Hi Stefan On Mon, Jul 20, 2009 at 11:54 AM, Stefan Reinauer<[email protected]> wrote: > On 20.07.2009 17:40 Uhr, Jason Wang wrote: >> Hi all, >> Since i am working with adding libpayload usb stack to option rom, >> i have been blocked with the usb_initialize function. I find that the >> size of libpayload is 104K, but the max size of option rom is oxff * >> 512bytes. which means 127.5K. >> I am afraid our option rom can not works fine with usb stack properly. >> Is there any other method for us to expand the rom size? > > I think only those parts of libpayload that are actually used are > getting linked into the final binary, since the objects are packed into > an "ar" archive. Do you exceed the rom size with libpayload and your > code already, or is this rather a theoretical issue?
No, not theoretical we have reached the max limit, even with a minimal configuration. > > -rw-r--r-- 1 stepan staff 114826 Jul 20 17:50 libpayload.a > > -rw-r--r-- 1 stepan staff 1192 Jul 20 17:50 colors.o > -rw-r--r-- 1 stepan staff 1624 Jul 20 17:50 console.o > -rw-r--r-- 1 stepan staff 1940 Jul 20 17:50 coreboot.o > -rw-r--r-- 1 stepan staff 1720 Jul 20 17:50 ctype.o > -rw-r--r-- 1 stepan staff 514 Jul 20 17:50 exec.S.o > -rw-r--r-- 1 stepan staff 692 Jul 20 17:50 exec.o > -rw-r--r-- 1 stepan staff 5928 Jul 20 17:50 getopt_long.o > -rw-r--r-- 1 stepan staff 896 Jul 20 17:50 head.S.o > -rw-r--r-- 1 stepan staff 703 Jul 20 17:50 ipchecksum.o > -rw-r--r-- 1 stepan staff 1384 Jul 20 17:50 keyboard.o > -rw-r--r-- 1 stepan staff 4592 Jul 20 17:50 lar.o > -rw-r--r-- 1 stepan staff 1328 Jul 20 17:50 lib.o > -rw-r--r-- 1 stepan staff 920 Jul 20 17:50 main.o > -rw-r--r-- 1 stepan staff 3224 Jul 20 17:50 malloc.o > -rw-r--r-- 1 stepan staff 1076 Jul 20 17:50 memory.o > -rw-r--r-- 1 stepan staff 1432 Jul 20 17:50 nvram.o > -rw-r--r-- 1 stepan staff 1608 Jul 20 17:50 options.o > -rw-r--r-- 1 stepan staff 2160 Jul 20 17:50 pci.o > -rw-r--r-- 1 stepan staff 6660 Jul 20 17:50 printf.o > -rw-r--r-- 1 stepan staff 816 Jul 20 17:50 rand.o > -rw-r--r-- 1 stepan staff 1756 Jul 20 17:50 readline.o > -rw-r--r-- 1 stepan staff 9016 Jul 20 17:50 sha1.o > -rw-r--r-- 1 stepan staff 1032 Jul 20 17:50 speaker.o > -rw-r--r-- 1 stepan staff 2648 Jul 20 17:50 string.o > -rw-r--r-- 1 stepan staff 712 Jul 20 17:50 sysinfo.o > -rw-r--r-- 1 stepan staff 1808 Jul 20 17:50 time.o > -rw-r--r-- 1 stepan staff 1448 Jul 20 17:50 timer.o > -rw-r--r-- 1 stepan staff 10224 Jul 20 17:50 tinycurses.o > -rw-r--r-- 1 stepan staff 11436 Jul 20 17:50 uhci.o > -rw-r--r-- 1 stepan staff 2384 Jul 20 17:50 uhci_rh.o > -rw-r--r-- 1 stepan staff 5560 Jul 20 17:50 usb.o > -rw-r--r-- 1 stepan staff 884 Jul 20 17:50 usb_dev.o > -rw-r--r-- 1 stepan staff 1576 Jul 20 17:50 usbinit.o > -rw-r--r-- 1 stepan staff 5704 Jul 20 17:50 usbmsc.o > -rw-r--r-- 1 stepan staff 524 Jul 20 17:50 util.S.o > -rw-r--r-- 1 stepan staff 2032 Jul 20 17:50 vga.o > -rw-r--r-- 1 stepan staff 2848 Jul 20 17:50 video.o > -rw-r--r-- 1 stepan staff 582 Jul 20 17:50 virtual.o > > > I would think there is some potential for optimization: lar, tinycurses, > vga, sha1, getopt_long, speaker, video, and probably some more stuff > could be removed? Removing most of libpayload we can take something around 100k, I think it`s not enough. Regards... -- (°= Leandro Dorileo //\ [email protected] - http://www.dorilex.net V_/ Software is a matter of freedom. -- coreboot mailing list: [email protected] http://www.coreboot.org/mailman/listinfo/coreboot
