diff options
| author | aliguori <aliguori@c046a42c-6fe2-441c-8c8c-71466251a162> | 2008-12-09 20:09:57 +0000 |
|---|---|---|
| committer | aliguori <aliguori@c046a42c-6fe2-441c-8c8c-71466251a162> | 2008-12-09 20:09:57 +0000 |
| commit | f65ed4c1529f29a7d62d6733eaa50bed24a4b2ed (patch) | |
| tree | 1d0351afc0542f8f709f9ed7f6a445474b137a9e /hw/e1000.c | |
| parent | d85dc283fa87353be10b11b463196d10eb49ca41 (diff) | |
KVM: Coalesced MMIO support
MMIO exits are more expensive in KVM or Xen than in QEMU because they
involve, at least, privilege transitions. However, MMIO write
operations can be effectively batched if those writes do not have side
effects.
Good examples of this include VGA pixel operations when in a planar
mode. As it turns out, we can get a nice boost in other areas too.
Laurent mentioned a 9.7% performance boost in iperf with the coalesced
MMIO changes for the e1000 when he originally posted this work for KVM.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@5961 c046a42c-6fe2-441c-8c8c-71466251a162
Diffstat (limited to 'hw/e1000.c')
| -rw-r--r-- | hw/e1000.c | 12 |
1 files changed, 12 insertions, 0 deletions
diff --git a/hw/e1000.c b/hw/e1000.c index f07936fdc1..67a062a318 100644 --- a/hw/e1000.c +++ b/hw/e1000.c @@ -1001,10 +1001,22 @@ e1000_mmio_map(PCIDevice *pci_dev, int region_num, uint32_t addr, uint32_t size, int type) { E1000State *d = (E1000State *)pci_dev; + int i; + const uint32_t excluded_regs[] = { + E1000_MDIC, E1000_ICR, E1000_ICS, E1000_IMS, + E1000_IMC, E1000_TCTL, E1000_TDT, PNPMMIO_SIZE + }; + DBGOUT(MMIO, "e1000_mmio_map addr=0x%08x 0x%08x\n", addr, size); cpu_register_physical_memory(addr, PNPMMIO_SIZE, d->mmio_index); + qemu_register_coalesced_mmio(addr, excluded_regs[0]); + + for (i = 0; excluded_regs[i] != PNPMMIO_SIZE; i++) + qemu_register_coalesced_mmio(addr + excluded_regs[i] + 4, + excluded_regs[i + 1] - + excluded_regs[i] - 4); } void |