October 2003 - U-Boot

[U-Boot-Users] I cann't initialize OR/BR registers in cpu_init.c?
by wang cheng 13 Oct '03

13 Oct '03

hello ! all I'm porting u-boot-0.4.0 to my WindRiver MDP PRO board now,some addressing specifications are listed below.I choose the fads823 as the model.I have vision ICE ii & vision CLICK to program flash and debug the code. I compile the code in the vmware linux environment.After "make myboard_config" and "make",I copy the u-boot(ELF) to the vision CLICK path in windows environment，then Convert to .BIN file, program to flash,run,use the hyperterminal as the console. Now in the cpu_init.c ,I masked all the OR/BR initialization clause,for when I released them there's nothing displayed on the hyperterminal ;but when masked them ,the program can run to the "=>"(after main_loop();)with the vision ICE II and to the relocate_code()(in the board_init_f();)without the visionICE II ,some register value are listed below in the vision CLICK BKM mode. **1.why I cann't initialize the OR/BR registers in the cpu_init.c **2.how to modify the dram initialization to pass the relocate_code() 3.can you tell me something details about the BCSR register in the FADS823 model. 4.are there any better model for WindRiver MDPRRO board ? Any suggests are very appreciated! specifications: Addressing: Total address range 4GB (on and off board) Flash Memory 1.0MB on board 16MB SIMM (4MB Std.) SRAM Memory 2MB (512K Std.) DRAM Memory 32MB SIMM (16MB Std.) EEPROM 8K >BKM>sc ****************** SIU ***************** IMMR xxxxxxxx FF000000 SIUMCR FF000000 00610000 SYPCR FF000004 FFFFFF88 SWSR FF00000E 0000 SIPEND FF000010 00000000 SIMASK FF000014 00400000 SIEL FF000018 00000000 SIVEC FF00001C 3C000000 TESR FF000020 00000000 SDCR FF000030 00004001 ****************** UPM_A ***************** UPMA_T0 FF00017C 0FF3FC00 UPMA_T1 FF00017C 1FF7FC07 UPMA_T2 FF00017C FFFFFFFF UPMA_T3 FF00017C FFFFFFFF ...... ****************** UPM_B ***************** UPMB_T0 FF00017C 0FF3CC24 UPMB_T1 FF00017C 0FF3CC04 UPMB_T2 FF00017C 0CF3CC04 UPMB_T3 FF00017C 00F3CC04 UPMB_T4 FF00017C 00F3CC00 UPMB_T5 FF00017C 37F7CC47 UPMB_T6 FF00017C FFFFFFFF UPMB_T7 FF00017C FFFFFFFF U ..... ****************** MEMC ***************** BR0 FF000100 FFF00801 OR0 FF000104 FFF00760 BR1 FF000108 00000000 OR1 FF00010C 00000000 BR2 FF000110 000000C1 OR2 FF000114 FF000800 BR3 FF000118 00000000 OR3 FF00011C 00000000 BR4 FF000120 03000081 OR4 FF000124 FBF80010 BR5 FF000128 10000401 OR5 FF00012C FFC00760 BR6 FF000130 00000000 OR6 FF000134 00000000 BR7 FF000138 00000000 OR7 FF00013C 00000000 MAR FF000164 00000000 MCR FF000168 0000001B MAMR FF000170 0C120111 MBMR FF000174 2FB20111 MSTAT FF000178 0000 MPTPR FF00017A 0800 MDR FF00017C FF0C0027 ****************** CLOCKS ***************** SCCRK FF000380 55CCAA33 SCCR FF000280 01800000 PLPRCRK FF000384 55CCAA33 PLPRCR FF000284 00400000 RSRK FF000388 55CCAA33 RSR FF000288 D0000000 ****************** SI ***************** SIMODE FF000AE0 00000000 SIGMR FF000AE4 00 SISTR FF000AE6 00 SICMR FF000AE7 00 SICR FF000AEC 00000000 SIRP FF000AF0 00000000

2 1

[U-Boot-Users] [PATCH] support for Intel IXP
by Kyle Harris 10 Oct '03

10 Oct '03

Wolfgang, Please find attached a patch for Intel CPU IXP425 and eval board IXDP425. This is applied against 0.4.0 (I tried several times to pull from CVS but failed, and finally gave up). Note: This patch does NOT include support for the PCI bridge :( Regards, Kyle.

