Jan Beulich [Fri, 10 Dec 2010 18:27:17 +0000 (18:27 +0000)]
libxc: correct bounce direction for debug-key handling
Signed-off-by: Jan Beulich <jbeulich@novell.com> Acked-by: Ian Campbell <Ian.Campbell@citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Jan Beulich [Fri, 10 Dec 2010 18:08:19 +0000 (18:08 +0000)]
xend: fix "xm block-detach 0 ..." for extended-ID devices
Simply taking stat()'s st_rdev doesn't work here, as the minor is
split into two parts, the major is present, and the "extended" bit
isn't set.
Rather than fixing this in a way that would likely be OS-dependent,
simply remove the access to the device file, and instead just parse
the provided string (as is done e.g. for block-attach).
Signed-off-by: Jan Beulich <jbeulich@novell.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Keir Fraser [Fri, 10 Dec 2010 11:32:19 +0000 (11:32 +0000)]
x86 acpi: Follow Windows behaviour more closely during reset.
This follows some changes proposed for upstream Linux:
1. Do not check the FADT reset register size/offset
2. Try ACPI poking twice during our reset attempt sequence
Hopefully this will help us reset reliably on a wider range of
platforms.
Keir Fraser [Fri, 10 Dec 2010 10:51:04 +0000 (10:51 +0000)]
credit2: Don't migrate cpus unnecessarily
Modern processors have one or two L3's per socket, and generally only
one core per L2. Credit2's design relies on having credit shared
across several. So as a first step to sharing a queue across L3 while
avoiding excessive cross-L2 migration
Step one: If the vcpu's current cpu is acceptable, just run it there.
Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>
Tim Deegan [Wed, 8 Dec 2010 10:46:31 +0000 (10:46 +0000)]
x86/mm: change ASSERTs to BUG_ONs in mem_sharing.c
These two ASSERTs have important side-effects so make them into BUG_ONs
consistent with the rest of the file.
Bug found by Jui-Hao Chiang <juihaochiang@gmail.com>.
Keir Fraser [Tue, 7 Dec 2010 18:32:04 +0000 (18:32 +0000)]
x86: remove BUG_ON() from QUIRK_IOAPIC_*_REGSEL handler
Since (non-pvops, 32-bit only up to 2.6.27) Linux would report "BAD"
unconditionally on all SiS chipset versions (it only looks for a PCI
device at 0000:00:00.0 with SiS as the vendor), we must not crash if
the report on a 64-bit hypervisor doesn't match the #define (which is
zero).
While we could honor the quirk indication even on 64-bit, it doesn't
seem worthwhile, as there's no evidence that newer SiS chipsets
(supporting 64-bit CPUs) are actually affected.
This should also address bug 1687 (mis-reported, however, afaict).
Keir Fraser [Tue, 7 Dec 2010 18:26:38 +0000 (18:26 +0000)]
amd xsave: Move xsave initialization code to a common place
This patch moves xsave/xrstor code to CPU common file. First of all,
it prepares xsave/xrstor support for AMD CPUs. Secondly, Xen would
crash on __context_switch() without this patch on xsave-capable AMD
CPUs. The crash was due to cpu_has_xsave reports true in domain.c
while xsave space wasn't initialized.
Keir Fraser [Tue, 7 Dec 2010 18:24:12 +0000 (18:24 +0000)]
x86 hvm: x2APIC emulation
This patch would enable Xen to handle x2APIC MSR accessing of HVM
guest, which is faster(avoid decoding of MMIO accessing). The credit
comes to Gleb Natapov who complete the work for KVM.
Have tested with 4 vcpus guest, with/without x2apic support.
From: Sheng Yang <sheng.yang@intel.com> Signed-off-by: Keir Fraser <keir@xen.org>
Keir Fraser [Fri, 3 Dec 2010 06:37:48 +0000 (06:37 +0000)]
minios: reverse layering of xc vs minios fd close
Having minios close() call back into the libxc core close routines is
backwards and unexpected. On every other OS the libxc core close
routine calls close().
