Jan Beulich [Thu, 29 Jan 2015 13:42:20 +0000 (13:42 +0000)]
x86/shadow: adjust mask shadow_audit_tables() passes to hash_foreach()
It so far having been ~1 made most of the code preceding the call
pointless, but I assume this wasn't meant to be that way. Also replace
the remaining hard coded ~1 with an expression documenting the
intention a little better.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Adjust again to use SHF_page_type_mask, at Jan's suggestion.
Jan Beulich [Thu, 29 Jan 2015 13:40:40 +0000 (13:40 +0000)]
x86/shadow: convert non-const statics
To make obvious that such statics are safe to use, they should be
const. In some of the cases, they wouldn't even need to be static, but
keep them so upon the maintainer's request.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Tim Deegan <tim@xen.org>
Jan Beulich [Thu, 29 Jan 2015 13:24:04 +0000 (14:24 +0100)]
x86: support SMBIOS v3
While presumably of primary use to ARM64 (once the code gets
generalized), we should still support this more modern variant,
allowing for the actual DMI data to reside in memory above 4Gb.
While based on draft version 3.0.0d, it is assumed that the final
version of the specification will not render this implementation
invalid (not the least because Linux 3.19 already makes the same
assumption).
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
David Vrabel [Thu, 29 Jan 2015 13:22:22 +0000 (14:22 +0100)]
grant-table: defer releasing pages acquired in a grant copy
Acquiring a page for the source or destination of a grant copy is an
expensive operation. A common use case is for two adjacent grant copy
ops to operate on either the same source or the same destination page.
Instead of always acquiring and releasing destination and source pages
for each operation, release the page once it is no longer valid for
the next op.
If either the source or destination domains changes both pages are
released as it is unlikely that either will still be valid.
XenServer's performance benchmarks show modest improvements in network
receive throughput (netback uses grant copy in the guest Rx path) and
no regressions in disk performance (using tapdisk3 which grant copies
as the backend).
Baseline Deferred Release
Interhost receive to VM 7.2 Gb/s ~9 Gbit/s
Interhost aggregate 24 Gb/s 28 Gb/s
Intrahost single stream 14 Gb/s 14 Gb/s
Intrahost aggregate 34 Gb/s 36 Gb/s
Aggregate disk write 900 MB/s 900 MB/s
Aggregate disk read 890 MB/s 890 MB/s
Signed-off-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Tim Deegan <tim@xen.org> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Dan Carpenter [Wed, 28 Jan 2015 15:50:08 +0000 (16:50 +0100)]
bunzip2: off by one in get_next_block()
"origPtr" is used as an offset into the bd->dbuf[] array. That array is
allocated in start_bunzip() and has "bd->dbufSize" number of elements so
the test here should be >= instead of >.
Later we check "origPtr" again before using it as an offset so I don't
know if this bug can be triggered in real life.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Trivial adjustments to make the respective Linux commit b5c8afe5be51078a979d86ae5ae78c4ac948063d apply to Xen.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Jan Beulich [Wed, 28 Jan 2015 15:38:20 +0000 (16:38 +0100)]
x86: skip further initialization for idle domains
While in the end not really found necessary, early versions of the
patches to follow pointed out that we needlessly set up paging for idle
domains. Arranging for that to be skipped made me notice that we can at
once skip vMCE setup for them. Leverage to adjustment to further
re-arrange the way FPU setup gets skipped.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Chao Peng [Wed, 28 Jan 2015 15:33:01 +0000 (16:33 +0100)]
x86: allow reading MSR_IA32_TSC with XENPF_resource_op
Memory bandwidth monitoring requires system time information returned
along with the monitoring counter to verify the correctness of the
counter value and to calculate the time elapsed between two samplings.
Add MSR_IA32_TSC to the read path and it returns scaled system time(ns)
instead of raw timestamp to elimanate the needs to convert. The return
time is obfuscated with booting random to eliminate the potential abuse
of it. RESOURCE_ACCESS_MAX_ENTRIES is also increased to 3 so MSR_IA32_TSC
can be used together with an MSR write/read operation pair.
