xen.git
17 years ago[IA64] allocate percpu area in the xen va area.
Isaku Yamahata [Thu, 7 Aug 2008 02:47:34 +0000 (11:47 +0900)]
[IA64] allocate percpu area in the xen va area.

To guarantee that the percpu is pinned down,
move its virtual address from the xen identity mapped area
to the xen va area which is pinned by DTR[IA64_TR_KERNEL].
Then unnecessary tlb miss fault will be avoided.
Sometimes per cpu area is accessed from very critial
point where tlb miss isn't allowed.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
17 years ago[IA64] Fix mca handler so as not to destroy ar
Isaku Yamahata [Thu, 31 Jul 2008 03:25:50 +0000 (12:25 +0900)]
[IA64] Fix mca handler so as not to destroy ar

This patchf ixes the mca handler so as not to destroy ar
and some bugs.
On mca the firmware saves non-banked and banked general registers
and the most of ar registers must be unchanged.
However before ia64_os_mca_proc_state_dump saves such registers,
the cited hunks unconditionally overwrite ar.k[36] so that
they are destroyed.

Signed-off-by: Kazuhiro Suzuki <kaz@jp.fujitsu.com>
17 years ago[IA64] fix fpswa mapping which was broken by EFI mapping.
Isaku Yamahata [Mon, 28 Jul 2008 08:29:09 +0000 (17:29 +0900)]
[IA64] fix fpswa mapping which was broken by EFI mapping.

This patch fixes the following panic reported by Kusamura when xen
VMM tries to use fpswa.
With the EFI mapping patch, all the firmware call is done in
the dedicated address space.
fpswa is EFI driver so that the address space must be switched
before/after calling fpswa.

> (XEN) $$$$$ PANIC in domain 2 (k6=0xf0000004f25e8000): *** xen_handle_domain_access: exception table lookup failed, iip=0xf40000000408cc30,
> addr=0xe0000004ffe62050, spinning...
> (XEN) d 0xf000000004138080 domid 2
> (XEN) vcpu 0xf0000004f25e8000 vcpu 0
> (XEN)
> (XEN) CPU 5
> (XEN) psr : 0000121008226038 ifs : 8000000000000716 ip  : [<f40000000408cc31>]
> (XEN) ip is at handle_fpu_swa+0x3c1/0x510
...

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
17 years ago[IA64] improve __cpu_die()
Isaku Yamahata [Mon, 28 Jul 2008 08:29:09 +0000 (17:29 +0900)]
[IA64] improve __cpu_die()

use udelay instead of nop loop.
This patch reduces __cpu_die() time much.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
17 years ago[IA64] typo in relocate_new_kernel.
Isaku Yamahata [Mon, 28 Jul 2008 08:29:09 +0000 (17:29 +0900)]
[IA64] typo in relocate_new_kernel.

typo in relocate_new_kernel.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
17 years agomerge with xen-unstable.hg
Isaku Yamahata [Wed, 23 Jul 2008 03:10:20 +0000 (12:10 +0900)]
merge with xen-unstable.hg

17 years ago[IA64] Don't perform implicit sync when vps_save/restore
Isaku Yamahata [Wed, 23 Jul 2008 02:21:47 +0000 (11:21 +0900)]
[IA64] Don't perform implicit sync when vps_save/restore

When calling vps_save/restore, pass 1 as the third parameter not to
perform implicit sync.

The third parameter of vps_save/restore is used to indidate whether
vps_save/restore do implicit vps_read_sync/vps_write_sync.
When the third parameter is 1, it doesn't perform implicit sync.
This parameter adds flexibility of vps_save/restore.
This feature was newly introduced by SDM specification update June 2008.

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
17 years agoioemu: Fix hvmloader erroneous debugging prints
Keir Fraser [Tue, 22 Jul 2008 15:03:45 +0000 (16:03 +0100)]
ioemu: Fix hvmloader erroneous debugging prints

Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
17 years agoioemu: fix vram tracking when !s->lfb_addr
Keir Fraser [Tue, 22 Jul 2008 14:57:19 +0000 (15:57 +0100)]
ioemu: fix vram tracking when !s->lfb_addr

When we don't have an LFB (standard VGA), we can not and do not need
vram tracking at all since we always get explicit dirtying.

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agox86/64: Account for allocation bitmap by making Xen heap appropriately larger.
Keir Fraser [Tue, 22 Jul 2008 12:36:56 +0000 (13:36 +0100)]
x86/64: Account for allocation bitmap by making Xen heap appropriately larger.

Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agox86: Fix the dom0 booting hang when VT-d is enabled.
Keir Fraser [Tue, 22 Jul 2008 10:56:26 +0000 (11:56 +0100)]
x86: Fix the dom0 booting hang when VT-d is enabled.

