Haozhong Zhang [Tue, 1 Mar 2016 13:37:53 +0000 (14:37 +0100)]
x86/hvm: replace architecture TSC scaling by a common function
This patch implements a common function hvm_scale_tsc() to scale TSC by
using TSC scaling information collected by architecture code.
Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com> Acked-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> for SVM bits Reviewed-by: Jan Beulich <jbeulich@suse.com>
Doug Goldstein [Mon, 29 Feb 2016 15:09:09 +0000 (16:09 +0100)]
build: consolidate CONFIG_HAS_VIDEO and CONFIG_VIDEO
No real advantage to keeping these separate. The use case of this from
Linux is when the platform or target board has support for something but
the user wants to be given the option to disable it.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Doug Goldstein [Mon, 29 Feb 2016 15:08:43 +0000 (16:08 +0100)]
build: consolidate CONFIG_HAS_VGA and CONFIG_VGA
No real advantage to keeping these separate. The use case of this from
Linux is when the platform or target board has support for something but
the user wants to be given the option to disable it.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Corneliu ZUZU [Mon, 29 Feb 2016 15:07:49 +0000 (16:07 +0100)]
arm/monitor vm-events: implement guest-request support
This patch adds ARM support for guest-request monitor vm-events.
Note: on ARM hypercall instruction skipping must be done manually
by the caller. This will probably be changed in a future patch.
Summary of changes:
== Moved to common-side:
* XEN_DOMCTL_MONITOR_EVENT_GUEST_REQUEST handling (moved from X86
arch_monitor_domctl_event to common monitor_domctl)
* hvm_event_guest_request->vm_event_monitor_guest_request
* hvm_event_traps->vm_event_monitor_traps (also added target vcpu as param)
* guest-request bits from X86 'struct arch_domain' (to common 'struct domain')
== ARM implementations:
* do_hvm_op now handling of HVMOP_guest_request_vm_event => calls
vm_event_monitor_guest_request (as on X86)
* arch_monitor_get_capabilities->vm_event_monitor_get_capabilities,
updated to reflect support for XEN_DOMCTL_MONITOR_EVENT_GUEST_REQUEST
* vm_event_init_domain (does nothing), vm_event_cleanup_domain
== Misc:
* vm_event_fill_regs, no longer X86-specific. ARM-side implementation of this
function currently does nothing, that will be added in a separate patch.
Signed-off-by: Corneliu ZUZU <czuzu@bitdefender.com> Acked-by: Jan Beulich <jbeulich@suse.com> Acked-by: Razvan Cojocaru <rcojocaru@bitdefender.com> Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Acked-by: Tamas K Lengyel <tamas@tklengyel.com>
Haozhong Zhang [Mon, 29 Feb 2016 15:06:40 +0000 (16:06 +0100)]
x86/hvm: setup TSC scaling ratio
This patch adds a field tsc_scaling_ratio in struct hvm_domain to record
the per-domain TSC scaling ratio, and sets it in tsc_set_info().
Before setting the per-domain TSC scaling ratio, we check its validity
in tsc_set_info(). If an invalid ratio is given, we will leave the
default value in tsc_scaling_ratio (i.e. ratio = 1) and setup guest TSC
as if no TSC scaling is used:
* For TSC_MODE_DEFAULT,
- if a user-specified TSC frequency is given, we will set the guest
TSC frequency to it; otherwise, we set it to the host TSC frequency.
- if guest TSC frequency does not equal to host TSC frequency, we will
emulate guest TSC (i.e. d->arch.vtsc is set to 1). In both cases,
guest TSC runs in the guest TSC frequency.
* For TSC_MODE_PVRDTSCP,
- we set the guest TSC frequency to the host TSC frequency.
- guest rdtsc is executed natively in the host TSC frequency as
before.
- if rdtscp is not available to guest, it will be emulated; otherwise,
it will be executed natively. In both cases, guest rdtscp gets TSC
in the host TSC frequency as before.
Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Acked-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Parth Dixit [Mon, 29 Feb 2016 15:06:07 +0000 (16:06 +0100)]
arm/acpi: read acpi memory info from uefi
ACPI memory is seperate from conventional memory and should be marked
as reserved while passing to DOM0. Create a new meminfo structure to
store all the acpi tables listed in uefi.