3 8

Re: [U-Boot-Users] Ramdisk for MPC8xx
by Wolfgang Denk 10 Oct '03

10 Oct '03

Dear Raghu, in message <20031010132205.90468.qmail(a)web40406.mail.yahoo.com> you wrote: > > > Well, did you read the README ? > Yes, I did but I am not very clear about the mem map. There is some documentation around, including the HOWTO (now at http://www.denx.de/twiki/bin/view/PPCEmbedded/Introduction) and mailing list archives. > > AFAIK this code never made it into the public Linux > > kernel tres. > > I think this is the same on on the www.kernel.org site I mean: we added this to the DENX linux kernel trees, but I don't expect to find it in the public trees. > > This is why I recommended our SELF bassed ramdisk > > image. > > I tried your suggeted ramdisk image and also set the > initrd_high = ffffffff and IT WORKED !!! > Thank you so much. You're welcome. > I have a dubious distinction of never getting the > busybox built by me up and running :-) See the SELF (either from the ELDK RPM or from our FTP server). Best regards, Wolfgang Denk -- Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd(a)denx.de "You shouldn't make my toaster angry." - Household security explained in "Johnny Quest"

1 0

[U-Boot-Users] Ramdisk for MPC8xx
by Raghu 10 Oct '03

10 Oct '03

Hello All, I am porting linux-2.4.22 to MPC852T based board. I have an issue with the Ramdisk. I have a total of 64MB SDRAM, and I am creating a ramdisk of 32MB. The compressed Ramdisk file is of 3.4MB size. Look like when I do the "bootm" command the Ramdisk is corrupting some important u-boot data. I have the console dump below. Is there a minimal Ramdisk that I can directly use ?? Warm Regards Raghu. ****************************************************** =>tftpload 1000000 uImage TFTP from server 10.208.1.90; our IP address is 10.208.1.247 Filename 'uImage'. Load address: 0x1000000 Loading: ####### done Bytes transferred = 649227 (9e80b hex) =>tftpload 2000000 ramdisk TFTP from server 10.208.1.90; our IP address is 10.208.1.247 Filename 'ramdisk'. Load address: 0x2000000 Loading:#### done Bytes transferred = 3685400 (383c18 hex) =>bootm 1000000 2000000 ## Booting image at 01000000 ... Image Name: Linux 2.4.22 Kernel Image Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 649163 Bytes = 633.9 kB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK Uncompressing Kernel Image ... OK ## cmdline at 0x007FFF00 ... 0x007FFF10 memstart = 0x00000000 memsize = 0x04000000 flashstart = 0xFE000000 flashsize = 0x00100000 flashoffset = 0x0001B000 sramstart = 0x00000000 sramsize = 0x00000000 immr_base = 0xFF000000 bootflags = 0x00000001 intfreq = 50 MHz busfreq = 50 MHz ethaddr = 00:00:00:00:00:02 IP addr = 10.208.1.247 baudrate = 38400 bps ## Loading RAMDisk Image at 02000000 ... Image Name: Linux Ramdisk Image Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 3685336 Bytes = 3.5 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK Loading Ramdisk to 03bcb000, end 03f4ebd8 ... ****************************************************** __________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com

2 3

[U-Boot-Users] Re: [PATCH] u-boot, support for dbau1x00 board
by Wolfgang Denk 09 Oct '03

09 Oct '03

In message <3F858A9D.5000002(a)corelatus.com> you wrote: > > here is the patch for dbau1x00 board. > > I left out the renaming of serial.* to asc_serial.* from > the main patch in case you want to do it yourself in > cvs. > I add a second patch for this in case you want it. Added both patches. > Comment: > I changed the mail address for Magnus Damm. He left Ericsson > some 5 years ago, so that address does not work anymore. Thanks. > dbau1x00 board will not compile unless 2 of the patches > for cmd_ide.c I mailed to list is added. > ( ATA read/write and CONFIG_IDE_8xx_DIRECT stuff. ) > > * Patch by Thomas Lange, 09 Oct 2003: > Support for new board dbau1x00 (mips32) added Added, thanks. Will show up on CVS soon. [You keep me pretty busy these days ;-) ] Best regards, Wolfgang Denk -- Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd(a)denx.de Little known fact about Middle Earth: The Hobbits had a very sophi- sticated computer network! It was a Tolkien Ring...

