In case cpu_schedule_up() is failing, it needs to undo all externally
visible changes it has done before.
Reason is that cpu_schedule_callback() won't be called with the
CPU_UP_CANCELED notifier in case cpu_schedule_up() did fail.
Fixes: 207589dbac ("xen/sched: move per cpu scheduler private data into struct sched_resource")
Reported-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 44a7d4f0a5
master date: 2024-07-31 14:50:18 +0200
The current code in ALT_CALL_ARG() won't successfully workaround the clang
code-generation issue if the arg parameter has a size that's not a power of 2.
While there are no such sized parameters at the moment, improve the workaround
to also be effective when such sizes are used.
Instead of using a union with a long use an unsigned long that's first
initialized to 0 and afterwards set to the argument value.
Reported-by: Alejandro Vallejo <alejandro.vallejo@cloud.com>
Suggested-by: Alejandro Vallejo <alejandro.vallejo@cloud.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 561cba38ff
master date: 2024-07-31 12:41:22 +0200
One of the error paths in the PV dom0 builder section that runs on the guest
page-tables wasn't restoring the Xen value of %cr3, neither removing the
mapcache override.
Fixes: 079ff2d32c ('libelf-loader: introduce elf_load_image')
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 1fc3f77113
master date: 2024-07-31 12:41:02 +0200
The XSM checks for XEN_DOMCTL_createdomain are problematic. There's a split
between xsm_domctl() called early, and flask_domain_create() called quite late
during domain construction.
All XSM implementations except Flask have a simple IS_PRIV check in
xsm_domctl(), and operate as expected when an unprivileged domain tries to
make a hypercall.
Flask however foregoes any action in xsm_domctl() and defers everything,
including the simple "is the caller permitted to create a domain" check, to
flask_domain_create().
As a consequence, when XSM Flask is active, and irrespective of the policy
loaded, all domains irrespective of privilege can:
* Mutate the global 'rover' variable, used to track the next free domid.
Therefore, all domains can cause a domid wraparound, and combined with a
voluntary reboot, choose their own domid.
* Cause a reasonable amount of a domain to be constructed before ultimately
failing for permission reasons, including the use of settings outside of
supported limits.
In order to remediate this, pass the ssidref into xsm_domctl() and at least
check that the calling domain privileged enough to create domains.
Take the opportunity to also fix the sign of the cmd parameter to be unsigned.
This issue has not been assigned an XSA, because Flask is experimental and not
security supported.
Reported-by: Ross Lagerwall <ross.lagerwall@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Daniel P. Smith <dpsmith@apertussolutions.com>
master commit: ee32b9b29a
master date: 2024-07-30 17:42:17 +0100
The decompression code parses a huffman tree and counts the number of
symbols for a given bit length. In rare cases, there may be >= 256
symbols with a given bit length, causing the unsigned char to overflow.
This causes a decompression failure later when the code tries and fails to
find the bit length for a given symbol.
Since the maximum number of symbols is 258, use unsigned short instead.
Fixes: ab77e81f65 ("x86/dom0: support bzip2 and lzma compressed bzImage payloads")
Signed-off-by: Ross Lagerwall <ross.lagerwall@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
master commit: 303d3ff85c
master date: 2024-07-30 11:55:56 +0200
Yet another clang code generation issue when using altcalls.
The issue this time is with using loop constructs around alternative_{,v}call
instances using parameter types smaller than the register size.
Given the following example code:
static void bar(bool b)
{
unsigned int i;
for ( i = 0; i < 10; i++ )
{
int ret_;
register union {
bool e;
unsigned long r;
} di asm("rdi") = { .e = b };
register unsigned long si asm("rsi");
register unsigned long dx asm("rdx");
register unsigned long cx asm("rcx");
register unsigned long r8 asm("r8");
register unsigned long r9 asm("r9");
register unsigned long r10 asm("r10");
register unsigned long r11 asm("r11");
asm volatile ( "call %c[addr]"
: "+r" (di), "=r" (si), "=r" (dx),
"=r" (cx), "=r" (r8), "=r" (r9),
"=r" (r10), "=r" (r11), "=a" (ret_)
: [addr] "i" (&(func)), "g" (func)
: "memory" );
}
}
See: https://godbolt.org/z/qvxMGd84q
Clang will generate machine code that only resets the low 8 bits of %rdi
between loop calls, leaving the rest of the register possibly containing
garbage from the use of %rdi inside the called function. Note also that clang
doesn't truncate the input parameters at the callee, thus breaking the psABI.
