Xen bitmaps can be 4 rather than 8 byte aligned, so use the appropriate type.
Otherwise the compiler can generate unaligned 8 byte accesses and cause traps.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
In this patch, we add some checks and restrictions in the relevant
p2m paths for p2m_is_foreign.
Signed-off-by: Mukesh Rathor <mukesh.rathor@oracle.com>
Acked-by: Tim Deegan <tim@xen.org>
A subsequent patch exposes rangesets to secondary emulators, so to allow a
limit to be placed on the amount of xenheap that an emulator can cause to be
consumed, the function rangeset_limit() has been created to set the allowed
number of ranges in a rangeset. By default, there is no limit.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Also needed to fix vsnprintf() et al so it can be called with a NULL buf
(and zero size, of course).
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
This patch only creates the ioreq server when the legacy HVM parameters
are read (by an emulator).
A lock is introduced to protect access to the ioreq server by multiple
emulator/tool invocations should such an eventuality arise. The guest is
protected by creation of the ioreq server only being done whilst the
domain is paused.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Collect together data structures concerning device emulation together into
a new struct hvm_ioreq_server.
Code that deals with the shared and buffered ioreq pages is extracted from
functions such as hvm_domain_initialise, hvm_vcpu_initialise and do_hvm_op
and consolidated into a set of hvm_ioreq_server manipulation functions. The
lock in the hvm_ioreq_page served two different purposes and has been
replaced by separate locks in the hvm_ioreq_server structure.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
To simplify creation of the ioreq server abstraction in a subsequent patch,
this patch centralizes all use of the shared ioreq structure and the
buffered ioreq ring to the source module xen/arch/x86/hvm/hvm.c.
The patch moves an rmb() from inside hvm_io_assist() to hvm_do_resume()
because the former may now be passed a data structure on stack, in which
case the barrier is unnecessary.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
This patch tidies up various parts of the code that following patches move
around. If these modifications were combined with the code motion it would
be easy to miss them.
There's also some function renaming to reflect purpose and a single
whitespace fix.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
This patch changes construct_dom0() to boot in pvh mode:
- Make sure dom0 elf supports pvh mode.
- Call guest_physmap_add_page for pvh rather than simple p2m setting
- Map all non-RAM regions 1:1 upto the end region in e820 or 4GB which
ever is higher.
- Allocate p2m, copying calculation from toolstack.
- Allocate shared info page from the virtual space so that dom0 PT
can be updated. Then update p2m for it with the actual mfn.
- Since we build the page tables for pvh same as for pv, in
pvh_fixup_page_tables_for_hap we replace the mfns with pfns.
Signed-off-by: Mukesh Rathor <mukesh.rathor@oracle.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
After commit 44e24f8567 (x86: don't call generic_identify() redundantly)
the struct cpu_dev don't need the c_identify.
Signed-off-by: Yi Li <peteryili@tencent.com>
For auto-translated guests the p2m code will do the necessary TLB
flushes, so there's no need to perform any TLB flushes in generic
grant table code.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Modify the function's name to help eventual backports involving this
function, and in one case where this is trivially possible also stop
ignoring its return value.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
With proper permission (and, for the I/O port case, wrap-around) checks
added (note that for the I/O port case a count of zero is now being
disallowed, in line with I/O memory handling):
XEN_DOMCTL_irq_permission:
XEN_DOMCTL_ioport_permission:
Of both IRQs and I/O ports there is only a reasonably small amount, so
there's no excess resource consumption involved here. Additionally
they both have a specialized XSM hook associated.
XEN_DOMCTL_iomem_permission:
While this also has a specialized XSM hook associated (just like
XEN_DOMCTL_{irq,ioport}_permission), it's not clear whether it's
reasonable to expect XSM to restrict the number of ranges associated
with a domain via this hook (which is the main resource consumption
item here).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
This statement only works using 64-bit arithmetic for the main
63
counter never changing by more then 2 . (Which is a boundary
case that should not happen in my life time.)
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Tim Deegan <tim@xen.org>
Based on the software-developers-hpet-spec-1-0a.pdf, the comparator
for a periodic timer will change to the new value when it matches
the master clock. The current code here uses a very standard
rounding formula of "((x + y - 1) / y) * y". This is wrong because
in this case you need to go to the next comparator value when "x"
equals "y". Not when "x + 1" equals "y". In this case "y" is the
period and "x" is the master clock.
The code lines:
elapsed = hpet_read_maincounter(h, guest_time) +
period - 1 - comparator;
comparator += (elapsed / period) * period;
are what matter here.