Shannon Zhao [Mon, 29 Feb 2016 15:05:32 +0000 (16:05 +0100)]
arm/acpi: add placeholder for efi and acpi load address
We will create EFI table, memory description table and some of acpi
tables and we're going to map them to kinfo->gnttab_start of Dom0.
Add placeholder for the starting address for loading in DOM0 and the
size of new added tables. Also add a placeholder to store the new
created tables.
Dario Faggioli [Mon, 29 Feb 2016 14:58:49 +0000 (15:58 +0100)]
credit1: avoid boosting vCPUs being "just" migrated
Moving a vCPU to a different pCPU means offlining it and
then waking it up, on the new pCPU. Credit1 grants BOOST
priority to vCPUs that wakes up, with the aim of improving
I/O latency. The net effect of this all is that vCPUs get
boosted when migrating, which shouldn't happen.
For instance, this causes scheduling anomalies and,
potentially, performance problems, as reported here:
http://lists.xen.org/archives/html/xen-devel/2015-10/msg02851.html
This patch fixes this by noting down (by means of a flag)
the fact that the vCPU is about to undergo a migration.
This way we can tell, later, during a wakeup, whether the
vCPU is migrating or unblocking, and decide whether or
not to apply the boosting.
Note that it is important that atomic-safe bit operations
are used when manipulating vCPUs' flags. Take the chance
and add a comment about this.
Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com> Reviewed-by: George Dunlap <george.dunlap@citrix.com>
tools/libxl: seperate device init/cleanup from checkpoint device layer
we call (init|cleanup)_subkind_nic and (init|cleanup)_subkind_drbd_disk
directly in checkpoint device. Move them to libxl_remus.c, Call them before
calling libxl__checkpoint_devices_setup() or after calling
libxl__checkpoint_devices_teardown().
it is pure refactoring and no functional changes.
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Acked-by: Ian Campbell <ian.campbell@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
Wen Congyang [Tue, 15 Dec 2015 05:59:52 +0000 (13:59 +0800)]
tools/libxl: move remus state into a seperate structure
Add a new structure remus state, and move concrete layer's private
member to remus state.
it is pure refactoring and no functional changes.
Init interval in libxl__remus_setup(). It is safe to move this initialisation,
because this value is only used for remus, and remus will use this value after
libxl__remus_setup().
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
tools/libxl: store remus_ops in checkpoint device state
Checkpoint device is an abstract layer to do checkpoint.
COLO can also use it to do checkpoint. But there are
still some codes in checkpoint device which touch remus.
This patch and:
tools/libxl: move remus state into a seperate structure
tools/libxl: seperate device init/cleanup from checkpoint device layer
will seperate remus from checkpoint device layer.
We use remus ops directly in checkpoint device. Store it
in checkpoint device state so that we do not aware of
remus_ops in the checkpoint device layer.
It is pure refactoring and no functional changes.
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn>
Acked-by:Ian Campbell <ian.campbell@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
The patch also fixes the following backword compatibility:
The error code ERROR_REMUS_XXX was introduced in Xen 4.5, and
changed to ERROR_CHECKPOINT_XXX after previous renaming.
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Reviewed-Lightly-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
Wen Congyang [Mon, 14 Dec 2015 07:01:44 +0000 (15:01 +0800)]
migration/save: pass checkpointed_stream from libxl to libxc
Pass checkpointed_stream from libxl to libxc.
It won't affact legacy migration because legacy migration
won't use this param.
Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> CC: Ian Campbell <Ian.Campbell@citrix.com> CC: Ian Jackson <Ian.Jackson@eu.citrix.com> CC: Wei Liu <wei.liu2@citrix.com> CC: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
Wen Congyang [Mon, 14 Dec 2015 06:14:28 +0000 (14:14 +0800)]
tools/libxl: introduce enum type libxl_checkpointed_stream
Introduce enum type libxl_checkpointed_stream in IDL.
rename the last argument of migrate_receive from "remus" to
"checkpointed" since the semantics of this parameter has
changed.
NOTE:
libxl_domain_restore_params and domain_create aren't changed here,
checkpointed_stream is still an int. Because we will pass the
value from libxl to libxc.
Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
Currently struct libxl__domain_suspend_state contains 2 type of states,
one is save state, another is suspend state. This patch separates those
two out.
The motivation of this is that COLO will need to do suspend/resume
continuously, we need a more common suspend state.
