Re: [RFC NO MERGE] f2fs: extent cache: support unaligned extent

[Chao Yu]

On 2021/8/3 9:25, Jaegeuk Kim wrote:
> On 08/03, Chao Yu wrote:
> > On 2021/8/3 2:03, Jaegeuk Kim wrote:
> > > On 08/01, Chao Yu wrote:
> > > > On 2021/7/31 3:20, Jaegeuk Kim wrote:
> > > > > On 07/07, Chao Yu wrote:
> > > > > > Currently, it only works in readonly format f2fs image.
> > > >
> > > > There wouldn't be any race condition because unaligned extent only works
> > > > for ro feature now?
> > >
> > > Isn't your patch proposing on writable partition?
> >
> > Please check description in patch message, now it was designed only for
>                                                   --
>                                                   what do you refer "it"?
>
> > compression case w/ ro feature, let's check and support rw partition later.
>
> Quite confused the patch description and code changes as well. You added some
> change with this as well which is for RW.
>
> +       if (is_inode_flag_set(dn->inode, FI_COMPRESSED_FILE) &&
> +                       !f2fs_sb_has_readonly(sbi)) {

My bad, I've updated in my dev branch, but forgot to resend it...:

https://git.kernel.org/pub/scm/linux/kernel/git/chao/linux.git/commit/?h=dev&id=c3a40f6a186ba064f95432b308173d0a8fe375dc

Thanks,

> > Thanks,
> >
> > > > > Is there any race condition between the address in extent_cache and the one in
> > > > > dnode? I feel that we could synchronize it by locking its dnode block.
> > > >
> > > > Thanks,