Fix this by turning the `e` element in the anonymous union into an array that
consumes the same space as an unsigned long, as this forces clang to reset the
whole %rdi register instead of just the low 8 bits.
Fixes: 2ce562b2a4 ('x86/altcall: use a union as register type for function parameters on clang')
Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: d51b2f5ea1
master date: 2024-07-23 13:59:30 +0200
set_cloexec() had a bit too much copy&pate from setnonblock(), and
insufficient testing on ancient versions of Linux...
As written (emulating ancient linux by undef'ing O_CLOEXEC), strace shows:
open("/dev/xen/xenbus", O_RDWR) = 3
fcntl(3, F_GETFL) = 0x8002 (flags O_RDWR|O_LARGEFILE)
fcntl(3, 0x8003 /* F_??? */, 0x7ffe4a771d90) = -1 EINVAL (Invalid argument)
close(3) = 0
which is obviously nonsense.
Switch F_GETFL -> F_GETFD, and fix the second invocation to use F_SETFD. With
this, strace is rather happer:
open("/dev/xen/xenbus", O_RDWR) = 3
fcntl(3, F_GETFD) = 0
fcntl(3, F_SETFD, FD_CLOEXEC) = 0
Fixes: bf7c146470 ("tools/libxs: Fix CLOEXEC handling in get_dev()")
Reported-by: Ross Lagerwall <ross.lagerwall@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Ross Lagerwall <ross.lagerwall@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
master commit: 37810b52d0
master date: 2024-07-09 15:32:18 +0100
Fix bug introduced by 0762e2502f ("x86/physdev: factor out the code to allocate and
map a pirq"). After that re-factoring, when pirq<0 and current_pirq>0, it means
caller want to allocate a free pirq for irq but irq already has a mapped pirq, then
it returns the negative pirq, so it fails. However, the logic before that
re-factoring is different, it should return the current_pirq that irq was already
mapped to and make the call success.
Fixes: 0762e2502f ("x86/physdev: factor out the code to allocate and map a pirq")
Signed-off-by: Jiqian Chen <Jiqian.Chen@amd.com>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Jiqian Chen <Jiqian.Chen@amd.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 0d2b87b5ad
master date: 2024-07-08 14:46:12 +0100
Forever since its introduction the main loop in the function dealing
with multi-vector MSI had error exit points ("break") with different
properties: In one case no IRQ descriptor lock is being held.
Nevertheless the subsequent error cleanup path assumed such a lock would
uniformly need releasing. Identify the case by setting "desc" to NULL,
thus allowing the unlock to be skipped as necessary.
This is CVE-2024-31143 / XSA-458.
Coverity ID: 1605298
Fixes: d1b6d0a024 ("x86: enable multi-vector MSI")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 57338346f2
master date: 2024-07-16 14:09:14 +0200
When backporting daa90dfea9 ("pirq_cleanup_check() leaks") I neglected
to pay attention to it depending on 13a7b0f9f7 ("restrict concept of
pIRQ to x86"). That one doesn't want backporting imo, so use / adjust
custom #ifdef-ary to address the immediate issue of pirq_cleanup_check()
not being available on Arm.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
While doing so is intentional when invoking the actual callback, to
mimic a hard-coded SYCALL_MASK / FMASK MSR, the same should not be done
when no handler is available and hence #UD is raised.
Fixes: ca6fcf4321 ("x86/pv: Inject #UD for missing SYSCALL callbacks")
Reported-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: d2fe9ab304
master date: 2024-07-02 12:01:21 +0200
PHYSDEVOP_pirq_eoi_gmfn_v<N> accepting just a single GFN implies that no
more than 32k pIRQ-s can be used by a domain on x86. Document this upper
bound.
To also enforce the limit, (ab)use both arch_hwdom_irqs() (changing its
parameter type) and setup_system_domains(). This is primarily to avoid
exposing the two static variables or introducing yet further arch hooks.
