| Age | Commit message (Collapse) | Author |
|---|
|
Update the kvm_is_tdp_enabled() function to include Zhaoxin CPUs by
querying checking "Intel compatible" when checking for EPT (Extended
Page Tables) support. This ensures that Zhaoxin CPUs use the
appropriate TDP (Two-Dimensional Paging) parameter check, similar to
Intel CPUs.
Signed-off-by: Frank Zhu <frankzhu@zhaoxin.com>
Link: https://patch.msgid.link/20260910115414.3015260-4-ewanhai-oc@zhaoxin.com
[sean: massage changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
kvm_is_mediated_pmu_enabled() returns kvm_amd's enable_mediated_pmu
module parameter unless the host CPU is Intel. Zhaoxin CPUs implement
VMX-compatible virtualization and are handled by KVM's VMX backend, so
on Zhaoxin hosts the helper wrongly reports kvm_amd's parameter.
Gate the check on host_cpu_is_intel_compatible so that Zhaoxin CPUs,
like Intel CPUs, query kvm_intel's enable_mediated_pmu instead. This
does not change behavior on Intel, and AMD and other non-Intel-
compatible CPUs still fall back to kvm_amd's parameter.
Signed-off-by: Frank Zhu <frankzhu@zhaoxin.com>
Link: https://patch.msgid.link/20260910115414.3015260-3-ewanhai-oc@zhaoxin.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add helpers to detect Zhaoxin CPUs using the "CentaurHauls" and
" Shanghai " vendor strings.
KVM handles Zhaoxin CPUs with its VMX backend. Add
host_cpu_is_intel_compatible for selftests that exercise behavior shared
by Intel and Zhaoxin CPUs, while retaining host_cpu_is_zhaoxin for
Zhaoxin-specific behavior.
Signed-off-by: Frank Zhu <frankzhu@zhaoxin.com>
Link: https://patch.msgid.link/20260910115414.3015260-2-ewanhai-oc@zhaoxin.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Skip the xAPIC IPI migration test instead of failing outright if migration
is forced by the user, but is unsupported due to lack of NUMA support.
Failing the test is arguably "fine" for opt-in behavior, but in practice
doesn't allow running the test in forced-migration mode in any kind of
automated test environment.
Reviewed-by: Shivank Garg <shivankg@amd.com>
Link: https://patch.msgid.link/20260903001625.2792367-8-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Rework the interface to the xAPIC IPI test to automatically run the
migration testcase if at least two NUMA nodes are found. The test exists
specifically to validate KVM's handling of migration of the APIC backing
page, i.e. forcing end users to opt-in to running the test in migration
mode largely defeats the purpose of the test.
Run both the "sleeping" and "migration" testcases by default, e.g. so that
the more basic testcase will fail if KVM completely breaks IPI delivery.
Keep the -m / migration parameter so that infrastructure that wants to
specifically test the migration case can do so and not get false passes.
Reviewed-by: Shivank Garg <shivankg@amd.com>
Link: https://patch.msgid.link/20260903001625.2792367-7-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Extract and slightly adopt the xAPIC IPI test's logic for getting the mask
and number of usable memory NUMA nodes into a common helper. To allow for
friendlier behavior when the underlying kernel doesn't support NUMA, or the
test was run without sufficient permissions, zero out the mask and return
"zero nodes" if get_mempolicy() fails with ENOSYS or EPERM respectively.
I.e. allow tests to skip (sub)tests instead of throwing an assert.
Suggested-by: Shivank Garg <shivankg@amd.com>
Reviewed-by: Shivank Garg <shivankg@amd.com>
Link: https://patch.msgid.link/20260903001625.2792367-6-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Compute the node masks for the source (from) and destination (to) NUMA
nodes in the xAPIC IPI test instead of pre-filling an array of masks with
node per mask. Computing the mask on-demand is technically slower, but
doesn't require a large-ish on-stack array, and more importantly allows the
test to use a generic "get next NUMA node" API without having to commit all
of KVM selftests to using a large array of single-bit nodemasks.