Using some numbers to help show the issue:
hpet_read_maincounter(h, guest_time) = 130252
period = 62500
comparator : 130252
elapsed : 62499
elapsed/period : 0
comparator_delta : 0
new comparator : 130252
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
This is based on software-developers-hpet-spec-1-0a.pdf saying:
When the main counter value matches the value in the timer's
comparator register, an interrupt can be generated. The hardware
will then automatically increase the value in the compare register
by the last value written to that register.
When the overall enable is off (the main count is halted), none of
the compare registers should change.
The code lines:
elapsed = hpet_read_maincounter(h, guest_time) +
period - 1 - comparator;
comparator += (elapsed / period) * period;
are what matter here. They will always adjust comparator to be no
more then one period away.
Using some numbers to help show the issue:
hpet_read_maincounter(h, guest_time) = 67752
period = 62500
comparator = 255252 == 67752 + 3 * 62500
comparator : 255252
elapsed : -125001
elapsed/period : -2
comparator_delta : -125000
new comparator : 130252
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
The software-developers-hpet-spec-1-0a.pdf says that the comparator
starts as all 1's. Also make the hidden register comparator64 the same.
Since only the hidden register comparator64 is used by hpet_save, it
needs to start out with the right value.
A disabled hpet (like when a guest is starting), should start with
the value the spec says. Both the guest (via reading the
comparator) and an administrator using xen-hvmctx, will see all 0's
not all 1's.
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
This changes save data to consistent/expected values. It is not
technically required because hpet_get_comparator() will adjust from
any value to the correct value. And hpet_get_comparator() is
effectivly called in hpet_load via hpet_set_timer.
However it does look strange to people that the output from
xen-hvmctx for the comparator values do not change when the master
clock does.
The software-developers-hpet-spec-1-0a.pdf says that the comparator
will allways be greater than master clock for a periodic timer.
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
When the master clock is not enabled, mc64 has the right value.
Basicly do the same thing as hpet_read_maincounter().
Signed-off-by: Don Slutz <dslutz@verizon.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
In the code section after the comment:
/*
* Clamp period to reasonable min/max values:
* - minimum is 100us, same as timers controlled by vpt.c
* - maximum is to prevent overflow in time_after() calculations
*/
The current maximum limit actually allows "bad" values like 0 and 1.
This is because it uses a mask not a maximum.
Signed-off-by: Don Slutz <dslutz@verizon.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
The current code sets both. If setting the comparator also set
comparator64 (the hidden version).
Based on:
software-developers-hpet-spec-1-0a.pdf
A write call should only change comparator or period, not both.
Signed-off-by: Don Slutz <dslutz@verizon.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
This call is expensive and will cause extra time to pass.
The software-developers-hpet-spec-1-0a.pdf does not say how long it
takes after the main clock is enabled before the first change of the
master clock. Therefore multiple calls to guest_time_hpet(h) are
not needed. Since each timer is started by a loop, each ones start
time will change on the multple calls. In the real hardware, there
is not delta based on which timer.
Without this change it is possible for an HVM guest running linux to
get the message:
..MP-BIOS bug: 8254 timer not connected to IO-APIC
On the guest console(s); and the guest will panic.
Also Xen hypervisor console with be flooded with:
vioapic.c:352:d1 Unsupported delivery mode 7
vioapic.c:352:d1 Unsupported delivery mode 7
vioapic.c:352:d1 Unsupported delivery mode 7
Signed-off-by: Don Slutz <dslutz@verizon.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Tim Deegan <tim@xen.org>
Add the code at tools/tests/vhpet.
See comment in tools/tests/vhpet/main.c for details on running
either in a xen source tree or elsewhere.
A basic in source tree usage is:
make -C tools/tests/vhpet run
Does repro the bug:
..MP-BIOS bug: 8254 timer not connected to IO-APIC
The make file includes coping hpet.c and hpet.h from the source
tree. hpet.c is then modifed to remove all include file and add the
emul.h include file.
The manual test code has only a few automatic checks that output
messages to stderr:
1) Possible ..MP-BIOS bug: 8254 timer...
if 1st period is not <= the expected value
2) hpet_set_mode(%ld): T%d Error: Set ...
if read of comparator != write of comparator in
3) hpet_check_stopped(%ld): T%d Error: Set ...
if read != write
4) main(%ld): With clock stopped mc64 changed: ...
if hpet_save returns different master clock values when called
more then once.
It also generates a lot of output, which is why the sugested way to
use includes a redirect of stdout to a file.
Signed-off-by: Don Slutz <dslutz@verizon.com>
Acked-by: Tim Deegan <tim@xen.org>
getBridge's method of returning a result was a little confused:
allocating a result buffer but never using it.
