November 2005 - U-Boot

[U-Boot-Users] Patch for drivers/cfi_flash.c; properly detect write errors.
by Marcus Hall 28 Feb '06

28 Feb '06

Hello.. It seems that there is a minor problem in cfi_flash.c that prevents it from detecting that a flash write error has occurred. The first change in the attached patch is trivial, it just makes the strings output for a protected and a non-protected sector have the same length so the output looks prettier.. The second change is in flash_full_status_check(). We have just called flash_status_check() and then did a switch() on the vendor. For intel, we *should* check if there are any error bits if the previous call returned ERR_OK (Otherwise we will have output an error message in flash_status_check() already.) The original code would only check for error bits if flash_status_check() returns ERR_TIMEOUT. -- Marcus Hall marcushall(a)lucent.com *** ../u-boot-1.1.3/drivers/cfi_flash.c Sat Aug 13 17:53:35 2005 --- u-boot-1.1.3/drivers/cfi_flash.c Tue Aug 23 14:47:06 2005 *************** *** 502,508 **** if ((i % 5) == 0) printf ("\n "); printf (" %08lX%s", ! info->start[i], info->protect[i] ? " (RO) " : " "); #endif } putc ('\n'); --- 502,508 ---- if ((i % 5) == 0) printf ("\n "); printf (" %08lX%s", ! info->start[i], info->protect[i] ? " (RO)" : " "); #endif } putc ('\n'); *************** *** 718,724 **** switch (info->vendor) { case CFI_CMDSET_INTEL_EXTENDED: case CFI_CMDSET_INTEL_STANDARD: ! if ((retcode != ERR_OK) && !flash_isequal (info, sector, 0, FLASH_STATUS_DONE)) { retcode = ERR_INVAL; printf ("Flash %s error at address %lx\n", prompt, --- 718,724 ---- switch (info->vendor) { case CFI_CMDSET_INTEL_EXTENDED: case CFI_CMDSET_INTEL_STANDARD: ! if ((retcode == ERR_OK) && !flash_isequal (info, sector, 0, FLASH_STATUS_DONE)) { retcode = ERR_INVAL; printf ("Flash %s error at address %lx\n", prompt,

2 1

[U-Boot-Users] [PATCH] cfi_flash.c patches
by Sangmoon Kim 28 Feb '06

28 Feb '06

Hi, The two patches attached are for drivers/cfi_flash.c. cfi_flash-protect.patch adds CFG_FLASH_PROTECT_CLEAR because for some flash memories(such as 28F320C3) all banks are protected after reset. cfi_flash-buffer.patch makes write_buff not to call flash_write_cfibuffer if buffer_size is1. Because for flash memories with buffer_size 1, buffer write is not supported. Regards, Sangmoon Kim

