Re: [PULL REQUEST] btrfs-progs updates, the main part

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 2019/4/16 上午12:27, David Sterba wrote:
> On Thu, Apr 11, 2019 at 01:54:44PM +0800, Qu Wenruo wrote:
>> Here is my pull request for the btrfs-progs:
>> https://github.com/adam900710/btrfs-progs/tree/for_devel
>>
>> The basis is the following commit from your devel branch:
>> commit a8389337c5a0e7c3fb32ff91d557889c036da73a
>> Author: Adam Borowski <kilobyte@xxxxxxxxxx>
>> Date:   Mon Feb 25 19:16:44 2019 +0100
>>
>>     btrfs-progs: defrag: open files RO on new enough kernels
> 
> Please rebase the branch on top of current devel, with "btrfs-progs:
> Remove get_argv0_buf". I had a few patches merged but the branch was not
> public so there's some duplication.
> 
>> This branch drops the following commit in your devel branch:
>> commit 4e8eb31f048f3fe9a68aacb58e4c0714df25b5c4 (david/devel)
>> Author: Qu Wenruo <wqu@xxxxxxxx>
>> Date:   Fri Sep 14 15:49:44 2018 +0800
>>
>>     btrfs-progs: Do metadata preallocation as long as we're not
>> modifying extent tree
> 
> This fixes issue #123, is there another fix? 

Not yet, the reason is known but new fix is still under development.
Discarding it at least won't cause new problem.


>I'll drop the patch from
> devel as it fails the test so the revert won't be needed.
> 
>> Due to test failure on looping chunk allocation.
>>
>> The branch passes all selftests, except misc/035 which is a known problem.
>>
>> This pull request include the following features:
>>
>> Core change:
>> - check --repair
>>   * Flush/FUA support to avoid breaking metadata CoW
>>       Now btrfs-progs crashing or transaction aborted won't cause
>>       new transid error.
>> - device scan:
>>   * Add the ability to forget a device
>>
>> Fixes and Enhancement:
>> - generic
>>   * Try best copy when reading tree blocks.
>>   * Skip unnecessary retry when one tree block copy fails.
>>   * Avoid back tree block to populate tree block cache.
>> - check
>>   * File extents repair no longer relies data in extent tree.
>>   * New ability to check and repair free space cache invalid inode mode.
>>   * Update backup roots when commit transaction.
>> - mkfs:
>>   * Enlarge metadata chunk size to follow kernel size.
>> - defrag
>>   * open file in RO for newer kernel.
>> - gitignore
>>   * Ignore hidden files.
>> - qgroup report
>>   * Handle unrecognized sort string.
>> - receive
>>   * New quite mode.
>> - Misc
>>   * More cleanup for fs_info <-> root parameters.
> 
> Nice, I'll put it to the merge commit once the branch is ready.
> 
[snip]
>>   btrfs-progs: qgroup show: report errors while parsing sort string
>>   btrfs-progs: qgroup show: report unrecognized format of sort string
>>   btrfs-progs: qgoup: propagate errors from
>>     btrfs_qgroup_parse_sort_string
> 
> For pull-based workflows there are some things that I'd require, based
> on what I do for kernel patches, so let's use this one to set the
> expectations:
> 
> The excerpt above from the shortlog (and other commits) was not intended
> to be pulled, so it should not be in the shortlog.

Got it.

> 
> Please note that all patches need to have signed-off-by by the
> committer. This is not true for

Just one minor question, if I'm the author, no need for another SoB line
right?

> 
> - btrfs-progs: fsck-test: enable lowmem repair for case 001
> - btrfs-progs: tests: add case for inode lose one file extent
> 
> The commits that do revert do not need the 'btrfs-progs:' prefix,
> without it it's obvious in the patch subject list that those are
> exceptions.
> 
> One thing that I edit in almost all commit is the order of tags, briefly
> documented at
> https://btrfs.wiki.kernel.org/index.php/Developer%27s_FAQ#Ordering

Oh, this is indeed different from my understand.

Especially for the reviewed-by and tested-by tags.

I'll definitely reorder those in next update.

Thanks,
Qu
> 
> I don't require a signed tag to pull as the amount of pull requests is
> not going to be high, review is feasible for me.
> 
> One thing that might be new for you is that you'll be the first get
> blamed for everything what's in the branch you send to pull, then the
> original authors :) It's good to favor quality over quantity.
> 
> There's probably more of lower importance, we'll find out along the way.
> 

Attachment: signature.asc
Description: OpenPGP digital signature


[Index of Archives]     [Linux Filesystem Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux