Message ID | 1460115329-21611-1-git-send-email-thuth@redhat.com |
---|---|
State | Superseded |
Headers | show |
On Fri, Apr 08, 2016 at 01:35:29PM +0200, Thomas Huth wrote: > This test can be used to check whether the SPR (special purpose > registers) of the PowerPC CPU are migrated right. It first fills > the various SPRs with some non-zero value, then reads the values > back into a first array, then waits for a key (with the '-w' option) > so that it is possible to migrate the VM, and finally reads the > values from the SPRs back into another array and then compares it > with the initial values. > Currently the test only supports the SPRs from the PowerISA v2.07 > specification (i.e. POWER8 CPUs), but other versions should be > pretty easy to add later. > > Signed-off-by: Thomas Huth <thuth@redhat.com> > --- > powerpc/Makefile.common | 5 +- > powerpc/sprs.c | 230 ++++++++++++++++++++++++++++++++++++++++++++++++ > powerpc/unittests.cfg | 4 + > 3 files changed, 238 insertions(+), 1 deletion(-) > create mode 100644 powerpc/sprs.c > > diff --git a/powerpc/Makefile.common b/powerpc/Makefile.common > index 4449aec..43b2e49 100644 > --- a/powerpc/Makefile.common > +++ b/powerpc/Makefile.common > @@ -8,7 +8,8 @@ tests-common = \ > $(TEST_DIR)/selftest.elf \ > $(TEST_DIR)/spapr_hcall.elf \ > $(TEST_DIR)/rtas.elf \ > - $(TEST_DIR)/emulator.elf > + $(TEST_DIR)/emulator.elf \ > + $(TEST_DIR)/sprs.elf > > all: $(TEST_DIR)/boot_rom.bin test_cases > > @@ -77,3 +78,5 @@ $(TEST_DIR)/spapr_hcall.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/spapr_hcall.o > $(TEST_DIR)/rtas.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/rtas.o > > $(TEST_DIR)/emulator.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/emulator.o > + > +$(TEST_DIR)/sprs.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/sprs.o > diff --git a/powerpc/sprs.c b/powerpc/sprs.c > new file mode 100644 > index 0000000..e99501f > --- /dev/null > +++ b/powerpc/sprs.c > @@ -0,0 +1,230 @@ > +/* > + * Test SPRs > + * > + * Copyright 2016 Thomas Huth, Red Hat Inc. > + * > + * This work is licensed under the terms of the GNU LGPL, version 2. > + * > + * The basic idea of this test is to check whether the contents of the Special > + * Purpose Registers (SPRs) are preserved correctly during migration. So we > + * fill in the SPRs with a well-known value, read the values back (since not > + * all bits might be retained in the SPRs), then wait for a key (if the '-w' > + * option has been specified) so that the user has a chance to migrate the VM, > + * and after a key has been pressed, we read back the values again and compare > + * them with the values before the migration, reporting errors for mismatches. > + * Note that we do not test all SPRs since some of the registers change their > + * content automatically, and some are only accessible with hypervisor privi- > + * leges, so we have to omit those registers. > + */ > +#include <libcflat.h> > +#include <util.h> > +#include <alloc.h> > +#include <asm/hcall.h> > + > +#define mfspr(nr) ({ \ > + uint64_t ret; \ > + asm volatile("mfspr %0,%1" : "=r"(ret) : "i"(nr)); \ > + ret; \ > +}) > + > +#define mtspr(nr, val) \ > + asm volatile("mtspr %0,%1" : : "i"(nr), "r"(val)); You may want these defines in processor.h to share with other tests. > + > +uint64_t before[1024], after[1024]; > + > +static int h_get_term_char(uint64_t termno) > +{ > + register uint64_t r3 asm("r3") = 0x54; /* H_GET_TERM_CHAR */ > + register uint64_t r4 asm("r4") = termno; > + register uint64_t r5 asm("r5"); > + > + asm volatile (" sc 1 " : "+r"(r3), "+r"(r4), "=r"(r5) > + : "r"(r3), "r"(r4)); > + > + return r3 == H_SUCCESS && r4 > 0 ? r5 >> 48 : 0; > +} You may want this in hcall.h to share with other tests. > + > +/* Common SPRs for all PowerPC CPUs */ > +static void set_sprs_common(uint64_t val) > +{ > + mtspr(1, val); /* XER */ > + mtspr(9, val); /* CTR */ > + mtspr(273, val); /* SPRG1 */ > + mtspr(274, val); /* SPRG2 */ > + mtspr(275, val); /* SPRG3 */ > +} > + > +/* SPRs from PowerISA 2.07 Book III-S */ > +static void set_sprs_book3s_207(uint64_t val) > +{ > + mtspr(3, val); /* DSCR */ > + mtspr(13, val); /* AMR */ > + mtspr(17, val); /* DSCR */ > + mtspr(18, val); /* DSISR */ > + mtspr(19, val); /* DAR */ > + mtspr(29, val); /* AMR */ > + mtspr(61, val); /* IAMR */ > + mtspr(152, val); /* CTRL */ > + mtspr(153, val); /* FSCR */ > + mtspr(157, val); /* UAMOR */ > + mtspr(159, val); /* PSPB */ > + mtspr(256, val); /* VRSAVE */ > + mtspr(272, val); /* SPRG0 */ > + mtspr(512, val); /* SPEFSCR */ > + mtspr(769, val); /* MMCR2 */ > + mtspr(770, val); /* MMCRA */ > + mtspr(771, val); /* PMC1 */ > + mtspr(772, val); /* PMC2 */ > + mtspr(773, val); /* PMC3 */ > + mtspr(774, val); /* PMC4 */ > + mtspr(775, val); /* PMC5 */ > + mtspr(776, val); /* PMC6 */ > + mtspr(779, val); /* MMCR0 */ > + mtspr(784, val); /* SIER */ > + mtspr(785, val); /* MMCR2 */ > + mtspr(786, val); /* MMCRA */ > + mtspr(787, val); /* PMC1 */ > + mtspr(788, val); /* PMC2 */ > + mtspr(789, val); /* PMC3 */ > + mtspr(790, val); /* PMC4 */ > + mtspr(791, val); /* PMC5 */ > + mtspr(792, val); /* PMC6 */ > + mtspr(795, val); /* MMCR0 */ > + mtspr(796, val); /* SIAR */ > + mtspr(798, val); /* SDAR */ > + mtspr(800, val); /* BESCRS */ > + mtspr(801, val); /* BESCCRSU */ > + mtspr(802, val); /* BESCRR */ > + mtspr(803, val); /* BESCRRU */ > + mtspr(804, val); /* EBBHR */ > + mtspr(805, val); /* EBBRR */ > + mtspr(806, val); /* BESCR */ > + mtspr(815, val); /* TAR */ > +} > + > +static void set_sprs(uint64_t val) > +{ > + uint32_t pvr = mfspr(287); /* Processor Version Register */ > + > + set_sprs_common(val); > + > + switch (pvr >> 16) { > + case 0x4b: /* POWER8E */ > + case 0x4c: /* POWER8NVL */ > + case 0x4d: /* POWER8 */ > + set_sprs_book3s_207(val); > + break; > + default: > + puts("Warning: Unknown processor version!\n"); > + } > +} > + > +static void get_sprs_common(uint64_t *v) > +{ > + v[1] = mfspr(1); /* XER */ > + v[9] = mfspr(9); /* CTR */ > + v[273] = mfspr(273); /* SPRG1 */ > + v[274] = mfspr(274); /* SPRG2 */ > + v[275] = mfspr(275); /* SPRG3 */ > +} > + > +static void get_sprs_book3s_207(uint64_t *v) > +{ > + v[3] = mfspr(3); /* DSCR */ > + v[13] = mfspr(13); /* AMR */ > + v[17] = mfspr(17); /* DSCR */ > + v[18] = mfspr(18); /* DSISR */ > + v[19] = mfspr(19); /* DAR */ > + v[29] = mfspr(29); /* AMR */ > + v[61] = mfspr(61); /* IAMR */ > + v[136] = mfspr(136); /* CTRL */ > + v[153] = mfspr(153); /* FSCR */ > + v[157] = mfspr(157); /* UAMOR */ > + v[159] = mfspr(159); /* PSPB */ > + v[256] = mfspr(256); /* VRSAVE */ > + v[259] = mfspr(259); /* SPRG3 (read only) */ > + v[272] = mfspr(272); /* SPRG0 */ > + v[512] = mfspr(512); /* SPEFSCR */ > + v[769] = mfspr(769); /* MMCR2 */ > + v[770] = mfspr(770); /* MMCRA */ > + v[771] = mfspr(771); /* PMC1 */ > + v[772] = mfspr(772); /* PMC2 */ > + v[773] = mfspr(773); /* PMC3 */ > + v[774] = mfspr(774); /* PMC4 */ > + v[775] = mfspr(775); /* PMC5 */ > + v[776] = mfspr(776); /* PMC6 */ > + v[779] = mfspr(779); /* MMCR0 */ > + v[780] = mfspr(780); /* SIAR (read only) */ > + v[781] = mfspr(781); /* SDAR (read only) */ > + v[782] = mfspr(782); /* MMCR1 (read only) */ > + v[784] = mfspr(784); /* SIER */ > + v[785] = mfspr(785); /* MMCR2 */ > + v[786] = mfspr(786); /* MMCRA */ > + v[787] = mfspr(787); /* PMC1 */ > + v[788] = mfspr(788); /* PMC2 */ > + v[789] = mfspr(789); /* PMC3 */ > + v[790] = mfspr(790); /* PMC4 */ > + v[791] = mfspr(791); /* PMC5 */ > + v[792] = mfspr(792); /* PMC6 */ > + v[795] = mfspr(795); /* MMCR0 */ > + v[796] = mfspr(796); /* SIAR */ > + v[798] = mfspr(798); /* SDAR */ > + v[800] = mfspr(800); /* BESCRS */ > + v[801] = mfspr(801); /* BESCCRSU */ > + v[802] = mfspr(802); /* BESCRR */ > + v[803] = mfspr(803); /* BESCRRU */ > + v[804] = mfspr(804); /* EBBHR */ > + v[805] = mfspr(805); /* EBBRR */ > + v[806] = mfspr(806); /* BESCR */ > + v[815] = mfspr(815); /* TAR */ > +} > + > +static void get_sprs(uint64_t *v) > +{ > + uint32_t pvr = mfspr(287); /* Processor Version Register */ > + > + get_sprs_common(v); > + > + switch (pvr >> 16) { > + case 0x4b: /* POWER8E */ > + case 0x4c: /* POWER8NVL */ > + case 0x4d: /* POWER8 */ > + get_sprs_book3s_207(v); > + break; > + } > +} > + > +int main(int argc, char **argv) > +{ > + int i; > + > + if (argc > 1) > + report_abort("Warning: Unsupported arguments!"); > + > + puts("Settings SPRs...\n"); > + set_sprs(0xcafefacec0debabeULL); > + > + memset(before, 0, sizeof(before)); > + memset(after, 0, sizeof(after)); before and after shouldn't need the memsets, as they're in the bss. > + > + get_sprs(before); > + > + if (argc > 0 && !strcmp(argv[0], "-w")) { > + puts("Now migrate the VM, then press a key...\n"); > + while (h_get_term_char(0) == 0) > + ; > + } else { > + puts("Parameter '-w' not specified - not waiting for a key.\n"); > + } > + > + get_sprs(after); > + > + puts("Checking SPRs...