7 28

[U-Boot-Users] PATCH for drivers/cfi_flash.c
by Tolunay Orkun 28 Feb '06

28 Feb '06

This patch reworks the patch submitted on May 6, 2005: Ref: http://sourceforge.net/mailarchive/message.php?msg_id=11674386 This patch (cfi_flash.patch.tolunay) is prepared against current CVS version + Peter Pearse Patch dated March 15, 2005. Ref: http://sourceforge.net/mailarchive/message.php?msg_id=11164812 CHANGELOG: * Patch by Tolunay Orkun, 02 July 2005: Fixes for drivers/cfi_flash.c: - Fix wrong timeout value usage in flash_status_check() - Round write_tout up when converting to msec in flash_get_size() - Remove clearing flash status at the end of flash_write_cfibuffer() which sets Intel 28F640J3 flash back to command mode on CSB472 ~~~ In case Peter's Patch is rejected for some reason please apply the alternate patch (cfi_flash.patch.tolunay2) instead. It makes one additional fix that I had submitted in my original patch. CHANGELOG (Alternate) * Patch by Tolunay Orkun, 02 July 2005: Fixes for drivers/cfi_flash.c: - Fix wrong timeout value usage in flash_status_check() - Round write_tout up when converting to msec in flash_get_size() - Remove clearing flash status at the end of flash_write_cfibuffer() which sets Intel 28F640J3 flash back to command mode on CSB472 - Fix incorrect if condition in flash_full_status_check() Best regards, Tolunay Orkun Index: drivers/cfi_flash.c =================================================================== RCS file: /cvsroot/u-boot/u-boot/drivers/cfi_flash.c,v retrieving revision 1.18 diff -p -u -r1.18 cfi_flash.c --- drivers/cfi_flash.c 16 May 2005 15:23:24 -0000 1.18 +++ drivers/cfi_flash.c 2 Jul 2005 23:45:50 -0000 @@ -694,7 +694,7 @@ static int flash_status_check (flash_inf /* Wait for command completion */ start = get_timer (0); while (flash_is_busy (info, sector)) { - if (get_timer (start) > info->erase_blk_tout * CFG_HZ) { + if (get_timer (start) > tout) { printf ("Flash %s timeout at address %lx data %lx\n", prompt, info->start[sector], flash_read_long (info, sector, 0)); @@ -1113,8 +1113,9 @@ static ulong flash_get_size (ulong base, info->erase_blk_tout = (tmp * (1 << flash_read_uchar (info, FLASH_OFFSET_EMAX_TOUT))); tmp = 1 << flash_read_uchar (info, FLASH_OFFSET_WBTOUT); info->buffer_write_tout = (tmp * (1 << flash_read_uchar (info, FLASH_OFFSET_WBMAX_TOUT))); - tmp = 1 << flash_read_uchar (info, FLASH_OFFSET_WTOUT); - info->write_tout = (tmp * (1 << flash_read_uchar (info, FLASH_OFFSET_WMAX_TOUT))) / 1000; + tmp = (1 << flash_read_uchar (info, FLASH_OFFSET_WTOUT)) * + (1 << flash_read_uchar (info, FLASH_OFFSET_WMAX_TOUT)); + info->write_tout = tmp / 1000 + (tmp % 1000 ? 1 : 0); /* round up when converting to ms */ info->flash_id = FLASH_MAN_CFI; if ((info->interface == FLASH_CFI_X8X16) && (info->chipwidth == FLASH_CFI_BY8)) { info->portwidth >>= 1; /* XXX - Need to test on x8/x16 in parallel. */ @@ -1281,7 +1282,6 @@ static int flash_write_cfibuffer (flash_ info->buffer_write_tout, "buffer write"); } - flash_write_cmd (info, sector, 0, FLASH_CMD_CLEAR_STATUS); return retcode; } #endif /* CFG_FLASH_USE_BUFFER_WRITE */ Index: drivers/cfi_flash.c =================================================================== RCS file: /cvsroot/u-boot/u-boot/drivers/cfi_flash.c,v retrieving revision 1.18 diff -p -u -r1.18 cfi_flash.c --- drivers/cfi_flash.c 16 May 2005 15:23:24 -0000 1.18 +++ drivers/cfi_flash.c 2 Jul 2005 23:52:15 -0000 @@ -694,7 +694,7 @@ static int flash_status_check (flash_inf /* Wait for command completion */ start = get_timer (0); while (flash_is_busy (info, sector)) { - if (get_timer (start) > info->erase_blk_tout * CFG_HZ) { + if (get_timer (start) > tout) { printf ("Flash %s timeout at address %lx data %lx\n", prompt, info->start[sector], flash_read_long (info, sector, 0)); @@ -719,7 +719,7 @@ static int flash_full_status_check (flas case CFI_CMDSET_INTEL_EXTENDED: case CFI_CMDSET_INTEL_STANDARD: if ((retcode != ERR_OK) - && !flash_isequal (info, sector, 0, FLASH_STATUS_DONE)) { + || !flash_isequal (info, sector, 0, FLASH_STATUS_DONE)) { retcode = ERR_INVAL; printf ("Flash %s error at address %lx\n", prompt, info->start[sector]); @@ -1113,8 +1113,9 @@ static ulong flash_get_size (ulong base, info->erase_blk_tout = (tmp * (1 << flash_read_uchar (info, FLASH_OFFSET_EMAX_TOUT))); tmp = 1 << flash_read_uchar (info, FLASH_OFFSET_WBTOUT); info->buffer_write_tout = (tmp * (1 << flash_read_uchar (info, FLASH_OFFSET_WBMAX_TOUT))); - tmp = 1 << flash_read_uchar (info, FLASH_OFFSET_WTOUT); - info->write_tout = (tmp * (1 << flash_read_uchar (info, FLASH_OFFSET_WMAX_TOUT))) / 1000; + tmp = (1 << flash_read_uchar (info, FLASH_OFFSET_WTOUT)) * + (1 << flash_read_uchar (info, FLASH_OFFSET_WMAX_TOUT)); + info->write_tout = tmp / 1000 + (tmp % 1000 ? 1 : 0); /* round up when converting to ms */ info->flash_id = FLASH_MAN_CFI; if ((info->interface == FLASH_CFI_X8X16) && (info->chipwidth == FLASH_CFI_BY8)) { info->portwidth >>= 1; /* XXX - Need to test on x8/x16 in parallel. */ @@ -1281,7 +1282,6 @@ static int flash_write_cfibuffer (flash_ info->buffer_write_tout, "buffer write"); } - flash_write_cmd (info, sector, 0, FLASH_CMD_CLEAR_STATUS); return retcode; } #endif /* CFG_FLASH_USE_BUFFER_WRITE */