1 0

[U-Boot-Users] [PATCH] memsize option
by Thomas Lange 09 Oct '03

09 Oct '03

Some versions of linux expect memsize in MB, other in bytes. This patch adds a new config option CONFIG_MEMSIZE_IN_BYTES that solves this. Also added a fix for mips that uses this new option. * Patch by Thomas Lange, 9 Oct 2003: New option CONFIG_MEMSIZE_IN_BYTES for passing memsize args to linux /Thomas diff -purN u-boot-cvs/README u-boot-new/README --- u-boot-cvs/README 2003-09-18 12:02:25.000000000 +0200 +++ u-boot-new/README 2003-10-09 12:37:08.000000000 +0200 @@ -405,7 +405,7 @@ The following options need to be configu CONFIG_8xx_GCLK_FREQ - if get_gclk_freq() can not work e.g. no 32KHz reference PIT/RTC clock -- Clock Interface: +- Linux Interface: CONFIG_CLOCKS_IN_MHZ U-Boot stores all clock information in Hz @@ -415,11 +415,16 @@ The following options need to be configu "clocks_in_mhz" can be defined so that U-Boot converts clock data to MHZ before passing it to the Linux kernel. - When CONFIG_CLOCKS_IN_MHZ is defined, a definition of "clocks_in_mhz=1" is automatically included in the default environment. + CONFIG_MEMSIZE_IN_BYTES + + When transfering memsize parameter to linux, some versions + expect it to be in bytes, others in MB. + Define CONFIG_MEMSIZE_IN_BYTES to make it in bytes. + - Console Interface: Depending on board, define exactly one serial port (like CONFIG_8xx_CONS_SMC1, CONFIG_8xx_CONS_SMC2, diff -purN u-boot-cvs/lib_mips/mips_linux.c u-boot-new/lib_mips/mips_linux.c --- u-boot-cvs/lib_mips/mips_linux.c 2003-06-27 23:32:47.000000000 +0200 +++ u-boot-new/lib_mips/mips_linux.c 2003-10-09 12:46:39.000000000 +0200 @@ -183,7 +183,18 @@ void do_bootm_linux(cmd_tbl_t *cmdtp, in linux_params_init (PHYSADDR(gd->bd->bi_boot_params), commandline); +#ifdef CONFIG_MEMSIZE_IN_BYTES + sprintf (env_buf, "%lu", gd->ram_size); +#ifdef DEBUG + printf ("## Giving linux memsize in bytes, %lu\n",gd->ram_size); +#endif +#else sprintf (env_buf, "%lu", gd->ram_size >> 20); +#ifdef DEBUG + printf ("## Giving linux memsize in MB, %lu\n",gd->ram_size >> 20); +#endif +#endif /* CONFIG_MEMSIZE_IN_BYTES */ + linux_env_set ("memsize", env_buf); sprintf (env_buf, "0x%08X", (uint)PHYSADDR(initrd_start));

2 1

[U-Boot-Users] [PATCH] ATA identity, Endian swap
by Thomas Lange 09 Oct '03

09 Oct '03

Reading of ATA drive id is broken in cmd_ide.c for big endian cpu:s that are not ppc. If I am not mistaken, the text strings should be endian swapped on big endian cpu:s, but not for little ones. * Patch by Thomas Lange, 10 Oct 2003: Endian swap ATA identity for all big endian cpu:s, not just ppc. /Thomas --- u-boot-0.4.8/common/cmd_ide.c 2003-07-01 23:07:07.000000000 +0200 +++ u-boot-0.4.8_new/common/cmd_ide.c 2003-10-08 23:56:06.000000000 +0200 @@ -164,9 +164,6 @@ static uchar ide_wait (int dev, ulong t static void __inline__ ide_outb(int dev, int port, unsigned char val); static unsigned char __inline__ ide_inb(int dev, int port); -#ifdef __PPC__ -static void input_swap_data(int dev, ulong *sect_buf, int words); -#endif static void input_data(int dev, ulong *sect_buf, int words); static void output_data(int dev, ulong *sect_buf, int words); static void ident_cpy (unsigned char *dest, unsigned char *src, unsigned int len); @@ -809,7 +806,13 @@ output_data_short(int dev, ulong *sect_b *pbuf = 0; } # endif /* CONFIG_AMIGAONEG3SE */ +#endif /* __PPC_ */ +/* We only need to swap data if we are running on a big endian cpu. */ +/* But Au1x00 cpu:s already swaps data in big endian mode! */ +#if defined(__LITTLE_ENDIAN) || defined(CONFIG_AU1X00) +#define input_swap_data(x,y,z) input_data(x,y,z) +#else static void input_swap_data(int dev, ulong *sect_buf, int words) { @@ -821,9 +824,7 @@ input_swap_data(int dev, ulong *sect_buf *dbuf++ = ld_le16(pbuf); } } -#else /* ! __PPC__ */ -#define input_swap_data(x,y,z) input_data(x,y,z) -#endif /* __PPC__ */ +#endif /* __LITTLE_ENDIAN || CONFIG_AU1X00 */ #ifdef __PPC__