\n"); > + for (i = 0; i < 1024; i++) { > + if (before[i] != 0 || after[i] != 0) > + report("SPR %d:\t0x%016lx <==> 0x%016lx", > + before[i] == after[i], i, before[i], after[i]); > + } > + > + return report_summary(); > +} > diff --git a/powerpc/unittests.cfg b/powerpc/unittests.cfg > index ed4fdbe..5563cbe 100644 > --- a/powerpc/unittests.cfg > +++ b/powerpc/unittests.cfg > @@ -50,3 +50,7 @@ groups = rtas > > [emulator] > file = emulator.elf > + > +[sprs] > +file = sprs.elf > +#extra_params = -append '-w' > -- > 1.8.3.1 > This is an interesting use of kvm-unit-tests. Ideally we'd keep avoid the key press, which could complicate an automated test. Above you say some SPRs change their content automatically. Are any of those changes something that can be predicted in the case of a migration? I.e. could we monitor them to see if a migration occurred? Thanks, drew -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08.04.2016 14:14, Andrew Jones wrote: > On Fri, Apr 08, 2016 at 01:35:29PM +0200, Thomas Huth wrote: >> This test can be used to check whether the SPR (special purpose >> registers) of the PowerPC CPU are migrated right. It first fills >> the various SPRs with some non-zero value, then reads the values >> back into a first array, then waits for a key (with the '-w' option) >> so that it is possible to migrate the VM, and finally reads the >> values from the SPRs back into another array and then compares it >> with the initial values. >> Currently the test only supports the SPRs from the PowerISA v2.07 >> specification (i.e. POWER8 CPUs), but other versions should be >> pretty easy to add later. >> >> Signed-off-by: Thomas Huth <thuth@redhat.com> >> --- >> powerpc/Makefile.common | 5 +- >> powerpc/sprs.c | 230 ++++++++++++++++++++++++++++++++++++++++++++++++ >> powerpc/unittests.cfg | 4 + >> 3 files changed, 238 insertions(+), 1 deletion(-) >> create mode 100644 powerpc/sprs.c [...] >> diff --git a/powerpc/sprs.c b/powerpc/sprs.c >> new file mode 100644 >> index 0000000..e99501f >> --- /dev/null >> +++ b/powerpc/sprs.c >> @@ -0,0 +1,230 @@ >> +/* >> + * Test SPRs >> + * >> + * Copyright 2016 Thomas Huth, Red Hat Inc. >> + * >> + * This work is licensed under the terms of the GNU LGPL, version 2. >> + * >> + * The basic idea of this test is to check whether the contents of the Special >> + * Purpose Registers (SPRs) are preserved correctly during migration. So we >> + * fill in the SPRs with a well-known value, read the values back (since not >> + * all bits might be retained in the SPRs), then wait for a key (if the '-w' >> + * option has been specified) so that the user has a chance to migrate the VM, >> + * and after a key has been pressed, we read back the values again and compare >> + * them with the values before the migration, reporting errors for mismatches. >> + * Note that we do not test all SPRs since some of the registers change their >> + * content automatically, and some are only accessible with hypervisor privi- >> + * leges, so we have to omit those registers. >> + */ >> +#include <libcflat.h> >> +#include <util.h> >> +#include <alloc.h> >> +#include <asm/hcall.h> >> + >> +#define mfspr(nr) ({ \ >> + uint64_t ret; \ >> + asm volatile("mfspr %0,%1" : "=r"(ret) : "i"(nr)); \ >> + ret; \ >> +}) >> + >> +#define mtspr(nr, val) \ >> + asm volatile("mtspr %0,%1" : : "i"(nr), "r"(val)); > > You may want these defines in processor.h to share with other tests. Ok, mfspr() is already used in spapr_hcall.c, too, so this is a good idea. >> + >> +uint64_t before[1024], after[1024]; >> + >> +static int h_get_term_char(uint64_t termno) >> +{ >> + register uint64_t r3 asm("r3") = 0x54; /* H_GET_TERM_CHAR */ >> + register uint64_t r4 asm("r4") = termno; >> + register uint64_t r5 asm("r5"); >> + >> + asm volatile (" sc 1 " : "+r"(r3), "+r"(r4), "=r"(r5) >> + : "r"(r3), "r"(r4)); >> + >> + return r3 == H_SUCCESS && r4 > 0 ? r5 >> 48 : 0; >> +} > > You may want this in hcall.h to share with other tests. Not sure whether it's of any use to another test, since they normally run automatically, without user interaction. It's just this test that needs to wait a little bit for the end of migration ... so I'd rather keep it here until we have another test that needs it. [...] >> +int main(int argc, char **argv) >> +{ >> + int i; >> + >> + if (argc > 1) >> + report_abort("Warning: Unsupported arguments!"); >> + >> + puts("Settings SPRs...\n"); >> + set_sprs(0xcafefacec0debabeULL); >> + >> + memset(before, 0, sizeof(before)); >> + memset(after, 0, sizeof(after)); > > before and after shouldn't need the memsets, as they're in the bss. Ok. >> + >> + get_sprs(before); >> + >> + if (argc > 0 && !strcmp(argv[0], "-w")) { >> + puts("Now migrate the VM, then press a key...\n"); >> + while (h_get_term_char(0) == 0) >> + ; >> + } else { >> + puts("Parameter '-w' not specified - not waiting for a key.\n"); >> + } >> + >> + get_sprs(after); >> + >> + puts("Checking SPRs...