[Automated-testing] Looking for a Debian kernel provisioning solution

[Bird, Timothy]

> -----Original Message-----
> From: Rees, Kevron 
> On Fri, Jan 26, 2018 at 12:38 PM,  <Tim.Bird at sony.com> wrote:
> > Does LAVA control grub over the serial console?
> >
> 
> Yes.  So either the board itself has to be able to output console
> information over serial, or grub has to be configured to do so.  I
> haven't personally done grub yet.  It's possible and it's on my list
> of things to figure out since we have some boards that do not have a
> bios that supports serial output.
Indeed.  I was trying to work out how to control grub if there's
no serial console.  I'm not sure if it's possible to try serial first,
then fall back to screen/keyboard-based operation, but that would
be nice.

> 
> > Does the grub for mbt have to be specially built for this, or is this
> > a feature of a grub binary that's already available?  That is,
> > does the default grub that comes in Intel images have
> > support for serial console?
> >
> 
> Grub has to be built with network support (and serial support if you
> use that).  We actually used a grub net boot efi file from ubuntu.
OK.

> 
> > Does LAVA already have the kernel and initramfs (pre-built) that I can
> > just grab, or is this something I have to build myself?
> >
> 
> They do.  We built our own, however.  It just needs to have network
> support, wget and dd.  We built ours with bmaptool also since it's
> faster at flashing than dd.
> 
> > Is this sequence of operations upstreamed anywhere?  Could I find
> > it in the LAVA repos, or is this Intel custom stuff?
> >
> 
> Everything except for the lava device-type are upstream.  We had to
> modify the x86-grub device type to make it work with this usb flash
> disk workflow.  We are working on upstreaming this new x86 device type
> with corresponding lava test job examples.
OK - I might try this approach, but write to SD card instead
of USB flash disk.

> 
> It is also possible you can supply your kernel and initramfs and run
> tests on the initramfs in memory.  That is a popular workflow with
> lava.  We needed to install to USB because I/O operations affect
> real-time latency in real life and we wanted our tests to reflect
> reality as much as possible.
Yeah - same here.  I want the testing to be on as close to a "production"
system as possible, including booting from the SD card, if possible.
I have filesystem performance tests that are meaningless if I'm 
running from a ramfs.

Thanks for the additional information.
 -- Tim