Implement said API as a common KVM NUMA API so that it can be used by other
tests, e.g. in guest_memfd tests. Deliberately make @from "exclusive" as
the anticipated usage in KVM selftests is to select the next, *different*
node, i.e. so that users don't have to copy+paste code to assert that the
found node is different than the starting node. The obvious downside is
that implementing the exclusive logic forces callers to pass -1 instead of
0 when the goal is to find the first node in the mask, but that arguably
yields more intuitive code anyways.
From an overall test functionality/coverage perspective, no functional
change intended (the walking pattern of node migration should be unchanged).
Reviewed-by: Shivank Garg <shivankg@amd.com>
Link: https://patch.msgid.link/20260903001625.2792367-5-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Replace the open-coded sizeof() * 8 calculations in do_migrations() with
BITS_PER_TYPE().
No functional change intended.
Signed-off-by: Shivank Garg <shivankg@amd.com>
Link: https://patch.msgid.link/20260903001625.2792367-4-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use MAXNODE_FOR_MASK() to compute the @maxnode argument when migrating
pages in the xAPIC IPI test, as the current code incorrectly passes the
number of bytes, not the number of bits, in the mask, and also fails to
account for the kernel's off-by-one bug. I.e. the test sets maxnode to 8,
and so the kernel only checks node IDs 0-6 even though the nodemask covers
node IDs 0-63.
Fixes: 678e90a349a4 ("KVM: selftests: Test IPI to halted vCPU in xAPIC while backing page moves")
Signed-off-by: Shivank Garg <shivankg@amd.com>
[sean: use MAXNODE_FOR_MASK(), rewrite changelog]
Link: https://patch.msgid.link/20260903001625.2792367-3-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add and use MAXNODE_FOR_MASK() to compute the "correct" maxnode value that
is passed to various NUMA-related syscalls, e.g. get_mempolicy(), mbind(),
migrate_pages(), etc. In quotes, because the kernel has an undocumented,
longstanding off-by-one bug that requires userspace to specify the number
of bits plus one, i.e. the max node plus two.
The kernel bug has been known since 2007[*]:
: And this is I think the reason why we can't change this now. I assume
: numactl allocates 1024 bits (0 to 1023) and passes 1025 to make sure all
: 1024 bits are processed. If we change it now, kernel will process 1025
: bits (0 to 1024) and overflow the allocated bitmask. If it happens to be
: at the border of mmaped vma, it's a segfault...
But unfortunately the manpages haven't yet been updated, e.g.
: The maxnode argument is the maximum node number in the bit mask plus one
Link: https://lore.kernel.org/all/63ccc890-fd57-118b-5997-e0259f507d28@suse.cz[*]
Reviewed-by: Shivank Garg <shivankg@amd.com>
Link: https://patch.msgid.link/20260903001625.2792367-2-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
run_test() returns as soon as it has confirmed that disabling NX huge
pages fails with -EPERM, without freeing the VM created a few lines
earlier.
Jump to the kvm_vm_free() at the end of the function instead, matching
the pattern used for the SEV smoke test in the previous patch.
Fixes: b774da3f2e57 ("KVM: selftests: Test disabling NX hugepages on a VM")
Signed-off-by: Gokul K <gokul02k@gmail.com>
Link: https://patch.msgid.link/20260811050029.530698-3-gokul02k@gmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
test_sev() returns directly out of the UCALL_DONE case rather than
leaving the loop, so the plain-SEV path skips the kvm_vm_free() at the
end of the function. The SEV-ES path is unaffected; it breaks out of
the loop and frees the VM correctly.
main() invokes test_sev() once per supported SEV VM type, so a full run
leaks a VM and its file descriptors. Nothing fails today because the
process exits shortly afterwards, which is presumably why this was not
noticed, but the leak also means the plain-SEV path never exercises VM
teardown.
Use a goto so UCALL_DONE joins the existing exit path. A plain break
would only leave the switch statement and spin the loop again.
