KVM. Avi Kivity decided to get a head start on the 2.6.31 merge window by filing a series of 43 patches to KVM. These include 9 core VMX updates, and a fairly large number of (IO)APIC and interrupt handling updates. Also on the subject of KVM, Gregory Haskins posted version 9 of his irqfd patch that aims to add support for abitrary interrupt injection and emulation into target KVM guests via signals. The most recent patch fixes a deadlock bug discovered by Marcelo Tosatti and has been rebased to a more recent git.
Xen. Jeremy Fitzhardinge, having previously asked about the fate of his recent 'Dom0' hypervisor patches, received more feedback. Ingo Molnar, in commenting on the /proc/mtrr implementation (which he didn't like because /proc/mtrr is obsolete and already on the way out in favor of PAT), pointed out that: 'That is a really broken model and design of virtualization: splitting the hypervisor into Xen and then a separate Linux dom0 entity because reality called home a few years ago and you needed actually working drivers and hardware support and a developer community to pull that off'. This wasn't the only frustration of the day, however.
Frustration. Linus expressed some frustration with recent pull requests that he has received. A recent patch entitled "avoid flexible array member inside struct" had been posted in order to avoid (ab)using a gcc extension in kernel code, as determined by Linus' own 'sparse' code verification tool. Quoting part of a mail he sent in response, Linus said "Yes, I'm upset. It's -rc6 and now two "please pull" requests have been totally unacceptable in very fundamental and obvious forms. I'm also upset becasue that obvious PIECE OF CRAP got two Acked-by's from people who should know better."
Miscalleneous. Ingo Molnar posted some core kernel fixes, including a patch that increases MAX_LOCKDEP_ENTRIES and MAX_LOCKDEP_CHAINS within the lockdep dependency checking code - the solution to failures seen by a number of people over the last few of the 2.6.30 RC kernels. Geoffrey Thomas posted a patch to sysrq handling such that panic_on_oom will not take effect if the user specifically requests OOM kill via magic sysrq.
Stephen Rothwell posted a linux-next tree for May 18th. Since Friday, the asm-generic tree has been re-added, the net fixes in the 'fixes' tree have been added to the 'net' tree, the arm tree lost its conflict, the genirq tree gained a conflict against Linus' tree, the rr tree lost its conflict, the block tree gained a conflict against the rr tree, and the asm-generic tree lost its build failure.
The latest kernel release is 2.6.30-rc6, which was released by Linus on Friday evening. A number of regressions, fixes, and other patches have already been proposed or discussed, and so a final 2.6.30 is still a little way off.