After this change, dss stands for libxl__domain_save_state,
dsps stands for libxl__domain_suspend_state.
Also introduce libxl__domain_suspend_init to initialise the
libxl__domain_suspend_state.
Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> CC: Ian Campbell <Ian.Campbell@citrix.com> CC: Ian Jackson <Ian.Jackson@eu.citrix.com> CC: Wei Liu <wei.liu2@citrix.com> CC: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by:Ian Campbell <ian.campbell@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
tools/libxl: move save/restore code into libxl_dom_save.c
This is purely code motion.
Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> CC: Ian Jackson <Ian.Jackson@eu.citrix.com> Acked-by: Ian Campbell <Ian.Campbell@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
After previous refactoring, we are now able to move all remus code
into a separate file libxl_remus.c.
Export following functions for internal use:
- setup/teardown Remus:
* libxl__remus_setup
* libxl__remus_teardown
* libxl__remus_restore_setup
Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> CC: Ian Campbell <Ian.Campbell@citrix.com> CC: Ian Jackson <Ian.Jackson@eu.citrix.com>
Acked-by:Ian Campbell <ian.campbell@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
Wen Congyang [Tue, 16 Feb 2016 03:41:16 +0000 (11:41 +0800)]
libxl/remus: init checkpoint callback in Remus setup callback
Init stream {read/write} state checkpoint_callback, suspend/resume/checkpoint
callback in Remus setup callback.
There's no functional change, it's just refactoring so that we can move
all remus code into one file.
Signed-off-by: Yang Hongyang <hongyang.yang@easystack.cn> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> CC: 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> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
Bob Moore [Fri, 26 Feb 2016 11:37:18 +0000 (12:37 +0100)]
ACPICA / Headers: Add support for CSRT and DBG2 ACPI tables
These tables are defined outside of the ACPI specification.
Signed-off-by: Bob Moore <robert.moore@intel.com>
[Linux commit 4e2f9c278ad84196991fcf6f6646a3e15967fe90]
[only port the DBG2 changes] Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org> Acked-by: Jan Beulich <jbeulich@suse.com>
Doug Goldstein [Fri, 26 Feb 2016 11:31:47 +0000 (12:31 +0100)]
build: consolidate CONFIG_HAS_ACPI and CONFIG_ACPI
No real advantage to keeping these separate. The use case of this from
Linux is when the platform or target board has support for something but
the user wants to be given the option to disable it.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
David Vrabel [Fri, 26 Feb 2016 11:30:11 +0000 (12:30 +0100)]
x86/hvm: add HVM_PARAM_X87_FIP_WIDTH
The HVM parameter HVM_PARAM_X87_FIP_WIDTH to allow tools and the guest
to adjust the width of the FIP/FDP registers to be saved/restored by
the hypervisor. This is in case the hypervisor hueristics do not do
the right thing.
Add this parameter to the set saved during domain save/migrate.
Signed-off-by: David Vrabel <david.vrabel@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
David Vrabel [Fri, 26 Feb 2016 11:16:13 +0000 (12:16 +0100)]
x86/fpu: add a per-domain field to set the width of FIP/FDP
The x86 architecture allows either: a) the 64-bit FIP/FDP registers to
be restored (clearing FCS and FDS); or b) the 32-bit FIP/FDP and
FCS/FDS registers to be restored (clearing the upper 32-bits).
Add a per-domain field to indicate which of these options a guest
needs. The options are: 8, 4 or 0. Where 0 indicates that the
hypervisor should automatically guess the FIP width by checking the
value of FIP/FDP when saving the state (this is the existing
behaviour).
The FIP width is initially automatic but is set explicitly in the
following cases:
- 32-bit PV guest: 4
- Newer CPUs that do not save FCS/FDS: 8
The x87_fip_width field is placed into an existing 1 byte hole in
struct arch_domain.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Fix build.
Jan Beulich [Fri, 26 Feb 2016 11:15:36 +0000 (12:15 +0100)]
vVMX: use latched VMCS machine address
Instead of calling domain_page_map_to_mfn() over and over, latch the
guest VMCS machine address unconditionally (i.e. independent of whether
VMCS shadowing is supported by the hardware).
