Other than when they're located in I/O port space, accessing them when
in MMIO space (currently) implies usage of some sort of global lock: In
-unstable this would be due to the use of vmap(), is older trees the
necessary locking was introduced by 2ee9cbf9 ("ACPI: fix
acpi_os_map_memory()"). This contention was observed to result in Dom0
kernel soft lockups during the loading of the ACPI processor driver
there on systems with very many CPU cores.
There are a couple of things being done for this:
- re-order elements of an if() condition so that the register access
only happens when we really need it
- turn off arbitration disabling only when the first CPU leaves C3
(paralleling how arbitration disabling gets turned on)
- only set the (global) bus master reload flag once (when the first
target CPU gets processed)
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
These are known to not support the feature, so we can save ourselves
from emitting the resulting #GP fault recovery related message (which
might worry people looking at the logs).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Liu Jinsong <jinsong.liu@intel.com>
Acked-by: Keir Fraser <keir@xen.org>
The function issues a newline itself.
Also correct two slightly misplaced __initdata annotations.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Keir Fraser <keir@xen.org>
Certain UEFI implementations reserve all memory below 1Mb at boot time,
making it impossible to properly allocate the chunk necessary for the
trampoline. Fall back to simply grabbing a chunk from EfiBootServices*
regions immediately prior to calling ExitBootServices().
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
The commit 58afa7ef "x86/hvm: Run the RTC periodic timer on a
consistent time series" aligns the RTC periodic timer to the VM's boot time.
However, it's aligned later again to the system time in create_periodic_time()
with vpt_align=1. The next tick might be skipped.
Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
"$N will be used for broadcast" is ambiguous between "$N timers" or "timer
$N", particuarly when N is 0.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Tim Deegan <tim@xen.org>
The domain destruction code, removes a domain from its cpupool
before attempting to destroy its scheduler information. Since
the scheduler framework uses the domain's cpupool information
to decide on which scheduler ops to use, this results in the
the wrong scheduler's destroy domain function being called
when the cpupool scheduler and the initial scheduler are
different.
Correct this by destroying the domain's scheduling information
before removing it from the pool.
Signed-off-by: Nathan Studer <nate.studer@dornerworks.com>
Reviewed-by: Juergen Gross <juergen.gross@ts.fujitsu.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: George Dunlap <george.dunlap@eu.citrix.com>
Acked-by: Keir Fraser <keir@xen.org>
When the FIFO-based ABI is in use, if an event is bound when the
corresponding event array page is missing any attempt to set the event
pending will lose the event (because there is nowhere to write the
pending state).
This wasn't initially considered an issue because guests were expected
to only bind events once they had expanded the event array, however:
1. A domain may start with events already bound (by the toolstack).
2. The guest does not know what the port number will be until the
event is bound (it doesn't know how many already bound events there
are), so it does not know how many event array pages are required.
This makes it difficult to expand in advanced (the current Linux
implementation expands after binding for example).
To prevent pending events from being lost because there is no array
page, temporarily store the pending state in evtchn->pending. When an
array page is added, use this state to set the port as pending.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
This patch flush cache when vmentry back to UC guest, to prevent
cache polluted by hypervisor access guest memory during UC mode.
The elegant way to do this is, simply add wbinvd just before vmentry.
However, currently wbinvd before vmentry will mysteriously trigger
lapic timer interrupt storm, hung booting stage for 10s ~ 60s. We still
didn't dig out the root cause of interrupt storm, so currently this
patch add flag indicating hypervisor access UC guest memory to prevent
interrupt storm -- though it still leaves aspects un-addressed, i.e.
speculative reads, and multi-vCPU issues, etc.
Whenever the interrupt storm got root caused and fixed, the protection
flag can be removed -- that would be final clean and elegant approach
dealing with cache flushing before vmentry.
This is CVE-2013-2212 / XSA-60.
Suggested-by: Jan Beulich <jbeulich@suse.com>
Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Liu Jinsong <jinsong.liu@intel.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Jun Nakajima <jun.nakajima@intel.com>
This patch solves XSA-60 security hole:
1. For guest w/o VT-d, and for guest with VT-d but snooped, Xen need
do nothing, since hardware snoop mechanism has ensured cache coherency.
2. For guest with VT-d but non-snooped, cache coherency can not be
guaranteed by h/w snoop, therefore it need emulate UC type to guest:
2.1). if it works w/ Intel EPT, set guest IA32_PAT fields as UC so that
guest memory type are all UC.
2.2). if it works w/ shadow, drop all shadows so that any new ones would
be created on demand w/ UC.
This patch also fix a bug of shadow cr0.cd setting. Current shadow has a
small window between cache flush and TLB invalidation, resulting in possilbe
cache pollution. This patch pause vcpus so that no vcpus context involved
into the window.
This is CVE-2013-2212 / XSA-60.
Signed-off-by: Liu Jinsong <jinsong.liu@intel.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jun Nakajima <jun.nakajima@intel.com>
Acked-by: Keir Fraser <keir@xen.org>
XSA-60 security hole comes from the problematic vmx_set_uc_mode.
This patch remove vmx_set_uc_mode logic, which will be replaced by
PAT approach at later patch.
This is CVE-2013-2212 / XSA-60.
