMRS_VALUE 0x000 #define MD_SHIFT 4 #define RAM_COMMAND_NONE 0x0 #define RAM_COMMAND_NOP 0x1 #define RAM_COMMAND_PRECHARGE 0x2 #define RAM_COMMAND_MRS 0x3 #define RAM_COMMAND_EMRS 0x4 #define RAM_COMMAND_CBR 0x6 #define RAM_COMMAND_NORMAL 0x7 #define RAM_CMD(command, offset) \ movl $(((offset) << (MD_SHIFT + 16))|((command << 4) & 0x70)), %ebx ; \ CALLSP(do_ram_command) #define RAM_NOP() RAM_CMD(RAM_COMMAND_NOP, 0) #define RAM_PRECHARGE() RAM_CMD(RAM_COMMAND_PRECHARGE, 0) #define RAM_CBR() RAM_CMD(RAM_COMMAND_CBR, 0) #define RAM_EMRS() RAM_CMD(RAM_COMMAND_EMRS, EMRS_VALUE) ram_cas_latency: .byte CAS_2_5, CAS_2_0 ram_mrs: /* Read the cas latency setting */ movl $0x78, %eax PCI_READ_CONFIG_BYTE /* Transform it into the form expected by SDRAM */ shrl $4, %eax andl $3, %eax movb ram_cas_latency(%eax), %al shll $(16+MD_SHIFT), %eax orl %eax, %ebx orl $((MODE_NORM | BURST_TYPE | BURST_LENGTH) << (16+MD_SHIFT)), %ebx jmp do_ram_command #define RAM_MRS(dll_reset) \ movl $((dll_reset << (8+MD_SHIFT+ 16))|((RAM_COMMAND_MRS <<4)& 0x70)), %ebx ; \ CALLSP(ram_mrs) #define RAM_NORMAL() \ movl $0x7c, %eax ;\ PCI_READ_CONFIG_BYTE ;\ andl $0x8F, %eax ;\ orb $(RAM_COMMAND_NORMAL << 4), %al ;\ movl %eax, %edx ;\ movl $0x7c, %eax ;\ PCI_WRITE_CONFIG_BYTE #define RAM_RESET_DDR_PTR() \ movl $0x88, %eax ;\ PCI_READ_CONFIG_BYTE ;\ orb $(1 << 4), %al ;\ movl %eax, %edx ;\ movl $0x88, %eax ;\ PCI_WRITE_CONFIG_BYTE ;\ movl $0x88, %eax ;\ PCI_READ_CONFIG_BYTE ;\ andb $~(1 << 4), %al ;\ movl %eax, %edx ;\ movl $0x88, %eax ;\ PCI_WRITE_CONFIG_BYTE do_ram_command: #if DEBUG_RAM_CONFIG >= 2 movl %esp, %esi movl %ebx, %edi CONSOLE_DEBUG_TX_CHAR($'P') CONSOLE_DEBUG_TX_CHAR($':') CONSOLE_DEBUG_TX_HEX8(%bl) CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') movl %edi, %ebx movl %esi, %esp #endif /* %ecx - initial address to read from */ /* Compute the offset */ movl %ebx, %ecx shrl $16, %ecx 1: /* Set the ram command */ CONSOLE_DEBUG_TX_CHAR($'Z') CONSOLE_DEBUG_TX_CHAR($':') movl $0x7c, %eax PCI_READ_CONFIG_BYTE movl %eax, %esi CONSOLE_DEBUG_TX_HEX8(%al) movl %esi, %eax andl $0x8F, %eax orb %bl, %al movl %eax, %esi CONSOLE_DEBUG_TX_HEX8(%al) CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') movl %esi, %edx movl $0x7c, %eax PCI_WRITE_CONFIG_BYTE /* Assert the command to the memory */ #if DEBUG_RAM_CONFIG >= 2 movl %esp, %esi movl %ebx, %edi CONSOLE_DEBUG_TX_CHAR($'R') CONSOLE_DEBUG_TX_CHAR($':') CONSOLE_DEBUG_TX_HEX32(%ecx) CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') movl %edi, %ebx movl %esi, %esp #endif CONSOLE_DEBUG_TX_CHAR($'S') CONSOLE_DEBUG_TX_CHAR($':') movl $0x7c, %eax PCI_READ_CONFIG_BYTE movb %al, %bh CONSOLE_DEBUG_TX_HEX8(%bh) CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') movl (%ecx), %eax CONSOLE_DEBUG_TX_CHAR($'T') CONSOLE_DEBUG_TX_CHAR($':') CONSOLE_DEBUG_TX_HEX8(%al) CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') /* Go to the next base address */ addl $0x02000000, %ecx /* Increment the counter */ incb %bh /* See if we are done */ cmpb $8, %bh jb 1b 3: /* The command has been sent to all dimms so get out */ RETSP #define ENABLE_REFRESH() \ movl $0x7c, %eax ;\ PCI_READ_CONFIG_DWORD ;\ orl $(1 << 