| Message ID | 1628612544-25130-4-git-send-email-pmorel@linux.ibm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | S390x: CPU Topology Information | expand |
On 8/10/21 6:22 PM, Pierre Morel wrote: > We check the PTF instruction. > > - We do not expect to support vertical polarization. KVM does not support vertical polarization and we don't expect it to be added in the future? > > - We do not expect the Modified Topology Change Report to be > pending or not at the moment the first PTF instruction with > PTF_CHECK function code is done as some code already did run > a polarization change may have occur. ENOPARSE > > Signed-off-by: Pierre Morel <pmorel@linux.ibm.com> > --- > s390x/Makefile | 1 + > s390x/topology.c | 99 +++++++++++++++++++++++++++++++++++++++++++++ > s390x/unittests.cfg | 3 ++ > 3 files changed, 103 insertions(+) > create mode 100644 s390x/topology.c > > diff --git a/s390x/Makefile b/s390x/Makefile > index 6565561b..c82b7dbf 100644 > --- a/s390x/Makefile > +++ b/s390x/Makefile > @@ -24,6 +24,7 @@ tests += $(TEST_DIR)/mvpg.elf > tests += $(TEST_DIR)/uv-host.elf > tests += $(TEST_DIR)/edat.elf > tests += $(TEST_DIR)/mvpg-sie.elf > +tests += $(TEST_DIR)/topology.elf > > tests_binary = $(patsubst %.elf,%.bin,$(tests)) > ifneq ($(HOST_KEY_DOCUMENT),) > diff --git a/s390x/topology.c b/s390x/topology.c > new file mode 100644 > index 00000000..a0dc3b9e > --- /dev/null > +++ b/s390x/topology.c > @@ -0,0 +1,99 @@ > +/* SPDX-License-Identifier: GPL-2.0-only */ > +/* > + * CPU Topology > + * > + * Copyright (c) 2021 IBM Corp > + * > + * Authors: > + * Pierre Morel <pmorel@linux.ibm.com> > + */ > + > +#include <libcflat.h> > +#include <asm/page.h> > +#include <asm/asm-offsets.h> > +#include <asm/interrupt.h> > +#include <asm/facility.h> > +#include <smp.h> > +#include <sclp.h> > + > +static int machine_level; > + > +#define PTF_REQ_HORIZONTAL 0 > +#define PTF_REQ_VERTICAL 1 > +#define PTF_REQ_CHECK 2 > + > +#define PTF_ERR_NO_REASON 0 > +#define PTF_ERR_ALRDY_POLARIZED 1 > +#define PTF_ERR_IN_PROGRESS 2 > + > +static int ptf(unsigned long fc, unsigned long *rc) > +{ > + int cc; > + > + asm volatile( > + " .insn rre,0xb9a20000,%1,0\n" > + " ipm %0\n" > + " srl %0,28\n" > + : "=d" (cc), "+d" (fc) > + : "d" (fc) > + : "cc"); > + > + *rc = fc >> 8; > + return cc; > +} > + > +static void test_ptf(void) > +{ > + unsigned long rc; > + int cc; > + > + report_prefix_push("Topology Report pending"); > + /* > + * At this moment the topology may already have changed > + * since the VM has been started. > + * However, we can test if a second PTF instruction > + * reports that the topology did not change since the > + * preceding PFT instruction. > + */ > + ptf(PTF_REQ_CHECK, &rc); > + cc = ptf(PTF_REQ_CHECK, &rc); > + report(cc == 0, "PTF check clear"); > + > + /* > + * In the LPAR we can not assume the state of the polarizatiom polarization > + * at this moment. > + * Let's skip the tests for LPAR. > + */ Any idea what happens on z/VM? We don't necessarily need to support z/VM but we at least need to skip like we do on lpar :-) Maybe also add a TODO, so we know we could improve the test? > + if (machine_level < 3) > + goto end; > + Add comments: We're always horizontally polarized in KVM. > + cc = ptf(PTF_REQ_HORIZONTAL, &rc); > + report(cc == 2 && rc == PTF_ERR_ALRDY_POLARIZED, > + "PTF horizontal already configured"); > + KVM doesn't support vertical polarization. > + cc = ptf(PTF_REQ_VERTICAL, &rc); > + report(cc == 2 && rc == PTF_ERR_NO_REASON, > + "PTF vertical non possible"); s/non/not/ > + > +end: > + report_prefix_pop(); > +} > + > +int main(int argc, char *argv[]) > +{ > + report_prefix_push("CPU Topology"); > + > + if (!test_facility(11)) { > + report_skip("Topology facility not present"); > + goto end; > + } > + > + machine_level = stsi_get_fc(); > + report_info("Machine level %d", machine_level); > + > + test_ptf(); > + > +end: > + report_prefix_pop(); > + return report_summary(); > +} > diff --git a/s390x/unittests.cfg b/s390x/unittests.cfg > index 9e1802fd..0f84d279 100644 > --- a/s390x/unittests.cfg > +++ b/s390x/unittests.cfg > @@ -109,3 +109,6 @@ file = edat.elf > > [mvpg-sie] > file = mvpg-sie.elf > + > +[topology] > +file = topology.elf >
