Re: [PATCH v3 1/1] ArmVirtPkg: Replace obsoleted network drivers from platform DSC/FDF.

[Ard Biesheuvel <ard.biesheuvel@linaro.org>]

On 6 November 2018 at 16:24, Laszlo Ersek <lersek@redhat.com> wrote:
> On 11/06/18 13:32, Ard Biesheuvel wrote:
>> On 6 November 2018 at 02:24, Fu Siyuan <siyuan.fu@intel.com> wrote:
>>> V3:
>>> Remove duplicate library added in v2 patch, since ArmVirtPkg.dsc.inc
>>> already have them. Just remove the if...end there is enough.
>>>
>>> V2:
>>> Add missing library instance for NetworkPkg iSCSI driver.
>>>
>>
>> Please don't put the patch revision history in the commit log. Put it
>> below the ---
>>
>>> This patch replaces the MdeModulePkg TCP, PXE and iSCSI driver with those
>>> ones in NetworkPkg. These 3 drivers in MdeModulePkg are not being actively
>>> maintained and will be removed from edk2 master soon.
>>>
>>> Cc: Laszlo Ersek <lersek@redhat.com>
>>> Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>>> Cc: Julien Grall <julien.grall@linaro.org>
>>> Contributed-under: TianoCore Contribution Agreement 1.1
>>> Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
>>> ---
>>
>> ... here ...
>>
>> The patch looks fine to me
>>
>> Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>>
>> but please don't merge it until after the next stable tag has been created
>
> This is not a bad idea (see also your discussion with Leif); however it
> does create a bit of inconsistency with how the other platform DSC/FDF
> files have been handled. (The changes have been pushed for those.)
>
> Again, I don't disagree, and I don't mind if ArmVirt is handled
> differently. It's just that we should have handled this more uniformly,
> I believe.
>

Yes - as I replied to Leif, I am not going to obsess about this. But
the point of stable tags is not to rush things in at the last minute.

>
> In retrospect, I would have also appreciated if the patches had
> referenced <https://bugzilla.tianocore.org/show_bug.cgi?id=1278>, even
> though they only implement "prep" work for now, on the platform DSC/FDF
> level, and not the actual driver removal.
>
> For example, the important explanation about MdeModulePkg's iSCSI driver
> implementing its own MD5 algo cannot be connected to the OVMF commit now
> (d2f1f6423bd1). I have copied the most relevant passage from the cover
> letter of this series into TianoCore BZ#1278, but the commit in question
> doesn't reference any BZ, so the link cannot be established.
>
> Thanks
> Laszlo