On Mon, Feb 08, 2016 at 04:57:46PM +0530, Jayachandran Chandrashekaran Nair wrote:

Lorenzo,

On Fri, Feb 5, 2016 at 3:17 PM, Lorenzo Pieralisi
[off-list ref] wrote:

quoted

On Fri, Feb 05, 2016 at 02:05:37PM +0530, Jayachandran Chandrashekaran Nair wrote:

[...]

quoted

pci_host_acpi.c is a generic implementation of these using a sysdata
pointing to acpi_pci_root_info, and using a pointer to the pci_mmcfg_region
to access ECAM area, Maybe I can rename this file to
pci_acpi_host_generic.c to reflect this better.

Maybe you should stop sending this series and work with Tomasz to
get this done, you are confusing everyone and I am really really
annoyed about this.

Do you realize there is no point in having two patch series doing
the same thing and wasting everyone's review time ?

Do you realize he started this work long before you and went through
several rounds of review already (I told you before but in case you
forgot) ?

Tomasz posted a version yesterday, integrating comments following months
of review and testing and I think it is ready to get upstream:

https://lkml.org/lkml/2016/2/4/646

Did you even consider reviewing his code or helping him instead of
churning out more patches doing the *SAME* thing ?

This is getting ridiculous, I had replied to your earlier mails on why
my patchset is NOT doing the exact same thing. I had also explained
why helping was not feasible.

The basic point again: I am trying to give a much simpler patchset
to solve the same problem, I take it that you haven't reviewed my
patchset before writing this mail. I would have appreciated
a technical discussion rather than this pointless flamefest.

If you have reviewed it, you can see that there are just 5 patches
instead of 23, and that overall it is a much simpler approach.

Ok, let's make it constructive. I think there is part of your
implementation that definitely makes sense (in particular the way you
cleaned-up the x86 MCFG unadulterated mess - patch 1), I will ask
Tomasz to integrate it, please work together on this.

I have nothing against your patchset, my point is that we can't keep
reviewing and testing two series (and I mean on ARM64 AND x86), please
understand my point, it is very time consuming to understand the
differences and make sure we don't break x86 in the process and I would
have to ask you to add all the code that I already reviewed in Tomasz's
set, I just do not want to do that.

I will reply to Tomasz, let's work together to have a single final
implementation please, I do not think I am asking too much here and
yes, by integrating part of your code I think Tomasz's patchset is
ready to go, obviously subject to Bjorn's review and opinion.

Thanks,
Lorenzo