June 2009 - U-Boot

Re: [U-Boot] Kirkwood (Sheevaplug) ping crashes(?) U-Boot
by Dieter Kiermaier 22 Jun '09

22 Jun '09

-------- Original-Nachricht -------- > Datum: Mon, 22 Jun 2009 10:10:20 -0700 > Von: Prafulla Wadaskar <prafulla(a)marvell.com> > An: "dk-arm-linux(a)gmx.de" <dk-arm-linux(a)gmx.de>, "André Schwarz" <andre.schwarz(a)matrix-vision.de> > CC: "u-boot(a)lists.denx.de" <u-boot(a)lists.denx.de> > Betreff: RE: [U-Boot] Kirkwood (Sheevaplug) ping crashes(?) U-Boot > > > > -----Original Message----- > > From: u-boot-bounces(a)lists.denx.de > > [mailto:u-boot-bounces@lists.denx.de] On Behalf Of Dieter Kiermaier > > Sent: Monday, June 22, 2009 9:21 PM > > To: André Schwarz > > Cc: u-boot(a)lists.denx.de > > Subject: Re: [U-Boot] Kirkwood (Sheevaplug) ping crashes(?) U-Boot > > > > Am Montag 22 Juni 2009 17:34:35 schrieb André Schwarz: > > André, > > > > > Dieter, > > > > > > On Mon, 2009-06-22 at 17:01 +0200, Dieter Kiermaier wrote: > > > > Am Montag 22 Juni 2009 15:08:20 schrieb Dieter Kiermaier: > > > > > > > > This seems to be an issue regarding Gigabit / 100MBits > > uplink with > > > > current marvell git u-boot. > > > > > > this gives lots of things worth looking at them : > > > > > > - are you running on a custom board, i.e. is MDIO known to work > > > properly > > > -> PHY adressing ? > > > - Does your chips require different clocks for 100M/1000M ? > > > - What about mis-strapped link advertisement of the PHY ? Maybe > > > 100M-FD disabled ? > > > > > > Have you tried mii diagnostics ? > > > Have a look at common/cmd_mii.c -> "mii info" and "mii > > dump" should be > > > your friends. > > > > > > > I don't think it's a hardware problem. > > An older u-boot version provided from marvell works without > > any problems. > Yes, this is not h/w problem, auto-speed negociation was disabled, hence > it was not working with 100baseT, now this problem is resolved, > Please use attached patches to resolve this for current u-boot.git > > The improved version of this driver is already mainlined you can find it > here > http://lists.denx.de/pipermail/u-boot/2009-June/054314.html > > I am sorry I could not update the thread, niether u-boot.git on > git.marvell.com :-(, > I am facing some problem at my end, hopefully I will update everything by > tomorrow. > > > > > > > > > > > > > > > > > type "save" here and "reset". > > > > > > > > > > > > > > > > > Environment is broken only with marvell git u-boot. > > > > > > > > > Preflashed u-boot find environment is ok. > > > > > > > > > Does sheevaplug u-boot from marvell git use other > > > > > > > > > environment settings than preflashed u-boot? > Yes Marvell u-boot and u-boot.git are different uses different address to > save environment variables, so if you altered any of u-boot, environment > will be resetted. > > And they are not same on these both u-boot. u-boot.git are simple and > minimal but enough to boot from nand. Thanks for clarifying these, too. > > > > > > > > > > > > > > > > > > > Marvell>> printenv > > > > > > > > > bootcmd=${x_bootcmd_kernel}; setenv bootargs > > ${x_bootargs} > > > > > > > > >${x_bootargs_root}; bootm 0x6400000; > > > > > > > > > bootdelay=3 > > > > > > > > > baudrate=115200 > > > > > > > > > x_bootargs=console=ttyS0,115200 > > > > > > > > > > >mtdparts=orion_nand:512k(uboot),3m@1m(kernel),1m@4m(psm),13 > > > > > > > > >m@5m (roo tf s) rw x_bootcmd_kernel=nand read 0x100000 > > > > > > > > >0x6400000 0x300000 > > x_bootargs_root=root=/dev/mtdblock3 rw > > > > > > > > > rootfstype=jffs2 stdin=serial stdout=serial > > > > > > > > >stderr=serial > > > > > > > > > ethaddr=00:50:43:e0:b9:d8 > > > > > > > > > ethact=egiga0 > > > > > > > > > ipaddr=192.168.2.75 > > > > > > > > > > > > > > > > > > (ipaddr is set after bootup with setenv ipaddr > > > > > > > > > 192.168.2.75) > > > > > > > > > > > > Did you try "bootp" or "dhcp" to set up the I/F ? > > > > > > > > > > Yes, but everytimes the same: > > > > > Marvell>> dhcp > > > > > BOOTP broadcast 1 > > > > > > > > > > ... machine hangs and only a reset helps... > Even this problem is resolved in new driver/attached patches. > Many thanks for the quick response! Dieter > Regards. > Prafulla . . -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

