From: Sean Christopherson <seanjc@google.com>
To: Manali Shukla <manali.shukla@amd.com>
Cc: kvm@vger.kernel.org, linux-kselftest@vger.kernel.org,
pbonzini@redhat.com, shuah@kernel.org, nikunj@amd.com
Subject: Re: [PATCH v1 2/4] KVM: selftests: Add an interface to read the data of named vcpu stat
Date: Thu, 19 Dec 2024 16:35:12 -0800 [thread overview]
Message-ID: <Z2S7wArwoBu4wBUb@google.com> (raw)
In-Reply-To: <20241021062226.108657-3-manali.shukla@amd.com>
On Mon, Oct 21, 2024, Manali Shukla wrote:
> From: Manali Shukla <Manali.Shukla@amd.com>
>
> The interface is used to read the data values of a specified vcpu stat
> from the currenly available binary stats interface.
>
> Add a concatenation trickery to trigger compiler error if the stat
> doesn't exist, so that it is not possible to pass a per-VM stat into
> vcpu_get_stat().
>
> Suggested-by: Sean Christopherson <seanjc@google.com>
> Signed-off-by: Manali Shukla <Manali.Shukla@amd.com>
> ---
> .../testing/selftests/kvm/include/kvm_util.h | 52 +++++++++++++++++++
> .../kvm/include/x86_64/kvm_util_arch.h | 36 +++++++++++++
> tools/testing/selftests/kvm/lib/kvm_util.c | 40 ++++++++++++++
> 3 files changed, 128 insertions(+)
>
> diff --git a/tools/testing/selftests/kvm/include/kvm_util.h b/tools/testing/selftests/kvm/include/kvm_util.h
> index bc7c242480d6..5dd3acf174f8 100644
> --- a/tools/testing/selftests/kvm/include/kvm_util.h
> +++ b/tools/testing/selftests/kvm/include/kvm_util.h
> @@ -531,6 +531,14 @@ void read_stat_data(int stats_fd, struct kvm_stats_header *header,
> struct kvm_stats_desc *desc, uint64_t *data,
> size_t max_elements);
>
> +#define DEFINE_CHECK_STAT(type, stat) \
> +static inline int check_##type##_##stat##_exists(void) \
> +{ \
> + return 1; \
> +} \
> +
> +#define STAT_EXISTS(type, stat) (check_##type##_##stat##_exists())
This is all unnecessary complicated. To trigger a compilation error, the set
of knnown stats just needs to be defined as _something_ and then referenced.
There's no need for layers of macros and a function for each stat. The fact that
a stat is defined is proof of its existence.
> +
> void __vm_get_stat(struct kvm_vm *vm, const char *stat_name, uint64_t *data,
> size_t max_elements);
>
> @@ -542,6 +550,50 @@ static inline uint64_t vm_get_stat(struct kvm_vm *vm, const char *stat_name)
> return data;
> }
>
> +#define DEFINE_GENERIC_VCPU_STAT \
> + DEFINE_CHECK_STAT(vcpu, halt_successfull_poll) \
> + DEFINE_CHECK_STAT(vcpu, halt_attempted_poll) \
> + DEFINE_CHECK_STAT(vcpu, halt_poll_invalid) \
> + DEFINE_CHECK_STAT(vcpu, halt_wakeup) \
> + DEFINE_CHECK_STAT(vcpu, halt_poll_success_ns) \
> + DEFINE_CHECK_STAT(vcpu, halt_poll_fail_ns) \
> + DEFINE_CHECK_STAT(vcpu, halt_wait_ns) \
> + DEFINE_CHECK_STAT(vcpu, halt_poll_success_hist) \
> + DEFINE_CHECK_STAT(vcpu, halt_poll_fail_hist) \
> + DEFINE_CHECK_STAT(vcpu, halt_wait_hist) \
> + DEFINE_CHECK_STAT(vcpu, blocking) \
> +
> +/*
> + * Define a default empty macro for architectures which do not specify
> + * arch specific vcpu stats
> + */
> +
> +#ifndef DEFINE_ARCH_VCPU_STAT
> +#define DEFINE_ARCH_VCPU_STAT
> +#endif
> +
> +DEFINE_GENERIC_VCPU_STAT
There's also no need to define macros in arch code just to expand them in common
code. Add simple macros in kvm_util_types.h and this goes away.
> +DEFINE_ARCH_VCPU_STAT
> +
> +#undef DEFINE_CHECK_STAT
> +#undef DEFINE_GENERIC_VCPU_STAT
> +#undef DEFINE_ARCH_VCPU_STAT
> +void __vcpu_get_stat(struct kvm_vcpu *vcpu, const char *stat_name, uint64_t *data,
> + size_t max_elements);
> +
> +#define vcpu_get_stat(vcpu, stat_name) \
> +({ \
> + uint64_t data; \
> + \
> + STAT_EXISTS(vcpu, stat_name); \
> + __vcpu_get_stat(vcpu, #stat_name, &data, 1); \
> + data; \
> +}) \
> +
> +#undef DEFINE_CHECK_STAT
> +#undef DEFINE_GENERIC_VCPU_STAT
> +
...
> +void __vcpu_get_stat(struct kvm_vcpu *vcpu, const char *stat_name, uint64_t *data,
> + size_t max_elements)
> +{
> + int vcpu_stats_fd;
> + struct kvm_stats_header header;
> + struct kvm_stats_desc *desc, *t_desc;
> + size_t size_desc;
> + int i;
> +
> + vcpu_stats_fd = vcpu_get_stats_fd(vcpu);
> + read_stats_header(vcpu_stats_fd, &header);
> +
> + desc = read_stats_descriptors(vcpu_stats_fd, &header);
> + size_desc = get_stats_descriptor_size(&header);
> +
> + for (i = 0; i < header.num_desc; ++i) {
> + t_desc = (void *)desc + (i * size_desc);
> +
> + if (strcmp(t_desc->name, stat_name))
> + continue;
> +
> + read_stat_data(vcpu_stats_fd, &header, t_desc,
> + data, max_elements);
> + break;
> + }
> +}
This is copy-pasted nearly verbatim from the VM-scoped code. It even has the
same bugs (doesn't assert the stat exists), along with new bugs (leaks the fd
and header).
It takes a bit of work, but not _that_ much work, to genericize the VM-scoped
infrastructure and reuse it for vCPU-scoped stats.
next prev parent reply other threads:[~2024-12-20 0:35 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-21 6:22 [PATCH v1 0/4] KVM selftests enhancements Manali Shukla
2024-10-21 6:22 ` [PATCH v1 1/4] KVM: selftests: Add safe_halt() and cli() helpers to common code Manali Shukla
2024-12-19 17:39 ` Sean Christopherson
2024-10-21 6:22 ` [PATCH v1 2/4] KVM: selftests: Add an interface to read the data of named vcpu stat Manali Shukla
2024-12-20 0:35 ` Sean Christopherson [this message]
2024-10-21 6:22 ` [PATCH v1 3/4] KVM: selftests: convert vm_get_stat to macro Manali Shukla
2024-10-21 6:22 ` [PATCH v1 4/4] KVM: selftests: Replace previously used vm_get_stat() " Manali Shukla
2024-12-20 0:42 ` Sean Christopherson
2024-11-28 15:06 ` [PATCH v1 0/4] KVM selftests enhancements Manali Shukla
2024-12-20 0:52 ` Sean Christopherson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Z2S7wArwoBu4wBUb@google.com \
--to=seanjc@google.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=manali.shukla@amd.com \
--cc=nikunj@amd.com \
--cc=pbonzini@redhat.com \
--cc=shuah@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).