2 1

[U-Boot-Users] cfi_flash.c patch
by richard＠uclinux.net 28 Feb '06

28 Feb '06

Afternoon... Adding 2 spaces in cfi_flash.c makes the output of "flinfo" much more readable if there are protected and unprotected sectors: rick ---------------------------------------------------------------------------- CHANGELOG: * Patch by Richard Klingler, 09 Apr 2005: correct column layout of flinfo output. --- u-boot.orig/drivers/cfi_flash.c 2005-01-09 19:12:33.000000000 +0200 +++ u-boot/drivers/cfi_flash.c 2005-04-09 16:35:04.000000000 +0300 @@ -483,7 +483,7 @@ void flash_print_info (flash_info_t * in if ((i % 5) == 0) printf ("\n "); printf (" %08lX%s", - info->start[i], info->protect[i] ? " (RO) " : " "); + info->start[i], info->protect[i] ? " (RO) " : " "); #endif } putc ('\n');

3 3

[U-Boot-Users] PATCH: minor bugs fix in cfi_driver.c
by Pierre AUBERT 28 Feb '06

28 Feb '06

Please find attached a little patch against current CVS. CHANGELOG: * Patch by Pierre Aubert 09 Dec 2004 - Fix flash information strings alignement in cfi_driver.c - Fix monitor size for flash protection in cfi_driver.c Best regards diff -pur u-boot/drivers/cfi_flash.c u-boot-paub/drivers/cfi_flash.c --- u-boot/drivers/cfi_flash.c 2004-10-11 01:27:35.000000000 +0200 +++ u-boot-paub/drivers/cfi_flash.c 2004-12-09 10:11:31.000000000 +0100 @@ -334,7 +334,7 @@ unsigned long flash_init (void) #if (CFG_MONITOR_BASE >= CFG_FLASH_BASE) flash_protect (FLAG_PROTECT_SET, CFG_MONITOR_BASE, - CFG_MONITOR_BASE + CFG_MONITOR_LEN - 1, + CFG_MONITOR_BASE + monitor_flash_len - 1, &flash_info[0]); #endif @@ -477,7 +477,7 @@ void flash_print_info (flash_info_t * in if ((i % 5) == 0) printf ("\n "); printf (" %08lX%s", - info->start[i], info->protect[i] ? " (RO) " : " "); + info->start[i], info->protect[i] ? " (RO)" : " "); #endif } putc ('\n');

2 1

[U-Boot-Users] [PATCH] Two System ACE CF fixes
by Peter Ryser 25 Jan '06

25 Jan '06

Attached a patch that adds support for the 16-bit wide bus of System ACE CF besides the existing support for the 8-bit wide bus. The patch also sets the FORCECFGMODE bit in the control register when the System ACE CF configuration controller is reset before reading a sector. Without this bit set System ACE CF reconfigures an FPGA if the CFGMODE pin is active whenever a sector is read. The third hunk fixes a typo. - Peter --- u-boot.orig/common/cmd_ace.c 2005-09-22 23:53:00.000000000 -0700 +++ u-boot/common/cmd_ace.c 2005-09-26 07:40:41.000000000 -0700 @@ -22,8 +22,9 @@ /* * The Xilinx SystemACE chip support is activated by defining * CONFIG_SYSTEMACE to turn on support, and CFG_SYSTEMACE_BASE - * to set the base address of the device. This code currently - * assumes that the chip is connected via a byte-wide bus. + * to set the base address of the device. CFG_SYSTEMACE_WIDTH + * selects the bus width the chip is connected with. Valid + * settings are 8 or 16. * * The CONFIG_SYSTEMACE also adds to fat support the device class * "ace" that allows the user to execute "fatls ace 0" and the @@ -137,6 +138,11 @@ block_dev_desc_t * systemace_get_dev(in systemace_dev.removable = 1; systemace_dev.block_read = systemace_read; + if(CFG_SYSTEMACE_WIDTH == 8) + ace_writew(0, 0); + else + ace_writew(1, 0); + init_part(&systemace_dev); } @@ -227,7 +233,7 @@ static unsigned long systemace_read(int /* Reset the configruation controller */ val = ace_readw(0x18); - val|=0x0080; + val|=0x0088; ace_writew(val, 0x18); retry = trans * 16; @@ -249,7 +255,7 @@ static unsigned long systemace_read(int retry -= 1; } - /* Clear the configruation controller reset */ + /* Clear the configuration controller reset */ val = ace_readw(0x18); val&=~0x0080; ace_writew(val, 0x18); * Add System ACE CF support for 16-bit bus width Add fix to ignore CFGMODE pin when resetting the System ACE CF configuration controller to avoid reconfiguration of FPGAs when reading sectors. Patch by Peter Ryser, 26 Sep 2005

