== Items to Merge == This a list of branches and patches which might be interesting for integration into the mainline or at least worth cherrypicking. The list is in no particular order. * GRUB port ([https://code.launchpad.net/~bremby/grub4helenos/main lp:~bremby/grub4helenos/main]) * Task checkpointing ([https://code.launchpad.net/~tomas-brambora/helenos/checkpoint lp:~tomas-brambora/helenos/checkpoint]) * Probably hard to integrate as-is, but worth cherrypicking (e.g. restarting of syscalls) * Statistical profiler ([https://code.launchpad.net/~ersin/helenos/measure2 lp:~ersin/helenos/measure2]) * RTC driver, statistical profiler * MMU-less operation ([https://code.launchpad.net/~jiri-tlach/helenos/nommu lp:~jiri-tlach/helenos/nommu]) * Worth integrating as-is, but a heavy maintenance burden * The kernel from the thesis was made fibril-aware. It was later agreed with JiriT that: >''the kernel shouldn't be fibril-aware, so rewriting it so that the fibril related >info resides in uspace should be a right step.'' * Better extending to a genuine hardware platform without MMU * NUMA support ([https://code.launchpad.net/~vojtech-horky/helenos/numa lp:~vojtech-horky/helenos/numa]) * Worth integrating as-is * Branch is up-to-date with mainline and runs fine but following shall be implemented prior merge * Proper handling of memory reservations (across nodes) * DMA implementation * RBAC and security (Stepan Henek) * Impossible to integrate as-is, but worth cherrypicking * NICF ([https://code.launchpad.net/~helenos-nicf/helenos/nicf lp:~helenos-nicf/helenos/nicf]) * Some items still worth investigating/reimplementing (binary logger, DDF DMA controller interface) * Xen support (Tomas Benhak) * No public repository, but worth cherrypicking * JJ comments: * workarounds for #3 no longer needed, kernel non-identity mapping available since Dec 31, 2011 * avoids use of devman which has been officially available since March 26, 2011 * GPL code reuse audit desirable * HelenOS as Xen (Jan Dolejs) * Worth cherrypicking * JJ comments: * workarounds for #3 no longer needed, kernel non-identity mapping available since Dec 31, 2011 * avoids using generic page table interfaces for PAE (most likely unnecessarily or requiring minor changes to the generic interface) * too monolithic? (doesn't seem to exploit the multiserver design for eg. the backend drivers) * xen awareness proliferation to generic code (elf_xen_parse_notes) * GPL code reuse audit desirable * WIP: PCI and USB on PowerPC (bzr://krabicka.net/orome/helenos/usb/) * Worth early integration