[LU-2800] build: clean out old autoconf options

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.6.0, Lustre 2.4.2
• Labels: patch
• Severity: 3
• Assignee: Bob Glossman (Inactive)
• Reporter: Jeff Mahoney (Inactive)

Description

It currently takes longer to run configure on the Lustre source than it does to build it. There are a ton of configure tests that can be removed in light of the support matrix for 2.4+

Issue Links

• is related to: LU-3080 Clean up code for ancient kernels <2.6.32

Comments

[James A Simmons]

The follow patches are the next series to merge. They will not conflict with master or each other. Please review if you haven't already.

http://review.whamcloud.com/#/c/5338
http://review.whamcloud.com/#/c/5340
http://review.whamcloud.com/#/c/5347
http://review.whamcloud.com/#/c/5359
http://review.whamcloud.com/#/c/5369
http://review.whamcloud.com/#/c/5411

[James A Simmons]

The follow matches make the next series to merge. They will not conflict with master as it stands or each other.

http://review.whamcloud.com/#/c/5336
http://review.whamcloud.com/#/c/5340
http://review.whamcloud.com/#/c/5343
http://review.whamcloud.com/#/c/5345
http://review.whamcloud.com/#/c/5347
http://review.whamcloud.com/#/c/5353
http://review.whamcloud.com/#/c/5384
http://review.whamcloud.com/#/c/5393
http://review.whamcloud.com/#/c/5411

[James A Simmons]

Luckily most of the smaller patches will not collide often with other non LU-2800 patches. The the big is http://review.whamcloud.com/#/c/5505 which can easily end up in the rebase treadmill.

[Prakash Surya (Inactive)]

Because of the "painful resynchronizing" process, it would be great if these can get some attention from Oleg before they become stale again. I'd hate to see cleanup like this die due to lack of attention.

[James A Simmons]

Finish the first round of the painful resynchronizing of the patches. All the patch for this ticket we be viewed with

http://review.whamcloud.com/#/q/status:open+message:LU-2800,n,z

Please be aware all patches are against master so if you attempted to apply them all some patches will collide. Having dependencies between patches delays the landing process and we don't know which patches will be merged first.

[Jeff Mahoney (Inactive)]

If you mean has it been refreshed, I have a refreshed version I can push. I've been holding off since Chris was talking about "false positives" but he hasn't commented since.

[Bob Glossman (Inactive)]

Alexey, Not sure I understand the question. Can you just look at the git history on ldiskfs/kernel_patches or lustre/kernel_patches and find what you want to know?

[Alexey Shvetsov (Inactive)]

Is there list of updated patches?

[Minh Diep]

Hi Jeff,

Did you try to compare the time it saved after the clean up?

[Jeff Mahoney (Inactive)]

No. I combined all the ones that only removed code. I marked the ones that I merged into 5505 as abandoned already. There are still a bunch more that do either slight or more involved modifications.

[Bob Glossman (Inactive)]

Jeff, Am I correct in thinking the myriad of previous autoconf patches are now combined into http://review.whamcloud.com/#change,5546 and http://review.whamcloud.com/#change,5505 making all the earlier ones obsolete?