Jan Beulich [Wed, 28 Jan 2015 15:29:46 +0000 (16:29 +0100)]
x86: also use tzcnt instead of bsf in __scanbit()
... when available, i.e. by runtime patching. This saves the
conditional move, having a back-to-back dependency on BSF's (EFLAGS)
result.
The need to include asm/cpufeatures.h from asm/bitops.h requires a
workaround for an otherwise resulting circular header file dependency:
Provide a mode by which the including site of the former header can
request to only get the X86_FEATURE_* defines (and very little more)
from it, allowing it to nevertheless be included in its entirety later
on.
While doing this I also noticed that the function's "max" parameter was
pointlessly "unsigned long" - the function only returning
"unsigned int", this can't be of any use, and hence gets converted at
once, along with the necessary adjustments to CMOVZ's output operands.
Note that while only alternative_io() is needed by this change (and
hence gets pulled over from Linux), for completeness its input-only
counterpart alternative_input() gets added as well.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Wei Liu [Wed, 28 Jan 2015 13:26:21 +0000 (13:26 +0000)]
libxl: correct function name
spaw_ -> spawn_
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Andrew Cooper [Tue, 27 Jan 2015 16:58:06 +0000 (16:58 +0000)]
tools/libxc: Disable CONFIG_MIGRATE in stubdom environments
The legacy save/restore infrastructure requires several function pointers from
the toolstack (libxl or Xend in the past) in order to work, and for HVM guests
also need to be able to play around in dom0's filesystem to move the device
model save record.
Migration v2 changes some of this, but is similarly dependent on
toolstack-provided function pointers.
Someone who wishes to re-architect the interaction of moving parts for running
a domain might be in a position to re-enabled this, but for now, explicitly
fail with ENOSYS (from xc_nomigrate.c) rather than failing with an error about
a missing function pointer (or indeed falling over a NULL pointer on certain
paths).
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: Wei Liu <wei.liu2@citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Andrew Cooper [Tue, 27 Jan 2015 20:38:11 +0000 (20:38 +0000)]
ocaml/xenctrl: Check return values from hypercalls
rather than blindly continuing and possibly using negative values.
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: Wei Liu <wei.liu2@citrix.com> CC: Dave Scott <dave.scott@eu.citrix.com> Acked-by: David Scott <dave.scott@citrix.com>
Create preinit_xen_time() and move to it minimum required
subset of operations needed to properly initialized
cpu_khz and boot_count vars. This is allow us to use udelay()
immediately after the call.
Quan Xu [Thu, 15 Jan 2015 09:21:50 +0000 (04:21 -0500)]
vTPM/TPM2: Support TPM 2.0 bind and unbind data
Bind data with TPM2_RSA_Encrypt, which performs RSA encryption using
the indicated padding scheme according to PKCS#1v2.1(PKCS#1). If the
scheme of keyHandle is TPM_ALG_NULL, then the caller may use inScheme
to specify the padding scheme.
Unbind data with TPM2_RSA_Decrypt, which performs RSA decryption using
the indicated padding scheme according to PKCS#1v2.1(PKCS#1).
Signed-off-by: Quan Xu <quan.xu@intel.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Quan Xu [Thu, 15 Jan 2015 09:21:48 +0000 (04:21 -0500)]
vTPM/TPM2: Support 'tpm2' extra command line.
Make vtpm-stubdom domain compatible to launch on TPM 1.x / TPM 2.0.
Add:
..
extra="tpm2=1"
..
to launch vtpm-stubdom domain on TPM 2.0, ignore it on TPM 1.x. for
example,
vtpm-stubdom domain configuration on TPM 2.0:
Quan Xu [Thu, 15 Jan 2015 09:21:47 +0000 (04:21 -0500)]
vTPM/TPM2: Add main entrance vtpmmgr2_init()
Accept commands from the vtpm-stubdom domains via the mini-os TPM
backend driver. The vTPM manager communicates directly with hardware
TPM 2.0 using the mini-os tpm2_tis driver.
Signed-off-by: Quan Xu <quan.xu@intel.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Quan Xu [Thu, 15 Jan 2015 09:21:46 +0000 (04:21 -0500)]
vTPM/TPM2: TPM2.0 TIS initialization and self test.
call the TPM 2.0 various registers that allow communication between
the TPM 2.0 and platform hardware and software. TPM2_SelfTest causes
the TPM 2.0 to perform a test of its capabilities.