While touching arch_hwdom_irqs() also mark it hwdom-init.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
amend 'cmdline: document and enforce "extra_guest_irqs" upper bounds'
Address late review comments for what is now commit 17f6d398f7:
- bound max_irqs right away against nr_irqs
- introduce a #define for a constant used twice
Requested-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 17f6d398f7
master date: 2024-07-02 12:00:27 +0200
master commit: 1f56accba3
master date: 2024-07-03 14:03:27 +0200
xs_fileno() opens a pipe on first use to communicate between the watch thread
and the main thread. Nothing ever sets CLOEXEC on the file descriptors.
Check for the availability of the pipe2() function with configure. Despite
starting life as Linux-only, FreeBSD and NetBSD have gained it.
When pipe2() isn't available, try our best with pipe() and set_cloexec().
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Anthony PERARD <anthony.perard@vates.tech>
master commit: a2ff677852
master date: 2024-07-02 10:52:59 +0100
get_socket() opens a socket, then uses fcntl() to set CLOEXEC. This is racy
with exec().
Open the socket with SOCK_CLOEXEC. Use the same compatibility strategy as
O_CLOEXEC on ancient versions of Linux.
Reported-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Anthony PERARD <anthony.perard@vates.tech>
master commit: 1957dd6aff
master date: 2024-07-02 10:51:11 +0100
Move the O_CLOEXEC compatibility outside of an #ifdef USE_PTHREAD block.
Introduce set_cloexec() to wrap fcntl() setting FD_CLOEXEC. It will be reused
for other CLOEXEC fixes too.
Use set_cloexec() when O_CLOEXEC isn't available as a best-effort fallback.
Fixes: f4f2f3402b ("tools/libxs: Open /dev/xen/xenbus fds as O_CLOEXEC")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Anthony PERARD <anthony.perard@vates.tech>
master commit: bf7c146470
master date: 2024-07-02 10:51:06 +0100
xc_dom_alloc_segment() is passed a size in bytes, calculates a size in pages
from it, then fills in the new segment information with a bytes value
re-calculated from the number of pages.
This causes the module information given to the guest (MB, or PVH) to have
incorrect sizes; specifically, sizes rounded up to the next page.
This in turn is problematic for Xen. When Xen finds a gzipped module, it
peeks at the end metadata to judge the decompressed size, which is a -4
backreference from the reported end of the module.
Fill in seg->vend using the correct number of bytes.
Fixes: ea7c8a3d0e ("libxc: reorganize domain builder guest memory allocator")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Anthony PERARD <anthony.perard@vates.tech>
master commit: 4c3a618b0a
master date: 2024-07-02 10:50:18 +0100
Its original introduction had two issues: For one the "common" part of
the checks (carried out in the macro) was inverted. And then after
removal from the radix tree the structure wasn't scheduled for freeing.
(All structures still left in the radix tree would be freed upon domain
destruction, though.)
For the freeing to be safe even if it didn't use RCU (i.e. to avoid use-
after-free), re-arrange checks/operations in evtchn_close(), such that
the pointer wouldn't be used anymore after calling pirq_cleanup_check()
(noting that unmap_domain_pirq_emuirq() itself calls the function in the
success case).
Fixes: c24536b636 ("replace d->nr_pirqs sized arrays with radix tree")
Fixes: 79858fee30 ("xen: fix hvm_domain_use_pirq's behavior")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: daa90dfea9
master date: 2024-07-02 08:35:56 +0200
There exists bitshifts in the IOAPIC code where signed integers are
shifted to the left by up to 31 bits, which is undefined behaviour.
This patch fixes this by changing the integers from signed to unsigned.
Signed-off-by: Matthew Barnes <matthew.barnes@cloud.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: c5746b021e
master date: 2024-06-21 15:09:26 +0100
There are no more users, and we want to disuade people from introducing new
users just for sd_notify() and friends. Drop the dependency.
We still want the overall --with{,out}-systemd to gate the generation of the
service/unit/mount/etc files.
Rerun autogen.sh, and mark the dependency as removed in the build containers.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Christian Lindig <christian.lindig@cloud.com>
tools: (Actually) drop libsystemd as a dependency
When reinstating some of systemd.m4 between v1 and v2, I reintroduced a little
too much. While {c,o}xenstored are indeed no longer linked against
libsystemd, ./configure still looks for it.
