Jan Beulich [Wed, 19 Jun 2019 07:19:50 +0000 (09:19 +0200)]
hvmloader: cater for lack of hvm_start_info
It wasn't obvious at all that the backport of commit 0d6968635c
("hvmloader: avoid tests when they would clobber used memory") has
further dependencies on changes all having appeared in the 4.8 dev
cycle. Granted I should have done a build test, yet I routinely do so
only on the hypervisor side ...
Jan Beulich [Mon, 17 Jun 2019 06:54:58 +0000 (08:54 +0200)]
hvmloader: avoid tests when they would clobber used memory
First of all limit the memory range used for testing to 4Mb: There's no
point placing page tables right above 8Mb when they can equally well
live at the bottom of the chunk at 4Mb - rep_io_test() cares about the
5Mb...7Mb range only anyway. In a subsequent patch this will then also
allow simply looking for an unused 4Mb range (instead of using a build
time determined one).
Extend the "skip tests" condition beyond the "is there enough memory"
question.
Reported-by: Charles Arnold <carnold@suse.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Tested-by: Gary Lin <glin@suse.com> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 0d6968635ce51a8ed7508ddcf17b3d13a462cb27
master date: 2017-05-19 16:04:38 +0200
Chris Patterson [Wed, 27 Jul 2016 20:01:26 +0000 (16:01 -0400)]
libxl: compilation warning fix for arm & aarch64
GCC 6 will warn on unused static const variables in c modules:
https://gcc.gnu.org/ml/gcc-patches/2015-09/msg00847.html
When compiling with LIBXL_HAVE_NO_SUSPEND_RESUME set (arm & aarch64),
the compiler emits the following errors:
xl_cmdimpl.c:101:19: error: 'migrate_report'
defined but not used [-Werror=unused-const-variable=]
xl_cmdimpl.c:99:19: error: 'migrate_permission_to_go'
defined but not used [-Werror=unused-const-variable=]
xl_cmdimpl.c:97:19: error: 'migrate_receiver_ready'
defined but not used [-Werror=unused-const-variable=]
xl_cmdimpl.c:95:19: error: 'migrate_receiver_banner'
defined but not used [-Werror=unused-const-variable=]
These unused const variables are only used in functions which exist between
the ifndef block:
#ifndef LIBXL_HAVE_NO_SUSPEND_RESUME
...
#endif
Wrap the same ifndef around these variables.
Signed-off-by: Chris Patterson <pattersonc@ainfosec.com> Acked-by: Wei Liu <wei.liu2@citrix.com>
(cherry picked from commit 5f28de0b0e474e01931b719fa27ca30b8aa446e0)
Wei Liu [Mon, 10 Oct 2016 12:50:58 +0000 (13:50 +0100)]
ipxe: update to newer commit
The current commit in tree is rather old. It has come to a point that
cherry-picking commits from upstream isn't trivial anymore.
There is long term plan to track ipxe upstream, but for 4.8 release, we
should just update ipxe to a newer commit (they are using rolling
release model now).
Forward-port the one boot prompt patch that is still relevant and retire
the rest which are already in upstream.
Reported-by: Juergen Schinker <ba1020@homie.homelinux.net> Signed-off-by: Wei Liu <wei.liu2@citrix.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
(cherry picked from commit 38ab99b26bf4298a33105ec66f3f6a3f7e05a326)
Juergen Gross [Mon, 31 Oct 2016 09:04:18 +0000 (10:04 +0100)]
stubdom: fix stubdom-vtpm build
stubdom-vtpm needs gmp and expects it under
stubdom/cross-root-x86_64/x86_64-xen-elf/lib while gmp seems to install
it under stubdom/cross-root-x86_64/x86_64-xen-elf/lib64 at least in an
openSUSE environment.
Modify gmp configure parameters to explicitly specify --libdir.
Signed-off-by: Juergen Gross <jgross@suse.com> Acked-by: Wei Liu <wei.liu2@citrix.com> Release-acked-by: Wei Liu <wei.liu2@citrix.com>
(cherry picked from commit 7791790c7ab97c85306dce749c6c8eb56d1dc0da)
Wei Liu [Sat, 29 Oct 2016 17:22:38 +0000 (18:22 +0100)]
stubdom: make GMP aware that it's being cross-compiled
Append --build and --host flags to GMP's configure script so that it
knows it is being cross-compiled.
This should fix the issue that GMP doesn't compile with gcc 6, because
configure script won't try to test the host environment anymore.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Reviewed-by: Juergen Gross <jgross@suse.com> Acked-by: Ian Jackson <ian.jackson@eu.citrix.com> Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org> Release-acked-by: Wei Liu <wei.liu2@citrix.com>
(cherry picked from commit 668e4edf92fcf7cb929eed221059a3eeb02722c3)
Jan Beulich [Tue, 24 Jan 2017 16:22:03 +0000 (16:22 +0000)]
x86emul/test: don't use *_len symbols
... as they don't work as intended with -fPIC.
I did prefer them over *_end ones at the time because older gcc would
cause .L* symbols to be public, due to issuing .globl for all
referenced externals. And labels at the end of instructions collide
with the ones at the start of the next instruction, making disassembly
harder to grok. Luckily recent gcc no longer issues those .globl
directives, and hence .L* labels, staying local by default, no longer
get in the way.
Reported-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> Tested-by: Wei Liu <wei.liu2@citrix.com>
(cherry picked from commit 9315fa0ef736d1153c98ce42bff5853da5ec697f)
Andrew Cooper [Wed, 12 Dec 2018 19:22:15 +0000 (19:22 +0000)]
x86/spec-ctrl: Introduce options to control VERW flushing
The Microarchitectural Data Sampling vulnerability is split into categories
with subtly different properties:
MLPDS - Microarchitectural Load Port Data Sampling
MSBDS - Microarchitectural Store Buffer Data Sampling
MFBDS - Microarchitectural Fill Buffer Data Sampling
MDSUM - Microarchitectural Data Sampling Uncacheable Memory
MDSUM is a special case of the other three, and isn't distinguished further.
These issues pertain to three microarchitectural buffers. The Load Ports, the
Store Buffers and the Fill Buffers. Each of these structures are flushed by
the new enhanced VERW functionality, but the conditions under which flushing
is necessary vary.
For this concise overview of the issues and default logic, the abbreviations
SP (Store Port), FB (Fill Buffer), LP (Load Port) and HT (Hyperthreading) are
used for brevity:
* Vulnerable hardware is divided into two categories - parts which suffer
from SP only, and parts with any other combination of vulnerabilities.
* SP only has an HT interaction when the thread goes idle, due to the static
partitioning of resources. LP and FB have HT interactions at all points,
due to the competitive sharing of resources. All issues potentially leak
data across the return-to-guest transition.
* The microcode which implements VERW flushing also extends MSR_FLUSH_CMD, so
we don't need to do both on the HVM return-to-guest path. However, the
Knights range of processors are immune to L1TF (therefore have no
MSR_FLUSH_CMD), but are vulnerable to MDS, so do require VERW on the HVM
path.
Note that we deliberately support mds=1 even without MD_CLEAR in case the
microcode has been updated but the feature bit not exposed.
This is part of XSA-297, CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, CVE-2019-11091.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 3c04c258ab40405a74e194d9889a4cbc7abe94b4)
Andrew Cooper [Wed, 12 Dec 2018 19:22:15 +0000 (19:22 +0000)]
x86/spec-ctrl: Infrastructure to use VERW to flush pipeline buffers
Three synthetic features are introduced, as we need individual control of
each, depending on circumstances. A later change will enable them at
appropriate points.
The verw_sel field doesn't strictly need to live in struct cpu_info. It lives
there because there is a convenient hole it can fill, and it reduces the
complexity of the SPEC_CTRL_EXIT_TO_{PV,HVM} assembly by avoiding the need for
any temporary stack maintenance.
This is part of XSA-297, CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, CVE-2019-11091.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 548a932ac786d6bf3584e4b54f2ab993e1117710)
Andrew Cooper [Wed, 12 Sep 2018 13:36:00 +0000 (14:36 +0100)]
x86/spec-ctrl: CPUID/MSR definitions for Microarchitectural Data Sampling
The MD_CLEAR feature can be automatically offered to guests. No
infrastructure is needed in Xen to support the guest making use of it.
This is part of XSA-297, CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, CVE-2019-11091.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit d4f6116c080dc013cd1204c4d8ceb95e5f278689)
Andrew Cooper [Wed, 12 Sep 2018 13:36:00 +0000 (14:36 +0100)]
x86/spec-ctrl: Misc non-functional cleanup
* Identify BTI in the spec_ctrl_{enter,exit}_idle() comments, as other
mitigations will shortly appear.
