On Sun, Jun 09, 2019 at 11:36:35PM +0200, Ahmad Fatoum wrote: > Hello Oleksij, > > On 7/6/19 09:06, Rouven Czerwinski wrote: > > From: Oleksij Rempel <linux@xxxxxxxxxxxxxxxx> > > > > With this option barebox will be build with breakpoint instruction > > in early pbl stage and on main barebox entry. Otherwise two nop > > instructions are generated which allow binary patching of existing > > barebox binaries. > > At least arm-linux-gnueabihf-gcc (Debian 8.3.0-2) does some funny > reordering with the result that the barebox header may no longer be > at the very start of the image. > This results e.g. in bootm not working for stm32 bareboxes, because > file_detect_type no longer detects them as valid barebox ARM images. > > Steps to reproduce with current next: > > $ make stm32mp1_defconfig > $ make > $ xxd -l64 -g1 images/start_stm32mp157c_dk2.pblb > 00000000: 04 46 0d 46 16 46 00 00 01 90 8f e2 19 ff 2f e1 .F.F.F......../. > 00000010: 00 f0 22 f8 fe e7 fe e7 fe e7 fe e7 fe e7 fe e7 .."............. > 00000020: fe e7 fe e7 fe e7 fe e7 62 61 72 65 62 6f 78 00 ........barebox. > 00000030: 00 00 00 00 00 00 00 00 55 55 55 55 55 55 55 55 ........UUUUUUUU > > Note how the new 8 bytes are not after the __barebox_arm_head as they should. > This leads to the barebox string being shifted from the > ARM_HEAD_MAGICWORD_OFFSET, where it's expected. > > Could you take a look? Yes, it doesn't work like this properly. I dropped 2/3 and 3/3 for now. Sascha -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox
