Patch Set 74:

Patch Set 74:

Patch Set 74:

I had a look on Supermicro's products list, especially on the X11 series, when I bought my X11SSM-F some days ago (I am currently working on porting coreboot to that board, too). I just realized that X11SSM-F and X11SSH-TF are basically the same board with just some minor differences (SSH has an m.2 slot and 10GBase-T).

I am wondering if it may be sensible to group those boards in a different scheme than suggested here (mainboards/x11ssh/tf), because we can support more boards this way more simple.

Many X11SS* boards share the same chipset and are more or less the same boards but have only minimal differences in internal/external interfaces.

My suggestion would be to group all these mainboards by chipset and make the rest configurable by selecting the specific board in Kconfig.

This could result in the following tree:

*snip*

Is this structure used somewhere else as well in the coreboot tree? I dont think we should reinvent something new and mix up chipsets with vendors and board variants.

I believe that renaming this board's folder to "x11ss" would be good enough to add both the X11SSH-TF and the X11SSM-F as variants, as well as any future variants. Looks like the 'X11' in the board names defines the generation (in this case, Intel Skylake), and the 'SS' part means "Single Socket", so all X11SS* board ports could be grouped together.

Good guess but wrong ;-) Unfortunately SS does not mean single socket. The first S means single socket, the second is the socket type, where S seems to stand for LGA-1151, while R is R4 (e.g. X11SRL-F). But even that is not 100% consistent, only about 95%...

That was the reason I had the idea of group by chipset.


However, I'd prefer if the "variants without baseboard" Kconfig mechanism (e.g., what gigabyte/ga-h61m-s2pv does) would be used instead. I don't think we have a way of knowing what the baseboard for these supermicro boards would be.

Right. There are X11SSM and X11SSM-TF for example, which are two different (but very similiar) boards!


I don't think splitting by chipset makes sense if all the boards were to run the same code, especially considering the additional nesting this would result in. However, I believe the X10 series is an older Intel platform than the X11 series, so code reuse between those series isn't worth implementing.

Not all X11 are C236 and thus do not run exactly the same code. Some are C232 and others. Mixing those could get very confusing


The resulting paths would be:
src/mainboard/supermicro/x11ss/variants/x11ssh-tf/
src/mainboard/supermicro/x11ss/variants/x11ssm-f/

View Change

To view, visit change 32734. To unsubscribe, or for help writing mail filters, visit settings.

Gerrit-Project: coreboot
Gerrit-Branch: master
Gerrit-Change-Id: I2edaa4a928de3a065e517c0f20e3302b4b702323
Gerrit-Change-Number: 32734
Gerrit-PatchSet: 74
Gerrit-Owner: Christian Walter <christian.walter@9elements.com>
Gerrit-Reviewer: Alexander Couzens <lynxis@fe80.eu>
Gerrit-Reviewer: Angel Pons <th3fanbus@gmail.com>
Gerrit-Reviewer: Christian Walter <christian.walter@9elements.com>
Gerrit-Reviewer: Felix Singer <felixsinger@posteo.net>
Gerrit-Reviewer: Keno Fischer <keno@alumni.harvard.edu>
Gerrit-Reviewer: Martin Roth <martinroth@google.com>
Gerrit-Reviewer: Name of user not set #1002358
Gerrit-Reviewer: Patrick Georgi <pgeorgi@google.com>
Gerrit-Reviewer: Patrick Rudolph <patrick.rudolph@9elements.com>
Gerrit-Reviewer: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
Gerrit-Reviewer: T. Hudson <trammell.hudson@gmail.com>
Gerrit-Reviewer: Trammell Hudson <hudson@trmm.net>
Gerrit-Reviewer: build bot (Jenkins) <no-reply@coreboot.org>
Gerrit-CC: Christoph Pomaska <github@aufmachen.jetzt>
Gerrit-CC: Felix Held <felix-coreboot@felixheld.de>
Gerrit-CC: Maxim Polyakov <max.senia.poliak@gmail.com>
Gerrit-CC: Michael Niewöhner
Gerrit-CC: Paul Menzel <paulepanter@users.sourceforge.net>
Gerrit-Comment-Date: Fri, 23 Aug 2019 19:08:18 +0000
Gerrit-HasComments: No
Gerrit-Has-Labels: No
Gerrit-MessageType: comment