qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: David Hildenbrand <david@redhat.com>
To: Thomas Huth <thuth@redhat.com>,
	qemu-devel@nongnu.org, Eduardo Habkost <ehabkost@redhat.com>,
	Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-ppc@nongnu.org, qemu-s390x@nongnu.org,
	Bharata B Rao <bharata@linux.vnet.ibm.com>
Subject: Re: [Qemu-devel] [PATCH 2/4] tests/cpu-plug-test: Check the CPU hot-plugging with device_add, too
Date: Wed, 18 Oct 2017 10:55:59 +0200	[thread overview]
Message-ID: <3caa6d80-0ef2-7019-8048-911e9bd2c40b@redhat.com> (raw)
In-Reply-To: <1508247148-25240-3-git-send-email-thuth@redhat.com>

On 17.10.2017 15:32, Thomas Huth wrote:
> Using 'device_add' instead of 'cpu-add' is the new way for
> hot-plugging CPUs, so we should test this regularly, too.
> 
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>  tests/cpu-plug-test.c | 42 ++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 40 insertions(+), 2 deletions(-)
> 
> diff --git a/tests/cpu-plug-test.c b/tests/cpu-plug-test.c
> index 4579119..5fef297 100644
> --- a/tests/cpu-plug-test.c
> +++ b/tests/cpu-plug-test.c
> @@ -16,6 +16,7 @@
>  struct PlugTestData {
>      char *machine;
>      const char *cpu_model;
> +    const char *device_model;
>      unsigned sockets;
>      unsigned cores;
>      unsigned threads;
> @@ -71,6 +72,34 @@ static void test_plug_without_cpu_add(gconstpointer data)
>      g_free(args);
>  }
>  
> +static void test_plug_with_device_add_x86(gconstpointer data)
> +{
> +    const PlugTestData *td = data;
> +    char *args;
> +    unsigned int s, c, t;
> +
> +    args = g_strdup_printf("-machine %s -cpu %s "
> +                           "-smp sockets=%u,cores=%u,threads=%u,maxcpus=%u",
> +                           td->machine, td->cpu_model,
> +                           td->sockets, td->cores, td->threads, td->maxcpus);
> +    qtest_start(args);
> +

I wonder if we should also query the hotpluggable CPUs via
query-hotpluggable-cpus and try to add all (maybe even in random order).
Could also be a second test case.

> +    for (s = td->sockets; s < td->maxcpus / td->cores / td->threads; s++) {
> +        for (c = 0; c < td->cores; c++) {
> +            for (t = 0; t < td->threads; t++) {
> +                char *id = g_strdup_printf("id-%i-%i-%i", s, c, t);
> +                qtest_qmp_device_add(td->device_model, id, "'socket-id':'%i', "
> +                                     "'core-id':'%i', 'thread-id':'%i'",
> +                                     s, c, t);
> +                g_free(id);
> +            }
> +        }
> +    }
> +
> +    qtest_end();
> +    g_free(args);
> +}
> +
>  static void test_data_free(gpointer data)
>  {
>      PlugTestData *pc = data;
> @@ -90,6 +119,7 @@ static void add_pc_test_case(const char *mname)
>      data = g_new(PlugTestData, 1);
>      data->machine = g_strdup(mname);
>      data->cpu_model = "Haswell"; /* 1.3+ theoretically */
> +    data->device_model = "Haswell-x86_64-cpu";
>      data->sockets = 1;
>      data->cores = 3;
>      data->threads = 2;
> @@ -105,19 +135,27 @@ static void add_pc_test_case(const char *mname)
>          (strcmp(mname, "pc-0.12") == 0) ||
>          (strcmp(mname, "pc-0.11") == 0) ||
>          (strcmp(mname, "pc-0.10") == 0)) {
> -        path = g_strdup_printf("cpu/%s/init/%ux%ux%u&maxcpus=%u",
> +        path = g_strdup_printf("cpu-plug/%s/init/%ux%ux%u&maxcpus=%u",
>                                 mname, data->sockets, data->cores,
>                                 data->threads, data->maxcpus);
>          qtest_add_data_func_full(path, data, test_plug_without_cpu_add,
>                                   test_data_free);
>          g_free(path);
>      } else {
> -        path = g_strdup_printf("cpu/%s/add/%ux%ux%u&maxcpus=%u",
> +        PlugTestData *data2 = g_memdup(data, sizeof(PlugTestData));

add an empty line?

> +        data2->machine = g_strdup(mname);
> +        path = g_strdup_printf("cpu-plug/%s/cpu-add/%ux%ux%u&maxcpus=%u",
>                                 mname, data->sockets, data->cores,
>                                 data->threads, data->maxcpus);
>          qtest_add_data_func_full(path, data, test_plug_with_cpu_add,
>                                   test_data_free);
>          g_free(path);
> +        path = g_strdup_printf("cpu-plug/%s/device-add/%ux%ux%u&maxcpus=%u",
> +                               mname, data2->sockets, data2->cores,
> +                               data2->threads, data2->maxcpus);
> +        qtest_add_data_func_full(path, data2, test_plug_with_device_add_x86,
> +                                 test_data_free);
> +        g_free(path);
>      }
>  }
>  
> 

device_model can usually be created from cpu_model (although not 100%
clean, e.g. s390 vs. s390x). But the distinction here actually allows us
to test that we can't plug mixed CPU models.

So we might add a test case later that tries to hotplug a different CPU.

Reviewed-by: David Hildenbrand <david@redhat.com>

-- 

Thanks,

David

  reply	other threads:[~2017-10-18  8:56 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-17 13:32 [Qemu-devel] [PATCH 0/4] Improve the CPU hot plug tests Thomas Huth
2017-10-17 13:32 ` [Qemu-devel] [PATCH 1/4] tests: Rename pc-cpu-test.c to cpu-plug-test.c Thomas Huth
2017-10-17 16:48   ` Philippe Mathieu-Daudé
2017-10-18  8:46   ` David Hildenbrand
2017-10-17 13:32 ` [Qemu-devel] [PATCH 2/4] tests/cpu-plug-test: Check the CPU hot-plugging with device_add, too Thomas Huth
2017-10-18  8:55   ` David Hildenbrand [this message]
2017-10-17 13:32 ` [Qemu-devel] [PATCH 3/4] tests/cpu-plug-test: Check CPU hot-plugging on ppc64, too Thomas Huth
2017-10-17 15:59   ` [Qemu-devel] [Qemu-ppc] " Daniel Henrique Barboza
2017-10-17 13:32 ` [Qemu-devel] [PATCH 4/4] tests/cpu-plug-test: Test CPU hot-plugging on s390x Thomas Huth
2017-10-18  8:58   ` David Hildenbrand
2017-10-17 16:27 ` [Qemu-devel] [Qemu-ppc] [PATCH 0/4] Improve the CPU hot plug tests Daniel Henrique Barboza
2017-10-17 16:36   ` Thomas Huth
2017-10-17 18:13     ` Daniel Henrique Barboza

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=3caa6d80-0ef2-7019-8048-911e9bd2c40b@redhat.com \
    --to=david@redhat.com \
    --cc=bharata@linux.vnet.ibm.com \
    --cc=ehabkost@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    --cc=qemu-s390x@nongnu.org \
    --cc=thuth@redhat.com \
    /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).