Since this requires altering the parameters of __[gs]et_vmcs{,_real}()
(and hence all their callers) anyway, take the opportunity to also drop
the bogus double underscores from their names (and from
__[gs]et_vmcs_virtual() as well).
Signed-off-by: Jan Beulich <jbeulich@suse.com> Tested-by: Liang Z Li <liang.z.li@intel.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
Jan Beulich [Fri, 26 Feb 2016 11:15:09 +0000 (12:15 +0100)]
x86emul: simplify IRET logic
Since we only handle real mode, we need to consider neither non-ring0
nor IOPL. Also for POPF the mode_iopl() check can really be inside the
not-ring-0 body.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Jan Beulich [Fri, 26 Feb 2016 11:14:39 +0000 (12:14 +0100)]
x86emul: limit-check branch targets
All branches need to #GP when their target violates the segment limit
(in 16- and 32-bit modes) or is non-canonical (in 64-bit mode). For
near branches facilitate this via a zero-byte instruction fetch from
the target address (resulting in address translation and validation
without an actual read from memory), while far branches get dealt with
by breaking up the segment register loading into a read-and-validate
part and a write one. The latter at once allows correcting some
ordering issues in how the individual emulation steps get carried out:
Before updating machine state, all exceptions unrelated to that state
updating should have got raised (i.e. the only ones possibly resulting
in partly updated state are faulting memory writes [pushes]).
Note that while not immediately needed here, write and distinct read
emulation routines get updated to deal with zero byte accesses too, for
overall consistency.
Reported-by: 刘令 <liuling-it@360.cn> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Tim Deegan <tim@xen.org>
Andrew Cooper [Thu, 25 Feb 2016 12:07:43 +0000 (13:07 +0100)]
x86: unilaterally remove .init mappings
Because of the new 2M alignment of .init and .bss, the existing memory
guarding infrastructure causes a shattered 2M superpage with non-present
entries for .init, and present entries for the alignment space.
Do away with the difference in behaviour between debug and non-debug builds;
always destroy the .init mappings, and reuse the space for xenheap.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <JBeulich@suse.com>
Andrew Cooper [Thu, 25 Feb 2016 12:07:14 +0000 (13:07 +0100)]
x86: use 2M superpages for text/data/bss mappings
This balloons the size of Xen in memory from 4.4MB to 8MB, because of the
required alignment adjustments.
However
* All mappings are 2M superpages.
* .text (and .init at boot) are the only sections marked executable.
* .text and .rodata are marked read-only.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Andrew Cooper [Thu, 25 Feb 2016 12:06:44 +0000 (13:06 +0100)]
x86: reorder .data and .init when linking
In preparation for using superpage mappings, .data and .bss will both want to
be mapped as read-write. By making them adjacent, they can share the same
superpage and will not require superpage alignment between themselves.
While making this change, fix a latent alignment bug whereby the alignment for
.bss.stack_aligned was in .init. __init_end only needs page alignment (due to
being reclaimed after boot), while .bss.stack_aligned really does needs
STACK_SIZE alignment.
Suggested-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Andrew Cooper [Thu, 25 Feb 2016 12:05:33 +0000 (13:05 +0100)]
memguard: drop memguard_init() entirely
The use of MAP_SMALL_PAGES causes shattering of the superpages making up the
Xen virtual region, and is counter to the purpose of this series.
Furthermore, it is not required for the memguard infrastructure to function
(which itself uses map_pages_to_xen() for creating holes).
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <JBeulich@suse.com> Acked-by: Stefano Stabellini <stefano.stabellini@citrix.com>
Andrew Cooper [Thu, 25 Feb 2016 12:03:43 +0000 (13:03 +0100)]
lockprof: move .lockprofile.data into .rodata
The entire contents of .lockprofile.data are unchanging pointers to
lock_profile structure in .data. Annotate the type as such, and link the
section in .rodata. As these are just pointers, 32byte alignment is
unnecessary.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
This is because grant_table.h contains this (note the
apostrophe): "granter\92s memory", and `grep -v', in version
2.23, stops processing the file (while, for instance,
until 2.22, this was not happening).
Although the above behavior is likely an issue in grep,
(https://debbugs.gnu.org/cgi/bugreport.cgi?bug=22461)
I think we better switch to using " ' " in that line
anyway, as we do basically everywhere else (even in
the same file).
