Re: [PATCH v3] powerpc: Add tests for sPAPR h-calls

[David Gibson <dgibson@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 6316 bytes --]

On Wed, 2 Mar 2016 11:07:55 +0100
Laurent Vivier <lvivier@redhat.com> wrote:

> On 02/03/2016 09:40, Thomas Huth wrote:
> > Introduce a test for sPAPR hypercalls, starting with the
> > three hypercalls H_SET_SPRG0, H_PAGE_INIT and H_RANDOM.
> > 
> > Signed-off-by: Thomas Huth <thuth@redhat.com>
> > ---
> >  v3:
> >  - Use report_xfail() in H_RANDOM test to report whether the
> >    h-call is available or not
> > 
> >  v2:
> >  - Rebased to the final version of Andrew's initial ppc64
> >    support patches that got merged yesterday
> >  - Added a test for the H_RANDOM hypercall
> > 
> >  Please note that you need the latest QEMU development version
> >  since some of the hypercalls have only been added recently.
> > 
> >  In case somebody wants to review the description of the h-calls, you
> >  can find them in LoPAPR, chapters 14.5.4.3.1, 14.5.4.3.3 and 14.15.1.
> >  See: https://members.openpowerfoundation.org/document/dl/469
> > 
> >  lib/powerpc/asm/hcall.h |   3 +
> >  powerpc/Makefile.common |   5 +-
> >  powerpc/spapr_hcall.c   | 167 ++++++++++++++++++++++++++++++++++++++++++++++++
> >  powerpc/unittests.cfg   |   3 +
> >  4 files changed, 177 insertions(+), 1 deletion(-)
> >  create mode 100644 powerpc/spapr_hcall.c
> > 
> > diff --git a/lib/powerpc/asm/hcall.h b/lib/powerpc/asm/hcall.h
> > index 750c655..f6f9ea8 100644
> > --- a/lib/powerpc/asm/hcall.h
> > +++ b/lib/powerpc/asm/hcall.h
> > @@ -15,8 +15,11 @@
> >  #define H_PRIVILEGE		-3
> >  #define H_PARAMETER		-4
> >  
> > +#define H_SET_SPRG0		0x24
> >  #define H_SET_DABR		0x28
> > +#define H_PAGE_INIT		0x2c
> >  #define H_PUT_TERM_CHAR		0x58
> > +#define H_RANDOM		0x300
> >  
> >  #ifndef __ASSEMBLY__
> >  /*
> > diff --git a/powerpc/Makefile.common b/powerpc/Makefile.common
> > index b526668..2ce6494 100644
> > --- a/powerpc/Makefile.common
> > +++ b/powerpc/Makefile.common
> > @@ -5,7 +5,8 @@
> >  #
> >  
> >  tests-common = \
> > -	$(TEST_DIR)/selftest.elf
> > +	$(TEST_DIR)/selftest.elf \
> > +	$(TEST_DIR)/spapr_hcall.elf
> >  
> >  all: $(TEST_DIR)/boot_rom.bin test_cases
> >  
> > @@ -63,3 +64,5 @@ generated_files = $(asm-offsets)
> >  test_cases: $(generated_files) $(tests-common) $(tests)
> >  
> >  $(TEST_DIR)/selftest.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/selftest.o
> > +
> > +$(TEST_DIR)/spapr_hcall.elf: $(cstart.o) $(reloc.o) $(TEST_DIR)/spapr_hcall.o
> > diff --git a/powerpc/spapr_hcall.c b/powerpc/spapr_hcall.c
> > new file mode 100644
> > index 0000000..46731e1
> > --- /dev/null
> > +++ b/powerpc/spapr_hcall.c
> > @@ -0,0 +1,167 @@
> > +/*
> > + * Test sPAPR hypervisor calls (aka. h-calls)
> > + *
> > + * Copyright 2016  Thomas Huth, Red Hat Inc.
> > + *
> > + * This work is licensed under the terms of the GNU LGPL, version 2.
> > + */
> > +#include <libcflat.h>
> > +#include <util.h>
> > +#include <alloc.h>
> > +#include <asm/hcall.h>
> > +
> > +#define PAGE_SIZE 4096
> > +
> > +#define H_ZERO_PAGE	(1UL << (63-48))
> > +#define H_COPY_PAGE	(1UL << (63-49))
> > +
> > +#define mfspr(nr) ({ \
> > +	uint64_t ret; \
> > +	asm volatile("mfspr %0,%1" : "=r"(ret) : "i"(nr)); \
> > +	ret; \
> > +})
> > +
> > +#define SPR_SPRG0	0x110
> > +
> > +/**
> > + * Test the H_SET_SPRG0 h-call by setting some values and checking whether
> > + * the SPRG0 register contains the correct values afterwards
> > + */
> > +static void test_h_set_sprg0(int argc, char **argv)
> > +{
> > +	uint64_t sprg0, sprg0_orig;
> > +	int rc;
> > +
> > +	if (argc > 1)
> > +		report_abort("Unsupported argument: '%s'", argv[1]);
> > +
> > +	sprg0_orig = mfspr(SPR_SPRG0);
> > +
> > +	rc = hcall(H_SET_SPRG0, 0xcafebabedeadbeefULL);
> > +	sprg0 = mfspr(SPR_SPRG0);
> > +	report("sprg0 = 0xcafebabedeadbeef",
> > +		rc == H_SUCCESS && sprg0 == 0xcafebabedeadbeefULL);
> > +
> > +	rc = hcall(H_SET_SPRG0, 0xaaaaaaaa55555555ULL);
> > +	sprg0 = mfspr(SPR_SPRG0);
> > +	report("sprg0 = 0xaaaaaaaa55555555",
> > +		rc == H_SUCCESS && sprg0 == 0xaaaaaaaa55555555ULL);
> > +
> > +	rc = hcall(H_SET_SPRG0, sprg0_orig);
> > +	sprg0 = mfspr(SPR_SPRG0);
> > +	report("sprg0 = 0x%llx",
> > +		rc == H_SUCCESS && sprg0 == sprg0_orig, sprg0_orig);
> > +}
> > +
> > +/**
> > + * Test the H_PAGE_INIT h-call by using it to clear and to copy a page, and
> > + * by checking for the correct values in the destination page afterwards
> > + */
> > +static void test_h_page_init(int argc, char **argv)
> > +{
> > +	u8 *dst, *src;
> > +	int rc;
> > +
> > +	if (argc > 1)
> > +		report_abort("Unsupported argument: '%s'", argv[1]);
> > +
> > +	dst = memalign(PAGE_SIZE, PAGE_SIZE);
> > +	src = memalign(PAGE_SIZE, PAGE_SIZE);
> > +	if (!dst || !src)
> > +		report_abort("Failed to alloc memory");
> > +
> > +	memset(dst, 0xaa, PAGE_SIZE);
> > +	rc = hcall(H_PAGE_INIT, H_ZERO_PAGE, dst, src);
> > +	report("h_zero_page", rc == H_SUCCESS && *(uint64_t*)dst == 0);
> > +
> > +	*(uint64_t*)src = 0xbeefc0dedeadcafeULL;
> > +	rc = hcall(H_PAGE_INIT, H_COPY_PAGE, dst, src);
> > +	report("h_copy_page",
> > +		rc == H_SUCCESS && *(uint64_t*)dst == 0xbeefc0dedeadcafeULL);
> > +
> > +	*(uint64_t*)src = 0x9abcdef012345678ULL;
> > +	rc = hcall(H_PAGE_INIT, H_COPY_PAGE|H_ZERO_PAGE, dst, src);
> > +	report("h_copy_page+h_zero_page",
> > +		rc == H_SUCCESS &&  *(uint64_t*)dst == 0x9abcdef012345678ULL);
> > +
> > +	rc = hcall(H_PAGE_INIT, H_ZERO_PAGE, dst + 0x123, src);
> > +	report("h_zero_page unaligned dst", rc == H_PARAMETER);
> > +
> > +	rc = hcall(H_PAGE_INIT, H_COPY_PAGE, dst, src + 0x123);
> > +	report("h_copy_page unaligned src", rc == H_PARAMETER);
> > +}
> > +
> > +static int h_random(uint64_t *val)
> > +{
> > +	register uint64_t r3 asm("r3") = H_RANDOM;
> > +	register uint64_t r4 asm("r4");
> > +
> > +	asm volatile (" sc 1 " : "+r"(r3), "=r"(r4) : "r"(r3));  
> 
> as you use "+r" with r3 in the output operands field, I think you don't
> have to declare it in the input operands field.

More importantly, I think a hypercall is allowed to clobber any of the
volatile registers (r0 & r3..r12), so they probably need to be declared
in the clobbers as well.

-- 
David Gibson <dgibson@redhat.com>
Senior Software Engineer, Virtualization, Red Hat

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]