* Use alternative_input() and cover the lack of memory cobber with a further
barrier.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 9b62eba6c429c327e1507816bef403ccc87357ae)
Andrew Cooper [Fri, 5 Apr 2019 12:26:30 +0000 (13:26 +0100)]
x86/boot: Detect the firmware SMT setting correctly on Intel hardware
While boot_cpu_data.x86_num_siblings is an accurate value to use on AMD
hardware, it isn't on Intel when the user has disabled Hyperthreading in the
firmware. As a result, a user which has chosen to disable HT still gets
nagged on L1TF-vulnerable hardware when they haven't chosen an explicit
smt=<bool> setting.
Make use of the largely-undocumented MSR_INTEL_CORE_THREAD_COUNT which in
practice exists since Nehalem, when booting on real hardware. Fall back to
using the ACPI table APIC IDs.
While adjusting this logic, fix a latent bug in amd_get_topology(). The
thread count field in CPUID.0x8000001e.ebx is documented as 8 bits wide,
rather than 2 bits wide.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit b12fec4a125950240573ea32f65c61fb9afa74c3)
Andrew Cooper [Fri, 5 Apr 2019 12:26:30 +0000 (12:26 +0000)]
x86/msr: Definitions for MSR_INTEL_CORE_THREAD_COUNT
This is a model specific register which details the current configuration
cores and threads in the package. Because of how Hyperthread and Core
configuration works works in firmware, the MSR it is de-facto constant and
will remain unchanged until the next system reset.
It is a read only MSR (so unilaterally reject writes), but for now retain its
leaky-on-read properties. Further CPUID/MSR work is required before we can
start virtualising a consistent topology to the guest, and retaining the old
behaviour is the safest course of action.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit d4120936bcd1695faf5b575f1259c58e31d2b18b)
Andrew Cooper [Wed, 12 Sep 2018 13:36:00 +0000 (14:36 +0100)]
x86/spec-ctrl: Reposition the XPTI command line parsing logic
It has ended up in the middle of the mitigation calculation logic. Move it to
be beside the other command line parsing.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit c2c2bb0d60c642e64a5243a79c8b1548ffb7bc5b)
Andrew Cooper [Mon, 18 Mar 2019 16:08:25 +0000 (17:08 +0100)]
x86/tsx: Implement controls for RTM force-abort mode
The CPUID bit and MSR are deliberately not exposed to guests, because they
won't exist on newer processors. As vPMU isn't security supported, the
misbehaviour of PCR3 isn't expected to impact production deployments.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 6be613f29b4205349275d24367bd4c82fb2960dd
master date: 2019-03-12 17:05:21 +0000
Cédric Bosdonnat [Thu, 10 Nov 2016 09:23:31 +0000 (10:23 +0100)]
Fix misleading indentation warnings
Gcc6 build reports misleading indentation as warnings. Fix a few
warnings in stubdom.
Signed-off-by: Cédric Bosdonnat <cbosdonnat@suse.com> Acked-by: Daniel De Graaf <dgdegra@tycho.nsa.gov> Acked-by: Quan Xu <xuquan8@huawei.com> Release-acked-by: Wei Liu <wei.liu2@citrix.com>
(cherry picked from commit 9fdffbbab3ada427bac07076f042f0265e5ae05f)
Wei Liu [Wed, 28 Nov 2018 17:43:33 +0000 (17:43 +0000)]
tools/firmware: update OVMF Makefile, when necessary
[ This is two commits from master aka staging-4.12: ]
OVMF has become dependent on OpenSSL, which is included as a
submodule. Initialise submodules before building.
Signed-off-by: Wei Liu <wei.liu2@citrix.com> Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
(cherry picked from commit b16281870e06f5f526029a4e69634a16dc38e8e4)
tools: only call git when necessary in OVMF Makefile
Users may choose to export a snapshot of OVMF and build it
with xen.git supplied ovmf-makefile. In that case we don't
need to call `git submodule`.
Jan Beulich [Tue, 5 Mar 2019 14:57:42 +0000 (15:57 +0100)]
x86/pv: _toggle_guest_pt() may not skip TLB flush for shadow mode guests
For shadow mode guests (e.g. PV ones forced into that mode as L1TF
mitigation, or during migration) update_cr3() -> sh_update_cr3() may
result in a change to the (shadow) root page table (compared to the
previous one when running the same vCPU with the same PCID). This can,
first and foremost, be a result of memory pressure on the shadow memory
pool of the domain. Shadow code legitimately relies on the original
(prior to commit 5c81d260c2 ["xen/x86: use PCID feature"]) behavior of
the subsequent CR3 write to flush the TLB of entries still left from
walks with an earlier, different (shadow) root page table.
Restore the flushing behavior, also for the second CR3 write on the exit
path to guest context when XPTI is active. For the moment accept that
this will introduce more flushes than are strictly necessary - no flush
would be needed when the (shadow) root page table doesn't actually
change, but this information isn't readily (i.e. without introducing a
layering violation) available here.
This is XSA-294.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Tested-by: Juergen Gross <jgross@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 329b00e4d49f70185561d7cc4b076c77869888a0
master date: 2019-03-05 13:54:42 +0100
Andrew Cooper [Tue, 5 Mar 2019 14:57:06 +0000 (15:57 +0100)]
x86/pv: Don't have %cr4.fsgsbase active behind a guest kernels back
Currently, a 64bit PV guest can appear to set and clear FSGSBASE in %cr4, but
the bit remains set in hardware. Therefore, the {RD,WR}{FS,GS}BASE are usable
even when the guest kernel believes that they are disabled.
The FSGSBASE feature isn't currently supported in Linux, and its context
switch path has some optimisations which rely on userspace being unable to use
the WR{FS,GS}BASE instructions. Xen's current behaviour undermines this
expectation.
In 64bit PV guest context, always load the guest kernels setting of FSGSBASE
into %cr4. This requires adjusting how Xen uses the {RD,WR}{FS,GS}BASE
instructions.
* Delete the cpu_has_fsgsbase helper. It is no longer safe, as users need to
check %cr4 directly.
* The raw __rd{fs,gs}base() helpers are only safe to use when %cr4.fsgsbase
is set. Comment this property.
* The {rd,wr}{fs,gs}{base,shadow}() and read_msr() helpers are updated to use
the current %cr4 value to determine which mechanism to use.
* toggle_guest_mode() and save_segments() are update to avoid reading
fs/gsbase if the values in hardware cannot be stale WRT struct vcpu. A
consequence of this is that the write_cr() path needs to cache the current
bases, as subsequent context switches will skip saving the values.
* write_cr4() is updated to ensure that the shadow %cr4.fsgsbase value is
observed in a safe way WRT the hardware setting, if an interrupt happens to
hit in the middle.
* load_segments() is updated to use the VMLOAD optimisation if FSGSBASE is
unavailable, even if only gs_shadow needs updating. As a minor perf
improvement, check cpu_has_svm first to short circuit a context-dependent
conditional on Intel hardware.
* pv_make_cr4() is updated for 64bit PV guests to use the guest kernels
choice of FSGSBASE.
This is part of XSA-293.
Reported-by: Andy Lutomirski <luto@kernel.org> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: eccc170053e46b4ab1d9e7485c09e210be15bbd7
master date: 2019-03-05 13:54:05 +0100
Andrew Cooper [Tue, 5 Mar 2019 14:56:33 +0000 (15:56 +0100)]
x86/pv: Rewrite guest %cr4 handling from scratch
The PV cr4 logic is almost impossible to follow, and leaks bits into guest
context which definitely shouldn't be visible (in particular, VMXE).
The biggest problem however, and source of the complexity, is that it derives
new real and guest cr4 values from the current value in hardware - this is
context dependent and an inappropriate source of information.
Rewrite the cr4 logic to be invariant of the current value in hardware.
First of all, modify write_ptbase() to always use mmu_cr4_features for IDLE
and HVM contexts. mmu_cr4_features *is* the correct value to use, and makes
the ASSERT() obviously redundant.
For PV guests, curr->arch.pv.ctrlreg[4] remains the guests view of cr4, but
all logic gets reworked in terms of this and mmu_cr4_features only.
Two masks are introduced; bits which the guest has control over, and bits
which are forwarded from Xen's settings. One guest-visible change here is
that Xen's VMXE setting is no longer visible at all.
pv_make_cr4() follows fairly closely from pv_guest_cr4_to_real_cr4(), but
deliberately starts with mmu_cr4_features, and only alters the minimal subset
of bits.
The boot-time {compat_,}pv_cr4_mask variables are removed, as they are a
remnant of the pre-CPUID policy days. pv_fixup_guest_cr4() gains a related
derivation from the policy.
