Fix a bug of xen-mceinj tool which used to test mce by software way.
Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
Committed-by: Keir Fraser <keir@xen.org>
PCI configuration access is allowed to any privileged domain
regardless of I/O port access restrictions; add XSM hooks for these
accesses.
Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Actions requiring IS_PRIV should also require some XSM access control
in order for XSM to be useful in confining multiple privileged
domains. Add XSM hooks for new hypercalls and sub-commands that are
under IS_PRIV but not currently under any access checks.
Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
The mmu_update hypercall can be used to manipulate the page tables of
a remote domain. Add a check for this in the XSM hook in addition to
the existing check on mapping pages of a remote domain.
Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
The log messages from dummy overrides appear on every boot with XSM
enabled, and are just noise when filling in the dummy_xsm_ops
structure.
Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
This information is useful when determining the cause of an AVC denial
caused by missing label on device memory or IRQs.
Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
oxenstored is an ocaml implementation of the xenstore daemon. It is faster,
more scalable and more reliable than the C xenstored.
In particular the transaction model in oxenstored does not involve taking a
complete copy of the database and aborting on any (even non-conflicting) other
change.
There is a paper on the design and implementation of oxenstored at
http://gazagnaire.org/pub/GH09.pdf which includes a performance evaluation and
comparison with the C daemon etc.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
The only cases where we might end up emulating fsincos (as any other
x87 operations without memory operands) are
- when a HVM guest is in real mode (not applicable on AMD)
- between two half page table updates in PAE mode (unlikely, and not
doing the emulation here does affect only performance, not
correctness)
- when a guest maliciously (or erroneously) modifies an (MMIO or page
table update) instruction under emulation (unspecified behavior)
Hence, in order to avoid the erratum to cause harm to the entire host,
don't emulate fsincos on the affected AMD CPU families.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
/var/lib/xen/qemu-save.<domid> is created when the stub domain is started
(connected to a stubdom console) and is used to save the device model state on
suspend/migrate but never cleaned up.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Otherwise at the default level we rarely log anything at all.
A completely empty log file is a good sign, but only if you know you are
looking in the right place...
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
First though:
- Move it to /etc/xen/oxenstored.conf.
- Use /var/run/xenstored.pid as default pid file
- Disable test-eagain "Randomly failed a transaction with EAGAIN. Used for
testing Xs user". Doesn't sound fun by default...
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Previous an unknown operation would be decoded as a Not_found exception which
would bubble all the way up to the try ... with surrounding the call to
main_loop where it would be logged and ignored.
This would leave the guest hanging waiting for a response to the invalid
request.
Instead introduce a specific "Invalid" operation. Higher level functionality,
such as Process.process_packet, already handles operations which are not
understood with an error reply due to the final wildcard entry in
Process.function_of_type but explicitly handle Invalid this way to make it
clear what is going on.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Remove hardcoded maximum size a microcode patch can have. This is
dynamic now.
The microcode patch for family15h can be larger than 2048 bytes and
gets silently truncated.
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Committed-by: Jan Beulich <jbeulich@suse.com>
When RIPV = EIPV = 0, it's a little bit tricky. It may be an asynchronic error, currently we have no way to precisely locate whether the error occur at guest or hypervisor.
To avoid handling error in wrong way, we treat it as unrecovered.
Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Committed-by: Jan Beulich <jbeulich@suse.com>
Currently Intel SDM add 2 kinds of MCE SRAR errors:
1). Data Load error, error code = 0x134
2). Instruction Fetch error, error code = 0x150
This patch add handler to these new SRAR errors.
It based on existed mce infrastructure, add code to handle SRAR specific error.
Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
Committed-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson.citrix.com>
Committed-by: Ian Jackson <ian.jackson.citrix.com>
It is unused.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson.citrix.com>
Committed-by: Ian Jackson <ian.jackson.citrix.com>
Since 23565:72eafe80ebc1 the xenstore entries for the stubdomain's PCI were
never created and therefore the stubdom ends up waiting forever for the devices
which it has been asked to insert to show up.
Since the stubdomain is already running when we call the libxl_device_pci_add
loop in do_domain_create we should treat it as if "starting == 0".
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson.citrix.com>
Committed-by: Ian Jackson <ian.jackson.citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
This blocks the main thread indefinitely if there is no PCI backend present in
xenstore.
Even in the case where there are passthrough devices configured libxl creates
the stubdom and waits for it to startup _before_ adding the backend. Since the
stub domains main thread is blocked before it can write the "running" state to
xenstore the toolstack eventually times out and kills everything.
