Re: [PATCH v4 4/6] migration-test: Make sure that multifd and cancel works

[Peter Maydell]

On Wed, 22 Jan 2020 at 11:20, Juan Quintela <quintela@redhat.com> wrote:
>
> Test that this sequerce works:
>
> - launch source
> - launch target
> - start migration
> - cancel migration
> - relaunch target
> - do migration again
>
> Signed-off-by: Juan Quintela <quintela@redhat.com>
> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> Signed-off-by: Juan Quintela <quintela@redhat.com>

A year-old commit, but we've just got around to running Coverity
on the tests/ directory, and it spotted this one:

>  static void migrate_set_capability(QTestState *who, const char *capability,
>                                     bool value)

The 3rd argument to migrate_set_capability() is a bool...


> +static void test_multifd_tcp_cancel(void)
> +{

> +    migrate_set_parameter_int(from, "downtime-limit", 1);
> +    /* 300MB/s */
> +    migrate_set_parameter_int(from, "max-bandwidth", 30000000);
> +
> +    migrate_set_parameter_int(from, "multifd-channels", 16);
> +    migrate_set_parameter_int(to, "multifd-channels", 16);
> +
> +    migrate_set_capability(from, "multifd", "true");
> +    migrate_set_capability(to, "multifd", "true");

...but here you pass it the character string '"true"' rather than
the boolean value 'true'.

This works by fluke since the implicit comparison of the literal string
against NULL will evaluate to true, but it isn't really right :-)

CID 1432373, 1432292, 1432288.

There seem to be 7 uses of the string "true" when the boolean
was intended; I don't know why Coverity only found 3 issues.

thanks
-- PMM