Remove the videoram from the memory target of the domains consistently,
leave the total amount of memory in maxmem.
[fixed up for conflicts with libxl__ naming policy changes -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Free the needed amount of memory before proceeding with the domain
build.
Use a filelock to prevent other xl instances from conflicting during
this operation.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Introduce a OS dependent global variable XEN_LOCK_DIR that points at the
system directory that contains lock files.
[fixed up for conflict in Config.mk -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Add a global configuration file: /etc/xen/xl.conf; the only option
currently parsed is autoballoon that is 1 by default.
[fixed up for conflicts with libxl__ naming policy changes -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Introduce a minimum value for the memory assigned to dom0.
[fixed up for conflicts with libxl__ naming policy changes -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
libxl_domain_need_memory: calculate how much memory a domain needs in
order to be built and start correctly.
libxl_get_free_memory: calculate the total free memory in the system.
libxl_wait_for_free_memory: wait for a certain amount of memory to
become free in the system.
libxl_wait_for_memory_target: wait for a domain to reach its memory
target.
[fixed up for conflicts with libxl__ naming policy changes -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Introduce a relative parameter to flag that target_memkb is a relative
amount of memory. The first time we are reading/writing dom0 memory
target, fill the informations in xenstore if they are missing.
Introduce libxl_get_memory_target.
[fixed up for conflicts with libxl__ naming policy changes -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
static-max represents the maximum memory the domain can ever have,
therefore it shouldn't be changed after domain creation.
[adjusted to cope with code movement since patch series written -iwj]
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
currently xcinfo2xlinfo reads tot_pages and uses that data to calculate
max_memkb, while tot_pages is the memory currently used by the domain
and max_pages is the theoretical maximum.
This patch calculates max_memkb from max_pages and introduces
current_memkb to store the current memory usage.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Mfns for PV domains were not properly checked, potentially
allowing a buggy or malicious PV guest to crash Xen. Also,
use get_page/put_page to claim a reference to the pages
so they can't disappear out from under tmem's feet.
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com>
It does not work reliably for a couple of reasons:
(1) page_lock() fails if a page is !PGT_validated, and a page can
remain in that state for unbounded time.
(2) in the kernel-side race that motivated this patch, pgd_pin() can
lose to vmalloc_sync_all() -- pgd_pin() can try to chaneg a pmd page's
type to l2_pagetable while
vmalloc_sync_all()->set_pmd()->do_mmu_update() has it temporarily
pinned as writable. This is hard to fix on the Xen side.
Hence I give up on this approach, revert the patch, and settle for
kernel-side patching only.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
The provided sample caml stubdom needs to use the ocaml cross
compiler, but if it hasn't been downloaded and compiled, then it won't
find it and the build will fail. This makes caml-stubdom depend on
the cross compiler, so it will automatically get built before the
stubdom.
Signed-off-by: Patrick Colp <pjcolp@cs.ubc.ca>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
xl_cfg_get_list can cope with the option not being a list, but
it always reports a warning in this case. Introduce a parameter
to omit this warning if needed. Future code can then decide what
to do with this information.
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Make Config.mk's definition of buildmakevars2file and
tools/libxl/Makefile's rule for _libxl_paths.h generate their files to
temporary files and only rename them into place if the intended file
has changed. This gets rid of unnecessarily compilation and link
steps.
Add *.tmp to .hgignore.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Keir Fraser <keir.fraser@citrix.com>
While earlier this year some work was done to eliminate libxc's
creation of nodes under /dev (as in some cases it actually got
things wrong), I just found that tapdisk2 does exactly the same.
While I'm not currently aiming at removing this code, I want to
at least suggest some adjustments to the provided rules file.
First of all, the present rule
KERNEL=="blktap[0-9]*", NAME="xen/%k"
matches both blktap1's devices and blktap2's ring devices,
resulting in whoever comes last replacing what was there
before (e.g. a tap2:aio: attach will replace bltap1's
/dev/blktap0 [i.e. the main control device] with the ring device
of the new virtual disk). Therefore we should add a subsystem
qualifier there.
Second, at some udev versions default to using 0660 as the
permissions on nodes it creates. The kernel (with devtmpfs)
defaults to 0600, and hence I'd suggest to also make this
explicit in the rules.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Boris Derzhavets <bderzhavets@yahoo.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
In b59f87f56b1e, the defaults for some of the directory config
variables implicitly changed. In this commit we arrange for the
defaults to be auto-detected (with the same logic as the run-time
scripts use).
This will make the build more likely to produce answers which the
builder intends. (And, in particular, it should un-break the
automated tests, which didn't set these variables explicitly and
therefore got broken by b59f87f56b1e.)
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
This can really happen -- all our PV Linux kernels have a race
between vmalloc_sync_all() and pgdir pinning/unpinning. The former is
protected by pgd_lock while the latter by mm->page_table_lock. Hence
they can happen concurrently, and vmalloc_sync_all() can attempt to
set_pmd() on a page directory which is in the process of being
pinned. This can confuse the hypervisor which may see a type change,
and hence fail do_mmu_update(). Until this patch. :-)
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
While removing the casts on the last arguments to __trace_var() I
noticed the bogus addition of 1 here.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
These structures are used by Xen, and hence guests must not be able
to fiddle with them.
qemu-dm currently plays with the MSI-X table, requiring Dom0 to
still have write access. This is broken (explicitly allowing the guest
write access to the mask bit) and should be fixed in qemu-dm, at which
time Dom0 won't need any special casing anymore.