1 0

[U-Boot] Flashing u-boot and X-loader on OMAP Zoom2
by Vivek Kutal 22 Jun '09

22 Jun '09

Hi all, I have accidentally erased the flash contents on the OMAP Zoom 2 device, now when I switch ON the board I only see the CLKREQ and the battery emulator leds ON, I have tried restoring x-loader and u-boot using the MMC card (as mentioned here https://omapzoom.org/gf/project/omapzoom/wiki/?pagename=BootingAndFlashing ) but it doesn't seem to work. I have also tried connecting JTAG as explained here https://omapzoom.org/gf/project/omapandroid/wiki/?pagename=T32Debugging but I dont see the green "running" indicator in the Trace32 software , instead in place of that I get "running (clock down)" in blue. Is my zoom2 device bricked or is there a way to reflash x-loader and u-boot ? -- Thanks Vivek

1 0

[U-Boot] Kirkwood (Sheevaplug) ping crashes(?) U-Boot
by Dieter Kiermaier 22 Jun '09

22 Jun '09

Hi all, I've tried the first time booting a self compiled u-boot from git.marvell.com U-Boot can be downloaded and started with openocd quite well. Boot messages (version is dirty due to a comment in board file): U-Boot 2009.06-rc2-02207-g14c575a-dirty (Jun 22 2009 - 13:40:45) Marvell-Sheevaplug SoC: Kirkwood 88F6281_A0 DRAM: 512 MB NAND: 512 MiB *** Warning - bad CRC or NAND, using default environment In: serial Out: serial Err: serial Net: egiga0 88E1116 Initialized on egiga0 Hit any key to stop autoboot: 0 Marvell>> Environment is broken only with marvell git u-boot. Preflashed u-boot find environment is ok. Does sheevaplug u-boot from marvell git use other environment settings than preflashed u-boot? Marvell>> printenv bootcmd=${x_bootcmd_kernel}; setenv bootargs ${x_bootargs} ${x_bootargs_root}; bootm 0x6400000; bootdelay=3 baudrate=115200 x_bootargs=console=ttyS0,115200 mtdparts=orion_nand:512k(uboot),3m@1m(kernel),1m@4m(psm),13m@5m(rootfs) rw x_bootcmd_kernel=nand read 0x100000 0x6400000 0x300000 x_bootargs_root=root=/dev/mtdblock3 rw rootfstype=jffs2 stdin=serial stdout=serial stderr=serial ethaddr=00:50:43:e0:b9:d8 ethact=egiga0 ipaddr=192.168.2.75 (ipaddr is set after bootup with setenv ipaddr 192.168.2.75) If I now try to ping another machine in the same network: Marvell>> ping 192.168.2.251 Using egiga0 device This is all what happens - no response from board, CTRL-C doesn't work. Only a hard reset will get the board up again. Do I have to change something in the default board configuration to succesfullly use networking on the sheevaplug? u-boot is build with following commands: make ARCH=arm sheevaplug_config make I get no errors during build. My gcc version is: arm-none-linux-gnueabi-gcc (GCC) 4.2.0 20070413 (prerelease) Any help would be appreciated, Dieter

2 7

[U-Boot] [PATCH] Atmel LCD driver GUARDTIME fix
by Mark Jackson 22 Jun '09

22 Jun '09

This patch correctly sets up the guard time parameter in the Atmel LCDC driver. This works on the AVR32 platform, but needs to be tested under the AT91 platform which shares the same code base. Signed-off-by: Mark Jackson <mpfj(a)mimc.co.uk> --- drivers/video/atmel_lcdfb.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/video/atmel_lcdfb.c b/drivers/video/atmel_lcdfb.c index 3a51cc7..14f578e 100644 --- a/drivers/video/atmel_lcdfb.c +++ b/drivers/video/atmel_lcdfb.c @@ -42,6 +42,7 @@ short console_row; /* configurable parameters */ #define ATMEL_LCDC_CVAL_DEFAULT 0xc8 #define ATMEL_LCDC_DMA_BURST_LEN 8 +#define ATMEL_LCDC_GUARD_TIME 2 #if defined(CONFIG_AT91SAM9263) || defined(CONFIG_AT91CAP9) #define ATMEL_LCDC_FIFO_SIZE 2048 @@ -69,7 +70,7 @@ void lcd_ctrl_init(void *lcdbase) /* Turn off the LCD controller and the DMA controller */ lcdc_writel(panel_info.mmio, ATMEL_LCDC_PWRCON, - 1 << ATMEL_LCDC_GUARDT_OFFSET); + ATMEL_LCDC_GUARD_TIME << ATMEL_LCDC_GUARDT_OFFSET); /* Wait for the LCDC core to become idle */ while (lcdc_readl(panel_info.mmio, ATMEL_LCDC_PWRCON) & ATMEL_LCDC_BUSY) @@ -150,7 +151,7 @@ void lcd_ctrl_init(void *lcdbase) lcdc_writel(panel_info.mmio, ATMEL_LCDC_DMACON, ATMEL_LCDC_DMAEN); lcdc_writel(panel_info.mmio, ATMEL_LCDC_PWRCON, - (1 << ATMEL_LCDC_GUARDT_OFFSET) | ATMEL_LCDC_PWR); + (ATMEL_LCDC_GUARD_TIME << ATMEL_LCDC_GUARDT_OFFSET) | ATMEL_LCDC_PWR); } ulong calc_fbsize(void)