\n"); >> + for (i = 0; i < 1024; i++) { >> + if (before[i] != 0 || after[i] != 0) >> + report("SPR %d:\t0x%016lx <==> 0x%016lx", >> + before[i] == after[i], i, before[i], after[i]); >> + } >> + >> + return report_summary(); >> +} [...] > This is an interesting use of kvm-unit-tests. Ideally we'd keep avoid > the key press, which could complicate an automated test. Above you say > some SPRs change their content automatically. Are any of those changes > something that can be predicted in the case of a migration? I.e. could > we monitor them to see if a migration occurred? I don't think that it is very easy to detect the end of the migration like this... the registers I was talking about are things like the decrementer or timebase registers. If we really want to automate this, we could maybe use something like the machine check interrupt instead (which can be triggered with the "nmi" HMP command in QEMU, I think) to let the guest know about the end of the migration. But all that would also require quite some additional logic in the runner script, I think. Not sure whether it's worth the effort just for this one small test, which can also be run manually from time to time... But if you think we should really go this way, let me know, then I can have a try... Thomas -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Apr 08, 2016 at 04:08:41PM +0200, Thomas Huth wrote: > > This is an interesting use of kvm-unit-tests. Ideally we'd keep avoid > > the key press, which could complicate an automated test. Above you say > > some SPRs change their content automatically. Are any of those changes > > something that can be predicted in the case of a migration? I.e. could > > we monitor them to see if a migration occurred? > > I don't think that it is very easy to detect the end of the migration > like this... the registers I was talking about are things like the > decrementer or timebase registers. > > If we really want to automate this, we could maybe use something like > the machine check interrupt instead (which can be triggered with the > "nmi" HMP command in QEMU, I think) to let the guest know about the end > of the migration. But all that would also require quite some additional > logic in the runner script, I think. Not sure whether it's worth the > effort just for this one small test, which can also be run manually from > time to time... But if you think we should really go this way, let me > know, then I can have a try... I'm OK with time-to-time manual tests. I wonder if we should add a run_tests.sh cmdline switch that enables all manual tests somehow. This would allow us to run the complete collection of manual tests easily, not forgetting any. drew -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, 8 Apr 2016 13:35:29 +0200 Thomas Huth <thuth@redhat.com> wrote: > This test can be used to check whether the SPR (special purpose > registers) of the PowerPC CPU are migrated right. It first fills > the various SPRs with some non-zero value, then reads the values > back into a first array, then waits for a key (with the '-w' option) > so that it is possible to migrate the VM, and finally reads the > values from the SPRs back into another array and then compares it > with the initial values. > Currently the test only supports the SPRs from the PowerISA v2.07 > specification (i.e. POWER8 CPUs), but other versions should be > pretty easy to add later. > > Signed-off-by: Thomas Huth <thuth@redhat.com> So, the main concern I have about this is that while writing an arbitrary value is ok for quite a few SPRs, there's a significant number where that's not safe: either because it could actually cause some sort of exception interrupting the test, or because the value will get masked or otherwise modifier on write. I think at the very least we should separate out the code dealing with the safe-to-write-anything SPRs from the others as a form of internal documentation. Then we probably want some sort of table of safe non-default values for the other SPRs. [snip] > +/* Common SPRs for all PowerPC CPUs */ > +static void set_sprs_common(uint64_t val) > +{ > + mtspr(1, val); /* XER */ > + mtspr(9, val); /* CTR */ > + mtspr(273, val); /* SPRG1 */ > + mtspr(274, val); /* SPRG2 */ > + mtspr(275, val); /* SPRG3 */ > +} > + > +/* SPRs from PowerISA 2.07 Book III-S */ > +static void set_sprs_book3s_207(uint64_t val) > +{ > + mtspr(3, val); /* DSCR */ > + mtspr(13, val); /* AMR */ > + mtspr(17, val); /* DSCR */ > + mtspr(18, val); /* DSISR */ > + mtspr(19, val); /* DAR */ > + mtspr(29, val); /* AMR */ AMR seems to be listed twice.. > + mtspr(61, val); /* IAMR */ > + mtspr(152, val); /* CTRL */ > + mtspr(153, val); /* FSCR */ > + mtspr(157, val); /* UAMOR */ > + mtspr(159, val); /* PSPB */ > + mtspr(256, val); /* VRSAVE */ > + mtspr(272, val); /* SPRG0 */ > + mtspr(512, val); /* SPEFSCR */ > + mtspr(769, val); /* MMCR2 */ > + mtspr(770, val); /* MMCRA */ > + mtspr(771, val); /* PMC1 */ > + mtspr(772, val); /* PMC2 */ > + mtspr(773, val); /* PMC3 */ > + mtspr(774, val); /* PMC4 */ > + mtspr(775, val); /* PMC5 */ > + mtspr(776, val); /* PMC6 */ > + mtspr(779, val); /* MMCR0 */ > + mtspr(784, val); /* SIER */ > + mtspr(785, val); /* MMCR2 */ > + mtspr(786, val); /* MMCRA */ > + mtspr(787, val); /* PMC1 */ > + mtspr(788, val); /* PMC2 */ > + mtspr(789, val); /* PMC3 */ > + mtspr(790, val); /* PMC4 */ > + mtspr(791, val); /* PMC5 */ > + mtspr(792, val); /* PMC6 */ > + mtspr(795, val); /* MMCR0 */ > + mtspr(796, val); /* SIAR */ > + mtspr(798, val); /* SDAR */ > + mtspr(800, val); /* BESCRS */ > + mtspr(801, val); /* BESCCRSU */ > + mtspr(802, val); /* BESCRR */ > + mtspr(803, val); /* BESCRRU */ > + mtspr(804, val); /* EBBHR */ > + mtspr(805, val); /* EBBRR */ > + mtspr(806, val); /* BESCR */ > + mtspr(815, val); /* TAR */ > +} At a glance SPRs above where writing an arbitrary value might not be safe include AMR, IAMR, UAMOR, and MMCR*.
On 11.04.2016 03:55, David Gibson wrote: > On Fri, 8 Apr 2016 13:35:29 +0200 > Thomas Huth <thuth@redhat.com> wrote: > >> This test can be used to check whether the SPR (special purpose >> registers) of the PowerPC CPU are migrated right. It first fills >> the various SPRs with some non-zero value, then reads the values >> back into a first array, then waits for a key (with the '-w' option) >> so that it is possible to migrate the VM, and finally reads the >> values from the SPRs back into another array and then compares it >> with the initial values. >> Currently the test only supports the SPRs from the PowerISA v2.07 >> specification (i.e. POWER8 CPUs), but other versions should be >> pretty easy to add later. >> >> Signed-off-by: Thomas Huth <thuth@redhat.com> > > So, the main concern I have about this is that while writing an > arbitrary value is ok for quite a few SPRs, there's a significant > number where that's not safe: either because it could actually cause > some sort of exception interrupting the test, or because the value will > get masked or otherwise modifier on write. If a write to an SPR caused any exception or hang of the guest, I did not include it in the list of SPRs to be tested here. Sorry, I should have mentioned that somewhere in the commit message or comment of the sprs.c file below. Anyway, working out a detailed list of values that should be written into each of the SPRs might get quite difficult/cumbersome since some of these SPRs are not that well documented in the PowerISA. So I'd like to keep the generic approach here instead. Anyway, if a write causes trouble, it's not a big issue to take that SPR out of the list again. And since the kvm-unit-tests are also a very isolated test (i.e. we're not running a Linux kernel here which might expect certain values in certain SPRs), it's IMHO also ok to mess up the state in the SPRs since the VM will get destroyed after the test anyway. About your second concern, that the values will get masked or modified: Yes, you're right, of course! But that's not a problem here since I read back the values from the SPRs before the migration - I do not use the original value that I wrote into the SPRs for comparison! > [snip] >> +/* Common SPRs for all PowerPC CPUs */ >> +static void set_sprs_common(uint64_t val) >> +{ >> + mtspr(1, val); /* XER */ >> + mtspr(9, val); /* CTR */ >> + mtspr(273, val); /* SPRG1 */ >> + mtspr(274, val); /* SPRG2 */ >> + mtspr(275, val); /* SPRG3 */ >> +} >> + >> +/* SPRs from PowerISA 2.07 Book III-S */ >> +static void set_sprs_book3s_207(uint64_t val) >> +{ >> + mtspr(3, val); /* DSCR */ >> + mtspr(13, val); /* AMR */ >> + mtspr(17, val); /* DSCR */ >> + mtspr(18, val); /* DSISR */ >> + mtspr(19, val); /* DAR */ >> + mtspr(29, val); /* AMR */ > > AMR seems to be listed twice.. That's because AMR is available via both SPR numbers - one time for userspace mode (which can be disabled by the kernel if desired), and one time for privileged mode. > At a glance SPRs above where writing an arbitrary value might not be > safe include AMR, IAMR, UAMOR, and MMCR*. Well, you should not think of this test as a nice, behaving guest kernel that only puts valid values into safe SPRs ... rather think of it as a stress test for the host with a bad, misbehaving guest kernel. So think what's the worst thing that could happen? Host crash? Right, this test already helped to find one of those bugs: http://www.spinics.net/lists/kvm/msg128989.html ... but then we've of course got to fix the host kernel (or QEMU), not this test. The other bad thing that could of course happen is a guest crash - but then, as mentioned above, we can simply adjust the list of SPRs that we test as soon as we see such a crash. With the current list, I do not get any guest crashes - at least not with kvm-hv. ... but kvm-pr and tcg are two other candidates that likely need some fixing for this test. Thomas