On 8/12/21 11:38 AM, Janosch Frank wrote: > On 8/10/21 6:22 PM, Pierre Morel wrote: >> We check the PTF instruction. >> >> - We do not expect to support vertical polarization. > > KVM does not support vertical polarization and we don't expect it to be > added in the future? OK > >> >> - We do not expect the Modified Topology Change Report to be >> pending or not at the moment the first PTF instruction with >> PTF_CHECK function code is done as some code already did run >> a polarization change may have occur. > > ENOPARSE OK I find another way to explain: " The Topology changes if the topology of the real CPUs backing the vCPUs changes. This can happen between the initialization of the VM and the start of the guest. As a consequence we can not expect the result of the first PTF instruction. " ... >> + /* >> + * In the LPAR we can not assume the state of the polarizatiom > > polarization yes > >> + * at this moment. >> + * Let's skip the tests for LPAR. >> + */ > > Any idea what happens on z/VM? > We don't necessarily need to support z/VM but we at least need to skip > like we do on lpar :-) No, I do not know. Then OK we skip the test for zVM too > > Maybe also add a TODO, so we know we could improve the test? > >> + if (machine_level < 3) >> + goto end; >> + > > Add comments: > We're always horizontally polarized in KVM. OK > >> + cc = ptf(PTF_REQ_HORIZONTAL, &rc); >> + report(cc == 2 && rc == PTF_ERR_ALRDY_POLARIZED, >> + "PTF horizontal already configured"); >> + > > KVM doesn't support vertical polarization. OK too > >> + cc = ptf(PTF_REQ_VERTICAL, &rc); >> + report(cc == 2 && rc == PTF_ERR_NO_REASON, >> + "PTF vertical non possible"); > > s/non/not/ yes, seems I forgot to change this. ... Thanks, Pierre
diff --git a/s390x/Makefile b/s390x/Makefile index 6565561b..c82b7dbf 100644 --- a/s390x/Makefile +++ b/s390x/Makefile @@ -24,6 +24,7 @@ tests += $(TEST_DIR)/mvpg.elf tests += $(TEST_DIR)/uv-host.elf tests += $(TEST_DIR)/edat.elf tests += $(TEST_DIR)/mvpg-sie.elf +tests += $(TEST_DIR)/topology.elf tests_binary = $(patsubst %.elf,%.bin,$(tests)) ifneq ($(HOST_KEY_DOCUMENT),) diff --git a/s390x/topology.c b/s390x/topology.c new file mode 100644 index 00000000..a0dc3b9e --- /dev/null +++ b/s390x/topology.c @@ -0,0 +1,99 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * CPU Topology + * + * Copyright (c) 2021 IBM Corp + * + * Authors: + * Pierre Morel <pmorel@linux.ibm.com> + */ + +#include <libcflat.h> +#include <asm/page.h> +#include <asm/asm-offsets.h> +#include <asm/interrupt.h> +#include <asm/facility.h> +#include <smp.h> +#include <sclp.h> + +static int machine_level; + +#define PTF_REQ_HORIZONTAL 0 +#define PTF_REQ_VERTICAL 1 +#define PTF_REQ_CHECK 2 + +#define PTF_ERR_NO_REASON 0 +#define PTF_ERR_ALRDY_POLARIZED 1 +#define PTF_ERR_IN_PROGRESS 2 + +static int ptf(unsigned long fc, unsigned long *rc) +{ + int cc; + + asm volatile( + " .insn rre,0xb9a20000,%1,0\n" + " ipm %0\n" + " srl %0,28\n" + : "=d" (cc), "+d" (fc) + : "d" (fc) + : "cc"); + + *rc = fc >> 8; + return cc; +} + +static void test_ptf(void) +{ + unsigned long rc; + int cc; + + report_prefix_push("Topology Report pending"); + /* + * At this moment the topology may already have changed + * since the VM has been started. + * However, we can test if a second PTF instruction + * reports that the topology did not change since the + * preceding PFT instruction. + */ + ptf(PTF_REQ_CHECK, &rc); + cc = ptf(PTF_REQ_CHECK, &rc); + report(cc == 0, "PTF check clear"); + + /* + * In the LPAR we can not assume the state of the polarizatiom + * at this moment. + * Let's skip the tests for LPAR. + */ + if (machine_level < 3) + goto end; + + cc = ptf(PTF_REQ_HORIZONTAL, &rc); + report(cc == 2 && rc == PTF_ERR_ALRDY_POLARIZED, + "PTF horizontal already configured"); + + cc = ptf(PTF_REQ_VERTICAL, &rc); + report(cc == 2 && rc == PTF_ERR_NO_REASON, + "PTF vertical non possible"); + +end: + report_prefix_pop(); +} + +int main(int argc, char *argv[]) +{ + report_prefix_push("CPU Topology"); + + if (!test_facility(11)) { + report_skip("Topology facility not present"); + goto end; + } + + machine_level = stsi_get_fc(); + report_info("Machine level %d", machine_level); + + test_ptf(); + +end: + report_prefix_pop(); + return report_summary(); +} diff --git a/s390x/unittests.cfg b/s390x/unittests.cfg index 9e1802fd..0f84d279 100644 --- a/s390x/unittests.cfg +++ b/s390x/unittests.cfg @@ -109,3 +109,6 @@ file = edat.elf [mvpg-sie] file = mvpg-sie.elf + +[topology] +file = topology.elf
We check the PTF instruction. - We do not expect to support vertical polarization. - We do not expect the Modified Topology Change Report to be pending or not at the moment the first PTF instruction with PTF_CHECK function code is done as some code already did run a polarization change may have occur. Signed-off-by: Pierre Morel <pmorel@linux.ibm.com> --- s390x/Makefile | 1 + s390x/topology.c | 99 +++++++++++++++++++++++++++++++++++++++++++++ s390x/unittests.cfg | 3 ++ 3 files changed, 103 insertions(+) create mode 100644 s390x/topology.c