Fixes: be250ff437fa ("KVM: selftests: Add a basic SEV smoke test")
Signed-off-by: Gokul K <gokul02k@gmail.com>
Link: https://patch.msgid.link/20260811050029.530698-2-gokul02k@gmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
validate_dst() is passed a starting offset and a byte count, but bounds
its loop with "i < nr_bytes", comparing the offset against the count
rather than against offset + nr_bytes. Only calls that pass offset 0
check the region they were asked to check. Every other call checks too
few bytes, and any call whose offset is >= the count returns without
checking anything at all.
Of the 1990 validate_dst() calls a full run makes per VM type, 740 (37%)
check zero bytes, only the 29 zero-offset calls are correct, and 62% of
the intended bytes are checked overall. The unaligned offsets the sweep
deliberately picks to exercise partial chunks are precisely the ones
that end up checking nothing.
The bug is latent today: validate_buffers() runs first and compares all
of src[] against dst[], so a genuine {en,de}crypt corruption is still
caught. That is presumably why this went unnoticed.
Bound the loop with offset + nr_bytes, and rename the parameter to
offset so it cannot be mistaken for a length again. The caller already
guarantees offset + nr_bytes <= BUFFER_SIZE, so the corrected loop stays
in bounds.
Fixes: 6edd35e77a42 ("KVM: selftests: Add a test to verify SEV {en,de}crypt debug ioctls")
Signed-off-by: Gokul K <gokul02k@gmail.com>
Link: https://patch.msgid.link/20260808061934.400892-1-gokul02k@gmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
vcpu_get_reg_list() allocates reg_list with calloc() and immediately
dereferences it via reg_list->n without checking for allocation failure,
unlike every other allocation in this file which is guarded by
TEST_ASSERT(). If calloc() returns NULL the test crashes with a NULL
pointer dereference instead of a clean failure message.
Add the missing TEST_ASSERT() check.
Fixes: fd02029a9e01 ("KVM: selftests: Add aarch64 get-reg-list test")
Signed-off-by: Chaithanya Lagisetty <nagachaithanya9911@gmail.com>
Link: https://patch.msgid.link/20260901090839.3338955-1-nagachaithanya9911@gmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
uffd_setup_demand_paging() passes the element size as the first argument
to calloc() and the element count as the second:
uffd_desc->pipefds = calloc(sizeof(int), num_readers);
calloc() expects the number of elements first and the size of each element
second. The product, and therefore the allocation size, is the same either
way, so this is not a functional change, but the arguments are misleading
and GCC 14 diagnoses them with -Wcalloc-transposed-args, which is enabled
by -Wextra.
Swap the arguments to match the calloc() prototype.
Signed-off-by: Chaithanya Lagisetty <nagachaithanya9911@gmail.com>
Reviewed-by: Gautam Menghani <gautam@linux.ibm.com>
Link: https://patch.msgid.link/20260904052033.3648259-1-nagachaithanya9911@gmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
demand_paging_test parses the -d option with strtoul() and then asserts
that the result is not negative:
p.uffd_delay = strtoul(optarg, NULL, 0);
TEST_ASSERT(p.uffd_delay >= 0, "A negative UFFD delay is not supported.");
p.uffd_delay is a useconds_t, which is an unsigned type, so the comparison
is always true and the assertion can never fire. GCC points this out with
-Wtype-limits, which is enabled by -Wextra.
As a result, "-d -1" is accepted and converted to a very large unsigned
delay, causing each demand paging fault to sleep for an unexpectedly long
time in usleep() instead of rejecting the argument up front. The return
value of strtoul() is not validated either, so a non-numeric argument
such as "-d abc" is silently treated as a zero delay.
Use atoi_non_negative() instead. It rejects negative values, unparsable
input, and trailing garbage. It is already used a few lines below for -v,
and hexadecimal input keeps working because atoi_paranoid() also passes a
base of 0 to strtol().
Fixes: 0119cb365c93 ("KVM: selftests: Add configurable demand paging delay")
Signed-off-by: Chaithanya Lagisetty <nagachaithanya9911@gmail.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Reviewed-by: Gautam Menghani <gautam@linux.ibm.com>
Link: https://patch.msgid.link/20260907045246.3783304-1-nagachaithanya9911@gmail.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Rename vmreadz() => vmread() (and vmread() => __vmread()) and have it
assert success instead of returning garbage on failure. While the eVMCS
path avoids clobbering *value on failure, the VMREAD path unconditionally
stores tmp to *value, without ever zeroing tmp. I.e. if VMREAD did fail,
*value would be loaded with random stack data, not '0'.
An audit of the code confirms that nothing is actually relying on vmreadz()
to zero the result on failure, as all paths either check for a non-zero
value, propagate the result into vmwrite(), or guard the vmread() with a
pre-check that the field exists.
Opportunistically fix the bug in what is now __vmread(), so that the code
isn't consuming uninitialized stack data. Begrudgingly keep zeroing
value in vmread(), as gcc-15 at least isn't smart enough to figure out
that value won't be used uninitialized (though to be fair, strictly
speaking it isn't actually unreachable code).
Link: https://patch.msgid.link/20260904002423.3154436-18-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Assert that vmwrite() succeeds instead of forcing all callers to do the
same (or as is much more common, ignoring VM-Fail and likely letting the
test fail in weird ways), opportunistically provide a more verbose error
message on failure.
All usage of vmwrite() "guarantees" success, either because the field in
question is unconditionally emulated by KVM, by virtue of earlier support
checks, or by falling over (later in the test) if the VMWRITE fails.
Cc: Yosry Ahmed <yosry@kernel.org>
Link: https://patch.msgid.link/20260904002423.3154436-17-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add __vmwrite() in anticipation of having vmwrite() assert success, and use
the "failure allowed" version in the common VMCS setup code, which is the
only code that wants to write fields that may or may not exist (although
arguably the code should check if the field exists instead of blindly
writing).
E.g. When doing initial configuration of a VMCS, the core library doesn't
want to assume secondary controls are supported, and so blindly does a
VMWRITE and pivots on success/failure.
All usage of vmwrite() outside of common setup already asserts success, or
100% relies on VMWRITE succeeding in order for the test to function
correctly.
No functional change intended.
Link: https://patch.msgid.link/20260904002423.3154436-16-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Assert success in the "outer" APIs for VMLAUNCH and VMRESUME to dedup a
pile of code, and switch to the double-underscores versions as necessary,
to make it more obvious which paths expect VM-Enter *failure*.
Link: https://patch.msgid.link/20260904002423.3154436-15-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Dedup the assembly code for VMLAUNCH vs. VMRESUME, for both architectural
VMCS and eVMCS, as the difference is literally only the actual VM-Enter
instruction for VMLAUNCH vs. VMRESUME, and VMWRITE vs. MOV for VMCS vs.
eVMCS.
Opportunistically place the implementations in vmx.c so that the lengthy
sequence (~580 bytes of code) aren't inlined for every single VM-Enter.
000000000041b910 <__vmlaunch>:
41b910: 80 3d 61 5e 10 00 00 cmpb $0x0,0x105e61(%rip) # 521778 <enable_evmcs>
41b917: 0f 85 13 01 00 00 jne 41ba30 <__vmlaunch+0x120>
41b91d: ba 14 6c 00 00 mov $0x6c14,%edx
41b922: b9 16 6c 00 00 mov $0x6c16,%ecx
41b927: 6a 00 push $0x0
41b929: 0f 79 d4 vmwrite %rsp,%rdx
41b92c: 48 8d 05 82 00 00 00 lea 0x82(%rip),%rax # 41b9b5 <__vmlaunch+0xa5>
41b933: 0f 79 c8 vmwrite %rax,%rcx
<xchg GPRs>
41b9ae: 0f 01 c2 vmlaunch
41b9b1: 48 ff 04 24 incq (%rsp)
<xchg GPRs>
41ba2d: 58 pop %rax
41ba2e: c3 ret
41ba2f: 90 nop
41ba30: 48 8b 35 39 5d 10 00 mov 0x105d39(%rip),%rsi # 521770 <current_evmcs>
41ba37: c7 86 38 03 00 00 00 00 00 00 movl $0x0,0x338(%rsi)
41ba41: 6a 00 push $0x0
41ba43: 48 89 a6 68 02 00 00 mov %rsp,0x268(%rsi)
41ba4a: 48 8d 05 83 00 00 00 lea 0x83(%rip),%rax # 41bad4 <__vmlaunch+0x1c4>
41ba51: 48 89 46 50 mov %rax,0x50(%rsi)
<xchg GPRs>
41bacd: 0f 01 c2 vmlaunch
41bad0: 48 ff 04 24 incq (%rsp)
<xchg GPRs>
41bb4c: 58 pop %rax
41bb4d: c3 ret
No functional change intended (though the code generation will be quite
different).
