It's not needed unless you are including headers in the current
directory _and_ you can't use "" instead of <> because the headers are
to be installed.
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>
Several places include it in CFLAGS even though it is never defined. A
few others use it as nothing more than a gathering point for
CFLAGS. Get rid of it.
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>
As well as the indirect dependency (via libxl) the xl binary also uses
libxenctrl directory (for xtl_createlogger_stdiostream etc) and
therefore must link against the library directly too.
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>
While c/s 22964:f71212f712fd and 23051:93c864c16ab1 fixed issues with
CPU onlining, they introduced a problem with resume: mcheck_init() is
also being called on that path, and hence checking whether it's
running on CPU 0, which is generally not a really good thing, is
particularly inappropriate here.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
First, $(shell ...) getting executed before the containing command gets
invoked doesn't have the intended effect when the inner command wants
variables of the outer command expanded. This needs to be done by the
shell executing the entire command.
Second, pattern rules for intermediate files with their targets not
mentioned explicitly result in them getting deleted. Since the .*.o.d
files get generated on the first make run and included for the second,
the second run sees them mentioned explicitly and re-builds them (and
the final *.init.o files), which is particularly annoying if the second
build is actually an install (as root) while normal builds get run as
ordinary user: Some of the output files will suddenly be root owned,
possibly causing permission issues on a later rebuild.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
A fair few things under tools link against libraries which they don't
even use.
Most of this appears to come from copy-and-pasting previous Makefile
snippets and cargo-culting plus the tendency to define global $(LIBS)
even for Makefiles which build multiple separate utilities or
libraries.
Identified by comparing a build with --as-needed to one without by
looking at the NEEDED header of all ELF objects.
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>
Tested by forcing --as-needed via tools/Rules.mk but this is included
since the intention is simply to support diustros which default to
--as-needed, not to enable it everywhere.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
In particular if binary A uses libB and libB uses libC entirely
internally then A does not need to link against libC only libB.
However when linking binary A the linker does need to have visibility
of the libraries which libB links against (libC in this example). For
out of tree uses this is achieved without fuss due because the
libraries are installed in a standard path. However in the case of
in-tree users the linker needs a hint in the form of the -rpath-link
option. Therefore a new class of build variable, $(SHLIB_FOO), is
introduced which includes the linker options needed to link against a
library which uses libFOO. The intention is that $(LDLIBS_bar) will
include the $(SHLIB_foo)s which it uses where necessary rather
requiring that users are aware of this.
For the python extensions this change appears particularly large since
previously each of python bindings were linked against the union of
all possible libraries used by all bindings instead of just what they
individually needed.
This change removes a dependency on libdl.so from nearly everything
in the system, only libxenctrl actually uses it.
In the context of xl/libxl the intention of libxl is to remove any
need for a user of libxl to know about libxenstore or libxenctrl,
however in the current build it is xl which links against those
libraries rather than libxl (which only links against libc). After
this change libxl correctly depends on the libraries it uses and xl
does not depend on libraries which it is not support to be required to
know about. Note that xl does depend on libxenctrl.so since it uses
xtl_* directly.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Clearly for the adjusted BUG_ON()s to not yield false positives
num_hpets_used (rather than num_chs_used, as done mistakenly in
23042:599ceb5b0a9b) must be incremented before setting up an IRQ (and
decremented back when the setup failed). To avoid further confusion,
just eliminate the local variable altogether.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
c/s 22964 fixed a mce memory leaks bug which may trigger xen crash
when cpu online. However, there are 2 mce memory leaks: 1 at mce
level (arch independent), 1 at mce_intel level (arch dependent). At
c/s 22964, it free both leaks at mce_intel level, which would has
problem under AMD arch.
This patch fix this issue.
It alloc/free poll_bankmask (arch independent) at mce level,=20 and
add a notifier block at mce level to avoid xmalloc risk when irq
disable.
With this patch, both Intel and AMD mce works OK in a clean way.
Signed-off-by: Liu, Jinsong <jinsong.liu@intel.com>
Allocate tracebuffers dynamically, based on the requested buffer size.
Calculate t_info_size from requested t_buf size.
Fix allocation failure path, free pages outside the spinlock.
Remove casts for rawbuf, it can be a void pointer since no math is
done.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Acked-by: George Dunlap <george.dunlap@eu.citrix.com>
Otherwise make can search the path relative to certain standard paths
such as /usr/include (e.g., the line '-include $(XEN_ROOT)/.config' in
Config.mk suffers from this).
Signed-off-by: Keir Fraser <keir@xen.org>
We can globally export it from xen/Makefile instead, as all hypervisor
builds have this Makefile at their root.
Signed-off-by: Keir Fraser <keir@xen.org>
This improves compatibility with xm style config files since xend does
this (see NetifController.getDeviceDetails() in
tools/python/xen/xend/server/netif.py) and
tools/hotplug/Linux/vif-setup expects an absolute path to the script.
This patch supports both absolute and relative paths in the
configuration. Also, if somehow nic->script==NULL, don't add write
the entry to xenstore at all (rather than crashing due to the new
check for absolute vs relative path).
