2008-02-05 22:38:44

by Stephen Rothwell

[permalink] [raw]
Subject: Ordering of commits and breaking bisecting

Hi all,

Can we be a bit more careful with the ordering of commits, please?
Commit d22a6966b8029913fac37d078ab2403898d94c63 (iommu sg merging: add
accessors for segment_boundary_mask in device_dma_parameters())
introduces the dma_get_seg_boundary accessor which is used in the
*preceding* commits fb3475e9b6bfa666107512fbd6006c26014f04b8 (for
powerpc), and 1b39b077789955c8389488d53d075518fdcd582e and
fde9a1094ddf2892188a8a0eccda527de47cba8e (for x86).

Thus we have another set of commits that will break bisection. This one
was easily avoidable.

A little more care is needed.
--
Cheers,
Stephen Rothwell [email protected]
http://www.canb.auug.org.au/~sfr/

Attachments:
(No filename) (684.00 B)
 (No filename) (189.00 B)
 Download all attachments

2008-02-06 23:22:53

by FUJITA Tomonori

[permalink] [raw]
Subject: Re: Ordering of commits and breaking bisecting

On Wed, 6 Feb 2008 09:38:17 +1100
Stephen Rothwell <[email protected]> wrote:

> Hi all,
>
> Can we be a bit more careful with the ordering of commits, please?
> Commit d22a6966b8029913fac37d078ab2403898d94c63 (iommu sg merging: add
> accessors for segment_boundary_mask in device_dma_parameters())
> introduces the dma_get_seg_boundary accessor which is used in the
> *preceding* commits fb3475e9b6bfa666107512fbd6006c26014f04b8 (for
> powerpc), and 1b39b077789955c8389488d53d075518fdcd582e and
> fde9a1094ddf2892188a8a0eccda527de47cba8e (for x86).
>
> Thus we have another set of commits that will break bisection. This one
> was easily avoidable.
>
> A little more care is needed.

Sorry about it.

I submitted the patches in the proper order but the order seemed to be
reversed in the -mm.

http://www.mail-archive.com/[email protected]/msg11919.html

I could have noticed the issue since it has been in the -mm for a long
time, sorry.