| /kvm-unit-tests/s390x/snippets/asm/ |
| H A D | pv-icpt-112.S | 5 * We setup and share a prefix at 0x0 and 0x8000 which the hypervisor
20 /* Let's tell the hypervisor we're ready to start */
25 * result in a 112. It will then import the lowcore again and we
34 /* Let's tell the hypervisor we're ready to start shared testing */
60 /* Let's tell the hypervisor we're ready to start shared testing */
|
| H A D | pv-diag-308.S | 20 * A valid PGM new PSW can be a real problem since we never fall out
22 * therefore we re-use the reset_psw as this has the short PSW
|
| /kvm-unit-tests/s390x/ |
| H A D | firq.c | 46 /* Stop CPU #2. It must succeed because we have at least 3 CPUs */ in test_wait_state_delivery()
51 * We're going to perform an SCLP service call but expect the in test_wait_state_delivery()
58 /* This must not fail because we have at least 3 CPUs */ in test_wait_state_delivery()
62 * We'd have to jump trough some hoops to sense e.g., via SIGP in test_wait_state_delivery()
68 * will take some time as well and smp_cpu_setup() returns when we're in test_wait_state_delivery()
|
| H A D | panic-loop-extint.c | 20 * Since we form a loop of ext interrupts, this code should never be in ext_int_cleanup()
21 * executed. In case it is executed, something went wrong and we want to in ext_int_cleanup()
25 * interrupt will fire every time we enable external interrupts, in ext_int_cleanup()
|
| H A D | migration-skey.c | 7 * verifies that the storage keys are still as we expect.
66 * This loop will set all 7 bits which means we set fetch in set_test_pattern()
101 * accurate nor that it won't be cleared by hardware. Hence we in verify_test_pattern()
155 * we left off and can do an additional verify round after migration finished. in set_skeys_thread()
185 /* Ensure we read result and thread_iters below from memory after thread exited */ in test_skey_migration_parallel()
194 * Verification of skeys occurs on the thread. We don't know if we in test_skey_migration_parallel()
|
| H A D | macros.S | 17 * we re-load the registers and load the old PSW.
36 /* Save registers on the stack (r15), so we can have stacked interrupts. */
44 /* Store the gr15 value before we allocated the new stack */
50 * Store CR0 and load initial CR0 so AFP is active and we can
|
| H A D | pv-diags.c | 42 * Check if we can inject a PGM operand which we are always in test_diag_500()
52 * Check if we can inject a PGM specification which we are always in test_diag_500()
95 * Check if we can inject a PGM spec which we are always in test_diag_288()
|
| H A D | pv-ipl.c | 41 * The snippet asked us for the subcode and we answer by in test_diag_308()
53 * We need to perform several UV calls to emulate the subcode in test_diag_308()
66 /* Mark the CPU as stopped so we can unshare and reset */ in test_diag_308()
117 * Check if we executed the iaddr of the reset PSW, we should in test_diag_308()
|
| H A D | sclp.c | 164 * copy the current lowcore to the future new location, otherwise we in test_sccb_prefix()
171 * save the current content of absolute pages 0 and 1, so we can in test_sccb_prefix()
172 * restore them after we trash them later on in test_sccb_prefix()
180 * testing with SCCB addresses in the lowcore; since we can't in test_sccb_prefix()
181 * actually trash the lowcore (unsurprisingly, things break if we in test_sccb_prefix()
190 * the SCLP should not even touch the memory, but we will write the in test_sccb_prefix()
192 * trashing them; we will need to restore them later. in test_sccb_prefix()
358 * Test addressing exceptions. We need to test SCCB addresses between the
378 /* test more addresses until we reach 1MB after end of memory; in test_addressing()
385 /* test the remaining addresses until we reach address 2GB; in test_addressing()
[all …]
|
| H A D | mvpg-sie.c | 49 * We need to execute the diag44 which is used as a blocker in test_mvpg_pei()
50 * behind the mvpg. It makes sure we fail the tests above if in test_mvpg_pei()
54 /* Make sure we intercepted for the diag44 and nothing else */ in test_mvpg_pei()
97 /* Enable MVPG interpretation as we want to test KVM and not ourselves */ in setup_guest()
|
| /kvm-unit-tests/lib/riscv/asm/ |
| H A D | isa.h | 8 * We assume and use several extensions, such as Zicsr and Zifencei.