On Mon, 11 Apr 2016 11:23:46 +0200 Thomas Huth <thuth@redhat.com> wrote: > On 11.04.2016 03:55, David Gibson wrote: > > On Fri, 8 Apr 2016 13:35:29 +0200 > > Thomas Huth <thuth@redhat.com> wrote: > > > >> This test can be used to check whether the SPR (special purpose > >> registers) of the PowerPC CPU are migrated right. It first fills > >> the various SPRs with some non-zero value, then reads the values > >> back into a first array, then waits for a key (with the '-w' option) > >> so that it is possible to migrate the VM, and finally reads the > >> values from the SPRs back into another array and then compares it > >> with the initial values. > >> Currently the test only supports the SPRs from the PowerISA v2.07 > >> specification (i.e. POWER8 CPUs), but other versions should be > >> pretty easy to add later. > >> > >> Signed-off-by: Thomas Huth <thuth@redhat.com> > > > > So, the main concern I have about this is that while writing an > > arbitrary value is ok for quite a few SPRs, there's a significant > > number where that's not safe: either because it could actually cause > > some sort of exception interrupting the test, or because the value will > > get masked or otherwise modifier on write. > > If a write to an SPR caused any exception or hang of the guest, I did > not include it in the list of SPRs to be tested here. Sorry, I should > have mentioned that somewhere in the commit message or comment of the > sprs.c file below. Hmm.. right, but that just means the particular value you're using doesn't break - other values could cause guest crashes. But as it appears now, it looks like the value written is arbitrary. > Anyway, working out a detailed list of values that should be written > into each of the SPRs might get quite difficult/cumbersome since some of > these SPRs are not that well documented in the PowerISA. So I'd like to > keep the generic approach here instead. Anyway, if a write causes > trouble, it's not a big issue to take that SPR out of the list again. > And since the kvm-unit-tests are also a very isolated test (i.e. we're > not running a Linux kernel here which might expect certain values in > certain SPRs), it's IMHO also ok to mess up the state in the SPRs since > the VM will get destroyed after the test anyway. Hm, yeah, I guess so. I'm just concerned about crashing the guest before the test can even finish. > About your second concern, that the values will get masked or modified: > Yes, you're right, of course! But that's not a problem here since I read > back the values from the SPRs before the migration - I do not use the > original value that I wrote into the SPRs for comparison! Ah, I missed that, sorry. Clever approach. > > [snip] > >> +/* Common SPRs for all PowerPC CPUs */ > >> +static void set_sprs_common(uint64_t val) > >> +{ > >> + mtspr(1, val); /* XER */ > >> + mtspr(9, val); /* CTR */ > >> + mtspr(273, val); /* SPRG1 */ > >> + mtspr(274, val); /* SPRG2 */ > >> + mtspr(275, val); /* SPRG3 */ > >> +} > >> + > >> +/* SPRs from PowerISA 2.07 Book III-S */ > >> +static void set_sprs_book3s_207(uint64_t val) > >> +{ > >> + mtspr(3, val); /* DSCR */ > >> + mtspr(13, val); /* AMR */ > >> + mtspr(17, val); /* DSCR */ > >> + mtspr(18, val); /* DSISR */ > >> + mtspr(19, val); /* DAR */ > >> + mtspr(29, val); /* AMR */ > > > > AMR seems to be listed twice.. > > That's because AMR is available via both SPR numbers - one time for > userspace mode (which can be disabled by the kernel if desired), and one > time for privileged mode. > > > At a glance SPRs above where writing an arbitrary value might not be > > safe include AMR, IAMR, UAMOR, and MMCR*. > > Well, you should not think of this test as a nice, behaving guest kernel > that only puts valid values into safe SPRs ... rather think of it as a > stress test for the host with a bad, misbehaving guest kernel. > > So think what's the worst thing that could happen? Host crash? Right, > this test already helped to find one of those bugs: > > http://www.spinics.net/lists/kvm/msg128989.html > > ... but then we've of course got to fix the host kernel (or