Another guest visible change here is that a 32bit PV guest can now flip
FSGSBASE in its view of CR4. While the {RD,WR}{FS,GS}BASE instructions are
unusable outside of a 64bit code segment, the ability to modify FSGSBASE
matches real hardware behaviour, and avoids the need for any 32bit/64bit
differences in the logic.
Overall, this patch shouldn't have a practical change in guest behaviour.
VMXE will disappear from view, and an inquisitive 32bit kernel can now see
FSGSBASE changing, but this new logic is otherwise bug-compatible with before.
This is part of XSA-293.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: b2dd00574a4fc87ca964177f8e752a968c27efb2
master date: 2019-03-05 13:53:32 +0100
Andrew Cooper [Tue, 5 Mar 2019 14:55:55 +0000 (15:55 +0100)]
x86/pv: Improve pv_cpuid()'s API
pv_cpuid()'s API is awkward to use. There are already two callers jumping
through hoops to use it, and a third is on its way.
Change the API to take each parameter individually (like its counterpart,
hvm_cpuid(), already does), and introduce a new pv_cpuid_regs() wrapper
implementing the old API.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Jan Beulich [Tue, 5 Mar 2019 14:55:25 +0000 (15:55 +0100)]
x86/mm: properly flush TLB in switch_cr3_cr4()
The CR3 values used for contexts run with PCID enabled uniformly have
CR3.NOFLUSH set, resulting in the CR3 write itself to not cause any
flushing at all. When the second CR4 write is skipped or doesn't do any
flushing, there's nothing so far which would purge TLB entries which may
have accumulated again if the PCID doesn't change; the "just in case"
flush only affects the case where the PCID actually changes. (There may
be particularly many TLB entries re-accumulated in case of a watchdog
NMI kicking in during the critical time window.)
Suppress the no-flush behavior of the CR3 write in this particular case.
Similarly the second CR4 write may not cause any flushing of TLB entries
established again while the original PCID was still in use - it may get
performed because of unrelated bits changing. The flush of the old PCID
needs to happen nevertheless.
At the same time also eliminate a possible race with lazy context
switch: Just like for CR4, CR3 may change at any time while interrupts
are enabled, due to the __sync_local_execstate() invocation from the
flush IPI handler. It is for that reason that the CR3 read, just like
the CR4 one, must happen only after interrupts have been turned off.
This is XSA-292.
Reported-by: Sergey Dyasli <sergey.dyasli@citrix.com> Reported-by: Andrew Cooper <andrew.cooper3@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Tested-by: Sergey Dyasli <sergey.dyasli@citrix.com>
master commit: 6e5f22ba437d78c0a84b9673f7e2cfefdbc62f4b
master date: 2019-03-05 13:52:44 +0100
Jan Beulich [Tue, 5 Mar 2019 14:54:29 +0000 (15:54 +0100)]
x86/mm: add explicit preemption checks to L3 (un)validation
When recursive page tables are used at the L3 level, unvalidation of a
single L4 table may incur unvalidation of two levels of L3 tables, i.e.
a maximum iteration count of 512^3 for unvalidating an L4 table. The
preemption check in free_l2_table() as well as the one in
_put_page_type() may never be reached, so explicit checking is needed in
free_l3_table().
When recursive page tables are used at the L4 level, the iteration count
at L4 alone is capped at 512^2. As soon as a present L3 entry is hit
which itself needs unvalidation (and hence requiring another nested loop
with 512 iterations), the preemption checks added here kick in, so no
further preemption checking is needed at L4 (until we decide to permit
5-level paging for PV guests).
The validation side additions are done just for symmetry.
This is part of XSA-290.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: bac4567a67d5e8b916801ea5a04cf8b443dfb245
master date: 2019-03-05 13:51:46 +0100
Jan Beulich [Tue, 5 Mar 2019 14:53:56 +0000 (15:53 +0100)]
x86/mm: also allow L2 (un)validation to be fully preemptible
Commit c612481d1c ("x86/mm: Plumbing to allow any PTE update to fail
with -ERESTART") added assertions next to the {alloc,free}_l2_table()
invocations to document (and validate in debug builds) that L2
(un)validations are always preemptible.
The assertion in free_page_type() was now observed to trigger when
recursive L2 page tables get cleaned up.
In particular put_page_from_l2e()'s assumption that _put_page_type()
would always succeed is now wrong, resulting in a partially un-validated
page left in a domain, which has no other means of getting cleaned up
later on. If not causing any problems earlier, this would ultimately
trigger the check for ->u.inuse.type_info having a zero count when
freeing the page during cleanup after the domain has died.
As a result it should be considered a mistake to not have extended
preemption fully to L2 when it was added to L3/L4 table handling, which
this change aims to correct.
The validation side additions are done just for symmetry.
This is part of XSA-290.
Reported-by: Manuel Bouyer <bouyer@antioche.eu.org> Tested-by: Manuel Bouyer <bouyer@antioche.eu.org> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 176ebf9c8bc2828f6637eb61cc1cf166e302c699
master date: 2019-03-05 13:51:18 +0100
George Dunlap [Tue, 5 Mar 2019 14:53:15 +0000 (15:53 +0100)]
xen: Make coherent PV IOMMU discipline
In order for a PV domain to set up DMA from a passed-through device to
one of its pages, the page must be mapped in the IOMMU. On the other
hand, before a PV page may be used as a "special" page type (such as a
pagetable or descriptor table), it _must not_ be writable in the IOMMU
(otherwise a malicious guest could DMA arbitrary page tables into the
memory, bypassing Xen's safety checks); and Xen's current rule is to
have such pages not in the IOMMU at all.
At the moment, in order to accomplish this, the code borrows HVM
domain's "physmap" concept: When a page is assigned to a guest,
guess_physmap_add_entry() is called, which for PV guests, will create
a writable IOMMU mapping; and when a page is removed,
guest_physmap_remove_entry() is called, which will remove the mapping.
Additionally, when a page gains the PGT_writable page type, the page
will be added into the IOMMU; and when the page changes away from a
PGT_writable type, the page will be removed from the IOMMU.
Unfortunately, borrowing the "physmap" concept from HVM domains is
problematic. HVM domains have a lock on their p2m tables, ensuring
synchronization between modifications to the p2m; and all hypercall
parameters must first be translated through the p2m before being used.
Trying to mix this locked-and-gated approach with PV's lock-free
approach leads to several races and inconsistencies:
* A race between a page being assigned and it being put into the
physmap; for example:
- P1: call populate_physmap() { A = allocate_domheap_pages() }
- P2: Guess page A's mfn, and call decrease_reservation(A). A is owned by the domain,
and so Xen will clear the PGC_allocated bit and free the page
- P1: finishes populate_physmap() { guest_physmap_add_entry() }
Now the domain has a writable IOMMU mapping to a page it no longer owns.
* Pages start out as type PGT_none, but with a writable IOMMU mapping.
If a guest uses a page as a page table without ever having created a
writable mapping, the IOMMU mapping will not be removed; the guest
will have a writable IOMMU mapping to a page it is currently using
as a page table.
* A newly-allocated page can be DMA'd into with no special actions on
the part of the guest; However, if a page is promoted to a
non-writable type, the page must be mapped with a writable type before
DMA'ing to it again, or the transaction will fail.
To fix this, do away with the "PV physmap" concept entirely, and
replace it with the following IOMMU discipline for PV guests:
- (type == PGT_writable) <=> in iommu (even if type_count == 0)
- Upon a final put_page(), check to see if type is PGT_writable; if so,
iommu_unmap.
In order to achieve that:
- Remove PV IOMMU related code from guest_physmap_*
- Repurpose cleanup_page_cacheattr() into a general
cleanup_page_mappings() function, which will both fix up Xen
mappings for pages with special cache attributes, and also check for
a PGT_writable type and remove pages if appropriate.
- For compatibility with current guests, grab-and-release a
PGT_writable_page type for PV guests in guest_physmap_add_entry().
This will cause most "normal" guest pages to start out life with
PGT_writable_page type (and thus an IOMMU mapping), but no type
count (so that they can be used as special cases at will).
Also, note that there is one exception to to the "PGT_writable => in
iommu" rule: xenheap pages shared with guests may be given a
PGT_writable type with one type reference. This reference prevents
the type from changing, which in turn prevents page from gaining an
IOMMU mapping in get_page_type(). It's not clear whether this was
intentional or not, but it's not something to change in a security
update.
This is XSA-288.
Reported-by: Paul Durrant <paul.durrant@citrix.com> Signed-off-by: George Dunlap <george.dunlap@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com>
master commit: fe21b78ef99a1b505cfb6d3789ede9591609dd70
master date: 2019-03-05 13:48:32 +0100
steal_page, in a misguided attempt to protect against unknown races,
violates both of these rules, thus introducing other races:
- Temporarily, the count_info has the refcount go to 0 while
PGC_allocated is set
- It explicitly returns the page PGC_allocated set, but owner == NULL
and page not on the page_list.
