On 13:40 Sun 20 Apr , Henrik Austad wrote: > On Saturday 19 April 2008 08:21:19 Michael Blizek wrote: > > There is a list on: http://kernelnewbies.org/KernelProjects > > Maybe some should have a look through and see if any of these are outdated? > At > least the balance raid-1 read seems to be fixed.
Yes it seems so. There are some other items which seem "interesting" as well: > /FasterRebuild2 > Increase speed for a build which updates a single file > Increase speed for a build wich updates a single file > We often edit a single file and then do a build. And the result is that we > spend 80% of the time linking the kernel. So an obvious improvement for the > kernel community would be to improve the speed of the linker (and decrease > memory footprint). What about http://lwn.net/Articles/274212/ ? > Asynchronous System Calls > The other issue with the current asynchronous APIs is that it is very > difficult to use. Going to a queue based asynchronous api with a > "select()"able method of managing the results will provide a more stable > programming environment. In effect, somthing similar to epoll may do the > trick. Signalfd (see http://lwn.net/Articles/225714/ ) probably solves this part. > /CorrectnessProofs > Proof that caching/buffering/transactions/syncs are handled correctly. How could anybody ever prove this? > /SelfOptimizingBlockDevice > Self-optimizing read-write block devices > There should be a mechanism that rearranges logical data blocks on the > physical block device according to access patterns: > Statistically evaluate access patterns. > Rearrange blocks on the physical block device to optimize disk accesses. > If there is enough free space on the physical device, one could even > duplicate blocks to support multiple access patterns. > To be more explicit: this is not a defragmentation mechanism. It goes even > further than that: in some situations it even fragments files in order to > improve performance. Any write access will change the disk and result in outdated statistics. It might be more interesting on the file system layer. -Michi -- programing a layer 3+4 network protocol for mesh networks see http://michaelblizek.homelinux.net -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
