Re: [PATCH 1/8] writeback: get rid of generic_sync_sb_inodes() export

[Jens Axboe <jens.axboe@oracle.com>]

On Tue, Sep 08 2009, Artem Bityutskiy wrote:
> On 09/08/2009 01:57 PM, Jens Axboe wrote:
>>> But well, if you do not want to carry my patch, then I'll have to
>>> re-base my tree later, fix stuff, and send a pull request. I mean,
>>> your stuff will for sure be merged first, because I send pull requests
>>> late, just because UBIFS is a minor thing in the kernel.
>>
>> You don't have to rebase, if my work is merged first then you just merge
>> Linus' tree into yours and fixup the conflict before asking Linus to
>> pull.
>
> I thought Linus asked to avoid merge commits in pull requests at some
> point, no? So I thought that I'd re-base then, which Linus also
> dislikes :-)

Pointless merges are discouraged, but this one isn't pointless since it
resolves a conflict. If you rebase, Linus will flame your ass to a
crisp, I know from personal experience :-)

>> It's a trivial conflict, I don't understand what the fuzz is about.
>
> Well, I was just thinking how to avoid merge commits. But I guess I can
> do what you suggest.

Just don't worry about it, things are fine as-is. When the conflict
happens with the mainline tree, fix it up. If it was a more involved
depdency chain, we could do something more about it. But for this, I'd
say it's a lot more work to attempt to "fix" something that is a 10s
merge issue at pull time.

-- 
Jens Axboe