On Saturday 10 January 2009 11:19, OGAWA Hirofumi wrote: > Hi, > > This is mainly for debugging buffer state, and cleanup for it. I think > we will play with buffer soon, to get bugs more early, I hope this is > useful for it. > > static-http://userweb.kernel.org/~hirofumi/tux3/ > > Please review.
Massive changes to buffer.c! I guess that means you like it. All this looks fine. It is still independent of tux3.h, which will be nice if we put it back into ddsnap. How should free_buffer work if there are mixed prealloc and randomly allocated buffers? Just free the prealloc vector and not worry about randomly allocated buffers, as that is just a hack to keep going when prealloc failed for some reason? Or maybe just eliminate random alloc when we are using prealloc. Instead have prealloc try again with a smaller allocation until it succeeds, which is useful when running in a tiny test environment like UML. There is no urgency to improving this. Sharing hexdump between kernel and userspace broke the kernel build because there is no definition of printf without including tux3.h, which I would like to avoid for this otherwise completely independent utility. We will remove this from kernel eventually anyway, there is already a much fancier hexdump in linux/lib/hexdump.c that we can use with a simple wrapper. (Notice there are various hexdumps littered around the kernel.) I partially reverted the hexdump move and pushed to public. Regards, Daniel _______________________________________________ Tux3 mailing list [email protected] http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
