Re: Fastmap update v2 (pile 1)

[Richard Weinberger <richard@nod.at>]

Am 09.01.2015 um 23:09 schrieb Ezequiel Garcia:
>> All of Pile1, 2, 3 and 4. :-)
>> One or tow patches are preparations for the real fix but obviously you'll need them
>> too.
>> The rest are enhancements and cleanups.
> 
> What do you mean by "the rest"?

Pile 5, 6, and 7.

>> As I wrote before I've structured the patch set in a way to make backporting easy.
>>
>>> For bugfixes, having a detailed explanation of the problem the commit is
>>> meant to fix would be better as well.
>>
>> Okay, I'll add the horror stories to these patches.
>>
> 
> I know it's a real pain, but if you can add a Fixes: tag, it would
> certainly help Artem track down the bug. The good thing is that you get
> the -stable hassle for free.

We cannot tag these as stable, first I have to inject old fastamp fixes
into -stable.
Two years ago Artem and I decided that fastmap as experimental feature does
not need -stable backports. It turned out that this was horrible wrong
and stupid.

>>> This patchset seems to have stalled, so perhaps having this information
>>> would help Artem to pick the ones that you point as fixes, before we
>>> miss another cycle.
>>
>> The question is, shall I wait for Artem or resend again?
> 
> Hm, well, given we are just a handful of developers, and we are all time
> constrained, maybe we could focus on the first two piles for now:
> 
> Pile 1, November 24, https://lkml.org/lkml/2014/11/24/324
> Pile 2, November 30, https://lkml.org/lkml/2014/11/30/50

Hmm, I'm not sure whether it is a good idea to resend Pile1 and Pile2.
I've currently around 50 patches on linux-mtd floating around (nett, without resends).
I fear it will just increase the mess we already have.

I really would like to hear what Artems plans are.
Actually I resent and split up the first series upon his request.

Thanks,
//richard