Signed-off-by: Quan Xu <quan.xu@intel.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Quan Xu [Thu, 15 Jan 2015 09:21:45 +0000 (04:21 -0500)]
vTPM/TPM2: Create and load SK on TPM 2.0
TPM2_Create is used to create an object that can be loaded into a
TPM using TPM2_Load(). If the command completes successfully, the
TPM will create the new object and return the object’s creation.
data (creationData), its public area (outPublic), and its encrypted
sensitive area (outPrivate). Preservation of the returned data is
the responsibility of the caller. The object will need to be loaded
(TPM2_Load()).
TPM2_Load is used to load objects into the TPM. This command is used
when both a TPM2B_PUBLIC and TPM2B_PRIVATE are to be loaded. If only
a TPM2B_PUBLIC is to be loaded, the TPM2_LoadExternal command is used.
Signed-off-by: Quan Xu <quan.xu@intel.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Quan Xu [Thu, 15 Jan 2015 09:21:44 +0000 (04:21 -0500)]
vTPM/TPM2: TPM 2.0 takes ownership and create SRK
TPM2_CreatePrimary is used to create a Primary Object under one of
the Primary Seeds or a Temporary Object under TPM_RH_NULL. The command
uses a TPM2B_PUBLIC as a template for the object to be created. The
command will create and load a Primary Object. The sensitive area is
not returned. Any type of object and attributes combination that is
allowed by TPM2_Create() may be created by this command. The constraints
on templates and parameters are the same as TPM2_Create() except that a
Primary Storage Key and a Temporary Storage Key are not constrained to
use the algorithms of their parents.
Signed-off-by: Quan Xu <quan.xu@intel.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Andrew Cooper [Tue, 27 Jan 2015 20:34:02 +0000 (20:34 +0000)]
tools/libxl: Correct static pattern rule for pkgconfig files
Attempting to build libxl causes Make to emit the following warnings
andrewcoop@andrewcoop:xen.git$ make -C tools/libxl all
...
Makefile:253: target `xenlight.pc' doesn't match the target pattern
Makefile:253: target `xlutil.pc' doesn't match the target pattern
...
because the static pattern rule is malformed. 'Makefile' as the only
prereq-pattern does not contain a pattern.
The rule ends up working because of the use of $@.in where $< should have been
used, but lacked any dependency between a $FOO.pc and its .in source file.
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: Wei Liu <wei.liu2@citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
libxl: Prevent qemu closing QMP socket on shutdown before libxl is done with it.
At present on shutdown when using pci-passthrough with qemu-xen, qemu
closes the QMP socket before libxl is done with it causing these
errors to be logged by libxl:
Waiting for domain test (domid 1) to die [pid 11568]
Domain 1 has shut down, reason code 0 0x0
Action for shutdown reason code 0 is destroy
Domain 1 needs to be cleaned up: destroying the domain
libxl: error: libxl_qmp.c:443:qmp_next: Socket read error: Connection reset by peer
libxl: error: libxl_qmp.c:701:libxl__qmp_initialize: Failed to connect to QMP
libxl: error: libxl_qmp.c:686:libxl__qmp_initialize: Connection error: Connection refused
libxl: error: libxl_dm.c:1588:kill_device_model: Device Model already exited
Done. Exiting now
Prevent this by using the qemu '-no-shutdown' parameter which is
described as doing:
"Don’t exit QEMU on guest shutdown, but instead only stop the emulation.
This allows for instance switching to monitor to commit changes to the disk image."
