linux-stable.git/tools/testing/selftests/kvm, branch v5.4

selftests: kvm: fix build with glibc >= 2.30

2019-11-13T14:48:00+00:00

Glibc-2.30 gained gettid() wrapper, selftests fail to compile:

lib/assert.c:58:14: error: static declaration of ‘gettid’ follows non-static declaration
   58 | static pid_t gettid(void)
      |              ^~~~~~
In file included from /usr/include/unistd.h:1170,
                 from include/test_util.h:18,
                 from lib/assert.c:10:
/usr/include/bits/unistd_ext.h:34:16: note: previous declaration of ‘gettid’ was here
   34 | extern __pid_t gettid (void) __THROW;
      |                ^~~~~~

Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: fix sync_regs_test with newer gccs

2019-10-22T11:31:18+00:00

Commit 204c91eff798a ("KVM: selftests: do not blindly clobber registers in
 guest asm") was intended to make test more gcc-proof, however, the result
is exactly the opposite: on newer gccs (e.g. 8.2.1) the test breaks with

==== Test Assertion Failure ====
  x86_64/sync_regs_test.c:168: run->s.regs.regs.rbx == 0xBAD1DEA + 1
  pid=14170 tid=14170 - Invalid argument
     1	0x00000000004015b3: main at sync_regs_test.c:166 (discriminator 6)
     2	0x00007f413fb66412: ?? ??:0
     3	0x000000000040191d: _start at ??:?
  rbx sync regs value incorrect 0x1.

Apparently, compile is still free to play games with registers even
when they have variables attached.

Re-write guest code with 'asm volatile' by embedding ucall there and
making sure rbx is preserved.

Fixes: 204c91eff798a ("KVM: selftests: do not blindly clobber registers in guest asm")
Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: vmx_dirty_log_test: skip the test when VMX is not supported

2019-10-22T11:31:17+00:00

vmx_dirty_log_test fails on AMD and this is no surprise as it is VMX
specific. Bail early when nested VMX is unsupported.

Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: consolidate VMX support checks

2019-10-22T11:31:16+00:00

vmx_* tests require VMX and three of them implement the same check. Move it
to vmx library.

Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: vmx_set_nested_state_test: don't check for VMX support twice

2019-10-22T11:31:16+00:00

vmx_set_nested_state_test() checks if VMX is supported twice: in the very
beginning (and skips the whole test if it's not) and before doing
test_vmx_nested_state(). One should be enough.

Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: synchronize .gitignore to Makefile

2019-10-22T11:31:13+00:00

Because "Untracked files:" are annoying.

Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: Fix libkvm build error

2019-10-03T10:07:58+00:00

Fix the following build error from "make TARGETS=kvm kselftest":

libkvm.a(assert.o): relocation R_X86_64_32 against `.rodata.str1.1' can not be used when making a PIE object; recompile with -fPIC

This error is seen when build is done from the main Makefile using
kselftest target. In this case KBUILD_CPPFLAGS and CC_OPTION_CFLAGS
are defined.

When build is invoked using:

"make -C tools/testing/selftests/kvm" KBUILD_CPPFLAGS and CC_OPTION_CFLAGS
aren't defined.

There is no need to pass in KBUILD_CPPFLAGS and CC_OPTION_CFLAGS for the
check to determine if --no-pie is necessary, which is the case when these
two aren't defined when "make -C tools/testing/selftests/kvm" runs.

Fix it by simplifying the no-pie-option logic. With this change, both
build variations work.

"make TARGETS=kvm kselftest"
"make -C tools/testing/selftests/kvm"

Signed-off-by: Shuah Khan 
Signed-off-by: Paolo Bonzini

KVM: selftests: x86: clarify what is reported on KVM_GET_MSRS failure

2019-09-27T16:05:25+00:00

When KVM_GET_MSRS fail the report looks like

==== Test Assertion Failure ====
  lib/x86_64/processor.c:1089: r == nmsrs
  pid=28775 tid=28775 - Argument list too long
     1	0x000000000040a55f: vcpu_save_state at processor.c:1088 (discriminator 3)
     2	0x00000000004010e3: main at state_test.c:171 (discriminator 4)
     3	0x00007fb8e69223d4: ?? ??:0
     4	0x0000000000401287: _start at ??:?
  Unexpected result from KVM_GET_MSRS, r: 36 (failed at 194)

and it's not obvious that '194' here is the failed MSR index and that
it's printed in hex. Change that.

Suggested-by: Sean Christopherson 
Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini

selftests: kvm: add test for dirty logging inside nested guests

2019-09-27T11:13:40+00:00

Check that accesses by nested guests are logged according to the
L1 physical addresses rather than L2.

Most of the patch is really adding EPT support to the testing
framework.

Signed-off-by: Paolo Bonzini

KVM: selftests: fix ucall on x86

2019-09-25T13:15:03+00:00

After commit e8bb4755eea2("KVM: selftests: Split ucall.c into architecture
specific files") selftests which use ucall on x86 started segfaulting and
apparently it's gcc to blame: it "optimizes" ucall() function throwing away
va_start/va_end part because it thinks the structure is not being used.
Previously, it couldn't do that because the there was also MMIO version and
the decision which particular implementation to use was done at runtime.

With older gccs it's possible to solve the problem by adding 'volatile'
to 'struct ucall' but at least with gcc-8.3 this trick doesn't work.

'memory' clobber seems to do the job.

Signed-off-by: Vitaly Kuznetsov 
Signed-off-by: Paolo Bonzini