2 2

[U-Boot-Users] EP852 : Kernel not booting ( Hangs after uncompressing)
by gautam borad 20 Jan '06

20 Jan '06

Hi, I'm trying to port u-boot-1.1.3 to EP852 board with mpc852T processor.The kernel im using is linuxppc 2.4.21. The board is working fine with u-boot , the problem is that kernel is not starting. I've made sure that the bd_info structure is same in kernel and u-boot. The clocks_in_mhz variable is not set and the IMMR value in kernel and u-boot matches. I run the test_burst program in u-boot and it worked fine.( so no prob with RAM) What else am i doing wrong?? The dump is : U-Boot 1.1.3 (Oct 28 2005 - 12:54:07) EP852 CPU: MPC852TxxZPnn at 50 MHz [25.0...125.0 MHz] 4 kB I-Cache 4 kB D-Cache FEC present Board: EP852 (Rev. e4). CPLD Rev: 00 DIP Switch: 1111 DRAM: 16 MB Top of RAM usable for U-Boot at: 01000000 Reserving 354k for U-Boot at: 00fa7000 Reserving 128k for malloc() at: 00f87000 Reserving 60 Bytes for Board Info at: 00f86fc4 Reserving 56 Bytes for Global Data at: 00f86f8c Stack Pointer at: 00f86f68 New Stack Pointer is: 00f86f68 Now running in RAM - U-Boot at: 00fa7000 FLASH: Flash addr:ff800000 AMD_ID_LV320B selected. before print_size 8 MB In: serial Out: serial Err: serial U-Boot relocated to 00fa7000 Net: FEC ETHERNET ### main_loop entered: bootdelay=8 ### main_loop: bootcmd="<UNDEFINED>" => printenv bootargs bootargs=console=ttyS0,9600 root=/dev/ram => tftp 0x00200000 pImage Using FEC ETHERNET device TFTP from server 10.129.87.5; our IP address is 10.129.87.250 Filename 'pImage'. Load address: 0x200000 Loading: ################################################################# ########################################################### done Bytes transferred = 631143 (9a167 hex) => bootm 0x00200000 ## Booting image at 00200000 ... Image Name: Linux-2.4.21 Created: 2005-10-27 13:15:01 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 631079 Bytes = 616.3 kB Load Address: 00100000 Entry Point: 00100000 Verifying Checksum ... OK Uncompressing Kernel Image ... OK before do_bootm_linux() argv = 16722772 addr = 16282928 ## Current stack ends at 0x00F86CA8 => set upper limit to 0x00800000 ## cmdline at 0x007FFF00 ... 0x007FFF20 No initrd ## Transferring control to Linux (at address 00100000) ... bi_memstart = 0 bi_memsize = 16777216 bi_flashstart = 4286578688 bi_flashsize = 8388608 bi_flashoffset = 343040 bi_sramstart = 0 bi_sramsize = 0 bi_immr_base = 4196401152 bi_bootflags = 1 bi_ip_addr = 176248826 bi_enetaddr[6] = bi_ethspeed = 114 bi_intfreq = 50000000 bi_busfreq = 50000000 bi_baudrate = 38400 // see the problem below The image was created with the following in arc/ppc/boot/Makefile : $(MKIMAGE) -A ppc -O linux -T kernel -C gzip -a 0x00100000 -e 0x00100000 \ -n 'Linux-$(VERSION).$(PATCHLEVEL).$(SUBLEVEL)$(EXTRAVERSION)' \ -d $< images/vmlinux.PPCBoot ln -sf vmlinux.PPCBoot images/pImage The same kernel was working fine before with the epbootloader. the partitions i made in kernel MTD are: ff80_0000 ------- ff82_0000 -> boot ff82_0000 ------- ff8c_0000 -> zImage ff8c_0000 ------- ff2c_0000 -> ramdisk ff2c_0000 ------- ffff_ffff -> UserFS (JFFS2) Some of the #defines in EP852 configs are : #define CFG_SDRAM_SIZE (16 * MB) #define CFG_FLASH_BASE 0xFF800000 #define CFG_FLASH_SIZE (8 * MB) #define CFG_BSCR_BASE 0xFA400000 #define CFG_IMMR 0xFA200000 #define CFG_JFFS2_FIRST_SECTOR 56 what should i give here? #define CFG_PRELIM_OR_AM 0xFF800000 #define CFG_OR0_PRELIM (CFG_PRELIM_OR_AM | CFG_OR_TIMING_FLASH) #define CFG_BR0_PRELIM 0xFF800001 #define CFG_OR1_PRELIM 0xFF000600 #define CFG_BR1_PRELIM ((CFG_SDRAM_BASE & BR_BA_MSK) | BR_MS_UPMA | BR_V ) #define CFG_BR3_PRELIM 0xFA400001 /* BCSR */ #define CFG_OR3_PRELIM 0xFF7F8970 #define CFG_BR4_PRELIM 0xFA000401 /* NVRAM&SRAM */ #define CFG_OR4_PRELIM 0xFFF80970 There is other minor(??) problem. When i set the value of CFG_SCCR to SCCR_COM00 | SCCR_DFSYNC00 | SCCR_DFBRG00 | SCCR_DFNL000 | SCCR_DFNH000 | SCCR_DFLCD000 | SCCR_DFALCD00 u-boot works at CONFIG_BAUDRATE / 4 speed i.e if CONFIG_BAUDRATE = 9600 i've to set minicom/kermit to 2400 . Now when i set CFG_SCCR to 0x03020800 ( the value given by embedded planet config file) the baud rate works fine but the FEC doesnt work properly. It works __VERY__ slowly, i get errors in fec_recv() ( enabling ET_DEBUG) How do I make both ( FEC and baudrate ) work fine and also is this related to kernel not booting? Thanks in advance. ----------- We would change the world..... But GOD wont give the source code.