Dom0 C/S 593 hooks the pci bus probe and remove function. When
probing a function at booting time, it will first add the device into
Dom0's list through a hypercall, then execute the driver's probe
function. If the probe function fails, another hypercall is called to
remove the device from Dom0's list. But for some RMRR devices, for
example, USB devices, they may still be in use for some operation by
BIOS at booting time. So when removing those kind of devices, we
should still keep the RMRR information.

Also add a small fix for "for_each_rmrr_device".

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Weidong Han <weidong.han@intel.com>
Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com>
17 years agoxend balloon: portability cleanup
Keir Fraser [Tue, 22 Jul 2008 10:55:06 +0000 (11:55 +0100)]
xend balloon: portability cleanup

Move the linux specific labels to osdep where they
belong. Modification on Solaris code ok'd by SUN (Ryan Scott).

Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agoFix a typo in the macro PCI_BDF2()
Keir Fraser [Tue, 22 Jul 2008 07:01:56 +0000 (08:01 +0100)]
Fix a typo in the macro PCI_BDF2()

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
17 years agofs-back: fix compilation error.
Keir Fraser [Tue, 22 Jul 2008 07:00:43 +0000 (08:00 +0100)]
fs-back: fix compilation error.

XC_PAGE_SIZE and XC_PAGE_MASK should be used instead of PAGE_SIZE
and PAGE_MASK.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
17 years agox86 hvm: stubdom and hvmloader always live under /usr/lib, never /usr/lib64
Keir Fraser [Tue, 22 Jul 2008 06:59:44 +0000 (07:59 +0100)]
x86 hvm: stubdom and hvmloader always live under /usr/lib, never /usr/lib64
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years ago[IA64] kexec: Map EFI regions into the same place they are maped into in Linux
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Map EFI regions into the same place they are maped into in Linux

Map EFI regions into the same place they are maped into in Linux

This is because of an unfortunate problem with the way that EFI interacts
with Kexec. The call to map the EFI regions may only be made once. This
means that after Kexec the EFI regions must be mapped into the same region
that they were mapped into prior to Kexec.

This is not usually a problem when kexecing from xen to xen or from linux
to linux, as the mapping will be the same. However when kexecing from xen
to linux or linux to xen, the mapping is different, and the problem
manifests.

So far Magnus Damm and I have come up with three different ideas for
resolving this problem.

1. Leave the EFI in physical mode
   - This is nice and simple
   - There is a potential performance hit, but PAL calls are not
     made very often, so it shouldn't be a problem
   - I have patches to do this, some of which are in the
     series that accompany this patch.
   - The SGI people tell me that it won't work on SN because
     it allows the OS to provide EFI (or SAL?) code.

2. Always map EFI into the space that Linux uses
   - Not so simple
   - Requires Xen to jump through some hoops
   - But leaves Linux unmodified
   - But it will break if Linux ever changes its mapping
   - This patch series implements this change

3. Always map EFI to some agreed space
   - Similar to 2. but less likely to break in the future
   - But it requires Xen and Linux to agree on a space to be used
   - Reqires both Xen and Linux to be modified

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Tristan Gingold <tgingold@free.fr>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Only map PAL when making EFI, PAL or SAL calls
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Only map PAL when making EFI, PAL or SAL calls

Move PAL code from the Xen identity mapped region to the
EFI identity mapped region, which overlaps with guest virtual space.

Make sure that PAL memory is only pinned into the TLB when making
EFI, PAL or SAL calls.

This seems to be nice as it provides a symmetrical approach to
mapping an unmapping pal code.

However it would be just as safe, and arguably simpler just to map
the PAL code (once?) when the EFI RR is active - for instance very
early on in boot, or when calling XEN_EFI_RR_ENTER. In other words,
unpining is XEN_EFI_RR_LEAVE shouldn't be neccessary, as the EFI RR
should protect the memory from unwanted accesses by guests (or
the hypevisor for that matter).

This patch is mostly the work of Yamahata-san.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: add __va_efi
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: add __va_efi

sal_desc_entry_point() converts physical addresses into virtual
addresses using __va() and these virtual addresses are used
to register the SAL and PAL handlers - which exist in firmware
memory.

As the mapping of firmware memory is being moved from
a PAGE_OFFSET of (0xf << 60) to a PAGE_OFFSET of (0xe << 60),
__va() does not provide the correct virtual address.

By adding __va_efi(), which uses EFI_PAGE_OFFSET=(0xe << 60),
and using this in sal_desc_entry_point(), the correct address is
used.