So Qemu will stop emulating, but keeps the QMP socket open and waits
for libxl to kill the qemu process when it is done, preventing the
race and resulting in this to be logged by libxl:
Waiting for domain test (domid 1) to die [pid 10859]
Domain 1 has shut down, reason code 0 0x0
Action for shutdown reason code 0 is destroy
Domain 1 needs to be cleaned up: destroying the domain
Done. Exiting now
Wei Liu [Sun, 25 Jan 2015 15:38:59 +0000 (15:38 +0000)]
tools/Makefile: fix qemu-xen-traditional build
In d9740237a ("tools: unhook blktap1 from the build and remove all
references to it"), one spot was left unchanged, which leads to failure
in building qemu-xen-traditional.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Wei Liu [Tue, 20 Jan 2015 11:47:46 +0000 (11:47 +0000)]
tools: generate systemd service files only when systemd is available
Though that's not in any way harmful but it is on the other hand not
very useful.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
[ ijc -- rerun autogen.sh ] Acked-by: Ian Campbell <ian.campbell@citrix.com>
Wei Liu [Tue, 20 Jan 2015 13:31:12 +0000 (13:31 +0000)]
tools: fix "make distclean"
The original rule to target "distclean" in tools/Rules.mk was in effect
"make clean". It should be "make distclean".
However not all Makefiles in subdirectories have distclean target
defined. So this patch also adds a bunch of distclean targets to various
Makefiles. They only depend on clean target and don't have any actions
in most cases.
With the patch applied, following command outputs 0 results:
Ocaml and libfsimage are known to have distclean defined in a dedicated
rules file.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Wei Liu [Tue, 20 Jan 2015 12:22:50 +0000 (12:22 +0000)]
libxl: provide xlutil.pc
Please rerun autogen.sh after applying this patch.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Wei Liu [Tue, 20 Jan 2015 12:22:49 +0000 (12:22 +0000)]
libxl: delete xenlight.pc.in in distclean
That file is generated by configure. Deleting it in "make clean" leads
to rerun configure. Move it under distclean target.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Julien Grall [Wed, 21 Jan 2015 13:25:44 +0000 (13:25 +0000)]
libxc: introduce a per architecture scratch pfn for temporary grant mapping
The code to initialize the grant table in libxc uses
xc_domain_maximum_gpfn() + 1 to get a guest pfn for mapping the grant
frame and to initialize it.
This solution has two major issues:
- The check of the return of xc_domain_maximum_gpfn is buggy because
xen_pfn_t is unsigned and in case of an error -ERRNO is returned.
Which is never catch with ( pfn <= 0 ).
- The guest memory layout maybe filled up to the end, i.e
xc_domain_maximum_gpfn() + 1 gives either 0 or an invalid PFN due to
hardware limitation.
Futhermore, on ARM, xc_domain_maximum_gpfn() is not implemented and
return -ENOSYS. This will make libxc to use always the same PFN which
may colapse with an already mapped region (see xen/include/public/arch-arm.h
for the layout).
This patch only address the problem for ARM, the x86 version use the same
behavior (ie xc_domain_maximum_gpfn() + 1), as I'm not familiar with Xen x86.
A new function xc_core_arch_get_scratch_gpfn is introduced to be able to
choose the gpfn per architecture.
For the ARM version, we use the GUEST_GNTTAB_GUEST which is the base of
the region by the guest to map the grant table. At the build time,
nothing is mapped there.
At the same time correctly check the return of xc_domain_maximum_gpfn
for x86.
Signed-off-by: Julien Grall <julien.grall@linaro.org> Cc: Jan Beulich <jbeulich@suse.com> Cc: Andrew Cooper <andrew.cooper3@citrix.com> Cc: Roger Pau Monné <roger.pau@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Wei Liu <wei.liu2@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
x86: vcpu_destroy_pagetables() must not return -EINTR
.. otherwise it has the side effect that: domain_relinquish_resources
will stop and will return to user-space with -EINTR which it is not
equipped to deal with that error code; or vcpu_reset - which will
ignore it and convert the error to -ENOMEM..
The preemption mechanism we have for domain destruction is to return
-EAGAIN (and then user-space calls the hypercall again) and as such we need
to catch the case of:
we need to return -ERESTART otherwise we end up returning -ENOMEM.
There are also other callers of vcpu_destroy_pagetables: arch_vcpu_reset
(vcpu_reset) are:
- hvm_s3_suspend (asserts on any return code),
- vlapic_init_sipi_one (asserts on any return code),
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Jan Beulich <jbeulich@suse.com>
Jan Beulich [Mon, 26 Jan 2015 11:50:21 +0000 (12:50 +0100)]
x86: use tzcnt instead of bsf
Following a compiler change done in 2012, make use of the fact that for
non-zero input BSF and TZCNT produce the same numeric result (EFLAGS
setting differs), and that CPUs not knowing of TZCNT will treat the
instruction as BSF (i.e. ignore what looks like a REP prefix to them).
The assumption here is that TZCNT would never have worse performance
than BSF.
Also extend the asm() input in find_first_set_bit() to allow memory
operands.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Andrew Cooper [Mon, 26 Jan 2015 11:48:38 +0000 (12:48 +0100)]
x86/HVM: improve EFER validation error messages
The previous error message was very little use in identifying the actual
problem after the fact. Now, hvm_efer_valid() will indicate the issue which
it objects to, which is far more useful for diagnosing issues from logs.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Boris Ostrovsky [Fri, 23 Jan 2015 16:53:49 +0000 (17:53 +0100)]
x86/VPMU: handle APIC_LVTPC accesses
Don't have the hypervisor update APIC_LVTPC when _it_ thinks the vector should
be updated. Instead, handle guest's APIC_LVTPC accesses and write what the guest
explicitly wanted.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Acked-by: Kevin Tian <kevin.tian@intel.com> Acked-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Dietmar Hahn <dietmar.hahn@ts.fujitsu.com> Tested-by: Dietmar Hahn <dietmar.hahn@ts.fujitsu.com>
Boris Ostrovsky [Fri, 23 Jan 2015 16:50:53 +0000 (17:50 +0100)]
x86/VPMU: manage VPMU_CONTEXT_SAVE flag in vpmu_save_force()
There is a possibility that we set VPMU_CONTEXT_SAVE on VPMU context in
vpmu_load() and never clear it (because vpmu_save_force() will see
VPMU_CONTEXT_LOADED bit clear, which is possible on AMD processors)
The problem is that amd_vpmu_save() assumes that if VPMU_CONTEXT_SAVE is set
then (1) we need to save counters and (2) we don't need to "stop" control
registers since they must have been stopped earlier. The latter may cause all
sorts of problem (like counters still running in a wrong guest and hypervisor
sending to that guest unexpected PMU interrupts).
Since setting this flag is currently always done prior to calling
vpmu_save_force() let's both set and clear it there.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Reviewed-by: Dietmar Hahn <dietmar.hahn@ts.fujitsu.com> Tested-by: Dietmar Hahn <dietmar.hahn@ts.fujitsu.com>
Roger Pau Monné [Fri, 23 Jan 2015 14:16:18 +0000 (15:16 +0100)]
x86: prevent access to HPET from Dom0
Prevent Dom0 from accessing HPET MMIO region by adding the HPET mfn to the
list of forbiden memory regions (if ACPI_HPET_PAGE_PROTECT4 or
ACPI_HPET_PAGE_PROTECT64 flag is set) or to the list of read-only regions.
Also provide an option that prevents adding the HPET to the read-only memory
regions called ro-hpet, in case there are systems that put other stuff in
the HPET page.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Don't loop over iomem_deny_access() for consecutive MFNs.
Put new command line option's doc entry in right spot.
Jan Beulich [Fri, 23 Jan 2015 14:13:05 +0000 (15:13 +0100)]
x86/HVM: make hvm_efer_valid() honor guest features
Following the earlier similar change validating CR4 modifications.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Tested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Jan Beulich [Fri, 23 Jan 2015 14:03:28 +0000 (15:03 +0100)]
x86/HVM: replace plain number in hvm_combine_hw_exceptions()
While doing so also take care of #VE here (even if we don't make use of
it yet). Note that contributory_exceptions, other than the original
0x7c01 constant, doesn't include #PF anymore, but the check where the
variable is used is after one that already filtered out #PF.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: TIm Deegan <tim@xen.org>
Jan Beulich [Fri, 23 Jan 2015 14:02:39 +0000 (15:02 +0100)]
arm64: fix fls()
It using CLZ on a 64-bit register while specifying the input operand as
only 32 bits wide is wrong: An operand intentionally shrunk down to 32
bits at the source level doesn't imply respective zero extension also
happens at the machine instruction level, and hence the wrong result
could get returned.