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
value is read when deciding whether to flush the VTd tables. The
effect is minor: in situations where the p2m entry is changed but the
mfn is the same, the VTd tables will be unnecessarily flushed.
old_entry is left untouched in the second case, since having a present
old_entry will cause the an unnecessary check to be taken at the end
of the function.
Spotted-by: Zhang, Peng Fei <zpfalpc23@gmail.com>
Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>
Acked-by: Tim Deegan <Tim.Deegan@citrix.com>
Also ensure that static and hidden functions use the libxl__ prefix
not just libxl_ (in the case of static functions only when they use a
libxl prefix to start with).
This follows the policy described in libxl.h "libxl memory
management".
Based on a manual audit of:
grep ^static tools/libxl/libxl*.[ch]| grep libxl_ctx
grep libxl__ tools/libxl/*.h| grep libxl_ctx
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
qemu is used as a fallback for DISK_BACKEND_TAP if no blktap is
available but if blktap is available, or for DISK_BACKEND_PHY, we
don't need a qemu process.
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>
Only call SCHED_OP(pick_cpu) if a previously picked cpu is not
suitable on the current iteration of the retry loop.
Signed-off-by: Juergen Gross <juergen.gross@ts.fujitsu.com>
Signed-off-by: Keir Fraser <keir@xen.org>
Clearly for the adjusted BUG_ON()s to not yield false positives
num_chs_used must be incremented before setting up an IRQ (and
decremented back when the setup failed).
Signed-off-by: Jan Beulich <jbeulich@novell.com>
At least to me, using ?: instead of the (a && ...) || (!a && ...)
construct is far easier to grok with a single look.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
For the bias to be really meaningful, it should be updated only when
the CPU selected will indeed be returned (and hence used for placing
the vCPU in question).
Signed-off-by: Jan Beulich <jbeulich@novell.com>
This fixes a fairly blatant bug I introduced in c/s 20377:cff23354d026
- I wonder how this went unnoticed for so long.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
This permits suspend/resume to work with 32bit dom0/tools when system
memory extends beyond 160GB (and up to 1TB).
AFAICT the limit to MACH2PHYS_COMPAT_NR_ENTRIES is redundant since
that refers to a limit in 32bit guest compat mappings under 64bit
hypervisors, not userspace where there may be gigabytes of useful
virtual space available for this.
Suggested-by: Ian Campbell <Ian.Campbell@eu.citrix.com>
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
c/s 21811:12f0618400de broke this by not properly initializing struct
cpuinfo_x86's x86_num_siblings member (other than Linux, where this is
a global variable, it is being maintained per CPU by Xen).
Hyper-threaded CPUs with CPUID leaf 0xb present had therefore all
cpu_sibling_map-s with just a single bit set, thus improperly feeding
the scheduler.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
cpuid results are possible to be changed now. For example, changing
CR4.OSXSAVE bit or setting MSR XCR_XFEATURE_ENABLED_MASK may change
XSAVE related cpuid leave return values.
The volatile prefix is required to avoid the second cpuid calls
following some possible changing operations being optimized in
incorrect way by compiler.
The sample bug is in xsave_init while debug=3Dn. The second call to
cpuid_count() may be optimized and lead to a BUG_ON case while compare
xsave_cntxt_size with ebx.
Signed-off-by: Wei Gang <gang.wei@intel.com>
Some compiler versions may choose to not inline certain functions, but
the check introduced in c/s 23003:768269c43914 wants .text to be
empty.
Also make sure an eventual error gets properly propagated even on the
first section of an object (.text typically being the first one), and
cover a broader set of sections.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
'unsigned int' is better suited as an array index on x86-64.
'u32' produces better code than 'unsigned long' on x86-64, so use the
former for storing 32-bit values read from the hardware.
this_cpu() uses an implicit smp_processor_id(), and hence using
per_cpu() when the result of smp_processor_id() is already available
is more efficient.
Fold one case of cpu_isset()+cpu_clear() into cpu_test_and_clear().
Drop the unused return value of evt_do_broadcast().
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Acked-by: Wei Gang <gang.wei@intel.com>
Typically there are far less than 32 counters available, and hence
there's no use in wasting the memory on (almost) every system.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Acked-by: Wei Gang <gang.wei@intel.com>
- separate init and resume code paths (so that the [larger] init parts
can go init .init.* sections)
- drop the separate legacy_hpet_event object, as we can easily re-use
the first slot of hpet_events[] for that purpose (the whole array is
otherwise unused when the legacy code is being used)
- use section placement attributes where reasonable
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Acked-by: Wei Gang <gang.wei@intel.com>
At least the legacy path can enter its interrupt handler callout while
initialization is still in progress - that handler checks whether
->event_handler is non-NULL, and hence all other initialization must
happen before setting this field.
Do the same to the MSI initialization just in case (and to keep the
code in sync).
Signed-off-by: Jan Beulich <jbeulich@novell.com>
This will allow us to use OS specific interfaces to ensure that the
allocated memory is safe for use as a hypercall buffer in the future.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Ian Campbell
Keir Fraser
Jan Beulich
Liu, Jinsong
Olaf Hering
Ian Jackson
Tim Deegan
Juergen Gross
Gianni Tedesco
Wei Gang
Stefano Stabellini