Re: [f2fs-dev] [PATCH v3] f2fs: fix to keep isolation of atomic write

2021-04-12 Thread Chao Yu
On 2021/4/13 11:27, Jaegeuk Kim wrote: On 04/12, Chao Yu wrote: As Yi Chen reported, there is a potential race case described as below: Thread AThread B - f2fs_ioc_start_atomic_write - mkwrite - set_page_di

Re: [f2fs-dev] [PATCH v3] f2fs: fix to keep isolation of atomic write

2021-04-12 Thread Jaegeuk Kim
On 04/12, Chao Yu wrote: > As Yi Chen reported, there is a potential race case described as below: > > Thread A Thread B > - f2fs_ioc_start_atomic_write > - mkwrite >- set_page_dirty >

Re: [f2fs-dev] [PATCH v2] f2fs: fix to avoid touching checkpointed data in get_victim()

2021-04-12 Thread Chao Yu
On 2021/4/13 10:59, Jaegeuk Kim wrote: @@ -2701,7 +2689,7 @@ static void change_curseg(struct f2fs_sb_info *sbi, int type, bool flush) reset_curseg(sbi, type, 1); curseg->alloc_type = SSR; - __next_free_blkoff(sbi, curseg, 0); + __next_free_blkoff(sbi, curseg->segno,

Re: [f2fs-dev] [PATCH v2] f2fs: fix to avoid touching checkpointed data in get_victim()

2021-04-12 Thread Chao Yu
On 2021/4/13 10:59, Jaegeuk Kim wrote: On 04/11, Chao Yu wrote: Hi Jaegeuk, Could you please help to merge below cleanup diff into original patch? or merge this separately if it is too late since it is near rc7. I didn't review this tho, this gives an error in xfstests/083. My bad, I hit th

Re: [f2fs-dev] [PATCH v2] f2fs: fix to avoid touching checkpointed data in get_victim()

2021-04-12 Thread Jaegeuk Kim
On 04/11, Chao Yu wrote: > Hi Jaegeuk, > > Could you please help to merge below cleanup diff into original patch? > or merge this separately if it is too late since it is near rc7. I didn't review this tho, this gives an error in xfstests/083. > > From 5a342a8f332a1b3281ec0e2b4d41b5287689c8ed M

[f2fs-dev] 答复: [f2fs-devel] Why does f2fs not reduce disk space when compression is on?

2021-04-12 Thread changfengnan
Thanks, I got it. -邮件原件- 发件人: Chao Yu 发送时间: 2021年4月13日 9:16 收件人: Fengnan Chang 抄送: [email protected]; [email protected]; [email protected] 主题: Re: [f2fs-dev] [f2fs-devel] Why does f2fs not reduce disk space when compression is on? On 2021/4/12 21:13, Fengnan Chang w

Re: [f2fs-dev] f2fs.resize makes errors in FS

2021-04-12 Thread Chao Yu
On 2021/4/12 23:40, beroal wrote: Hi. Sorry for the delay. The 1.14.0 version with the patches "fix wrong sit/nat bitmap during rebuild_checkpoint()" and "fix to check free space before shrink" completes my test case successfully. Thanks for the test. ;) Jaegeuk, could you please check these t

Re: [f2fs-dev] [f2fs-devel] Why does f2fs not reduce disk space when compression is on?

2021-04-12 Thread Chao Yu
On 2021/4/12 21:13, Fengnan Chang wrote: I'm testing the f2fs data compression feature. I created a new FS with the flags extra_attr and compression, use chattr -R +c on the test dir to enable compression for the directory. I found that when compression was turned on, the files did not take up a

Re: [f2fs-dev] f2fs.resize makes errors in FS

2021-04-12 Thread beroal
Hi. Sorry for the delay. The 1.14.0 version with the patches "fix wrong sit/nat bitmap during rebuild_checkpoint()" and "fix to check free space before shrink" completes my test case successfully. On 06.04.21 04:44, Chao Yu wrote: Hi beroal, Could you please apply below patch based on last fi

[f2fs-dev] [f2fs-devel] Why does f2fs not reduce disk space when compression is on?

2021-04-12 Thread Fengnan Chang
I'm testing the f2fs data compression feature. I created a new FS with the flags extra_attr and compression, use chattr -R +c on the test dir to enable compression for the directory. I found that when compression was turned on, the files did not take up any less space. Using trace, I see that th

[f2fs-dev] [PATCH v3] f2fs: fix to keep isolation of atomic write

2021-04-12 Thread Chao Yu
As Yi Chen reported, there is a potential race case described as below: Thread AThread B - f2fs_ioc_start_atomic_write - mkwrite - set_page_dirty - f2fs_set_page_private(page,