Drop this too.
Fixes: ae26101f6b ("tools: Drop libsystemd as a dependency")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: ae26101f6b
master date: 2024-05-23 15:04:40 +0100
master commit: 6ef4fa1e7f
master date: 2024-05-30 12:15:25 +0100
... in order to avoid linking against the whole of libsystemd.
Only minimal changes to the upstream copy, to function as a drop-in
replacement for sd_notify() and as a header-only library.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Christian Lindig <christian.lindig@cloud.com>
master commit: 78510f3a15
master date: 2024-05-23 15:04:40 +0100
We are about to import code licensed under MIT-0. It's compatible for us to
use, so identify it as a permitted license.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Acked-by: Christian Lindig <christian.lindig@cloud.com>
master commit: 219cdff3fb
master date: 2024-05-23 15:04:40 +0100
In case a test is failing in test-xenstore, let the tool exit with an
exit status other than 0.
Fix a typo in an error message.
Reported-by: Andrew Cooper <andrew.cooper3@citrix.com>
Fixes: 3afc5e4a5b ("tools/tests: add xenstore testing framework")
Signed-off-by: Juergen Gross <jgross@suse.com>
master commit: 2d4ba20559
master date: 2024-05-02 18:15:46 +0100
Today test-xenstore will write nodes with 3000 bytes node data. This
size is exceeding the default quota for the allowed node size. While
working in dom0 with C-xenstored, OCAML-xenstored does not like that.
Use a size of 2000 instead, which is lower than the allowed default
node size of 2048.
Fixes: 3afc5e4a5b ("tools/tests: add xenstore testing framework")
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 642005e310
master date: 2024-05-02 18:15:31 +0100
On the BSP, shadow stacks are enabled only relatively late in the
booting process. They in particular aren't active yet when initcalls are
run. Keep the testing there, but invoke that testing a 2nd time when
shadow stacks are active, to make sure we won't regress that case after
addressing XSA-451.
While touching this code, switch the guard from NDEBUG to CONFIG_DEBUG,
such that IS_ENABLED() can validly be used at the new call site.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: cfe3ad6712
master date: 2024-02-27 13:49:52 +0100
fixup_irqs() is used to evacuate interrupts from to be offlined CPUs. Given
the CPU is to become offline, the normal migration logic used by Xen where the
vector in the previous target(s) is left configured until the interrupt is
received on the new destination is not suitable.
Instead attempt to do as much as possible in order to prevent loosing
interrupts. If fixup_irqs() is called from the CPU to be offlined (as is
currently the case for CPU hot unplug) attempt to forward pending vectors when
interrupts that target the current CPU are migrated to a different destination.
Additionally, for interrupts that have already been moved from the current CPU
prior to the call to fixup_irqs() but that haven't been delivered to the new
destination (iow: interrupts with move_in_progress set and the current CPU set
in ->arch.old_cpu_mask) also check whether the previous vector is pending and
forward it to the new destination.
This allows us to remove the window with interrupts enabled at the bottom of
fixup_irqs(). Such window wasn't safe anyway: references to the CPU to become
offline are removed from interrupts masks, but the per-CPU vector_irq[] array
is not updated to reflect those changes (as the CPU is going offline anyway).
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: e2bb28d621
master date: 2024-06-20 12:09:32 +0200
[ This is a minimal backport of commit 71cacfb035 ("x86/cpuid: Fix handling
of XSAVE dynamic leaves") to fix the bugs without depending on the large
rework of XSTATE handling in Xen 4.19 ]
First, if XSAVE is available in hardware but not visible to the guest, the
dynamic leaves shouldn't be filled in.
Second, the comment concerning XSS state is wrong. VT-x doesn't manage
host/guest state automatically, but there is provision for "host only" bits to
be set, so the implications are still accurate.
In Xen 4.18, no XSS states are supported, so it's safe to keep deferring to
real hardware.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 71cacfb035
master date: 2024-06-19 13:00:06 +0100
The clobbering of this_cpu(xcr0) and this_cpu(xss) to architecturally invalid
values is to force the subsequent set_xcr0() and set_msr_xss() to reload the
hardware register.
