On 14 November 2010 21:29, Andrew <ni...@seti.kr.ua> wrote: > 14.11.2010 23:15, davidMbrooke пишет: >> On Sat, 2010-11-13 at 20:50 +0200, Andrew wrote: >>> 13.11.2010 12:24, davidMbrooke пишет: >>>> On Sat, 2010-11-13 at 11:11 +0200, Andrew wrote: >>>>> 12.11.2010 20:24, davidMbrooke пишет: >>>>>> I see that we are using "CONFIG_MODPROBE_SMALL" rather than >>>>>> "CONFIG_MODPROBE" for BusyBox. Is there a good reason for that? Would >>>>>> switching to CONFIG_MODPROBE be a practical option? >>>>>> >>>>>>> From comparing the source code (modprobe.c versus modprobe-small.c) >>>>>>> it >>>>>> seems to me that modprobe.c would give us (more) standard Debian >>>>>> behaviour... >>>>>> >>>>>> dMb >>>>> As I remembered, standard modprobe requires presence of System.map (file >>>>> with kernel symbols) which has huge size. But I can mistake. >>>> Hi Andrew, >>>> >>>> Is it fairly easy to check on that? >>>> >>>> I will soon be off-line until tomorrow, and I know you wanted a quick >>>> answer on this, so my thinking is: >>>> >>>> - If CONFIG_MODPROBE is a practical option (e.g. no need for >>>> System.map) then I would prefer to change to that. >>>> >>>> - If CONFIG_MODPROBE is *not* a practical option I am OK with >>>> patching BusyBox - kp convinced me :-) >>>> >>> Best way - to look in code, or to assemble bb, copy it somewhere on >>> running distro, and run in's modprobe. >> I finally got chance to try this. I rebuilt everything from scratch (I >> was due a rebuild anyway) and changed the BusyBox .config to select >> CONFIG_MODPROBE rather than CONFIG_MODPROBE_SMALL. Everything seems to >> work fine, and no need for System.map. >> > If it works good and nothing is broken - you can commit it into main > tree :)
Hi Andrew, I won't promise that *nothing* is broken, but everything I tried works OK. I will commit when I get home from work (+6 hours from now). I have been reading some more. Seems that "_SMALL" saves 14KB. I think we can spare 14KB to use non-SMALL... :-) I think I will turn *off* FEATURE_LSMOD_PRETTY_2_6_OUTPUT because that says it makes the code bigger; I don't think we need it. Same for CHECK_TAINTED_MODULE. dMb >>> CONFIG_FEATURE_CHECK_TAINTED_MODULE=y >>> CONFIG_FEATURE_MODUTILS_ALIAS=y >>> CONFIG_FEATURE_MODUTILS_SYMBOLS=y >> I didn't realize that these last three "common" options were not set >> before. They seemed like a good idea when I was answering the config >> prompts. Perhaps we can leave these as they are. >> > They are available only for full-weight modutils; simplified modutils > act like with enabled 2 last options. >> I have not investigated the module options handling but comments in >> modprobe.c indicate that /etc/modprobe.conf (or /etc/modules.conf) are >> searched for module option settings, e.g. >> options tulip2 irq=4 io=0x308 >> >> Does that let us do what we need with module options? >> >> dMb >> >> > Yes, IMHO it'll be enough. > ------------------------------------------------------------------------------ Centralized Desktop Delivery: Dell and VMware Reference Architecture Simplifying enterprise desktop deployment and management using Dell EqualLogic storage and VMware View: A highly scalable, end-to-end client virtualization framework. Read more! http://p.sf.net/sfu/dell-eql-dev2dev _______________________________________________ leaf-devel mailing list leaf-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/leaf-devel