Add suitable inline assembly abstraction so that the function can
remain shared between arm32 and arm64. The need to include asm_defns.h
in bitops.h makes it necessary to adjust processor.h though - it is
generally wrong to include public headers without making sure that
integer types are properly defined. (I didn't innvestigate or try
whether the possible alternative of moving the public/arch-arm.h
inclusion down in the file would also work.)
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Jan Beulich [Fri, 23 Jan 2015 13:59:37 +0000 (14:59 +0100)]
make fls() and ffs() consistent across architectures
Their parameter types differed between ARM and x86.
Along with generalizing the functions this fixes
- x86's non-long functions having long parameter types
- ARM's ffs() using a long intermediate variable
- generic_fls64() being broken when the upper half of the input is
non-zero
- common (and in one case also ARM) code using fls() when flsl() was
meant
Also drop ARM's constant_fls() in favor of the identical generic_fls().
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Andrew Cooper [Thu, 22 Jan 2015 11:59:14 +0000 (12:59 +0100)]
x86/vlapic: express x2apic msr readability with a bitmap
The x2apic MSR space is currently defined between 0x800 and 0x83f, which
conveniently fits in a 64 bit wide bitmap. This is far more efficient than
the cascade comparisons generated by the switch statement, which can't be
optimised because of the case ranges used for the ISR, TMR and IRR blocks.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Convert 0xffUL to ((1UL << (NR_VECTORS / 32)) - 1) and drop a couple of
clearly superfluous parentheses.
Jan Beulich [Thu, 22 Jan 2015 11:57:27 +0000 (12:57 +0100)]
VMX: drop VMCS *_HIGH enumerators
Most of them have been unused since the dropping of 32-bit support, and
the few remaining cases are more efficiently dealt with using a generic
macro (and probably things should have been done that way from the
beginning).
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
Jan Beulich [Thu, 22 Jan 2015 11:56:33 +0000 (12:56 +0100)]
VMX: dump further control state
A few relevant control state fields did not get dumped so far; in
particular, VM_ENTRY_INSTRUCTION_LEN got printed twice (instead of also
printing VM_EXIT_INSTRUCTION_LEN). Where suitable (to reduce the amount
of output) make some of the dumping conditional upon guest settings
(this isn't required for correctness as vmr() already uses
__vmread_safe(), i.e. it is fine to access non-existing fields).
Also drop casts.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
Jan Beulich [Thu, 22 Jan 2015 11:55:56 +0000 (12:55 +0100)]
VMX: dump full host state
A few host state fields did not get dumped so far. Where suitable (to
reduce the amount of output) make some of the dumping conditional upon
guest settings (this isn't required for correctness as vmr() already
uses __vmread_safe(), i.e. it is fine to access non-existing fields).
Also drop casts - many of them haven't been needed anymore since the
dropping of 32-bit support.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
Jan Beulich [Thu, 22 Jan 2015 11:54:49 +0000 (12:54 +0100)]
VMX: dump full guest state
Several guest state fields did not get dumped so far. Where suitable
(to reduce the amount of output) make some of the dumping conditional
upon guest settings (this isn't required for correctness as vmr()
already uses __vmread_safe(), i.e. it is fine to access non-existing
fields).
Move CR3_TARGET_* and TSC_OFFSET processing into the control state
section, at once making the upper bound of CR3_TARGET_VALUEn printed
depend on CR3_TARGET_COUNT (which architecturally can be higher than
4).
Also rename GUEST_PDPTRn to GUEST_PDPTEn (matching the SDM naming) and
group them as well as CR3_TARGET_VALUEn similar to EOI_EXIT_BITMAP.
Finally, drop casts - they haven't been needed anymore since the
dropping of 32-bit support (and some of them were not really needed in
the first place). Introduce vmr16() and vmr32() helper macros to avoid
the "l" printk format modifier and at the same time validate that only
16-/32-bit fields get accessed this way.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
Jan Beulich [Thu, 22 Jan 2015 11:47:56 +0000 (12:47 +0100)]
x86: don't expose XSAVES capability to PV guests
As done by the recent Linux commit b65d6e17fe ("kvm: x86: mask out
XSAVES") for KVM, we should also mask out XSAVES from what PV guests
get to see as long as we don't emulate accesses to MSR_IA32_XSS.
