On Tue, Jun 27, 2023 at 5:09 PM Andrew Jones [off-list ref] wrote:

On Fri, Jun 23, 2023 at 06:40:10PM +0800, Haibo Xu wrote:

quoted

No functional changes. Just move the reject_set check logic to a
function so we can check for specific errno for specific register.
This is a preparation for support reject_set in riscv.

Suggested-by: Andrew Jones <redacted>
Signed-off-by: Haibo Xu <redacted>
---
 tools/testing/selftests/kvm/aarch64/get-reg-list.c | 8 ++++++++
 tools/testing/selftests/kvm/get-reg-list.c         | 7 ++++++-
 2 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/kvm/aarch64/get-reg-list.c b/tools/testing/selftests/kvm/aarch64/get-reg-list.c
index aaf035c969ec..4e2e1fe833eb 100644
--- a/tools/testing/selftests/kvm/aarch64/get-reg-list.c
+++ b/tools/testing/selftests/kvm/aarch64/get-reg-list.c

@@ -27,6 +27,14 @@ bool filter_reg(__u64 reg)
      return false;
 }

+bool reject_set_fail(__u64 reg)
+{
+     if (reg == KVM_REG_ARM64_SVE_VLS)
+             return (errno != EPERM);
+
+     return false;
+}

I think we should pass errno in as a parameter and I prefer positive
predicate functions, so I'd name this check_reject_set() and reverse
the logic. Also, we don't want to check for KVM_REG_ARM64_SVE_VLS,
because that duplicates the rejects set. I see in a later patch
that riscv needs to check reg because different errors are used
for different registers, but that's because KVM_REG_RISCV_TIMER_REG(state)
was erroneously added to the rejects set. KVM_REG_RISCV_TIMER_REG(state)
doesn't belong there. That register can be set, but it only supports
certain input, otherwise, it correctly, results in EINVAL. We'll need
the concept of a "skip set" to avoid tripping over that one.

So, I think arm's function should be

 bool check_reject_set(int errno)
 {
     return errno == EPERM;
 }

and riscv's should be

 bool check_reject_set(int errno)
 {
     return errno == EOPNOTSUPP;
 }

Sure, will add a new 'skips_set' member to 'struct vcpu_reg_sublist' and
move KVM_REG_RISCV_TIMER_REG(state) reg to it.

quoted

+
 #define REG_MASK (KVM_REG_ARCH_MASK | KVM_REG_SIZE_MASK | KVM_REG_ARM_COPROC_MASK)

 #define CORE_REGS_XX_NR_WORDS        2

diff --git a/tools/testing/selftests/kvm/get-reg-list.c b/tools/testing/selftests/kvm/get-reg-list.c
index f6ad7991a812..b956ee410996 100644
--- a/tools/testing/selftests/kvm/get-reg-list.c
+++ b/tools/testing/selftests/kvm/get-reg-list.c

@@ -98,6 +98,11 @@ void __weak print_reg(const char *prefix, __u64 id)
      printf("\t0x%llx,\n", id);
 }

+bool __weak reject_set_fail(__u64 reg)
+{
+     return false;
+}
+
 #ifdef __aarch64__
 static void prepare_vcpu_init(struct vcpu_reg_list *c, struct kvm_vcpu_init *init)
 {

@@ -216,7 +221,7 @@ static void run_test(struct vcpu_reg_list *c)
                      if (s->rejects_set && find_reg(s->rejects_set, s->rejects_set_n, reg.id)) {
                              reject_reg = true;
                              ret = __vcpu_ioctl(vcpu, KVM_SET_ONE_REG, &reg);
-                             if (ret != -1 || errno != EPERM) {
+                             if (ret != -1 || reject_set_fail(reg.id)) {
                                      printf("%s: Failed to reject (ret=%d, errno=%d) ", config_name(c), ret, errno);
                                      print_reg(config_name(c), reg.id);
                                      putchar('\n');
--

2.34.1

Thanks,
drew

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel