Commit 0537d246f8 ("mm: add 'is_special_page' inline function...")
moved the is_special_page() checks first in its respective changes to
PoD code. While this is fine for p2m_pod_zero_check_superpage(), the
validity of the MFN is inferred in both cases from the p2m_is_ram()
check, which therefore also needs to come first in this 2nd instance.
Take the opportunity and address latent UB here as well - transform
the MFN into struct page_info * only after having established that
this is a valid page.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
There are no negative sizes. Make the function's parameter as well as
that of its derivates "unsigned int". Similarly make its local "count"
variable "unsigned int", and drop "todo" altogether. Don't use min_t()
anymore to calculate "count". Restrict its scope as well as that of
other local variables of the function.
While at it I've also noticed that {copy_{from,to},clear}_user_hvm()
have been returning "unsigned long" for no apparent reason, as their
respective "size" parameters have already been "unsigned int". Adjust
this as well as a slightly wrong comment there at the same time.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <pdurrant@amzn.com>
Implement hypercall that allows a fork to shed all memory that got allocated
for it during its execution and re-load its vCPU context from the parent VM.
This allows the forked VM to reset into the same state the parent VM is in a
faster way then creating a new fork would be. Measurements show about a 2x
speedup during normal fuzzing operations. Performance may vary depending how
much memory got allocated for the forked VM. If it has been completely
deduplicated from the parent VM then creating a new fork would likely be more
performant.
Signed-off-by: Tamas K Lengyel <tamas.lengyel@intel.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
VM forking is the process of creating a domain with an empty memory space and a
parent domain specified from which to populate the memory when necessary. For
the new domain to be functional the VM state is copied over as part of the fork
operation (HVM params, hap allocation, etc).
Signed-off-by: Tamas K Lengyel <tamas.lengyel@intel.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Julien Grall <jgrall@amazon.com>
We haven't used mini-os master for about 2 years now due to a stubdom
test failing [1]. Booting a guest with mini-os master used for building
stubdom didn't reveal any problem, so use master for unstable in order
to let OSStest find any problems not showing up in the local test.
[1]: https://lists.xen.org/archives/html/minios-devel/2018-04/msg00015.html
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Wei Liu <wl@xen.org>
All callers pass &this_cpu(cpu_sig) for the cpu_sig parameter, and all
implementations unconditionally return 0. Simplify it to be void.
Drop the long-stale comment on the AMD side, whose counterpart in
start_update() used to be "collect_cpu_info() doesn't fail so we're fine".
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
With the newly cleaned up vendor logic, each struct microcode_patch is a
trivial object in memory with no dependent allocations.
This is unlikely to change moving forwards, and function pointers are
expensive in the days of retpoline. Move the responsibility to xfree() back
to common code. If the need does arise in the future, we can consider
reintroducing the hook.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
No paths to apply_microcode() pass a NULL pointer, and other hooks don't
tolerate one in the first place. We can expect the core logic not to pass us
junk, so drop the checks.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
It turns out there are no callers of the hook(). The only callers are the
local, which can easily be rearranged to use the appropriate internal helper.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
The CPUID instruction is expensive. No point executing it twice when once
will do fine.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
This fixes the fact that qemu-deprivilege.md, non-cooperative-migration.md and
xenstore-migration.md don't currently get rendered at all, and are therefore
missing from xenbits.xen.org/docs
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Reviewed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Backport: 4.12
Commit 562a1c0f7e ("tools/xenstore: dont unlink connection object
twice") introduced a potential use after free problem in
domain_cleanup(): after calling talloc_unlink() for domain->conn
domain->conn is set to NULL. The problem is that domain is registered
as talloc child of domain->conn, so it might be freed by the
talloc_unlink() call.