There is already a separate pcifront thread which waits for the backend to
appear and calls init_pcifront at the appropriate time should a backend ever
appear.
Unfortunately I don't have any free test boxes with VT-d so I haven't been able
to test the cases where PCI deivces are passed through but I obviously have
tested that I can now start an HVM domain with stub qemu without PCI devices
passed through which I couldn't do before so this is an improvement. This stuff
is a bit like pushing the lump around the carpet :-/
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson.citrix.com>
Committed-by: Ian Jackson <ian.jackson.citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
When user wants to change entry in grub2 menu in pygrub, there's no
response in case of appending command line arguments ('a' key).
Append malfunction is caused by change of keyword used for kernel
record. Grub uses 'kernel' for line with linux kernel but grub2 uses
'linux' instead. This patch adds checking for both grub 1 and 2 keywords.
Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
When user wants to change entry in grub2 menu in pygrub, there
may be crash of pygrub in case of editing item ('e' key).
Crash on editing is caused longer entry list in case of grub2. As entry
window is 10 lines high, it can hold only 8 entries (2 lines for border).
Adding line outside of windows high causes crash. Patch add handling
for longer lists and scrolling through them.
Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
GNU make prior to 3.81 doesn't support $(realpath ...). This fixes a
regression introduced in 23368:0f670f5146c8 (the option tested via
cc-option-add got interpreted as the argument of the -I compiler
option, as its intended argument was blank, and hence the compiler was
falsely considered to support *any* option in the pciutils sub-tree).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
This extends create_irq() to take a node parameter, allowing the
resulting IRQ to have its destination set to a CPU on that node right
away, which is more natural than having to post-adjust this (and
get e.g. a new IRQ vector assigned despite a fresh one was just
obtained).
All other callers of create_irq() pass NUMA_NO_NODE for the time being.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
It has been hiding actual mistakes, and there are not too many changes
necessary to make things build without suppressing this warning.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
Two definitions (the first register and a macro to calculate the
register for a given bank) are sufficient per kind of register.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
With the recent hotplug changes to the Xen part of the microcode
loading, this allows the kernel driver to avoid unnecessary calls into
the hypervisor during pCPU hot-enabling: Knowing that the hypervisor
retains the data for already booted CPUs, only data for CPUs with a
different signature needs to be passed down. Since the microcode
loading code can be pretty verbose, avoiding to invoke it can make the
log much easier to look at in case of problems.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
... to indicate counting from the end of the modules list.
Suggested by Tim Deegan.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Keir Fraser <keir@xen.org>
Replace
libxl__gc gc = LIBXL_INIT_GC(ctx);
...
libxl__free_all(&gc);
with
GC_INIT(ctx);
...
GC_FREE;
throughout with a couple of perl runes.
We must then adjust uses of the resulting gc for pointerness, which is
mostly just replacing all occurrences of "&gc" with "gc". Also a
couple of unusual uses of LIBXL_INIT_GC needed to be fixed up by hand.
Here are those runes:
perl -i -pe 's/\Q libxl__gc gc = LIBXL_INIT_GC(ctx);/ GC_INIT(ctx);/' tools/libxl/*.c
perl -i -pe 's/\Q libxl__free_all(&gc);/ GC_FREE;/' tools/libxl/*.c
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
libxl_internal.h now #includes libxl.h and various system headers.
This
1. makes the order of header inclusion more predictable
2. explicitly allows libxl_internal.h to use objects defined in libxl.h
3. removes the need for individual files to include these headers
Also
- remove some unnecessary #includes of libxl_utils.h,
flexarray.h, etc. in some libxl*.c files,
- include libxl_osdeps.h at the top of libxl_internal.h
- add missing includes of libxl_osdeps.h to a couple of files
- change libxl.h to libxl_internal.h in a couple of files
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Provide some macros which are useful shorthands for use within libxl:
* GC_INIT to initialise a gc from a ctx and GC_FREE to free it
* CTX(gc) to give you back the ctx
* LIBXL_TAILQ_INSERT_SORTED for inserting things into sorted lists
These will be used by later patches.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
GCC and C99 allow declarations to be mixed with code. This is a good
idea because:
* It allows variables to be more often initialised as they are
declared, thus reducing the occurrence of uninitialised variable
errors.
* Certain alloca-like constructs (arrays allocated at runtime on the
stack) can more often be written without a spurious { } block.
Such blocks are confusing to read.
* It makes it easier to write and use macros which declare and
initialise formulaic variables and do other function setup code,
because there is no need to worry that such macros might be
incompatible with each other or have strict ordering constraints.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Liu, Jinsong