Actually, go beyond that: Just like for leaf 7, switch from
blacklisting to whitelisting, i.e. only allow XSAVEOPT and XSAVEC for
the time being. And do these overrides consistently for both Dom0 and
DomU-s.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Andrew Cooper [Thu, 22 Jan 2015 11:46:43 +0000 (12:46 +0100)]
drop redundant MAX_VIRT_CPUS bounds checks
In all 4 cases, visible in the context are bounds check against d->max_vcpus.
Domain building will ensure that d->max_vcpus never exceeds an appropriate
bound. In the x86 case, different types of domains have different maxima for
vcpus, making the checks wrong as opposed to simply redundant.
For vpsci in ARM, 'vcpuid' is an unsigned type so could never be less than 0.
For the common changes to do_{,compat}_vcpu_op(), these changes do result in a
guest visible change, but only in so far as certain invalid vcpu ids will now
fail with -ENOENT rather than -EINVAL.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Andrew Cooper [Thu, 22 Jan 2015 11:46:10 +0000 (12:46 +0100)]
sched/arinc653: remove MAX_VIRT_CPUS bounds check
The arinc653 interface is capable of specifying a domain in the schedule (from
the toolstack) before the domain itself exists, or is present in the cpupool
(The domain is identified by UUID rather than domid). As a result, the
schedule can't be validated at this point.
The vcpu_id from userspace is only ever used to compare against a list of real
vcpus available to the scheduler, which prevents ill-specified vcpus from
actually being scheduled.
Remove the MAX_VIRT_CPUS test, as it is not an appropriate bound for vcpu_id.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by Robert VanVossen <robert.vanvossen@dornerworks.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Andrew Cooper [Thu, 22 Jan 2015 11:45:13 +0000 (12:45 +0100)]
evtchn: reduce the size of the poll_mask where possible
Use domain_max_vcpus(d) in preference to MAX_VIRT_CPUS when allocating the
poll mask. This allows x86 HVM guests to have a poll mask of 128 bits rather
than 8k bits.
While changing this, use xzalloc_array() in preference to xmalloc_array() to
avoid needing the subsequent call to bitmap_zero().
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Andrew Cooper [Thu, 22 Jan 2015 11:44:03 +0000 (12:44 +0100)]
introduce domain_max_vcpus() helper and implement per arch
This allows the common XEN_DOMCTL_max_vcpus handler to lose some x86-specific
architecture knowledge.
It turns out that Xen had the same magic number twice in-tree with different
names (HVM_MAX_VCPUS and MAX_HVM_VCPUS). This removes all use of
MAX_HVM_VCPUS, and x86 uses HVM_MAX_VCPUS from the public headers.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Jan Beulich [Thu, 22 Jan 2015 11:41:50 +0000 (12:41 +0100)]
make error codes a formal part of the ABI
Now that we have two cases where patches against hvmloader got
submitted needing to include the hypervisor's errno.h (for the host's
system header not necessarily reflecting the correct numbers), take
this as a strong sign that we need to make the error return values part
of the hypervisor ABI (which de-fact they've always been).
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Jan Beulich [Tue, 20 Jan 2015 09:46:19 +0000 (10:46 +0100)]
x86: latch current->domain in do_physdev_op()
... and drop global latching of current, as being needed more than once
only in PHYSDEVOP_set_iopl and PHYSDEVOP_set_iobitmap, and not at all
in all other cases.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
We don't really need the MFN in more than one place (after dropping
mfn_to_page() translations where we know the result already), so no
need to have a local variable for it.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Andrew Cooper [Tue, 20 Jan 2015 09:42:26 +0000 (10:42 +0100)]
xsm/evtchn: never pretend to have successfully created a Xen event channel
Xen event channels are not internal resources. They still have one end in a
domain, and are created at the request of privileged domains. This logic
which "successfully" creates a Xen event channel opens up undesirable failure
cases with ill-specified XSM policies.
