>From irc: <hirofumi> static-http://userweb.kernel.org/~hirofumi/tux3/ <hirofumi> flips, please review it later <hirofumi> I'll sleep
OK, I will comment by email to make efficient use of time and share this with the email archive. For anybody who wants to read along with me: hg clone static-http://userweb.kernel.org/~hirofumi/tux3/ cd tux3 && hg view Introduce tux_error() for fs corruption: We should add sb to all warn, error and trace and assert macros, then tux_error can just be warn. The difference between warn and error in kernel is that error would put the fs into read-only mode, as should assert (which calls error), instead of just oopsing. There may be a few places where I have assumed error will not return, these should be hunted down and changed to error returns. With the additional sb parameter, we can turn tracing output on and off at runtime. When we start hitting bugs on big filesystems, that kind of control can be a real help. I don't think we need to call these macros tux_, they are pretty generic. Share MAX_GROUP_ENTRIES: You cleaned up my cleanup :-) Clear the existent magic first, then overwrite: Yes, more sensible. Does MD really want us to clear 64 KB at the top of the volume? That would seem rather sloppy of MD. Really, it should rely on the just the final sector of a volume to know if it is an MD volume. Cleanup/Fix kernel/trace.h: Wow, that gcc-specific command deletion lore is obscure :-) Fix ->i_nlink in ->rename(), and some bugs: When the structure of this settles down we can find more descriptive names for __helper functions. Introduce ARRAY_SIZE(): This macro fits the pattern of vecset, veccpy and vecmove already in tux3.h, so we can rename it to vecsize. Pulled to public. As usual, nothing broke. Regards, Daniel _______________________________________________ Tux3 mailing list [email protected] http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