Andrew Cooper [Thu, 25 Feb 2016 12:01:01 +0000 (13:01 +0100)]
work around Clang generating .data.rel.ro section for init-only files
Clang-3.8 generates several .data.rel.ro sections when compiling Xen. As
these contain no global symbols, they should be .data.rel.ro.local. This
breaks the SPECIAL_DATA_SECTIONS check when converting the transition units to
being init-only.
For alternatives.c, explicitly move the nops arrays into __initconst. For efi
boot.c, manually create the optimisation performed by Clang by collapsing the
switch statement into a lookup table. The double use of const is required to
avoid breaking the ARM build by creating a section type conflict with
fdt_guid.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Haozhong Zhang [Thu, 25 Feb 2016 12:00:11 +0000 (13:00 +0100)]
x86/hvm: collect information of TSC scaling ratio
Both VMX TSC scaling and SVM TSC ratio use the 64-bit TSC scaling ratio,
but the number of fractional bits of the ratio is different between VMX
and SVM. This patch adds the architecture code to collect the number of
fractional bits and other related information into fields of struct
hvm_function_table so that they can be used in the common code.
Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Doug Goldstein [Wed, 24 Feb 2016 11:06:28 +0000 (12:06 +0100)]
xenoprof: drop unnecessary macro
This macro doesn't really provide a benefit. When support is added the
implementer can implement this how it needs to be and not conform to the
macro. Additionally this change limits the output of the warning to just
once instead of nrpages worth. While editing this area I dropped
trailing whitespace.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Juergen Gross [Wed, 24 Feb 2016 11:05:58 +0000 (12:05 +0100)]
use XEN_SYSCTL_SCHEDOP_* for sysctl operation checks
In flask_sysctl_scheduler_op() and sched_adjust_global() the test for
the desired operation is done with the wrong constants. While the
values are correct, the names are not.
Correct the error message for the case of an unknown operation in
flask_sysctl_scheduler_op(), too.
Signed-off-by: Juergen Gross <jgross@suse.com> Reviewed-by: George Dunlap <george.dunlap@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Huaitong Han [Wed, 24 Feb 2016 11:04:50 +0000 (12:04 +0100)]
x86/hvm: add xstate support for pkeys
The XSAVE feature set can operate on PKRU state only if the feature set is
enabled (CR4.OSXSAVE = 1) and has been configured to manage PKRU state
(XCR0[9] = 1). And XCR0.PKRU is disabled on PV mode without PKU feature
enabled.
Signed-off-by: Huaitong Han <huaitong.han@intel.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Huaitong Han [Wed, 24 Feb 2016 11:04:19 +0000 (12:04 +0100)]
x86/hvm: add pkeys support for guest_walk_tables
Protection keys define a new 4-bit protection key field(PKEY) in bits 62:59 of
leaf entries of the page tables.
PKRU register defines 32 bits, there are 16 domains and 2 attribute bits per
domain in pkru, for each i (0 = i = 15), PKRU[2i] is the access-disable bit for
protection key i (ADi); PKRU[2i+1] is the write-disable bit for protection key
i (WDi). PKEY is index to a defined domain.
A fault is considered as a PKU violation if all of the following conditions are
true:
1.CR4_PKE=1.
2.EFER_LMA=1.
3.Page is present with no reserved bit violations.
4.The access is not an instruction fetch.
5.The access is to a user page.
6.PKRU.AD=1
or The access is a data write and PKRU.WD=1
and either CR0.WP=1 or it is a user access.
Signed-off-by: Huaitong Han <huaitong.han@intel.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> Reviewed-by: George Dunlap <george.dunlap@citrix.com>
Dario Faggioli [Wed, 24 Feb 2016 11:03:32 +0000 (12:03 +0100)]
credit1: trace vCPU boost/unboost
Add tracepoints and a performance counter for
boosting and unboosting in Credit1.
Note that they (the trace points) do not cover
the case of the idle vCPU being boosted to run
a tasklet, as there already is
TRC_CSCHED_SCHED_TASKLET for that.
Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com> Reviewed-by: George Dunlap <george.dunlap@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
Dario Faggioli [Wed, 24 Feb 2016 11:02:37 +0000 (12:02 +0100)]
sched: get rid of static private schedulers' structures
In fact, they look rather useless: they are never
referenced neither directly, nor via the sched_data
pointer, as a dynamic copy that overrides them is
allocated as the very first step of a scheduler's
initialization.
While there, take the chance to also reset the sched_data
pointer to NULL, upon scheduler de-initialization.
Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com> Acked-by: George Dunlap <george.dunlap@citrix.com>
Jan Beulich [Mon, 22 Feb 2016 16:38:34 +0000 (17:38 +0100)]
common: re-arrange struct kernel_param fields
Even if placed in .init.* there's no reason to needlessly bloat the
binary due to padding fields the compiler needs to insert on 64-bit
architectures.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Tamas K Lengyel [Mon, 22 Feb 2016 16:24:15 +0000 (17:24 +0100)]
x86/vm_event: consolidate hvm_event_fill_regs and p2m_vm_event_fill_regs
Currently the registers saved in the request depend on which type of event
is filling in the registers. In this patch we consolidate the two versions
of register filling function as to return a fix set of registers irrespective
of the underlying event.
Signed-off-by: Tamas K Lengyel <tlengyel@novetta.com> Acked-by: George Dunlap <george.dunlap@citrix.com> Acked-by: Razvan Cojocaru <rcojocaru@bitdefender.com>
Jan Beulich [Mon, 22 Feb 2016 16:23:08 +0000 (17:23 +0100)]
x86: drop register reload from INT80 malicious MSI guard
None of the restored registers are actually of interest to the
subsequent code (as opposed to the similar construct on the compat
mode hypercall path).
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Roger Pau Monné [Mon, 22 Feb 2016 16:20:37 +0000 (17:20 +0100)]
x86/PVHv2: add XEN_ prefix to HVM_START_MAGIC_VALUE
Reported by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Andrew Cooper [Mon, 22 Feb 2016 16:17:18 +0000 (17:17 +0100)]
init: annotate all command line parameter infrastructure as const
There is no reason for any of it to be modified. Additionally, link
.init.setup beside the other constant .init data.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
[jb: reduce alignments to 8] Reviewed-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Doug Goldstein [Fri, 19 Feb 2016 02:57:03 +0000 (20:57 -0600)]
m4/python: fix typo in LDFLAGS variable name
[ also, reran autogen.sh ]
Reported-by: Jonathan Creekmore <jonathan.creekmore@gmail.com> Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Doug Goldstein [Fri, 19 Feb 2016 19:55:49 +0000 (13:55 -0600)]
MAINTAINERS: add Doug Goldstein for Travis CI config
Add myself as the maintainer for the Travis CI config.
Signed-off-by: Doug Goldstein <cardoe@cardoe.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Andrew Cooper [Fri, 12 Feb 2016 19:06:48 +0000 (19:06 +0000)]
tools/xenalyze: Fix build with clang
1) EXIT_REASON_EXCEPTION_NMI is 0, and Clang complains:
xenalyze.c:513:33: error: initializer overrides prior initialization of this subobject [-Werror,-Winitializer-overrides]
[EXIT_REASON_EXCEPTION_NMI]="EXCEPTION_NMI",
^~~~~~~~~~~~~~~
xenalyze.c:512:11: note: previous initialization is here
[0] = "NONE",
^~~~~~
2) cr3_time_compare(), eip_compare(), ipi_send() and cr3_compare_start() are
declared as nested functions, which is a GCCism not supported by Clang.
As they don't actually make use of the interesting feature offered by
nested functions (i.e. dynamic scoping), move them to just being normal
functions.
3) clear_interval_summary(), update_cpi() and clear_interval_cpi() are all
unused. The former isn't reference anywhere, so is deleted, while the other
two are called from currently #if 0'd code.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: George Dunlap <george.dunlap@citrix.com>
Ian Campbell [Wed, 17 Feb 2016 15:39:55 +0000 (15:39 +0000)]
xenpaging: don't try to log via xch if xc_interface_close fails
Since xch may not be valid (enough) any longer, xc_interface_close
already logs anything of any use before it tears down the integrated
logger so there is no need to log any further in the application via
that path.
Ian Campbell [Wed, 17 Feb 2016 14:30:38 +0000 (14:30 +0000)]
tools: gtracestat: make all functions and global data static
After "Drop unused functions do_cstate and single_cstate helper" make
all the remaining functions and global data static and in the process
allow the compiler to notice that cond_rec_init() is also unused, thus
remove it.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Campbell [Wed, 17 Feb 2016 14:30:37 +0000 (14:30 +0000)]
tools: gtracestat: Drop unused functions do_cstate and single_cstate helper
These have always been dead code since the code was added AFAICT.