Cc: Yosry Ahmed <yosry@kernel.org>
Link: https://patch.msgid.link/20260904002423.3154436-14-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Drop the return codes from evmcs_vmptrld() and load_evmcs() as they are
hardcoded to return '0' and "true" respectively.
No functional change intended.
Link: https://patch.msgid.link/20260904002423.3154436-13-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add macros to build helpers for VMXON, VMPTRLD, and VMCLEAR, i.e. the VMX
instructions that take a VMCS address as input, and don't have outputs.
Getting the params right is annoying, especially when using KVM_ASM_SAFE().
As a bonus, this adds non-asserting double-underscores versions of VMXON
and VMCLEAR for "free".
Link: https://patch.msgid.link/20260904002423.3154436-12-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Now that load_vmcs() either asserts or returns success, drop the return
code and the many asserts in its callers.
Link: https://patch.msgid.link/20260904002423.3154436-11-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Assert that vmptrld() succeeds instead of forcing all callers to do the
same, and opportunistically provide a macro to print (hopefully) useful
information on VMX instruction failures.
Leave load_vmcs() as-is for the moment to minimize the amount of churn
since the vmptrld() changes aren't completely trivial.
Link: https://patch.msgid.link/20260904002423.3154436-10-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Continue removing dead error "handling", and assert that VMCLEAR
succeeds instead of forcing callers of load_vmcs() to assert success.
For all intents and purposes, no functional change intended.
Link: https://patch.msgid.link/20260904002423.3154436-9-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
succeed
Assert that vmxon() succeeds, and by extension prepare_for_vmx_operation().
Forcing every caller to assert that a core preparation helper succeeds is
absurd and generates hard-to-read code.
For all intents and purposes, no functional change intended.
Link: https://patch.msgid.link/20260904002423.3154436-8-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Assert success in vmptrst() instead of punting to the caller, as literally
every user expects VMPTRST to succeed (and because '0' is a valid physical
address).
Link: https://patch.msgid.link/20260904002423.3154436-7-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Check that VMPTRLD when eVMCS is active results in #UD. This matches
genuine Hyper-V's behavior.
Use KVM_ASM_SAFE framework to handle #UD from VMPTRLD. Unfortunately, the
same trick cannot be applied to the existing #UD check on VMLAUNCH as
VMLAUNCH clobbers all registers which KVM_ASM_SAFE depends on. Keep
VMLAUNCH handling separately. Deliberately use a double-underscores
prefix instead of a "safe" postfix for the helper, to communicate that the
inner/outer helpers handle more than just exceptions, they also handle
VM-Fail conditions.
Assisted-by: Claude:claude-sonnet-4-6
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
[sean: use __ instead of _safe]
Link: https://patch.msgid.link/20260904002423.3154436-6-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Previously, VMPTRST was forbidden with eVMCS and selftests were mocking the
correct behavior in vmptrst() by returning enlightened vmptr directly.
Since KVM's behavior has changed to match genuine Hyper-V, adjust evmcs
test accordingly.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Link: https://patch.msgid.link/20260904002423.3154436-5-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use a straight MOV instead of an XOR to zero R9 in the non-faulting path of
__KVM_ASM_SAFE() so as not to clobber RFLAGS. XOR generates smaller code,
but optimizing performance by minimizing the code footprint is just about
the last thing that KVM selftests care about.