The second one meant that page_get_owner_and_reference() could return
NULL even after having successfully grabbed a reference on the page,
leading the caller to leak the reference (since "couldn't get ref" and
"got ref but no owner" look the same).
Furthermore, rather than grabbing a page reference to ensure that the
owner doesn't change under its feet, it appears to rely on holding
d->page_alloc lock to prevent this.
Unfortunately, this is ineffective: page->owner remains non-NULL for
some time after the count has been set to 0; meaning that it would be
entirely possible for the page to be freed and re-allocated to a
different domain between the page_get_owner() check and the count_info
check.
Modify steal_page to instead follow the appropriate access discipline,
taking the page through series of states similar to being freed and
then re-allocated with MEMF_no_owner:
- Grab an extra reference to make sure we don't race with anyone else
freeing the page
- Drop both references and PGC_allocated atomically, so that (if
successful), anyone else trying to grab a reference will fail
- Attempt to reset Xen's mappings
- Reset the rest of the state.
Then, modify the two callers appropriately:
- Leave count_info alone (it's already been cleared)
- Call free_domheap_page() directly if appropriate
- Call assign_pages() rather than open-coding a partial assign
With all callers to assign_pages() now passing in pages with the
type_info field clear, tighten the respective assertion there.
This is XSA-287.
Signed-off-by: George Dunlap <george.dunlap@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com>
master commit: 3d4868a481eebed232eeacba36ea28e5dee5e946
master date: 2019-03-05 13:48:08 +0100
Jan Beulich [Tue, 5 Mar 2019 14:52:19 +0000 (15:52 +0100)]
IOMMU/x86: fix type ref-counting race upon IOMMU page table construction
When arch_iommu_populate_page_table() gets invoked for an already
running guest, simply looking at page types once isn't enough, as they
may change at any time. Add logic to re-check the type after having
mapped the page, unmapping it again if needed.
This is XSA-285.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Tentatively-Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 1f0b0bb7773d537bcf169e021495d0986d9809fc
master date: 2019-03-05 13:47:36 +0100
Jan Beulich [Tue, 5 Mar 2019 14:51:44 +0000 (15:51 +0100)]
gnttab: set page refcount for copy-on-grant-transfer
Commit 5cc77f9098 ("32-on-64: Fix domain address-size clamping,
implement"), which introduced this functionality, took care of clearing
the old page's PGC_allocated, but failed to set the bit (and install the
associated reference) on the newly allocated one. Furthermore the "mfn"
local variable was never updated, and hence the wrong MFN was passed to
guest_physmap_add_page() (and back to the destination domain) in this
case, leading to an IOMMU mapping into an unowned page.
Ideally the code would use assign_pages(), but the call to
gnttab_prepare_for_transfer() sits in the middle of the actions
mirroring that function.
This is XSA-284.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: George Dunlap <george.dunlap@citrix.com>
master commit: 6d4f36c3fecc0a6a0991716199612c81d909316e
master date: 2019-03-05 13:45:58 +0100
Jan Beulich [Fri, 23 Nov 2018 10:54:35 +0000 (11:54 +0100)]
VMX: allow migration of guests with SSBD enabled
The backport of cd53023df9 ("x86/msr: Virtualise MSR_SPEC_CTRL.SSBD for
guests to use") did not mirror the PV side change into the HVM (VMX-
specific) code path.
Jan Beulich [Tue, 20 Nov 2018 15:05:36 +0000 (16:05 +0100)]
x86/shadow: shrink struct page_info's shadow_flags to 16 bits
This is to avoid it overlapping the linear_pt_count field needed for PV
domains. Introduce a separate, HVM-only pagetable_dying field to replace
the sole one left in the upper 16 bits.
Note that the accesses to ->shadow_flags in shadow_{pro,de}mote() get
switched to non-atomic, non-bitops operations, as {test,set,clear}_bit()
are not allowed on uint16_t fields and hence their use would have
required ugly casts. This is fine because all updates of the field ought
to occur with the paging lock held, and other updates of it use |= and
&= as well (i.e. using atomic operations here didn't really guard
against potentially racing updates elsewhere).
This is part of XSA-280.
Reported-by: Prgmr.com Security <security@prgmr.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Tim Deegan <tim@xen.org>
master commit: 789589968ed90e82a832dbc60e958c76b787be7e
master date: 2018-11-20 14:59:54 +0100
Jan Beulich [Tue, 20 Nov 2018 15:04:55 +0000 (16:04 +0100)]
x86/shadow: move OOS flag bit positions
In preparation of reducing struct page_info's shadow_flags field to 16
bits, lower the bit positions used for SHF_out_of_sync and
SHF_oos_may_write.
Instead of also adjusting the open coded use in _get_page_type(),
introduce shadow_prepare_page_type_change() to contain knowledge of the
bit positions to shadow code.
This is part of XSA-280.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Tim Deegan <tim@xen.org>
master commit: d68e1070c3e8f4af7a31040f08bdd98e6d6eac1d
master date: 2018-11-20 14:59:13 +0100
Andrew Cooper [Tue, 20 Nov 2018 15:04:10 +0000 (16:04 +0100)]
x86/mm: Don't perform flush after failing to update a guests L1e
If the L1e update hasn't occured, the flush cannot do anything useful. This
skips the potentially expensive vcpumask_to_pcpumask() conversion, and
broadcast TLB shootdown.
More importantly however, we might be in the error path due to a bad va
parameter from the guest, and this should not propagate into the TLB flushing
logic. The INVPCID instruction for example raises #GP for a non-canonical
address.
This is XSA-279.
Reported-by: Matthew Daley <mattd@bugfuzz.com> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 6c8d50288722672ecc8e19b0741a31b521d01706
master date: 2018-11-20 14:58:41 +0100
Jan Beulich [Tue, 20 Nov 2018 15:03:32 +0000 (16:03 +0100)]
AMD/IOMMU: suppress PTE merging after initial table creation
The logic is not fit for this purpose, so simply disable its use until
it can be fixed / replaced. Note that this re-enables merging for the
table creation case, which was disabled as a (perhaps unintended) side
effect of the earlier "amd/iommu: fix flush checks". It relies on no
page getting mapped more than once (with different properties) in this
process, as that would still be beyond what the merging logic can cope
with. But arch_iommu_populate_page_table() guarantees this afaict.
Roger Pau Monné [Tue, 20 Nov 2018 15:02:46 +0000 (16:02 +0100)]
amd/iommu: fix flush checks
Flush checking for AMD IOMMU didn't check whether the previous entry
was present, or whether the flags (writable/readable) changed in order
to decide whether a flush should be executed.
Fix this by taking the writable/readable/next-level fields into account,
together with the present bit.
Along these lines the flushing in amd_iommu_map_page() must not be
omitted for PV domains. The comment there was simply wrong: Mappings may
very well change, both their addresses and their permissions. Ultimately
this should honor iommu_dont_flush_iotlb, but to achieve this
amd_iommu_ops first needs to gain an .iotlb_flush hook.
Also make clear_iommu_pte_present() static, to demonstrate there's no
caller omitting the (subsequent) flush.
This is part of XSA-275.
Reported-by: Paul Durrant <paul.durrant@citrix.com> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com>
master commit: 1a7ffe466cd057daaef245b0a1ab6b82588e4c01
master date: 2018-11-20 14:52:12 +0100
Jan Beulich [Wed, 7 Nov 2018 08:54:57 +0000 (09:54 +0100)]
x86: work around HLE host lockup erratum
XACQUIRE prefixed accesses to the 4Mb range of memory starting at 1Gb
are liable to lock up the processor. Disallow use of this memory range.
Unfortunately the available Core Gen7 and Gen8 spec updates are pretty
old, so I can only guess that they're similarly affected when Core Gen6
is and the Xeon counterparts are, too.
This is part of XSA-282.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: cc76410d20aff2cc07b268b0713dc1d2740c6e12
master date: 2018-11-07 09:33:24 +0100
Ian Jackson [Tue, 18 Sep 2018 10:25:20 +0000 (11:25 +0100)]
stubdom/grub.patches: Drop docs changes, for licensing reasons
The patch file 00cvs is an import of a new upstream version of
grub1 from upstream CVS.
Unfortunately, in the period covered by the update, upstream changed
the documentation licence from a simple permissive licence, to the GNU
"Free Documentation Licence" with Front and Back Cover Texts.
The Debian Project is of the view that use the Front and Back Cover
Texts feature of the GFDL makes the resulting document not Free
Software, because of the mandatory redistribution of these immutable
texts. (Personally, I agree.)
This is awkward because Debian do not want to ship non-free content.