This eliminates the code containing CID 10567079, 10567080, 10567081
and 10567082 (all apparently some confusion between max_cx_num vs
MAX_CX_NR, but given the lack of callers its hard to tell what was
intended)
Signed-off-by: Ian Campbell <ian.campbell@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Ian Campbell [Wed, 17 Feb 2016 14:04:15 +0000 (14:04 +0000)]
xl: create: close restore_fd_to_close on error
Currently the fd is opened and then later closed and
restore_fd_to_close set back to -1, however there are several goto out
and goto error_out paths in the interim.
Since the code resets restore_fd_to_close to -1 it is OK to check this
and close on the out path too.
Ian Jackson [Thu, 18 Feb 2016 12:37:04 +0000 (12:37 +0000)]
tools: libxl: Simplify logic in libxl__realloc
Replace the loop exit and separate test for loop overrun with an
assert in the loop body.
This simplifies the code. It also (hopefully) avoids Coverity
thinking that gc->alloc_maxsize might change, resulting in the loop
failing to find the right answer but also failing to abort.
(gc->alloc_maxsize can't change because gcs are all singlethreaded:
either they are on the stack of a specific thread, or they belong to
an ao and are covered by the ctx lock.)
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com> Acked-by: Ian Campbell <ian.campbell@citrix.com>
Corneliu ZUZU [Thu, 18 Feb 2016 16:47:36 +0000 (17:47 +0100)]
x86/monitor: minor left-shift undefined behavior checks
This minor patch adds a range-check to avoid left-shift caused undefined
behavior. Also replaces '1 <<' w/ '1U <<' @ x86 monitor.h in an effort to avoid
a future potential '1 << 31' that would cause a similar issue.
Corneliu ZUZU [Thu, 18 Feb 2016 14:08:25 +0000 (15:08 +0100)]
x86/hvm_event: fix uninitialized struct field usage introduced by c/s f5365e6
c/s f5365e6: "xen/vm-events: Move parts of monitor_domctl code to common-side",
introduced a use without initialization issue.
hvm_event_breakpoint calls hvm_event_traps(&req) and if sync is true that
ors some bits into req->flags which was never initialised.
Reported by Coverity Scan.
Initializes req @ hvm_event_breakpoint entry.
Coverity-ID: 1353192 Signed-off-by: Corneliu ZUZU <czuzu@bitdefender.com> Acked-by: Razvan Cojocaru <rcojocaru@bitdefender.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Jan Beulich [Thu, 18 Feb 2016 14:07:33 +0000 (15:07 +0100)]
x86: drop failsafe callback invocation from assembly
Afaict this was never necessary on a 64-bit hypervisor, and was instead
just blindly cloned over from 32-bit code: We don't fiddle with (and
hence don't reload) any of DS, ES, FS, or GS, and an exception on IRET
itself can equally well be reported to the guest as that very exception
on the target of that IRET.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Jan Beulich [Thu, 18 Feb 2016 14:05:34 +0000 (15:05 +0100)]
x86emul: fix rIP handling
Deal with rIP just like with any other register: Truncate to designated
width upon entry, write back the zero-extended 32-bit value when
emulating 32-bit code, and leave the upper 48 bits unchanged for 16-bit
code.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Dario Faggioli [Thu, 18 Feb 2016 14:04:23 +0000 (15:04 +0100)]
RTDS: pack trace data better for xentrace_format
when tracing runstate changes, the vcpu and domain IDs
are encoded in the lower and higher, respectively, parts
of a 32 bits integer. When decoding a trace with
xentrace_format, this makes it possible to display
such events like this:
For consistency, we should do the same when displaying
the events coming from the RTDS scheduler (when using
the same tool), and to do that, we need to invert the
order in which the fields are being put in the trace
struct right now.
While there, we also:
- fix the use of TRC_RTDS_SCHED_TASKLET (it should
only be involved when a tasklet is scheduled, not
_every_ time rt_schedule() is invoked!);
- remove a very chatty and useless (nothing has been
picked!) use of TRC_RTDS_RUNQ_PICK.
In fact, one can already figure out when nothing has been
picked from the runqueue, by looking at when cpu_idle
is invoked --which is the same thing one would do if on
Credit or Credit2.
Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: George Dunlap <george.dunlap@citrix.com>
projects / people / liuw / xen.git / log