The bug is relatively benign in the current code base, e.g. it marginally
degrades the coverage in the fastops test, but otherwise does no harm. But
the bug will become actively problematic if/when the "safe assembly" macros
are used for instructions that communicate results via RFLAGS, e.g. for VMX
instructions, which signal failure via RFLAGS.CF and RFLAGS.ZF.
Opportunistically use R9D instead of the full R9, as bits 63:32 of GPRs are
zeroed on writes to bit 31:0.
Reported-by: Sashiko Bot <sashiko-bot@kernel.org>
Closes: https://lore.kernel.org/all/20260818165258.2613603-5-vkuznets@redhat.com
Cc: Vitaly Kuznetsov <vkuznets@redhat.com>
Link: https://patch.msgid.link/20260904002423.3154436-4-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add an arch hook to communicate to the core physical page allocator that
page table allocations need to be naturally aligned. The hook will be used
for upcoming PowerPC support, which has MMU tables that are larger than the
guest base page size and need to be naturally aligned.
Bury the hook in the physical page allocator to avoid having to plumb a
rarely used param into the main APIs.
Cc: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-21-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Automatically choose the minimum GPA for physical page allocations based on
the region type instead of sprinkling the logic over various wrappers and
tests. All usage falls into three categories: (a) don't care, just use the
bare minimum GPA, (b) page tables, use a slightly higher min to keep low
memory available, (c) custom memslot, use the base of the memslot. I.e.
there isn't a strong need to allow completely custom minimums.
Opportunistically move the vGIC LPI stress test's "gpa_base" variable into
setup_memslot() as the only reason it was a global variable was so that it
could be passed as the minimum GPA for allocations from the extra memslot.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-20-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use a proper selftests TEST_ASSERT() in vm_get_mem_region() so that bad
input is guaranteed to fail the test, and with the standard error message
format.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-19-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Take the memory region type instead of the memslot in all page allocator
APIs, except for the innermost quad-underscores function. This will allow
automatically selecting the minimum GPA based on the allocation type, which
can't be done using the memslot due to all regions sharing a single memslot
by default. PowerPC support will also leverage the type to ensure that
page table allocations are naturally aligned.
No functional change intended.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-18-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use the TEST_EXTRA region/memslot in the "set memory region" test instead
of a custom memslot, so that the primary vm_phy_pages_alloc() APIs can be
reworked to take a region type instead of a raw memslot.
For all intents and purposes, no functional change intended, as
vm_override_mem_region() is barely more than a wrapper for
vm_userspace_mem_region_add().
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-17-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use the TEST_EXTRA region/memslot for x86's smaller MAXPHYADDR test instead
of a custom memslot, so that the primary vm_phy_pages_alloc() APIs can be
reworked to take a region type instead of a raw memslot.
For all intents and purposes, no functional change intended, as
vm_add_extra_mem_region() is barely more than a wrapper for
vm_userspace_mem_region_add().
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-16-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use the TEST_EXTRA region/memslot for arm64's vGIC LPI stress test instead
of a custom memslot, so that the primary vm_phy_pages_alloc() APIs can be
reworked to take a region type instead of a raw memslot.
For all intents and purposes, no functional change intended, as
vm_override_mem_region() is barely more than a wrapper for
vm_userspace_mem_region_add().
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-15-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
And another memory region type to deal with extra, one-off memory regions,
and use the new type to manage x86's SMRAM memslot, as another step towards
taking the region type instead of the raw memslot in the physical page
allocator APIs.
Alternatively, SMRAM setup could simply use the quad-underscore API to
continue passing in the memslot, but a surprising number of tests use an
"extra" memslot for a variety of reasons. I.e. allocating memory from one
(and exactly one) extra memslot isn't all that rare, and so should be
treated as normal behavior, not as something extraordinary, as
quad-underscore functions typically suggest.
Opportunistically add comments to document the intended usage of the types,
as the difference between DATA, TEST_DATA, and TEST_EXTRA in particular
isn't exactly obvious.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-14-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add and use APIs to override memory region types instead of open coding the
logic in tests, partly to dedup test code, but mostly to add some amount of
protection against conflicting overrides. This will allow converting more
tests to override memory regions without taking on too much risk of bugs in
the long term, e.g. due to silently overriding a region multiple times.
