* Sync AP TSCs with BP at startup
* Only halt BP TSC when descheduled
* Correctly handle IPIs on timer vector
Signed-off-by: Xiaowei Yang <xiaowei.yang@intel.com>
As a program that runs in dom0 which serves users from guests,
the qemu drivers need to be vigilant to the input that comes
from the guests since they may be malicious.
As it is there are multiple ways to get ne2000 to read/write
memory beyond the 48K buffer that it has allocated for each
adapter.
This patch checks the addresses and prevents this from occuring.
The boundary is checked each time since it's changed for every
packet received while the other parameters are only changed
(by the guest) during setup.
Signed-off: Herbert Xu <herbert@gondor.apana.org.au>
the fly. We cannot rely on the length contained in the gzip trailer to determine
the length of the decompressed data because images have been observed which have
trailing junk.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
HVM guests have a memory hole below 4GB, so can't just check whether
addresses are < the amount of RAM allocated.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
This will now be used only by older guests who do not understand
the feature-rx-{copy,flip} feature flags.
Signed-off-by: Keir Fraser <keir@xensource.com>
support PV-on-HVM drivers for older kernel releases.
The compatability code is included in the xen-platform-pci module
since this module underpins all the others.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
Signed-off-by: K. Y. Srinivasan <ksrinivasan@novell.com>
Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
non-privileged guest.
This is useful for PV-on-HVM because it removes code which would
otherwise require extra code in the compatability shim.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
only build it when CONFIG_XEN_BACKEND is set.
This removes unused code from frontend only configurations and also
makes PV-on-HVM drivers on older kernels simpler by removing code
which would otherwise require extra code in the compatability shim.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
compatability with kernels from before Kbuild files were introduced.
This follows best practice as described in
Documentation/kbuild/modules.txt
Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
Signed-off-by: K. Y. Srinivasan <ksrinivasan@novell.com>
Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
files rather than relying on them being included indirectly.
This is required because these headers are not pulled in indirectly on
older kernels.
Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
This patch modifies memory ops to use the NUMA-aware page allocator
functions. We use the target domain's VCPU0 placement to determine
which node's memory to use. We expect the system administrator to
utilize the exposed NUMA topology information to help craft guest
config files that are NUMA-friendly (use only processors and memory values
that will fit within a given node).
Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
This patch adds a per-node bucket to the heap structure in Xen.
During heap initialization the patch determines which bucket to place
the memory. We reserve guard pages between node boundaries in the case
that said boundary isn't already guarded by the MAX_ORDER boundary to
prevent the buddy allocator from merging pages between nodes.
Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
We found Vista polls ACPI timer very frequently (about 15 times
averagely) when handling timer (RTC @ 64HZ) interrupt routine. Though
the exact reason is known, it should be related to system time
adjustment. When it's in Qemu, the overhead is big. After moving,
Vista's idle overhead decreases dramatically from ~10% to 0.9%.=20
Another benefit is that Vista can only pass Performance rating with
this patch. The root cause is that ACPI timer in Qemu isn't
synchronous with other platform timer in HV, which results in Vista
complains "can't measure TSC frequency".
This patch changes vpit.h to vpt.h, for it not only has pit structure
in it, but other platform timer's structure. Another change is moving
ACPI timer and related address from acpi.h to ioreq.h, which can be shared
by HV and ACPI firmware.
Signed-off-by: Xiaowei Yang <xiaowei.yang@intel.com>
Allow enough shadow memory to avoid thrashing the shadow pages,
rather than just enough for safety.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
As it stands grant entries for the actual data are mapped in bulk.
If one of the earlier entries fail to be mapped, it can cause all
subsequent entries to be leaked if they were successfully mapped.
This patch changes it so that we will continue to reap grant entries
even when an error is detected.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
kaf24@localhost.localdomain
Emmanuel Ackaouy
Ian Campbell
Tim Deegan
kfraser@localhost.localdomain
Ewan Mellor
awilliam@xenbuild.aw