Export minios specific functions from the minios libxc code to
implement fd closing for each type of xc file handle and simply call
close() in the core close routine.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Keir Fraser [Wed, 1 Dec 2010 20:10:27 +0000 (20:10 +0000)]
x86: fix IRQ migration when using directed EOI (broken with c/s 20465)
In directed-EOI mode, there is no chance to do the migration in
mask_and_ack_level_ioapic_irq(), as the remote IRR bit can't possibly
be clear after issuing the EOI to the LAPIC. Consequently, there's no
point to even try. Instead, migration must be done in
end_level_ioapic_irq(), and it requires masking the interrupt source
prior to issuing the EOI to the IO-APIC.
Keir Fraser [Tue, 30 Nov 2010 11:34:08 +0000 (11:34 +0000)]
x86 hvm: Do not overwrite boot-cpu capability data on VMX/SVM startup.
Apparently required back in the earliest days of Xen, we now properly
initialise CPU capabilities early during bootstrap. Re-writing
capability data later now causes problems if specific features have
been deliberately masked out.
Thanks to Weidong Han at Intel for finding such a bug where XSAVE
feature is masked out by default, but then erroneously written back
during VMX initialisation. This would cause memory corruption problems
during boot for XSAVE-capable systems.
Keir Fraser [Mon, 29 Nov 2010 14:40:55 +0000 (14:40 +0000)]
x86: tighten filter on ptwr_do_page_fault()
Even not-so-recent Linux may, due to post-2.6.18 changes to the
process creation code, cause quite a number (depending on environment
and argument size) of faulting accesses to user space originating from
kernel mode. Generally those happen for non-present pages and would
lead to a nested page fault from guest_get_eff_l1e(). They can be
avoided by checking for PFEC_page_present as long as the guest isn't
running on shadow page tables.
Signed-off-by: Jan Beulich <jbeulich@novell.com> Signed-off-by: Keir Fraser <keir@xen.org>
Keir Fraser [Mon, 29 Nov 2010 14:34:32 +0000 (14:34 +0000)]
x86-64: don't crash Xen upon direct pv guest access to GDT/LDT mapping area
handle_gdt_ldt_mapping_fault() is intended to deal with indirect
accesses (i.e. those caused by descriptor loads) to the GDT/LDT
mapping area only. While for 32-bit segment limits indeed prevent the
function being entered for direct accesses (i.e. a #GP fault will be
raised even before the address translation gets done, on 64-bit even
user mode accesses would lead to control reaching the BUG_ON() at the
beginning of that function.
Fortunately the fix is simple: Since the guest kernel runs in ring 3,
any guest direct access will have the "user mode" bit set, whereas
descriptor loads always do the translations to access the actual
descriptors as kernel mode ones.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Further, relax the BUG_ON() in handle_gdt_ldt_mapping_fault() to a
check-and-bail. This avoids any problems in future, if we don't
execute x86_64 guest kernels in ring 3 (e.g., because we use a
lightweight HVM container).
Keir Fraser [Fri, 26 Nov 2010 14:25:30 +0000 (14:25 +0000)]
xenpaging: increase recently used pages from 4MB to 64MB
Increase recently used pages from 4MB to 64MB. Keeping more pages in
memory allows the guest to make more progress if the paging file spans
the entire guest memory.
Keir Fraser [Fri, 26 Nov 2010 14:23:31 +0000 (14:23 +0000)]
xenpaging: handle temporary out-of-memory conditions during page-in
p2m_mem_paging_prep() should return -ENOMEM if a new page could not be
allocated. This can be handled in xenpaging to retry the
page-in. Right now such condition would stall the guest because the
requested page will not come back, xenpaging simply exits. So
xenpaging could very well retry the allocation forever to rescue the
guest.
Keir Fraser [Fri, 26 Nov 2010 14:22:11 +0000 (14:22 +0000)]
xenpaging: optimize p2m_mem_paging_populate
p2m_mem_paging_populate will always put another request in the ring.
