KVM

From KVM

(Difference between revisions)
Revision as of 09:53, 2 August 2007
AviKivity (Talk)
(guest time accounting)
← Previous diff
Revision as of 12:36, 2 August 2007
AviKivity (Talk)
(add cpuid masking)
Next diff →
Line 13: Line 13:
* The spinlock->mutex conversion exposes many possiblities of code simplifications, where previously atomic context is now regular code. * The spinlock->mutex conversion exposes many possiblities of code simplifications, where previously atomic context is now regular code.
* Store command line options in an empty snapshot in a qcow2 format image file. This allows an image file to be self contained: <code><nowiki>qemu file.img</nowiki></code> should be enough to launch a fully configured virtual machine. * Store command line options in an empty snapshot in a qcow2 format image file. This allows an image file to be self contained: <code><nowiki>qemu file.img</nowiki></code> should be enough to launch a fully configured virtual machine.
 +* Allow all cpuid features supported by kvm to be passed to the guest; add a command line option to specifiy the cpuid explicitly to allow migration in a heterogeneous server farm.
MMU related: MMU related:

Revision as of 12:36, 2 August 2007

TODO


The following items need some love. Please post to the list if you are interested in helping out:

  • Real mode support: VT support for real mode is terrible, so we need to do it in software. This means extending the x86 emulator (x86_emulate.c) to handle more instructions, and changing the execution loop to call the emulator for real mode (in progress).
  • Move the apic/ioapic/pic/pit emulation into the kernel. This will simplify in-kernel paravirtualized devices (as there is no need to exit to userspace to inject an interrupt) and guest SMP (in progress).
  • Emulate the architectural performance monitor counters/msrs, for Linux nmi watchdog support.
  • Implement smp_call_function_mask() on x86_64 and use it in kvm_flush_remote_tlbs()
  • Add "guest time" accounting to the scheduler. Currently guest time is accounted as system time.

The following smaller scale tasks can be a nice entry point to someone wishing to get involved:

  • The spinlock->mutex conversion exposes many possiblities of code simplifications, where previously atomic context is now regular code.
  • Store command line options in an empty snapshot in a qcow2 format image file. This allows an image file to be self contained: qemu file.img should be enough to launch a fully configured virtual machine.
  • Allow all cpuid features supported by kvm to be passed to the guest; add a command line option to specifiy the cpuid explicitly to allow migration in a heterogeneous server farm.

MMU related:

  • Attach kvm memory to a Linux address_space so that guest memory can be paged out (patch exists for alternative approach to guest paging).
  • Support large pages (in conjunction with the item above) so that if the guest uses a large page mapping and the guest memory is backed by hugetlbfs, a large-page pte is created.
  • Improve mmu page eviction algorithm (currently FIFO, change to approximate LRU).
  • Add a read-only memory type.
  • Implement AM20 for dos and the like.
  • Implement direct page tables on paravirt_ops enabled Linux guests (in progress)

x86 emulator updates:

  • Trap #UD and emulate sysenter/syscall/sysret/sysexit. These instructions don't exist on all cpus in all modes, so they hinder cross-vendor migration (in progress)
  • Consolidate the inb/outb emulation helpers into x86_emulate.c (in progress)
  • Currently, the emulator is called twice for each instruction that reads mmio memory: once to fetch the operand
and once to execute.  Split the emulator into two functions: one to decode into the emulation context, and
the other to actually execute the instruction.

Interactivity improvements:

  • If for several frames in a row a large proportion of the framebuffer pages are
changing, then for the next few frames don't bother to get the dirty page log
from kvm, but instead assume that all pages are dirty.  This will reduce page
fault overhead on highly interactive workloads.
  • When detecting keyboard/video/mouse activity, scale up the frame rate; when
activity dies down, scale it back down (applicable to qemu as well).

For the adventurous:

  • Emulate the VT and SVM instructions, so that kvm can run in a virtual machine. Test by running a VM in a VT guest in an SVM guest on VT hardware, as well as running a VM in an SVM guest in a VT guest on SVM hardware.
  • Emulate the VT and SVM instruction sets on qemu. This would be very beneficial to debugging kvm.
  • Keep this TODO up to date


Views Article Discussion Edit History
Personal tools:  Log in / create account
Toolbox What links here Related changes Upload file Special pages Printable version