Signed-off-by: Liu Jinsong <jinsong.liu@intel.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Acked-by: Jun Nakajima <jun.nakajima@intel.com>
Recently Oracle developers found a Xen security issue as DOS affecting,
named as XSA-60. Please refer http://xenbits.xen.org/xsa/advisory-60.html
Basically it involves how to handle guest cr0.cd setting, which under
some environment it consumes much time resulting in DOS-like behavior.
This is a preparing patch for fixing XSA-60. Later patch will fix XSA-60
via PAT under Intel EPT case, which depends on cpu_has_vmx_pat.
This is CVE-2013-2212 / XSA-60.
Signed-off-by: Liu Jinsong <jinsong.liu@intel.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Acked-by: Jun Nakajima <jun.nakajima@intel.com>
These viridian MSRs are read-only sources of the TSC and APIC frequency (in
units of Hz)
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
This viridian MSR is a read-only source of time (in units of 100ns) since the
domain started.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
This allow building of XEN from source on a system without internet
access using either the environment or .config
Do the same with XEN_EXTFILES_URL, QEMU_REMOTE, IPXE_GIT_URL, and
IPXE_TARBALL_URL as QEMU_UPSTREAM_URL, SEABIOS_UPSTREAM_URL, etc.
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
libxl currently refuses to execute hotplug scripts if the backend
domid of a device is different than LIBXL_TOOLSTACK_DOMID. This will
prevent libxl from executing hotplug scripts when running on a domain
different than LIBXL_TOOLSTACK_DOMID, we should instead check if
backend_domid is different than current domid.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
This patch writes Dom0 domid on xenstore (like it's done for other
guests), and adds a libxl helper function to fetch that domid from
xenstore.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
[ ijc -- dropped xencommons hunk, same change was made independently
in 02ebea7768 ]
These are in the same page as the UART which is used as the Xen console. We are
not currently smart enough to avoid passing them through to the guest,
accidentally giving the guest access to the Xen console UART.
we blacklist them all, if necessary in the future we can split the list into
two halves and make a per platform decision about which half should be
blacklisted on a given platform depending on which UART is wired up as the
console.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Julien Grall <julien.grall@linaro.org>
Namely, don't overwrite all the other flags when twiddling O_NONBLOCK,
and add basic error handling.
Coverity-ID: 1055041
Signed-off-by: Matthew Daley <mattjd@gmail.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Reviewed-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Change the qemu parameters for upstream qemu vgas to the
new ones (-device), introduced some years ago.
Signed-off-by: Fabio Fantoni <fabio.fantoni@m2r.biz>
VMs using Cirrus graphics have always had 4 MB of video RAM on XenServer, while
libxl currently does not allow values less than 8 MB.
Signed-off-by: Rob Hoes <rob.hoes@citrix.com>
Tested-by: Fabio Fantoni <fabio.fantoni@m2r.biz>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
The ARINC653 scheduler alloc_vdata function does not add the
idle cpu to its internal vcpu_list, but when the free_vdata
function is called, the scheduler attempted to remove the vcpu
from its internal vcpu_list, regardless of whether or not
the vcpu was the idle vcpu. Since the idle vcpu's list field
was never initialized, a NULL pointer was passed to list_del.
When using cpupools, this resulted in a crash when moving a cpu
from an arinc653 scheduler pool.
Signed-off-by: Nathan Studer <nate.studer@dornerworks.com>
Acked-by: George Dunlap <george.dunlap@eu.citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
The ARINC653 scheduler was directly assigning and manipulating
the sched_priv field of a vcpu in its alloc_vdata function.
When creating a cpu pool, this resulted in the corruption
of the sched_priv field of the vcpu, which was then passed
to the initial scheduler's free_vdata function with
disastrous results.
Signed-off-by: Nathan Studer <nate.studer@dornerworks.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: George Dunlap <george.dunlap@eu.citrix.com>
This is really a boolean_param() hidden inside a hand-coded attempt to
replicate boolean_param(), which misses the 'no-' prefix semantics
expected with Xen boolean parameters.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Memory blocks outside of the always visible 1:1 mapping range get
passed to the allocator separately (once enough other setup was done).
Skipping non-RAM regions, however, was forgotten in adc5afbf ("x86:
support up to 16Tb").
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
Other than for the legacy APIC, the x2APIC MADT entries have valid
ranges possibly extending beyond what our internal arrays can handle,
and hence we need to guard ourselves against corrupting memory here.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Keir Fraser <keir@xen.org>
Coverity ID 1087189
Correct a lock order reversal between a domains page allocation and grant
table locks.
This is CVE-2013-4494 / XSA-73.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Consolidate error handling.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Keir Fraser <keir@xen.org>
Tested-by: Matthew Daley <mattjd@gmail.com>
When calculating the virtual address of the UART,
add the page offset of the UART to the mapped address.
Create define EARLY_UART_VIRTUAL_ADDRESS and use instead
of FIXMAP_ADDR(FIXMAP_CONSOLE)
Signed-off-by: Jon Fraser <jfraser@broadcom.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
This error case can only be triggered by gzread returning 0 (and having
not read anything), so move it there.
Coverity-ID: 1056076
Signed-off-by: Matthew Daley <mattjd@gmail.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Jan Beulich
Matthew Daley
Kouya Shimura
Andrew Cooper
Nathan Studer
David Vrabel
Liu Jinsong
Paul Durrant
Don Slutz
Roger Pau Monne
Ian Campbell
Bamvor Jian Zhang
Fabio Fantoni
Rob Hoes
Jon Fraser