To reduce pressure on the ring, place only required requests in the
ring. If the gfn was already processed by another thread, and the
current vcpu does not need to be paused, p2m_mem_paging_resume will do
nothing with the request. And also xenpaging will drop the request if
the vcpu does not need a wakeup.
Keir Fraser [Fri, 26 Nov 2010 14:21:33 +0000 (14:21 +0000)]
xenpaging: when populating a page, check if populating is already in progress
p2m_mem_paging_populate can be called serveral times from different
vcpus. If the page is already in state p2m_ram_paging_in and has a new
valid mfn, invalidating this new mfn will cause trouble later if
p2m_mem_paging_resume will set the new gfn/mfn pair back to state
p2m_ram_rw. Detect this situation and keep p2m state if the page is
in the process of being still paged-out or already paged-in. In fact,
p2m state p2m_ram_paged is the only state where the mfn type can be
invalidated.
Keir Fraser [Fri, 26 Nov 2010 14:20:39 +0000 (14:20 +0000)]
xenpaging: allow negative num_pages and limit num_pages
Simplify paging size argument. If a negative number is specified, it
means the entire guest memory should be paged out. This is useful for
debugging. Also limit num_pages to the guests max_pages.
Keir Fraser [Fri, 26 Nov 2010 14:20:10 +0000 (14:20 +0000)]
xenpaging: notify policy only on resume
If a page is requested more than once, the policy is also notified
more than once about the page-in. However, a page-in happens only
once. Any further resume will only unpause the other vcpu. The
multiple notify will put the page into the mru list multiple times and
it will unlock other already resumed pages too early. In the worst
case, a page that was just resumed can be evicted right away, causing
a deadlock in the guest.
Keir Fraser [Fri, 26 Nov 2010 14:17:56 +0000 (14:17 +0000)]
xenpaging: populate paged-out pages unconditionally in grant code
Populate a page unconditionally to avoid missing a page-in request.
If the page is already in the process of being paged-in, the this vcpu
will be stopped and later resumed once the page content is usable
again.
This matches other p2m_mem_paging_populate usage in the source tree.
Keir Fraser [Fri, 26 Nov 2010 14:15:59 +0000 (14:15 +0000)]
xenpaging: Open paging file only if xenpaging_init() succeeds
Open paging file only if xenpaging_init() succeeds. It can fail if the
host does not support the required virtualization features such as EPT
or if xenpaging was already started for this domain_id.
Signed-off-by: Olaf Hering <olaf@aepfle.de> Already-Acked-by: Patrick Colp <pjcolp@cs.ubc.ca> Already-Acked-by: Keir Fraser <keir.fraser@citrix.com>
Keir Fraser [Fri, 26 Nov 2010 14:15:21 +0000 (14:15 +0000)]
xenpaging: break endless loop during inital page-out with large pagefiles
To allow the starting for xenpaging right after 'xm start XYZ', I
specified a pagefile size equal to the guest memory size in the hope
to catch more errors where the paged-out state of a p2mt is not
checked.
While doing that, xenpaging got into an endless loop because some
pages cant be paged out right away. Now the policy reports an error if
the gfn number wraps.
Signed-off-by: Olaf Hering <olaf@aepfle.de> Already-Acked-by: Patrick Colp <pjcolp@cs.ubc.ca> Already-Acked-by: Keir Fraser <keir.fraser@citrix.com>
Keir Fraser [Fri, 26 Nov 2010 10:10:40 +0000 (10:10 +0000)]
Allow assign_irq_vector to return an old vector while moving an irq
The guest calls assign_irq_vector() to assign one if it doesn't have
one, and to find out the vector if it does have one.
If the cpu mask passed intersects with the existing mask, the old
vector is simply returned.
However, if the irq happens to be in transit at the time, this returns
EBUSY. This is unnecessary if, as soon as the irq migration succeeds,
the logic would just return the old vector anyway.
This patch makes two changes:
* Switch the checks, so if the mask overlaps it always returns
* Return -EAGAIN instead of -EBUSY for moving irqs, to let the caller
know that the failure is temporary and may work if repeated.
This fixes a bug where on certain hardware, using the credit2
scheduler, a pvops kernel with multiple vcpus doesn't boot.
Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>
Tim Deegan [Wed, 24 Nov 2010 10:20:03 +0000 (10:20 +0000)]
x86/mm: remove incorrect BUG_ON.
This BUG_ON tests a property of an effectively random PFN in the guest,
and is explicitly _not_ seeing the MFN that's known to be owned.
Ian Jackson [Tue, 23 Nov 2010 19:36:14 +0000 (19:36 +0000)]
tools/xl: show shutdown reason code, improve xl list heading
Previously, xl list would not reveal the shutdown reason code unless
it was SHUTDOWN_crashed. This is unfortunate; it makes it hard for
scripts which use xl to tell what's going on.
In this patch:
* xl list shows the reason code as a single letter if it is
any of the defined values from sched.h:
- poweroff or domain not shut down
r reboot
s suspend
c crashed
w watchdog
This is not 100% backward-compatible with xm but I think it's a
justifiable improvement. It would be nice to make the same change
to xm.
* xl list -v shows the full numeric reason code in hex, or "-" if the
domain is not shut down.
* xl list -v has column headings for the UUID and numeric reason
code. The heading for the reason code overlaps with the UUID a bit.
These headings are intended for human readers; scripts can parse
the output by breaking on whitespace.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Tim Deegan [Tue, 23 Nov 2010 19:31:29 +0000 (19:31 +0000)]
tools/xl: VMs that are started paused shouldn't reboot paused too.
Otherwise a VM that's been migrated won't ever reboot cleanly again.
Signed-off-by: Tim Deegan <Tim.Deegan@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Tim Deegan [Tue, 23 Nov 2010 19:30:27 +0000 (19:30 +0000)]
tools/xl: make it explicit that we migrate from stdin.
No semantic changes, just makes things a bit less confusing.
Signed-off-by: Tim Deegan <Tim.Deegan@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Campbell [Tue, 23 Nov 2010 19:29:13 +0000 (19:29 +0000)]
tools: fetch remote changesets when force refetching/resetting qemu
This makes "make tools/ioemu-dir-force-update" usable for picking up
an entirely new QEMU_TAG.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Campbell [Tue, 23 Nov 2010 19:28:03 +0000 (19:28 +0000)]
ocaml: install built modules
Previously the install target was having no effect because it ended up
calling the default target in the subdir Makefile instead of the
install target.
Resolve this by tying the tools/ocaml Makefiles into the generic
handling done by tools/Rules.mk.
Other changes arising in one way or another from this:
- Add libs/xl/META.in
- Update .hgignore for META files
- Create leading directories
- Remove existing module before installation in install targer
(worksaround what appears to be a quirk of "ocamlfind install")
- Use the globally defined $(DESTDIR)
- Move "ocamlfind printfconf destdir" to a common variable,
repurposing exising unused OCAMLDESTDIR, incorporating $(DESTDIR) at
the same time.
- Drop a few unused variabe definitions (mainly to avoid deciding if
$(DESTDIR) made sense for them or not.
- Pass -destdir to ocamlfind in uninstall target for symmetry with
install target.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
libxl, xl: Account for shadow memory for PV guests too
We need to account for the memory needed by shadow pagetables even for PV
guests, because in that case shadow pagetables are used during live
migration.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Jan Beulich [Tue, 23 Nov 2010 16:43:38 +0000 (16:43 +0000)]
qemu-xen: build adjustments to support out-of-tree builds
QEMU by itself can be built outside of its source directory. With the
qemu repository being separate from the hypervisor/tools one it seems
to make sense to make use of this feature, but doing so requires a
couple of adjustments to the Xen changes to it. Basically, if
CONFIG_QEMU is found to indicate an existing directory, this directory
will be used rather than cloning the git repo into the build tree.