The changes are made under the assumption that p2m_mmio_direct will
only ever be used for order 0 pages.
An open question is whether dealing with pv guests (including the
IOMMU-less case) is necessary, as handling mappings a domain may
already have in place at the time the first interrupt gets set up
would require scanning all of the guest's L1 page table pages.
Currently a hole still remains allowing PV guests to map these ranges
before actually setting up any MSI-X vector for a device.
An alternative would be to determine and insert the address ranges
earlier into mmio_ro_ranges, but that would require a hook in the
PCI config space writes, which is particularly problematic in case
MMCONFIG accesses are being used.
A second alternative would be to require Dom0 to report all devices
(or at least all MSI-X capable ones) regardless of whether they would
be used by that domain, and do so after resources got determined/
assigned for them (i.e. a second notification later than the one
currently happening from the PCI bus scan would be needed).
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Acked-by: Jiang, Yunhong <yunhong.jiang@intel.com>
The credit scheduler ties to keep work balanced, even on a mostly idle
system. Unfortunately, if you have one VM burning cpu and another VM
idle, the effect is that the busy VM will flip back and forth between
sockets.
This patch addresses this, by only migrating to a different socket if
the number of idle processors is twice that of the socket the vcpu is
currently on.
This will only affect mostly-idle systems; as the system becomes more
busy, other load-balancing code will come into effect.
Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>
Introduce CONFIG_LEAF_DIR, SUBSYS_DIR, INITD_DIR variables to cope
better with variation between distributions.
See docs/misc/distro_mapping.txt for details.
Signed-off-by: Bruce Edge <bruce.edge@gmail.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Building in a tree with read-only source files failed because the flex/
bison generated files are left in place.
All of the source files (instead of all of the object files) were
depending on libxl.h (in particular forcing the needless re-generation
of the flex/bison generated files).
For _libxl_types.[ch] a single non-pattern rule was used. For non-
pattern rules with multiple targets, other than for pattern ones, the
body gets executed once for each target, causing problems when building
with -j<n>, <n> >= 2. Convert this rule to a pattern one.
The flex/bison generated headers pointlessly depended on the
corresponding source files. Use multiple-target pattern rules instead.
It would also have been nice if the minimal flex/bison versions needed
to successfully build libxl were documented: neither flex 2.5.31 nor
bison 2.1 (as found e.g. on SLE10) is sufficient.
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>
Several seconds of backward time drift per minute can be seen on a
RHEL6 HVM guest by switching the clocksource to 'acpi_pm' and then
running gettimeofday() in a loop. This is due to the accumulation
of small inaccuracies that are caused by shifting out the lower 32
bits when pmt_update_time() computes 'tmr_val'.
The patch makes sure that the lower 32 bits of the computed value
are not lost. They are saved in a new field 'not_accounted' in the
PMTState structure and are accounted the next time pmt_update_time()
is called.
From: Ulrich Obergfell <uobergfe@redhat.com>
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
Add a cpuid parameter into libxl_domain_build_info and use
it's content while setting up the domain. This is a only paving the way,
the real functionality is implemented in the later patches.
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>
At least one of my compilers complains about the broken const
qualifier in the fork_exec_bootloader() call (despite the cast).
Fix this properly by propagating the const attribute down the
call stack.
Signed-off-by: Andre Przywara <andre.przywara@amd.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Introduce python binding for libxl. The binding is not yet complete but
serveral methods are implemented and tested. Those which are implemented
provide examples of the two or three basic patterns that most future
methods should follow.
Over 5,000 lines of boilerplate is automatically generated to wrap and
export all relevant libxl structure definitions. There are a few places
where such code cannot be fully auto-generated and special hooks are
declared and stubbed where, for example, conversion between
libxl_file_reference and a python file object is required.
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Firstly remove an anonymous union in libxl_device_pci structure which
was making auto-generating language bindings more complicated than
necessary and exporting random bits of low level ABI that libxl that
would rather hide anyway. There is a corresponding (untested) change to
the ocaml binding which maintains previous ml API.
Secondly make the libxl_file_reference type a Builtin. This is a
'semantic
correctness' issue in that libxl ABI/API won't change. But it makes it
so that when the IDL is used to generate language bindings that a
file_reference type is not exported.
Also implement a Numeric type which all integers are derived from. Make
sure a boolean signed/unsigned attribute is set accordingly. This is
required to allow language bindings to correctly handle the sign bit in
environments with arbitrarily long integers.
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
There is an errata in some of Intel processors.
AAJ72. EOI Transaction May Not be Sent if Software Enters Core C6
During an Interrupt Service Routine
If core C6 is entered after the start of an interrupt service routine
but before a write to the APIC EOI register, the core may not send an
EOI transaction (if needed) and further interrupts from the same
priority level or lower may be blocked.
This patch fix this issue, by checking if ISR is pending before enter
deep Cx state. If so, it would use power->safe_state instead of deep
Cx state to prevent the above issue happen.
Signed-off-by: Sheng Yang <sheng@linux.intel.com>
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
Somewhere between 3.2 and 3.3 this got dropped (as I noticed on a
[rarely used] box that didn't crash on me since then).
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Stefano Stabellini
Ian Jackson
Keir Fraser
Patrick Colp
Jan Beulich
Andre Przywara
Ian Campbell
Juergen Gross
Jun Zhu
Gianni Tedesco