29), %eax ;\ movl %eax, %ecx ;\ mov $0x7c, %eax ;\ PCI_WRITE_CONFIG_DWORD /* * Table: constant_register_values */ .p2align 3 constant_register_values: /* SVID - Subsystem Vendor Identification Register * 0x2c - 0x2d * [15:00] Subsytem Vendor ID (Indicates system board vendor) */ /* SID - Subsystem Identification Register * 0x2e - 0x2f * [15:00] Subsystem ID */ .long 0x2c, 0, (0x1849 << 0) | (0x2560 << 16) /* 40 - 4F */ .long 0x40, 0, (05 << 8) | (0xbc << 0 ) .long 0x44, 0, (0x41 << 0) | (0x20 << 8) | (0x10 << 16)|(0x20<<24) .long 0x48, 0, (0x04 << 0) | (0x01 << 8) .long 0x4c, 0, (0x1b << 0) | (0x08 << 8) | (0x10 << 16) /* GC - Graphic Control Register */ /* 0x52 */ .long 0x52, 0, 0x00000044 /* DRB - DRAM Row Boundary Registers * 0x60 - 0x63 * An array of 8 byte registers, which hold the ending * memory address assigned to each pair of DIMMS, in 32MB * granularity. */ /* Conservatively say each row has 32MB of ram, we will fix this up later */ // .long 0x60, 0x00000000, (0x01 << 0) | (0x02 << 8) | (0x03 << 16) | (0x04 << 24) /* Aggresively say each row has max ram = 1G, we will fix this up later */ .long 0x60, 0x00000000, (0x08 << 0) | (0x08 << 8) | (0x08 << 16) | (0x08 << 24) /* DRA - DRAM Row Attribute Register * 0x70 Row 0,1 * 0x71 Row 2,3 * [7:7] reserved * [6:4] Row Attributes for Odd numbered rows * 000 == 2KB * 001 == 4KB * 010 == 8KB * 011 == 16KB * Others == Reserved * [3:3] reserved * [2:0] Row Attributes for Even numbered rows * 000 == 2KB * 001 == 4KB * 010 == 8KB * 011 == 16KB (This page size appears broken) * Others == Reserved */ .long 0x70, 0x00000000, 0x00000002 /* DRT - DRAM Time Register * 0x78-0x7B * [31:18] Reserved * [17:15] * 000: infinite * 001: 0 * 010: 8 DRAM clocks * 011: 16 DRAM clocks * 100: 64 DRAM clocks * others: * [14:12] Reserved * [11] * 0 == 120us * 1 == reserved * [10:9] * 00 == 8 clocks * 01 == 7 clocks * 10 == 6 clocks * 11 == 5 clocks * [8:7] Reserved * [6:5] DDR CL * 00 == 2.5 * 01 == 2 * 10 == Reserved * 11 == Reserved * [4] Reserved * [3:2] RAS# to CAS# delay * 01 == 3 DRAM clocks * 10 == 2 DRAM Clocks * 11 == Reserved * [1:0] DRAM RAS# precharge * 00 == Reserved * 01 == 3 DRAM Clocks * 10 == 2 DRAM Clocks * 11 == Reserved */ #define DRT_CAS_2_5 (0<<5) #define DRT_CAS_2_0 (1<<5) #define DRT_CAS_MASK (3<<5) #if CAS_LATENCY == CAS_2_5 #define DRT_CL DRT_CAS_2_5 #elif CAS_LATENCY == CAS_2_0 #define DRT_CL DRT_CAS_2_0 #endif /* Most aggressive settings possible */ // .long 0x78, 0x00007190, (0x2b<<24)|(0x41<<16)|(0x82<<8)|0x05 .long 0x78, 0x00000000, (0x00<<24)|(0x00<<16)|(0x00<<8)|0x00 /* FIXME why was I attempting to set a reserved bit? */ /* 0x0100040f */ /* DRC - DRAM Contoller Mode Register * 0x7c * [31:30] Reserved * [29:29] Initialization Complete * 0 == Not Complete * 1 == Complete * [28:28] Dynamic Power-Down Mode Enable * 0 == disable * 1 == enable * [27:10] Reserved * [09:07] Refresh mode select * 000 == Reserved * 001 == Refresh interval 15.6 usec * 010 == Refresh interval 7.8 usec * 011 == Refresh interval 64 usec * 111 == Refresh interval 64 clocks * others == Reserved * [06:04] Mode Select (SMS) * 000 == Self Refresh Mode * 001 == NOP Command * 010 == All Banks Precharge * 011 == Mode Register Set * 100 == Extended Mode Register Set * 101 == Reserved * 110 == CBR Refresh * 111 == Normal Operation * [03:01] Reserved * [00:00] DRAM type(RO) * 0 == SDR * 1 == DDR */ // .long 0x7c, 0x0000007f, (0x20<<24)|(0xc0<<8)|(1 << 8) /* 0x80 - 8B */ .long 0x80, 0, (00<<24)|(0xaf <<16)|(0x01<<8)|(0x29) .long 0x84, 0, 0xad .long 0x88, 0, 0x01 /* FDHC - Fixed DRAM Hole Control * 0x97 * [7:7] Hole_Enable * 0 == No memory Hole * 1 == Memory Hole from 15MB to 16MB * [6:0] Reserved * * PAM - Programmable Attribute Map * 0x90 [3:0] Reserved * 0x90 [7:4] 0xF0000 - 0xFFFFF * 0x91 [1:0] 0xC0000 - 0xC3FFF * 0x91 [5:4] 0xC4000 - 0xC7FFF * 0x92 [1:0] 0xC8000 - 0xCBFFF * 0x92 [5:4] 0xCC000 - 0xCFFFF * 0x93 [1:0] 0xD0000 - 0xD3FFF * 0x93 [5:4] 0xD4000 - 0xD7FFF * 0x94 [1:0] 0xD8000 - 0xDBFFF * 0x94 [5:4] 0xDC000 - 0xDFFFF * 0x95 [1:0] 0xE0000 - 0xE3FFF * 0x95 [5:4] 0xE4000 - 0xE7FFF * 0x96 [1:0] 0xE8000 - 0xEBFFF * 0x96 [5:4] 0xEC000 - 0xEFFFF * 00 == DRAM Disabled (All Access go to memory mapped I/O space) * 01 == Read Only (Reads to DRAM, Writes to memory mapped I/O space) * 10 == Write Only (Writes to DRAM, Reads to memory mapped I/O space) * 11 == Normal (All Access go to DRAM) */ .long 0x90, 0x00000000, 0x03311110 .long 0x94, 0x00000000, 0x00110000 .long 0x98, 0x0000ffff, 0x00000445 .long 0x9c, 0xff0000ff, 0x00380a00 constant_register_values_end: /* * Routine: ram_set_registers * Arguments: none * Results: none * Trashed: %eax, %ebx, %ecx, %edx, %esi, %eflags * Effects: Do basic ram setup that does not depend on serial * presence detect information. * This sets PCI configuration registers to known good * values based on the table: * constant_register_values * Which are a triple of configuration regiser, mask, and value. * */ /* DDR RECOMP/SCOMP table */ .p2align 3 .p2align 3 ram_set_registers: movb $0xAD, %al outb %al, $0x80 #if DEBUG_RAM_CONFIG CALLSP(dumpnorth) #endif movl $constant_register_values, %ebx jmp ram_set_one_register_start ram_set_one_register: #if DEBUG_RAM_CONFIG movl %ebx, %esi CONSOLE_DEBUG_TX_CHAR($'C') CONSOLE_DEBUG_TX_CHAR($':') movl 0(%esi), %eax CONSOLE_DEBUG_TX_HEX32(%eax) CONSOLE_DEBUG_TX_CHAR($':') movl 4(%esi), %eax CONSOLE_DEBUG_TX_HEX32(%eax) CONSOLE_DEBUG_TX_CHAR($':') movl 8(%esi), %eax CONSOLE_DEBUG_TX_HEX32(%eax) CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') movl %esi, %ebx #endif /* DEBUG_RAM_CONFIG */ movl 0(%ebx), %eax /* Read the original value to preserve the reserved bits */ PCI_READ_CONFIG_DWORD movl 4(%ebx), %edx /* Reserved bits mask */ andl %edx, %eax /* Preserve only the reserved bits */ movl 8(%ebx), %ecx /* Read the new values into %ecx */ notl %edx andl %edx, %ecx /* Keep only the unreserved bits */ orl %eax, %ecx /* Put the two sets of bits together */ movl 0(%ebx), %eax /* Refetch the address to write */ PCI_WRITE_CONFIG_DWORD addl $12, %ebx ram_set_one_register_start: cmpl $constant_register_values_end, %ebx jb ram_set_one_register #if DEBUG_RAM_CONFIG CALLSP(dumpnorth) #if DEBUG_RAM_CONFIG CALLSP(dumplpc) CALLSP(dumppcibridge) #endif #endif RET_LABEL(ram_set_registers) ram_3aada: movl %esp, %ebp /* RAM NOP */ movb $0x7C, %al andl $0x000000ff, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x10, %al movl %eax, %ebx movb $0x7C, %al andl $0x000000ff, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) #if 0 movb $0x7C, %al andl $0x000000ff, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) outb %al, $0x80 #endif #if 1 movb $0xeb, %al outb %al,$0x80 #endif movl (%esi), %eax movb $0xec, %al outb %al, $0x80 movl $0x0C8, %eax CALLSP(ram_3cb60) CONSOLE_DEBUG_TX_CHAR($'R') CONSOLE_DEBUG_TX_CHAR($'1') CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') /* All Banks Pre-Charge */ xorl %esi, %esi movb $0x7c, %al andl $0x000000ff, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x20, %al movl %eax, %ebx movb $0x7c, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) movl (%esi), %eax CONSOLE_DEBUG_TX_CHAR($'R') CONSOLE_DEBUG_TX_CHAR($'2') CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') /* EMRS */ movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x40, %al mov %eax, %ebx movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) movl (%esi), %eax CONSOLE_DEBUG_TX_CHAR($'R') CONSOLE_DEBUG_TX_CHAR($'3') CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') /* MRS */ movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x30, %al movl %eax, %ebx movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) addl $0x350, %esi orl $0x800, %esi movl (%esi), %eax andl $0xFE000000, %esi /* precharge */ movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) movb $0x8F, %al orb $0x20, %al movl %eax, %ebx movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) movl (%esi), %eax /* CBR Refresh Enable */ movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x60, %al movl %eax, %ebx movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) movl (%esi), %eax movl (%esi), %eax movl (%esi), %eax movl (%esi), %eax movl (%esi), %eax movl (%esi), %eax movl (%esi), %eax movl (%esi), %eax addl $0x350, %esi movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x30, %al movl %eax, %ebx movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) movl (%esi), %eax andl $0xFE000000, %esi movl %ebp, %esp RETSP /* * Routine: sdram_read_paired_byte * Arguments: %esp return address * %bl device on the smbus to read from * %bh address on the smbus to read * Results: * zf clear * byte read in %al * On Error: * zf set * %eax trashed * * Preserved: %ebx, %esi, %edi * * Trashed: %eax, %ecx, %edx, %ebp, %esp, %eflags * Used: %eax, %ebx, %ecx, %edx, %esp, %eflags * * Effects: Reads two spd bytes from both ram channesl * and errors if they are not equal. * It then returns the equal result. */ spd_read_paired_byte: movl %esp, %ebp CALLSP(smbus_read_byte) setnz %cl movb %al, %ch /* zhu */ CONSOLE_INFO_TX_CHAR($'Z') CONSOLE_INFO_TX_CHAR($':') CONSOLE_INFO_TX_HEX8(%ch) CONSOLE_INFO_TX_CHAR($'\r') CONSOLE_INFO_TX_CHAR($'\n') movb %ch, %al /* Clear the zero flag */ testb %cl, %cl spd_verify_byte_out: movl %ebp, %esp RETSP /* * Routine: spd_verify_dimms * Arguments: none * Results: none * Preserved: none * Trashed: %eax, %ebx, %ecx, %edx, %ebp, %esi, %edi, %esp, %eflags * Used: %eax, %ebx, %ecx, %edx, %ebp, %esi, %edi, %esp, %eflags * * Effects: * - Verify all interesting spd information * matches for both dimm channels. * - Additional error checks that can be easily done * here are computed as well, so I don't need to * worry about them later. */ spd_verify_dimms: movl $(SMBUS_MEM_DEVICE_START), %ebx spd_verify_dimm: /* Verify this is DDR SDRAM */ movb $2, %bh CALLSP(spd_read_paired_byte) jz spd_verify_next_dimm cmpb $7, %al jne invalid_dimm_type /* Verify the row addresses */ movb $3, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $0x0f, %al jz spd_invalid_data /* Column addresses */ movb $4, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $0xf, %al jz spd_invalid_data /* Physical Banks */ movb $5, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data cmp $1, %al jb spd_invalid_data cmp $2, %al ja spd_invalid_data /* Module Data Width */ movb $7, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data cmpb $0, %al jne spd_invalid_data movb $6, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data cmpb $64, %al je 1f cmpb $72, %al je 1f jmp spd_unsupported_data 1: /* Cycle time at highest CAS latency CL=X */ movb $9, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data /* SDRAM type */ movb $11, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data /* Refresh Interval */ movb $12, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data /* SDRAM Width */ movb $13, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data andb $0x7f, %al cmpb $4, %al je 1f cmpb $8, %al je 1f jmp spd_unsupported_data 1: /* Back-to-Back Random Column Accesses */ movb $15, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb %al, %al jz spd_invalid_data cmpb $4, %al ja spd_unsupported_data /* Burst Lengths */ movb $16, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $(1<<2), %al jz spd_unsupported_data /* Logical Banks */ movb $17, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb %al, %al jz spd_invalid_data /* Supported CAS Latencies */ movb $18, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $(1 << 1), %al /* CL 1.5 */ jnz 1f testb $(1 << 2), %al /* CL 2.0 */ jnz 1f testb $(1 << 3), %al /* CL 2.5 */ jnz 1f jmp spd_unsupported_data 1: /* Cycle time at Cas Latency (CLX - 0.5) */ movb $23, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data /* DRAM Minimum Clock Cycle when CL is Derated by One Clock*/ movb $25, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data /* Cycle time at Cas Latency (CLX - 1.0) */ movb $26, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data /* tRP Row precharge time */ movb $27, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $0xfc, %al jz spd_invalid_data /* tRCD RAS to CAS */ movb $29, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $0xfc, %al jz spd_invalid_data /* tRAS Activate to Precharge */ movb $30, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb %al, %al jz spd_invalid_data /* Module Bank Density */ movb $31, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data testb $(1<<2), %al /* 16MB */ jnz spd_unsupported_data testb $(1<<3), %al jnz spd_unsupported_data /* 32MB */ /* Address and Command Hold Time After Clock */ movb $33, %bh CALLSP(spd_read_paired_byte) jz spd_missing_data spd_verify_next_dimm: /* go to the next DIMM */ addb $(SMBUS_MEM_DEVICE_INC), %bl /* increment the smbus device */ cmpb $SMBUS_MEM_DEVICE_END, %bl jbe spd_verify_dimm spd_verify_dimms_out: RET_LABEL(spd_verify_dimms) /* Estimate that SLOW_DOWN_IO takes about 50&76us*/ /* delay for 200us */ #define DO_DELAY \ movl $16, %edi ; \ 1: SLOW_DOWN_IO ; \ decl %edi ; \ jnz 1b #define EXTRA_DELAY DO_DELAY #define ERRFUNC(x, str) \ .section ".rom.data" ;\ x##_str: ;\ .string str ;\ .ascii str ;\ .previous ;\ x: ;\ movl $x##_str, %esi ;\ jmp mem_err ERRFUNC(invalid_dimm_type, "Invalid dimm type") ERRFUNC(spd_missing_data, "Missing sdram spd data") ERRFUNC(spd_invalid_data, "Invalid sdram spd data") ERRFUNC(spd_unsupported_data, "Unsupported sdram spd value") ERRFUNC(unsupported_page_size, "Unsupported page size") ERRFUNC(sdram_presence_mismatch, "DIMM presence mismatch") ERRFUNC(sdram_value_mismatch, "spd data does not match") ERRFUNC(unsupported_refresh_rate, "Unsuported spd refresh rate") ERRFUNC(inconsistent_cas_latencies, "No cas latency supported by all dimms") ERRFUNC(unsupported_rcd, "Unsupported ras to cas delay") #undef ERRFUNC .section ".rom.data" mem_err_err: .string "ERROR: " mem_err_pair: .string " on dimm pair " mem_err_byte: .string " spd byte " crlf: .string "\r\n" .previous mem_err: movl %ebx, %edi CONSOLE_ERR_TX_STRING($mem_err_err) CONSOLE_ERR_TX_STRING(%esi) CONSOLE_ERR_TX_STRING($mem_err_pair) movl %edi, %ebx subb $(SMBUS_MEM_DEVICE_START), %bl CONSOLE_ERR_TX_HEX8(%bl) CONSOLE_ERR_TX_STRING($mem_err_byte) movl %edi, %ebx CONSOLE_ERR_TX_HEX8(%bh) jmp mem_stop mem_stop: CONSOLE_ERR_TX_STRING($crlf) 1: hlt /* stick here.. */ jmp 1b intel_i845gv_out: CALL_LABEL(spd_verify_dimms) CALL_LABEL(ram_set_registers) DO_DELAY EXTRA_DELAY //00039EDB: 6633F6 xor esi,esi xorl %esi, %esi //00039EDE: 32C9 xor cl,cl xorb %cl, %cl //00039EE0: B060 mov al,060 ; movb $0x60, %al //00039EE2: 6625FF000000 and eax,0000000FF ; andl $0x000000FF, %eax //00039EE8: 660D00000080 or eax,080000000 ; orl $0x80000000, %eax //00039EEE: 660FCC bswap esp //00039EF1: BCF79E mov sp,09EF7 ; //00039EF4: E9DFFA jmp 0000399D6 -------- (1) //00039EF7: 660FCC bswap esp CALLSP(pci_read_dword) 2: //00039EFA: 38C8 cmp al,cl cmpb %cl, %al //00039EFC: 7412 je 000039F10 -------- (2) je 1f //00039EFE: 668BF8 mov edi,eax movl %eax, %edi //00039F01: 660FCC bswap esp //00039F04: BC0A9F mov sp,09F0A ; //00039F07: E9D00B jmp 00003AADA -------- (3) //00039F0A: 660FCC bswap esp CALLSP(ram_3aada) //00039F0D: 668BC7 mov eax,edi movl %edi, %eax 1: //00039F10: 8AC8 mov cl,al movb %al, %cl //00039F12: 6633F6 xor esi,esi xorl %esi, %esi //00039F15: 660FB6F1 movzx esi,cl movzx %cl, %esi //00039F19: 66C1E619 shl esi,019 ; shll $0x19, %esi //00039F1D: 66C1E808 shr eax,008 ; shrl $0x08, %eax //00039F21: 6683F800 cmp eax,000 ; cmpl $0x00, %eax //00039F25: 75D3 jne 000039EFA -------- (3) jne 2b /* normal mode */ movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_read_dword) andb $0x8F, %al orb $0x70, %al movl %eax, %ebx movb $0x7C, %al andl $0x000000FF, %eax orl $0x80000000, %eax CALLSP(pci_write_dword) CONSOLE_DEBUG_TX_CHAR($'E') CONSOLE_DEBUG_TX_CHAR($'N') CONSOLE_DEBUG_TX_CHAR($'\r') CONSOLE_DEBUG_TX_CHAR($'\n') hlt