QEMU), not > this test. > > The other bad thing that could of course happen is a guest crash - but > then, as mentioned above, we can simply adjust the list of SPRs that we > test as soon as we see such a crash. With the current list, I do not get > any guest crashes - at least not with kvm-hv. ... but kvm-pr and tcg are > two other candidates that likely need some fixing for this test. I'm still a little concerned about crashing the guest before the test can complete. But as you say generating sane values for all SPRs is quite a lot of work, and in the meantime the SPRs which could crash the guest (or the host) are the ones that are especially important to test migration of. So, you've convinced me that this is a good idea on balance. Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
diff --git a/powerpc/Makefile.common b/powerpc/Makefile.common index 4449aec..43b2e49 100644 --- a/powerpc/Makefile.common +++ b/powerpc/Makefile.common @@ -8,7 +8,8 @@ tests-common = \ $(TEST_DIR)/selftest.elf \ $(TEST_DIR)/spapr_hcall.elf \ $(TEST_DIR)/rtas.elf \ - $(TEST_DIR)/emulator.elf + $(TEST_DIR)/emulator.elf \ + $(TEST_DIR)/sprs.elf all: $(TEST_DIR)/boot_rom.bin test_cases @@ -77,3 +78,5 @@ $(TEST_DIR)/spapr_hcall.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/spapr_hcall.o $(TEST_DIR)/rtas.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/rtas.o $(TEST_DIR)/emulator.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/emulator.o + +$(TEST_DIR)/sprs.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/sprs.o diff --git a/powerpc/sprs.c b/powerpc/sprs.c new file mode 100644 index 0000000..e99501f --- /dev/null +++ b/powerpc/sprs.c @@ -0,0 +1,230 @@ +/* + * Test SPRs + * + * Copyright 2016 Thomas Huth, Red Hat Inc. + * + * This work is licensed under the terms of the GNU LGPL, version 2. + * + * The basic idea of this test is to check whether the contents of the Special + * Purpose Registers (SPRs) are preserved correctly during migration. So we + * fill in the SPRs with a well-known value, read the values back (since not + * all bits might be retained in the SPRs), then wait for a key (if the '-w' + * option has been specified) so that the user has a chance to migrate the VM, + * and after a key has been pressed, we read back the values again and compare + * them with the values before the migration, reporting errors for mismatches. + * Note that we do not test all SPRs since some of the registers change their + * content automatically, and some are only accessible with hypervisor privi- + * leges, so we have to omit those registers. + */ +#include <libcflat.h> +#include <util.h> +#include <alloc.h> +#include <asm/hcall.h> + +#define mfspr(nr) ({ \ + uint64_t ret; \ + asm volatile("mfspr %0,%1" : "=r"(ret) : "i"(nr)); \ + ret; \ +}) + +#define mtspr(nr, val) \ + asm volatile("mtspr %0,%1" : : "i"(nr), "r"(val)); + +uint64_t before[1024], after[1024]; + +static int h_get_term_char(uint64_t termno) +{ + register uint64_t r3 asm("r3") = 0x54; /* H_GET_TERM_CHAR */ + register uint64_t r4 asm("r4") = termno; + register uint64_t r5 asm("r5"); + + asm volatile (" sc 1 " : "+r"(r3), "+r"(r4), "=r"(r5) + : "r"(r3), "r"(r4)); + + return r3 == H_SUCCESS && r4 > 0 ? r5 >> 48 : 0; +} + +/* Common SPRs for all PowerPC CPUs */ +static void set_sprs_common(uint64_t val) +{ + mtspr(1, val); /* XER */ + mtspr(9, val); /* CTR */ + mtspr(273, val); /* SPRG1 */ + mtspr(274, val); /* SPRG2 */ + mtspr(275, val); /* SPRG3 */ +} + +/* SPRs from PowerISA 2.07 Book III-S */ +static void set_sprs_book3s_207(uint64_t val) +{ + mtspr(3, val); /* DSCR */ + mtspr(13, val); /* AMR */ + mtspr(17, val); /* DSCR */ + mtspr(18, val); /* DSISR */ + mtspr(19, val); /* DAR */ + mtspr(29, val); /* AMR */ + mtspr(61, val); /* IAMR */ + mtspr(152, val); /* CTRL */ + mtspr(153, val); /* FSCR */ + mtspr(157, val); /* UAMOR */ + mtspr(159, val); /* PSPB */ + mtspr(256, val); /* VRSAVE */ + mtspr(272, val); /* SPRG0 */ + mtspr(512, val); /* SPEFSCR */ + mtspr(769, val); /* MMCR2 */ + mtspr(770, val); /* MMCRA */ + mtspr(771, val); /* PMC1 */ + mtspr(772, val); /* PMC2 */ + mtspr(773, val); /* PMC3 */ + mtspr(774, val); /* PMC4 */ + mtspr(775, val); /* PMC5 */ + mtspr(776, val); /* PMC6 */ + mtspr(779, val); /* MMCR0 */ + mtspr(784, val); /* SIER */ + mtspr(785, val); /* MMCR2 */ + mtspr(786, val); /* MMCRA */ + mtspr(787, val); /* PMC1 */ + mtspr(788, val); /* PMC2 */ + mtspr(789, val); /* PMC3 */ + mtspr(790, val); /* PMC4 */ + mtspr(791, val); /* PMC5 */ + mtspr(792, val); /* PMC6 */ + mtspr(795, val); /* MMCR0 */ + mtspr(796, val); /* SIAR */ + mtspr(798, val); /* SDAR */ + mtspr(800, val); /* BESCRS */ + mtspr(801, val); /* BESCCRSU */ + mtspr(802, val); /* BESCRR */ + mtspr(803, val); /* BESCRRU */ + mtspr(804, val); /* EBBHR */ + mtspr(805, val); /* EBBRR */ + mtspr(806, val); /* BESCR */ + mtspr(815, val); /* TAR */ +} + +static void set_sprs(uint64_t val) +{ + uint32_t pvr = mfspr(287); /* Processor Version Register */ + + set_sprs_common(val); + + switch (pvr >> 16) { + case 0x4b: /* POWER8E */ + case 0x4c: /* POWER8NVL */ + case 0x4d: /* POWER8 */ + set_sprs_book3s_207(val); + break; + default: + puts("Warning: Unknown processor version!