Hi Michael,

On 01/09/2015 23:58, Michael Heimpold wrote:

Running mxsboot on a big-endian system produces a sd image which cannot be started by the i.MX28 ROM. It complains on the debug uart as following:

I confess that I never tested mxsboot and/or mkimage on a big endian host. I am expecting there are similar issues with mkimage and imximage, too....

0x8020a009 0x80502008 0x8020a009 0x80502008 ...

Enforcing all fields within the BCB to little-endian make the image bootable again.

Signed-off-by: Michael Heimpold mhei@heimpold.de

tools/mxsboot.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-)

diff --git a/tools/mxsboot.c b/tools/mxsboot.c index 185b327..8284e95 100644 --- a/tools/mxsboot.c +++ b/tools/mxsboot.c @@ -7,6 +7,7 @@

• SPDX-License-Identifier: GPL-2.0+

*/

+#include <endian.h> #include <fcntl.h> #include <sys/stat.h> #include <sys/types.h> @@ -569,15 +570,15 @@ static int mx28_create_sd_image(int infd, int outfd)

cb = (struct mx28_sd_config_block *)buf;

• cb->signature = 0x00112233;
• cb->primary_boot_tag = 0x1;
• cb->secondary_boot_tag = 0x1;
• cb->num_copies = 1;
• cb->drv_info[0].chip_num = 0x0;
• cb->drv_info[0].drive_type = 0x0;
• cb->drv_info[0].tag = 0x1;
• cb->drv_info[0].first_sector_number = sd_sector + 4;
• cb->drv_info[0].sector_count = (size - 4) / 512;

• cb->signature = htole32(0x00112233);

• cb->primary_boot_tag = htole32(0x1);

• cb->secondary_boot_tag = htole32(0x1);

• cb->num_copies = htole32(1);

• cb->drv_info[0].chip_num = htole32(0x0);

• cb->drv_info[0].drive_type = htole32(0x0);

• cb->drv_info[0].tag = htole32(0x1);

• cb->drv_info[0].first_sector_number = htole32(sd_sector + 4);

• cb->drv_info[0].sector_count = htole32((size - 4) / 512);

  wr_size = write(outfd, buf, size); if (wr_size != size) {

Looks ok to me !

Acked-by: Stefano Babic sbabic@denx.de

Best regards, Stefano Babic