"Something _totally_ different"

Jun 18, 2007 14:26 GMT  ·  By

Unlike others who've wasted time doing nothing, Mr. Linus Torvalds has actually done some pretty good work lately. Therefore, before the promptitude proven on the Linux Kernel Mailing List (LKML) by taking part in all the interesting threads related to the Linux kernel and the forthcoming GPLv3, Linus has also sweated on creating a "brand-spanking-new" Linux 2.6.22- RC5 release.

One of the main interesting aspects in this RC would be the solution to numerous annoying regressions. For example, the suspend/resume problems in the ATA layer, which were introduced by some previous cleanups, would no longer piss you off here. That's not all. According to the post made by Linus on LKML this version also features: ?a bonus round of random architecture fixes, and an extra-bonus round of Blackfin updates."

Regarding this aspect, Mr. Tovarlds also mentioned:

"Ok, so I can't really call the Blackfin fixes for regressions, but let's face it, it's not like anybody is actually going to care. So doing the Blackfin merge at this point may not be strictly proper -rc policy, but hey, it's a new architecture. It certainly won't cause any more regressions, if only because there's not a whole lot to regress there"

Highlights:

- hugetlb: the get_policy for stacked shared memory files - fixed - pi-futex: exit races and locking problems - fixed - update checkpatch.pl to version 0.03 - update feature-removal-schedule.txt to include deprecated functions - update checkpatch.pl to version 0.04 - the MIPS machines enjoy here a fix in the system calls with long long arguments - KVM: prevent guest fpu state from leaking into the host - ibmveth : fixed the h_free_logical_lan error on pool resize

Improvements and bug-fixes can be found on most processors architectures, from PowerPC, Parc64 to Mips machines or Xtnsa. Many packages and libraries got their welcomed updates, and their behavior has been also corrected to act as smoothly as possible.