With Xenstore being single threaded there are normally no concurrent
memory allocations running and freeing a virtual memory area normally
doesn't result in that area no longer being accessible. A problem
could occur only in case either a signal received results in some
memory allocation done in the signal handler (SIGHUP is a primary
candidate leading to reopening the log file), or in case the talloc
framework would do some internal memory allocation during freeing of
the memory (which would lead to clobbering of the freed domain
structure).
Fixes: 562a1c0f7e ("tools/xenstore: dont unlink connection object twice")
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Julien Grall <jgrall@amazon.com>
The domain is being passed in - no need to obtain it from p2m->domain.
Also drop a pointless cast and simplify expressions while touching this
code anyway.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
As per SDM rev 071 Cannon Lake has
- no CC3 residency MSR at 3FC,
- a CC1 residency MSR ar 660 (like various Atoms),
- a useless (always zero) CC3 residency MSR at 662.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
Use the DIV_ROUND_UP macro to replace open-coded divisor calculation
(((n) + (d) - 1) / (d)) to improve readability.
Signed-off-by: Simran Singhal <singhalsimran0@gmail.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
There's an outer scope rc already, and its use for the mem-sharing logic
does not conflict with its use elsewhere in the function.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
guest_physmap_remove_page() gets handed an MFN from the outside, yet
takes the necessary lock to prevent further changes to the GFN <-> MFN
mapping itself. While some callers, in particular guest_remove_page()
(by way of having called get_gfn_query()), hold the GFN lock already,
various others (most notably perhaps the 2nd instance in
xenmem_add_to_physmap_one()) don't. While it also is an option to fix
all the callers, deal with the issue in p2m_remove_page() instead:
Replace the ASSERT() by a conditional and split the loop into two, such
that all checking gets done before any modification would occur.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
It's not very nice to return from guest_physmap_add_entry() after
perhaps already having made some changes to the P2M, but this is pre-
existing practice in the function, and imo better than ignoring errors.
Take the liberty and replace an mfn_add() instance with a local variable
already holding the result (as proven by the check immediately ahead).
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
We're deliberately bringing XEN_COMPILE_ARCH and XEN_TARGET_ARCH out of
sync in this case, and hence HOSTCC won't get set from CC. Therefore
without this addition HOSTCC would not match a possible make command
line override of CC, but default to "gcc", likely causing the build to
fail for test_x86_emulator.c on systems with too old a gcc.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
The run* targets can be used to test whatever the tool chain is capable
of building, as long as at least the main harness source file builds.
Don't probe the tools chain, in particular to avoid issuing the warning,
in this case. While looking into this I also noticed the wording of the
respective comment isn't quite right, which therefore gets altered at
the same time.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
There is a bug in commit 5e4b419966 ("xen: credit2: only reset
credit on reset condition"). In fact, the aim of that commit was to
make sure that we do not perform too many credit reset operations
(which are not super cheap, and in an hot-path). But the check used
to determine whether a reset is necessary was the wrong one.
In fact, knowing just that some vCPUs have been skipped, while
traversing the runqueue (in runq_candidate()), is not enough. We
need to check explicitly whether the first vCPU in the runqueue
has a negative amount of credit.
Since a trace record is changed, this patch updates xentrace format file
and xenalyze as well
This should be backported.
Signed-off-by: Dario Faggioli <dfaggioli@suse.com>
Acked-by: George Dunlap <george.dunlap@citrix.com>
There have been report of stalls of guest vCPUs, when Credit2 was used.
It seemed like these vCPUs were not getting scheduled for very long
time, even under light load conditions (e.g., during dom0 boot).
Investigations led to the discovery that --although rarely-- it can
happen that a vCPU manages to run for very long timeslices. In Credit2,
this means that, when runtime accounting happens, the vCPU will lose a
large quantity of credits. This in turn may lead to the vCPU having less
credits than the idle vCPUs (-2^30). At this point, the scheduler will
pick the idle vCPU, instead of the ready to run vCPU, for a few
"epochs", which often times is enough for the guest kernel to think the
vCPU is not responding and crashing.
An example of this situation is shown here. In fact, we can see d0v1
sitting in the runqueue while all the CPUs are idle, as it has
-1254238270 credits, which is smaller than -2^30 = −1073741824:
(XEN) Runqueue 0:
(XEN) ncpus = 28
(XEN) cpus = 0-27
(XEN) max_weight = 256
(XEN) pick_bias = 22
(XEN) instload = 1
(XEN) aveload = 293391 (~111%)
(XEN) idlers: 00,00000000,00000000,00000000,00000000,00000000,0fffffff
(XEN) tickled: 00,00000000,00000000,00000000,00000000,00000000,00000000
(XEN) fully idle cores: 00,00000000,00000000,00000000,00000000,00000000,0fffffff
[...]
(XEN) Runqueue 0:
(XEN) CPU[00] runq=0, sibling=00,..., core=00,...
(XEN) CPU[01] runq=0, sibling=00,..., core=00,...
[...]
(XEN) CPU[26] runq=0, sibling=00,..., core=00,...
(XEN) CPU[27] runq=0, sibling=00,..., core=00,...
(XEN) RUNQ:
(XEN) 0: [0.1] flags=0 cpu=5 credit=-1254238270 [w=256] load=262144 (~100%)
We certainly don't want, under any circumstance, this to happen.
Let's, therefore, define a minimum amount of credits a vCPU can have.
During accounting, we make sure that, for however long the vCPU has
run, it will never get to have less than such minimum amount of
credits. Then, we set the credits of the idle vCPU to an even
smaller value.
NOTE: investigations have been done about _how_ it is possible for a
vCPU to execute for so much time that its credits becomes so low. While
still not completely clear, there are evidence that:
- it only happens very rarely,
- it appears to be both machine and workload specific,
- it does not look to be a Credit2 (e.g., as it happens when
running with Credit1 as well) issue, or a scheduler issue.
This patch makes Credit2 more robust to events like this, whatever
the cause is, and should hence be backported (as far as possible).
Reported-by: Glen <glenbarney@gmail.com>
Reported-by: Tomas Mozes <hydrapolic@gmail.com>
Signed-off-by: Dario Faggioli <dfaggioli@suse.com>
Reviewed-by: George Dunlap <george.dunlap@citrix.com>
cpu_request_microcode() is still a confusing mess to follow, with sub
functions responsible for maintaining offset. Rewrite it so all container
structure handling is in this one function.
Rewrite struct mpbhdr as struct container_equiv_table to aid parsing. Drop
container_fast_forward() entirely, and shrink scan_equiv_cpu_table() to just
its searching/caching logic.
container_fast_forward() gets logically folded into the microcode blob
scanning loop, except that a skip path is inserted, which is conditional on
whether scan_equiv_cpu_table() thinks there is appropriate microcode to find.
With this change, we now scan to the end of all provided microcode containers,
and no longer give up at the first applicable one.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
With all the necessary cleanup now in place, fold struct microcode_header_amd
into struct microcode_patch and drop the struct microcode_amd temporary
ifdef-ary.
This removes the memory allocation of struct microcode_amd which is a single
pointer to a separately allocated object, and therefore a waste.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Just as on the Intel side, there is no point having
get_ucode_from_buffer_amd() make $N memory allocations and free $N-1 of them.
Delete get_ucode_from_buffer_amd() and rewrite the loop in
cpu_request_microcode() to have 'saved' point into 'buf' until we finally
decide to duplicate that blob and return it to our caller.
Introduce a new struct container_microcode to simplify interpreting the
container format. Doubly indent the logic to substantially reduce the churn
in a later change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
To simplify future cleanup, rename this variable.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Although it is logically a step in the wrong direction overall, it simplifies
the rearranging of cpu_request_microcode() substantially for microcode_fits()
to take struct microcode_header_amd directly, and not require an intermediate
struct microcode_amd pointing at it.
Make this change (taking time to rename 'mc_amd' to its eventual 'patch' to
reduce the churn in the series), and a later cleanup will make it uniformly
take a struct microcode_patch.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
We only stash the microcode blob size so it can be audited in
microcode_fits(). However, the patch size check depends only on the CPU
family.
Move the check earlier to when we are parsing the container, which avoids
caching bad microcode in the first place, and allows us to avoid storing the
size at all.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
We currently copy the entire equivalency table, and the single correct
microcode. This is not safe to heterogeneous scenarios, and as Xen doesn't
support such situations to begin with, can be used to simplify things further.
The CPUID.1.EAX => processor_rev_id mapping is fixed for an individual part.
We can cache the single appropriate entry on first discovery, and forgo
duplicating the entire table.
Alter install_equiv_cpu_table() to be scan_equiv_cpu_table() which is
responsible for checking the equivalency table and caching appropriate
details. It now has a check for finding a different mapping (which indicates
that one of the tables we've seen is definitely wrong).
A return value of -ESRCH is now used to signify "everything fine, but nothing
applicable for the current CPU", which is used to select the
container_fast_forward() path.
Drop the printk(), as each applicable error path in scan_equiv_cpu_table()
already prints diagnostics.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
... rather than collecting it repeatedly in microcode_fits(). This brings the
behaviour in line with the Intel side.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
All code works fine with it having its correct type, and it even allows us to
drop two casts in a printk().
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
The microcode revision of whichever CPU runs cpu_request_microcode() is not
necessarily applicable to other CPUs.
If the BIOS left us with asymmetric microcode, rejecting updates in
cpu_request_microcode() would prevent us levelling the system even if only up
to the final level. Also, failing to cache microcode misses an opportunity to
get beyond the final level via the S3 path.
Move check_final_patch_levels() earlier and use it in apply_microcode().
Reword the error message to be more informative, and use -ENXIO as this corner
case has nothing to do with permissions.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
cpu_request_microcode() doesn't know the buffer is at least 4 bytes long
before inspecting UCODE_MAGIC.
install_equiv_cpu_table() doesn't know the boundary of the buffer it is
interpreting as an equivalency table. This case was clearly observed at one
point in the past, given the subsequent overrun detection, but without
comprehending that the damage was already done.
Make the logic consistent with container_fast_forward() and pass size_left in
to install_equiv_cpu_table().
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
With all the necessary cleanup now in place, fold struct
microcode_header_intel into struct microcode_patch and drop the struct
microcode_intel temporary ifdef-ary.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Rewrite the size checks in a way which doesn't depend on Xen being compiled as
64bit.
Introduce a check missing from the old code, that total_size is a multiple of
1024 bytes, and drop unnecessary defines/macros/structures.
No practical change in behaviour.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Implement a new get_ext_sigtable() helper to abstract the logic for
identifying whether an extended signature table exists. As part of this,
rename microcode_intel.bits to data and change its type so it can be usefully
used in combination with the datasize header field.
Also, replace the sigmatch() macro with a static inline with a more useful
API, and an explanation of why it is safe to drop one of the previous
conditionals.
No practical change in behaviour.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Every caller actually passes a struct microcode_header_intel *, but it is more
helpful to us longterm to take struct microcode_patch *. Implement the
helpers with proper types, and leave a comment explaining the Pentium Pro/II
behaviour with empty {data,total}size fields.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
cpu_request_microcode() needs to scan its container and duplicate one blob,
but the get_next_ucode_from_buffer() helper duplicates every blob in turn.
Furthermore, the length checking is only safe from overflow in 64bit builds.
Delete get_next_ucode_from_buffer() and alter the purpose of the saved
variable to simply point somewhere in buf until we're ready to return.
This is only a modest reduction in absolute code size, but avoids making
memory allocations for every blob in the container.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
microcode_sanity_check()'s callers actually call it with a mixture of
microcode_intel(/patch) and microcode_header_intel pointers, which is fragile.