So the Debian maintainers need to launder the upstream source code, to
remove the troublesome files. This is an extra step when
incorporating new upstream versions. It's particularly annoying for
security response, which often involves rebasing onto a new upstream
release.
grub1 is obsolete and the last change to Xen's PV grub1 stubdom code
was in 2016. Furthermore, the grub1 documentation is not built and
installed by the Xen pv-grub stubdom Makefiles.
Therefore, remove all docs changes from stubdom/grub.patches. This
means that there are now no longer any GFDL-licenced grub docs in
xen.git.
There is no user impact, and Debian is helped. This change would
complicate any attempts to update to a new version of upstream grub1,
but it seems unlikely that such a thing will ever happen.
Wei Liu [Tue, 7 Aug 2018 14:35:34 +0000 (15:35 +0100)]
xl.conf: Add global affinity masks
XSA-273 involves one hyperthread being able to use Spectre-like
techniques to "spy" on another thread. The details are somewhat
complicated, but the upshot is that after all Xen-based mitigations
have been applied:
* PV guests cannot spy on sibling threads
* HVM guests can spy on sibling threads
(NB that for purposes of this vulnerability, PVH and HVM guests are
identical. Whenever this comment refers to 'HVM', this includes PVH.)
There are many possible mitigations to this, including disabling
hyperthreading entirely. But another solution would be:
* Specify some cores as PV-only, others as PV or HVM
* Allow HVM guests to only run on thread 0 of the "HVM-or-PV" cores
* Allow PV guests to run on the above cores, as well as any thread of the PV-only cores.
For example, suppose you had 16 threads across 8 cores (0-7). You
could specify 0-3 as PV-only, and 4-7 as HVM-or-PV. Then you'd set
the affinity of the HVM guests as follows (binary representation):
In order to make this easy, this patches introduces three "global affinity
masks", placed in xl.conf:
vm.cpumask
vm.hvm.cpumask
vm.pv.cpumask
These are parsed just like the 'cpus' and 'cpus_soft' options in the
per-domain xl configuration files. The resulting mask is AND-ed with
whatever mask results at the end of the xl configuration file.
`vm.cpumask` would be applied to all guest types, `vm.hvm.cpumask`
would be applied to HVM and PVH guest types, and `vm.pv.cpumask`
would be applied to PV guest types.
The idea would be that to implement the above mask across all your
VMs, you'd simply add the following two lines to the configuration
file:
Signed-off-by: George Dunlap <george.dunlap@citrix.com> Signed-off-by: Wei Liu <wei.liu2@citrix.com>
(cherry picked from commit aa67b97ed34279c43a43d9ca46727b5746caa92e)
PVH guest type in toolstack is not available in this version of Xen.
Change code and manpage to cope. Also xl is still part of libxl in
thsi version, manually backport code to relevant places.
Jan Beulich [Mon, 13 Aug 2018 11:07:23 +0000 (05:07 -0600)]
x86: Make "spec-ctrl=no" a global disable of all mitigations
In order to have a simple and easy to remember means to suppress all the
more or less recent workarounds for hardware vulnerabilities, force
settings not controlled by "spec-ctrl=" also to their original defaults,
unless they've been forced to specific values already by earlier command
line options.
This is part of XSA-273.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
(cherry picked from commit d8800a82c3840b06b17672eddee4878bbfdacc6d)
Andrew Cooper [Tue, 29 May 2018 17:44:16 +0000 (18:44 +0100)]
x86/spec-ctrl: Introduce an option to control L1D_FLUSH for HVM HAP guests
This mitigation requires up-to-date microcode, and is enabled by default on
affected hardware if available, and is used for HVM guests
The default for SMT/Hyperthreading is far more complicated to reason about,
not least because we don't know if the user is going to want to run any HVM
guests to begin with. If a explicit default isn't given, nag the user to
perform a risk assessment and choose an explicit default, and leave other
configuration to the toolstack.
This is part of XSA-273 / CVE-2018-3620.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 3bd36952dab60290f33d6791070b57920e10754b)
Andrew Cooper [Fri, 13 Apr 2018 15:34:01 +0000 (15:34 +0000)]
x86/msr: Virtualise MSR_FLUSH_CMD for guests
Guests (outside of the nested virt case, which isn't supported yet) don't need
L1D_FLUSH for their L1TF mitigations, but offering/emulating MSR_FLUSH_CMD is
easy and doesn't pose an issue for Xen.
The MSR is offered to HVM guests only. PV guests attempting to use it would
trap for emulation, and the L1D cache would fill long before the return to
guest context. As such, PV guests can't make any use of the L1D_FLUSH
functionality.
This is part of XSA-273 / CVE-2018-3646.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit fd9823faf9df057a69a9a53c2e100691d3f4267c)
Andrew Cooper [Wed, 28 Mar 2018 14:21:39 +0000 (15:21 +0100)]
x86/spec-ctrl: CPUID/MSR definitions for L1D_FLUSH
This is part of XSA-273 / CVE-2018-3646.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 3563fc2b2731a63fd7e8372ab0f5cef205bf8477)
x86/pv: Force a guest into shadow mode when it writes an L1TF-vulnerable PTE
See the comment in shadow.h for an explanation of L1TF and the safety
consideration of the PTEs.
In the case that CONFIG_SHADOW_PAGING isn't compiled in, crash the domain
instead. This allows well-behaved PV guests to function, while preventing
L1TF from being exploited. (Note: PV guest kernels which haven't been updated
with L1TF mitigations will likely be crashed as soon as they try paging a
piece of userspace out to disk.)
This is part of XSA-273 / CVE-2018-3620.
Signed-off-by: Juergen Gross <jgross@suse.com> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Tim Deegan <tim@xen.org> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 06e8b622d3f3c0fa5075e91b041c6f45549ad70a)
Andrew Cooper [Mon, 23 Jul 2018 06:11:40 +0000 (08:11 +0200)]
x86/mm: Plumbing to allow any PTE update to fail with -ERESTART
Switching to shadow mode is performed in tasklet context. To facilitate this,
we schedule the tasklet, then create a hypercall continuation to allow the
switch to take place.
As a consequence, the x86 mm code needs to cope with an L1e operation being
continuable. do_mmu{,ext}_op() may no longer assert that a continuation
doesn't happen on the final iteration.
To handle the arguments correctly on continuation, compat_update_va_mapping*()
may no longer call into their non-compat counterparts. Move the compat
functions into mm.c rather than exporting __do_update_va_mapping() and
{get,put}_pg_owner(), and fix an unsigned long/int inconsistency with
compat_update_va_mapping_otherdomain().
This is part of XSA-273 / CVE-2018-3620.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit c612481d1c9232c6abf91b03ec655e92f808805f)
x86/shadow: Infrastructure to force a PV guest into shadow mode
To mitigate L1TF, we cannot alter an architecturally-legitimate PTE a PV guest
chooses to write, but we can force the PV domain into shadow mode so Xen
controls the PTEs which are reachable by the CPU pagewalk.
Introduce new shadow mode, PG_SH_forced, and a tasklet to perform the
transition. Later patches will introduce the logic to enable this mode at the
appropriate time.
To simplify vcpu cleanup, make tasklet_kill() idempotent with respect to
tasklet_init(), which involves adding a helper to check for an uninitialised
list head.
This is part of XSA-273 / CVE-2018-3620.
Signed-off-by: Juergen Gross <jgross@suse.com> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Tim Deegan <tim@xen.org> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit b76ec3946bf6caca2c3950b857c008bc8db6723f)
Andrew Cooper [Mon, 23 Jul 2018 13:46:10 +0000 (13:46 +0000)]
x86/spec-ctrl: Introduce an option to control L1TF mitigation for PV guests
Shadowing a PV guest is only available when shadow paging is compiled in.
When shadow paging isn't available, guests can be crashed instead as
mitigation from Xen's point of view.
Ideally, dom0 would also be potentially-shadowed-by-default, but dom0 has
never been shadowed before, and there are some stability issues under
investigation.
This is part of XSA-273 / CVE-2018-3620.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 66a4e986819a86ba66ca2fe9d925e62a4fd30114)
Andrew Cooper [Wed, 25 Jul 2018 12:10:19 +0000 (12:10 +0000)]
x86/spec-ctrl: Calculate safe PTE addresses for L1TF mitigations
Safe PTE addresses for L1TF mitigations are ones which are within the L1D
address width (may be wider than reported in CPUID), and above the highest
cacheable RAM/NVDIMM/BAR/etc.
All logic here is best-effort heuristics, which should in practice be fine for
most hardware. Future work will see about disentangling the SRAT handling
further, as well as having L0 pass this information down to lower levels when
virtualised.