2 1

[U-Boot-Users] [PATCH] cmd_ide.c, ifdef code only needed by 8xx_direct
by Thomas Lange 09 Oct '03

09 Oct '03

A big chunk of code in cmd_ide.c is only valid for boards that set CONFIG_IDE_8xx_DIRECT That makes it harder than neccesary to follow the flow since alot of calculations/variables are pointless. * Patch by Thomas Lange, 10 Oct 2003: code only needed by CONFIG_IDE_8xx_DIRECT ifdef:ed /Thomas --- u-boot-0.4.8/common/cmd_ide.c 2003-07-01 23:07:07.000000000 +0200 +++ u-boot-0.4.8_new/common/cmd_ide.c 2003-10-09 00:02:42.000000000 +0200 @@ -65,6 +65,7 @@ #if (CONFIG_COMMANDS & CFG_CMD_IDE) +#ifdef CONFIG_IDE_8xx_DIRECT /* Timings for IDE Interface * * SETUP / LENGTH / HOLD - cycles valid for 50 MHz clk @@ -101,6 +102,8 @@ static int pio_mode = CFG_PIO_MODE; #define PCMCIA_MK_CLKS( t, T ) (( (t) * (T) + 999U ) / 1000U ) +#endif /* CONFIG_IDE_8xx_DIRECT */ + /* ------------------------------------------------------------------------- */ /* Current I/O Device */ @@ -180,8 +183,6 @@ ulong atapi_read (int device, ulong blkn #ifdef CONFIG_IDE_8xx_DIRECT static void set_pcmcia_timing (int pmode); -#else -#define set_pcmcia_timing(a) /* dummy */ #endif /* ------------------------------------------------------------------------- */ @@ -472,9 +473,8 @@ int do_diskboot (cmd_tbl_t *cmdtp, int f void ide_init (void) { - DECLARE_GLOBAL_DATA_PTR; - #ifdef CONFIG_IDE_8xx_DIRECT + DECLARE_GLOBAL_DATA_PTR; volatile immap_t *immr = (immap_t *)CFG_IMMR; volatile pcmconf8xx_t *pcmp = &(immr->im_pcmcia); #endif @@ -502,6 +502,7 @@ void ide_init (void) WATCHDOG_RESET(); +#ifdef CONFIG_IDE_8xx_DIRECT /* Initialize PIO timing tables */ for (i=0; i <= IDE_MAX_PIO_MODE; ++i) { pio_config_clk[i].t_setup = PCMCIA_MK_CLKS(pio_config_ns[i].t_setup, @@ -518,6 +519,7 @@ void ide_init (void) pio_config_ns[i].t_length, pio_config_clk[i].t_length, pio_config_ns[i].t_hold, pio_config_clk[i].t_hold); } +#endif /* CONFIG_IDE_8xx_DIRECT */ /* Reset the IDE just to be sure. * Light LED's to show @@ -528,12 +530,12 @@ void ide_init (void) #ifdef CONFIG_IDE_8xx_DIRECT /* PCMCIA / IDE initialization for common mem space */ pcmp->pcmc_pgcrb = 0; -#endif /* start in PIO mode 0 - most relaxed timings */ pio_mode = 0; set_pcmcia_timing (pio_mode); +#endif /* * Wait for IDE to get ready. * According to spec, this can take up to 31 seconds!