Simplify by instead allowing a result buffer to be passed in and
modifying the single usage to match.
Signed-off-by: Matthew Daley <mattd@bugfuzz.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Commit 1438d36f ("xenstat: Fix buffer over-run with new_domains being
negative.") attempted to fix the handling of a negative error result
from xc_domain_getinfolist in xenstat_get_node. However, it forgot to
change the result variable from an unsigned type to a signed one.
Do so, allowing the error result to be handled properly.
Signed-off-by: Matthew Daley <mattd@bugfuzz.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Document the multi-queue feature in terms of XenStore keys to be written
by the backend and by the frontend.
Signed-off-by: Andrew J. Bennieston <andrew.bennieston@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
- some of the case statements are effectively or mostly special cases
of others, so there's no good reason not to share the code
- in the "get" function, a variable can be made case-wide instead of
having multiple instance of it (and those even with a pointless
initializer)
- minor formatting adjustments
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
This patch move stuff unchanged from asm-x86/x86_64/asm_defns.h to
asm-x86/asm_defns.h
Signed-off-by: Feng Wu <feng.wu@intel.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
This patch moves label common_interrupt from asm_defns.h to entry.S and
convert SAVE_ALL from a C to an assembler macro.
Signed-off-by: Feng Wu <feng.wu@intel.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
This patch defines macros CPUINFO_features and CPUINFO_FEATURE_OFFSET.
CPUINFO_features can be used as the base of the offset for cpu features,
while CPUINFO_FEATURE_OFFSET is used to define the right offset for
specific CPU feature.
Signed-off-by: Feng Wu <feng.wu@intel.com>
Some further cleanup (both to the patch and to surrounding code).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Each family has a stipulated max patch_size. Use this as
additional sanity check before we apply it.
Also, tone down the amount of debug messages and
Follow microcode_intel's implementation of pr_debug.
While at it, fix comment at very top to indicate we support ucode
patch loading from fam10h and higher.
Signed-off-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@amd.com>
Reviewed-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
* Don't mix uint32_t and unsigned int between prototype and definition
* Don't bitwise or with 0
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Commit b3e024f3 ("x86/NPT: don't walk page tables when changing types
on a range") neglected the fact that p2m_next_level() replaces the
previous level's mapping with the new level's one, hence dereferencing
a stale pointer the translation for which may no longer be available
(timing dependent). Add a parameter to that function allowing the
caller to request that the mapping be retained (the unmapping will be
taken care of by the caller then).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Ian Campbell <ian.campbell@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Tim Deegan <tim@xen.org>
Document and implement a new option that permits disk backends which
would otherwise have to avoid O_DIRECT (because of the network memory
lifetime bug) to use it anyway. This is:
direct-io-safe in the xl domain disk config specification
direct_io_safe in the libxl disk API
direct-io-safe in the backend xenstore interface
Add a reference to xen/include/public/io/blkif.h in
docs/misc/vbd-interface.txt.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Tested-by: Felipe Franciosi <felipe@paradoxo.org>
This updates libxlu_cfg_y.[ch] to code generated by bison from
Debian wheezy (1:2.5.dfsg-2.1 i386).
There should be no functional change since there is no change to the
source file, but we will inherit bugfixes and behavioural changes from
the new version of bison. So this is more a matter of hope than
knowledge.
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
The current function name vcpu_domain_init doesn't reflect what the function
does and might be misused.
Rename it into domain_vtimer_init.
Signed-off-by: Julien Grall <julien.grall@linaro.org>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
If there only is 1 NUMA node, no need to go through placement
candidate selection, etc., we can just bail.
Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
* Don't walk off the end of p2m_table under the mistaken impression that it
contains toolstack unsigned longs. Despite its array type it contains guest
unsigned longs so unconditionally needs casting to the guest width to use
correctly. Furthermore, a 64bit toolstack must be extra careful when it
finds a 32bit guest's INVALID_MFN.
* Drop 'mapped' and 'pinned' descriptions. This are both bogus, including all
uses of the is_mapped() macro.
* Rearrange the type name printing to be more concise.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
CC: Ian Campbell <Ian.Campbell@citrix.com>
CC: Ian Jackson <Ian.Jackson@eu.citrix.com>
CC: Dario Faggioli <dario.faggioli@citrix.com>
Reviewed-by: Dario Faggioli <dario.faggioli@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Rather than using the internal MPI format for the Diffie-Hellman group,
whose representation depends on the size of the t_uint type, store the
value as a big-endian integer and use mpi_read_binary to convert it in
an architecture-independent manner. This patch also removes the
unnecessary range check on the exponent which ended up being different
between 32- and 64-bit code.
Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Andrew Cooper