This is part of XSA-273 / CVE-2018-3620.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit b03a57c9383b32181e60add6b6de12b473652aa4)
Christian Lindig [Mon, 13 Aug 2018 16:26:56 +0000 (17:26 +0100)]
tools/oxenstored: Make evaluation order explicit
In Store.path_write(), Path.apply_modify() updates the node_created
reference and both the value of apply_modify() and node_created are
returned by path_write().
At least with OCaml 4.06.1 this leads to the value of node_created being
returned *before* it is updated by apply_modify(). This in turn leads
to the quota for a domain not being updated in Store.write(). Hence, a
guest can create an unlimited number of entries in xenstore.
The fix is to make evaluation order explicit.
This is XSA-272.
Signed-off-by: Christian Lindig <christian.lindig@citrix.com> Reviewed-by: Rob Hoes <rob.hoes@citrix.com>
(cherry picked from commit 73392c7fd14c59f8c96e0b2eeeb329e4ae9086b6)
Andrew Cooper [Mon, 18 Jun 2018 08:12:39 +0000 (16:12 +0800)]
x86/vtx: Fix the checking for unknown/invalid MSR_DEBUGCTL bits
The VPMU_MODE_OFF early-exit in vpmu_do_wrmsr() introduced by c/s 11fe998e56 bypasses all reserved bit checking in the general case. As a
result, a guest can enable BTS when it shouldn't be permitted to, and
lock up the entire host.
With vPMU active (not a security supported configuration, but useful for
debugging), the reserved bit checking in broken, caused by the original
BTS changeset 1a8aa75ed.
From a correctness standpoint, it is not possible to have two different
pieces of code responsible for different parts of value checking, if
there isn't an accumulation of bits which have been checked. A
practical upshot of this is that a guest can set any value it
wishes (usually resulting in a vmentry failure for bad guest state).
Therefore, fix this by implementing all the reserved bit checking in the
main MSR_DEBUGCTL block, and removing all handling of DEBUGCTL from the
vPMU MSR logic.
This is XSA-269.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 2a8a8e99feb950504559196521bc9fd63ed3a962)
Andrew Cooper [Tue, 14 Aug 2018 10:20:53 +0000 (11:20 +0100)]
common/gnttab: Introduce command line feature controls
This patch was originally released as part of XSA-226. It retains the same
command line syntax (as various downstreams are mitigating XSA-226 using this
mechanism) but the defaults have been updated due to the revised XSA-226
patched, after which transitive grants are believed to functioning
properly.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit dc96c65ed6d7ffd4c95487373df708d97443cf77)
Jan Beulich [Thu, 19 Jul 2018 09:54:45 +0000 (11:54 +0200)]
VMX: fix vmx_{find,del}_msr() build
Older gcc at -O2 (and perhaps higher) does not recognize that apparently
uninitialized variables aren't really uninitialized. Pull out the
assignments used by two of the three case blocks and make them
initializers of the variables, as I think I had suggested during review.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
(cherry picked from commit 97cb0516a322ecdf0032fa9d8aa1525c03d7772f)
Andrew Cooper [Mon, 7 May 2018 10:57:00 +0000 (11:57 +0100)]
x86/vmx: Support load-only guest MSR list entries
Currently, the VMX_MSR_GUEST type maintains completely symmetric guest load
and save lists, by pointing VM_EXIT_MSR_STORE_ADDR and VM_ENTRY_MSR_LOAD_ADDR
at the same page, and setting VM_EXIT_MSR_STORE_COUNT and
VM_ENTRY_MSR_LOAD_COUNT to the same value.
However, for MSRs which we won't let the guest have direct access to, having
hardware save the current value on VMExit is unnecessary overhead.
To avoid this overhead, we must make the load and save lists asymmetric. By
making the entry load count greater than the exit store count, we can maintain
two adjacent lists of MSRs, the first of which is saved and restored, and the
second of which is only restored on VMEntry.
For simplicity:
* Both adjacent lists are still sorted by MSR index.
* It undefined behaviour to insert the same MSR into both lists.
* The total size of both lists is still limited at 256 entries (one 4k page).
Split the current msr_count field into msr_{load,save}_count, and introduce a
new VMX_MSR_GUEST_LOADONLY type, and update vmx_{add,find}_msr() to calculate
which sublist to search, based on type. VMX_MSR_HOST has no logical sublist,
whereas VMX_MSR_GUEST has a sublist between 0 and the save count, while
VMX_MSR_GUEST_LOADONLY has a sublist between the save count and the load
count.
One subtle point is that inserting an MSR into the load-save list involves
moving the entire load-only list, and updating both counts.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
(cherry picked from commit 1ac46b55632626aeb935726e1b0a71605ef6763a)
Andrew Cooper [Mon, 7 May 2018 10:57:00 +0000 (11:57 +0100)]
x86/vmx: Support remote access to the MSR lists
At the moment, all modifications of the MSR lists are in current context.
However, future changes may need to put MSR_EFER into the lists from domctl
hypercall context.
Plumb a struct vcpu parameter down through the infrastructure, and use
vmx_vmcs_{enter,exit}() for safe access to the VMCS in vmx_add_msr(). Use
assertions to ensure that access is either in current context, or while the
vcpu is paused.
Note these expectations beside the fields in arch_vmx_struct, and reorder the
fields to avoid unnecessary padding.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 80599f0b770199116aa753bfdfac9bfe2e8ea86a)
Andrew Cooper [Mon, 7 May 2018 10:57:00 +0000 (11:57 +0100)]
x86/vmx: Factor locate_msr_entry() out of vmx_find_msr() and vmx_add_msr()
Instead of having multiple algorithms searching the MSR lists, implement a
single one. It has the semantics required by vmx_add_msr(), to identify the
position in which an MSR should live, if it isn't already present.
There will be a marginal improvement for vmx_find_msr() by avoiding the
function pointer calls to vmx_msr_entry_key_cmp(), and a major improvement for
vmx_add_msr() by using a binary search instead of a linear search.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
(cherry picked from commit 4d94828cf11104256dccea1fa7762f00575dfaa0)
Andrew Cooper [Mon, 7 May 2018 10:57:00 +0000 (11:57 +0100)]
x86/vmx: Internal cleanup for MSR load/save infrastructure
* Use an arch_vmx_struct local variable to reduce later code volume.
* Use start/total instead of msr_area/msr_count. This is in preparation for
more finegrained handling with later changes.
* Use ent/end pointers (again for preparation), and to make the vmx_add_msr()
logic easier to follow.
* Make the memory allocation block of vmx_add_msr() unlikely, and calculate
virt_to_maddr() just once.
No practical change to functionality.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
(cherry picked from commit 94fda356fcdcc847662a4c9f6cc63511f25c1247)
Andrew Cooper [Mon, 7 May 2018 10:57:00 +0000 (11:57 +0100)]
x86/vmx: API improvements for MSR load/save infrastructure
Collect together related infrastructure in vmcs.h, rather than having it
spread out. Turn vmx_{read,write}_guest_msr() into static inlines, as they
are simple enough.
Replace 'int type' with 'enum vmx_msr_list_type', and use switch statements
internally. Later changes are going to introduce a new type.
Rename the type identifiers for consistency with the other VMX_MSR_*
constants.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
(cherry picked from commit f54b63e8617ada823be43d60467a43c8224b7909)
Andrew Cooper [Mon, 28 May 2018 14:02:34 +0000 (15:02 +0100)]
x86/vmx: Defer vmx_vmcs_exit() as long as possible in construct_vmcs()
paging_update_paging_modes() and vmx_vlapic_msr_changed() both operate on the
VMCS being constructed. Avoid dropping and re-acquiring the reference
multiple times.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Acked-by: Kevin Tian <kevin.tian@intel.com>
(cherry picked from commit f30e3cf34042846e391e3f8361fc6a76d181a7ee)
Andrew Cooper [Thu, 9 Aug 2018 16:22:17 +0000 (17:22 +0100)]
x86/spec-ctrl: Yet more fixes for xpti= parsing
As it currently stands, 'xpti=dom0' is indistinguishable from the default
value, which means it will be overridden by ARCH_CAPABILITIES_RDCL_NO on fixed
hardware.
Switch opt_xpti to use -1 as a default like all our other related options, and
clobber it as soon as we have a string to parse.
In addition, 'xpti' alone should be interpreted in its positive boolean form,
rather than resulting in a parse error.
(XEN) parameter "xpti" has invalid value "", rc=-22!
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Juergen Gross <jgross@suse.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 2a3b34ec47817048ab59586855cf0709fc77487e)
Andrew Cooper [Mon, 30 Jul 2018 12:13:09 +0000 (14:13 +0200)]
x86/spec-ctrl: Fix the parsing of xpti= on fixed Intel hardware
The calls to xpti_init_default() in parse_xpti() are buggy. The CPUID data
hasn't been fetched that early, and boot_cpu_has(X86_FEATURE_ARCH_CAPS) will
always evaluate false.