On an HP rx2600 this eliminates the problem where the SAL rendezvous address
can't be registered and subsequently the boot fails. I suspect it
solves similar problems that have been seen on other HP machines too.

Actually, its rather amazing that the HP rx2620 and Tiger2 that
I have been using work without this fix.

Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Use a separate RID for EFI
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Use a separate RID for EFI

This activates the use of the EFI RID.

The basic idea is to switch to this RID, which is in the range reserved
for the hypervisor, before making EFI, PAL or SAL calls. The page fault
handler where the identity mapping checks for this RID, if present it
does the identity mapping, else it just follows the normal mapping
rules. In this way, VMX domains should not be able to access this
memory, and they should be able to use the virtual addresses that are
used by EFI for their own purposes.

Subsequent patches move EFI memory such that faults to it will
be protected by the EFI RID.

Cc: Tristan Gingold <tgingold@free.fr>
Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Move sal.h to asm-ia64/linux-xen/asm as it needs to be modified
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Move sal.h to asm-ia64/linux-xen/asm as it needs to be modified

A subsequent patch will modify sal.h, so it needs to be moved from
xen/include/asm-ia64/linux-xen/asm/sal.h to
xen/include/asm-ia64/linux/asm/sal.h and the relevant README.origin
files need to be updated accordingly.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Handle EFI UC area correctly in vmx_alt_dtlb
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Handle EFI UC area correctly in vmx_alt_dtlb

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Acked-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Add identity mapping of EFI memory to alt_dtlb_miss
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Add identity mapping of EFI memory to alt_dtlb_miss

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Tristan Gingold <tgingold@free.fr>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Allow EFI_RID to be used in ivt.S
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Allow EFI_RID to be used in ivt.S

This is a preliminary patch to allow itentity mapping
of EFI memory to be handled in the asm page fault handlers.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Tristan Gingold <tgingold@free.fr>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Repining for EFI RID
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Repining for EFI RID

A cut down version of set_one_rr (and ia64_new_rr7) for
use when switching to the EFI RID for SAL, PAL and EFI calls.

There seems to be no need to repin: palcode, mapped_regs or vhpt in this
case. If it turns they do need to be repinned then special care needs to
betaken to track the correct value to repin.  That is generally the values
that were most recently pinned by ia64_new_rr7.

ia64_new_rr7_efi can probably be merged with ia64_new_rr7,
as they are quite similar, but for testing purposes it seems
easier to keep them separate.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Header changes in preparation for EFI RID
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Header changes in preparation for EFI RID

The EFI RID patches require pal.h to (directly or indirectly)
have access to GRANULE_SIZE which is defined in pgtable.h.
This effectively causes a header loop as pgtable.h includes
system.h and system.h includes pal.h. This patch breaks
that loop by not including pal.h in system.h if XEN is defined,
which is the only time the loop will occur.

There are two side effects of this:

1. regionreg.c makes use of some symbols declared in
   pal.h but does not include it directly. This is
   resolved by including it, which doesn't seem to
   cause any additional problems.

2. system.h makes use of ia64_pal_halt_light which is defined in pal.h.

   #define safe_halt()         ia64_pal_halt_light()

   This is probably the reason that pal.h is included in system.h.
   However this does not seem to manifest as any sort of build problem,
   presumably because either nothing in xen uses safe_halt,
   or because those that do include pal.h by some other means.

   In any case the change seems safe, though hackish.

Cc: Tristan Gingold <tgingold@free.fr>
Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Unpin shared_info, mapped_regs and VPD TR in ia64_do_tlb_purge
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Unpin shared_info, mapped_regs and VPD TR in ia64_do_tlb_purge

Unpinning shared_info, mapped_regs and VPD seems to be missing
from ia64_do_tlb_purge and seems to be needed for kexec.

Like VHPT, the pinned value is recored in a percpu variable
so that the correct value can be unpinned.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Unpin the correct VHPT TR in ia64_do_tlb_purge
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Unpin the correct VHPT TR in ia64_do_tlb_purge

GET_VA_VCPU_VHPT_MADDR(r2,r3) does not give the value of the VHPT
pinned into the TLB. I believe that this is because
current is changed between pinning and calling play_dead,
though I am not sure of the exact scemantics.

In any case, by recording the pinned value in a percpu variable,
and unpinning this value, the TR entry is removed and all is well.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Add a wrapper to vmx_switch_rr7()
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Add a wrapper to vmx_switch_rr7()

Add a C wrapper around the assembly vmx_switch_rr7().
This will be filled out with code that saves inserted TLB values
by subsequent patches.

Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: define EFI offsets for identity mapping
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: define EFI offsets for identity mapping

This is used by paches that move the EFI runtime regions into what is
normally guest space.  A description of why this mapping is made is
included in the patch that makes the mapping.

Cc: Tristan Gingold <tgingold@free.fr>
Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years ago[IA64] kexec: Define macros for EFI RID
Isaku Yamahata [Tue, 22 Jul 2008 03:15:02 +0000 (12:15 +0900)]
[IA64] kexec: Define macros for EFI RID

Macros to be called by PAL, SAL and EFI to switch into
and out of EFI RID.

Cc: Tristan Gingold <tgingold@free.fr>
Cc: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: Alex Williamson <alex.williamson@hp.com>
Cc: Aron Griffis <aron@hp.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
17 years agodoc: Fix multi-core amd example
Keir Fraser [Mon, 21 Jul 2008 15:24:19 +0000 (16:24 +0100)]
doc: Fix multi-core amd example

Signed-off-by: Jean Guyader <jean.guyader@eu.citrix.com>
17 years agotools: cpuid inputs must be 32 character long if hexadecimal value is
Keir Fraser [Mon, 21 Jul 2008 13:03:26 +0000 (14:03 +0100)]
tools: cpuid inputs must be 32 character long if hexadecimal value is
not used.

Signed-off-by: Jean Guyader <jean.guyader@eu.citrix.com>
17 years agolibxc: add xc_gnttab_map_domain_grant_refs.
Keir Fraser [Mon, 21 Jul 2008 13:02:53 +0000 (14:02 +0100)]
libxc: add xc_gnttab_map_domain_grant_refs.

xc_gnttab_map_domain_grant_refs permits to simply map several pages
from the same domain.

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agox86: Reintroduce clocksource=tsc.
Keir Fraser [Mon, 21 Jul 2008 09:39:15 +0000 (10:39 +0100)]
x86: Reintroduce clocksource=tsc.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agoRevert accidental commit from 18096:fa66b33f975a8f
Keir Fraser [Mon, 21 Jul 2008 09:00:18 +0000 (10:00 +0100)]
Revert accidental commit from 18096:fa66b33f975a8f

17 years agoFix debug key t to avoid irq jitter.
Keir Fraser [Mon, 21 Jul 2008 08:49:40 +0000 (09:49 +0100)]
Fix debug key t to avoid irq jitter.

From: Dan Magenheimer <dan.magenheimer@oracle.com>
Siigned-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agoxend: Fix keymap handling
Keir Fraser [Mon, 21 Jul 2008 08:47:07 +0000 (09:47 +0100)]
xend: Fix keymap handling

Patch allows keymap to be specified inside vfb description and
allows VM configured keymap setting to override XenD default

Signed-off-by: Pat Campbell <plc@novell.com>
17 years agox86: More accurate PIT frequency estimation (1193182Hz)
Keir Fraser [Mon, 21 Jul 2008 08:46:17 +0000 (09:46 +0100)]
x86: More accurate PIT frequency estimation (1193182Hz)

In practice this has no impact since variance between crystals will be
substantially greater than 1-2ppm. But we may as well be as accurate
as possible.

From: Dan Magenheimer <dan.magenheimer@oracle.com>
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years ago[XSM][FLASK] Remove unused code from XSM and Flask
Keir Fraser [Mon, 21 Jul 2008 08:42:30 +0000 (09:42 +0100)]
[XSM][FLASK] Remove unused code from XSM and Flask

The complete_init hook has been unused since the introduction of the
create secure interface for domain_create.

Signed-off-by: George Coker <gscoker@alpha.ncsc.mil>
17 years ago[XSM][FLASK] Argument handling bugs in XSM:FLASK
Keir Fraser [Mon, 21 Jul 2008 08:41:36 +0000 (09:41 +0100)]
[XSM][FLASK] Argument handling bugs in XSM:FLASK

Addresses a number of argument handling bugs in the flask_op hypercall
in the XSM:Flask module.  Thanks to Rafal Wojtczuk at McAfee for
reporting the issues and Tim Deegan at Citrix for providing an
initial patch.