While XCR0 is reloaded in xstate_init(), MSR_XSS isn't. This causes
get_msr_xss() to return the invalid value, and logic of the form:
old = get_msr_xss();
set_msr_xss(new);
...
set_msr_xss(old);
to try and restore said invalid value.
The architecturally invalid value must be purged from the cache, meaning the
hardware register must be written at least once. This in turn highlights that
the invalid value must only be used in the case that the hardware register is
available.
Fixes: f7f4a52392 ("x86/xstate: reset cached register values on resume")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 9e6dbbe8bf
master date: 2024-06-19 13:00:06 +0100
struct type_descriptor is arranged with a NUL terminated string following the
kind/info fields.
The only reason this doesn't trip UBSAN detection itself (on more modern
compilers at least) is because struct type_descriptor is only referenced in
suppressed regions.
Switch the declaration to be a real flexible member. No functional change.
Fixes: 00fcf4dd8e ("xen/ubsan: Import ubsan implementation from Linux 4.13")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: bd59af9970
master date: 2024-06-18 14:55:04 +0100
Currently there's logic in fixup_irqs() that attempts to prevent
_assign_irq_vector() from failing, as fixup_irqs() is required to evacuate all
interrupts from the CPUs not present in the input mask. The current logic in
fixup_irqs() is incomplete, as it doesn't deal with interrupts that have
move_cleanup_count > 0 and a non-empty ->arch.old_cpu_mask field.
Instead of attempting to fixup the interrupt descriptor in fixup_irqs() so that
_assign_irq_vector() cannot fail, introduce logic in _assign_irq_vector()
to deal with interrupts that have either move_{in_progress,cleanup_count} set
and no remaining online CPUs in ->arch.cpu_mask.
If _assign_irq_vector() is requested to move an interrupt in the state
described above, first attempt to see if ->arch.old_cpu_mask contains any valid
CPUs that could be used as fallback, and if that's the case do move the
interrupt back to the previous destination. Note this is easier because the
vector hasn't been released yet, so there's no need to allocate and setup a new
vector on the destination.
Due to the logic in fixup_irqs() that clears offline CPUs from
->arch.old_cpu_mask (and releases the old vector if the mask becomes empty) it
shouldn't be possible to get into _assign_irq_vector() with
->arch.move_{in_progress,cleanup_count} set but no online CPUs in
->arch.old_cpu_mask.
However if ->arch.move_{in_progress,cleanup_count} is set and the interrupt has
also changed affinity, it's possible the members of ->arch.old_cpu_mask are no
longer part of the affinity set, move the interrupt to a different CPU part of
the provided mask and keep the current ->arch.old_{cpu_mask,vector} for the
pending interrupt movement to be completed.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 369558924a
master date: 2024-06-18 15:15:10 +0200
Given the current logic it's possible for ->arch.old_cpu_mask to get out of
sync: if a CPU set in old_cpu_mask is offlined and then onlined
again without old_cpu_mask having been updated the data in the mask will no
longer be accurate, as when brought back online the CPU will no longer have
old_vector configured to handle the old interrupt source.
If there's an interrupt movement in progress, and the to be offlined CPU (which
is the call context) is in the old_cpu_mask, clear it and update the mask, so
it doesn't contain stale data.
Note that when the system is going down fixup_irqs() will be called by
smp_send_stop() from CPU 0 with a mask with only CPU 0 on it, effectively
asking to move all interrupts to the current caller (CPU 0) which is the only
CPU to remain online. In that case we don't care to migrate interrupts that
are in the process of being moved, as it's likely we won't be able to move all
interrupts to CPU 0 due to vector shortage anyway.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 817d1cd627
master date: 2024-06-18 15:14:49 +0200
Intel CPUs have a MSR bit to limit CPUID enumeration to leaf two. If
this bit is set by the BIOS then CPUID evaluation does not work when
data from any leaf greater than two is needed; early_cpu_init() in
particular wants to collect leaf 7 data.
Cure this by unlocking CPUID right before evaluating anything which
depends on the maximum CPUID leaf being greater than two.
Inspired by (and description cloned from) Linux commit 0c2f6d04619e
("x86/topology/intel: Unlock CPUID before evaluating anything").
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: fa4d026737
master date: 2024-06-18 15:12:44 +0200
mfn_valid() is RAM-focused; it will often return false for MMIO. Yet
access to actual MMIO space should not generally be restricted to UC
only; especially video frame buffer accesses are unduly affected by such
a restriction.