If a domain is permitted to create ioreq servers or memevent listeners, but
not to create event channels, the ioreq/memevent creation will succeed but
attempting to bind the returned event channel will fail without any indication
of a permission error.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Andrew Cooper [Tue, 20 Jan 2015 09:41:18 +0000 (10:41 +0100)]
x86/cpuid: correct parameter types for cpuid_count()
About half of the cpuid space has the top bit of op set, and op it always
specified with unsigned integers. There are no problematic uses in tree at
the moment.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Wei Liu [Mon, 19 Jan 2015 16:13:59 +0000 (16:13 +0000)]
libxl: check HVM direct boot parameters
We should honour -initrd and -append iff -kernel is specified, because
that's how QEMU works.
Check direct boot parameters in libxl__domain_build_info_setdefault.
Return failure if kernel is missing while ramdisk (-initrd) and / or
cmdline (-append) is present.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Chunyan Liu <cyliu@suse.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Julien Grall [Thu, 15 Jan 2015 21:06:18 +0000 (21:06 +0000)]
tools/configure: Check if pixman is present on the system when building QEMU
QEMU upstream requires the use of pixman. When pixman is not present the
system, the configure of QEMU will fail with:
ERROR: pixman not present. Your options:
(1) Preferred: Install the pixman devel package (any recent
distro should have packages as Xorg needs pixman too).
(2) Fetch the pixman submodule, using:
git submodule update --init pixman
I think we can use by default the version on the system. So check it
a Xen configuration time to avoid a build issue later.
Signed-off-by: Julien Grall <julien.grall@linaro.org> Acked-by: Ian Campbell <ian.campbell@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
[ ijc -- reran autogen.sh, adjusted comment ]
Julien Grall [Thu, 15 Jan 2015 20:23:40 +0000 (20:23 +0000)]
xen/arm: vgic: Rename nr_lines into nr_spis
The field nr_lines in the arch_domain vgic structure contains the number of
SPIs for the emulated GIC. Using the nr_lines make confusion with the GIC
code, where it means the number of IRQs. This can lead to coding error.
Also introduce vgic_num_irqs to get the number of IRQ handled by the emulated
GIC.
Finally drop the initialization of nr_spis in both gicv2v_init and gicv3_init
as it's already initialized in vgic common code.
Wei Liu [Fri, 9 Jan 2015 14:32:18 +0000 (14:32 +0000)]
libxl: provide xenlight.pc
A pkg-config file for libxl. It also contains two variables
(xenfirmwaredir and libexec_bin) so that tools that are very keen on
knowing the locations of Xen binaries (say, libvirt) can use them to
determine the location of the binaries.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
[ ijc -- ran autogen.sh ]
Paul Durrant [Mon, 19 Jan 2015 11:18:03 +0000 (12:18 +0100)]
x86/hvm: add per-vcpu evtchn upcalls
HVM guests have always been confined to using the domain callback
via (see HVM_PARAM_CALLBACK_IRQ) to receive event notifications.
This is usually an IOAPIC vector and is only used if the event
channel is bound to vcpu 0.
PVHVM Linux uses a pre-defined interrupt vector for the event
channel upcall, set using HVM_PARAM_CALLBACK_IRQ by ORing in a
special bit (bit 57) into the value (see params.h). However, it
does not assert the interrupt via the emulated local APIC.
This mechanism is not suitable in the general case since Windows
(and potentially other OSes) because they:
- cannot guarantee the same vector for all VCPUs
- do require the upcall to be asserted via the local APIC
This patch adds a new HVM op allowing a guest to specify a local
APIC vector to use as an upcall notification for a specific vcpu
therefore coping with the case of differing vector numbers.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Andrew Cooper [Mon, 19 Jan 2015 11:16:44 +0000 (12:16 +0100)]
hvm/hpet: correctly gate the virtual HPET on HVM_PARAM_HPET_ENABLE
c/s 3f8e22de7 "x86 hvm: Allow HPET to be configured as a per-domain config
option" introduced the parameter to conditionally enable the HPET.
However, having the check in hpet_range() does not have the intended effect.
As currently implemented, when the HPET is disabled, the range is not claimed
and an ioreq is forwarded to qemu, which implements an HPET itself.
Properly disable the HPET by always claiming the range, dropping writes and
reading ~0.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
projects / xen.git / log