3 4

[U-Boot] FW: [plagnioj@jcrosoft.com: [RFC PATCH 1/6] integratorap: splitpci support]
by Peter Pearse 22 Jun '09

22 Jun '09

ACK tghis patch Regards Peter > -----Original Message----- > From: Jean-Christophe PLAGNIOL-VILLARD [mailto:plagnioj@jcrosoft.com] > Sent: 22 June 2009 13:36 > To: Peter Pearse > Subject: Re: [plagnioj(a)jcrosoft.com: [RFC PATCH 1/6] > integratorap: splitpci support] > > On 08:16 Mon 22 Jun , Peter Pearse wrote: > > Jean-Christophe > > > > Looks fine to me > please ack them to the ML > > Best Regards, > J. >

1 0

[U-Boot] [PATCH] ppc4xx: Fix FDT EBC mappings on Canyonlands
by Felix Radensky 22 Jun '09

22 Jun '09

This patch fixes 2 problems with FDT EBC mappings on Canyonlands. First, NAND EBC mapping was missing, making Linux NAND driver unusable on this board. Second, NOR remapping code assumed that NOR is always on CS0, however when booting from NAND NOR is on CS3. --- board/amcc/canyonlands/canyonlands.c | 38 ++++++++++++++++++++++++++------- include/configs/canyonlands.h | 2 + 2 files changed, 32 insertions(+), 8 deletions(-) diff --git a/board/amcc/canyonlands/canyonlands.c b/board/amcc/canyonlands/canyonlands.c index 2b74689..78c32b0 100644 --- a/board/amcc/canyonlands/canyonlands.c +++ b/board/amcc/canyonlands/canyonlands.c @@ -577,20 +577,42 @@ int misc_init_r(void) #if defined(CONFIG_OF_LIBFDT) && defined(CONFIG_OF_BOARD_SETUP) void ft_board_setup(void *blob, bd_t *bd) { - u32 val[4]; int rc; + int i; + u32 bxcr; + u32 ranges[EBC_NUM_BANKS * 4]; + u32 *p = ranges; + char *ebc_path = "/plb/opb/ebc"; ft_cpu_setup(blob, bd); + /* + * Read 4xx EBC bus bridge registers to get mappings of the + * peripheral banks into the OPB/PLB address space + */ + for (i = 0; i < EBC_NUM_BANKS; i++) { + mtdcr(ebccfga, EBC_BXCR(i)); + bxcr = mfdcr(ebccfgd); + + if ((bxcr & EBC_BXCR_BU_MASK) != EBC_BXCR_BU_NONE) { + *p++ = i; + *p++ = 0; + *p++ = bxcr & EBC_BXCR_BAS_MASK; + *p++ = EBC_BXCR_BANK_SIZE(bxcr); + } + } + /* Fixup NOR mapping */ - val[0] = 0; /* chip select number */ - val[1] = 0; /* always 0 */ - val[2] = CONFIG_SYS_FLASH_BASE_PHYS_L; /* we fixed up this address */ - val[3] = gd->bd->bi_flashsize; - rc = fdt_find_and_setprop(blob, "/plb/opb/ebc", "ranges", - val, sizeof(val), 1); + i = CONFIG_SYS_NOR_CS * 4; + ranges[i++] = CONFIG_SYS_NOR_CS; /* chip select number */ + ranges[i++] = 0; /* always 0 */ + ranges[i++] = CONFIG_SYS_FLASH_BASE_PHYS_L; /* we fixed up this address */ + ranges[i++] = gd->bd->bi_flashsize; + + rc = fdt_find_and_setprop(blob, ebc_path, "ranges", ranges, + (p - ranges) * sizeof(u32), 1); if (rc) { - printf("Unable to update property NOR mapping, err=%s\n", + printf("Unable to update property EBC mapping, err=%s\n", fdt_strerror(rc)); } diff --git a/include/configs/canyonlands.h b/include/configs/canyonlands.h index d814012..48c5198 100644 --- a/include/configs/canyonlands.h +++ b/include/configs/canyonlands.h @@ -132,9 +132,11 @@ */ #if !defined(CONFIG_NAND_U_BOOT) && !defined(CONFIG_NAND_SPL) #define CONFIG_ENV_IS_IN_FLASH 1 /* use FLASH for environment vars */ +#define CONFIG_SYS_NOR_CS 0 /* NOR chip connected to CSx */ #define CONFIG_SYS_NAND_CS 3 /* NAND chip connected to CSx */ #else #define CONFIG_ENV_IS_IN_NAND 1 /* use NAND for environment vars */ +#define CONFIG_SYS_NOR_CS 3 /* NOR chip connected to CSx */ #define CONFIG_SYS_NAND_CS 0 /* NAND chip connected to CSx */ #define CONFIG_ENV_IS_EMBEDDED 1 /* use embedded environment */ #endif -- 1.5.4.3

