[dpdk-dev] [PATCH v2] app/test: enhance power manager unit tests

Pattan, Reshma reshma.pattan at intel.com
Tue Apr 24 14:51:25 CEST 2018


Hi,

> -----Original Message-----
> From: Hunt, David
> Sent: Tuesday, April 24, 2018 1:10 PM
> To: Pattan, Reshma <reshma.pattan at intel.com>; Richardson, Bruce
> <bruce.richardson at intel.com>; Thomas Monjalon <thomas at monjalon.net>
> Cc: Parthasarathy, JananeeX M <jananeex.m.parthasarathy at intel.com>;
> dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] app/test: enhance power manager unit
> tests
> 
> 
> On 24/4/2018 12:23 PM, Pattan, Reshma wrote:
> > Hi,
> >
> >> -----Original Message-----
> >> From: Richardson, Bruce
> >> Sent: Tuesday, April 24, 2018 11:59 AM
> >> To: Thomas Monjalon <thomas at monjalon.net>
> >> Cc: Parthasarathy, JananeeX M <jananeex.m.parthasarathy at intel.com>;
> >> dev at dpdk.org; Pattan, Reshma <reshma.pattan at intel.com>; Hunt, David
> >> <david.hunt at intel.com>
> >> Subject: Re: [dpdk-dev] [PATCH v2] app/test: enhance power manager
> >> unit tests
> >>
> >> On Mon, Apr 23, 2018 at 11:04:27PM +0200, Thomas Monjalon wrote:
> >>> 11/04/2018 16:14, Reshma Pattan:
> >>>> Unit Testcases are added for power_acpi_cpu_freq,
> power_kvm_vm_test
> >>>> to improve coverage
> >>>>
> >>>> Signed-off-by: Jananee Parthasarathy
> >>>> <jananeex.m.parthasarathy at intel.com>
> >>>> Acked-by: David Hunt <david.hunt at intel.com>
> >>> Applied, thanks
> >>>
> >> Sadly, this patch seems to break shared library builds. If you try
> >> doing "make test-build" with shared libraries on it will fail, or if
> >> you do a meson build using shared libraries you will get the same result.
> >>
> >> The root cause is that the function guest_channel_host_connect() is a
> >> private function and so is not listed in the shared library map file,
> >> preventing the test app from linking.
> >>
> > Any action from my side required? Let me know.
> 
> Reshma,
>      Looking at this, I think this particular unit test needs to be removed. The
> way it is at the moment, it's "faking" the connect, then any commands that
> are sent to the dummy host are only really to test to see if the API breaks,
> which is going to be captured by compilation tests anyway. I don't see the
> value of this unit test unless you have the full host setup underneath is, in
> which case it's no longer a unit test.
> Also, we don't want to make these functions public, as they are only of use to
> the library internally, and there is no use for them publicly (unless a guest
> wants to fake a connection to a non-existent host).
> 
> What do you think?

Fine, we are reverting the changes and will send the patch soon.

Thanks,
Reshma




More information about the dev mailing list