This patch addresses the following issues:
 - bounds checking and validation on input arguments to flask_op
 - updated ABI/API, size and cmd are now uint32_t
 - updated userspace tools and libraries to account for ABI/API
 changes
 - implemented all copies using from/to guest, better portability
 - implemented upper bounds checking on op->cmd, op->size
 - implemented sanity checking on op->size and op->buf
 - implemented bit vector for checking from/to usage on op->cmd

Signed-off-by: George Coker <gscoker@alpha.ncsc.mil>
17 years agoUpdate .hgignore
Keir Fraser [Mon, 21 Jul 2008 08:40:37 +0000 (09:40 +0100)]
Update .hgignore

17 years agomini-os: fix FS frontend index reading
Keir Fraser [Mon, 21 Jul 2008 08:38:33 +0000 (09:38 +0100)]
mini-os: fix FS frontend index reading

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agomini-os: fix fs-front.c debugging prints
Keir Fraser [Mon, 21 Jul 2008 08:38:19 +0000 (09:38 +0100)]
mini-os: fix fs-front.c debugging prints

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agofs-back: enable build by default
Keir Fraser [Fri, 18 Jul 2008 16:11:27 +0000 (17:11 +0100)]
fs-back: enable build by default

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agomini-os: export allocate_ondemand
Keir Fraser [Fri, 18 Jul 2008 14:45:06 +0000 (15:45 +0100)]
mini-os: export allocate_ondemand

allocate_ondemand can be used to allocate addresse space. Primarily
used for mapping MFNs, it can also be used e.g. to map grant refs.

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agomini-os is missing list_top in its list.h
Keir Fraser [Fri, 18 Jul 2008 14:44:39 +0000 (15:44 +0100)]
mini-os is missing list_top in its list.h

Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agostubdom: fix build dependencies
Keir Fraser [Fri, 18 Jul 2008 14:03:55 +0000 (15:03 +0100)]
stubdom: fix build dependencies

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agostubdom: fix build dependency
Keir Fraser [Fri, 18 Jul 2008 13:23:41 +0000 (14:23 +0100)]
stubdom: fix build dependency

newlib now depends on mini-os header links

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agoxend: portability cleanup
Keir Fraser [Fri, 18 Jul 2008 13:14:12 +0000 (14:14 +0100)]
xend: portability cleanup

Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agomini-os: add stack walking debug
Keir Fraser [Fri, 18 Jul 2008 13:09:14 +0000 (14:09 +0100)]
mini-os: add stack walking debug

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agoamd iommu: support device add and remove for recent dom0 kernel
Keir Fraser [Fri, 18 Jul 2008 11:46:52 +0000 (12:46 +0100)]
amd iommu: support device add and remove for recent dom0 kernel
Signed-off-by: Wei Wang <wei.wang2@amd.com>
17 years agoCheck phy: block device existence
Keir Fraser [Fri, 18 Jul 2008 11:44:58 +0000 (12:44 +0100)]
Check phy: block device existence

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
17 years agostubdom: fix qemu keymap paths
Keir Fraser [Fri, 18 Jul 2008 11:27:45 +0000 (12:27 +0100)]
stubdom: fix qemu keymap paths

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agox86: Remove clocksource=tsc for now.
Keir Fraser [Fri, 18 Jul 2008 11:26:49 +0000 (12:26 +0100)]
x86: Remove clocksource=tsc for now.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agox86: Fix APERF/MPERF query bug on non-current cpu
Keir Fraser [Fri, 18 Jul 2008 10:24:13 +0000 (11:24 +0100)]
x86: Fix APERF/MPERF query bug on non-current cpu

Currently xen hypervisor cpufreq can only query APERF/MPERF on running
cpu, which will result in system broken when query on non-current
cpu. This patch fix the APERF/MPERF query bug on non-current cpu.

Signed-off-by: Liu Jinsong <jinsong.liu@intel.com>
17 years agoX86: Fix cpufreq _psd HW_ALL coordination bug
Keir Fraser [Fri, 18 Jul 2008 10:23:25 +0000 (11:23 +0100)]
X86: Fix cpufreq _psd HW_ALL coordination bug

Currently xen cpufreq has a bug when handleing _psd HW_ALL,
which will result in system broken when _psd HW_ALL.
This patch fix this bug by handling _psd HW_ALL in same way as
SW_ALL coordiantion, for the seek of safety.

Signed-off-by: Liu Jinsong <jinsong.liu@intel.com>
17 years agopvscsi: Install vscsi script.
Keir Fraser [Fri, 18 Jul 2008 10:22:12 +0000 (11:22 +0100)]
pvscsi: Install vscsi script.

Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
Signed-off-by: Jun Kamada <kama@jp.fujitsu.com>
17 years agoblktap scripts: deletion of double check and permission of using /dev/sdxx
Keir Fraser [Fri, 18 Jul 2008 10:21:06 +0000 (11:21 +0100)]
blktap scripts: deletion of double check and permission of using /dev/sdxx

Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
17 years agostubdom: add example config file
Keir Fraser [Fri, 18 Jul 2008 10:19:17 +0000 (11:19 +0100)]
stubdom: add example config file

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agomerge with xen-unstable.hg
Isaku Yamahata [Fri, 18 Jul 2008 03:12:41 +0000 (12:12 +0900)]
merge with xen-unstable.hg

17 years agostubdom: use lwip 1.3.0 release + patch instead of checking out from cvs
Keir Fraser [Thu, 17 Jul 2008 14:37:09 +0000 (15:37 +0100)]
stubdom: use lwip 1.3.0 release + patch instead of checking out from cvs

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agoPV-GRUB: use the 0.97 snapshot + patch instead of checking out from cvs
Keir Fraser [Thu, 17 Jul 2008 14:33:15 +0000 (15:33 +0100)]
PV-GRUB: use the 0.97 snapshot + patch instead of checking out from cvs

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agostubdom: add functions for caml runtime
Keir Fraser [Thu, 17 Jul 2008 12:21:37 +0000 (13:21 +0100)]
stubdom: add functions for caml runtime

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agoioemu: fix the devices loop.
Keir Fraser [Thu, 17 Jul 2008 10:20:01 +0000 (11:20 +0100)]
ioemu: fix the devices loop.

Signed-off-by: Jean Guyader <jean.guyader@eu.citrix.com>
Signed-off-by: Yuji Shimada <shimada-yxb@necst.nec.co.jp>
17 years agoxsm: Missing function in dummy module
Keir Fraser [Thu, 17 Jul 2008 10:17:32 +0000 (11:17 +0100)]
xsm: Missing function in dummy module

This adds a missing function to the dummy module to make it compile.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
17 years agox86 Cx: Add option hpetbroadcast to force enabling hpet_broadcast.
Keir Fraser [Wed, 16 Jul 2008 12:57:34 +0000 (13:57 +0100)]
x86 Cx: Add option hpetbroadcast to force enabling hpet_broadcast.

This option can be used for test & experiment purpose.

Signed-off-by: Wei Gang <gang.wei@intel.com>
17 years agox86: Adjust handle_hpet_broadcast to let it run better before broadcast exit
Keir Fraser [Wed, 16 Jul 2008 12:54:02 +0000 (13:54 +0100)]
x86: Adjust handle_hpet_broadcast to let it run better before broadcast exit

Since hpet_broadcast_exit has been moved after interrupt enabled in C3
case, so adjust the handler of hpet broadcast to adapt to this.
Meanwhile, remove a freqently executed debug print line to simplify
the serial output.

Signed-off-by: Wei Gang <gang.wei@intel.com>
17 years agoPV-GRUB: allow passing menu.lst as a ramdisk
Keir Fraser [Wed, 16 Jul 2008 12:50:47 +0000 (13:50 +0100)]
PV-GRUB: allow passing menu.lst as a ramdisk

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agox86 shadow: Re-init per-page spinlock before freeing p2m pages.
Keir Fraser [Wed, 16 Jul 2008 10:25:06 +0000 (11:25 +0100)]
x86 shadow: Re-init per-page spinlock before freeing p2m pages.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agoioemu: Fix issues when passthrough device is hot-removed from HVM domain.
Keir Fraser [Wed, 16 Jul 2008 10:16:48 +0000 (11:16 +0100)]
ioemu: Fix issues when passthrough device is hot-removed from HVM domain.

When passthrough device is hot removed from HVM domain by "xm
pci-detach" command, following issues occur. The patch fixes them.

   - Allocated memory is not deallocated.
   - Unbind interrupt with invalid interrupt pin.
   - MSI-X memory mapped register area is not unmapped.

Signed-off-by: Yuji Shimada <shimada-yxb@necst.nec.co.jp>
17 years agovt-d: Disable VT-d table output on DMA page faults
Keir Fraser [Wed, 16 Jul 2008 10:15:14 +0000 (11:15 +0100)]
vt-d: Disable VT-d table output on DMA page faults

Printing VT-d tables requires map_domain_page() which can not be
called from interrupt context (on x86_32).

Signed-off-by: Espen Skoglund <espen.skoglund@netronome.com>
17 years agofs-backend: do not expose file descriptors to frontend
Keir Fraser [Wed, 16 Jul 2008 10:13:21 +0000 (11:13 +0100)]
fs-backend: do not expose file descriptors to frontend

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agoxend: Fix destruction of VMMetrics objects for managed domains
Keir Fraser [Wed, 16 Jul 2008 10:12:36 +0000 (11:12 +0100)]
xend: Fix destruction of VMMetrics objects for managed domains

While plugging a xend memory lead (c/s 18030), I introduced a bug in
the case of managed domains.  Managed domains continue to exist after
shutdown and thus should not have their associated VMMetrics object
destroyed until deleted.