Deliberately "unroll" the loops in __vm_create() and the CMMA test's
create_main_memslot() to avoid filling all memslots in anticipation of
adding an "extra" memslot that is for test use and *only* test use, i.e.
shouldn't be configured by core library code.
Converting tests to overriding memory regions will allow taking the region
type instead of the raw memslot in the physical page allocator APIs.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Acked-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-13-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Initialize vm->memslots[] with an invalid memslots value during very
initial VM creation so that failure to configure the core memory regions
generates an immediate assert instead of potential weirdness, e.g. if slot0
is coincidentally valid. This will also allow hardening the region APIs to
guard overriding a region's memslot multiple times.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Acked-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-12-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Add vm_alloc_page_table_pages() and use it throughout arch code to replace
open coded equivalents.
No functional change intended.
Reviewed-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Acked-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Acked-by: Anup Patel <anup@brainfault.org>
Acked-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-11-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use the quad-underscores page allocator in s390's IRQ routing test in
anticipation of dropping the @min_gpa param from the primary APIs.
Note, it's not entirely clear that the test *needs* to use a very magical
@min_gpa, i.e. it's possible a more elegant solution exists. But that's a
problem for a different day.
No functional change intended.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Acked-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-10-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use the quad-underscores physical page allocator API in the memslot perf
test in anticipation of converting the outer versions to take the memory
region type instead of the raw memslot. Stating the obvious, the memslot
perf test creates and allocates from multiple memslots, and so can't use
MEM_REGION_TEST_DATA to communicate which region/slot to allocate from.
No functional change intended.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-9-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Move the single-page allocator APIs to kvm_util.h as inline functions, as
the are trivial wrappers, and unlike __vm_phy_pages_alloc(), are expected
to remain that way for the foreseeable future.
No functional change intended.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-8-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Extend the page allocator to support naturally aligned allocations, as some
flavors of PowerPC page tables require multiple guest pages per table, and
those tables need to be naturally aligned.
Take care to align the potential base address *inside* the retry loop, i.e.
the base needs to be (re)aligned on every attempt.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-7-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
physical pages
Rework the core loop of the physical page allocator to use a goto instead
of do-while loop when retrying with a new base address. The do-while loop
obviously works, but it obfuscates the success vs. failure paths, and in
general is convoluted.
For all intents and purposes, no functional change intended.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[Rebased to latest mainline tree]
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
[sean: separate from alignment change, use enomem label, write changelog]
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Acked-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-6-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Rename the "num" param in the physical page allocator APIs to "nr_pages" to
make the unit and role of the param more obvious.
Opportunistically rework the function comment to drop the superfluous
kerneldoc style, and to very explicitly call out that success is guaranteed.
No functional change intended.
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-5-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Now that LoongArch uses KVM_GUEST_PAGE_TABLE_MIN_PADDR for its page table
page allocations, use the common vm_alloc_page_table() instead of open
coding the same.
Opportunistically drop the assert that the allocation succeeded, as the
allocator itself guarantees success.
No functional change intended.
Reviewed-by: Bibo Mao <maobibo@loongson.cn>
Tested-by: Itaru Kitayama <itaru.kitayama@fujitsu.com>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-4-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Use the common minimum GPA for page tables (0x180000) on LoongArch instead
of LoongArch's custom minimum of 0x200000, so that LoongArch can use
vm_alloc_page_table() instead of open coding a rough equivalent. There are
no known reasons why LoongArch can't use a lower, less-aligned minimum.
The more-aligned value was originally selected to support huge pages for
the guest, but that only requires the address of the final allocation to be
aligned, the addresses of page table pages is irrelevant on that front.
Reviewed-by: Bibo Mao <maobibo@loongson.cn>
Tested-by: Itaru Kitayam <itaru.kitayama@fujitsu.com>
Link: https://patch.msgid.link/20260902164123.2546762-3-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
