linux/include/xen, branch v2.6.28

xen: remove unused balloon.h

2008-10-03T08:04:10Z

The balloon driver doesn't have any externally callable functions at the moment, so remove the (effectively empty) header. We can add it back if we need to. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar

xen: save previous spinlock when blocking

2008-08-21T11:52:57Z

A spinlock can be interrupted while spinning, so make sure we preserve the previous lock of interest if we're taking a lock from within an interrupt handler. We also need to deal with the case where the blocking path gets interrupted between testing to see if the lock is free and actually blocking. If we get interrupted there and end up in the state where the lock is free but the irq isn't pending, then we'll block indefinitely in the hypervisor. This fix is to make sure that any nested lock-takers will always leave the irq pending if there's any chance the outer lock became free. Signed-off-by: Jeremy Fitzhardinge Acked-by: Jan Beulich Signed-off-by: Ingo Molnar

xen: implement Xen-specific spinlocks

2008-07-16T09:15:53Z

The standard ticket spinlocks are very expensive in a virtual environment, because their performance depends on Xen's scheduler giving vcpus time in the order that they're supposed to take the spinlock. This implements a Xen-specific spinlock, which should be much more efficient. The fast-path is essentially the old Linux-x86 locks, using a single lock byte. The locker decrements the byte; if the result is 0, then they have the lock. If the lock is negative, then locker must spin until the lock is positive again. When there's contention, the locker spin for 2^16[*] iterations waiting to get the lock. If it fails to get the lock in that time, it adds itself to the contention count in the lock and blocks on a per-cpu event channel. When unlocking the spinlock, the locker looks to see if there's anyone blocked waiting for the lock by checking for a non-zero waiter count. If there's a waiter, it traverses the per-cpu "lock_spinners" variable, which contains which lock each CPU is waiting on. It picks one CPU waiting on the lock and sends it an event to wake it up. This allows efficient fast-path spinlock operation, while allowing spinning vcpus to give up their processor time while waiting for a contended lock. [*] 2^16 iterations is threshold at which 98% locks have been taken according to Thomas Friebel's Xen Summit talk "Preventing Guests from Spinning Around". Therefore, we'd expect the lock and unlock slow paths will only be entered 2% of the time. Signed-off-by: Jeremy Fitzhardinge Cc: Jens Axboe Cc: Peter Zijlstra Cc: Christoph Lameter Cc: Petr Tesarik Cc: Virtualization Cc: Xen devel Cc: Thomas Friebel Cc: Nick Piggin Signed-off-by: Ingo Molnar

xen64: fix !HVC_XEN build dependency

2008-07-16T09:06:48Z

fix: arch/x86/xen/built-in.o: In function `set_page_prot': enlighten.c:(.text+0x111d): undefined reference to `xen_raw_printk' arch/x86/xen/built-in.o: In function `xen_start_kernel': : undefined reference to `xen_raw_console_write' arch/x86/xen/built-in.o: In function `xen_start_kernel': : undefined reference to `xen_raw_console_write' Signed-off-by: Ingo Molnar

xen64: define asm/xen/interface for 64-bit

2008-07-16T08:56:18Z

Copy 64-bit definitions of various interface structures into place. Signed-off-by: Jeremy Fitzhardinge Cc: Stephen Tweedie Cc: Eduardo Habkost Cc: Mark McLoughlin Signed-off-by: Ingo Molnar

xen: add xen_arch_resume()/xen_timer_resume hook for ia64 support

2008-07-16T08:55:50Z

add xen_timer_resume() hook. Timer resume should be done after event channel is resumed. add xen_arch_resume() hook when ipi becomes usable after resume. After resume, some cpu specific resource must be reinitialized on ia64 that can't be set by another cpu. However available hooks is run once on only one cpu so that ipi has to be used. During stop_machine_run() ipi can't be used because interrupt is masked. So add another hook after stop_machine_run(). Another approach might be use resume hook which is run by device_resume(). However device_resume() may be executed on suspend error recovery path. So it is necessary to determine whether it is executed on real resume path or error recovery path. Signed-off-by: Isaku Yamahata Cc: Stephen Tweedie Cc: Eduardo Habkost Cc: Mark McLoughlin Signed-off-by: Ingo Molnar

xen: implement ptep_modify_prot_start/commit

2008-06-25T13:17:23Z

Xen has a pte update function which will update a pte while preserving its accessed and dirty bits. This means that ptep_modify_prot_start() can be implemented as a simple read of the pte value. The hardware may update the pte in the meantime, but ptep_modify_prot_commit() updates it while preserving any changes that may have happened in the meantime. The updates in ptep_modify_prot_commit() are batched if we're currently in lazy mmu mode. The mmu_update hypercall can take a batch of updates to perform, but this code doesn't make particular use of that feature, in favour of using generic multicall batching to get them all into the hypervisor. The net effect of this is that each mprotect pte update turns from two expensive trap-and-emulate faults into they hypervisor into a single hypercall whose cost is amortized in a batched multicall. Signed-off-by: Jeremy Fitzhardinge Acked-by: Linus Torvalds Acked-by: Hugh Dickins Signed-off-by: Ingo Molnar

Merge commit 'v2.6.26-rc8' into x86/xen

2008-06-25T10:16:51Z

Conflicts: arch/x86/xen/enlighten.c arch/x86/xen/mmu.c Signed-off-by: Ingo Molnar

x86: Make xen use the paravirt clocksource structs and functions

2008-06-24T18:02:32Z

This patch updates the xen guest to use the pvclock structs and helper functions. Signed-off-by: Gerd Hoffmann Acked-by: Jeremy Fitzhardinge Signed-off-by: Avi Kivity

xen: add new Xen elfnote types and use them appropriately

2008-06-02T11:25:51Z

Define recently added XEN_ELFNOTEs, and use them appropriately. Most significantly, this enables domain checkpointing (xm save -c). Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar