Karel Gardas commented: https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/5110#note_110798 Indeed, this is classical and well expected argument. Practice however shows that there is a single testcase from ~600 tests from RTEMS testsuite I compiled with and without the option which resulted in different binary completely. All others have the same text/bss/data sizes so assumption is do not differ or differ much. One different was tftpfs. Anyway. AFIAK there is a no way how to decide if code is using FMC or it is not, besides looking into RTEMS configuration itself. E.g. once user set linkcmds.*sdram, then this is clear FMC is in use. On the other hand this is probably not easily detectable in ini itself.... So would you like to have configure option like: `STM32H7_ERRATA_ENABLE_FMC_282 = False/True` in ini file which would drive addition of -mno-unaligned-access option? Fine by me, but what would be default? Here I would really appreciate `True` as this is about stability and even ST itself recommends that! Once someone feels a need to be brave, then she/he can switch to `False` of course, but he/she will/should understand consequences well. -- View it on GitLab: https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/5110#note_110798 You're receiving this email because of your account on gitlab.rtems.org.
_______________________________________________ bugs mailing list [email protected] http://lists.rtems.org/mailman/listinfo/bugs