[ This changeset is the xen-unstable part of the patch but also
includes the QEMU_TAG update to pull in the qemu part. -iwj ]
Signed-off-by: Jan Beulich <jbeulich@novell.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Jan Beulich [Fri, 19 Nov 2010 18:39:33 +0000 (18:39 +0000)]
stubdom: allow building with read-only sources
Signed-off-by: Jan Beulich <jbeulich@novell.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Andre Przywara [Fri, 19 Nov 2010 18:15:14 +0000 (18:15 +0000)]
libxc: fix tracing (broken with hypercall buffers)
the attached patch makes Xen tracing work again, after the introduction
of the hypercall buffers broke it. Just a missing line.
Thanks to Uwe Dannowski for reporting this.
Signed-off-by: Andre Przywara <andre.przywara@amd.com> Acked-by: Ian Campbell <ian.campbell@citrix.com> Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Keir Fraser [Fri, 19 Nov 2010 13:45:08 +0000 (13:45 +0000)]
Introduce PHYSDEVOP_get_free_pirq
Introduce a new physdev_op called PHYSDEVOP_get_free_pirq to allow a
guest to get a free pirq number from Xen; the hypervisor would keep
that pirq free for the guest to use in a mapping.
Keir Fraser [Fri, 19 Nov 2010 13:43:24 +0000 (13:43 +0000)]
Interrupt remapping to PIRQs in HVM guests
This patch allows HVM guests to remap interrupts and MSIs into pirqs;
once the mapping is in place the guest will receive the interrupt (or
the MSI) as an event. The interrupt to be remapped can either be an
interrupt of an emulated device or an interrupt of a passthrough
device and we keep track of that.
Keir Fraser [Fri, 19 Nov 2010 13:24:00 +0000 (13:24 +0000)]
x86-64: one more adjustment for Fam10 MMCONF enabling
The BASE_VALID() macro needs adjustment to match the other changes
done to the original Linux code (which are all queued to be merged
into Linus' tree), plus per Hypertransport specification the range
0xff00000000-0xffffffffff also needs to be excluded.
Keir Fraser [Fri, 19 Nov 2010 13:21:09 +0000 (13:21 +0000)]
x86/mm: Allocate log-dirty bitmaps from shadow/HAP memory.
Move the p2m alloc and free functions back into the per-domain paging
assistance structure and allow them to be called from the log-dirty
code. This makes it less likely that log-dirty code will run out of
memory populating the log-dirty bitmap.
Keir Fraser [Thu, 18 Nov 2010 12:28:31 +0000 (12:28 +0000)]
x86 hvm: Refuse to perform __hvm_copy() work in atomic context.
Soon we will properly handle paged out memory in this function by
sleeping in hypervisor context. This will require that all callers can
sleep.
If this check is too strong, we can reduce it to only applying to
guests with paging enabled (which also currently implies only guests
using Intel EPT). However my brief testing seems to indicate it works
okay.
Keir Fraser [Thu, 18 Nov 2010 12:26:27 +0000 (12:26 +0000)]
rcu_lock(current->domain) does not need to disable preemption.
If the guest sleeps in hypervisor context, it should not be destroyed
until execution reaches a safe point (i.e., guest context). This is
not implemented yet. :-) But the next patch will rely on it, to allow
an HVM guest to execute hypercalls that indirectly invoke __hvm_copy()
within an rcu_lock_current_domain() region.
Keir Fraser [Wed, 17 Nov 2010 17:28:17 +0000 (17:28 +0000)]
x86/hvm/pmtimer: improving scalability of virtual time update
Mitigate the heavy contention on handle_pmt_io when running a HVM
configured with many cores (e.g., 32 cores). As the virtual time of a
domain must be fresh, there should be someone updating it,
periodically. But it is not necessary to let a VCPU update the virtual
time when another one has been updating it. Thus the update can be
skipped when the VCPU finds someone else is updating the virtual time.
So every time a VCPU invoke handle_pmt_io to update the current
domain's virtual time, it will first try to acquire the pmtimer lock.
If it succeeds, it will update the virtual time. Otherwise, it can
skip the update, waits for the pmtimer lock holder to finish updating
the virtual time and returns the updated time.
Signed-off-by: Xiang Song <xiangsong@fudan.edu.cn> Signed-off-by: Keir Fraser <keir@xen.org>
projects / xen.git / log