9 * Here we only track extensions which we don't assume and the
|
| /kvm-unit-tests/lib/s390x/ |
| H A D | interrupt.c | 144 /* If we have an error on SIE we directly move to sie_exit */ in fixup_pgm_int()
154 * was produced intentionally and we should return to the in fixup_pgm_int()
163 * We branched to the instruction that caused in fixup_pgm_int()
164 * the exception so we can use the return in fixup_pgm_int()
203 * responsible instruction. Forward the PSW so we don't loop. in fixup_pgm_int()
271 * If we fall out of SIE before loading the host registers, in print_pgm_info()
272 * then we need to do it here so we print the host registers in print_pgm_info()
283 * We copy 0 and 1 here and 2 - 15 with the memcopy below. in print_pgm_info()
293 /* Dump stack doesn't end with a \n so we add it here instead */ in print_pgm_info()
308 /* Force sclp_busy to false, otherwise we will loop forever */ in handle_pgm_int()
|
| H A D | sie.c | 29 * 0xffff will never be returned by SIE, so we can indicate a in sie_get_validity()
59 /* When a pgm int code is set, we'll never enter SIE below. */ in sie()
66 /* Reset icptcode so we don't trip over it below */ in sie()
71 * in home space all the time, we switch every time in sie() because: in sie()
87 /* leave SIE when we have an intercept or an interrupt so the test can react to it */ in sie()
118 /* We assume that cpu 0 is always part of the vm */ in sie_guest_sca_create()
135 /* For non-PV guests we re-use the host's ASCE for ease of use */ in sie_guest_create()
165 * Thus we use the next multiple of 4GB after the end of physical in sie_guest_alloc()
|
| H A D | snippet.h | 35 * addresses which means we need a lot more memory than other
80 * We want to exit on PGM exceptions so we don't need in snippet_init()
111 /* Unpack works on guest addresses so we only need off */ in snippet_pv_init()
|
| /kvm-unit-tests/s390x/snippets/c/ |
| H A D | sie-dat.c | 23 /* tell the host the page's physical address (we're running DAT off) */ in main()
30 /* indicate we've written all pages */ in main()
37 /* indicate we've written the non-allowed page (should never get here) */ in main()
|
| /kvm-unit-tests/powerpc/ |
| H A D | run | 62 # run_qemu treats that as a failure exit and returns 1, so we need
64 # The second fixup is also a FIXME, because once we add chr-testdev
65 # support for powerpc, we won't need the second fixup.
|
| /kvm-unit-tests/lib/ |
| H A D | abort.c | 11 * status, we further limit exit codes to those resulting in qemu
12 * exiting with a status < 128. We give abort() the highest (127),
|
| /kvm-unit-tests/lib/riscv/ |
| H A D | timer.c | 59 * we used ULONG_MAX, then we wouldn't stop the timer at all, in timer_stop()
61 * QEMU to do what we want. in timer_stop()
|
| H A D | io.c | 22 * having earlier printf support. We'll overwrite it with the real
23 * base address that we read from the device tree later. This is
24 * the address we expect the virtual machine manager to put in
145 * be in a0 when we halt. That gives us a final chance to see the exit
|
| /kvm-unit-tests/arm/ |
| H A D | micro-bench.c | 253 * We use a 10msec timer to test the latency of PPI, in timer_post()
254 * so we subtract the ticks of 10msec to get the in timer_post()
270 * FIXME: We need an MMIO address that we can safely read to test in mmio_read_user_prep()
272 * address (and one we could write to too), but until it does we in mmio_read_user_prep()
273 * use a virtio-mmio transport address. FIXME2: We should be getting in mmio_read_user_prep()
275 * but so far we lazily hardcode it. in mmio_read_user_prep()
|
| /kvm-unit-tests/x86/ |
| H A D | intel-iommu.c | 104 /* We are good as long as we reach here */ in vtd_test_ir()
127 /* We are good as long as we reach here */ in vtd_test_ir()
|
| H A D | kvmclock.c | 203 * forward. If we are less than that, we should not be much smaller. in pvclock_clocksource_read()
204 * We assume there is an error margin we're inside, and then the in pvclock_clocksource_read()
209 * We just need to make sure we are not seeing a backwards event. in pvclock_clocksource_read()
|
| /kvm-unit-tests/lib/linux/ |
| H A D | const.h | 7 * C code. Therefore we cannot annotate them always with
8 * 'UL' and other type specifiers unilaterally. We
|
| /kvm-unit-tests/lib/arm/ |
| H A D | io.c | 24 * having earlier printf support. We'll overwrite it with the real
25 * base address that we read from the device tree later. This is
26 * the address we expect the virtual machine manager to put in
139 * be in x0/r0 when we halt. That gives us a final chance to see the exit
|
| /kvm-unit-tests/lib/powerpc/asm/ |
| H A D | hcall.h | 39 * hcall_have_broken_sc1 checks if we're on a host with a broken sc1.
40 * Returns 0 if we're not.
|