The bug actually manifested by a new -> start -> shutdown -> delete
sequence of operations.  Shutdown destroyed the VMMetrics object and
delete tried to do the same :-):

Signed-off-by: Jim Fehlig <jfehlig@novell.com>
17 years agofs-backend: fix ioctl(BLKGETSIZE) call on 64bit
Keir Fraser [Tue, 15 Jul 2008 17:00:13 +0000 (18:00 +0100)]
fs-backend: fix ioctl(BLKGETSIZE) call on 64bit

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agofs-backend: Fix freelist implementation
Keir Fraser [Tue, 15 Jul 2008 16:59:31 +0000 (17:59 +0100)]
fs-backend: Fix freelist implementation

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agox86: Report TSC skew on boot at higher log level.
Keir Fraser [Tue, 15 Jul 2008 16:58:39 +0000 (17:58 +0100)]
x86: Report TSC skew on boot at higher log level.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agox86: Ensure TSC calibration stats are same across all CPUs when
Keir Fraser [Tue, 15 Jul 2008 16:57:00 +0000 (17:57 +0100)]
x86: Ensure TSC calibration stats are same across all CPUs when
running with clocksource=tsc.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agox86: No need to vcpu_update_system_time() on map_vcpu_info().
Keir Fraser [Tue, 15 Jul 2008 16:56:25 +0000 (17:56 +0100)]
x86: No need to vcpu_update_system_time() on map_vcpu_info().
Also, ensure update_vcpu_system_time() triggers at least once for any
given vcpu, even if the TSC stamp is zero.

Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agoxen: Small cleanups towards allowing stricter compile warnings.
Keir Fraser [Tue, 15 Jul 2008 15:13:46 +0000 (16:13 +0100)]
xen: Small cleanups towards allowing stricter compile warnings.
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agoiommu: disable by default. 'iommu' on cmdline to enable.
Keir Fraser [Tue, 15 Jul 2008 14:36:50 +0000 (15:36 +0100)]
iommu: disable by default. 'iommu' on cmdline to enable.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agotools: Make functions static which should not be exported.
Keir Fraser [Tue, 15 Jul 2008 14:03:58 +0000 (15:03 +0100)]
tools: Make functions static which should not be exported.
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agox86: Add clocksource=tsc option.
Keir Fraser [Tue, 15 Jul 2008 13:04:02 +0000 (14:04 +0100)]
x86: Add clocksource=tsc option.

This option should only be used on machines where TSC is known to be
synchronized across all processors.  A future TODO is to dynamically
determine if this is the case.

Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com>
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agox86: platform-timer read function returns 64 bits.
Keir Fraser [Tue, 15 Jul 2008 12:36:22 +0000 (13:36 +0100)]
x86: platform-timer read function returns 64 bits.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agotools: Declare functions static where they should be, and provide
Keir Fraser [Tue, 15 Jul 2008 12:19:26 +0000 (13:19 +0100)]
tools: Declare functions static where they should be, and provide
proper prototypes for others as required.

Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agostubdom: enable compilation/installation by default
Keir Fraser [Mon, 14 Jul 2008 14:21:03 +0000 (15:21 +0100)]
stubdom: enable compilation/installation by default

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agovt-d: Disable VT-d if parsing ACPI DMAR fails
Keir Fraser [Mon, 14 Jul 2008 14:20:35 +0000 (15:20 +0100)]
vt-d: Disable VT-d if parsing ACPI DMAR fails

Signed-off-by: Espen Skoglund <espen.skoglund@netronome.com>
17 years agoxentop: declare signal handler as static
Keir Fraser [Mon, 14 Jul 2008 12:26:45 +0000 (13:26 +0100)]
xentop: declare signal handler as static
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agoRevert 18025:07c7aef164 -- go back to 8MB VRAM.
Keir Fraser [Mon, 14 Jul 2008 12:13:39 +0000 (13:13 +0100)]
Revert 18025:07c7aef164 -- go back to 8MB VRAM.

We may well want it in future anyway, and changing it makes
save/restore compatibility difficult.

Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agox86: PIT broadcast to fix local APIC timer stop issue for Deep C state
Keir Fraser [Mon, 14 Jul 2008 09:43:32 +0000 (10:43 +0100)]
x86: PIT broadcast to fix local APIC timer stop issue for Deep C state

Local APIC timer may stop at deep C state (C3/C4...) entry. Initial
HPET broadcast working in legacy replacing mode, broke RTC intr, so
was bypassed. This patch add the logic that use platform timer (PIT)
to reenable local APIC timer at C state entry/exit.

Currently, only keep PIT enabled with 100Hz freq. The next step is
trying to dynamically enable/disable PIT while needed, and give it
lower freq.