\n"); + } +} + +static void get_sprs_common(uint64_t *v) +{ + v[1] = mfspr(1); /* XER */ + v[9] = mfspr(9); /* CTR */ + v[273] = mfspr(273); /* SPRG1 */ + v[274] = mfspr(274); /* SPRG2 */ + v[275] = mfspr(275); /* SPRG3 */ +} + +static void get_sprs_book3s_207(uint64_t *v) +{ + v[3] = mfspr(3); /* DSCR */ + v[13] = mfspr(13); /* AMR */ + v[17] = mfspr(17); /* DSCR */ + v[18] = mfspr(18); /* DSISR */ + v[19] = mfspr(19); /* DAR */ + v[29] = mfspr(29); /* AMR */ + v[61] = mfspr(61); /* IAMR */ + v[136] = mfspr(136); /* CTRL */ + v[153] = mfspr(153); /* FSCR */ + v[157] = mfspr(157); /* UAMOR */ + v[159] = mfspr(159); /* PSPB */ + v[256] = mfspr(256); /* VRSAVE */ + v[259] = mfspr(259); /* SPRG3 (read only) */ + v[272] = mfspr(272); /* SPRG0 */ + v[512] = mfspr(512); /* SPEFSCR */ + v[769] = mfspr(769); /* MMCR2 */ + v[770] = mfspr(770); /* MMCRA */ + v[771] = mfspr(771); /* PMC1 */ + v[772] = mfspr(772); /* PMC2 */ + v[773] = mfspr(773); /* PMC3 */ + v[774] = mfspr(774); /* PMC4 */ + v[775] = mfspr(775); /* PMC5 */ + v[776] = mfspr(776); /* PMC6 */ + v[779] = mfspr(779); /* MMCR0 */ + v[780] = mfspr(780); /* SIAR (read only) */ + v[781] = mfspr(781); /* SDAR (read only) */ + v[782] = mfspr(782); /* MMCR1 (read only) */ + v[784] = mfspr(784); /* SIER */ + v[785] = mfspr(785); /* MMCR2 */ + v[786] = mfspr(786); /* MMCRA */ + v[787] = mfspr(787); /* PMC1 */ + v[788] = mfspr(788); /* PMC2 */ + v[789] = mfspr(789); /* PMC3 */ + v[790] = mfspr(790); /* PMC4 */ + v[791] = mfspr(791); /* PMC5 */ + v[792] = mfspr(792); /* PMC6 */ + v[795] = mfspr(795); /* MMCR0 */ + v[796] = mfspr(796); /* SIAR */ + v[798] = mfspr(798); /* SDAR */ + v[800] = mfspr(800); /* BESCRS */ + v[801] = mfspr(801); /* BESCCRSU */ + v[802] = mfspr(802); /* BESCRR */ + v[803] = mfspr(803); /* BESCRRU */ + v[804] = mfspr(804); /* EBBHR */ + v[805] = mfspr(805); /* EBBRR */ + v[806] = mfspr(806); /* BESCR */ + v[815] = mfspr(815); /* TAR */ +} + +static void get_sprs(uint64_t *v) +{ + uint32_t pvr = mfspr(287); /* Processor Version Register */ + + get_sprs_common(v); + + switch (pvr >> 16) { + case 0x4b: /* POWER8E */ + case 0x4c: /* POWER8NVL */ + case 0x4d: /* POWER8 */ + get_sprs_book3s_207(v); + break; + } +} + +int main(int argc, char **argv) +{ + int i; + + if (argc > 1) + report_abort("Warning: Unsupported arguments!"); + + puts("Settings SPRs...\n"); + set_sprs(0xcafefacec0debabeULL); + + memset(before, 0, sizeof(before)); + memset(after, 0, sizeof(after)); + + get_sprs(before); + + if (argc > 0 && !strcmp(argv[0], "-w")) { + puts("Now migrate the VM, then press a key...\n"); + while (h_get_term_char(0) == 0) + ; + } else { + puts("Parameter '-w' not specified - not waiting for a key.\n"); + } + + get_sprs(after); + + puts("Checking SPRs...\n"); + for (i = 0; i < 1024; i++) { + if (before[i] != 0 || after[i] != 0) + report("SPR %d:\t0x%016lx <==> 0x%016lx", + before[i] == after[i], i, before[i], after[i]); + } + + return report_summary(); +} diff --git a/powerpc/unittests.cfg b/powerpc/unittests.cfg index ed4fdbe..5563cbe 100644 --- a/powerpc/unittests.cfg +++ b/powerpc/unittests.cfg @@ -50,3 +50,7 @@ groups = rtas [emulator] file = emulator.elf + +[sprs] +file = sprs.elf +#extra_params = -append '-w'
This test can be used to check whether the SPR (special purpose registers) of the PowerPC CPU are migrated right. It first fills the various SPRs with some non-zero value, then reads the values back into a first array, then waits for a key (with the '-w' option) so that it is possible to migrate the VM, and finally reads the values from the SPRs back into another array and then compares it with the initial values. Currently the test only supports the SPRs from the PowerISA v2.07 specification (i.e. POWER8 CPUs), but other versions should be pretty easy to add later. Signed-off-by: Thomas Huth <thuth@redhat.com> --- powerpc/Makefile.common | 5 +- powerpc/sprs.c | 230 ++++++++++++++++++++++++++++++++++++++++++++++++ powerpc/unittests.cfg | 4 + 3 files changed, 238 insertions(+), 1 deletion(-) create mode 100644 powerpc/sprs.c