On Sat, 2014-04-26 at 21:36 +0200, Wolfgang Denk wrote: > Dear York Sun, > > In message <535be09f.80...@freescale.com> you wrote: > > >> Change return value for mac_read_from_eeprom() when mismatch happens to > > >> prevent calling hang(). > > > > > > You mean, you just ignore the error? This is a change of the cpolicy > > > that has nothing to do with generic board support, right? Why should > > > this be done now, i. e. why has it been accepted and considered to be > > > working before? > > > > This function is helpful but not critical. If reading fails, the board > > should > > continue to boot then users will have a chance to fix it. The new generic > > board > > treats this as other functions in board_init_r. Any error will cause > > hanging. > > You repeat yourself, but you do not answer my questions.
Isn't this the first time that mac_read_from_eeprom() in board/freescale/common/sys_eeprom.c was used with common/board_r.c? Previously it was called in arch/powerpc/lib/board.c which only used the initfunc array approach for board_init_f(). board_init_r() called mac_read_from_eeprom() and ignored the return. It does seem excessive to hang on initcall failure, rather than print an error and continue. -Scott _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