Since, as of 777c71d313 ("x86/EPT: avoid marking non-present entries
for re-configuring"), the function won't be called with INVALID_MFN or,
worse, truncated forms thereof anymore, we call fully drop that check.
Fixes: 81fd0d3ca4 ("x86/hvm: simplify 'mmio_direct' check in epte_get_entry_emt()")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 4fdd8d7556
master date: 2024-06-13 16:55:22 +0200
For non-present entries EMT, like most other fields, is meaningless to
hardware. Make the logic in ept_set_entry() setting the field (and iPAT)
conditional upon dealing with a present entry, leaving the value at 0
otherwise. This has two effects for epte_get_entry_emt() which we'll
want to leverage subsequently:
1) The call moved here now won't be issued with INVALID_MFN anymore (a
respective BUG_ON() is being added).
2) Neither of the other two calls could now be issued with a truncated
form of INVALID_MFN anymore (as long as there's no bug anywhere
marking an entry present when that was populated using INVALID_MFN).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 777c71d313
master date: 2024-06-13 16:54:17 +0200
mfn_valid() granularity is (currently) 256Mb. Therefore the start of a
1Gb page passing the test doesn't necessarily mean all parts of such a
range would also pass. Yet using the result of mfn_to_page() on an MFN
which doesn't pass mfn_valid() checking is liable to result in a crash
(the invocation of mfn_to_page() alone is presumably "just" UB in such a
case).
Fixes: ca24b2ffdb ("x86/hvm: set 'ipat' in EPT for special pages")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 5540b94e81
master date: 2024-06-13 16:53:34 +0200
The current check used in fixup_irqs() to decide whether to move around
interrupts is based on the affinity mask, but such mask can have all bits set,
and hence is unlikely to be a subset of the input mask. For example if an
interrupt has an affinity mask of all 1s, any input to fixup_irqs() that's not
an all set CPU mask would cause that interrupt to be shuffled around
unconditionally.
What fixup_irqs() care about is evacuating interrupts from CPUs not set on the
input CPU mask, and for that purpose it should check whether the interrupt is
assigned to a CPU not present in the input mask. Assume that ->arch.cpu_mask
is a subset of the ->affinity mask, and keep the current logic that resets the
->affinity mask if the interrupt has to be shuffled around.
Doing the affinity movement based on ->arch.cpu_mask requires removing the
special handling to ->arch.cpu_mask done for high priority vectors, otherwise
the adjustment done to cpu_mask makes them always skip the CPU interrupt
movement.
While there also adjust the comment as to the purpose of fixup_irqs().
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: c7564d7366
master date: 2024-06-12 14:30:40 +0200
Due to the current rwlock logic, if the CPU calling get_cpu_maps() does
so from a cpu_hotplug_{begin,done}() region the function will still
return success, because a CPU taking the rwlock in read mode after
having taken it in write mode is allowed. Such corner case makes using
get_cpu_maps() alone not enough to prevent using the shorthand in CPU
hotplug regions.
Introduce a new helper to detect whether the current caller is between a
cpu_hotplug_{begin,done}() region and use it in send_IPI_mask() to restrict
shorthand usage.
Fixes: 5500d265a2 ('x86/smp: use APIC ALLBUT destination shorthand when possible')
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: 171c52fba5
master date: 2024-06-12 14:29:31 +0200
When adjusting move_cleanup_count to account for CPUs that are offline also
adjust old_cpu_mask, otherwise further calls to fixup_irqs() could subtract
those again and create an imbalance in move_cleanup_count.
Fixes: 472e0b74c5 ('x86/IRQ: deal with move cleanup count state in fixup_irqs()')
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
master commit: e63209d3ba
master date: 2024-06-10 10:33:22 +0200
When the revision in hardware is newer than anything Xen has to hand,
'microcode_cache' isn't set up. Then, `xen-ucode` initiates the update
because it doesn't know whether the revisions across the system are symmetric
or not. This involves the patch getting all the way into the
apply_microcode() hooks before being found to be too old.
This is all a giant mess and needs an overhaul, but in the short term simply
adjust the apply_microcode() to return -EEXIST.
Also, unconditionally print the preexisting microcode revision on boot. It's
relevant information which is otherwise unavailable if Xen doesn't find new
microcode to use.
Fixes: 648db37a15 ("x86/ucode: Distinguish "ucode already up to date"")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 977d98e67c
master date: 2024-05-16 13:59:11 +0100
There's no point in forcing a system wide update of the MTRRs on all processors
when there are no changes to be propagated. On AP startup it's only the AP
that needs to write the system wide MTRR values in order to match the rest of
the already online CPUs.
We have occasionally seen the watchdog trigger during `xen-hptool cpu-online`
in one Intel Cascade Lake box with 448 CPUs due to the re-setting of the MTRRs
on all the CPUs in the system.
While there adjust the comment to clarify why the system-wide resetting of the
MTRR registers is not needed for the purposes of mtrr_ap_init().
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Release-acked-by: Oleksii Kurochko <oleksii.kurochko@gmail.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: abd00b037d
master date: 2024-05-15 19:59:52 +0100
In compare_cpu_pct(), there is a double -> unsigned long long converion when
calling compare(). In C, this discards the fractional part, resulting in an
out-of order sorting such as:
NAME STATE CPU(sec) CPU(%)
xendd --b--- 4020 5.7
icecream --b--- 2600 3.8
Domain-0 -----r 1060 1.5
neon --b--- 827 1.1
cheese --b--- 225 0.7
pizza --b--- 359 0.5
cassini --b--- 490 0.4
fusilli --b--- 159 0.2
bob --b--- 502 0.2
blender --b--- 121 0.2
bread --b--- 69 0.1
chickpea --b--- 67 0.1
lentil --b--- 67 0.1
Introduce compare_dbl() function and update compare_cpu_pct() to call it.
Fixes: 49839b535b ("Add xenstat framework.")
Signed-off-by: Leigh Brown <leigh@solinno.co.uk>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: e27fc7d15e
master date: 2024-05-15 19:59:52 +0100
The function itself properly handles and hands onwards failure from
create_perdomain_mapping(). Therefore its caller should respect possible
failure, too.
Fixes: 4b28bf6ae9 ("x86: re-introduce map_domain_page() et al")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 7270fdc7a0
master date: 2024-05-15 19:59:52 +0100
If xenstored runs out of memory it is possible for it to fail operations
that should succeed. libxl wasn't robust against this, and could fail
to ensure that the TTY path of a non-initial console was created and
read-only for guests. This doesn't qualify for an XSA because guests
should not be able to run xenstored out of memory, but it still needs to
be fixed.
Add the missing error checks to ensure that all errors are properly
handled and that at no point can a guest make the TTY path of its
frontend directory writable.
Signed-off-by: Demi Marie Obenour <demi@invisiblethingslab.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
master commit: 531d3bea5e
master date: 2024-05-11 00:13:43 +0100
libxl passes some information to libacpi to create the ACPI table for a PVH
guest, and among that information it's a bitmap of which vCPUs are online
which can be less than the maximum number of vCPUs assigned to the domain.
While the population of the bitmap is done correctly for HVM based on the
number of online vCPUs, for PVH the population of the bitmap is done based on
the number of maximum vCPUs allowed. This leads to all local APIC entries in
the MADT being set as enabled, which contradicts the data in xenstore if vCPUs
is different than maximum vCPUs.
Fix by copying the internal libxl bitmap that's populated based on the vCPUs
parameter.
Reported-by: Arthur Borsboom <arthurborsboom@gmail.com>
Link: https://gitlab.com/libvirt/libvirt/-/issues/399
Reported-by: Leigh Brown <leigh@solinno.co.uk>
Fixes: 14c0d328da ('libxl/acpi: Build ACPI tables for HVMlite guests')
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Tested-by: Leigh Brown <leigh@solinno.co.uk>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
master commit: 5cc7347b04
master date: 2024-05-11 00:13:43 +0100
Right now, Xen returns -ENOENT for both "the provided blob isn't correct for
this CPU", and "the blob isn't newer than what's loaded".
This in turn causes xen-ucode to exit with an error, when "nothing to do" is
more commonly a success condition.
Handle EEXIST specially and exit cleanly.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
master commit: 648db37a15
master date: 2024-05-09 18:19:49 +0100
Juergen Gross