2 1

[U-Boot] NAND FDT mapping on Canyonlands
by Felix Radensky 22 Jun '09

22 Jun '09

Hi, Canyonlands board port has its own version of ft_board_setup(), overriding the generic one from cpu/ppc4xx/fdt.c. Canyonlands version is missing the code that sets OPB/PLB mappings in device tree. As a result, Linux NAND driver is unusable on this board, as it fails to find a relevant NAND EBC mapping. Will a patch adding OPB/PLD mapping code to Canyonlands version of ft_board_setup() be accepted ? Felix.

3 2

[U-Boot] Question regarding FPGA firmware load in u-boot
by Dieter Kiermaier 22 Jun '09

22 Jun '09

Hello List, I want to boot a (Lattice) FPGA from u-boot by using Slave Serial configuration on a Marvell Kirkwood device. As seen there is an (Xilinx and Altera) FPGA loader driver allready available in u-boot. Please can somebody explain how I have to provide a FPGA bitstream image to u-boot? It seems I have to create a image with mkimage? This image have to be loaded into ram and the address has to be provided to the load command? Is there a maintainer for the FPGA code inside u-boot? I'm willing to contribute some extensions for Lattice FPGAs if I'm successful! Thanks to all, Dieter

2 6

Re: [U-Boot] Question regarding FPGA firmware load in u-boot
by Dieter Kiermaier 22 Jun '09

22 Jun '09

Prafulla, > > -----Original Message----- > > From: u-boot-bounces(a)lists.denx.de > > [mailto:u-boot-bounces@lists.denx.de] On Behalf Of Dieter Kiermaier > > Sent: Monday, June 22, 2009 2:30 PM > > To: André Schwarz > > Cc: u-boot(a)lists.denx.de > > Subject: Re: [U-Boot] Question regarding FPGA firmware load in u-boot > > > > André, > > > > > Dieter, > > > > > > On Mon, 2009-06-22 at 08:19 +0200, Dieter Kiermaier wrote: > > > > André, > > > > > > > > > Dieter, > > > > > > > > > > On Fri, 2009-06-19 at 17:46 +0200, Dieter Kiermaier wrote: > > > > > > Hello List, > > > > > > > > > > > > I want to boot a (Lattice) FPGA from u-boot by using Slave > > > > > > Serial configuration on a Marvell Kirkwood device. > > > > > > > > > > Should be no problem as long as you'll have access to GPIO. > > > > > > > > I'm working on it :) > > > > But more likely I would have the SPI controller to do the > > > > serializing of the bitstream. But maybe this would be the > > > > second step. > > Hi > Pls see my patch, hopefully this is useful for you > http://lists.denx.de/pipermail/u-boot/2009-June/053605.html Thanks for pointing to it. But as I tried to explain, the common SPI driver is useless for my board due to the kirkwood pin multiplexing. Thats why I asked for a TDM module based SPI driver in the past. (Please see my comment on the openplug wiki from today morning for a more detailed explanation) Dieter > Regards.. > Prafulla . .

1 0

[U-Boot] ARM mach-types.h sync request
by Daniel Gorsulowski 22 Jun '09

22 Jun '09

Dear Jean-Christophe PLAGNIOL-VILLARD, please sync include/asm-arm/mach-types.h with mach-types.h from latest kernel source to enable support for boards MEESC (number 2165) and OTC570 (number 2166). Thanks! Best regards, Daniel Gorsulowski

1 1