As a result, the default case won't disable XPTI on Intel hardware which
advertises ARCH_CAPABILITIES_RDCL_NO.
Simplify parse_xpti() to solely the setting of opt_xpti according to the
passed string, and have init_speculation_mitigations() call
xpti_init_default() if appropiate. Drop the force parameter, and pass caps
instead, to avoid redundant re-reading of MSR_ARCH_CAPS.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Juergen Gross <jgross@suse.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> Acked-by: Jan Beulich <jbeulich@suse.com>
master commit: be5e2ff6f54e0245331ed360b8786760f82fd673
master date: 2018-07-24 11:25:54 +0100
Jan Beulich [Mon, 30 Jul 2018 12:12:39 +0000 (14:12 +0200)]
x86: command line option to avoid use of secondary hyper-threads
Shared resources (L1 cache and TLB in particular) present a risk of
information leak via side channels. Provide a means to avoid use of
hyperthreads in such cases.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: d8f974f1a646c0200b97ebcabb808324b288fadb
master date: 2018-07-19 13:43:33 +0100
Jan Beulich [Mon, 30 Jul 2018 12:11:42 +0000 (14:11 +0200)]
x86: possibly bring up all CPUs even if not all are supposed to be used
Reportedly Intel CPUs which can't broadcast #MC to all targeted
cores/threads because some have CR4.MCE clear will shut down. Therefore
we want to keep CR4.MCE enabled when offlining a CPU, and we need to
bring up all CPUs in order to be able to set CR4.MCE in the first place.
The use of clear_in_cr4() in cpu_mcheck_disable() was ill advised
anyway, and to avoid future similar mistakes I'm removing clear_in_cr4()
altogether right here.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com>
master commit: 8797d20a6ec2dd75195585a107ce345c51c0a59a
master date: 2018-07-19 13:43:33 +0100
Jan Beulich [Mon, 30 Jul 2018 12:10:58 +0000 (14:10 +0200)]
x86: distinguish CPU offlining from CPU removal
In order to be able to service #MC on offlined CPUs, the GDT, IDT,
stack, and per-CPU data (which includes the TSS) need to be kept
allocated. They should only be freed upon CPU removal (which we
currently don't support, so some code is becoming effectively dead for
the moment).
Note that for now park_offline_cpus doesn't get set to true anywhere -
this is going to be the subject of a subsequent patch.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 2e6c8f182c9c50129b1c7a620242861e6ad6a9fb
master date: 2018-07-19 13:43:33 +0100
Jan Beulich [Mon, 30 Jul 2018 12:10:30 +0000 (14:10 +0200)]
x86/AMD: distinguish compute units from hyper-threads
Fam17 replaces CUs by HTs, which we should reflect accordingly, even if
the difference is not very big. The most relevant change (requiring some
code restructuring) is that the topoext feature no longer means there is
a valid CU ID.
Take the opportunity and convert wrongly plain int variables in
set_cpu_sibling_map() to unsigned int.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Brian Woods <brian.woods@amd.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 9429b07a0af7f92a5f25e4068e11db881e157495
master date: 2018-07-19 09:42:42 +0200
Jan Beulich [Mon, 30 Jul 2018 12:09:57 +0000 (14:09 +0200)]
cpupools: fix state when downing a CPU failed
While I've run into the issue with further patches in place which no
longer guarantee the per-CPU area to start out as all zeros, the
CPU_DOWN_FAILED processing looks to have the same issue: By not zapping
the per-CPU cpupool pointer, cpupool_cpu_add()'s (indirect) invocation
of schedule_cpu_switch() will trigger the "c != old_pool" assertion
there.
Clearing the field during CPU_DOWN_PREPARE is too early (afaict this
should not happen before cpu_disable_scheduler()). Clearing it in
CPU_DEAD and CPU_DOWN_FAILED would be an option, but would take the same
piece of code twice. Since the field's value shouldn't matter while the
CPU is offline, simply clear it (implicitly) for CPU_ONLINE and
CPU_DOWN_FAILED, but only for other than the suspend/resume case (which
gets specially handled in cpupool_cpu_remove()).
By adjusting the conditional in cpupool_cpu_add() CPU_DOWN_FAILED
handling in the suspend case should now also be handled better.
Jan Beulich [Mon, 30 Jul 2018 12:09:19 +0000 (14:09 +0200)]
allow cpu_down() to be called earlier
The function's use of the stop-machine logic has so far prevented its
use ahead of the processing of the "ordinary" initcalls. Since at this
early time we're in a controlled environment anyway, there's no need for
such a heavy tool. Additionally this ought to have less of a performance
impact especially on large systems, compared to the alternative of
making stop-machine functionality available earlier.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 5894c0a2da66243a89088d309c7e1ea212ab28d6
master date: 2018-07-16 15:15:12 +0200
Jan Beulich [Mon, 30 Jul 2018 12:08:42 +0000 (14:08 +0200)]
x86/spec-ctrl: command line handling adjustments
For one, "no-xen" should not imply "no-eager-fpu", as "eager FPU" mode
is to guard guests, not Xen itself, which is also expressed so by
print_details().
And then opt_ssbd, despite being off by default, should also be cleared
by the "no" and "no-xen" sub-options.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: ac3f9a72141a48d40fabfff561d5a7dc0e1b810d
master date: 2018-07-10 12:22:31 +0200
Jan Beulich [Mon, 30 Jul 2018 12:06:44 +0000 (14:06 +0200)]
cmdline: fix parse_boolean() for NULL incoming end pointer
Use the calculated lengths instead of pointers, as 'e' being NULL will
otherwise cause undue parsing failures.
Reported-by: Karl Johnson <karljohnson.it@gmail.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
Jan Beulich [Thu, 28 Jun 2018 09:31:48 +0000 (11:31 +0200)]
x86/HVM: don't cause #NM to be raised in Xen
The changes for XSA-267 did not touch management of CR0.TS for HVM
guests. In fully eager mode this bit should never be set when
respective vCPU-s are active, or else hvmemul_get_fpu() might leave it
wrongly set, leading to #NM in hypervisor context.
{svm,vmx}_enter() and {svm,vmx}_fpu_dirty_intercept() become unreachable
this way. Explicit {svm,vmx}_fpu_leave() invocations need to be guarded
now.
With no CR0.TS management necessary in fully eager mode, there's also no
need anymore to intercept #NM.
Reported-by: Charles Arnold <carnold@suse.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 488efc29e4e996bb3805c982200f65061390cdce
master date: 2018-06-28 09:07:06 +0200
Ian Jackson [Thu, 28 Jun 2018 09:31:34 +0000 (11:31 +0200)]
libxl: restore passing "readonly=" to qemu for SCSI disks
A read-only check was introduced for XSA-142, commit ef6cb76026 ("libxl:
relax readonly check introduced by XSA-142 fix") added the passing of
the extra setting, but commit dab0539568 ("Introduce COLO mode and
refactor relevant function") dropped the passing of the setting again,
quite likely due to improper re-basing.
Restore the readonly= parameter to SCSI disks. For IDE disks this is
supposed to be rejected; add an assert. And there is a bare ad-hoc
disk drive string in libxl__build_device_model_args_new, which we also
update.
This is XSA-266.
Reported-by: Andrew Reimers <andrew.reimers@orionvm.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
master commit: dd64d3c41a2d15139c3a35d22d4cb6b78f4c5c59
master date: 2018-06-28 09:05:06 +0200
Ian Jackson [Thu, 28 Jun 2018 09:31:21 +0000 (11:31 +0200)]
libxl: qemu_disk_scsi_drive_string: Break out common parts of disk config
The generated configurations are identical apart from, in some cases,
reordering of the id=%s element. So, overall, no functional change.
This is part of XSA-266.
Reported-by: Andrew Reimers <andrew.reimers@orionvm.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
master commit: 724e5aa31b58d1e430ad36b484cf0ec021497399
master date: 2018-06-28 09:04:55 +0200
Andrew Cooper [Thu, 28 Jun 2018 09:30:56 +0000 (11:30 +0200)]
x86: Refine checks in #DB handler for faulting conditions
One of the fix for XSA-260 (c/s 75d6828bc2 "x86/traps: Fix handling of #DB
exceptions in hypervisor context") added some safety checks to help avoid
livelocks of #DB faults.
While a General Detect #DB exception does have fault semantics, hardware
clears %dr7.gd on entry to the handler, meaning that it is actually safe to
return to. Furthermore, %dr6.gd is guest controlled and sticky (never cleared
by hardware). A malicious PV guest can therefore trigger the fatal_trap() and
crash Xen.
Instruction breakpoints are more tricky. The breakpoint match bits in %dr6
are not sticky, but the Intel manual warns that they may be set for
non-enabled breakpoints, so add a breakpoint enabled check.