2 1

[U-Boot-Users] [PATCH] fix ATA for non ppc
by Thomas Lange 09 Oct '03

09 Oct '03

AFAIK, cmd_ide.c is broken for non ppc boards. The read/write functions does not add ATA base address, which causes failure. This patch solves that. It also contains a patch for shannon board that will be broken with this change. * Patch by Thomas Lange, 10 Oct 2003: fixes for ide on non ppc boards /Thomas diff -purN u-boot-0.4.8/common/cmd_ide.c u-boot-0.4.8_new/common/cmd_ide.c --- u-boot-0.4.8/common/cmd_ide.c 2003-07-01 23:07:07.000000000 +0200 +++ u-boot-0.4.8_new/common/cmd_ide.c 2003-10-09 00:06:50.000000000 +0200 @@ -42,8 +42,12 @@ #ifdef CONFIG_STATUS_LED # include <status_led.h> #endif -#ifdef __I386__ +#ifndef __PPC__ #include <asm/io.h> +#ifdef __MIPS__ +/* Macros depend on this variable */ +static unsigned long mips_io_port_base = 0; +#endif #endif #ifdef CONFIG_SHOW_BOOT_PROGRESS @@ -116,9 +120,8 @@ ulong ide_bus_offset[CFG_IDE_MAXBUS] = { #endif }; -#ifdef __PPC__ + #define ATA_CURR_BASE(dev) (CFG_ATA_BASE_ADDR+ide_bus_offset[IDE_BUS(dev)]) -#endif #ifndef CONFIG_AMIGAONEG3SE static int ide_bus_ok[CFG_IDE_MAXBUS]; @@ -763,7 +766,7 @@ ide_outb(int dev, int port, unsigned cha static void __inline__ ide_outb(int dev, int port, unsigned char val) { - outb(val, port); + outb(val, ATA_CURR_BASE(dev)+port); } #endif /* __PPC__ */ @@ -785,7 +788,7 @@ ide_inb(int dev, int port) static unsigned char __inline__ ide_inb(int dev, int port) { - return inb(port); + return inb(ATA_CURR_BASE(dev)+port); } #endif /* __PPC__ */ @@ -846,7 +849,7 @@ output_data(int dev, ulong *sect_buf, in static void output_data(int dev, ulong *sect_buf, int words) { - outsw(ATA_DATA_REG, sect_buf, words<<1); + outsw(ATA_CURR_BASE(dev)+ATA_DATA_REG, sect_buf, words<<1); } #endif /* __PPC__ */ @@ -870,7 +873,7 @@ input_data(int dev, ulong *sect_buf, int static void input_data(int dev, ulong *sect_buf, int words) { - insw(ATA_DATA_REG, sect_buf, words << 1); + insw(ATA_CURR_BASE(dev)+ATA_DATA_REG, sect_buf, words << 1); } #endif /* __PPC__ */ @@ -1420,14 +1423,14 @@ input_data_shorts(int dev, ushort *sect_ static void output_data_shorts(int dev, ushort *sect_buf, int shorts) { - outsw(ATA_DATA_REG, sect_buf, shorts); + outsw(ATA_CURR_BASE(dev)+ATA_DATA_REG, sect_buf, shorts); } static void input_data_shorts(int dev, ushort *sect_buf, int shorts) { - insw(ATA_DATA_REG, sect_buf, shorts); + insw(ATA_CURR_BASE(dev)+ATA_DATA_REG, sect_buf, shorts); } #endif /* __PPC__ */ diff -purN u-boot-0.4.8/include/configs/shannon.h u-boot-0.4.8_new/include/configs/shannon.h --- u-boot-0.4.8/include/configs/shannon.h 2002-11-03 19:03:56.000000000 +0100 +++ u-boot-0.4.8_new/include/configs/shannon.h 2003-10-08 23:45:24.000000000 +0200 @@ -204,13 +204,13 @@ #define CFG_ATA_BASE_ADDR CFG_PCMCIA_ATTRB_ADDR /* Offset for data I/O */ -#define CFG_ATA_DATA_OFFSET (CFG_ATA_BASE_ADDR) +#define CFG_ATA_DATA_OFFSET 0 /* Offset for normal register accesses */ -#define CFG_ATA_REG_OFFSET (CFG_ATA_BASE_ADDR) +#define CFG_ATA_REG_OFFSET 0 /* Offset for alternate registers */ -#define CFG_ATA_ALT_OFFSET (CFG_ATA_BASE_ADDR) +#define CFG_ATA_ALT_OFFSET 0 /*----------------------------------------------------------------------- */

2 1

[U-Boot-Users] Re: minor patch for u-boot
by Wolfgang Denk 09 Oct '03

09 Oct '03

In message <3F83826E.4080605(a)corelatus.com> you wrote: > > Since it only affects some comments, maybe you dont need to add it > to CHANGELOG? Indeed. > ------ > * Corrected old file location comments in common.h > ------ Thanks, added. Will show up on CVS soon. Best regards, Wolfgang Denk -- Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd(a)denx.de Uncertain fortune is thoroughly mastered by the equity of the calcu- lation. - Blaise Pascal

1 0