Signed-off-by: Yu Ke <ke.yu@intel.com>
Signed-off-by: Tian Kevin <kevin.tian@intel.com>
Signed-off-by: Wei Gang <gang.wei@intel.com>
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
17 years agoxm: vt-d: Add a command "xm pci-list-assignable-devices" to list all
Keir Fraser [Mon, 14 Jul 2008 09:12:07 +0000 (10:12 +0100)]
xm: vt-d: Add a command "xm pci-list-assignable-devices" to list all
the assignable devices.

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
17 years agoxend: vt-d: improved FLR logic for pass-thru PCI devices
Keir Fraser [Mon, 14 Jul 2008 09:11:39 +0000 (10:11 +0100)]
xend: vt-d: improved FLR logic for pass-thru PCI devices

1) If the device is PCIe endpoint and supports PCIe FLR, we use that;
else

2) if the device is PCIe endpoint, and all functions on the
device are assigned to the same guest, we use the immediate parent
bus's Secondary Bus Reset to reset all functions of the device (here,
actually we require all the functions of the device be assigned to the
same guest); else

3) if the device is PCI endpoint and is on a host bus
(e.g. integrated devices), and if the device supports PCI Advanced
Capabilities, we use that for FLR; else

4) we use the Secondary Bus Reset (if the PCI device is behind a
PCI/PCI-X bridge, then all devices behind the uppermost such PCI/PCI-X
bridge above this device must be co-assigned).

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
17 years agovt-d: Remove the FLR logic in Xen.
Keir Fraser [Mon, 14 Jul 2008 09:10:14 +0000 (10:10 +0100)]
vt-d: Remove the FLR logic in Xen.

The current simple logic has some issues: 1) Dstate transition is not
guaranteed to properly clear the device state; 2) the current code for
PCIe FLR is actually buggy: PCI_EXP_DEVSTA_TRPND doesn't mean the
completion of FLR; according to the PCIe spec, after issuing FLR, we
should wait at least 100ms.

The improved FLR logic will be added into xend.

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
17 years agoxend: A small cleanup to the find_sysfs_mnt() of pci.py
Keir Fraser [Mon, 14 Jul 2008 09:09:25 +0000 (10:09 +0100)]
xend: A small cleanup to the find_sysfs_mnt() of pci.py
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
17 years agoioemu: sdl without OpenGl fix
Keir Fraser [Mon, 14 Jul 2008 09:07:11 +0000 (10:07 +0100)]
ioemu: sdl without OpenGl fix
Signed-off-by: Frederic Guihery <sygus@cat-lan.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
17 years agoChoice of network interface for establishing a bridge on if NFSROOT is used
Keir Fraser [Mon, 14 Jul 2008 09:04:41 +0000 (10:04 +0100)]
Choice of network interface for establishing a bridge on if NFSROOT is used

This patch fixes a problem related to machines that are booted using
nfsroot ( '/' provided via nfs). Previously the code was assuming that
nfsroot would be provided via eth0. Now this additional code checks
over which interface the nfsroot is provided after detecting that
nfsroot is actually being used. To determine from where nfsroot is
mounted I am reading the kernel command line (/proc/cmdline) and
filter for an argument starting with 'nfsroot=' and determine the nfs
server's IP address by assuming the format 'nfsroot=<ip
adddress>:<path to root>' - if there's a better way of doing this,
please let me know.  After that I determine the interface over which
this IP address would be accessed using 'ip route get <address>'. Then
I compare that interface against a previously determined default
interface and if they are equal return a value that causes an
alternative interface to be chosen.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
17 years agoblktap: portability cleanup
Keir Fraser [Mon, 14 Jul 2008 09:03:09 +0000 (10:03 +0100)]
blktap: portability cleanup
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agostubdom: missing fixes for old gcc
Keir Fraser [Mon, 14 Jul 2008 09:00:47 +0000 (10:00 +0100)]
stubdom: missing fixes for old gcc

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agostubdom: fixes for old gcc & binutils
Keir Fraser [Fri, 11 Jul 2008 14:37:31 +0000 (15:37 +0100)]
stubdom: fixes for old gcc & binutils

Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
17 years agoblktap: cleanup
Keir Fraser [Fri, 11 Jul 2008 14:36:40 +0000 (15:36 +0100)]
blktap: cleanup

Make functions w/o a prototype static and remove redundant
declaration of xs_fire_next_watch().

Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
17 years agolibxc: remove redundant declaration of xc_copy_to_domain_page().
Keir Fraser [Fri, 11 Jul 2008 14:35:50 +0000 (15:35 +0100)]
libxc: remove redundant declaration of xc_copy_to_domain_page().
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>