Beyond that, because of the restriction on the linear addresses PV guests can
set, and the fault (rather than trap) nature of instruction breakpoints
(i.e. can't be deferred by a MovSS shadow), there should be no way to
encounter an instruction breakpoint in Xen context. However, for extra
robustness, deal with this situation by clearing the breakpoint configuration,
rather than crashing.
This is XSA-265
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 17bf51297220dcd74da29de99320b6b1c72d1fa5
master date: 2018-06-28 09:04:20 +0200
Jan Beulich [Thu, 28 Jun 2018 09:29:48 +0000 (11:29 +0200)]
x86/EFI: further correct FPU state handling around runtime calls
We must not leave a vCPU with CR0.TS clear when it is not in fully eager
mode and has not touched non-lazy state. Instead of adding a 3rd
invocation of stts() to vcpu_restore_fpu_eager(), consolidate all of
them into a single one done at the end of the function.
Rename the function at the same time to better reflect its purpose, as
the patches touches all of its occurences anyway.
The new function parameter is not really well named, but
"need_stts_if_not_fully_eager" seemed excessive to me.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Paul Durrant <paul.durrant@citrix.com>
master commit: 23839a0fa0bbe78c174cd2bb49083e153f0f99df
master date: 2018-06-26 15:23:08 +0200
Jan Beulich [Thu, 28 Jun 2018 09:29:21 +0000 (11:29 +0200)]
x86/EFI: fix FPU state handling around runtime calls
There are two issues. First, the nonlazy xstates were never restored
after returning from the runtime call.
Secondly, with the fully_eager_fpu mitigation for XSA-267 / LazyFPU, the
unilateral stts() is no longer correct, and hits an assertion later when
a lazy state restore tries to occur for a fully eager vcpu.
Fix both of these issues by calling vcpu_restore_fpu_eager(). As EFI
runtime services can be used in the idle context, the idle assertion
needs to move until after the fully_eager_fpu check.
Introduce a "curr" local variable and replace other uses of "current"
at the same time.
Reported-by: Andrew Cooper <andrew.cooper3@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Tested-by: Juergen Gross <jgross@suse.com>
master commit: 437211cb696515ee5bd5dae0ab72866c9f382a33
master date: 2018-06-21 11:35:46 +0200
Jan Beulich [Thu, 28 Jun 2018 09:28:39 +0000 (11:28 +0200)]
x86: correct default_xen_spec_ctrl calculation
Even with opt_msr_sc_{pv,hvm} both false we should set up the variable
as usual, to ensure proper one-time setup during boot and CPU bringup.
This then also brings the code in line with the comment immediately
ahead of the printk() being modified saying "irrespective of guests".
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: d6239f64713df819278bf048446d3187c6ac4734
master date: 2018-05-29 12:38:52 +0200
Andrew Cooper [Thu, 7 Jun 2018 16:00:37 +0000 (17:00 +0100)]
x86/spec-ctrl: Mitigations for LazyFPU
Intel Core processors since at least Nehalem speculate past #NM, which is the
mechanism by which lazy FPU context switching is implemented.
On affected processors, Xen must use fully eager FPU context switching to
prevent guests from being able to read FPU state (SSE/AVX/etc) from previously
scheduled vcpus.
This is part of XSA-267 / CVE-2018-3665
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 243435bf67e8159495194f623b9e4d8c90140384)
Andrew Cooper [Thu, 7 Jun 2018 16:00:37 +0000 (17:00 +0100)]
x86: Support fully eager FPU context switching
This is controlled on a per-vcpu bases for flexibility.
This is part of XSA-267 / CVE-2018-3665
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
(cherry picked from commit 146dfe9277c2b4a8c399b229e00d819065e3167b)
Jan Beulich [Wed, 30 May 2018 06:37:19 +0000 (08:37 +0200)]
x86: re-enable XPTI/PCID as needed in switch_native()
Additionally avoid accessing d->arch.pv_domain for PVH domains (running
in a HVM container).
Reported-by: Sergey Dyasli <sergey.dyasli@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Avoid flushing the complete TLB when switching %cr3 for mitigation of
Meltdown by using the PCID feature if available.
We are using 4 PCID values for a 64 bit pv domain subject to XPTI and
2 values for the non-XPTI case:
- guest active and in kernel mode
- guest active and in user mode
- hypervisor active and guest in user mode (XPTI only)
- hypervisor active and guest in kernel mode (XPTI only)
We use PCID only if PCID _and_ INVPCID are supported. With PCID in use
we disable global pages in cr4. A command line parameter controls in
which cases PCID is being used.
As the non-XPTI case has shown not to perform better with PCID at least
on some machines the default is to use PCID only for domains subject to
XPTI.
With PCID enabled we always disable global pages. This avoids having to
either flush the complete TLB or do a cycle through all PCID values
when invalidating a single global page.
Signed-off-by: Juergen Gross <jgross@suse.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
xen/x86: use flag byte for decision whether xen_cr3 is valid
Today cpu_info->xen_cr3 is either 0 to indicate %cr3 doesn't need to
be switched on entry to Xen, or negative for keeping the value while
indicating not to restore %cr3, or positive in case %cr3 is to be
restored.
Switch to use a flag byte instead of a negative xen_cr3 value in order
to allow %cr3 values with the high bit set in case we want to keep TLB
entries when using the PCID feature.
This reduces the number of branches in interrupt handling and results
in better performance (e.g. parallel make of the Xen hypervisor on my
system was using about 3% less system time).
Signed-off-by: Juergen Gross <jgross@suse.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
xen/x86: disable global pages for domains with XPTI active
Instead of flushing the TLB from global pages when switching address
spaces with XPTI being active just disable global pages via %cr4
completely when a domain subject to XPTI is active. This avoids the
need for extra TLB flushes as loading %cr3 will remove all TLB
entries.
In order to avoid states with cr3/cr4 having inconsistent values
(e.g. global pages being activated while cr3 already specifies a XPTI
address space) move loading of the new cr4 value to write_ptbase()
(actually to switch_cr3_cr4() called by write_ptbase()).
This requires to use switch_cr3_cr4() instead of write_ptbase() when
building dom0 in order to avoid setting cr4 with cr4.smap set.
Signed-off-by: Juergen Gross <jgross@suse.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Instead of switching XPTI globally on or off add a per-domain flag for
that purpose. This allows to modify the xpti boot parameter to support
running dom0 without Meltdown mitigations. Using "xpti=no-dom0" as boot
parameter will achieve that.
Move the xpti boot parameter handling to xen/arch/x86/pv/domain.c as
it is pv-domain specific.
Signed-off-by: Juergen Gross <jgross@suse.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Juergen Gross [Tue, 29 May 2018 08:42:44 +0000 (10:42 +0200)]
x86/xpti: avoid copying L4 page table contents when possible
For mitigation of Meltdown the current L4 page table is copied to the
cpu local root page table each time a 64 bit pv guest is entered.
Copying can be avoided in cases where the guest L4 page table hasn't
been modified while running the hypervisor, e.g. when handling
interrupts or any hypercall not modifying the L4 page table or %cr3.
So add a per-cpu flag indicating whether the copying should be
performed and set that flag only when loading a new %cr3 or modifying
the L4 page table. This includes synchronization of the cpu local
root page table with other cpus, so add a special synchronization flag
for that case.
A simple performance check (compiling the hypervisor via "make -j 4")
in dom0 with 4 vcpus shows a significant improvement:
- real time drops from 112 seconds to 103 seconds
- system time drops from 142 seconds to 131 seconds
Signed-off-by: Juergen Gross <jgross@suse.com> Reviewed-by: Jan Beulich <jbeulich@suse.com>
Jan Beulich [Tue, 23 Jan 2018 09:43:39 +0000 (10:43 +0100)]
x86: move invocations of hvm_flush_guest_tlbs()
Their need is not tied to the actual flushing of TLBs, but the ticking
of the TLB clock. Make this more obvious by folding the two invocations
into a single one in pre_flush().
Also defer the latching of CR4 in write_cr3() until after pre_flush()
(and hence implicitly until after IRQs are off), making operation
sequence the same in both cases (eliminating the theoretical risk of
pre_flush() altering CR4). This then also improves register allocation,
as the compiler doesn't need to use a callee-saved register for "cr4"
anymore.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Jan Beulich [Tue, 29 May 2018 08:40:37 +0000 (10:40 +0200)]
x86/XPTI: fix S3 resume (and CPU offlining in general)
We should index an L1 table with an L1 index.
Reported-by: Simon Gaiser <simon@invisiblethingslab.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 6b9562dac1746014ab376bd2cf8ba400acf34c6d
master date: 2018-05-28 11:20:26 +0200
projects / xen.git / log