4 6

[U-Boot-Users] [PATCH]Fix cc-option macro in config.mk
by Paolo Broggini 21 Dec '05

21 Dec '05

The following patch fix the compilation problem discussed in thread "Problem building U-Boot for ARM target" Best regards -P.Broggini Signed-off-by: Paolo Broggini <pbroggini(a)softool.ch> CHANGELOG: Fix cc-option macro in config.mk, $(CFLAGS) dropped. Patch by Paolo Broggini, 12 Oct 2005 diff --git a/config.mk b/config.mk --- a/config.mk +++ b/config.mk @@ -90,7 +90,7 @@ HOSTSTRIP = strip # Option checker (courtesy linux kernel) to ensure # only supported compiler options are used # -cc-option = $(shell if $(CC) $(CFLAGS) $(1) -S -o /dev/null -xc /dev/null \ +cc-option = $(shell if $(CC) $(1) -S -o /dev/null -xc /dev/null \ > /dev/null 2>&1; then echo "$(1)"; else echo "$(2)"; fi ;) #

3 3

[U-Boot-Users] 83xx TSEC
by David Updegraff 02 Dec '05

02 Dec '05

Hi. Anyone know of any registers to look at in the TSEC stuff that might tell a guy why a frame is not moving? Our PHY looks fine; 'mii dump' is fine, link up, etc. but in tsec_send() the buffer of a ping is loaded up, and simply never goes anywhere. ->tstat remains clear (no halt bit), but rtx.txbd[txIdx].status never changes from assigned value. Stats claim no frames; which is correct. I need some help in understanding why TSEC feels it cannot feed anything out the inerface. Thanx. -dbu.

2 2

[U-Boot-Users] IceCube MPC5200 PCI crash - need help
by Nick Droogh 30 Nov '05

30 Nov '05

Hi everyone, I am using the IceCube MPC5200 evaluation board from Freescale. Uboot 1.1.4 is operational on the board. With no PCI card install, I get the following on startup: PCI: Bus Dev VenId DevId Class Int 00 1a 1057 5803 0680 00 In: serial Out: serial Err: serial Net: FEC ETHERNET When I plug in a PCI card, the board crashes with the following message: PCI: Bus Dev VenId DevId Class Int Bus Fault @ 0x03fa3cc0, fixup 0x00000000 Machine check in kernel mode. Caused by (from msr): regs 03f4fd20 Unknown values in msr NIP: 03FA3CC0 XER: 00000000 LR: 03FA3CB4 REGS: 03f4fd20 TRAP: 0200 DAR: 20446576 MSR: 00003000 EE: 0 PR: 0 FP: 1 ME: 1 IR/DR: 00 Has anyone else come across this problem? If you had, what can be done to fix it? I don't have a BDM at the moment so it is a little more diffucult to trace. Thanks in advance, Nick Droogh

2 1