@@ -354,7 +354,7 @@ void *guest_flat_to_host(struct kvm *kvm, u64 offset)
return bank->host_addr + (offset - bank_start);
}
- pr_warning("unable to translate guest address 0x%llx to host",
+ pr_debug("unable to translate guest address 0x%llx to host",
(unsigned long long)offset);
return NULL;
}
@@ -371,7 +371,7 @@ u64 host_to_guest_flat(struct kvm *kvm, void *ptr)
return bank->guest_phys_addr + (ptr - bank_start);
}
- pr_warning("unable to translate host address %p to guest", ptr);
+ pr_debug("unable to translate host address %p to guest", ptr);
return 0;
}
guest_flat_to_host() and host_to_guest_flat return NULL, respectively 0, when the address is not found in the existing memslots. It is expected that the calling code will handle this error. However, both functions also print an error message containing the offending address. This can be redundant, if the calling code also prints the address, or even misleading, if the calling code can gracefully handle the failure, like is the case in kvm__dump_mem(). Change the warning to a debug, since knowing the address might still be useful for those call sites where the address isn't printed, or if the error isn't handled at all. Before, when running the PMU test from kvm-unit-tests using the test runner, which redirects stdout (where kvm__dump_mem() writes) and stderr (where pr_warning() writes): Error: KVM exit reason: 9 ("KVM_EXIT_FAIL_ENTRY") Warning: unable to translate guest address 0x0 to host Registers: <snip> *lr: 0x00000000: <unknown> 0x00000008: <unknown> 0x00000010: <unknown> 0x00000018: <unknown> The error is caused by the VCPU migrating to a physical CPU with a different PMU. In the example, the warning not only is unnecessary, but is quite a distance away from the offending address and can confuse the person looking at the code, like it happened to the patch author. Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com> --- kvm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)