Rework it to take struct microcode_patch *, which in turn requires
microcode_update_match()'s type to be altered.
No functional change - compiled binary is identical.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Currently, each cpu_request_microcode() allocates a struct microcode_patch,
which is a single pointer to a separate allocated structure. This is
wasteful.
Fixing this is complicated because the common microcode_free_patch() code is
responsible for freeing struct microcode_patch, despite this being asymmetric
with how it is allocated.
Make struct microcode_patch fully opaque to the common logic. This involves
moving the responsibility for freeing struct microcode_patch fully into the
free_patch() hook.
In each vendor logic, use some temporary ifdef-ary (cleaned up in subsequent
changes) to reduce the churn as much as possible, and forgo allocating the
intermediate pointer in cpu_request_microcode().
Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
The orders of start and end are inverted in order to calculate the
size of the copy operation.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
This is to augment SYSCALL, which we've been supporting for quite some
time.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
AMD CPUs permit the insn everywhere (even outside of protected mode),
while Intel ones restrict it to 64-bit mode. While at it also comment
about the apparently missing CPUID bit check.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
The involved comparison was, very likely inadvertently, converted from
>= to > when making changes unrelated to the actual family range.
Fixes: 9841eb71ea ("x86/cpuid: Drop a guests cached x86 family and model information")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Paul Durrant <paul@xen.org>
The parameter 'params' of xc_set_parameters() should never be modified.
So mark it as const.
Signed-off-by: Julien Grall <jgrall@amazon.com>
Reviewed-by: Ian Jackson <ian.jackson@eu.citrix.com>
OCaml is using a string to describe the parameter 'keys' of
xc_send_debug_keys(). Since Ocaml 4.06.01, String_val() will return a
const char * when using -safe-string. This will result to a build
failure because xc_send_debug_keys() expects a char *.
The function should never modify the parameter 'keys' and therefore the
parameter should be const. Unfortunately, this is not directly possible
because DECLARE_HYPERCALL_BOUNCE() is expecting a non-const variable.
A new macro DECLARE_HYPERCALL_BOUNCE_IN() is introduced and will take
care of const parameter. The first user will be xc_send_debug_keys() but
this can be used in more place in the future.
Reported-by: Dario Faggioli <dfaggioli@suse.com>
Signed-off-by: Julien Grall <jgrall@amazon.com>
Reviewed-by: Ian Jackson <ian.jackson@eu.citrix.com>
The fields set_parameter.params and debug.keys should never be modified
by the hypervisor. So mark them as const.
Signed-off-by: Julien Grall <jgrall@amazon.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
arm*/head.o isn't in obj-y or extra-y, so make don't load the
associated .*.d file (or .*.cmd file when if_changed will be used).
There is a workaround where .*.d file is added manually into DEPS.
Changing DEPS isn't needed, we can simply add head.o into extra-y and
the dependency files will be loaded.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Acked-by: Julien Grall <jgrall@amazon.com>
At the moment, early printk can only be configured on the make command
line. It is not very handy because a user has to remove the option
everytime it is using another command other than compiling the
hypervisor.
Furthermore, early printk is one of the few odds one that are not
using Kconfig.
So this is about time to move it to Kconfig.
The new kconfigs options allow a user to eather select a UART driver
to use at boot time, and set the parameters, or it is still possible
to select a platform which will set the parameters.
If CONFIG_EARLY_PRINTK is present in the environment or on the make
command line, make will return an error.
Signed-off-by: Julien Grall <julien.grall@arm.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Tested-by: Stefano Stabellini <sstabellini@kernel.org>
Reviewed-by: Julien Grall <jgrall@amazon.com>
Jan Beulich
Tamas K Lengyel
Juergen Gross
Andrew Cooper
Simran Singhal
Dario Faggioli
Roger Pau Monné
Julien Grall
Anthony PERARD