4 4

RE: Makefile changes for symbols
by Rodney.Kittleson＠bench.com Oct. 22, 2004

Oct. 22, 2004

Thanks Ron, I'm a little slow today. Steve, The linuxbios_c.map file should help you find the offset you need for importing linuxbios_c.o into SourcePoint. The map file itself cannot be imported. Rod -----Original Message----- From: Kittleson, Rodney C. Sent: Friday, October 22, 2004 10:47 AM To: 'Ronald G. Minnich'; Kimball, Stephen Cc: linuxbios(a)clustermatic.org; Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols Yep. The format isn't compatible with SourcePoint. Rod -----Original Message----- From: Ronald G. Minnich [mailto:rminnich@lanl.gov] Sent: Friday, October 22, 2004 10:33 AM To: Kimball, Stephen Cc: Kittleson, Rodney C.; linuxbios(a)clustermatic.org; Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols On Fri, 22 Oct 2004 Stephen.Kimball(a)bench.com wrote: > Rod, > > I'm using Source Point 7.0.0. I can see the source files. I can > display source, disassembly or mixed. However, the disassembly > doesn't match up with the source code. There is a way to give Source > Point an offset when the symbols are loaded, but I can't seem to get > the right value. I know the address of _start and __protected_start > but these labels are not in linuxbios_c.o. I need to find the > location of a symbol in linuxbios_c.o to compute the offset. there is a file we create called linuxbios.map, did you check that out? ron

1 0

RE: Makefile changes for symbols
by Dave Aubin Oct. 22, 2004

Oct. 22, 2004

Hi, I don't know if this is your problem or not, but I have seen Issues when using #if 0 or #ifdef <something no set> causing Alignment issues. What you co do is run the file through the Pre-procesor to generate output as what the compiler sees and then Use this as your source. Windows .net actuality is the one I've Seen this alignment issue with. That might be it. Dave -----Original Message----- From: linuxbios-admin(a)clustermatic.org [mailto:linuxbios-admin@clustermatic.org] On Behalf Of Ronald G. Minnich Sent: Friday, October 22, 2004 11:33 AM To: Stephen.Kimball(a)bench.com Cc: Rodney.Kittleson(a)bench.com; linuxbios(a)clustermatic.org; Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols On Fri, 22 Oct 2004 Stephen.Kimball(a)bench.com wrote: > Rod, > > I'm using Source Point 7.0.0. I can see the source files. I can > display source, disassembly or mixed. However, the disassembly > doesn't match up with the source code. There is a way to give Source > Point an offset when the symbols are loaded, but I can't seem to get > the right value. I know the address of _start and __protected_start > but these labels are not in linuxbios_c.o. I need to find the > location of a symbol in linuxbios_c.o to compute the offset. there is a file we create called linuxbios.map, did you check that out? ron _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios

1 0

RE: changes today.
by YhLu Oct. 22, 2004

Oct. 22, 2004

Ron/Eric/Stephan/Ollie, Please check the patch. 1. some tyan volatile messing because of ROMCC changes. 2. S2882 interrupt line assignment in mptable.c 3. exclude range for flash_rom. 4. better ht_setup_chainx() in incoherent_ht.c ---> change share bus 0 to different bus for different incoherent HT link. 5. fix one bug about io and mem allocation for multi ht links. --- in northbridge.c Mark it if used before assign final value. 6. enable amd/quartet to init multi incoherent HT link in auto.c --- not test, no MB on hand. 7. enable S2885 to init multi incoherent HT link in auto.c Regards YH

5 5

RE: Makefile changes for symbols
by Rodney.Kittleson＠bench.com Oct. 22, 2004

Oct. 22, 2004

I have an offset of 4000 for "linuxbios_c.o". "linuxbios_c" has the correct addresses and as far as I can tell "linuxbios_c" has all the symbols "linuxbios_c.o" has. I just upgraded SP to 7.0.0, source still isn't loaded. Could you give me a step by step? Thanks, Rod -----Original Message----- From: Kimball, Stephen Sent: Friday, October 22, 2004 8:04 AM To: Kittleson, Rodney C.; linuxbios(a)clustermatic.org Cc: Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols Rod, I'm using Source Point 7.0.0. I can see the source files. I can display source, disassembly or mixed. However, the disassembly doesn't match up with the source code. There is a way to give Source Point an offset when the symbols are loaded, but I can't seem to get the right value. I know the address of _start and __protected_start but these labels are not in linuxbios_c.o. I need to find the location of a symbol in linuxbios_c.o to compute the offset. Steve -----Original Message----- From: Kittleson, Rodney C. Sent: Thursday, October 21, 2004 5:42 PM To: Kimball, Stephen; linuxbios(a)clustermatic.org Cc: Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols Steve, What version of Source point are you using? The symbols are coming in but SourcePoint doesn't allow me to reference the source files. I mapped the linuxbios directory on my Win2000 system for SourcePoint to access. (I have samba running on my linux system) I just need to get SourcePoint working. Thanks, Rod -----Original Message----- From: Kimball, Stephen Sent: Wednesday, October 20, 2004 1:04 PM To: Kittleson, Rodney C.; linuxbios(a)clustermatic.org Cc: Jeff_Pitts(a)arium.com Subject: RE: Makefile changes for symbols The file under your target in normal/linuxbios_c.o seems to have all the symbols if you add "-gdwarf-2" to the $CFLAGS in the Makefile. Then when you open the first C module the SourcePoint software asks where the file exists, after you tell SourcePoint it seems to find the rest by itself. I recreated the same Linux directory structure on Windows. Thanks Rod! Steve -----Original Message----- From: Kittleson, Rodney C. Sent: Wednesday, October 20, 2004 11:47 AM To: Kimball, Stephen; linuxbios(a)clustermatic.org Subject: RE: Makefile changes for symbols Linuxbios_c, from the build directory, can be sucked into Sourcepoint to extract the symbols. This allows the disassembled instructions to display the symbols inline. It makes it easier to set breakpoints and follow the code, but it's a far cry from source level debugging. I haven't attempted to link the source with sourcepoint yet. I may try it later this week, let me know if you have any success. Rod -----Original Message----- From: Kimball, Stephen Sent: Tuesday, October 19, 2004 3:33 PM To: linuxbios(a)clustermatic.org Subject: Makefile changes for symbols Can someone give me the Makefile changes to make a LinuxBIOS with symbols? I tried adding "-gdwarf-2" to CFLAGS, but I'm not sure where the unstripped output would be. Are there any features in the code that are useful in debugging LinuxBIOS? Any defines that should change as well? Someone must have some experience with source code level debugging of LinuxBIOS with an American Arium ECM-50. Thanks. Steve Kimball Benchmark Electronics Hudson, NH _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios

1 0

Dunno if you guys have seen this or not...
by Joshua Wise Oct. 22, 2004

Oct. 22, 2004

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Dunno if you guys have seen this or not, but slashdot is reporting on a discussion on LKML about the possibility of an open-hardware video card with an open-source BIOS. http://linux.slashdot.org/article.pl?sid=04/10/22/021213&threshold=0&tid=15… joshua -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFBeKBYPn9tWOqA4LMRAmrwAKCbMsyCqWxpT8S4880lGvqA/Xc/XQCdFZYk SI0xgsnM9JLGVsgZeayqpUk= =ULzN -----END PGP SIGNATURE-----

2 1