According to the EFI spec, firmware may describe ACPI tables loaded
at runtime in EFI_RESERVED_TYPE or EFI_ACPI_MEMORY_NVS memory ranges.
This patch adds these to the list of memory types we pass through for
dom0. This allows Xen to boot on HP Superdomes.
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
clean up vmx_vcpu_tpa() and mfn_to_gmfn(). don't access mpt_table directly.
use get_gpfn_from_mfn()
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
add memory barrier to domain_flush_vtlb_range().
vtlb purge must be visible before vhpt invalidation.
added some BUG_ON().
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Fix a bug. potential miss of page assignment for domain.
If mpaddr and size are not aligned well, the last page might be not
assigned for the domain.
Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>
Here's a first pass at some sparse tree cleanups based on feedback
we've gotten on linux-ia64. I haven't taken all of the comments into
account yet, this is just a start. Specific changes:
- CONFIG_ARCH_XEN is no longer necessary
- simple is_running_on_xen() checks can be handled w/o CONFIG_XEN
- pulled forward a drivers/xen/Makefile change that broken make clean
- changed some comments from C++ style to C
- turned HYPERVISOR_ioremap into a macro for !CONFIG_XEN
- made hypervisor.h and privop.h include-able for !CONFIG_XEN
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
Merge efi_emul.c into fw_emul.c
Clean up: replace some goto with return.
Declare translate_mpa_to_imva in mm.h and as void *.
Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
Clean up event wait sequence in io emulation path, as a critical
fix to a severe issue where VTI domain may fall into block state
when doing bulk network transfer.
Signed-off-by Kevin Tian <kevin.tian@intel.com>
Clean-up: fw_emul.c created.
More definitions moved to mm.c
process.c is lighter and renamed to faults.c
Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
This patch pass the vcpus number to VT-i domain guest firmware,
so that guest firmware can set SMP info accordingly, mostly in
ACPI MADT table.
This patch is required by VT-i SMP support.
Signed-off-by: Yu Ke <ke.yu@intel.com>
After enabling dom0_vp mode, we lost the VGA console.
The reason is that VGA frame buffer(0xa0000-0xc0000) was set
to conventional memory not IO in dom0's p2m table.
Low conventional memory ranges are also given MDT entries for
dom0 to allow Linux to properly detect whether VGA is present.
Signed-off-by: Zhang xiantao <xiantao.zhang@intel.com>
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
check memory descriptor over lap in dom_fw_init() and
assign page to dom0 precisely.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
paravirtualize vdso area.
introduce hyperprivop HYPERPRIVOP_RSM_BE and HYPERPRIVOP_GET_PSR.
and paravirtualize vdso area using them.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
introduce vcpu_get_domain_bundle() and replace __get_domain_handle call
in priv_handle_op with it.
priv_handle_op() uses __get_domain_handle to domain's bundle.
it directly access guest ip with guest virtual address which may results
in data tlb miss which cause some trobles.
This patch also cleans up vcpu_translate().
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
This patch intends to fix isr setting before injecting fault to it.
With this small fix, CPU2000 in VTi can pass now.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Zhang xiantao <xiantao.zhang@intel.com>
This patch fixes several minor issues, as a prepare step to support VGA
acceleration for VTI domain:
- shared vram buffer needs to be mapped as WB in both sides, because
the shared vram is the true WB memory though VTI domain is told as UC
- Introduced a new pte_mem to indicate p2m entry containing valid mfn
when replace p2m entry. Pte_none doesn't work for VTI case, since IO
type is encoded into p2m entry which however doesn't contain valid mfn.
After above change, guest_physmap_add/remove_page is available to
be used by qemu to setup linear buffer for VTI domain.
Signed-off-by Kevin Tian <kevin.tian@intel.com>
1. Make Mini-OS start from 0x0.
2. Fixes the pagetable builder to handle half full, but already mapped
pt frames.=20
3. Add a bounds check to ensure than Mini-OS does not try to use Xen
virtual space.
Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>
Signed-off-by: Aravindh Puthiyaparambil
<aravindh.puthiyaparambil@unisys.com>
